Talk:American Journal of Anatomy 11 (1910-11)

From Embryology

THE AMERICAN JOURNAL OF ANATOMY


EDITORIAL BOARD


Charles R. Bardeen

University of Wisconsin

Henry H. Donaldson

The Wistar Institute

Thomas Dwight

Harvard University

Simon H. Gage

Cornell University

G. Carl Huber

The Wistar Institute


George S. PIijntington

Columbia University

Franklin P. Mall

Johns Hoi)kin.s University

J. Playfair McMurrich

University of Toronto

Charles S. Minot

Harvard University

George A. Piersol

University of Pennsylvania


Henry McE. Knower, Secretary

University of Cincinnati


VOLUME 11 1910-1911


THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY

PHILADELPHIA, PA.


^<fb


1


1X2']


COMPOSED AND PRINTED AT THt

WAVERLY PRESS

15y Tin; Williams & Wilkins Company

Baltimcbe, U. S. a.


CONTENTS

11)10-1911


No. 1. NOVEMBER, 1910

William F. Allen. Distribution of the lymphatics in the tail region of Scor psenichthys marmoratus. Twelve figures 1

Leonard W. Williams. The somites of the chick. Nineteen figures 55


No. 2. JANUARY, 1911

J. Gordon Wilson. The nerves and nerve endings in the membrana tympani in man. Three plates 101

Florence R. Sabin. Description of a model showing the tracts of fibres

medullated in a new-born baby's brain. Nine plates 113

W. E. Dandy and Emil Goetsch. The blood supply of the pituitary body.

Four figures 13'

Jeremiah S. Ferguson. The anatomy of the thyroid gland of Elasmobranchs, with remarks upon the hypobranchial circulation in these fishes. Twenty fi2;ures 1^1


No. 3. MARCH, 1911

Franklin P. Mall. On the muscular architecture of the ventricles of the

human heart. Twenty-two figures 21 1

J. F. Gudernatsch. Hermaphroditismus verus in man. Seven figures 267

Albert Kuntz. The development of the sympathetic nervous system in

turtles. Thirteen figures 279


No. 4. MAY, 1911

John U'aukkx. Tlic (icvclopincnt of the paraphysis and pineal region in

Kept ilia. Thirteen plates 313

John Lewis Bremer. Morphology of the tubules of the human testis and

epididymis. Twelve figures 393

Charles Russell B.\hdeen. Further studies on the variation in susceptibility of amphibian ova to the X-rays at different stages of development. Eighteen figures 419


DISTRIBUTION OF THE LYMPHATICS IN THE TAIL REGION OF SCORP.ENICHTHYS MARMORATUS

WILLIAM F. ALLEN

From the Herzstein Laboratory of the Universilij of California, New Monterey,

California

TWELVE FIGURES

INTRODUCTION

This paper is a continuation of an earlier study. Since Scorpsenichthys differs notably in several important details from any of the many forms described by Favaro in his most comprehensive work it seems desirable to complete the distribution of this system of vessels in this specialized species.

Material and method of procedure. — Scorpsenichthys marmoratus, on which all the dissections were made, is one of the common rockfish found all along the Pacific coast. The tails of Scorpsenichthys which were to be injected were severed a little anterior of the caudal peduncle, and were so arranged in a pan that the cut end was considerably higher than the tail end. The caudal artery was then injected with a carmine gelatin mass, after which the subcutaneous or lymphatic canals were filled with a Berlin blue gelatin mass or in one or two cases, with India ink, from the longitudinal neural trunk. In order to somewhat check up the work microscopically, the small tide pool cottid, Clinocottus anahs, a closely related species, was sectioned. As with Scorpansenichthys the tails of full grown adults were severed transversely a little anterior of the caudal peduncle; when they were killed, fixed, and macerated in Tellyesniczky's potassium bichromate-acetic mixture for a period of about two weeks; after which they were embedded in paraffin, cut 10 microns, stained in Heidenhain's

THE AMERICAN JOURNAL OF ANATOMY, VOL. 11, NO. 1.


2 WILLIAM V. ALLEN

iron hcTiiiatoxyliii, and counter stained in a concentrated alcoholic solution of orange G plus a little acid fuchsin. Some microscopic observations were also made on the tail of a living Phanerodon atripes (viviparous perch) embryo. This study was made at the Herzstein IVIarine Laboratory of the University of California, New Monterey, California.

Literature. — In addition to the bibliographical lists given in my previous papers several important works have come to my notice; among which Favaro's monograph is of special interest.

In this the author gives a detailed description of the distribution of the subcutaneous vessels in the tail region of a great number of species from Petromyzon to the most specialized of the Teleostomi. Furthermore he gives the development of these vessels in Acanthias vulgaris, Torpedo ocellata, Belone acus, and Squalius cavedanus. In the Cyclostomes, Selachians, and Acipenser sturio they are portrayed as veins; while in the Teleosts they are represented as lymphatics, Favaro considering the lymphatic system of the Teleosts as phylogenetically derived from the corresponding subcutaneous veins of the lower orders of fishes. In fig. 1 56 the gradual evolution of the vasa intermedia of the Selachians to the longitudinal haemal lymphatic trunk of the Teleosts is graphically shown. With Belone acus, Favaro finds an embryonic condition in the haemal canal comparable to the vasa intermedia of the Selachians. The following is a translation of the last paragraph of Favaro's paper, which is a concise summary of the authors conclusions. Beyond (cephalad) the heart, likewise certain lymphatics, as for example the haemal, and indirectly to a certain extent the others, are derived from the embryonal venous system so that it is possible in fishes to recognize the close relationship, not only phylogenetic, but ontogenetic as well, between the lymphatic and venous systems."

If frecjuent reference to this most excellent work was not to be made later, much more would be quoted here.

By a most unfortunate circumstance no note of Favaro's monograph was made in my last paper, although it appeared some time prior to may publication. It is true I saw notice of Favaro's work in the Bibliographia Zoologica, and i^laced an order for it


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS 6

with an European firm. It did not, however, arrive until after my paper appeared, and I was unable to gain access to another copy.

A year earlier Favaro published a most interesting paper on the caudal heart of the eel, and as was the case with the previous mentioned paper I was unable to gain access to this until quite recently. The author finds the caudal heart of the eel to be a lymphatic heart situated at the posterior end of the last vertebra and to consist of two cavities. As I had long suspected the first, the atrium cordis caudalis is said to be in communication anteriorly with the longitudinal haemal trunk, tronco linfatico subvertebrale, and posteriorly with the trunk from the tail. Both openings are guarded by valves opening into the atrium. The atrium is connected mesad with the second cavity, ventriculus cordis caudalis, the orifice being guarded by valves opening into the ventricle, and anteriorly the ventricle empties into the caudal vein, the orifice, likewise having valves opening into the vein.

Kellicott in his monograph on the vascular system of Ceratodus does not allude to the lymphatics or subcutaneous system further than to say that a pair of well developed lateral cutaneous veins are found beneath the skin at a level with the lateral line. Posteriorly they are said to anastomose with the caudal vein and anteriorly they open into the subscapular veins. Kellicott likens these vems throughout to the lateral cutaneous veins of Mustelus as described by Parker.

The recent studies of Sabin, Lewis, Huntington and McClure, Hoyer, Knower, Baetjer, Heuer and Clark on the ontogeny of the lymphatic system concede that the primary or deep seated lymphatics arise as sacs or hearts from transformed veins, and that the superficial or secondary lymphatic system originates from an endothelial sprouting from these sacs or hearts. Sala while admitting the derivation of the lymphatic hearts or sacs in the chick from the veins still holds to the old view that the ducts are formed from the mesenchyme cells. Marcus states that the segmental lymphatic hearts in the snake-like Amphibian, Hypogeophis, are formed from the coelom or body cavity epithelium and not from veins.


4 WILLIAM F. ALLEN

Concerning the phylogeny of the lymphatics, Favaro and I have shown that considerable anatomical data supports the hypothesis that the lymphatic system of the higher or more specialized orders of fishes have their homologuo in veins in the lower or more generalized orders of fishes.

BLOOD-VASCULAR SUPPLY FOR THE TAIL-REGION

To a\'oid confusion it may be well to consider first the distribution of the blood vessels before taking up the lymphatics.

Caudal artery. — As in other fishes the caudal artery in Scorpa^nichthys (figs. 4, 5, 7-10, C.A.) traverses the haemal canal directly below the centra. This trunk was also seen in nearly all the transverse sections of Clinocottus. In these sections it should not, however, be confused wdth the minor caudal artery with which it runs parallel and is a branch. Beneath the last vertebra the caudal artery separates into a major and a minor fork. Sometimes the main stem is the left fork, but more often it is the right. The minor stem (figs. 9 and 10, R.C.A.) supplies the musculature of the side of the fin; while the major stem or caudal artery proper (figs. 4-8, 10, and 11, C.A.) continues caudad in the space between the two hypural bones to the posterior ends of these bones, where it bifurcates to form a dorsal and a ventral caudal fin artery. Immediately behind the last vertebra the posterior neural artery (figs. 6 and 7, P.Neu.A.) is given off from the major stem of the caudal artery to pass dorsad in a median line, a little in front of the superior hypural bone. In the specimen from which fig. 7 was drawn it passed across the base of the left side of the superior hypural and along the left side of the last interspinal bones to break up in a deep network in the connective tissue covering these bones, but in most of the other dissections it had a similar course on the opposite or right side. A little caudad of the last vertebra a pair of hypural arteries (figs. 4-7, Hyp. A.) are sent off to either side of the superior hypural bone, which they cross obliquely and break up in a capillary net work on the posterior and dorsal surfaces of the bone. These arteries may have arisen from the fusion of several of the embry


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS 5

onic neural arteries and if so should be considered as the posterior neural arteries instead of the hypural arteries. In the Clinocottus series the posterior neural artery was noticed, but the hypurals were not observed.

The course of the caudal fin arteries (figs. 4-6, and 8, C.F.A.' and C.F.A.") is either dorsad or ventrad in the basal canal of the caudal fin.^ In the transverse series of Clinocottus the division of the caudal artery into the caudal fin arteries was clearly seen, and in fig. 11 the forking of the caudal artery in a 30 mm. Phane rodon embryo is shown. In Scorpaenichthys, Clinocottus, and Phanerodon a small branch, the caudal fin ray artery (figs. 4, 8, and 11, C.R.A.) is given off to the center of each ray. In Scorpaenichthys and Clinocottus after continuing caudad a short distance in the center of a ray this artery forks, one branch passing along the dorsal surface of the ray and the other along the ventral surface, each giving off a network, to the ray and to the fin ray membrane. In the 30 mm. Phanerodon embryo (fig. 11) the first ventral caudal ray artery forked as they all do in Scorpaenichthys, but the remaining ones did not bifurcate within the rays; those from the dorsal half of the fin traverse the dorsal surfaces of the rays, and those from the ventral half of the fin, excepting the first, pass along the ventral surfaces of the rays. In the living Phanerodon embryo the red corpuscles could be clearly seen leaving these arteries to enter a network of capillaries in the fin njembrane, and become collected on the opposite side by the caudal ray vein.

Minor caudal artery. (Figs. 5-8, 10, and 12, C.A.d)). — In the caudal peduncle region of Scorpaenichthys and in all sections of Clinocottus this vessel was found running parallel with the caudal artery, sometimes lying to the side, and again below it. At frequent intervals this artery gives off branches (fig. 10, C.A' .(d), which cross the lower surface of the caudal artery. Often these branches have as great a caliber as the main stem, and so far as

1 The basal canal of the caudal fin is a canal formed at the base of the caudal fin at the point where the two halves of the caudal rays separate to become attached to the hypural bones. This canal would therefore pass dorso-ventrad through the proximal ends of the caudal rays.


(•) WILLIAM F. ALLEN

could l)e ascertained they were destined to supply the blood vessels of the haMnal canal. For their branches were observed going to and breaking up on the surfaces of the caudal and intersegmental arteries. In the dissection from which fig. 10 was drawn the main stem of the minor caudal artery ran along the right side of the caudal artery to the fifth vertebra from the last, when it crossed to the opposite side and continued caudad on the left side of the caudal artery to the end of the last vertebra, where it bent dorsad with the caudal artery to the interval between the two hypural bones. In the single specimen in which the origin of the minor caudal artery was traced, it was found to branch off from the lef c side of the dorsal aorta a few millimeters cephalad of the posterior end of the kidney. The main stem passed caudad along the left side of the aorta from which a short branch was given off cephalad. Shortly after leaving the body cavity the minor caudal artery bends to the ventral surface of the caudal artery where it separates into two forks of about equal size, the main stem traveling caudad on the right side of the caudal artery and the minor stem on the left. Beside the capillary branching to the blood vessels, as mentioned above, there are frequent cross branches between these two stems of the minor caudal artery, producing a ladder-like appearance. In some dissections of Scorpsenichtys, as in fig. 5, the minor caudal artery in traversing the hypural interval was situated ventrad of the caudal artery, but in other dissections of Scorpaenichthys and in the Clinocottus series, as shown in figs. 7 and 8, the minor caudal artery traveled for the most part dorsad of the caudal artery. In both Scorpaenichthys and Clinocottus the minor caudal artery divides into a dorsal and a ventral minor caudal fin artery (fig., 8 C.F.A.'n) and C.F.A."{x)). As a rule these branches traverse the basal canal of the fin with, but cfeudad of, the corresponding caudal fin arteries, and like them, send off a branch, the minor caudal ray artery (fig. 8, C.R.A.d)). These minor branches always fork before the caudal ray arteries do, and like them follow the dorsal and .ventral surfaces of each ray, but so far as could be determined they did not extend caudad of the intrinsic muscles of the caudal fin. They appear, however, to furnish the principle supply for these muscles; while the cau


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS


dal fin ray arteries supply the fin rays and the membrane connecting them.

Favaro describes the minor caudal artery in numerous species as the arteria o arteriae longitudinales vasorum intermediorum. His fig. 156 graphically shows the evolution of this vessel. In Squalus it is represented simply as branches from the segmental arteries, which anastomose with the vasa intermedia. In Raja they also have their origin from branches of the segmental arteries, but run caudad in the haemal canal for a short distance before anastomosing w^ith the vasa intermedia. In Acipenser they likewise arise from the segmental arteries and anastomose with the vasa intermedia, and moreover by continuing in the haemal canal to anastomose with the successive segmental arteries they form a continuous trunk which runs parallel with the caudal artery and the vasa intermedia, and at regular intervals, between the intersegmental arteries it sends off anastomosing branches to the vasa intermedia ; while in the Teleosts the arrangement is identical to Acipenser, except that the direct connections of the vasa intermedia or longitudinal haemal lymphatic trunk with the caudal vein are lost.

As stated above in Scorpaenichthys no connections with the intersegmental arteries were observed; hence the minor caudal artery in this species is still further differentiated ; for it is simply a vessel arising from the aorta and passing caudad with it to the tail.

Caudal vein (figs. 3-6, 9 and 12, C.V.). In Scorpaenichthys and Clinocottus, as in other fishes, this vein traverses the haemal canal immediately below the caudal and the minor caudal arteries. When distended its caliber is much greater than the artery, while its walls are much thinner. Unlike Lepisosteus (p. 51) it does not expand under the last vertebra in a sinus (Sinus venous caudalis of Favaro). As represented in Ophiodon elongatus (p. 107) the caudal vein of Scorpaenichthys and Clinocottus bifurcates below the last vertebra into a right and a left branch, which are of different lengths and different relative importance. Sometimes the main stem is the right branch, but more often it is the left; while in one instance they were of equal lengths and importance.


8 WILLIAM F. ALLEN

The iniiioi- fork of the caudal vein (figs. 3, 6, and 9, R.C.V.) curves u]) around the right posterior side of the last vertebra in front of the main stem of the caudal artery. When the median line is reached, it usually divides; one branch passes laterad to the periphery, and the other goes caudad a short distance between the superficial and deep muscles of the caudal fin. In one dissection the minor fork of the caudal vein equaled the major fork in length and in caliber. It continued caudad to the caudal fin, where it })enetrated the ventral basal canal of the caudal fin, and collected the blood from that half of the fin; while the blood from the dorsal half of the fin was collected by the other or main fork of the caudal vein.

The main fork of the caudal vein or the caudal vein proper (figs. 4-5, C.V., and figs. 3 and 9, L.C.V.) passes up the opposite and usually the left posterior side of the last vertebra to the median line; where it receives the posterior neural vein (figs. 4 and 5, P.Neu.V.), which in curving around to the dorsal side of the superior hypural bone crosses directly mesad of the posterior portion of the lateral trunk, to eventually follow up the left side of the last neural spines. Meanwhile the main stem continues caudad between the fascia or superficial muscles and the deep muscles of the caudal fin. In its course between these muscles it runs parallel with and mesad to the caudal fin nerve, and receives branches from these muscles. At the base of the tail this vein takes a position quite close to the periphery, in fact, it travels for some little distance directly mesad of the posterior portion of the lateral lymphatic trunk, but never communicates with the same. It then bends mesad to enter the basal canal of the caudal fin, where it soon separates into the dorsal and ventrar caudal fin veins (figs. 4-6, and 11, C.F.V/ and C.F.V."). In Scorpsenichthys the bifurcation of the main trunk of the caudal vein occurs usually anterior to that of the caudal artery, nerve, or the caudal lymphatic trunk, and for the most part the caudal fin veins lie anterior to the other vessels in the basal canal of the fin. In the Phancrodon embryo (fig. 11) the caudal artery forked at the base of the tail before the caudal vein did, and the caudal fin veins were posterior in position to the caudal fin arteries in the


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS 9

basal canal of the fin; while in the Clinocottus series the order of bifurcation was: first the minor acudal artery, then the caudal artery, the caudal vein, and finally the caudal lymphatic trunk. With Scorpaenichthys both caudal fin veins receive numerous caudal ray veins (fig. 4, C.R.V.) from the caudal rays. Their arrangement, however, is not so regular as the caudal ray arteries. For example, the first of the dorsal caudal ray veins (see fig. 4) in addition to receiving a branch from the dorsal and ventral surface of the first dorsal caudal ray collects a third branch which traverses the ventral surface of the second dorsal caudal ray; while the second dorsal caudal ray vein takes its source solely from one stem, which passes along the dorsal surface of the second dorsal caudal ray vein. These veins collect the capillary networks from the fin membranes and from the rays themselves. The veins arising from the intrinsic muscles of the caudal fin empty separately into both lateral sides of the caudal fin veins.

In the Clinocottus series I was unable to trace the main caudal vein much beyond its branching in the basal canal of the tail, and no caudal ray veins were seen unless the caudal ray canals (fig. 3, C.R.T.) function both for lymphatics and veins. Possibly the injection method would have revealed caudal ray veins emptying into the caudal fin veins. Since, the caudal ray arteries and canals are distinctly visible in all the sections through the rays, one could hardly claim the section method faulty for not showing these vessels.

As stated above, the caudal fin veins in a 30 mm. Phanerodon lie posterior to the caudal fin arteries in traveling through the basal canal of the tail. Only one caudal ray vein (fig. 11, C.R. V.) was observed coming from a caudal ray. Those from the dorsal half of the fin ran along the ventral side of the rays; while those from the ventral half of the fin followed along the dorsal surface of the rays. With the exception of the first ventral caudal ray vein, none of these vessels forked ; this vein, however, divided, one branch came from the dorsal side of the second ventral caudal ray, and the other from the dorsal surface of the third ventral ray. The vascular suj^ply for the caudal fin of a young Phanerodon embryo is therefore very simple. An artery traverses one side


10 WILLIAM F. ALLEN

of a ray, while a vein follows along the opposite side of the adjacent ray. In the membrane connecting these two rays there is a network of capillaries, through which one could readily trace corpuscles going from the caudal ray arteries to the caudal ray

veins.

Vogt in Salmo, Hyrtl in Esox ( = Lucius) and Leuciscus, Emery in Fierasfer, Parker in Mustelus, Hopkins in Amia ( =Amiatus), and Vogt and Yung in Perca did not trace the caudal vein further caudal than the last vertebra, that is to the caudal sinus. Jones . (p. 676) describes the great vein in the tail of the eel as being formed from two branches, a larger and a smaller. The larger stem is said to receive the venous radicals from the terminal parts of the tail; while the smaller stem collects the venous radicals from the dorsal part of the tail, and near the junction with the former it receives the caudal heart. Sappey (p. 46; pi. xi, fig. 6 and pi. xii, fig. 3) finds that the caudal vein in the carp and pike arises from a dorsal or superior and a ventral or inferior branch in the base of the fin. At the level of the last vertebra they unite to again divide into a right and left branches, which reunite at the end of the last vertebra in forming the caudal vein. Shortly before anastomosing, each of the above branches are said to receive a papilla from the lateral lymphatic trunk. According to McKenzie the venous system of Amiurus catus takes its origin in the tail from two vessels of unequal size. Silvester (p. 109) describes and figures the caudal vein in the tile-fish as having its source from two branches from the caudal fin. Favaro in numerous Teleosts notes practically the same arrangement as described above for Scorpa^nichthys. In most species he represents the caudal vein as beginning as a sinus, sinus venosus caudalis, under the last vertera, which not only receives the vein from the tail, but also the caudal lymphatic sinus, ventriculus cordis caudalis.

Intersegmental or intercostal vessels. (Figs. 4 and 4a, Neu.A., Neu.V.yHcB.A., and //ce.F.; figs. 9 and 10, Neu. andL.A.,Neu.and L.V .,L.A.,L.V .,H(]e.A.,2iTid Hce.V.). — Apparently these blood vessels are practically the same in all species of Ganoids and Teleosts. They are destined to supply the body musculature, the vertebral


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS 11


column, the myelon, all of the fins except the caudal, and the connective tissue surrounding the vertebral column, the body muscles, and the periphery. Ordinarily, as shown in fig. 9, below each vertebra, two lateral branches and one ventral branch are given off or received by the longitudinal blood vessels of the haemal canal. One of the lateral branches is a vein and the other is an artery; while half of the ventral branches are arteries and half are veins, there being an artery for each alternate vertebra and a vein for the intermediate alternate vertebrae. Upon examining fig. 9 from cephalad to caudad it will be seen that the arteries and veins alternately shift from one side to the other. For example, under vertebra numbered 14 the artery passes to the right side and the vein to the left; while under vertebra numbered 13 the artery goes to the left and the vein to the right. Furthermore, all of the arteries which pass in a lateral direction, upon leaving the haemal canal curve around to the side of the centra; where one-half of them, the lateral arteries (fig. 9, L.A.), follow the intermuscular septa laterad to the periphery, and the other half (fig. 9, Neu. and L.A.) bifurcate; one branch, the lateral arteries (figs. 4 and 4a, L.A.), follow the intermuscular septa laterad to the periphery, and the other branch, the neural arteries (figs. 4 and 4a, Neu.A.) pass dorsad along the neural spines to the periphery. The same correlation occurs with the veins that pasg in a lateral direction ; one-half of them (fig. 9, L.V.) to trace backward, simply pass laterad to the periphery; while the other half, (fig. 9, Neu. and L.V.) to trace backward, bifurcate, one branch (fig. 4a, L.V.) going lateral to the periphery, and the other (fig 4a, Neu. V.) dorsad to the periphery. The arrangement of the ventral branches of the haemal trunks is less complicated. Below one centrum an artery (figs. 9 and 4a, Hce.A.) passes ventrad along its haemal spine to the periphery, and from under the next vertebra to trace backward a vein (figs 9 and 4a, Hce. V.), passes ventrad along the haemal spine to the periphery, and so on, the veins alternating with the arteries.

The outcome of this complex arrangement is, that each alternate dorsal intermuscular septum receives a neural artery, and from every intermedian dorsal intermuscular septa there comes


12 WILLIAM F. ALLEN

a neural vein. Likewise each alternate ventral intermuscular septum receives a haemal artery, and from every intermediate ventral sejita there comes a haemal vein. Furthermore each alternate lateral intermuscular septum receives a lateral artery, and from every intermedian septa there comes a lateral vein. A glance at fig. 9 demonstrates that one intermuscular septum does not, however, possess a neural, haemal, and two lateral arteries, but, on the contrary, that septum may receive, as shown opposite vertebra numbered 9 in fig. 9, a neural, a lateral, and a haemal vein and a lateral artery ; while the intermuscular septum opposite vertebra numbered 8 in the same figure has a neural, a lateral and a haemal artery and a lateral vein. Occasionally, however, some irregularities may occur, as for example, opposite vertebra numbered 10 in fig. 9 and under the posterior vertebrae.

Of the two classes of arteries arising from the side of the caudal artery, those designated as lateral arteries (figs. 9 and 10, L.A.; indicated but not lettered in fig. 4), after leaving the haemal canal, curve around to the side of the centra, where each bends at right angles to pass laterad in the intermuscular septum. When about half way between the vertebra and the skin it divides into a dorsal and a ventral branch. In their course these branches supply the myotome in front and the one behind, and upon arriving at the surface, above and below the lateral lymphatic trunk they usually follow dorsad or ventrad a short distance on the surface of the intermuscular septum to give off branches to the connective tissue between the skin and the muscles.

The other lateral branches of the caudal artery (figs. 4 and 4a, Neu.A.,; figs. 9 and 10, Neu. and L.A.) are larger and more important vessels. Like the lateral arteries described above they curve around to the side of the centra ; where each separates into a lateral and a neural artery. The former (figs. 4 and 4a, L.A.) is identical to the lateral artery described above, and the latter (figs. 4 and 4a, Neu. A.) curves around to the dorsal side of the vertebra, where it follows up along the anterior surface of the corresponding neural spine. In crossing the vertebra it gives off a myelonal artery, which presumably passes through the spinal nerve formamen, to supply the myelon or the spinal cord; this


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS 13

point, however, was not determined for a certainty. At the apex of the neural spine a dorsal lateral artery (figs. 4a, D.L.A.) is sent off lateral along the intermuscular septum to the periphery, supplying the dorsal portion of the two adjacent myotomes. At this point the neural artery makes a cephalic bend to cross the bases of the depressor and levator muscles of the adjacent dorsal ray. After which it takes another turn to pass dorsad between this levator muscle and the next depressor muscle to the level of extrinsic muscles of the dorsal fin; where it divides into an anterior and a posterior branch. The main stem and these branches supply the intrinsic and extrinsic muscles mentioned above, and also send off two dorsal ray arteries (fig. 4a, D.R.A.), which follow up the posterior side of their respective rays, supplying each and their fin membrane.

The ventral branches of the caudal artery or the haemal arteries (figs. 4, 4a, 9 and 10, Hce. A.) after leaving the haemal canal pursue a similar course ventrad to that of the neural arteries dorsad. At the apex of each haemal spine a ventral lateral artery (figs. 4, 4a, V.L.A.) is given off to the periphery. It passes along the intermuscular septum, between two myotomes, supplying each. The main stem of the haemal artery crosses a depressor and a levator muscle of the adjacent anal ray, and then continues ventrad to the extrinsic muscles of the anal fin between the above levator muscle and the next depressor muscle. Like the corresponding neural trunks upon reaching the extrinsic muscles it separates into an anterior and a posterior branch, which supplies the extrinsic muscles and the levator and two depressor muscles, and sends off two anal ray arteries (figs. 4a, A.R.A.) to the posterior surfaces of two rays, supplying them and thefr fin membrane.

Interseg^nental or intercostal veins. (Figs. 4 and 4:a,Neu. and Hce. V. fig. 9, Neu. and L.V.; L.V., and Hce. V.). — The arrangement and distribution of these veins is identically the same as the corresponding arteries, with this difference that they lie in the intermediate alternate intermuscular septa and terminate in the caudal vein. So that the above description of the distribution of the neural, haemal, lateral, dorsal or ventral lateral, and the dorsal


14 WILLIAM F. ALLEN

or anal ray arteries would apply equally well to the distribution of the corresponding neural, haemal, lateral, dorsal or ventral lateral, and the dorsal or anal ray veins (figs. 4, 4a, 9 and 10, Xeu. v., Hce.V., L.V., D.L.V., V.L.V., and D.R.V.).

In general with the primitive fishes, Cyclostomes and Selachians, but one kind of intersegmental artery or vein is portrayed by Mayer and Favaro. It is represented as forking immediately after leaving the haemal canal into a dorsal or neural and a ventral or hsemal vessel; both of which send out numerous branches to the muscles, and a mesal branch to the spinal cord. In some cases, however, the haemal or ventral v^essels may arise from or empty into the main longitudinal trunks. In the Ganoid Lepisosteus (p. 52) these intersegmental vessels- were found to be practically the same as described above for Scorpaenichthys. With the Teleosts, Vogt, Sappey, Mc.Kenzie, Vogt and Yung, and Favaro found the neural and haemal vessels to be entirely separated, arising or emptying directly into the caudal artery or the caudal vein. They, however, portrayed incorrectly a neural and a hsemal artery and a neural and a hsemal vein for each intermuscular septum. Silvester, however, described the correct relationship of these vessels in the tile-fish.

LYMPHATICS OF THE TAIL

A transverse section, as shown by fig. 12, through the caudal peduncle of Scorpaenichthys or Clinocottus severs six great longitudinal lymphatic canals. Four of which are superficial or subcutaneous, namely, the dorsal, ventral, and lateral lymphatic trunks , and two are deep seated, namely, the longitudinal neural and hsemal lymphatic trunks. These lymphatic trunks collect superficial and deep networks, which are decidedly lymphatic in the character of their meshes, from all tissues that are supplied with blood vessels. So far as could be determined these lymphatic canals had no connection with the blood vessels, either capillary, or direct with the caudal vein.'- Nevertheless, as will be noted in

-' In an etirliui' paper on (lie lynijjliat ics of the head region of 8eorj)ainichthys it was erroneoufily statod that the conihinod trunks formed by the union of the longi


LYMPHATICS IN TAIL REGION", SCORP^NICHTHYS 15

detail later on, the lymphatics came into close touch with the caudal vein, but did not anastomose. Furthermore an injection of the longitudinal neural trunk often fills the caudal vein, and conversely an injection of either the caudal artery or vein often fills the entire lymphatic system. These occurrences I attribute to an extravasation of the injection mass through the thin walls separating these vessels rather than to a direct communication.

SUBCUTANEOUS SYSTEM

Lateral subcutaneous lymphatic trunks. (Figs. 1-2, 4-7, and 12, L.T.). — As in other fishes these two trunks in Ophiodon, Scorpsenichthys, and Clinocottus travel parallel with the lateral line, directly within the skin, in the tough connective tissue sheath that binds the dorsal half of the great lateral muscle to the ventral portion. Its position is clearly portrayed in fig. 12, L. 7". A transverse section of this canal in an adult Scorpaenichthys shows its walls to be composed mostly of fibrous tissue, containing a few smooth muscle fibers, and a peculiar papilla-like endothelial lining. Furthermore in consideration of the enoromous caliber of this trunk it contained relatively very few corpuscles; strange to say the red greatly outnumbered the white, the ratio in the field examined being 26 to 4. The anterior termination of this trunk was given in detail in an earlier paper (p. 51). When about opposite the last vertebra its course becomes deeper (region drawn in outline in figs. 1 and 2), and here it receives or is continuous with a posterior lateral trunk or a posterior portion of the lateral trunk, resulting in the formation of the profundus lateral trunk or the profundus portion of the lateral trunk, which passes mesad to unite with its fellow trunk forming the longitudinal neural lymphatic trunk.

tudinal neural and the lateral lymphatic trunks discharged their contents into the right and left forks of the caudal vein. The cause of this error was due to the fact that before a careful study of the tail region had been made, the posterior neural vein was taken to be a lymphatic vessel, and in crossing under the posterior portion of the lateral lymphatic trunk close to its union with the anterior portion it was thought to anastomose with the former.


l(i WILLIAM F. ALLEN

The posterior jiortion of the lateral trunk (figs. 1-3, 4-7, L.T.ixi) takes its origin from a conspicuous network on the fascia (figs. 1 and 2, F.) and from the intrinsic muscles of the caudal fin. In crossing the fascia it makes a slight ventral bend, and is distinctly a superficial vessel. It receives a prominent dorsal branch and one or two ventral branches. In Ophiodon there is a noticeable fan-shaped network (fig. 2, Net.) between this dorsal branch and the main stem. Also at the base of the tail in Ophiodon there is a mesal connection (fig. 2, L.T.^s)), which joins a swelling of the caudal trunk, designated as the posterior caudal sinus. This communication was clearly shown in an earlier paper (pi. 1, fig. 7). The same relationship is likewise found in Clinocottus (fig. 3, L.T.is)). Nothing of the kind, however, was found at the base of the tail in Scorpaenichthys. In its place a branch of the caudal artery was seen going to the periphery.

What is designated as the profundus or transverse portion of the lateral lymphatic trunk (figs. 4-7, L.T.{2)), namely, the trunk formed by the union of the main stem and the posterior portion of the lateral trunk, passes mesad to the posterior lateral surface of the last centrum, where it bends dorsad in front or behind the posterior neural blood vessels, to anastomose with its fellow trunk on the opposite side, thus forming the source of the great longitudinal neural lymphatic trunk. At some point in its course it receives the posterior neural lymphatic trunk (figs. 4-7, P.Neu.T.), which to trace backward, travels dorsad, parallel with the corresponding blood vessel. On the side of the vertebral column from which a posterior neural vein has its course (figs. 4 and 5) this posterior neural trunk runs along the anterior surface of the vein, and terminates in the corresponding profundus portion of the lateral trunk at the point ^vhere the latter anastomoses with its fellow to form the longitudinal neural lymphatic trunk. On the opposite side of the vertebral column, that is, where the posterior neural trunk follows along a posterior neural artery (figs. 6 and 7), the lymphatic vessel traverses the posterior surface of the artery, and after crossing the artery it culminates in the profundus portion of the lateral trunk at the point of its dorsal bend on the surface of the last centrum. On one side, but


LYMPHATICS IN TAIL REGION, SCORP.^NICHTHYS 17

not usuall}' on both, as shown in figs. 4, 6 and 7, there is a communication between either the posterior neural lymphatic trunk or the profundus portion of the lateral trunk and the caudal lymphatic trunk. In several dissections of Scorpsenichthys and in all the Clinocottus series the profundus portion of the lateral trunk spreads out in a sinus on the side of the last centrum, which for the most part extends cephalad and ventrad. In a few dissections of Scorpsenichthys this sinus extended to the center of the next to the last centrum, where it sent off a lateral communication to the lateral lymphatic trunk. In these dissections of Scorpsenichthys the ventral prolongation of this sinus comes into such close touch with a fork of the caudal vein that it might be taken to anastomose with it, but such is not the case. In the Clinocottus series from which fig. 3 was reconstructed this sinus was seen to pass between the last vertebra and the left caudal vein, being separated from the vein by a single thin layer. In another Clinocottus series in one section, this thin layer was torn in such a manner as to give the appearance of the lymphatic vessel emptying into the vein, and the orifice being guarded by two valves; but in the anterior and posterior sections this layer was complete, there being no evidence of any communication; nor was any to be found in any of the other series, or in any of the dissections of Scorpsenichthys.

As in other fishes the lateral lymphatic trunk collects a dorsal and ventral intermuscular or transverse lymphatic vessel (figs. 1 and 2, Intm.T.) from every intermuscular septum. These vessels are situated superficially directly beneath the ski , in the intermuscular speta, and collect a network from the connective tissue situated between the myotomes and the skin. In the caudal peduncle region they are not continued dorsad or ventrad to anastomose with the dorsal and ventral lymphatic trunks as they do farther cephalad, but anastomose with two vessels which empty into these trunks.

Except for its caudal ending the distribution of the lateral trunk seems to be about the same for all species, but concerning its caudal termination there appears to be considerable difference in various species, and often for the same species as described by different authors.

THE AMERICAN JOURNAL OF ANATOMY. VOL. 11, NO. 1.


18 WILLIAM F. ALLEN

Of the older writers who studied this system in fishes, Vogt in the case of Sahiio is the only author to describe the lateral trunk as extending to the caudal fin. In addition to terminating in a caudal sinus situated at the end of the last vertebra, he finds (pp. 135-6) each trunk extending caudad to the base of the caudal fin, where it separates into a dorsal and a ventral sinus (pi. K; fig. 3, 67), each of which is said to communicate with a corresponding sinus on the opposite side. In the Ganoid, Amia (=Amitus), Hopkins found in one specimen that a branch (p. 373 and fig. 11, t.) connected a large trunk at the base of the tail with the lateral trunk, close to the union of the latter with the caudal sinus. No such caudal continuation of the lateral trunk was observed in the Ganoid, Lepisosteus. Favaro finds in Tinea vulgaris, Esox ( = Lucius) lucius, and Coricus rostratus that the lateral lymphatic trunk is continued to the tail. In Tinea they are represented as anastomosing at the base cf the tail with the sinus lymphaticus caudalis, in this relationship agreeing with Clinocottus and Ophiodon; while in the other two species no anastomosis was recorded, thus agreeing with Scorpsenichthys. Hyrtl who also studied Lucius must have overlooked this posterior continuation of the lateral trunk. In the Dipnoi, Kellicott (p. 208) portrays the lateral subcutaneous veins of Ceratodus as being similar to the lateral cutaneous veins as described by Parker for Mustelus, and posteriorly they are said to anastomose with the caudal \ein.

According to Fohmann, Jones, Jossifov, and Favaro there are no lateral trunks in the fresh water eel, Anguilla.

C/Oncerning the termination of the lateral trunk in Selachians, Parker (p. 721) notes that the lateral cutaneous vein in Mustelus anastomoses with the dorsal and the caudal veins. Sappey (p. 38) finds in Squalus that the lateral lymphatic trunk expands into a fibrous sinus that terminates in the caudal vein. Mayer (pp. 316-7) describes the vena lateralis as receiving the vena dorsalis, and after joining the vena ventralis the latter empties into the vena caudalis.

As for the Ganoids, Hopkins (p. 372) represents each lateral lyjuphatic trunk of Amiatus as terminating under the last verte


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS 19

brae in a caudal sinus that empties into the caudal vein. In Lepisosteus I found in the caudal peduncle region (p. 62, and figs. 1-5) that the lateral trunk bent mesad, and in anastomosing with a haemal trunk, formed a sinus on the vertebral column, designated as sinus (x) , which opened into a caudal sinus that emptied into the caudal vein.

Regarding the Teleosts, Hyrtl (pp. 233-5, and figs. 1-5) and Vogt (pp. 135-6, and pi. K; figs. 3-5, 66) found in Lucius, Leuciscus, and Salmo that each lateral lymphatic trunk terminated behind the last vertebra in a caudal sinus. The former observed that the caudal sinus received the longitudinal neural lymphatic trunk in addition, and both found these sinuses to empty into the caudal vein. Likewise Stannius (p. 254) portrayed the lateral and the longitudinal neural lymphatic trunks as ending in a caudal sinus, which emptied into the caudal vein. Trois (pp. 5, 9, 20-22, 39, 51-2) described in great detail the distribution of the lateral trunks in Lophius piscatorius, Uranoscopus scaber, and in several of the Pleuronectidae, but concerning the caudal ending, has nothing to add to Hyrtl's description. Sappey (pp. 46-7; pi. xi, figs. 4-5, and pi. xii, fig. 2) represented each lateral trunk in the carp and pike as terminating in papilla opposite the last vertebra, which communicates with a fork of the caudal vein.

Taking Tinea vulgaris as a type, Favaro (pp. 182-6, and fig. 84) portrays an entirely different termination of the lateral lymphatic trunk than has been observed before in Teleosts. He finds them to end in two caudal sinuses, one situated behind the last vertebra, designated as the atrium cordis caudalis, and the other situated at the base of the tail, the sinus lymphaticus caudalis, which through the medium of the caudal lymphatic trunk communicates with the former sinus. The atrium is said to be connected mesad with a parallel sinus, the ventriculus cordis caudalis, which discharges its contents into a swelling of the caudal vein, the sinus venosus caudalis. With Lucius lucius, Favaro's representation (pp. 198-9 and fig. 107) of the ending of the lateral lymphatic trunk is quite different from his type described above and from Sappey 's description of the same. Here the lateral trunks are continued to the tail, but have no communication with the sinus lymphaticus


20 W II. LI AM F. ALLEN

caudalis or the caudal lymphatic trunk. Opposite the last vertebra the lateral trunks bend mesad, anastomose, and at the point of union they receive the longitudinal hsemal lymphatic trunk, and the combined trunk passes caudad to enter the atrium. The atrium also receives the caudal trunk and communicates mesad with the ventricle, which empties into the sinus venosus caudalis of the caudal vein. In Coricus rostratus, Favaro finds (pp. 208-9, and fie;. 121) the termination of the lateral trunks more like Scorpsenichthys than in either of the above species. The lateral trunk is prolonged to the caudal fin, but does not anastomose with the sinus lymphaticus caudalis (posterior caudal sinus of Scorpsenich, thys). As in Scorpsenichthys each lateral trunk bends inward, receives a fork of the longitudinal neural trunk; but then, contrary to Scorpa^nichthys, after anastomosing in the median line with its fellow the caudal trunk is received from the rear and the combined ti'unk discharges its contents into the sinus venosus caudalis of the caudal vein.

Dorsal subcutaneous lym/phaiic trunk. (Figs. 1, 4 and 12, D.T.) — In Scorpa^nichthys the dorsal lymphatic trunk proper, not taking into consideration its posterior continuation, which will be described later as the dorsal caudal fin Ijanphatic trunk, is a very inconspicuous vessel compared to what it is in the cephalic region or to what it is in the caudal region of other fishes. The territory usually drained by this trunk is collected by other canals in Scorpsenichthys, namely the neural lymphatic vessels. In no dissection was it possible to trace a continuous dorsal trunk from the head to the tail. In an earlier paper (pp. 54-5) a description of the anterior portion of this trunk and its cephalic ending was given. About all that can be said for a certainty concerning the posterior portion of the dorsal lymphatic trunk in Scorpa;nichthys and Clinocottus is that it takes its origin on either side of the posterior end of the second dorsal fin. At the extremity of the fin the forks fuse to form a single trunk, which continues caudad for a short distance along the dorso-median line of the caudal peduncle directly below the skin in a tough connective tissue sheath that binds the two great lateral muscles together. When a short distance from the base of the caudal fin it anasto


LYMPHATICS IN TAIL REGION, SCORP.^NICHTHYS 21

moses with both the dorsal caudal fin lymphatic trunk and the greatly enlarged next to the last neural lymphatic vessel (fig. 4, C.T. (1) and Neu. T. d)), and through the latter reaches the longitudinal neural trunk. In certain portions of the posterior dorsal fin two neural lymphatic branches fuse at the base of the fin between the extrinsic and intrinsic muscles, forming for a short distance an analogous vessel to the lateral dorsal lymphatic trunks of the anterior dorsal fin. The structure of this trunk is identical with that of the ventral trunk, which will be described in the next paragraph.

Ventral subcutaneous lyrnphatic trunk. (Figs. 1, 4 and 12, V.T.). — This trunk pursues a course along the ventro-median line of the caudal peduncle identical with that of the dorsal trunk along the dorsal side of the caudal peduncle, and like the dorsal trunk it is not a continuous trunk posteriorly. In an earlier paper (p. 57) its cephalic course and termination was given. In front of the vent it divides into a right and left fork, which pass to either side of the vent and can be traced for a short distance along the bases of the anal fin, between the extrinsic and intrinsic muscles. In a like manner it arises again from two forks, coming from the posterior bases of the anal fin. Immediately behind the anal fin these forks unite, forming a single trunk, which traverses the ventro-median line of the caudal peduncle, directly beneath the skin in a tough fibrous tissue sheath that binds the two great lateral muscles together. Near the base of the caudal fin it anastomoses with both the ventral caudal fin lymphatic trunk and the posterior haemal lymphatic trunk or the two last haemal lymphatic trunks (fig. 4, C.T. (2) and Hce.T.a) and (2)), and through the latter reach the longitudinal haemal lymphatic trunk.

A transverse section through the ventral trunk of an adult Scorpaenichthys presents the appearance of a large cavity within the ventral intermuscular septum, which consists of a mass of fibrous tissue ; the ventral portion of which is the inner layer of the skin, and here the fibers have a tendency to run in one direction and to form bundles. The inner layer is endothelium, within which there are but few corpuscles; the red, however, greatly predominate over the white, in a ratio of about 4 to 1 .


22 WILLIAM F. ALLEN

Caudal Lymphatic trunks.— Of these vessels, those coming from the basal canal of the caudal fin, namely, the dorsal and ventral caudal fin lymphatic trunks (figs. 3-7, C.T. o and (2)) are very large and important vessels occupying most of the space in this canal, and for a considerable distance they run parallel to corresponding blood vessels and nerves. In the median line the dorsal and ventral trunks anastomose, and the trunk thus formed extends cephalad a short distance between the posterior ends of the hypural bones, where it expands into a small sinus, designated as the posterior caudal sinus. As previously stated the dorsal caudal lymphatic trunk unites with the dorsal lymphatic trunk a little anterior of the base of the caudal fin, and at the point of union they are joined by a greatly enlarged neural trunk{hgA,Neu.T.(i)), which occupies a large part of the space between the neural spines of the third and fourth vertebrae from the last, and terminates in the longitudinal neural lymphatic trunk opposite the third vertebra from the last. A similar arrangement is found in connection with the ventral caudal fin lymphatic trunk, which together with the ventral lymphatic trunk fuse as shown in fig. 4 with the last two haemal lymphatic trunks {Hoe.T. (d and (2)). Possibly a more correct rendering of these relationships would be to consider the next to the last neural trunk (fig. 4, Neu.T.a)), and the last two haemal trunks (figs. 4, Hw.T. u) and (2)) as continuations of the dorsal and ventral caudal fin lymphatic trunks, which empty directly into the longitudinal neural and hasmal lymphatic trunks. With this interpretation the dorsal lymphatic trunk would be said to empty into the dorsal caudal fin lymphatic trunk, and the ventral lymphatic trunk into the ventral caudal fin lymphatic trunk.

Throughout their course through the basal canal of the caudal fin both caudal fin lymphatic trunks receive numerous caudal ray lymi)hatic trunks (figs. 3-5, and 7, C.R.T.) ; two of which traverse the dorsal and ventral surfaces of each caudal ray. They collect a rather course network, decidedly lymphatic, in the character of its meshes, and which so far as could be determined had no connections with the arterial system, from the fin membrane connecting two rays. Ordinarily, especially in the extreme dorsal


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS 23

and ventral portion of the fin, two caudal ray lymphatic vessels from adjacent rays would anastomose a short distance from the basal canal of the caudal fin, and the combined trunk would terminate in the caudal fin lymphatic trunk. Toward the center of the fin, as shown in figs. 5 and 7 where this region was dissected out more carefully than in fig. 4 what at first appeared to be a continuous caudal fin trunk is in reality composed of several diverticula into which several of the caudal ray trunks emptied. One of these, as shown in fig. 5, is continued so as to empty directly into the posterior caudal sinus.

Posterior caudal sinus. (Figs. 3-7, C.S.). — The sinus so named can hardly be compared to the paired caudal sinuses of Lepisosteus (p. 67 and figs. 1-7, R. a smdL.C.S.) and other fishes, which are situated behind the last vertebra, collect the lymph from the tail region and eject it into the caudal vein, but is evidently homologous to what Favaro (p. 184 and fig. 84) describes and figures as the sinus lymphaticus caudalis in Tinea vulgaris; for in this fish he also describes and figures a paired caudal lymphatic heart situated behind the last vertebra consisting of two cavities, designated as the atrium cordis caudalis and the ventriculus cordis caudalis. While these caudal hearts or sinuses as described in Tinea have somewhat different connections than are found in Lepisosteus and other fishes, still they have much in common and must be homologous. In Clinocottus the so-called posterior caudal sinus (fig. 3, C.S.) is partially paired throughout most of its length and each reservoir is in communication with the posterior portion of a lateral lymphatic trunk. With Scorpsenichthys most diligent search was made to find a connection between the posterior portion of the lateral trunk and the posterior caudal sinus, but none was found. A branch of the caudal artery is given off to the periphery, which should not be confused with the above mentioned lymphatic connection. In the anterior end of the posterior caudal sinus there are at least two orifices into which the caudal lymphatic trunks open.

These trunks (figs. 4-7, C.T.) are very slender and delicate vessels in Scorpsenichthys, which occupy most of the space between the caudal artery and the hypural bones. Upon reaching


24 WILLIAM F. ALLEN

the last vertebra the ventral stem after receiving a communication from tlie dorsal stem, enters the haemal canal with the caudal and the minor caudal arteries to form the beginning of the longitudinal haemal lymphatic trunk. Near the last vertebra the dorsal stem of the caudal lymphatic trunk receives two hypural lymphatic vessels (figs. 4-7, Hyp.T.) coming from either side of the superior hypural bone, which follow the course of their corresponding arteries, often nearly surrounding them, and each collects a parallel network from the dorsal and posterior lateral, surfaces of the superior hypural bone, which is continuous with a similar network that is gathered by the posterior neural lymphatic vessel. It should be noted, however, that no vein was seen collecting the arterial network from the superior hypural bone, and it may be that the so-called hypural lymphatic trunk receives this network. Still, however, no connections were found between the arterial network and the lymphatic network ; they ran parallel to each other, and appeared to be two separate systems. On one side of the superior hypural bone, but seldom on both, there is an interlinking lymphatic canal (figs. 4, 6, and 7) connecting the dorsal stem of the caudal trunk with either profundus portion of the lateral lymphatic trunk or the posterior neural lymphatic trunk.

Concerning the distribution and caudal ending of the dorsal, ventral, and caudal lymphatic trunks in the tail region as described by different authors there is great variation. In many species the posterior portions of the dorsal and ventral subcutaneous trunks are large and important canals, and in place of the small caudal lymphatic trunks between the hypural bones as was described for Scorpienichthys there are large paired caudal sinuses or hearts, which collect the lymph from the entire tail region and discharge it directly into the caudal vein.

In the Selachians, Parker (p. 721) found in Mustelus that the posterior ventral cutaneous vein formed loops about the anal fin and the cloaca, and posteriorly the lateral cutaneous veins were said to anastomose with the dorsal cutaneous and the caudal veins. Sappey observed that Squalus (pp. 38-9) possessed a dorsal and a ventral lymphatic trunk. The dorsal trunk was said


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS 25

to have its origin behind the posterior dorsal fin and to encircle the bases of both dorsal fins. On the anterior fin a secondary elliptical stem (pi. x, fig. 3, 7) is represented as rising from the posterior end of these elliptical vessels and after crossing the lateral surface of the fin terminates in the anterior end of the same vessel from which it takes its origin. Mayer (pp. 316-7) states that Parker's description of the caudal ending of these cutaneous trunks is insufficient; for he finds that the dorsal vein empties into laterals, which in turn join the ventrals, and the latter being paired at the origin of the tail soon terminate in the caudal vein. Mayer likewise portrayed the dorsal subcutaneous vein (pp. 333-4) as encircling the bases of the dorsal fins as vense circulares (pi. xvii; fig. 17, vcirc.) At the posterior insertion of the fin, namely at the point of division of the dorsal fin, a reservoir of considerable size was formed into which emptied a vena postica, coming from the posterior part of the fin, and the vena profunda (pi. xvii; figs. 21, 22, and 24, vproj.), which, in addition to communicating with this sinus also joined the caudal vein.

With the Ganoids, Hopkins noted with Amiatus that the dorsal lymphatic trunk after leaving the dorsal fin separated into two branches, which anastomosed with the lateral trunks near their termination in the caudal sinuses. The ventral lymphatic trunk was portrayed (pp. 372-3 and fig. 11, v.) as beginning as a large canal (fig. 11, o.) at the base of the caudal fin, which in one instance was said to send off a communicating branch (t) to the lateral trunk. As stated previously the lateral lymphatic trunks, into which all the lymphatics of the tail region were discharged, emptied into two caudal sinuses (fig. 11, s.) which were situated below the last vertebrae, communicated with each other, and culminated in the caudal vein. In an earlier paper I found in Lepisosteus that there were two conspicuous caudal sinuses situated under the caudal vertebrae, which emptied into the caudal vein. Posteriorly one of these sinuses received a caudal trunk from the caudal fin, which in reality is a continuation of the ventral subcutaneous trunk through the basal canal of the caudal fin. Into each of the caudal sinuses, a sinus disignated as sinus (x) emptied its contents. The latter were described and figured


26 WILLIAM F. ALLEN

as passing along the side of the last vertebrae, where each collected a lateral and a longitudinal haemal trunk; and one or the other of them, the dorsal subcutaneous trunk. In the tail region of Lepisosteus the dorsal and ventral subcutaneous trunks are enormous canals, and contrary to the Selachians they do not divide upon reaching the dorsal and anal fins, but pass directly through their basal canals and collect branches from their rays.

In the teleosts Hyrtl and Vogt noted two caudal hearts behind the last vertebra in Lucius, Leuciscus, and Salmo, which termi- . nated in the caudal vein. Since the dorsal and ventral lymphatic t runks were not described they must have been overlooked Trois described a dorsal and a ventral lymphatic trunk in Lophius (pp. (3_8)^ Uranoscopus (pp. 23-24), and in the Pleuronectida? (pp. 40-41). In Lophius and Uranoscopus these trunks were said to trifurcate in the region of the dorsal and the anal fins, one trunk passing through the basal canals of the fins and the other two to either side, and the median trunk was represented as collecting two ray vessels from each ray. Nothing of especial interest was said concerning the caudal ending of these trunks, but in Uranoscopus, according to fig. 4, the dorsal and ventral lymphatic trunks are continued to the tail, where they apparently fuse with the longitudinal neural and haemal lymphatic trunks. They comnmnicate superficially with the lateral trunk through the intermuscular vessels, and deeply with the longitudinal neural and haemal trunks through the neural and haemal vessels. In all these fish Trois portrayed the superficial lymphatic trunks to be well developed posteriorly. Sappey (p. 471, and pi. xi, fig. 5, and pi. xii,fig. 2) found a somewhat similar arrangement of the dorsal and ventral lymphatic trunks of the carp and pike to that of Trois for Lophius and Uranoscopus; except that the dorsal and ventral trunks in the caudal region are less important canals, for they did not extend clear to the tail. In Pleuronectes the dorsal and ventral lymphatic trunks were portrayed by Sappey (pp. 50 and p. xii, fig. 4) as being continued through the basal canal of the caudal fin, where they anastomosed and thus formed an elliptical trunk about the body, which emptied dorso-cephalad in the jugular and ventro-cephalad in the ductus of Cuvier.


LYMPHATICS IN TAIL REGION, SCORP-ENICHTHYS 27

One of the Teleosts that swims by a snake-like movement, namely, the eel (Anguilla) has a pulsating heart in the tail. The old view of this heart as presented by Jones (pp. 676-9 and figs. 1 and 2) was that it consisted of a single contractle reservoir situated near the end of the tail. No lymphatics were portrayed as emptying into it, but it was represented as discharging its contents into the dorsal or minor fork of the caudal vein. In 1905 Favaro (p. 571, fig. 1) and in 1906 (pp. 157-168, figs. 64, 70 and 71) sets forth a very different arrangement. He finds this heart paired, consisting of an atrium or auricle and a ventricle, which communicate with each other mesad. Anteriorly the atrium is represented as receiving the longitudinal haemal lymphatic trunk and posteriorly the caudal lymphatic trunk, both orifices being guarded by valves; while the ventricle has but one opening, which is cephalad into the caudal vein, and is likewise guarded by valves. This heart is said to possess three tunics, an internal one of endothelium, a median of elastic fibers, and an external of striated muscles.

In Tinea vulgaris, Favaro (pp. 181-6 and fig. 84, II) finds an arrangement quite similar to the eel. Behind the last vertebra on either side of hypural interval there are two caudal sinuses, which are connected with each other through this interval. One of them the ventricle or ventriculus cordis caudalis opens anteriorly into a swelling of the caudal vein, the sinus venosus caudalis ; while the atrium or atrium cordis caudalis has two orifices, the anterior receives the profundus portions of the lateral trunks and the posterior the caudal trunk. The latter trunk arises from a sinus, sinus lymphaticus caudalis, situated in the hypural interval at the base of the caudal fin. This sinus is also said to receive the posterior portion of the lateral trunks, and the caudal fin trunks. The latter collect the caudal ray branches from the fin, and are continuous cephalad with the dorsal and ventral lymphatic trunkS; which trunks are said to be but little developed and as in Scorpsenichthys they communicate with the lateral trunks through transverse branches. In Lucius lucius quite a different arrangement of the caudal lymphatic trunks is given by Favaro (pp. 1969 and fig. 107) than was recorded by Sappey, which is to a great


28 WILLIAM F. ALLEN

extent similar to Anguilhi and Tinea. An atrium and ventricle are present and they communicate with each other. The ventricle empties anteriorly into the sinus venosus caudalis of the caudal vein. Anteriorly the atrium receives a common trunk formed by the union of the longitudinal hsemal and the lateral lymphatic trunks; while posteriorly it receives the caudal lymphatic trunk, which according to Favaro in this species simply passes through the hypural interval, without expanding into any sinus lymphaticus caudalis that communicate with the posterior portions of the lateral trunks. It nevertheless continues through the basal canal of the caudal fin, and is continuous with the dorsal and ventral trunks, which are said to be rather rudimentary. Favaro's account of the arrangement of the caudal lymphatic vessels in Coricus rostratus (pp. 208-10 and fig. 121) is more like Scorpsenichthys than any of the above mentioned species. In Coricus no sinuses corresponding to the atrium and ventricle were described. There is, however, a longitudinal neural trunk, which anastomoses with the lateral and caudal trunks, and the common trunk thus formed empties into the sinuses venosus caudalis of the caudal vein. The caudal trunk is said to have a little swelling at the base of the tail, without assuming the appearance of a sinus lymphaticus caudalis. Nothing is said concerning the dorsal and ventral lymphatic trunks, but the longitudinal haemal trunk of other species is absent in Coricus.

PROFUNDUS LYMPHATIC SYSTEM

In Scorpsenichthys and Clinocottus the profundus system, which consists of the longitudinal neural and haemal lymphatic trunks together with their branches are very large and important canals, far more so than the subcutaneous trunks.

Longitudinal neural or superior vertebral lymphatic trunk. (Figs. 3-7, and 12 L.Neu.T.). — This enormous trunk is situated in the neural canal directly above the myelon or spinal cord, being separated from it by a tough fibrous tissue septum. Its exact position is shown in transverse section (fig. 12, L.Neu.T.). Since this trunk in Scorpsenichthys and Clinocottus has no direct con


LYMPHATICS IN TAIL REGION, SCORP.ENICHTHYS 29

nection with the caudal vein it must be said to take its origin above the last vertebra from the anastomosis of the two profundus portions of the lateral trunks (figs. 3-7, L.T.(2)) and the posterior neural trunks (figs. 4, 6, and 7, P.Neu.T.). As stated previously the latter vessels collect a network from the sides of the last interspinal bones; while the profundus portions of the lateral trunks would in all probability collect most of the lymph from the posterior portions of the lateral trunks (figs. 4, L.T. (d) and some from the main lateral trunks (fig. 4, L.T .), but as the latter trunk increases in caliber caudo-cephalad the resistance must be less in that direction, with the result that most of the lymph flows in that direction. In Clinocottus the posterior portions of the lateral trunks are also in communication with the posterior caudal sinus, in which case the lymph received by that sinus could find its way into the posterior portions of the lateral trunk and thence into the longitudinal neural trunk, or the lymph from the posterior portions of the lateral trunks could go in the opposite direction, namely, into the posterior caudal sinus, in which case it would ultimately find its way into the longitudinal neural or haemal lymphatic trunks through the dorsal or ventral caudal fin lymphatic trunks. On one side of the tail in Scorpsenichthys the caudal lymphatic trunk communicates with a profundus portion of one of the lateral trunks, hence a possible supply from that source. A most important accession, however, is the next to the last neutral trunk or the dorsal caudal fin trunk (fig. 4, Neu.T. (d) which joins the longitudinal neural trunk opposite the third vertebra from the last. It is an enormous trunk formed by the anastomosis of the small dorsal trunk (fig. 4, D.T.) and the large dorsal caudal fin trunk (fig. 7, C.T. (d). As previously stated the latter trunk collected the lymph from the upper half of the caudal fin and doubtless some from the posterior caudal sinus. In other words then, a glance at fig. 4 shows the arrangement of the lymphatics to be such, that it would be possible for the lymph from the entire tail region to reach the longitudinal neural trunk, but under ordinary circumstances, however, it is probable that a large portion of the lymph reaches the longitudinal haemal trunk through the ventral caudal fin trunk. The flow of h^mph in the longi


30 WILLIAM F. ALLEN

tutliiial neural trunk is therefore cephalad, where it is discharged into the jugular veins after the manner described in an earlier paper (p]). 59-61).

A trans\Trse section through the longitudinal neural lymphatic trunk of Clinocottus shows this trunk to more than ecjual the size of the myelon or spinal cord. It is composed of fibrous tissue and lined with endothelium, and strange to say contained very few corpuscles. In a representative section but two were noted and they were red.

Throughout its entire course the longitudinal neural lymphatic trunk receives a neural lymphatic vessel (fig. la, Neu.T.) opposite each centrum, that is, one for each segment. It should be recalled in this connection that but one neural artery or vein was found for every two segments, the arteries alternating with the veins. Consequently^ then, since these neural lymphatic vessels run parallel, but distal, to the blood vessels in their course along the neural spines, each alternate neural lymphatic vessel would follow a neural artery, and every intermediate alternate neural lymphatic vessel would follow a neural vein. There are, however, a few exce])tions to this plan. Tracing a neural lymphatic vessel dorsad, one sees it leave the apex of the neural spine with the corresponding blood vessel to cross the depressor and levator muscles of the dorsal rays, and at this level it sends off a cephalic branch to anastomose with the next neural lymphatic vessel, thus forming a sort of irregular longitudinal lymphatic trunk (represented in fig. 4 a, but not lettered) homologous to a similar secondary dorsal lymphatic trunk described by Trois in Pleuronectes and by Sappey for Lucius lucius. The neural lymphatic vessel proper continues dorsad with the corresponding blood vessel between the levator and the next dei)ressor dorsal ray muscles to the extrinsic muscles of the fin. where it bifurcates with the corresponding blood vessel between the extrinsic and intrinsic muscles. Sometimes two of these branches from adjacent neural vessels anastomose, forming a continuous longitudinal vessel for a short distance, which is analogous to the lateral dorsal lymphatic trunks described in an earlier paper (p. 55) for the anterior dorsal fin. These forks of the neural vessel collect the dorsal ray lymphatic canals


LYMPHATICS IN TAIL REGION, SCORP^ENICHTHYS 31


(fig. 4 a, D.R.T.), which traverse the posterior surfaces of the raja's and receive a network from the fin membrane.

The arrangement of the lymphatics in the posterior dorsal or the posterior portion of the dorsal fin is perceptibly different from the anterior dorsal or the anterior portion of the dorsal fin ; where there were three longitudinal dorsal lymphatic trunks, two of which passed along the sides of the base of the fin and the other traveled through the center of its basal canal. The central trunk received two dorsal spine canals from each spine and communicated with the lateral trunks through numerous cross branches ; while the lateral dorsal trunks were connected with the longitudinal neural trunk through the neural vessels and with the lateral trunk through the intermuscular vessels.

Longitudinal hcemal or inferior vertebral lyjnphatic trunlx. (Figs. 4, 7, and 12, L.Hce.T.). — This great profundus trunk is of almost equal importance to the longitudinal neural lymphatic trunk. In Scorpsenichthys it may be said to take its origin in the haemal canal as a small vessel under the last vertebra, which is continuous with the caudal lymphatic trunk. The latter has been described as traversing the hypural interval, and beside several minor connections it opens into the posterior caudal sinus at the base of the caudal fin, which also is continuous with the dorsal and ventral caudal fin lymphatic trunks. Under the last vertebra the longitudinal haemal trunk is a ver}^ inconspicuous vessel, and often in sections of Clinocottus through the region of the last vertebra it was invisible. It doubtless receives very little, if any, lymph from the caudal trunk because the resistance in that vessel is evidently less caudad, that is toward the posterior caudal sinus. This being the case a portion of its lymph would ultimately reach the longitudinal haemal trunk through the ventral caudal fin trunk, which forms the principal accession to the longitudinal haemal, trunk, in fact, it might be said to constitute its source. As stated previousl}^ the ventral caudal fin trunk after leaving the basal canal of the caudal fin, receives the small ventral lymphatic trunk, and here separates into two trunks (fig. 4 Hce.T. (1) and (2)) which enter the haemal canal opposite the third and fourth vertebrae from the last, and anastomose with the small


32 WILLIAM F. ALLEN

longitudinal liaenial lymphatic trunk. From this point cephalad, this longitudinal trunk may be a single canal or it may have resolved itself into several, in which case they appear in transverse section like rather large cavities in the spongy connective tissue supporting the blood vessels. In this trunk the corpuscles areextremely scarce, and as was noted for the other lymphatic trunks, the red greatly outnumber the white. Upon entering the body cavity this trunk takes on more the form of a sinus, which follows the aorta, cephalad, between the vertebral column and the kidney, and where the kidney separates into right and left lobes at the insertion of the retractor muscles of the pharyngeal bones, it joins the great abdominal sinus, situated below the kidney. In an earlier paper (pp. 62-3) the cephalic termination of this sinus has been fully given.

From the tail to the body cavity the longitudinal haimal trunk received a htemal lymphatic trunk (fig. 4a, Hoe. T.) from between each two segments, which amounts to, one for each segment. Exactly the same correlation was established between these haemal lymphatic vessels and the blood vessels as was shown between the neural lymphatics and the neural blood vessels. Along the anterior surface of a haemal spine there traveled a haemal lymphatic vessel and a haemal artery, and along the next spine a lymphatic vessel and a vein, this relationship being even more constant than was the case with the neural vessels. In the region of the anal fin the haemal lymphatic vessels arose from two branches from between the extrinsic and intrinsic muscles of the fin, which occasionally anastomose with the corresponding branch ahead or behind, thus forming a short longitudinal trunk on the side of the base of the fin. These branches also collect the anal ray lymphatic canals (fig. 4a, A.R.T.), which traverse the posterior surfaces of the rays and gather a network from the fin membrane. Thus formed a haemal lymphatic trunk crosses a pair of intrinsic anal ray muscles, but upon reaching the apex of a haemal spine it does not send off a connecting branch to the adjacent haemal trunk, to form a secondary ventral longitudinal trunk. Its course was then dorsad along the haemal spine, immediately distad of the corresponding blood vessels, and entering the haemal canal it joins the longitudinal haemal lymphatic trunk.


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS 33

At various intervals interlinking lympathic vessels were seen passing around ttie centra, and connecting the longitudinal haemal trunk with the longitudinal neural trunk.

As regards these longitudinal profundus lymphatic trunks in other fishes, especially inTeleosts, as protrayed by different authors there appears to be great variation. In some species one trunk is said to be absent ; in another, the other ; while in still others both are absent or else have been overlooked.

In the Selachians there are no longitudinal neural trunks, but in Scyllium, Mustelus, and Squatina, Mayer (p. 320) describes a vasa vasorum in the haemal canal, which consists of a longitudinal vein traveling along between the caudal artery and vein. It is said to receive direct communications from each intercostal or intersegmental artery, and to give off at regular intervals branches to the caudal vein. Favaro finds a similar arrangement in Squalus, where he designates the longitudinal vein running between the caudal artery and vein as the vasa intermedia, the connecting branches between the segmental or intersegmental arteries and the vasa intermedia are termed the arteria segmentalis vasorum intermediorum ; while the connecting vessels between the vasa vasorium and the caudal vein are plainly figured but are not named. In fig. 156 Favaro graphically shows the evolution of the longitudinal haemal trunk of Teleosts from the vasa intermedia of Squalus. The arrangement for Squalus was just given above. With Raja he portrays these connecting arteries as passing caudad a short distance before anastomosing with the vasa intermedia; while the arrangement of the connecting vessels between the vasa intermedia and the caudal vein remain as in Squalus. With Acipenser these connecting arteries are continued caudad a short distance as in Raja, but here they fork, one stem joins the vasa intermedia and the other is continued caudad to anastomose with the following segemental artery, thus forming a continuous longitudinal arterial trunk, designated as the arteria longitudinales vasorium intermediorum (Minor caudal artery of Scorpaenichthys) . The venous connections between the vasa intermedia and the caudal vein are the sam e as in Squalus and Raja. With the Teleosts the arterial arrangement islden THE AMERICAN JOURNAL OF ANATOMY, VOL. 11, NO. 1.


34 WILLIAM F. ALLEN

tically the same as in Acipenser; while there are no connecting vessels between the vasa vasorum and the caudal vein, the former becoming the longitudinal haemal lymphatic trunk. It is also of interest to note that in the haemal canal of an embryo of Belone acus, Favaro finds a similar arrangement to the vasa intermedia as represented above for Squalus.

With the Ganoids, Hopkins did not protray any longitudinal neural or haemal trunks for Amiatus. In Lepisosteus there is no longitudinal neural trunk, but I found (p. 64) two longitudinal haemal trunks in the haemal canal of the tail region, which passed laterad under one of the posterior vertebrae and anastomosed with the lateral trunks in forming the two sinuses (x), which emptied into caudal sinuses that terminated in the, caudal vein. Strange to say no haemal vessels were observed.

In the Teleosts Vogt does not mention either of these profundus longitudinal lymphatic trunks in Salmo. Hyrtl notes a longitudinal neural trunk in Leuciscus, which collects lymphatic vessels from the dorsal fin, but has nothing to say concerning a longitudinal haemal trunk. Trois finds in Lohpius (pp. 11-12) and in Uranoscopus (pp. 25-6) that the superior and inferior longitudinal spinal trunks are well developed. His fig. 4 shows these trunks to fuse behind the last vertebra, and apparently also with the dorsal and ventral lymphatic trunks. Behind the abdomen, interspinal vessels are said to connect these trunks with the dorsal and ventral trunks. In the Pleuronectidae, Rhombus maximus, R. laevis, and Pleuronectes grohmannij Trois represents (pp. 43-4, and fig. 2, D^ and D-) in addition to the above mentioned profundus trunks, that two other longitudinal trunks travel along on a level with the apices of the neural and haemal spines, and anastomose with the interspinal vessels. These he claims to have described in an earlier paper, prior to Sappey's description of them for Lucius lucius. In P. grohamanni cross branches were observed by Trois passing from the inferior longitudinal spinal trunk to the superior longitudinal spinal trunk. Sappey's description of these profundus lymphatic trunks in Lucius and Pleuronectes is similar to that given by Trois, except that these trunks are not portrayed as extending to the last vertebra. Favaro finds in Tinea vulgaris


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS 35

(p. 181) that the longitudinal neural trunk is absent, and in his fig. 84 no longitudinal haemal trunk appears. With Lucius lucius, Favaro noticed both profundus longitudinal trunks; the longitudinal neural did not extend caudad far enough to empty into the atrium of the caudal heart; while the longitudinal hsema] trunk is well developed, and after anastomosing with the profundus portions of the lateral trunks, the combined trunk emptied into the atrium of the caudal heart. In Coricus rostratus (p. 208) there is said to be no longitudinal hsemal trunk, but a longitudinal neural trunk is represented as dividing and each fork anastomosing with a profundus portion of a lateral trunk, the combined trunks thus formed, fuse, and at the point of union, receive the caudal trunk, and the common vessel thus formed does not empty into the caudal heart, but terminates directly into the sinus venosas of the caudal vein. In the eel, Anguilla, Favaro states (pp. 157-8) that the longitudinal hsemal trunk is the only profundus lymphatic trunk observed, and it is said to terminate directly in the cephalic end of the atrium of the caudal heart.

SUMMARY AND GENERAL CONSIDERATIONS

In the tail region of the Cottids, Scorpsenichthys marmoratus and Clinocottus analis there is a distinct system of longitudinal canals, which has no counterpart in the arterial system. It consists of four longitudinal subcutaneous trunks and two longitudinal profundus trunks. None of these terminate behind the last vertebra in caudal sinuses, which empty into the caudal vein, as is the case with many fishes. In fact, so far as could be determined there is no direct communication of the lymphatics with the caudal vein or with the arterioles in the periphery. It should be noted, however, that in four places, namely, on either side of the last vertebra and at the base of the tail this system comes into close touch with the two forks of the caudal vein, and in one section of Clinocottus the thin wall separating the profundus portion of the lateral trunk and the right fork of the caudal vein was torn in such a manner as to give the appearance of the lymphatic trunk empting into the vein and the orifice being guarded by


36 WILLIAM F. ALLEN

two valves: but in no other series or in any other section in this series or in any dissection of Scorpsenichthys was any direct opening found. As stated in the text the filling of the blood vessels from an injection of the longitudinal neural trunk and conversely the filling the lymphatics from an injection of the caudal artery or vein was attributed to the rupturing of the delicate walls separating these systems, resulting in the extravasation of the injection mass from one system to the other, rather than to a direct passage from one to the other through an open communication.

Since no caudal connection was established between these systems, and the fact that all the longitudinal trunks increase in caliber anteriorly, the flow of lymph must be cephalad to enter the jugular vein after the manner described in an earlier paper. Two forces evidently contribute to propel the lymph forward; one is the difference of pressure between these two systems, forming a sort of suction in the lymphatics; while the other, which is probably the main factor, is the lateral movement of the tail and body against the great wall of water, which presses the lymphatic trunks against the muscles and the bones.

In these Cottids the profundus longitudinal lymphatic trunks are larger and more important than the subcutaneous trunks. The longitudinal neural trunk is without doubt the main lymphatic conduit of the body.

Both the longitudinal neural and the longitudinal haemal trunks begin at the last vertebra, the former from the anastomosis of the profundus portions of the lateral trunks, and the latter as a continuation of the caudal trunk. Each receives from between every two segments, a neural or a haemal branch after the manner previously described.

Two of these, the posterior neural and the posterior ha-mal lymphatic trunks are of enormous caliber. They are formed by the anastomosis of the small dorsal and the large dorsal caudal fin lymphatic trunks, and the small ventral and the large ventral caudal fin lymphatic trunks respectively.

With the Cottids the dorsal and ventral caudal fin lympnatic trunks are enormous canals which occupy a greater part of the basal canal of the caudal fin. From the dorsal and ventral sur


LYMPHATICS IN TAIL REGION, SCORPiENICHTHYS 37

faces of each ray they collect a caudal ray lymphatic vessel. In the median line, between the two hypural bones, at the base of the fin the two caudal fin lymphatic trunks unite in the small posterior caudal sinus.

This sinus is partially paired in Clinocottus. With Clinocottus and Ophiodon it has conspicuous connections with the posterior portion of the lateral trunks, but in Scorpainichthys there are no such communicacions. Anteriorly two or more caudal lymphatic trunks leave this sinus to foUow the caudal artery into the haemal canal, where they anastomose and form the longitudinal haemal trunk.

The dorsal and the ventral lymphatic trunks in the caudal region are very inconspicuous vessels in comparison to what they are cephalad or to what they are in the caudal region of other fishes. The tract usually drained by these vessels in the posterior dorsal and the anal fins is collected in Scorpsenichthys by branches of the neural and haemal trunks. In the case of Scor psenichthys the dorsal and ventral trunks of the caudal region may be said to have their origin from the posterior ends of the second dorsal and the anal fins, and after traversing the dorsal and ventral median septa of the caudal peduncle they anastomose with the dorsal and the ventral caudal fin lymphatic trunks, which culminate in the longitudinal neural and the longitudinal haemal lymphatic trunks.

In the tail region the lateral lymphatic trunks are far more important canals than either the dorsal or the ventral. Each can be divided into a main portion, a posterior portion, and a profundus portion.

The main portion travels immediately beneath the lateral line, in a tough connective sheath that binds the two halves of the lateral muscle together. It receives a dorsal and a ventral intermuscular or transverse lymphatic canal from the surface of each intermuscular sepum.

The posterior portion of the lateral trunk is confined to the subcutaneous region between the last vertebra and the base of the caudal fin. It collects a network from the fascia and several from the intrinsic muscles of the caudal fin. With Ophiodon


38 WILLIAM F. ALLEN

(fig. 2) this network in the fascia assumes the form of a fan. In Scorpsenichthys there is no commimication at the base of the caudal fin between this trunk and the posterior caudal sinus ^ but in Clinocottus and Ophiodon there are very distinct connections.

Opposite the last vertebra the main and the posterior portions of the lateral trunks may be said to fuse and form the profundus portions, which pass mesad and anastomose at the beginning of the neural canal to form the great longitudinal neural lymphatic trunk.

The structure of the longitudinal lymphatic trunks are very much the same, they are made up for the most part of a dense fibrous layer, which is lined with endothelium. Notwithstanding the enormous size of these canals they contain but very few corpuscles, and strange to say the red greatly outnumber the white.

Respecting the distribution of the blood vessels previously given in detail, a few points are deserving of special mention.

The so-called minor caudal artery has an entirely different origin in Scorpsenichthys than was described for it by Favaro in numerous Teleosts, where it was said to arise from branches of the segmental arteries. So far as could be determined in Scorpa^nichthys ic nad no connections with the intersegmental arteries, but rather arose from the dorsal aorta in the body cavity, and continued parallel with it and the caudal artery to the tail ; where it separated into a dorsal and ventral minor caudal fin arteries, which gave off a minor caudal ray artery for each ray. These branches, however, could not be traced caudad of the intrinsic muscles. In places within the haemal canal, the minor candal arteries gave ofif branches of almost equal caliber to itself. The function of these branches appears to be to furnish nutrient arteries for the blood vessels and possibly for the lymphatics of the ha>mal canal.

There are three distinct varieties of intersegmental arteries and veins, arising from the caudal artery or terminating in the caudal vein. The first or lateral vessels simply curve around to the side of the centra and then pass laterad in the intermuscular septa to the periphery. The second or combined lateral and neural trunks


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS 39

likewise curve around to the side of the centra, where each divides into a lateral vessel which has an identical course to the lateral vessel described above, and a neural vessel which follows the neural spine to the periphery. The third or haemal vessels pursue a ventral course along the haemal spines to the periphery. The outcome of this arrangement is that every two myotomes are supplied by a neural, hsemal, and two lateral arteries and veins.

Resume. — In connection with this system of canals in the tail region of these two Cottids two views are tenable. One is, that it is a separate lymphatic system, probably more closely related to the blood-vascular system than in the higher Vertebrata; the other is, that it is a separate venous system, which has no counterpart in the arterial system, but which may function also for lymphatics.

Of these two views the former seems more plausible for the following reasons. — So far as could be deterpiined in the tail region there w^as no direct connection between this system and the caudal vein as is found in most fishes or with the arterial system in the periphery. The smaller branches usually follow the arteries, often nearly surrounding them, and their networks are decidedly lymphatic in the character of their meshes. The caliber of these trunks are also much larger than one would expect veins to be. When these canals are severed in a living specimen no blood is expelled, in fact, in a series of microscopical sections one frequently has to look at several sections to find a corpuscle in one of these trunks. With a single possible exception the entire tail region, both superficial and deep, are amply supplied with veins that have mutual relations with the arteries ; while the lymphatic system is always isolated. It was shown that along each neural spine there extended a neural lymphatic vessel, which received lymph from the dorsal fin and the dorsal musculature, but none from the lateral periphery (that region being drained by the lateral lymphatic trunk), and emptied into the longitudinal neural lymphatic trunk. Also along each alternate neural spine there traveled a neural artery, which arose from the caudal artery, and supplied the dorsal and lateral musculature, the dorsal fin and its musculature, and the dorsal and lateral periphery. Likewise


(■

. 40 WII)lIAM F. ALLEN

along the iiiterinediate alternate neural spines there pass neural veins, which have identical courses to the corresponding arteries, draining the same regions they supplied, and terminating in the caudal vein. Assuming for a moment the neural lymphatic vessels to be neural veins; there would be a neural artery and vein for each alternate neural spine, and two neural veins, having almost identical courses for every intermediate alternate neural spine; thus constituting a very unlikely arrangement. Exactly the same correlation can be shown in connection with the hsemal vessels.

The following observations to a certain extent favor the supposition that this sysiem is a venous system. — On both sides of the superior hypural bone in Scorpsenichthys there was noted an arterial and a lymphatic system, but no vein was observed coming from that locality, unless the hypural lymphatic vessel also function as a vein. Likewise in Clinocottus the caudal vein was traced only to its point of bifurcation in the basal canal of the caudal fin. No caudal ray veins were seen emptying into these short caudal fin veins, but as both caudal ray arteries and lymphatics were clearly defined, the absence of such vessels could hardly be attributed to faulty technique; tience the caudal ray lymphatic vessels and the caudal fin lymphatic trunks in Clinocottus may function as both lymphatics and veins. All of the microscopic preparations clearly demonstrate that the red corpuscles greatly outnumber the white in all the longitudinal lymphatic trunks.

This study to a considerable extent supports the hypothesis set forth in 1907 (p. 93) and in 1908 (pp. 72-3) and also prev^iously championed by Favaro that the lymphatics of fishes have evolved from veins, the evidence of course being derived solely from a study of comparative anatomy. In the Selachians these trunks have every indication of being veins for there are numerous connections with the veins throughout the entire body; while with the Teleosts these trunks are undoubtedly lymphatics. The Ganoids appear to be a sort of intermediary; Polyodon, a cartilaginous Ganoid, leaning toward the Selachians ; while Lepisosteus, a bony Ganoid, inclines toward the Teleosts. It was shown in these Ganoids that connections with the venous system are quite nunier


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS 41

ous in the head region; while in Lepisosteus there are but two in the tail region. With most Teleosts there are two connections in the head and two in the tail, but in the specialized Scorpa^nichthys these communications are confined solely to the head.


42 WILLIAM F. ALLEN


BIBLIOGRAPHY

Agassiz ET VoGT. 1846 Anatomic des Salmones. Memoires de la soci6t6 des sciences naturelles de Neuchatel.

Allen, W. F. 1905 The blood-vascular system of the Loricati, the mail-checked fishes. Proc. Wash. Acad, of Sci., vol. 7.

1906 Distribution of the lymphatics in the head, and in the dorsal, pectoral, and ventral fins of Scorpoenichthys marmoratus. Proc. Wash. Acad, of Sci., vol. 8.

1907 Distribution of the subcutaneous vessels in the head region of the ganoids Polyodon and Lepisosteus. Proc. Wash. Acad, of Sci., vol. 9.

1908 Distribution of the subcutaneous vessels in the tail region of Lepisosteus. Am. Jour. Anat., vol. 8.

Baetjer, W. a. 1908 On the origin of the mesenteric sac and the thoracic duct in the embryo pig. Am. Jour. Anat., vol. 3.

Budge, A. 1880 Ueberein Canalsystem im Mesoderm von Hlihnerembryonen. Arch. f. Anat. u. Phys.

1887 Untersuchungen liber die Entwicklung des Lymphsystems beim Hiihnerembryo. Arch. f. Anat. u. Phys.

Clarke, E. R. 1909 Observations on the living growing lymphatics in the tail of the frog larva. Anat. Rec, vol.8.

Emery, C. 1880 Fierasfer. Studi interno alia sistematica, I'anatomia e la biologia delle specie mediterranee di questo genere. Atti della Reale Accademia dei Lincei.

Favaro, Giuseppe. 1905 Note fisiologiche intorno al cuore caudale dei Murenoidi (Tipo Anguilla vulgaris, Turt.). Estratto dall 'Archivio di Fisiologia.

1906 Ricerche interno alia morfologia ed alio sviluppo dei vasi, seni e cuori caudali nei Ciclostomi e nei Pesci. Atti del Reale Istituto Veneto di Scienze, Lettere ed Arti. Venezia.

FoHMAJsjN, V. 1827 Das Saugadersystem der AVirbelthiere. Heidelberg und Leipzig.

Greene, C. W. 1900 Contributions to the physiology of the California hagfish, Polistotrema stouti: 1. The anatomy and i)hysiology of the caudal heart. Am. Jour, of Physiol., vol. 3.

Heuer, G. 1909 The development of the lymphatics in the small intestine of the pig. Am. Jour. Anat., vol. 9.

HocHSTETTER, F. 1887 Beitriigc zur vergleichenden Anatomic und Entwickelungsgcschichte des Venensystems der Amphibien und Fische. Morph. Jahrb. Bd. IS.


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS 43

Hopkins, G.S. 1893 The lymphatic and entericepithelium of Amiacalva. The Wilder Quarter-Century Book. Ithaca, N. Y.

HoYER, M. H. 1904 UeberdieLymphherzenderFrosche. Bulletin international de Tacademie des sciences de Cracovie., no. 5.

1905 Untersuchungen iiber das Lymphgefiisssystem der Froschlarven. 1. Teil. Bulletin international de I'academie des sciences de Cracovie.

1908 Untersuchungen iiber das Lymphgefasssystem der Froschlarven. 2. Teil. Bulletin international de I'academie des sciences de Cracovie.

Huntington, G. S. The phylogenetic relations of the lymphatic and blood vascular systems in vertebrates. Anat. Rec, vol. 4, no. 1.

Huntington and McClure. 1908 The anatomy and development of the j ugular lymph sacs in the domestic cat. (Felis domestica). Anat. Rec, vol. 2.

Hyrtl, J. 1843 Ueber die Caudal und Kopf-Sinuse der Fische, und das damit zusammenhangende Seitengefass-System. Archiv ftir Anatomie und Physiologie.

Jones, T. W. 1868 The caudal heart of the eel a lymphatic heart. Philosophical Transactions of the Royal Society of London.

JossiFOV, S. M. 1906 Sur les voies principales et les organes de propulsion de la lymphe chez certains poissons. Archives d'anatomie microcopique, Tome 8.

Jourdain, S. 1881 Sur les sacs sous-cutan^s et les sinus lymphatiques de la region cephalique dans la Rana temporia. Comptesrendushebdomadaires des Seances de I'Academie des Sciences. Paris.

Kellicott, W. E. 1905 The development of the vascular and respiratory systems of Ceratodus. New York Academy of Sciences. Memoirs, vol. 2, Pt. 4.

Klinckowsteom, a. 1908 Beitrage zur Kenntnis des Verlaufes der Darm-und Lebervenen bei Myxine glutinosa. Verhandlungen des biologischen Vereins in Stockholm.

Knower, H. McE. 1908 The origin and development of the anterior lymph hearts and the subcutaneous lymph sacs in the frog. Anat. Rec.

Langer, C. 1866 Ueber das Lymphgefasssystem des Frosches. Sitzungsberichte Akademie der Wissenschaften. Wien.

Lewis, F. T. 1905 The development of the lymphatic system in rabbits. Am. Jour. Anat., vol. 5.

McClure and Silvester. 1909 A comparative study of the lymphatico- venous communications in adult mammals. Part 1. Anat. Rec, Vol. 3, no. 10.


44 WILLIAM F. ALLEN

McKenzie, T. 1884 The blood-vascular system of Amiurus catus. Proceedings of the Canadian Institute.

Marcus, H. 1908 Ueber intersegmentale lymphherzen nebst Bermerkungen iiber das Lymphsystem. Morph. Jahrb., Ed. 38.

Mater, P. 1888 Ueber Eigenthtimlichkeiten in den Kreislaufsorganen der Selachier. Mitthcillungen aus der zoologischen Station zu Neapel, 8 Bd.

MtJLLER, J. 1833 On the existence of four distinct hearts having regular pulsations connected with the lymphatic system in certain amphibious animals. Phil. Trans.

1839 Ueber das Gefasssystem der Fische. Abhandl. d. Berlin. Akad. Also vol. 4, Vergleichende Anatomic der Myxinoiden, Berlin, 1841.

Owen, R. 1866 Anatomy and physiology of vertebrates. London, vol. 1.

Panizza, B. 1834 Ueber die lymphherzen der Amphibien. Arch. f. Anat. u. Physiol.

Parker, J. T. 1886 On the blood-vessels of Mustelus antarticus: A contribution to the morphology of the vascular system in the Vertebrata. Philosophical Transactions of the Royal Society of London.

Priestly, J. 1897 An account of the anatomy and physiology of thebatrachian lymph hearts. Jour, of Physiol.

Ranvier, L. 1897 Morphologic et d6veloppement des vaisseaux lymphatiques chez les mammiferes. Archives d'anatomie microscopique.

Retzius, Gustaf. 1895 Biologische Untersuchungen. II. Das hintere Ende des Riickenmarkes und das Caudalskelet der Myxine glutinosa. Jena. Neue Folge 7.

Sabin, Florence R. 1902 On the origin of the lymphatic system from the veins and the development of the lymph hearts and thoracic duct in the pig. Am. Jour. Anat.

1909 The lymphatic system in human embryos with a consideration of the morphology of the system as a whole. Am. Jour. Anat., vol. 9.

Sala, L. 1900 Sullo sviluppo dei cuori linfatici e dei dotti toracici nell'emljrione di polio. Ricerche fatte nel lab. di Anat. norm. Roma, vol. 7.

Sappey, p. C. 1880 Etudes sur I'appareil mucipare et sur le systeme lymphatiquc des poissons. Paris.

Stannius, H. 1854 Handbuch der Anatomic der VVirbelthicre. Berlin.

Silvester, C. F. 1904 The blood-vascular system of the tile-fish, Lopholatilus chama)leonticeps. Bulletin of the Bureau of Fisheries, vol. 24.


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS 45

Trois, E. F. 1878-1880 Contribuzione alio studio del sistema linfatico dei Teleostei. I. Lophius piscatorius, II. Uranoscopus scaber, III. Rhombus maximus e. R. lavis. IV. Vario Pleuronectidae. Atti del R. Istituto Veneto di scienze, lettere, ed arti .

VoGT UND Yung 1894 Lehrbuch der praktischen vcrgleichenden Anatomic. Braunschweig, Bd. 2, pp. 526-531.


46


WILLIAM F. ALLEN


EXPI. AXA'I'IOX OF FIGT'RES

1 iiiul 1 lU wcic drawn to a scale from injected dissect ions of Scorpa^nichthj'^s murmoratus tails. Fig. 11 is a microscopic view of a portion of the tail of a 30 mm. living Phanerodon atripcs (Viviparous perch) embyro, and figs 3 and 12 are from transverse sections taken through the caudal peduncle of a 28 mm. adult. C'linocottus analis. In general the subcutaneous or lymphatic canals are drawn in bhiik, the veins are cross-barred, and the arteries are stippled or draAvn in outline. A vessel represented in dotted outline signifies that it passes within or on the opjiosite side of a muscle, bone, or other vessel. All microscopical outlines were made with the aid of a camera lucida and the details were filled in afterward.


LIST OF ABBREVIATIONS USED IN THE FIGURES

A. or p. prefixed to an abbreviation signifiies anterior or posterior; R. or L. right or left; the vertebrae are numbered from caudad to cephalad, and the caudal fm ravs are numbered from the centre of the fin, dorsad or ventrad.


A.


Ex


. M.


A.


R.



A.


R.


A.


A


.R.


T.


A.


R.


V.


C.


A.



C.


A.


(1)


C.


A.


'. (1)


C.


F.


A.


C.


F.


A'.


C. F. A".

C. F. A. (1) C. F. A'. (1)

C. F. A". (1)


c.


F.


N.


c.


F.


V.


c.


F.


v.


c.


F.


V"


C. In. M.


Anal tin extrinsic muscle.

Anal ray.

Anal i-ay arterj-.

Anal ray lJ^nphatic trunk.

Anai ray vein.

Caudal artery.

Minor caudal arterJ^

Branch of minor caudal artery in the ha:'mal canal.

Caudal fin artery.

Dorsal caudal fin artery.

Ventral caudal fin artery.

Minor caudal fin artery.

Dorsal minor caudal fin artery.

Ventral minor caudal fin arter3^

Caudal fin nerve.

Caudal fin vein.

Dorsal caudal fin vein.

Ventral caudal fin vein.

Caudal fin intrinsic muscle.


C.R. C. R. A. C.R. A. (1)

C.R. T.

C.R. V. C. S. C. T.

C. T.

C. T. (1)

C. T. (2j


C. V. Der.

D. Ex. M.

D. L. A. D. L. V. D.M. A. R.

D. M. D. R.


Caudal ray.

Caudal ray artery.

Minor caudal ray artery.

Caudal ray Ij'inphatic trunk.

Caudal ray vein.

Posterior caudal sinus.

Caudal lymphatic trunk.

Caudal lymphatic

trunk.

Dorsal or superior caudal fin lymphatic trunk.

Ventral or inferior caudal fin lymphatic trunk.

Caudal vein.

Dermis.

Dorsal fin I'xtrinsic muscle.

Dorsal lateral artery.

Dorsal lateral vein.

Depressor muscle of an anal ray.

Dei)ressor muscle of a dorsal rav.


LYMPHATICS IX TAIL REGION, SCORP.ENICHTHYS


47


D. R. D. R. A. D. R. T.

D. R. V. D. T.

Ep.

f.

Hae. A. Hae. S. Hae. T.

Hae. T. (Ij & {2)


Hge. V. Hyp. (1)

Hyp. (2)


Hyp.


A.


Hyp.


T.


Intm


. T.


L.


A.



L.


C.


A.


L.


C.


V.


L.


Hse. T.


L.


L.



L.


M,


. A. R,


L.


M,


, D. R


L.


Neu. T.


L.


T.



L.


T.


(1)


Dorsal ray. Dorsal ray artery. Dorsal ray lymphatic

trunk. Dorsal ray vein. Dorsal subcutaneous lymphatic trunk. Epidermis. Fascia.

Ha?mai artery. Haemal spine. Ha;mal lymphatic

trunk. Last two ha;'mal lymphatic trunks or ventral caudal fin lym]>liatic trunks. Hicmal vein. Dorsal or superior hy pural bone. Ventral or inferior hy j)ural bone. Hypural artery. Hj'purul lymphatic trunk.

Intermuscular or transverse lymphatic trunk. Lateral artery. Left fork of the caudal

artery. Left fork of the caudal

vein. Longitudinal haemal

lymphatic trunk. Lateral line canal. Levator muscle of an

anal ray. Levator muscle of a

dorsal ray. Longitudinal neural lymphatic trunk. Lateral subcutaneous lymphatic trunk. Posterior portion of the lateral lymphatic trunk.


L. T. (2) L. T. (3)


M. F.

My.

Myo.

Net.


Neu. A. Neu. Ar. Neu. S. Neu. T.

Neu. T. [1]


Neu. V. Neu. & L. A.


Neu. & L. V.

P. Hae. A.

P. Neu. A.

P. Neu. T.

P. Neu. V. P. Ver.

P. Ver.


Profundus portion of the lateral lymphatic trunk. In Clinocottus and Ophiodon communication between the posterior portion of the lateral lymphatic trunk and the posterior caudal sinus (Figs. 2 and 3). Muscule fibers. Meylon or spinal cord. Myotomes.

Fan-shaped lymphatic network on fascia of Ophiodon (Fig. 2). Neural artery. Neural arch. Neural spine. Neural lymphatic

trunk. Next to the last neural lymphatic trunk or dorsal caudal fin I y m p h a t i c trunk. Neural vein. Combined neural and lateral artery (Figs. 9 and lOj C^ombined neural and lateral vein (Fig. 9). Posterior haemal or inferior hypural artery. Posterior neural artery. Posterior neural lymphatic trunk. Posterior neural vein. Posterior or caudal

vertebra. Posterior or caudal vertebra.


48


WILLIAM F. ALLEN


R.


c.


A.


Right fork of the caudal artery.


R.


c.


V.


Right fork of the caudal vein.


Ve


r.



Vertebra.


V.


L.


A.


Ventral lateral artery.


V.


L.


V.


Ventral lateral vein.


V.


T.



Ventral subcutaneous lymphatic trunk.


X.




In fig. 8 branch of the caudal artery to the tail musculature.


z.


1-14.


In fig. 3 this line denotes the beginning of the caudal fin in Clinocottus analis.

In figs. 9 and 10 posterior vertebra; are numbered from caudad to cephalad.


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS


49





VT . : CI,

UtmT l-T« 



RCV.


LT...J LT-w


LX,


L.N...T.



1. Represents a general superticial lateral view of the tail of a median sized Scorpffinichthys as viewed from the left side when the skin only is removed. The origin of the lateral lymphatic trunk from the surface of the myotomes and the posterior portion of the lateral lymphatic trunk from the fascia and the intrinsic muscles of the caudal fin is clearl}' shown. XI. Reduced 1

2. Is a somewhat similar dissection of a median sized Ophiodon elongatus tail as seen from the right side. Introduced mainly to illustrate the peculiar fanshaped network of the posterior portion of the lateral lymphatic trunk on the fascia. X 2. Reduced ^

3. Dorsal view of a diagrammatic reconstruction of the longitudinal neural, lateral, and cau<lal lymphatic trunks taken from a transverse series of sections through the tail of a 28 mm. adult Clinocottus analis. This species differs from Scorpainichthys in having the posterior portions of the lateral lymphatic trunks in direct communication with the posterior caudal sinus. X 50. Reduced i


THE .\MERIC.\X JOUUNAI, OF ANATOMY, VOl,. 11, NO. 1.


50


WILLIAM F. ALLEN



-*^


1


>1


M


«|N



11


t4


<i)





"o


o


0)


« 


^^


d


CJ


JS


-!


-.*


rt


3


■^



' i— J


h


T-l




3


^


O


S


!



Si



"3



t-H


1


rt


(-•


1



CO




3


X


c3


^


-tJ


03


a


>>




tn



aJ


-M


fi>



iH


-tJ


«J


rl



Tl


^


^


+i


T3



.H o

Pi <u


o o


03 e3


(U


■*^ c3

'-' ^

O X2

a ^

.2 53

o o

^ a;


M c3

O

S-i

K « 

c -g

'S c3

^ a

a « 

y5 -— <^

-1

o — (

r-; 03

o3 U

cu ;> >

r-, <V ^ r^

g_1 Cm

o3 O


o o

§.2


o

-^

3

2 2

Si ,(_

""^ c M

S 2

■^ q;

r-" 03

  • ^ CO

14-1 (U

O ^

2 2


J3 CO

o aj

>> a

03 M

« o

^ §

c3 S

OJ a

A o

o

li

c


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS


51



crv •

CFA-CFA'


crV



PN«,T


GA


CT LH-T: PH«/V. CT


GS : C'T*

C-F-V-'


CRT


GR


5. A deep dissection of the base of the tail of a large Scorpsenichthys as seen from the left side. With the exception of the posterior portion of the lateral trunk all of the so-called subcutaneous or lymphatic vessels are sketched. The left or main fork of the caudal vein is cut immediately behind the last vertebra. A small portion of each of the dorsal and ventral caudal fin veins are shown. Each of the caudal arteries are cut close to their origin ; while the minor caudal artery is severed between the hypurals shortly before reaching the caudal sinus. X 2. Reduced ^

6. Similar view taken from the right or opposite side of the same specimen as fig. 5. X 2. Reduced §

7. From a deep dissection of the dorsal portion of the base of a large Scorpaenichthys tail as seen from the left side. Introduced mainly to illustrate a different termination of the caudal ray trunks in the caudal fin trunk than is shown in fig. 5. X 2. Reduced i


52


WILLIAAI F. ALLEN


GR-A-t/


PNeuA'


CRA


CR-A


CA PN?u-A iC-V TVer.

LA LV—

L-V.-.

HxeV

Neu.^L-A

HaeA

I\lei,.J>.L.A.HaeA- ■

Neu.i.LV. HaeV LV-.

LA HaeA- Neu-VL A HacA

MeuiLV- «==^ HoeV

LAHoeA


R-CA RCV — P.HmA


LV Hae-ANeu-&.LA


LV


1



-Neu-5.LA Hae-A ■OA


CT

CA-, •

P-Neu-A- -••^Ir RCA

LHaeT



Neu.5,.L-A'


8. Diagrammatic drawing showing the distribution of the caudal and the minor caudal arteries as found in the tail of a large Scorpienichthj's viewed from the left side.

9. Ventral view of the ha'inal canal of a largo Scor])a'nichthys; the body musculature and the haemal spines being completely removed. Designed to show the relation of the neural arteries to the neural veins and to the corresponding hajmal vessels, and also the dorsal intersegmental vessels that liave both dorsal and lateral branches to those that have only lateral branches. The vertebrae are numbered from caudad to cephalad. XI. Reduced |

10. Deeper dissection of the same specimen as fig. 9, the caudal vein and its branches being entirely removed . It shows the relation of the minor caudal artery to the caudal artery and the branching of each, together with a little of the ha>mal lymphatic or subcutaneous system. XI. Reduced ^


LYMPHATICS IN TAIL REGION, SCORP^NICHTHYS


53



L. Hae-T.


Hoe -Til)

-VT


11. Portion of the tail of a 30 mm. living embryo Phanerodon atripes (Viviparous perch) as seen in a watch glass of sea water under a microscope . The blood corpuscles were plainly seen passing out of the caudal.artery into the dorsal and ventral caudal fin arteries, thence into the caudal fin ray arteries, Irom which they entered a minute capillary network lying in the membrane connecting two rays. On the opposite side these networks were collected by corresponding caudal fin ray arteries, which emptied into the dorsal and ventral caudal fin veins, the latter uniting to form the main caudal vein. X 33. Reduced |

12. Transverse section through the caudal peduncle region of a 28 mm. adult Clinocottus analis. All of the main longitudinal trunks are seen in section. The posterior neural and haemal trunks are also seen in section midway between the longitudinal neural and the dorsal trunks, and between the longitudinal hajmal and the ventral trunks. X 25. Reduced |


THE SOMITES OF THE CHICK

LEONARD W. WILLIAMS From the Harvard Medical School, Boston

NINETEEN FIGURES

The first consistent account of the history of the somites of the chick, in which was presented the much discussed theory of the resegmentation of the vertebral column, was published by Remak in 1855. He believed that the somites are originally hollow cubical masses of cells which, as the medullary groove deepens, become triangular prisms with dorsal, medial, ventral and end walls. The medial ventral edge of each somite elongates and, reaching the notochord, divides into two leaf-like processes which, uniting with those of the opposite side, grow around the notochord and form the tissue of the perichordal sheath or Wirbelkorpersdule. From the same edge, he believed, there grows into the ca\ ity of the somite a mass of cells, the core Urwirbelkern, which greatly reduces the cavity. The core of the somite soon fuses with the neighboring walls with the exception of the dorsal wall. Each somite is now divided into an epithelial or epithelioid upper wall and a mesenchymal mass formed by the fusion of the core and walls of the somite. Remak named the former the Riickentafel or Muskelplatte. The latter, the sclerotome, he named the Wii^belkernmasse. There soon appears a contrast between the anterior and posterior parts of the sclerotome. The spinal nerve with its ganglion and roots appears in the anterior portion of the sclerotome from which Remak believed that it arose. The posterior part becomes condensed forming what Remak and others have called the vertebral arch. The correctness of this interpretation will be discussed later. Remak believed that toward the end of the fourth da} this vertebral arch is pushed backward so that its posterior edge is covered by the dorsal lamella or Rilcken THE AMEUICAN JOURNAL OF ANATOMY, VOL. II, NO. 1.


56 LEONARD W. WILLIAMS

lafel of the following somite, to which it becomes attached. Since each arch retains its connection with the dorsal lamella of its own somite, it is now attached to the dorsal lamellae of two adjacent somites. There now forms in the perichordal sheath at the center of each segment, a conspicuous condensation which is separated from the condensations of the adjacent segments by light transverse zones, which Remak believed were clefts. These condensations, however, do not correspond with the vertebral centra and, to distinguish them from the latter, Remak called them the primitive vertebral centra.

During the fifth and sixth days, a striking change occurs in the perichordal sheath. This consists in the apparent resegmentation, Neue Gliederung, of this sheath, by which the definitive vertebral centra are formed. Since the primitive centra arise from the whole medial ventral edge of the somite, and the vertebral arch comes from the posterior part of each somite, the arch is consequently attached to the posterior part of the primitive centrum. The spinal nerve and ganglion occupy a space between two vertebral arches which corresponds with the anterior part of the primitive centrum. However, in a later stage the relation between these structures is reversed: the arch is now attached to the anterior end of the centrum and the ganglion lies in a space which corresponds with the posterior end of the centrum. This change in the position of the nerve in relation to the vertebral arch is the basis of Remak's assertion that there is a resegmentation of the vertebral centra, or more accurately of the perichordal sheath (Wirbelkorpersdule.) It will be remembered that Remak worked with whole or dissected embryos, consequently it is not surprising that he mistook the primary perichordal condensations which form the intervertebral ligaments for the vertebral centra. This error was corrected by Gegenbaur ('62) who, while accepting Remak's theory as a whole, introduced certain modifications. Unfortunately, conceiving the sclerotome or Wirbelkerjimasse to be identical with the primitive vertebral centrum of Remak or the intervertebral condensation, Gegenbaur introduced some confusion which is scarcely yet cleared away, but he rightly maintained that the apparent spaces which separate successive midsegmental


THE SOMITE8 OF THE CHICK 57

or intervertebral condensations are really masses of loose tissue which become the vertebral centra. Gegenbaiir also saw that in birds and reptiles (as is also true in mammals) the tissue around the notochord forms a membranous, skeletogenous" or "perichordal" sheath which is replaced by a cartilaginous or precartilaginous sheath, which Minot named the chondrostyle. Resegmentation is effected by the formation of joints in the chondrostyle.

In 1868 His published a new account of the somites of the chick. He found that each somite is primarily a flat quadrangular body consisting of a core and cortex; the former is a small cluster of irregular rounded cells; the latter is composed of fusiform radiating cells attached to one another only at base. Each cell bears peripherally a free projecting process. The more posterior somites differ in several respects from the anterior or first formed somites. They are cubical, or nearly so, their walls have more epithelial characters, and their cores are larger. His like Remak believed that in the chick the dorsal lamella is entirely converted into voluntary muscle, and Bardeen maintains that this is also true in the pig; nevertheless, this view has not been generally accepted. The credit of showing that the spinal ganglia arise from the neural crest, not from the sclerotome, belongs to His but he failed to see that the sympathetic ganglia arise in the same manner. He believed wrongly that the core of the somite forms the sympathetic ganglion and that the ventral wall of the somite forms only the muscular coat of the aorta. All connective tissues, according to his now abandoned parablast theory, arise from the extraembryonic mesoderm and migrate along the blood vessels into the spaces between the entoderm, ectoderm, neural tube, notochord, and somites. Consequently His believed that Remak's theory of resegmentation of the vertebral column was wholly without foundation. Goette, however, in 1875 pointed out that the tissues of the vertebral column really do arise from the sclerotomes.

Froriep ('83) found in each segment a lyre-shaped mass of dense mesenchyma, which Bardeen has recently named the scleromere. This extends from a midsegmental point below the


58 LEONARD W. AVILLIAMS

uotochord laterally, dorsally and backward to the following intersegmental fissure. Its central portion (the {)riinitive vertebral centrum of Remak) soon encircles the notochord and then differentiates into a fibrous ring, surrounding the notochord, and a transverse bar of cartilage. The former becomes the intervertebral disc, or takes part in the formation of the intervertebral joint; the latter, the intercentrum or subnotochordal bar {hypochordale Spnnge) usually degenerates without first losing its connection with the lateral parts of the scleromere, but in connection with the first two cervical vertebrae it persists, forming the body of the atlas and a part of that of the axis. The vertebral centrum, arising in the loose tissue between the midsegmental condensations, fuses, Froriep believed, with the preceding scleromere, forming in this way the definitive vertebra. The centrum of the atlas, however, does not fuse with the preceding but with the succeeding scleromere, that of the axis. The only criticism I have of Froriep 's work is the one made in a former paper, namely, that the scleromere which ultimately gives rise to the intervertebral disc (or ligament), the intercentrum, the neural arches, the ribs, and the myoseptum, cannot be regarded as a morphological unit. The only actual units with which we are here concerned are the centrum, the neural arches, the ribs, and the intercentrum. Froriep's description of these and of their relation to the definitive vertebrae is correct.

Remak's theory received new support from Von Ebner ('88) whose discovery of a midsegmental diverticulum of the cavity of the somite, which divides the sclerotome into essentially equal anterior and posterior parts, reopened the whole question of vertebral formation. Von Ebner found this fissure, which he named the intervertebral fissure, in the lizard, chick, mouse, and bat. Schultze ('96) claimed that in birds the fissure arises independently of the cavity of the segment and forms a connection with it later, but in mammals is entirely without connection with it. I have pointed out elsewhere that there is really no fissure in mammals and the same is true, I find, in birds. Von Ebner, like Gegenl)aur, thought that the sclerotome is the structure which Remak called the primary vertebral centrum; consequently the fissure


THE SOMITES OF THE CHICK 59

seemed to him to be the only evidence needed to prove the correctness of Remak's theory. The anterior half of one sclerotome he believed fuses with the posterior half of the preceding sclerotome to form an intersegmental vertebra.

This view was attacked by Corning ('97) who pointed out the weakness of Von Ebner's position. He maintained rightly that a midsegmental vertebral centrum does not exist and that the sclerotome is not the primitive vertebral centrum of Remak This criticism elicited a reply from Von Ebner ('92) which contained the fundamental truth that the dense mass of tissue forming the greater portion of the posterior part of the sclerotome IS a composite structure. The primitive arches," i.e., the lateral portions of the scleromeres, are merely segmental structures which contain the anlagen of several diverse structures.

A very different conception of the structure of the vertebral column is that of Goette who in a series of papers, particularly one m 1896, presents the theory that each segment contains primarily the anlagen of one haemal and two neural arches, an intercentrum, and a centrum. In the tail of a well advanced embryo of Lacerta, there is a gradual transition in the structure of the neural arch. Anteriorly each half of the arch is a broad plate of bone; in the middle of the tail it is divided by a deep vertical groove, or by a narrow sht, into a large anterior and a small posterior arm; and still farther back it is represented by two bars of bone, of which the anterior is the larger. In adult lizards, however, the neural arch of every caudal vertebra is an undivided broad plate. The transverse processes of the caudal vertebrae of Lacerta, and also of certain mammals, show a similar tendency to divide into anterior and posterior portions. The dense tissue in the middle of each segment represents the intercentrum and the loose tissue between every two intercentra and between the right and left intersegmental arteries forms the primary centrum. The bases of the neural arches broaden and fuse with the primary centrum, forming the secondary centrum. The haemal arch extends from below the intercentrum backward and downward into the intermyotomic septum. It may fuse with the posterior end of the preceding vertebra, as it does in Lacerta, or with the anterior


00 LEONARD W. WILLIAMS

end of the following centrum. Goette believes that these elements, namely the centrum, the intercentrum, the two pairs of neural and the single pair of haemal arches, unite with one another in several ways with the abortion or enlargement of certain units so as to form the various types of vertebrae occurring in Digitates and in the Amiidae among fishes!

Manner ('99) found two pairs of cartilaginous neural arches in Angius and Lacerta; and Schauinsland ('00) discovered two pairs of cartilaginous neural arches in the caudal region of embryos of Hatteria. In embryos which were about to hatch, he found conditions similar to those described by Goette in Lacerta. However, in Hatteria not only are the neural arches and transverse processes double, but the centrum also shows indications of the same structure because, as Schauinsland believes, it arises from parts of the sclerotomes of two adjacent segments. Goette, however, does not maintain that the centrum is double, but that, corresponding to each segment, there are two axial elements, — the centrum, which is intersegmental, and the intercentrum, which is midsegmental in position. Schauinsland regards the subnotochordal bar (hypochordale Spange) of Froriep as the haemal arch.

After Remak had shown that the Urwirbel or protovertebra contains anlagen of the body musculature as well as of the vertebrae, the name became inappropriate; nevertheless, it was used without question until Goette in 1875 proposed to use the term segment. A few years later in the second edition of Foster and Balfour's "Textbook of Embryology" ('83) the term mesoblastic somite, or somite, was used as a substitute for the name protovertebra. It seems best to denote by the word somite one of the blocks of mesoderm formed by the segmentation of the vertel)ral or somitic plate and to make the segment include a pair of somites with the corresponding nephrotomes.

While Von Ebner was discussing the intervertebral fissure and resegmentation, Kabl ('88), demonstrated that the dorsal lamella or muscle plate of Remak becomes a two-layered plate, the Haiitmuskelplaiie (dermomyotome) and that the inner layer only, the Muskellamelle (myotome) forms muscle, whereas its outer layer.


THE SOMITES OF THE CHICK 61

the CutislameUe (dermatome) is converted into the connective tissue of the dermis. At the same time Hatschek proposed to call Remak's Wirbelkernmasse the sclerotome.

In the same year Paterson ('88) discovered that the ventral edges of the dermomyotomes of the trunk of the chick grow downward into the memhrana re^iidens inferior and so supply the musculature of the abdominal and thoracic walls, but that the musculature of the limbs does not arise directly from the dermomyotomes. Paterson saw that the dermomyotome is composed of two lamellae but did not discover that the outer lamella contributes to the dermis.

Kaestner ('90) upon insufficient grounds attacked the work of Rabl and Paterson. He believed that some of the cells of the cutis plate become myoblasts and that the remainder form an epithelium which is destroyed during the third day by parablastic tissue. He further maintained that the Muskelknospen, or growing ventral edges of the dermomyotomes, described first by Paterson, contribute directly to the musculature of the limbs.

Kollmann ('91) describing the somites of human embryos, asserted that whereas the muscle plate, or inner layer of the dermomyotome, supplies the dorsal musculature of the trunk, the cutis plate gives rise to the ventral body musculature and to that of the limbs. He admitted, however, that dermal connective tissue arises from the myotomes (dermomyotomes?) of the trunk.

Fischel ('95) took a somewhat intermediate position between Rabl and Paterson on the one hand and Kaestner and Kollmann on the other. He pictured the growing dorsal and ventral edges of the dermomyotome in the chick and showed that the prebrachial dermomyotomes do not have ventral growing edges. The ventral buds or growing edges of the dermomyotomes of the trunk, he thought, break up into loose tissue which mingles with the mesenchyma of the nephrotome and lateral plate. The muscles of the limbs and of the ventral part of the body wall arise in the mass of tissue formed in this way, and Fischel judges from analogy that these muscle masses probably arise from the dermomyotome.


62 LEONARD W. WILLIAMS

Engert ('00) in the chick and Bardeen and Lewis ('01) in human embryos, proved beyond reasonable doubt that Rahl and Paterson were in the right and that the ventral body musculature does arise from the ventral part of the myotome or muscle lamella but that the muscles of the limbs arise independently of the myotome. Engert shows that all of the cutis plate except its extreme upper and lower edges forms dermal mesenchyma and that these edges become transformed en masse into myoblasts forming the much thickened dorsal and ventral edges of the myotome. Bardeen ('00) however, maintained, I believe wrongly, that the cutis plate in t\]e pig gives rise only to myoblasts.

My intention, when I began this paper, was to study the later development of the notochord of the chick and the relation of the notochord to the vertebrae, as a continuation of the work upon the notochord published in 1907. I soon found, however, that more knowledge of the structure and development of the somites was necessary. I have therefore followed with great care the history of one of the two somites of the second segment up to the time of its transformation into the sclerotome, myotome, and dermatome. The differences at the time of their origin between the second and several other segments have been pointed out. A brief account of the history of the tenth segment, and of the relation of its sclerotomes to the vertebrae is given, and finally the history of the twenty-fifth and forty-fourth segments is described, in order to emphasize the differences, most of which are well known, between the occipital, cervical, trunk, and caudal segments.


THE SECOND SOMITE

Rabl ('89) in his extensive work upon the early history of the mesoderm has shown that, shortly before the origin of the first segment, the mesoderm of the chick is represented by a sheet of syncytial tissue which extends in all directions from the primitive streak. Centrally it is thick and contains numerous closely packed but irregularly arranged nuclei. Peripherally it gradually becomes thinner until, near the inner edge of the area vasculosa,


THE SOMITES OF THE CHICK 63

it is an exceedingly thin and imperfect sheet, composed of small irregular clusters of stellate cells or cell-like masses of protoplasm. The portion of the mesoderm in front of the primitive streak, which Rabl names the gastral mesoderm, is bisected by the notochord. Small isolated cavities appear in the larger clusters of cells of the antero-lateral portion of the gastral mesoderm and, gradually enlarging, unite with one another, forming on either side the parietal or amnio-cardiac cavity. The portion of the gastral mesoderm below the medullary plate and beside the notochord is thicker than its lateral part, and the first intersegmental cleft appears in this thick medial edge at a point nearly midway between the anterior extremity of the embryo and the primitive streak. This cleft, as is well known, is somewhat V-shaped, beingprolonged on each side of the median line laterally and slightly backward. Patterson ('07) and Miss Hubbard ('08) have proven experimentally that this cleft is the most anterior of the entire series and not, as Von Baer thought, the third from the head. It forms the posterior boundary of the first segment, which, being continuous anteriorly with the unsegmented mesoderm of the head, is incomplete. Each somite of this segment produces approximately half as much muscle and mesenchyma as the following somites. A second intersegmental fissure appears quickly and cuts off the first pair of complete somites, those of the second segment. Each somite of this segment (fig. 1) is a very irregular flattened mass of cytoplasm containing many nuclei, which are rounded or oval and are without definite arrangement. There is observable, however, a very faint indication of a division of the somite into upper and lower layers.

While the third intersegmental fissure is forming, there appears a longitudinal constriction which lies parallel to, and a short distance from the notochord, and divides the gastral mesoderm into a narrow medial zone, the somitic plate, and a broad lateral region, the lateral plate. This constriction extends forward so as to form the lateral boundary of the first and second somites and also extends backward some distance behind the third intersegmental cleft. The somitic plate, like all other portions of the mesodermal sheet, gradually becomes thicker toward the primitive streak.


04 LEONARD \V. WILLIAMS

In embryos of three segments, the second somite (fig. 2) is quadrangular and is considerably thicker than before. It is now distinctly divided into upper and lower layers, which are continuous with one another at the edges of the somite. The longest axes of the oval nuclei radiate from the center of the somite. Its surface is still very irregular.

New somites are constantly forming, and each differs more or less from the others, as will be seen in figs. 10 to 14, which represent transverse sections of several newly formed somites. It seems wisest, therefore, to study the development of the somites of a single segment, the second, and then to compare these with the others.

In embryos of six segments (fig. 3) the edges of the medullary plate are considerably elevated above the second segment and each of the somites of this segment has become an irregular triangular prism with dorso-medial, dorso-lateral, lower, and anterior and posterior surfaces. The first two of these arise from the upper surface of the earlier somite. The dorso-medial or medial surface is concave and is molded against the convex lower surface of the medullary plate. The dorso-lateral (which will soon become the upper) surface is quite irregular. The lower surface is convex. The medial edge of the somite is almost in contact with the notochord and its anterior end is prolonged forward so that the anterior end or wall of the somite slopes from above downward and forward. The somite now contains a flattened core consisting of a few rounded nuclei and a small amount of cytoplasm.

The coelom expands rapidly and has now divided the greater part of the lateral plate into an upper or parietal and a lower or visceral layer. It appears also in the thicker medial edge of the lateral plate, forming a slight expansion (C) and finally sends a small tortuous prolongation through the constricted area or stalk which connects the second somite and the lateral plate. Thus a small prolongation of the coelom enters the somite and separates the dorsal wall of the somite from the core. It is now clear that each somite of the second segment for a time has no central cavity and is without a core. This fact was overlooked by Remak who believed that each newly formed somite has a central cavity, and by His who states that from the first the somite has a core.


THE SOMITES OF THE CHICK 65

His states that the most anterior somites do not contain cavities, but there are cavities in the first three or four somites and the communications between them and the coelom in embryos of from ten to twenty segments were discovered by Dexter ('91). This cavity and communication appear in the second somite of embryos of five or six segments. Bonnet shows that the first four somites of sheep embryos have a similar structure and connection with the coelom.

In embryos of nine segments, the neural tube is closed in the region of the second segment and consequently the shape of each of the second somites (fig. 4) is somewhat altered. Its medial surface is nearly vertical, the upper surface is larger and is nearly horizontal, and the lower surface is marked by a groove for the aorta, which has now appeared and is rapidly enlarging. The cortex or wall of the somite, particularly the dorsal wall, has a very remarkable structure which has been recognized, I believe, only by Held^ w^hose figures show this structure very beautifully and accurately. While having the appearance of a simple or stratified columnar epithelium, the cortex is really an epithelioid syncytium of unusual character. The oval or elliptical nuclei are imbedded in columns of cytoplasm which proximally, i.e., toward the core, are continued into a dense basal layer of cytoplasm and distally end in irregular sparingly branched processes. Certain of these distal processes unite so as to form a faint external boundary of the cortex, but others extend freely into the space between the somite and the ectoderm and other adjacent structures. Nuclei preparing to divide withdraw to the basal layer of cytoplasm, leaving conspicuous gaps in the outer part of the wall. I have been unable to find a single mitotic figure elsewhere in the somitic cortex than in close proximity to the basal layer, and it is interesting to note that practically all of the axes of the mitotic figures are parallel to the surfaces of the cortex. The structure described is quite unmistakable in the first five or six somites which have relatively few cells, but posteriorly it is somewhat obscured by the much greater number of cells in the cortex.

' Figures 91-96 of birds, 98-99 of the rabbit. The American Journal of Anatomy, Vol. 11. No. 1.


66 LEONARD W. WILLIAMS

I believe, however, that the cortex of all the somites of the chick and, probably, of all birds and mammals, is an epithelioid syncytium of this peculiar type.

The upper wall of the somite is thin laterally where it is continuous with the upper layer of the lateral plate. Medially it gradually becomes thicker, its medial edge being nearly twice as thick as its lateral edge. The medial wall of the somite immediately underlying the thick edge of the upper wall is quite thin and is indented by a small groove-like evagination of the cavity of the somite, (fig. 4, G.U.), for which I propose the name upper myotomic groove. The appearance of this groove is the first indication of the formation of the myotome. The cells at its base apparently change from a cylindrical to an oval or spherical form in preparation for their subsequent longitudinal elongation as they become definitely recognizable myoblasts. The floor and the lower part of the medial and anterior walls have fused with the core of the somite to form a mass of mesenchyma, the sclerotome, or since it subsequently receives a very considerable addition from other parts of the cortex, the primary sclerotome. The fusion between the core and cortex begins first at the anterior end of the ventro-medial edge of the somite and, owing partly to the expansion which accompanies the transformation of the tissue of the somite into mesenchyma, this angle grows forward and medially more rapidly than the remainder of this edge of the somite.

In embryos of twelve segments, a second groove appears (fig. 5, G.L.) in the medial wall of the second somite. This, however, is an invagination of the wall and lies at a slightly lower level than the upper myotomic groove. Since it marks the lower edge of the muscle plate, I propose to call it the lower myotomic groove.

The upper myotomic groove, I believe, has never been described or figured. It is, however, neither as constant nor as conspicuous as the lower groove which, although often figured, has not been described or named. A cord of vascular cells partly fills the lower myotomic groove.

The myotome is represented by a narrow zone between the two myotomic grooves which differs considerably from the


THE SOMITES OF THE CHICK 67

adjacent portions of the cortex of the somite. The cells or celllike elements of the myotome are shorter and are more closely packed than those of the rest of the cortex. Moreover, the cells of the myotome radiate from the upper myotomic groove, upward, dorso-medially, medially, and ventro-medially. The upper wall of the somite is somewhat thicker and its cells are more closely placed than before. Its medial portion is overlaid by a plate of neural crest cells.

The communication between the coelom and the cavity of the somite is larger than in younger embryos and is approximately circular in cross section, i.e., as seen in sagittal series. It has now reached perhaps its greatest development. The stalk of the segment is deeply constricted both on its upper and its lower surface, in front and behind the communication.

The sclerotome is slightly larger, owing partly to the transformation of more of the anterior and medial walls of the somite into mesenchyma. The posterior wall and the posterior ventral edge of the somite are still epithelioid and have not contributed to the formation of the sclerotome. The amount of expansion of the sclerotome is shown in the model (text fig. 1) and can be indicated precisely by the example of the second somite of an embryo of thirteen somites whose sclerotome in sagittal section is longer by nearly one-fourth than the dorsal wall, and is one-tenth longer than the distance between the centers of the adjacent intersegmental clefts. The aorta divides the lower part of the sclerotome into two keel-shaped processes, the aortic or lateral (fig. 5, P. A.), which projects downward and medialty between the aorta and the pharynx, and the notochordal or medial process (fig. 5, P.N'.), which, projecting toward the notochord, separates the aorta from the neural tube.. Fig. 119 (p. 204) in Minot's Human Embryology" shows that there is a similar projection forward of the lower part of the sclerotome of the rabbit.

In emhryos of fifteen segments (fig. 6) the most conspicuous alteration of the second somite as compared with that of embryos of twelve segments, is the deepening of the lower myotomic groove and its extension upon the anterior and, to a slight extent, upon the posterior surface of the somite. The groove is deep and nar


68 LEONARD W. WILLIAMS

row anteriorly but is broad and shallow posteriorly. The muscle plate underlies the medial fourth or fifth of the upper wall of the somite and its nuclei are gradually becoming elongated in the longitudinal plane, instead of as before in the transverse plane. The boundaries of the cells of the muscle plate are indistinct but are probably present.

The further description of this somite requires a study of the adjacent blood-vessels which have been best described by Evans. His figure ( '09, 2, fig. 3b) of the blood-vessels of the head of an embryo of fifteen segments shows that the anterior cardinal vein consists of three portions, namely, a long slender vessel, lying at the side of the neural tube and between it and the unsegmented mesoderm of the head, the vena capitis medialis; a short transverse portion, which I find lies between the first and second somites; and finally an irregular trunk which passes obliquely laterally and backward upon the upper surface of the j^arietal plate to a point opposite the third somite, where it opens into the common cardinal vein or duct of Cuvier. The latter is a minute vertical vessel which passes through the lateral plate in the mesocardium later ale (Koelliker) and joins the vitelline vein. The first indication of the anterior cardinal vein is seen in an embryo of six segments (H.E.C. no. 639). The aortae are just established but are still small and irregular vessels. Each aorta, however, gives off into the first intersegmental cleft a branch of nearly its own size which extends dorso-medially to the side of the neural tube. The aorta and its branches are connected with a delicate cellular network which extends between the somites, between them and the neural tube, and between the medial portion of the parietal plate and the ectoderm. The cells of this network can often be distinguished from the ordinary mesenchymal cells by slight differences in the intensity of the stain. They are shown by their connection with the aorta and by their subsequent history to be vascular cells. Two strands of these vascular cells are of particular and immediate interest; one of these extends from the aorta upward through the second intersegmental fissure where it connects with the second strand which extends from the termination of the aortic branch in the first fissure, forward between the neural


THE SOMITES OF THE CHICK 69

tube and the unsegmented mesoderm of the head and backward between the former and the somites. Both of these strands are quicklj' replaced by blood-vessels; the former in an embrj^o of eight segments (H. E. C. no. 642) is transformed into the first intersegmental artery ; and the latter forms the posterior end of the vena capitis medialis and, behind the first intersegmental cleft, the beginning of a chain of anastomoses between the distal ends of the intersegmental arteries. An apparently isolated vessel represents the beginning of the third or oblique portion of the anterior cardinal vein.

In an embryo of fourteen segments, each aorta bears a dorsolateral branch in the first intersegmental cleft. One of these is connected with a small and somewhat tortuous transverse vessel which, joining the posterior end of the vena capitis jnedialis and the anterior end of the oblique portion of the anterior cardinal vein, forms the transverse portion of the latter vessel. On the other side of the embryo, the posterior end of the vena capitis medialis and the distal end of the dorso-lateral branch of the aorta are bound together only by strands of vascular cells. Similar dorso-lateral branches of the aorta occur in the second intersegmental cleft and are connected with the intersegmental arteries by transverse vessels.

The anterior cardinal veins of two embryos of fifteen segments in the Harvard embryological collection (nos. 1444 and 1460) are in the main like those of the embryo of the same number of segments figured and described by Evans. The transverse portion of this vein in both embryos is connected with the aorta by a quite direct vessel or in one case by two vessels. Both embryos are probably younger than the one studied by Evans for in one (no. 1460) the common cardinal veins are not present, and in the other they are very minute. The second intersegmental cleft on each side contains a single T-shaped branch of the aorta which is apparently formed by the fusion of the preceding vessels. One arm of the T anastomoses with the posterior prolongation of the vena capitis medialis, the other extends upward between the stalks of the second and third somites and then expands in an irregular flattened vessel which, uniting by minute anastomoses with simi


70 LEONARD W. AVILLIAMS

lar vessels in the adjacent intersegmental clefts and with irregular vessels upon the parietal plate, »forms the oblique or third portion of the anterior cardinal vein.

The posterior part of the anterior cardinal vein is formed, as Evans has shown to be the case with certain other veins, from several separate branches of the aorta. The first intersegmental artery gives rise to the posterior part of the vena capitis medialis, leaving that of the second intersegmental fissure to serve as the first definitive intersegmental artery.

The dorsal wall of the second somite of embryos of fifteen segments is almost separated from the lateral plate by the enlargement of the vessels of the network which gives rise to the oblique portion of the anterior cardinal vein. A very narrow bridge, however, connects the dorsal wall of the somite and the parietal plate and forms the roof of the much reduced communication between the cavity of the segment and the coelom.

The floor and the lower part of the walls of the somite fuse with its core and are quickly converted into sclerotomic tissue. The upper part of the walls of the somite, however, has a different history, for instead of fusing with the sclerotome this part of the cortex of the somite forms a center of growth from which much of the mesenchyma not only of the sclerotome but also of the dermis proliferates. Moreover, in addition to mesenchyma this part of the somite also produces the myotome. It is important, therefore, to distinguish the roof of the somite, with its bordering zone of growth, from the sclerotome, and for this purpose Reinak's name Ruckentajel and its English equivalent, dorsal lamella, will serve.

The lower or lateral edge of the myotome merges gradually with the sclerotome at the bottom of the lower myotomic groove, but I cannot find a center of growth here, nor can I determine whether or not there is a migration of cells to or from the myotome.

The plate of neural crest cells is migrating laterally and now overlies the adjacent edges of the somite and parietal plate.

I71 embryos of eighteen segments (fig. 7) one sees a continuation of the several processes described above, namely, the deepening of the lower myotomic groove and its further extension upon the


THE SOMITES OF THE CHICK 71

ends of the somite, the enlargement of the third part of the anterior cardinal vein and the resulting separation of the dorsal lamella and the parietal plate, the migration laterally of the neural crest cells and the proliferation from the edges of the dorsal lamella of cells which contribute to the growth of the myotome and sclerotome. The lower myotomic groov^e, which is now scarcely more than a cleft, and the myotome underlie the medial half of the dorsal lamella and consequently the sclerotome is connected with the dorsal lamella only by a stalk whose diameter is somewhat less than half the length of the somite.

The sclerotome of this segment now begins to fuse with the adjacent sclerotomes, and in all later embryos it is impossible to distinguish the boundaries of the sclerotomes of the first five or six segments. It is important, therefore, to note the relation of the sclerotome to surrounding structures. Dorsally it is bounded by the dorsal lamella. Laterally it is attached to the somatic or parietal and the splanchnic or visceral layers of the lateral plate. Medially it is in contact with the neural tube and the notochord. Its ventral surface is in contact with the aorta and bears the notochordal and aortic processes which come into contact with the entoderm on each side of the aorta. The dorsal part of the anterior surface of the sclerotome is in contact with the transverse portion of the anterior cardinal vein, its ventral part slopes forward and downward so that the notochordal process (Text-fig. 1.) projects forward beyond the plane of the upper part of the intersegmental fissure, one third the length of the somite. The posterior surface of the sclerotome inclines sharply forward and downward so as to allow the notochordal process of the following sclerotome to project forward under the second somite.

The capillary vessels between the sclerotomes and between them and the neural tube have begun to sink into the sclerotomes, and thus the sclerotomic tissue receives its first blood-vessels.

In embryos of tweniy-five segments the dorsal lamella (fig. 8) inclines downward laterally at an angle of about 45° with the median plane. The myotome has extended laterally so as to unite with the greater part of the turned under lateral edge of the


72


LEONARD W. WILLIAMS


dorsal lamella. A small stalk of mesenchyma still binds the sclerotome to the essentially complete dermomyotome and prevents the final fusion of the edges of the myotome and cutis plate or dermatome. The nuclei of the lower part of the myotome are distinctly larger than those of its upper or medial edge. Mitoses are abundant in all parts of the mj^otome.


D.M


O.L.



RA


Text fig. 1. Model of the left somite of the second segments of an embryo of eighteen segments. The medial, anterior and part of the dorsal surface are seen. X 400. H. E. C. no. 146G.


The dermomyotome is barely complete when its disintegration begins. This process, however, is only slightly indicated at this time in a small area just lateral to the center of the cutis plate.


THE SOMITES OF THE CHICK 73

As Rabl and others have shown, the cells of this area begin to separate slightly and to send out free distal protoplasmic processes toward the ectoderm. At the same time, I find that a few cells not in mitosis withdraw to the basal layer of cytoplasm which as before is separated from the muscle plate by a cleft, the remains of the cavity of the somite.

The third portion of the anterior cardinal vein (fig. 8, C.A.) is still farther from the ectoderm and is now separated from the aorta only by a rather thin sheet of mesenchyma. It extends outward only to the level of the upper surface of the muscle plate, and it is separated from the ectoderm by the lower edge of the dermal plate and, ventrally, by a part of the mass of neural crest cells which is forming the ganglion of the vagus. The lateral, anterior, and posterior boundaries of the somite and later of the cutis plate, are indicated by a small acute ridge (fig. 8, R.) upon the inner surface of the ectoderm. This ridge now marks the boundary between the cutis plate (fig. 8, D) and the neural crest cells (N.C.) A few isolated neural crest cells can be seen migrating laterallj' from the roof of the neural tube.

Evans states that "The center of each sclerotome is, on its upper surface, supplied by a sheet of closely anastomosed capillaries; but the outer divisions of the sclerotome are not so supplied. There capillaries are absent for a considerable time, so that the vertebral column presents a succession of vascular and non-vascular zones, the former areas in each case overlying the segmental vessels" ('09, 2, p. 515). This statement does not hold good for the sclerotomes of the four cephalic segments owing doubtless to the fact that the cephalic sclerotomes differ considerably in structure (compare p. 77) from the spinal sclerotomes. The sclerotomic mesenchyma of the head is divided into a vascular outer and upper zone above the aorta and lateral to the intersegmental arteries and a non- vascular zone lying beneath the neural tube and the notochord. Evans' figure (cf. '09, 1, ng. 3) shows that the richest capillary plexus lies in a longitudinal zone at the side of the neural tube and between it and the medial or dorso-medial surface of the sclerotomes. The capillaries of this plexus, like other capillaries around the sclerotomes, begin to sink into, or to


74 LEONARD W. WILLIAMS

be surrounded by, the selerotomic tissue when the embryo has about twenty segments, and in embryos of twenty-five segments they are quite surrounded by selerotomic tissue.

The notochordal processes of the sclerotomes now unite beneath the notochord, separating it and the neural tube from the entoderm, and also unite beneath the aorta with the aortic processes, separating the aorta of each side from the entoderm.

In embryos of forty segments (fig. 9) the dermomj^otome is much larger than before and is now of irregular thickness, being more than twice as thick at the junction of its middle and lower thirds as in its upper fourth. This is due to the rapid expansion of the tissue of the cutis plate which accompanies its transformation from an epithelioid into a reticular or mesenchymal form. This transformation can be readily followed in transverse or frontal sections, for, beginning at a point somewhat lateral to the center of the cutis plate, the area of disintegration rapidly spreads in all directions until it reaches the lower, anterior, and posterior edges of the cutis plate. The history of the lower edge of the cutis plate is brief, for, as Fischel has shown, the dermomyotome of each of the prebrachial somites does not develop along its lateral or lower edge a zone of growth ; consequently this edge is quickly transformed into mesenchyma. The zones of growth along the anterior and posterior edges of the dermomyotome continue for some time to produce streams of cells which augment the mass of mesenchyma, between the ectoderm and the myotome. The cells arising from the two centers of growth on the posterior edge of one and on the anterior edge of the following myotome form a dense mass of mesenchyma which in frontal sections is very conspicuous.

The upper part (approximately one-fourth) of the cutis plate retains an epithelioid aspect as is well shown in figure 121 in Minot's "Human Embryology." It is differentiated, however, into a loose distal portion and a thin dense basal layer containing both dividing and resting nuclei. The latter layer forms the outer boundary of the cavity of the segment which is now reduced to a narrow cleft in the upper edge of the dermomyotome. I have not been able to follow in detail the further history of the upper


THE SOMITES OF THE CHICK 75

edge of the lateral layer of the dermomyotonie of this somite, but it seems to be converted into myoblasts like the upper and lower edges of the lateral layers of the dermomyotomes of the trunk.

The myotome has a clavicular form in vertical section, its upper edge being slightly bent toward the neural tube and its lower edge being strongly inclined laterally. The rapid expansion of the mesenchyma arising from the cutis plate has apparently caused the overlying ectoderm to bulge outward. The inner surface of the myotome is quite sharp and regular and is only loosely attached to the sclerotomic tissue. The outer surface, on the contrary, is closely connected with the overlying mesenchyma and is somewhat irregular, owing to the formation of large irregular spaces. These are most abundant somewhat below the center of the plate, and appear to indicate the beginning of the separation of the myoblasts to allow the interpolation of connective tissue elements. Maurer's figure ('04, fig. 24) of the dorsal edge of one of the dermomyotomes of the middle of the trunk of a chick embryo of five days differs in certain respects from sections of this and of all other dermomyotomes of the chick that I have seen. The outer surface of the muscle plate is represented as regular and well defined, not the inner surface; the medial or inner surface is irregular, owing to the migration of sclerotomic tissue into the muscle plate. Finally, the upper edge of the dermomyotonie is very thin and acute, while in all the segments that I have seen it is thick and rounded. I cannot explain the wide divergence between Maurer's observations and my own.

The lower edge of the muscle plate is not well defined but merges gradually with the adjacent connective tissue. It is slightly indicated, however, by a blood-vessel which passes close under it to supply the mesenchyma of the dermatome and which anastomoses with other vessels that pass between the slightly divergent lower ends of the adjacent nwotomes.

The nuclei of the lower part of the myotome, as before, are larger and less closely packed than those of its upper part. The majority of them are elongated longitudinally, but a few, particularly those near the middle of the inner surface of the plate, are rounded or are elongated in another direction. Similarly the axes of the


76 LEONARD W. WILLIAMS

vast majority of mitotic figures are longitudinal, but, as will be seen in the figure, a few are transverse or are directed otherwise than longitudinally. Bardeen ('00) maintained against the general opinion that the mesenchyma between the myotome and the ectoderm, in the pig at least, does not arise from the cutis plate whose cells migrate in a mass and without losing their epithelial arrangement, from the middle of the anterior border of the myotome, backward, upward, and downward toward the corresponding edges of the myotome where the cutis plate cells turn over the edge of the muscle plate and, with the exception of a few which degenerate, become myoblasts. The proof of this migration lies in the gradual reduction in the size of the cutis plate; in the gradual transition in structure between the muscle plate and the cutis plate; and finally in the existence of an external membrane upon the periphery of the cutis plate. Nevertheless, there is no doubt that the cutis plate in the chick gives rise to the mass of mesenchyma in question, for not only can the gradual transformation of the epithelioid cutis plate into mesenchyma be followed, but there is also no path open by which this tissue can migrate from the parietal plate. The ganglion of the vagus and the anterior cardinal vein quite fill up the narrow slit between the lower edge of the muscle plate and the ectoderm through which alone the tissue could migrate. In the majority of the segments of the chick, a flat wedge-shaped plate of mesenchyme projects upward between the lower part of the cutis plate and the ectoderm, just as is the case in mammals; but in these segments the structure and history of the cutis plate is the same as in the second segment, except that the cutis plate has many more cells, and there is every reason to believe that it produces mesenchyma. I have not had an opportunity to follow the development of the cutis plate in any mammal with sufficient care to warrant a positive assertion, nevertheless, the similarity in structure and development of the cutis plate in birds and mammals leads me to doubt Bardeen's conclusions.

The sclerotome of this somite has fused completely with the adjacent sclerotomes, but its anterior boundary is indicated by the first intersegmental vein and its posterior boundary by the first


THE SOMITES OF THE CHICK 77

definitive intersegmental artery and the second vein. Occasionally a small artery occurs on one or both sides in the place of the normally aborted first intersegmental artery. The outer part of the sclerotome is somewhat denser than its medial portion.

The four occipital segments have only rudiments of spinal ganglia, but the ganglionic commissure of the vagus extends backward through the somites of each side to the aborting ganglia of the fifth and sixth segments. No ventral nerve root forms in the first segment but, as Chiarugi has shown, the hypoglossal nerve has three roots belonging respectively to the second, third, and fourth segments.

Frontal sections of the myotome show that its upper and lower edges differ considerably in form. Its upper part, in an embryo of forty-four segments (H. E. C. no. 98) is of nearly uniform thickness and at each end narrows sharply to an acute edge which is separated by a narrow hyaline zone from the next myotome. The middle of the myotome is thicker than its upper part and is symmetrically convex. Its anterior, and to a less degree its posterior edge are obliquely truncated so that an acute wedge of sclerotomic tissue projects between the ends of the adjacent myotomes. The inner surface of the lower part of the myotome is more convex than its outer surface, and wedges of sclerotomic tissue, apparently the rudiments of the myosepta, separate this part of the myotome from the adjacent myotomes.

Transverse sections through the middle of the last formed somite of embryos of two, five, ten, fifteen, twenty-five, thirty, and fortj^-f our segments are represented in figures 1 and 10 to 15. The somites of the different regions of the body show certain differences which, though not large, have not been sufficiently emphasized, for it is assumed that all somites have the same structure and history. Certain investigators upon this assumption have tried to bridge over gaps in the development of a particular segment by taking the structure of a more anterior or more posterior one as a later or an earlier stage of development. Lillie, for example, says (p. 185): The manner of origin of these parts (of the somite) can be studied fully in an embryo of twenty-five or thirty somites, by comparing the most posterior somites in which


78 LEONARD W. WILLIAMS

the process is beginning with somites of intermediate or anterior positions in the series which show successively later stages." Kollmann expresses the same idea more specifically (p. 43 and 44) :- A myotome from the posterior part of the trunk will be represented because its condition immediately follows that of the mj^otome of the embryo of two weeks, which is not the case with the myotomes of the anterior part of the trunk. There, in correspondence with the more advanced development of the anterior part of the body, the myotomes are considerably more advanced than those of the posterior part of the trunk. One can therefore make out various stages of development of this muscle organ in a single embryo."

Our attention is at once arrested by the great variation in the size of the somites. There is, from in front backward to the thirtieth, a gradual increase in the size of the somites; beyond this, that is, in the caudal region, the somites gradually become smaller. The fifth somite is apparently an exception for it is nearly as large as the fifteenth.

A distinct core appears first in the ninth or tenth somite and each succeeding somite as far back as the thirtieth or thirty-fifth segment has a larger core. The greater size of the more posterior somites is due largely to the fact that the somitic plate from which they are formed is thicker posteriorly. The structures of the anterior part of the embryo grow more rapidly than those of its posterior part and consequently each somite is always larger than those behind it. In other words, the anterior somites, which at the start are smaller and more primitive than the posterior somites, because of their more rapid growth always keep in advance of the latter. Thus we see that the second somite (fig. 5) of an embryo of twelve segments is larger, as well as more advanced, than the fif


- "Es 8oll hior zunjiclist cin Myotoin aus dcin liintei'cn Rumpfabschnitt dieses Embryo geschildert wcrden, weil sich dessen Verhallen unmittclbar an dasjenige des zwei Wochen alten Embryo anschliesst, was mit den Myotomen im Vorderrumpf nicht der Fall ist. Dort sind sie in Uebereinstimmung mit der ganzen vorgeschrittenen Ausbildung dcs Vorderkorpcrs lietraclitlich dcnen dcs Hinterrumpfes voraus. Man kann also verschiedene Entwickelungsstufen dieses Muskelorganes kennen lernen an einem iind demselben Embryo."


THE SOMITES OF THE CHICK


79


teenth somite of a slightly older embryo (fig. 12). The growth of the segments is shown in the longitudinal sections as well as in the transverse sections, as will be seen from the following table of measurements.


NUMBER OF



LENGTH IN MICRONS



DXiliMJljNTS Or EMBRYO


Second somite


Fifth somite


Tenth somite


Fifteenth somite


2


71





3


78





4


93





5


93


93




7


99


96




8


109


99




9


87


93




9


99


87




10


109


96


68



10


93


93


68



11


99


99


74



12


78


99


68



13


99


93


71



15


99


96


68


78


15


93


93


74


68


17


99


93


78


81


17


68


90


68


78


19


99,


93


84


81


In embryos of thirty segments, all of the somites are from 139 to 156 microns long.

Furthermore the later somites do not recapitulate the development of the earlier somites; on the contrary, they merely omit the initial phases of the development of the earlier somites. Thus the core gradually forms in the second somite whereas that of the tenth and later somites is present from the first. In the same way the upper myotomic groove appears in the second somite after the formation of the ninth segment, but in the twenty-fifth and thirtieth somites it appears almost at once. So also the nephrotome is in a more advanced stage of development in the newlj formed posterior segments than in the anterior segments at the time of their origin.


80 LEONARD W. WILLIAMS

It is not my purpose to follow in detail the development of each segment, but merely to mention the most conspicuous points of difference, most of which are already known, between the somites of the head, neck, trunk, and tail, and to determine as far as possible the relation of the somite to the sclerotome and of the sclerotome to the vertebra.


THE TENTH SOMITE

The tenth somite will serve as a type of the somites of the neck, just as the second has served as an example of an occipital segment. The structure of this somite at the time of its origin (fig. 10) needs no description, except to note that the intersegmental fissures which bound it are inclined slightly forward below\

While the three following segments are forming, the tenth somite of each side grows considerably and, owing apparently to the pressure of the surrounding structures, becomes somewhat pentagonal, with nearly equal dorsal, medial, and ventro-lateral surfaces and smaller lateral and ventro-medial surfaces. The single important structural change is the thinning out of the medial and anterior walls and the formation at the top of the former of the upper myotomic groove.

An ejctensive fusion of the core with the antero-ventral angle of the floor of the tenth somite is present in embryos of fifteen segments and in embryos of seventeen and eighteen segments the lower myotomic groove appears on both the medial and the anterior walls of the somite. The newly formed notochordal process now replaces the ventro-medial surface of the somite and the ridge between the dorsal and lateral surfaces has disappeared. Consequently the somite is now triangular. A mesenchymal mass, the aortic process, projects downward between each aorta and the corresponding posterior cardinal vein, but it apparently arises entirely from the nephrotome, not from the sclerotome. The anterior end of the notochordal process projects some distance forward and consequently the medial portion of the intersegmental fissure is inclined forward ventrally so as to enclose


THE SOMITES OF THE CHICK 81

with the ectoderm of the dorsal surface of the embryo an angle of 76.° This part of the sclerotome is thus carried forward one-third or one-fourth the length of the somite. The succeeding intersegmental fissures are more and more vertical, the fifteenth being the first that is without any inclination forward. The sclerotomes of the first fifteen somites, except the first which, it will be remembered, is not separated from the unsegmented mesoderm of the head, differ from those of the remaining somites in having this peculiar forward prolongation which, although it can be observed readily in transverse and sagittal and often in frontal sections, has never before been described. The obliquity of the sclerotomes of the first, and presumably the most primitive segments coupled with the well-known obliquity of the first one or two intersegmental clefts gives color to the suggestion that the segments of the chick are typically oblique rather than transverse structures and, consequently, that each vertebra arises from a single segment.

The lower myotomic groove deepens rapidly and, in embryos of twenty segments, appears on the posterior wall of the somites of the tenth segment. Neural crest cells, migrating downward, gather in the deep anterior part of this groove. The right and left aortae are now moving toward the median plane and consequently separate the notochord more widely from the entoderm.

The cutis plate of the tenth somite of embryos of twenty-five segments has nearly the same structure as that of the second somite of embryos of eighteen segments, except that, being composed of a greater number of cells, it more closely resembles a stratified columnar epithelium.

The deepening of the lower myotomic groove has now reduced the connection between the dermomyotome and the sclerotome to a slender stalk. This process accompanies the differentiation of the scelerotome into two conspicuous but not sharply defined regions. The medial portion of the sclerotome becomes less dense owing in part at least to its expansion into the space beside and below the notochord; its lateral part, on the other hand, becomes denser. The right and left aortae have united in this segment forming the median cylindrical aorta, which forces the

The American Joohxaf. of Axatomy, Vol. II, No. 1.


82 LEONARD W. WILLIAMS

notochordal processes of the sclerotome into a nearly horizontal position. The aortic process has been carried inward and is now a nearly vertical septum between the aorta and the posterior cardinal vein.

In embryos of thirty segments, the derniomyotome of the tenth somite is quite separated from the sclerotome and is inclined downward and laterally approximately at an angle of 45° with the sagittal plane of the embryo. Its breadth (255 microns) is nearly twice its length (150 microns). The ventro-medial surface of the derniomyotome is strongly convex except at the upper and lower edges where it is barely concave. Its dorso-lateral surface is slightly convex, both at the edge and at the center, and the intermediate zone is flat or barely concave. The central convexity of both surfaces of the dermomyotome is due apparently to the expansion of the tissue of the dermatome as it becomes mesenchyma. A small central area of the dermatome is entirely converted into mesenchyma and a circular zone as broad as the dermomyotome is long and somewhat more than half its breadth is partly so. The lower edge of the dermomyotome rests upon the posterior cardinal vein and is partly separated from the ectoderm by a thin sheet of compact mesenchyma belonging to the parietal plate but its slight contact with the ectoderm forms a definite boundary which proves beyond doubt that the mesenchyma in the concavity of the muscle plate cannot have migrated from the parietal plate. The increased breadth of the dermomyotomes seems to be due to two factors— intrinsic growth and the addition of new material at its dorsal growing edge. The cavity of the segment is represented by small cavities in the upper and lower edges of the dermomyotome.

The sclerotome is fusing with the adjacent sclerotomes and its tissue is becoming vascularized.

Figure 17 represents a frontal section through the tenth somite of an embryo of forty segments. It is remarkably like figure 120, p. 205,of a somite of the rabbit inMinot's "Human Embryology." The dermomyotome, which in embryos of thirty segments was inclined approximately at an angle of 45° with the median plane, is now nearly vertical. The plane of the section is slightly inclined


THE SOMITES OF THE CHICK 83

upward laterally and, as the disintegration of the cutis plate has proceeded but slowly, the upper part of the area of partial disintegration is shown in the section. The spinal ganglion is now well defined and is quite surrounded by mesenchyma. The spinal nerve extends downward nearly into the section figured. The sclerotome is bounded anteriorly and posteriorly by the intersegmental vessels {I. A., I.V.), laterally by the muscle plate (M.), medially by the notochord (A'"), and at a higher level by the neural tube, ventrally by the aorta and the posterior cardinal vein. The sheet of sclerotomic tissue between the aorta and the notochord has become considerably thicker, and a similar but very thin sheet separates the notochord from the neural tube. The loose tissue surrounding the notochord, and extending as far laterally as the intersegmental arteries (I. A.), forms a continuous perichordal sheath in which no visible condensations occur. The intersegmental arteries mark the outer limit of the future vertebral centra and the body of each vertebra will form in this originally homogeneous perichordal sheath which was first described by Gegenbaur. No portion of this sheath can be assigned with accuracy to any particular somite or sclerotome, for we have seen that the anterior ends of the sclerotomes of the second to the fifteenth segments extend far forward under the preceding sclerotomes. Laterally to the intersegmental arteries the sclerotomic tissue rapidly becomes denser except in the middle of the sclerotome where a thin vertical zone of loose tissue, the "intervertebral fissure," divides the denser tissue into distinct anterior and posterior portions which take the form of square columns. These two columns of dense sclerotomic tissue are generally, but I believe wrongly, regarded as morphological entities. There is, however, some difference of opinion as to their exact meaning. Those who hold to Remak's theory of vertebral resegmentation as modified by Von Ebner, consider them to be the anterior and posterior halves of the right or left half of the primitive vertebra." Others, however, who accept Schauinsland's theory (which, as was pointed out above, is closely related to Goette's theory) that there are two primitive vertebrae" in each segment, regard each column of dense sclerotomic tissue as simply one-half of the anterior


84 LEONARD W. WILLIAMS

or the posterior vertebra of the segment. These cohuiins are, I beHeve, nothing more than centers of growth which have been cut off by the lower myotomic groove from the zones of growth of the anterior and posterior edges of the dorsal lamella of the somite. They certainly contain (as Von Ebner pointed out in 1892, and as I have maintained elsewhere) the anlagen of a large number of structures. The anterior, however, differs largely from the posterior column, owing, I believe, to the intrusion of the spinal ganglion and nerve into it and the consequent interference with its development.

The cutis plate of each somite of the tenth segment of an embryo of forty-four segments (H. E. C, no. 98) has given rise to a large mass of loose dermal mesenchyma and to a wall-like peripheral zone of denser mesenchyma which extends from the zone of proliferation at the anterior, posterior, and dorsal edges of the myotome and from the remnant of the lower edges of the cutis plate to the ectoderm. The septum of denser mesenchyma is attached to the ectodermal thickening or ridge along the anterior, posterior, and lateral (or ventral) boundaries of the somite, and seems to draw the ectoderm toward the myotome, causing a deep invagination of the ectoderm. The existence of this invagination around three sides of the somites accounts for the greater conspicuousness of the somites of entire embryos of three or four days than of those of younger embryos.

The outer lamella of the rapidly growdng upper edge of the dermomyotome is not affected by the disintegration of the cutis plate, for as Engert, has shown, this edge of the. dermomyotome becomes enlarged, appearing in transverse sections like a pothook or a shepherd's crook, and is transformed bodily into myoblasts.

The sclerotomic tissue, both of the perichordal sheath and of the sclerotomic columns, is considerably denser than before. Sympathetic ganglia are now present in the tenth segment.

In embryos of forty-eight to fifty-two segments (H. E. C. nos. 478, 483, and 526) the dense peripheral portion of the dermal mesenchyma has become much less conspicuous except along the dorsal and the upper part of the anterior and posterior edges of


THE SOMITES OF THE CHICK 85

the myotome. The corresponding ectodermal ridge, however, has been elevated, especially its vertical portions, into an acuminate septiim-like structure which in one case projects medially from the bottom of the ectodermal invagination 50 microns, or more than one-third of the distance from the bottom of the ectodermal invagination to the myotome.

The cutaneous blood vessels pass to the skin through the gaps between the lower ends of the myotomes.

The myotome is considerably thicker than before, and, except at its upper and lower edges, is now triangular in frontal section. Its outer surface is flat or slight^ convex and its medial surface is now divided into two equal surfaces which, facing slightly forward and backward, meet in a large obtuse angle that varies in different levels from about 120° to 160°. The myoblasts of the lateral part of the myotome are in a more advanced stage of development than those of the medial portion, and they seem to stretch from the antero-medial to the posterior medial surface of the myotome. The myoblasts of the medial part of the myotome are shorter and are more irregularly arranged than the more lateral myoblasts. Those in immediate contact with the medial surfaces near the medial angle of the myotome form an epithelioid layer.

The transverse diameter or breadth of the intervertebral fissure is considerably reduced owing to the encroachment of the medial angle of the myotome.

The axial mesenchyma is denser than in embryos of fortyfour segments, and there has now appeared an extensive midsegmental subnotochordal condensation which is continuous laterally with the dense anterior and posterior sclerotomic columns.

The presence of the large spinal nerve and the sympathetic ganglion in the anterior sclerotomic column interferes with the original continuity of the anterior column and therefore Froriep and Bardeen have conceived that the dense tissue of each segment forms a simple lyre-shaped mass extending from the median midsegmental point laterally, upward, and backward. Froriep calls this structure the primitive vertebral arch, and Bardeen names it the scleromere. Remak names its central portion the primitive vertebral centrum, and its lateral portions the verte


86 LEONARD \V. WILLIAMS

bral arches. Froriep shows that the primitive vertebral centrum soon extends around the notochord and differentiates into the perichordale Faserring and the cartihi^inous hypochordale Spange. The former takes part in the formation of the intervertebral joint or in mammals forms the intervertebral disc — the latter is tlie intercentrum (Mannich, '02) or the haemal arch (Schauinsland, '05). Froriep further shows that the vertebral centra are formed from the loose tissue of the perichordal sheath between the successive midsegmental condensations. I believe it will be found that in the chick, as in the pig, cartilage arises from relativeh' loose, not from dense mesenchyma, or that when it does arise in dense tissue it does so only after a previous loosening up of the dense tissue. This is certainly true of the centrum and intercentrum. We ought, I believe, to regard the scleromere as a composite structure, not as a morphological or structural unit. It is unwise, therefore, to try to homologize the sclerotomic columns or halves or the scleromere with the vertebra or with any of its elements.


THE TWENTY-FTFTH SOMITE

The twenty-fifth segment (fig. 13) at the time of its formation is in a more advanced state of development than the second or the tenth when they first appear, and it passes with great rapidity through the first steps of its later development. In an embryo of twenty-seven segments (H. E. C. no. 1520) each somite of this segment is already divided into the sclerotome and the dorsal lamella, and both myotomic grooves have appeared. The upper myotomic groove, however, is very shallow and is so indistinct that it is not readily seen. The lower groove entirely encircles the somite and neural crest cells are already moving down into its deep anterior end. The newly formed sclerotome has fused with the preceding and following sclerotomes, but not with the opposite one. It is continuous laterally with the mesenchyma of the nephrotome and through it with that of the lateral plate. The aorta, the posterior cardinal vein and the connecting interseg


THE SOMITES OF THE CHICK 87

mental vessels are present. The aorta lies directly beneath the somite, but it is almost in contact with the lateral plate. The posterior cardinal vein lies directly above the Wolffian duct which in turn lies in a groove between the lateral plate and the nephrotomic mesenchyma.

The twenty-fifth somite of each side in embryos of thirty-three segments (H. E. C. nos. 97 and 1587) is triangular. The aortae have moved together and have fused. The entire nephrotome with the posterior cardinal vein has moved toward the median plane, and consequently the sclerotome is now bounded ventrally by the aorta, the mesonephric tubules, the Wolffian duct, and the posterior cardinal vein. The dorsal surface of the somite, owing to the lateral inclination downward of the dorsal lamella, makes an angle of 45° with the sagittal plane.

The lateral edge of the dorsal lamella has grown ventro-laterally some distance beyond the ectodermal ridge which formerly marked the lateral boundary of the somite and now is pushing into the dense mesenchyma of the Wolffian ridge. The myotome now underlies all of the dorsal lamella but a small area in the middle of its ventro-lateral edge. The sclerotome is definitely divided into a loose medial and a dense lateral part. The two sclerotomes of this segment have united in the median plane so that a nearly continuous sheet of sclerotomic tissue separates the aorta and the notochord. Neural crest cells continue to migrate into the spinal ganglion.

After the first phase of rapid development, changes in the structure of the twenty-fifth segment are less rapid. The dermomyotome of each of the twenty-fifth somites of an embryo of fortj^-four segments is essentially as far advanced as that of the second somite of an embrj^'o of forty segments. There is, however, a striking difference between the two (compare fig. 16 with fig. 9). The ventral edge of the dermomyotome of the twenty-fifth somite contains a center of growth and extends downward until it comes in contact with the coelomic epithelium. This edge of the dermomyotome thus comes to lie medial to a large wedge-shaped process of the mesenchyma of the Wolffian ridge. This process extends upward between the dermomyotome and the ectoderm to


88 LEONARD W. WILLIAMS

the longitudinal ectodermal invagination which marks the former lateral boundary of the somite. The center of the cutis plate, comprising about one-third of the whole, is now entirely transformed into dermal mesenchyma, and the adjacent portion of the cutis plate, including perhaps another third, is partially transformed into mesenchyma. The dermal mesenchyma of this somite is quite dense. It extends from the myotome to the ectoderm and it also projects in the form of a thin sheet between the partly transformed area of the cutis plate and the ectoderm. Ventrally the dermal mesenchyma reaches nearly to the upper edge of the mesenchyma of the Wolffian ridge, from which, however, it is separated by a small space that intervenes between the ectodermal ridge and invagination that mark the former lateral boundary of the somite and of the dermomyotome. The presence of this gap between the mesenchyma of the cutis plate and that of the parietal plate makes it clear that the former does not arise from the latter but that the two mesenchyma masses arise separately. The later development of the myotomes of the trunk has been followed by Engert.


THE FORTY-FOURTH SOMITE

The history of the forty-fourth segment does not differ much from that of the twenty-fifth segment except that shortly after its differentiation into muscle, and dermal and axial connective tissue, it degenerates. In an embryo of forty-nine segments (H. E. C'. no. 478) the core of each somite of this segment has fused with the lower part of its cortex and both the upper and the lower myotomic grooves are present. The upper groove is more distinct than in the twenty-fifth segment, and the lower groove is deeper on the lateral than on the medial surface of each somite.

The dermomyotomes of the forty-fourth segment of an embryo of fifty-one segments (H. E. C. no. 526, 5 days), are complete, but are not cut off from the sclerotomes. The sclerotomes of this segment have fused with the large mass of hypomerous mesenchyma and with the adjacent sclerotomes.


THE SOMITES OF THE CHICK 89

In an embryo of fifty-two segments (H. E. C. no. 344, 5 days 16 hours) which, however, is considerably older than the one of fifty-one segments, the mesenchymal mass (fig. 18, S) formed by the fusion of the sclerotomes with the hypomerous mesenchyma almost surrounds the dermomyotomes of the forty-fourth segment and meets above the neural tube. The dermomyotome is still intact notwithstanding the conversion of the central part of the cutis plate into mesenchyma. The lower edge of the dermomyotome and that of the myotome are somewthat enlarged. The upper edges of both are thin.

In transverse sections we find, in the upper part of the sections, four layers of tissue between the ectoderm and the notochord namely, a layer of mesenchyma (fig. 18, S) which extends upward between the dermatome (D) and the ectoderm nearly or quite to the top of the dermomyotome ; a second layer of mesenchyma representing the central part of the cutis plate (D) ; the myotome (M) ; and a thick layer of sclerotomic tissue which is divided into a lateral zone, of dense and a medial zone of looser tissue.

In a somewhat older embryo (H. E. C. no. 485, 5 days, 18 hours) the dermal mesenchyma has fused with that between the dermomyotome and the ectoderm, and the resulting mass of tissue is supplied with a rich plexus of blood vessels. The lower edge of the myotome bears a small epithelioid cap which represents a much reduced ventral zone of growth. The upper edge of the myotome bears a somewhat rounded zone of growth.

SUMMARY AND CONCLUSIONS

Although the somites have the same fundamental structure in all parts of the body, they differ greatly in many respects.

The first somite at its origin is the smallest of the entire series, with the possible exception of some of the degenerate caudal somites, and each succeeding somite of the neck and trunk is larger than the preceding somite.

Every somite, owing to the rapid growth of the anterior part of the embryo, is larger at any one time than any of the following somites.


90 LEONARD AV. WILLIAMS

A given somite does not recapitulate the development of each or any preceding somite; on the contrary, each succeeding somite is in a more advanced structural condition at the time of its formation than the preceding somites.

Each somite divides first into the primary sclerotome, formed by the fusion of the core with the floor and the lower part of the walls of the somite, and the dorsal lamella which consists of the roof or dorsal wall of the somite and a bordering zone of growth developed in the upper part of the medial, anterior, and posterior walls and, in the somites of the trunk, of the lateral wall also.

The dorsal lamella gives rise to the dermomyotome and also, I believe, to two large dense masses of sclerotomic mesenchyma, the anterior and posterior "halves" of the sclerotome, or, as I perfer to call them, the anterior and posterior sclerotomic columns.

The myotome, or muscle plate, is formed in a large measure, particularly in the occipital and cervical regions, by proliferation from the medial edge of the dorsal lamella but in the trunk and to a less degree in the tail, the lateral edge of the dorsal lamella contributes to its growth.

The anterior and posterior edges of the dorsal lamella contribute but slightly to the myotome. They produce, however, a large amount of mesenchyma which is added to the primary sclerotome and which forms the dense anterior and posterior columns of the secondary sclerotome.

The sclerotomic columns, which are usually regarded as morphological units that represent one or more elements of a primarj^ or secondary vertebra, are actually composite structures in which the several morphological elements cannot be distinguished.

The perichordal sheath, in which the centra and intercentra arise, is formed by the fusion of the notochordal processes of the primary sclerotomes. The notochordal processes of the first fourteen pairs of sclerotomes, exclusive of those of the first segment, grow far forward under the preceding segments; consequently parts of the perichordal sheath can be only vaguely assigned to particular segments. Each centrum arises in an intersegmental part of the perichordal sheath and is formed from loose tissue ; each intercentrum, on the contrary, arises in a midsegmen


THE SOMITES OF THE CHICK 91

tal condensation. The ribs in the same way arise in the dense tissue of the posterior sclerotomic columns. There is, therefore, much reason to believe that in the chick cartilage arises froni relatively loose as well as from dense tissue and also that dense tissue always becomes somewhat looser before cartilage forms in it. I have maintained in a former paper that cartilage arises in this manner in the pig.

There is a possibility that the forward growth of the lower part of the sclerotome is a typical condition in vertebrates and, consequently that the vertebrae are truly segmental structures.

Paterson showed that the dorsal and ventral edges of the dermomyotomes of the trunk grow rapidly upward and downward respectively into the membrana reuniens superior and inferior. They are ultimately converted, as Engert shows, en masse into myoblasts and form the thick dorsal and ventral edges of the myotome.

Rabl was right in asserting that the cutis plate in the chick gives rise to dermal connective tissue, and I question the correctness of Bardeen's theory that the cutis plate in the pig produces only myoblasts.

Finally, I wish to call attention to the circumstance that the anterior cardinal vein arises in part from a branch of the aorta that is comparable to the first intersegmental artery.


92 LEONARD W. WILLIAMS

BIBLIOGRAPHY

Bardeen, C. R. 1900 The development o' the musculature of the body wall in the pig. Johns Hopkins Hospital Reports, vol. 9, pp. 367-399, 10 pis. 1905 The development of the thoracic vertebrae in man. Am. Jour. Anat., vol. 4, pp. 163-1 M, 7 pis.

Bardeen, C. R., and Lewis, W. H. 1901 Development of the limbs, body-wall and back in man. Am. Jour. Anat., vol. 1, pp. 1-35, pis. 1-9.

Bonnet, R. 1884 Beitrage zur Embryologie der Wiederkiiuer gewonnen am Schafei. I Arch. f. Anat. u. Physiol., Anat., Abth., pp. 170-230, pis. 9-11. 1889 II Ibid., pp. 1-106, pis. 1-6.

Chiarugi, G. 1889 Lo sviluppo dei nervi vago, accessorio, ipoglosso e primi cervicali nei sauropsidi e nei mammiferi. Atti. Soc. Tosc. Sc. Nat., Pisa, vol. 10, pp. 149-24J, pis. 11-12.

Corning, H. K. 1891 Ueber die sog. Neugliederung der Wirbelsaule und iiber das Schicksal der Urwirbelhohle bei Reptilien. Morphol. Jahrb., vol. 17, pp. 611-622, pi. 30.

Dexter, S. 1891 The somites and the coelome in the chick. Anat. Anz., vol. 6, pp. 284-^89, 4 figs.

Von Ebner, V. 1888 Urwirbel und Neugliederung der Wirbelsaule. Sitzb. d. k. Akad. d. Wiss., Wien, 97, Abth. 3, pp. 194-206, pis. 1-2. 1892 Ueber die Beziehungen der Wirbel zu den Urwirbel. Sitzb. d. k. Akad. d. Wiss., Wien, 101, Abth. 3, pp. 235-260, pi. 1.

Engert, H. 1900 Die Entwicklung der ventralen Rumpfmuskulatur bei Vogeln. Morphol. Jahrb. 29, pp. 169-186, pis. 8-10.

Evans, H. M. 1909 On the earliest bloodvessels in the anterior limb buds of birds and their relation to the primary subclavian artery. Am. Jour. Anat., vol. 9, pp. 281-319, 20 figs.

1909 On the development of the aortae, cardinal and umbilical veins and other blood vessels of vertebrate embryos from capillaries. Anat. Rec, vol. 3, pp. 498-518, 21 figs.

FiscHEL, A. 1895 Zur Entwickelung der ventralen Rumpf- und Extremitatenmuskulatur der Vogel und Saugetiere. Morphol. Jahrb., vol. 23, pp. 544-561, pi. 28.

Froriep, a. 1883 Zur Entwickelungsgeschichte der Wirbelsaule, insbesondcre des Atlas und Epistrophius und der Occipital Region. Arch, f. Anat. u. Physiol., Anat. Abth., pp. 177-23 1, pis. 7-9.

Gegenbaur, C. 1862 Untersuchungen zur vergleichenden Anatomic der Wirbelsaule bei Amphibien und Reptilien. Leipzig, 72 pp., 4 pis.

Goette,A. 1875 Die Entwickelungsgeschichte der Unke. Leipzig, 965 pp. Atlas. 1896 Ueber den Wirbelbau bei den Reptilien und cinigcn anderen Wirbeltieren. Zeitschr. f. wiss. Zool., vol. 62, pp. 343-394, pis. 15-17.


THE SOMITES OF THE CHICK 93

Held, H. 1909 Die Entwicklung des Nervengewebes bei den Wirbeltieren. Leipzig, 378 pp., 53 pis., 275 figs.

His, W. 1868 Untersuchungen iiber die erste Anlage des Wirbeltierleibes. Die erste Entwickelung des Hiihnchens im Ei. Leipzig, 16+237 pp., 12 pis.

Hubbard, M. E. 1908 Some experiments on the order of succession of the somites of the chick. Am. Nat., vol. 42, pp. 466-471, 2 figs.

KoLLiKER, A. 1879 Entwickelungsgeschichte des Menschen iind der hoheren Tiere. 2 ed., Leipzig.

KoLLMANN., J. 1891 Die Rumpfsegmente menschlicher Embryonen von 13 bis 35 Urwirbeln. Arch. f. Anat. u. Physiol., Anat. Abth., pp. 39-88, pis. 3-5.

LiLLiE, F. R. 1908 The development of the chick. New York, 472 pp., 250 figs.

Manner, H. 1899 Beitrage zur Entwickelungsgeschichte der Wirbelsaule bei Reptilien. Zeitschr. f. wiss. ZooL, vol. 66, pp. 43-68, pis. 4-7.

Mannich, H. 1902 Beitrage zur Entwickelung der Wirbelsaule von Eudyptes chrysocome I. D., Univ. Leipzig, 46 pp., 1 pi.

Maurer, F. 1904 Entwickelung des Muskelsystems und electrischen Organe. Hertwig's Handbuch der vergl. u. exper. Entwickelungslehre der Wirbeltiere, Bd. 3, Teil 1, 1-80, 41 figs.

Minot, C. S. 1892 Human embryology. New York, 815 pp.

Paterson, a. M. 1888 On the fate of the muscle-plate and the development of the spinal nerves and limb plexuses in birds and mammals. Quart. Journ. Microscop. Sci., vol. 28, pp. 109-129, pis. 7-8.

Patterson, T. J. 1907 The order of appearance of the anterior somites of the chick. Biol. Bull., vol. 13, pp. 121-133.

Rabl, C. 1888 Ueber die Differenzierung des Mesoderms. Verb. d. Anat. Ge^., Wtirzburg, vol. 2, pp. 140-146, 8 figs.

1889 Theorie des Mesoderms, I. Morphol. Jahrb., vol. 15, pp. 113252, pis. 7-10.

Remak, R. 1855 Untersuchungen fiber die Entwicklung der Wirbeltiere. Berlin, 37 +194 pp., 12 pis.

Schauinsland, H. 1900 H. Weitere Beitrage zur Entwicklungsgeschichte der Hatteria. Arch. f. mikr. Anat., vol. 56, pp. 747-867, pis. 32-34. 1905 Die Entwickelung der Wirbelsaule nebst Rippen und Brustbein. Hertwig's Handbuch d. vergl. u. exper. Entwickelungslehre der Wirbeltiere, Bd. 3, Teil 2, pp. 339-572, 157 figs.

Williams, L. W. 1908 The later development of the notochord in mammals. Amer. Jour. Anat., vol. 8, pp. 251-284, 20 figs.


94


LEONARD W. WILLIAMS


EXPLANATION OF FIGURES

All fifiiuros arc from series in the Harvard Ein})r3-ological Collection. All figures except the text figure, are of tlie same magnification, 280 diameters.

Figures 1 to 9 represent transverse sections through the middle of one somite of the second segment of chick embryos.

1 From an embryo of two segments. Series 627, section 107.

2 From an embryo of three segments. Series 629, section 100.

3 From an embryo of six segments. Series 632, section 120.

4 From an embryo of nine segments. Series 645, section 149.

5 From an embryo of twelve segments. Series 1449, section 183.

6 From an embryo of fifteen segments. Series 1460, section 103.

7 From an embryo of eighteen segments. Series 1466, section 170.

8 From an embryo of twenty-five segments. Series 89, section 144.

9 From an embryo of forty segments. Series 100, section 150.

Figures 10 to 15 represent similar sections through one of the somites of the last segment of embryos of different ages.

10 From an embryo of five segments. Series 632, section 141.

11 From an embryo of ten segments. Series 42, section 257.

12 F'rom an embryo of fifteen segments. Series 1460, section 314.

13 From an embryo of twenty-five segments. Series 89, section 408.

14 From an embryo of thirty segments. Series 95, section 395.

15 From an embryo of forty-four segments. Series 98, section 287.

16 Somewhat oblique transverse section of one dermomyotome of the twentysixth segment of an embryo of forty-four segments. Series 98, section 257.

17 Frontal section through the left somite of the tenth segment of an embryo of forty segments. Series 100, section 50.

18 Transverse section of one dermomyotome of the forty-fourth segment of an embryo of fifty-two segments (5 days, 16 hours). Series 344, section 228.


A


Aorta.


/.


V


Intersegmental vein.


,^


Coelom.



M


Myotome.


ri


Remnant of the cavity of the






somite.


M.


I'


Myocardial process of the late


C. A


Anterior cardinal vein.




ral plate.


C. I'


Posterior cardinal vein.



N


Notochord.


D


Dermatome.


N.


C


Neural crest.


I). M


Dermomyotome.


P.


A


Aortic process of sclerotome.


E


Ectoderm.


P.


N


Notochordal process of sclero


a


Entoderm.




tome.


a. L


Lower myotomic groove.



S


Hj'^pomerous mesenchyma.


G. U


Upper myotomic groove.



V


Vascular cells.


I. A


Intersegmental artery.





THE SOMITES OF THE CHICK


95



0__:S




96


LEONARD W. WILLIAMS






THE SOMITES OF THE CHICK


97



THE AMERICAN JOURNAL OF ANATOMY, VOL. 11, NO. 1.


98


LEONARD \V. WILLIAMS



THE SOMITES OF THE CHICK


99



5Vpo







15


100


LEONARD W. WILLIAMS



I.A


'.1


Jb%


&,


'^"-'^^im-^j^'^-


3Mm


■r§S&


a^s^^rtS-^^^^--:



& ®


4


fP^^m ■tc?.t% ^ZS: ^"^




.0^


'r.r Q> "«>. ^


I.A



18




17


THE NERVES AND NERVE ENDINGS IN THE MEIMBRANA TYMPANI OF MAN

J. GORDON WILSON

Department of Otology, Northwestern University Medical School

SIX FIGURES^ THREE PLATES

The membrana tympani though recognized by the anatomist to be a structure admirably adapted to play an important part in the mechanism of sound conduction, has not, until recently, received from the neurologist the attention it would appear to merit. At present the description of the nerve distribution in this membrane in mammals, with the exception of man, may be said to be in the main satisfactory. In man the account is not only meagre but lacking in many essential details. It is contained in the work of Kessel pubhshed in 1872, when the technic for nerves and nerve endings was less satisfactory than it now is. So far as I know, excepting a drawing by Kessel, believed both by Jacques and myself to be inaccurate, there have not appeared any illustrations of the nerve distribution. It is difficult to account for this indifference. While it may be due in some small measure to a lack of appreciation of the importance of this membrane, it is mainly to be accounted for by the difficulty of the technic inherent in its structure. In a former paper ('07 a) I described the mode of distribution and the varieties of endings found in the membrana tympani of the rabbit, dog, cat and monkey; in this paper I propose to extend these investigations to the membrana tympani of man.

1 The drawings have been made by Herr R. Schilling of Freiburg, and Miss Hill of the University of Chicago, to whom I wish to express my sincere thanks.

THE AMERICAN JOURNAL JF ANATOMY, VOL,. 11, NO. 2 JANUARY, 1911


102 J. GORDON WILSON

The literature of the nerve distribution in this structure in mammals is by no means voluminous. In addition to the work quoted above, it is contained in papers by Kessel, Jacques, Calamida and Deinike. Of these the only author who has described the nerves in the membrane of man is Kessel. Jacques' work was done in the membrane of the cat and dog; Calamida's investigations were carried out on several of the lower animals — horse, cat, goat, etc.; Deinike limited his work to the ox and horse. The papers of Jacques, Calamida and Deinike are not here reviewed as they do not bear directly on the nerves in man and further since this has been done sufficiently in a former paper. Within the last year an article has appeared by Gemelli on the nerves in the cat, horse, dog and ape.

Kessel found that in man the principal nerve, composed of medullated fibers, passes from the external auditory meatus, on to the membrane at the upper part of the posterior segment close to and behind the artery. In the onward course of the nerve branches are given off which accompany the vascular twigs. Corresponding to the forking of the artery over the manubrium the nerve divides into two branches of which one supplies the anterior, the other the posterior and lower part of the manubrium. Besides the main trunk several smaller nerves, accompanying blood vessels, pass to the membrane from various parts of the periphery. All these nerves in their course give ofT branches which lie between the cutis and membrana propria forming what he calls the ground plexus. Fibers also enter the membrana tympani from the plexus tympanicus; these fibers reinforce the nerve supply which reaches the mucous membrane from the cuticular side. The ultimate distribution of these nerves is to be found :

(a) around the capillaries — forming a capillary plexus,

(b) under the stratum Malpighii — forming a subepithelial plexus,

(c) under the mucous membrane — forming a submucous plexus. As a result of staining with chloride of gold he describes

(d) in the course of the capillaries single nerve fibers consisting of axis cylinders in which are nodal swellings, containing a nucleus,


NEKVES IN THE MEMBRANA TYMPANI OF MAN 103

at which two or more branches may be given off ; these he regarded as probably ganglion cells;

(e) in the plexus under the stratum Malpighii many bi- and multipolar ganglion cells.

It is no easy task to obtain good results in investigating the nerves of this membrane. Like most of the investigators — Jacques, Calamida and Deinike — I find that ordinary staining methods are inapplicable. Osmic acid, apart from its limitations to medullated nerves, so blackens the tissue as to make it difficult to follow the nerve. Gold and silver salts cause so great a precipitate as to make them unsuitable for satisfactory work. Methylene blue is here an ideal method. The very thinness of the membrane, so objectionable for the other methods, renders it for this the more suitable. It is the method I have chiefly used; at times I have found a useful substitute in a combination of methylene blue and osmic acid.

The method employed in this research was as follows: after removal of the brain, the petrous temporal bone was loosened from the adjacent parts by means of a small chisel and then carefully lifted out, cutting with a knife its fibrous and muscular attachments below. It comes away easily, leaving the tympanic membrane intact in the tympanic bone with the malleus attached to it. Usually the incus remains attached to the malleus; the stapes always comes away with the petrous portion. It is of some interest from the operative point of view to note that the attachment of the foot plate of the stapes to the foramen ovale is firmer than that of the incus to the stapes.

The edges of the membrana tympani and the adjacent part of the covering of the external auditory meatus are now loosened and removed leaving the malleus still attached. By this means there can be detached the whole drum membrane with as much of its mucous layer extension to the middle ear and of the cuticular layer to the external meatus as may be desired. These are immersed in a weak solution of methylene blue (three to six drops of a J per cent solution in 20 cc. of normal salt solution), which has been warmed to a temperature of 37° C. The tissue in this solution is placed in a thermostat at 37° C. for three to eight minutes.


104 J. GORDON "WILSON

Tliis warm solution dissolves the fatty material which lies on the surface of the membrane, loosens the epithelial scales, and increases the action of the dye on the nerves. The effect may be increased by brushing the external surface of the membrane with a small clean camel's hair brush dipped in the fluid. Finally the membrane is exposed to the air on a clean glass slide with the side uppermost on which one wishes to get the better representation of the nerves, usually the external. During this process the tissue is kept moist with a solution of methlyene blue at a temperature of 37° C. of the following strength :

Methylene blue (nach Ehrlich) 0.5 per cent sol 10 cc. or 5 cc.

Salt solution (0.75 per cent) 90 cc. or 95 cc.

The time at which the nerves begin to appear varies with the period after death at which the tissue is obtained, the sooner after death the earlier they appear, but I have obtained results six or eight hours after death when the body has been kept in a cold chamber. If nerves do not appear within one hour it may be regarded as useless to persevere. The dye is fixed in the nerve by immersion of the tissue in an 8 per cent ammonium molybdate solution. The subsequent treatment consisting of washing in water, passing through alcohol and xylol, has been fully described in a recent paper ClOb). The membrane may be mounted entire in Canada balsam, or, the malleus being removed, it may be immersed in paraffin and cut. Counterstaining when required, appears to me to be best obtained by a weak alcoholic solution of orange G. acid fuchsin.

I find what may be called the principal nerve of the membrana tympani, n. tympanicus major (n. t.m.), enters as a broad bundle of fibers from the posterior edge of the membrana flaccida, on the upper posterior segment of the membrana tensa (fig. 1).^ It has an intimate relationship to the artery of the manubrium (a. m.t.), the vessel at first lies posterior to the nerve but as the ves ' In a single methylene blue preparation not all the nerves take up the dye; in one a particular group will appear well, while in a second it may be a different group. This is particularly so in the human membrana tympani where the tissue is not usually obtained till some hours after death. In the preparation from which this drawing was made many of the nerves in the posterior superior and in the inferior anterior quadrant did not stain.


NJERVES IN THE MEMBBANA TYMPANT OF MAN 105

sel and nerve approach the manubrium the artery passes under the nerve and in the rest of its course hes adjacent to it, but nearer, to the manubrium. As the nerve bundle passes downward it frequently gives off branches which pass outwards towards the limbus, on the anterior side passing over and external to the manubrium. At a point inferior to the middle of the manubrium the. nerve bundle spreads out and while at first the majority of its fibers still have a downward direction, the general tendency and ultimate course is towards the limbus. A considerable number pass over the manubrium; of these some may be traced to the anterior limbus, but the greater number break up and get lost in the plexus over the manubrium or anterior to it. At the apex of the manubrium the fibers radiate out ; some curving round the apex,: join fibres which have crossed the manubrium, and pass out to the. anterior limbus. In short, while the main nerve is passing downwards along the manubrium it is constantly sending off on each side branches which have a general direction towards the periphery. A smaller but well marked nerve bundle which has branched off from the main trunk in the external auditory meatus enters the membrana flaccida slightly superior to the main bundle. It is directed towards the anterior superior segment, and passes sometimes directly over the processus lateralis, but always within a short distance of it.

In addition to these, numerous smaller bundles pass in from the external auditory meatus not only over the membrana flaccida, but all around the periphery. At the limbus these fibers can be noted at short irregular intervals entering under the epidermal prolongation. While the general direction of their main stem is towards the center of the membrana tensa, both external and internal to the limbus, branches are given off which form a well marked plexus, the zonular plexus, from which fibers radiate toward the center.

The radiating fibers whether coming from the nerves entering at the membrana flaccida or from the limbus give off numerous twigs and gradually get smaller. These twigs, in the main non^ medul'ated or with a very faint medullary sheath, after usually a long course during which they repeatedly divide, pass into :


106 J- GORDON WILSON

(a) a wide meshed plexus in the fibrous tissue, through wh.ich pass most of the fibers which help to form either,

(b) a subepithelial plexus under the cuticular layer, or

(c) a subepithelial plexus under the mucous layer.

From these plexuses the fibers can be traced to various endings, in some cases looping through the plexus they pass into another nerve stem and so reach their endings.

There are fibers which enter from the tympanic cavity. These, few in number compared with those from the meatus acusticus externus, come from the plexus tympanicus. They ultimately enter the plexuses in the fibrous tissue and under the mucous layer.

(a) The wide meshed plexus, the Grundgeflecht of Kessel and Deinike, is abundantly distributed throughout the whole fibrous tissue. It consists chiefly of the ramifications and interlacings of the nerves which enter from the meatus externus though fibers also reach it from the tympanic cavity. From it fibers are distributed to the subcutaneous and submucous plexuses, as well as to the endings in the connective tissue. On account of the wide meshes of which the plexus is composed it is easy to follow a single branch for a long distance. Thus in fig. 2 a branch given ofT from a nerve n entering from the meatus externus can be traced through repeated dichotomus divisions till it finally ends in a simple branched ending (c) in the sub-epithelial plexus. In addition to these some very complex endings are seen in the fibrous tissue consisting of an intricate interweaving of a frequently dividing nerve fiber (fig. 3). Into these endings there is often seen entering a second nerve (s), very fine and varicose. I have been unable to identify in man any of the plate-like endings described and figured by Deinike in the horse which lie between the radiating and circular fibers; or those endings figured by myself in the dog lying in the connective tissue in this area.

In the fibrous tissue near the periphery there are seen several varieties of endings whose size, shape and poorly developed capsule enable one to classify as modified vater-pacinian corpuscles (figs. 4 and 5). They lie immediately under the epithelium, no papillae being present in this area or in any part of the membrana tympani. In all these capsulated endings the interlacings are so


NERVES IN THE MEMBRANA TYMPANI OF MAN 107

complex and lie at such varying levels that it is impossible to give any delineation which would exactly represent what exists. To do so would only result in inextricable confusion. So it has been found advisable to draw only the most evident of these windings. Inside the capsule in fixed preparations there appears a clear space which is due to a retraction of the fine fibrils during the process of fixation. The impression given by unfixed preparations is that the ending is lying in a clear semi-fluid substance enclosed within the capsule; the clear space within the capsule does not appear. The sheath of Henle blends with the capsule. At the distal end a fine twig may pierce the capsule, divide and end in the epithelium immediately above the ending (fig. 5). At times a second, very fine varicose fibril can be seen to enter the capsule. These endings are frequently to be found in man though I have failed to find similar endings in the dog, cat or monkey, nor are they described by any of the other writers. As is known modified vater-pacinian corpuscles are widely distributed in the skin. They have been chiefly described in the tela subcutanea of the genital organs and in the conjunctiva. It is interesting to note their presence in the latter where they lie immediately under the epithelium, no papillae being present, since as will be referred to later there is a close correspondence between the nerve distribution in the cornea and membrana tympani.

The subepithelial plexus, ausseres oberfliichliches Geflecht of Deinike, consists of interlacing bundles of very fine varicose fibrils lying directly under the deepest layer of the epidermis. Man}^ of these fibrils end after a long course in the terminal part of which they run in the deeper layer of the epidermis (fig. 6.) There is thus formed an intra-epithelial plexus from which fibers pass upward towards the surface and end between the cells as fine points. These endings often appear as knobs, but it appears to me that the knob-like processes are probably artifacts produced by the dye at the terminal point, for in the best stained and sharply defined preparations the fine points are chiefly seen, whereas in the less satisfactory preparations the bulb-like point predominates. It is no unusual thing to trace a fiber a long distance through the subepidermal plexus and even through the intra-epithelial plexus,


108 J. GORDON WILSON

then back into a different nerve l)undle from the one it originally came from. These fine or bulb pointed endings are the only ones to be found in the epidermis. I have never seen any touch corpuscles. Lying in the connective tissue under the epidermis, in close proximity to the subepithelial plexus, are the branched endings shown in fig. 2.

The submucous plexus, the inneres oberflachliches Geflecht of Deinike, consists of interlacing fibers lying in the mucous layer under the epithelial lining. In close relation to it are also branched endings similar to those shown in fig. 2.

The blood vessels are abundantly supplied by non-medullated, varicose nerve fibrils. These enter along with the main vessels both over the pars flaccida (Shrapnell's membrane) and at the periphery and can be traced to the smaller vessels, forming the well known vaso-motor plexuses.

Contrary to the opinion of Kessel no ganghon cells are to be seen anyw^here in the membrana tjrmpani. Such swellings as at times appear after gold chloride impregnation are to be identified as nodal points or sheath cells.

The close analogy between the nerve distribution in the membrana tympani and the cornea was pointed out by Jacques. It will be noted that we have the zonular plexus of the former corresponding to the annular plexus of the latter, the ground plexus to the fundamental plexus, and a subepithelial and intra-epithelial plexus in both. In a former paper ('07b) I ventured the following remark: "One is tempted to carry the analogy further and to say that as in the cornea pain and not touch appears to be the sensation evoked, so also in the membrana tympani one might expect that the slightest pressure would evoke unpleasant sensations, passing into pain, a fact well borne out by clinical observations." Since then I have made many observations to test this hypothesis and find that it is undoubtedly true. By lightly touching the membrana tympani either with a small piece of cotton wool or a fine hair mechanical stimulation from the threshold possesses unpleasantness. My results have been briefly stated in a recent paper ('10a), as follows : "if this membrane be touched with a fine


NERVES IN THE MEMBRANA TYMPANI OF MAN 109

hair according as the pressure stroke of the hair is increased so do we pass from unpleasantness through acute pain in the ear to pain radiating along the n. auriculo-temporalis. This it appears to me is the impress put upon the organ through its phylogenetic history that injury will entail serious results to the individual, a case illustrating what Sherrington would probably call 'a selective adaptation attached to a specific sense of its own injuries.'" As in the cornea so in the membrana tympani there is in the epithelium but one morphological variety of nerve ending, namely, free endings lying near the surface between the epithelial cells. Moreover we have in the membrana tympani one other modified skin structure in which pain is the only species of sensation which can be evoked. From this it would apjDcar that in this structure there is additional support for Sherrington's hypothesis that the noci-ceptive organs of the skin are probably naked nerve endings.

In a former paper ('07a) there were stated the experimental data on which I based my claim that the nerve supply of the membrana tympani came chiefly from the n. mandibularis through the n. auriculo-temporalis and to a less extent from the n. vagus. Briefly these were that in dogs and monkeys, after section of the n. mandibularis at its exit from the foramen ovale and of the n. auriculotemporalis under the mandible, degeneration was observed in the nerves of the meatus acusticus externus adjacent to the membrana tympani as well as in the nerves of that membrane. Recently it has been asserted by Hunt that the chief nerve supply comes from the gangUon geniculi of the n. facialis. The chief arguments which Hunt advances in favor of this view are based on:

1. The findings of comparative anatomy that the very considerable afferent distribution of the facial in the lower vertebrates has "in the course of phylogenetic development undergone a considerable shrinkage and displacement by the n, trigeminus. A vestigial remnant in the mouth is still demonstrable and an important sensory innervation of facial origin still exists in the middle ear and in the external ear." But there is no proof that in the lower vertebrates above the cyclostomes there are sensory


110 J. GORDON WILSO^'

fibers from the facial to the skin except in fishes for the innervation of special sense organs belonging to the gustatory or lateral line system. Herrick has frequently asserted that there is no proof that the geniculate ganglion ever sends general sensory fibers to the skin; in all cases where general sensory fibers distribute to the skin by branches of the facialis they either have a separate root of their own (as in cyclostomes) or enter the facialis distal to the geniculate ganglion by anastomosing branches from the gasserian ganglion of the trigeminus.

2. The experimental results of Amabolino who after cutting the n. facialis at the stylo-mastoid foramen found retrograde degeneration in the cells of the geniculate ganglion. These results do not show as Hunt asserts that these fibers "are destined for the cutaneous distribution of the facial to the external ear." It simply proves that the facial contains afferent fibers, which may be and certainly some are nerves of deep sensation.

3. Clinical observations: (a) in the loss of sensation accompanying facial paralysis; (b) in herpes oticus with paralysis of the n. facialis.

(a) Tests for anaesthesia in this area must necessarily be unsatisfactory because of the very considerable overlapping of sensory nerves. As a result of my own observations as well as from a study of the cases quoted by Hunt I am convinced that such evidence cannot give data accurate enough to decide the point at issue.

(b) The frequency of the association of herpes oticus with paralysis of the n. facialis and with auditory symptoms appears to lend considerable weight to Hunt's hypothesis and cannot be sunnnarily dismissed. Although herpes oticus is comparatively rare yet Hunt has carefully collected a number of cases in which this association is a marked feature. But as he points out facial paralysis also occurs with herpes facialis and herpes occipito-collaris. He explains this by an associated inflammation of the geniculate ganglion "based on the well recognized tendency of this affection to produce inflammatory changes in a series of spinal ganglia. The Gasserian, geniculate and upper cervical consti


NERV'ES IN THE MEMBRANA TYMPANT OF MAN 111

tute such a serial chain" (p. 84). As the "ganglion of the auditory nerve may be primarily involved in zona" (p. 340), and so would be embraced within this series, it only remains to explain the implication of the facial from inflammatory changes in the geniculate.

I recognize the cogency of Hunt's arguments but feel that something further is required to elucidate the point in question. There are two obvious possibilities, namely, that the fibers which I have shown reach the external auditory meatus by the n. auriculotemporalis may come from the geniculate through the n. petrosus superficialis minor and the further possibility that fibers may pass from the n. facialis to the auricular branch of the n. vagus and so reach the meatus. In order to arrive at a conclusion in this possible peripheral distribution I have recently deslroyed the geniculate in dogs and monkeys so as not to involve the ninth or tenth cranial nerves nor the branches from the second and third cervical. When the series are completed I hope to publish the results.

CONCLUSIONS

The membrana tympani of man is chiefly supplied by nerves which enter from the external auditory meatus. These pass in (1) as one large trunk along with the principal artery; (2) as numerous small branches around the periphery.

These form a plexus in the fibrous tissue from which branches are distributed to a sub-epithelial and a sub-mucous plexus. In addition there are to be distinguished a zonular and an intra-epithelial plexus.

There are nerves, fewer in number, which enter from the tympanic cavity.

The blood vessels are well supplied by vaso-motor nerves.

Only one variety of nerve ending is seen in the epithelium. In the fibrous tissue both subcutaneous and submucous there are found nerve arborisations; at the periphery modified vater-pacinian corpuscles are present.

No ganglia are to be seen.


112 J. GORDON WILSON

The nerve supply comes from the n. auriculo-temporalis and the n. vagus. We have not sufficient evidence to prove that nerves reached it from the ganglion geniculi.

The sensation produced by lightly touching the membrane is pain and this is to be associated with irritation of the particular nerve endings found in the epithelium.

BIBLIOGRAPHY

Calamida, U. 1901 Terminazioni nervose nella membrana timpanica. Arch. Hal. di. Otologia, vol. 11, p. 326-329.

Deinike, D. 1905 Ueber die Nerven des Trommelfells. Arch. J. mikr. Anat.^ Bd. 66, S. 116-120.

DoGiEL. 1904 Zeitschrift. f. Wissenschaft. ZooZ., v. 45, S. 61.

Gemelli, a. 1909 Les rierfs et les terminaisons nerveuses de la membrane du tympan. La Cellula, t. 25, fasc. 1.

Hunt, T. Ramsay. 1907 The sensory system of the facial nerve and its systematology. Jour, of Nerv. and Ment. Dis.. vol. 36, p. 321-349, 1909; and vol.34.

Jacques, P. 1900 De la fine innervation de la membrane du tympan. 13 Cong, internat. du medicin. Sect. d'Otologie, p. 46, Paris.

Kessel, J. 1872 Das aussere Ohr. Handbuch der Lehre von den Geweben des Menschen; herausgeben von S. Strieker, Bd. 2, S. 853, Leipsig.

Sherrington, C. S. 1906 The integrative action of the nervous system. Pp. 227 and 319, New York.

Wilson, J. G. 1907a The nerves and nerve endings in the membrana tympani. Jour. Camp. Neur. Psych., vol. 17, p. 459-468.

19071) Jour. Coin]). Neurol. Psych., vol. 17, p. 466.

1910a I'ain in the ear and its diagnostic significance. Quarterly Bull., N. W. U. M. S., vol. 11, pp. 211.

lUlOlj Intravitan staining with methylene blue. The .\naloinicnl Record, vol. 4, p. 267-277.


PLATES


in.AT]<: 1

EXPLANATION OK FIGUKK

1 Nerve distribution in the right membrana tyinpani of man. The pars flaccida with the head of the malleus has been removed. The nerves were stained with methylene blue and the specimen was mounted in Canada balsam. During the process of mounting the manul>riui mallei was twisted so that the processus lateralis (/^/.) is tilted posteriorly. The main nerve trunk {n.t.m.) together with the artery (a.m.t.) is seen entering over the posterior superior cjuadrant. Not all the peripheral nerves {p.n.) took up the dye but a sufficient number to give an ade(|uate representation of their mode of distribution at the limbus.

/. a. — limbus membranae tympani anterior

I. p. — limbus membranae tympani posterior


NERVES IN THE MEMBRANA TYMPANI IN MAN

J. GORDON WILSON


PLATE 1


ntm amt



THE AMERICAN JOURNAL OF ANATOMY, VOL. 11, NO. 2


PLATE 2

EXPLANATION OF FIGURES

2 Course of a nerve passing in from the periphery to end in the suii-ej)ithelial plexus. From the peripheral meduUated nerve (n) a branch (6) is given off near the linil)us. From (6) numerous twigs branch off in the fibrous tissue with a common direction towards the sub-epithelial plexus. In the course of division the medullary sheath gets fainter and finally disappears. Terminal branches (c) were seen Ijreaking up in the sub-epithelial plexus. Zeiss comp. oc. 4, obj. 8 mm. and obj. 2 mm.

3 A branched non-capsulated ending in the fibrous tissue of the membrana tympani. The nerve (n) breaks up into a dense mesh work of fibers from which off-shoots go forward into the adjacent connective tissue. A fine varicose hbril (.s) passes into the network and is there lost. Zeiss comp. oc. 4, obj . 3 mm.


NERVES IN THE MEMBRANA TYMPANI IN MAN

J. GORDON WILSON


PLATE 2


n



X«tk^r. ne H . \\


THE AMERICAN JOURNAL OF ANATOMY, VOL, 11, NO, 2


PLATE 3

EXPLANATION OF FIGURES

4 Modified vater- pacinian corpuscle at the periphery of niembrana tj'mpani of man. A nerve (n) is seen entering the capsule surrounded by the sheath of Henie (.// ) which blends with the connective tissue of the capsule (c) . The nerve as soon as it enters the capsule begins to send oflf numerous fine fibrils, which wind in and out at varying depths interlacing freely among themselves. The main trunk, gradually getting thinner, can be followed to near the distal end in a tortuous course through the intricate interlacing of the fibrils. Before entering the capsule the nerve has a faint medullary sheath which it loses before piercing the capsule. The capsulated sheath is poorly developed, but can be easily distinguished in the orilinary preparation. It is well brought out with eosin or acid fuchsin.

5 A modifipd vater-pacinian corpuscle. A fiber (r?) with at first a faint medullary sheath runs into the thin capsulated ending. The breaking up is by dichotomous divisions; the first division takes place about one-fourth of the length of the ending from its proximal part. The numerous rami freely interlace and finally end as fine bu]l)()us points, apparently just under the capsule. The form of the ending is long and spindle shaped. It is divided into two nearly equal parts by a connect ive tissue diaphragm (d) which passes from the inner side of the capsule. The proximal half is entirely supplied by the first division of the nerve; the other half is distributed to the distal half. From the distal end a fine fibril (n') passes out through (he (;af)sule to end, after dividing, in the sub-epit hiMial tissue.

6 Sensor}' endings in the epithelium of the membrana tympani of man. Several non-medullatc(l nerves, some. of which can be followed for a considerable distance in the sub-epithelial plexus, run between the deeper layers of the epithelial cells on the external surface of the membrana tympani, interlacing and forming an intra-epithelial plexus from which branches go to the surface as fine or bulb-shaped points. Zeiss comp. oc. 4, obj. 2 mm.


NERVES IN THE MEMBRANA TYMPANI IN MAN

J. GORDON WILSON


PLATE 3


n



M..i"a-;n« Hill


THE AMERICAN JOURNAL OF ANATOMY, VOL. 11, NO. 2


DESCRIPTION OF A MODEL SHOWING THE TRACTS OF FIBRES MEDULLATED IN A NEW-BORN BABY'S BRAIN

FLORENCE R. SABIN

From the Anatomical Laboratory of the Johns Hopkins University

ELEVEN FIGURES

In 1900 I published a description of a model of the medulla, pons and mid-brain made from a series of sections of the brain stem of a new-born babe.^

In this model were shown first the form of the main nuclear masses and secondly, the form of the fibre tracts that had received their medullary sheaths, thus the study was based on the fact that the tracts of the central nervous system are successively medullated, as has been worked out by Flechsig, and was the result of the idea that at birth so few of the tracts have received their medullary sheaths, that it is possible to model them out and obtain a picture of their form. A clear picture of the form of such tracts as the lemniscus medialis and fasciculus longitudinalis medialis, even though it represent only one stage of development is of value from two standpoints, first in pointing the way toward a picture of the tract in the complicated relations of the adult brain, and secondly in leading toward an understanding of the modifications in form of the tracts as they shift in their relations in development.


1 Sabin, A Model of the Medulla Oblongata, Pons and Mid-brain of a New-Born Babe. Contributions to the Science of Medicine dedicated by his pupils to William Henry Welch, Johns Hopkins Press, 1900; and Johns Hopkins Hospital Reports, vol. 9; and Atlas of the Medulla and Midbrain, The Friedenwald Company, Baltimore, 1901.


The American Journai. of Anatomy, Vol. 11, No. 2


114 FLORENCE R. SABIN

The series of sections from which the model was made, extended only just through the region of the mid-brain. It was the hope at that time, to procure a series of sections of the entire brain of the new-born babe so as to complete the modeling of the tracts into the thalamus and cerebrum. The especial goal was to throw some light on the form relation of the internal capsule. Since that time five series have been cut in this laboratory, three of the newborn brain, and two of the brain stem and basal ganglia of the adult. Of these series, three are in the sagittal plane, while one of the baby's brain and one of the adult were cut in transverse series. The sagittal series of one of the adult specimens was cut by Miss Gertrude Stein. It has proved most valuable. Miss Stein made a model of the medullated tracts in one of the series of the new-born brain, but it proved that the gaps in the series were sufficient to prevent an interpretation of the model . From the series on which the present study is based, two models have been made, the first one was by Dr. E. G. Gowans, of Salt Lake City, who modeled the thalamus and basal ganglia on one side. Unfortunatel}^ Dr. Gowans' work was unavoidabl}^ interrupted and as there was little chance that he would be able to take it up again, I have completed it. In studying Dr. Gowans' model I found it hard to interpret the bundles in the thalamus without including those of the brain stem, and hence made a new model of the brain stem and thalamus of the other side of the same brain. The figures are all taken from my model, but the study covers all the work done in this laboratory on this topic by Miss Stein, Dr. Gowans and myself. I am indebted to Miss Stein for series she prepared, and to Dr. Gowans both for the model and for his study of the subject by which I benefit. The specimen from which the model was made was brittle and while in thick cellodin the brain stem cracked from the thalamus. The cellodin, however, was thick enough to hold the two parts in exact apposition, as can be seen in figs. 5 and 7. The plane of the section is oblique as can be seen in the same two figures, since the raphe is included in the sagittal section (see no. 61, fig. 5). This only increases the difficulty in the pihng of the model but does not affect its accuracy. I believe that the model does throw some


MODEL OF MEDULLATED TRACTS IN BABY's BRAIN 115

light on two points, first, the form relations of the medial lemniscus and the fibres of the red nucleus in their relations to the thalamus, and secondly, on the form of the internal capsule in relation to the thalamus and corpus striatum, notwithstanding the fact that much of the internal capsule is non-medullated.

The model is shown in a series of five drawings, three from the mesial aspect and two from the lateral. Fig. 1 represents a mesial view of the entire model and fig. 9 a lateral view. The other figures, namely, 3, 4 and 1 1 , represent dissections of the model to show the various structures. The nuclei are represented in a solid tone, the non-medullated bundles are faintly streaked, while the medullated bundles have a blue tone. Fig. 2 is a drawing of the mesial aspect of a baby's brain, given for the purpose of orienting the first three figures of the model, while fig. 10 is a view of a dissection of the pyramidal tract given to orient the lateral views of the model. Both figs. 2 and 10 are from a baby's brains of which, however, we have no records of the age. It is obviously older than the one from which the model was made, in which the pyramidal tract is not at all medullated. Four figures of sections are given, two from the baby's brain from which the model was made, and two taken at about the same level from an adult brain.

In the description of the model I shall follow this plan. First, I shall take up the medullated tracts beginning with the lemniscus medialis and the group of tracts associated with it. The various nuclei will be mentioned in connection with the fibre bundles. Secondly will follow a description of the form of the internal capsule, both the non-medullated and the medullated part, and thirdly, an account will be given of the nuclei of the thalmus, hypothalamus and corpus striatum, with the associated medullated bundles. The following table gives a list of all the medullated tracts found in the series in approximately the order in which they are taken up.


116 FLORENCE R. SABIN

A. A group of fibres making up the main sensory tract to the cortex together

with certain tracts associated with it by j)osition: I. Seen from the mesial aspect —

1. Lemniscus medialis and rubro-lenticular tract (or fasciculus

lenticularis of Forel)

2. Fasciculus gracilis and fasciculus cuneatus, — not modeled

3. Tracts from no. 2, forming the lemniscus medialis, namely:

(a) Decussatio lemniscorum

(b) Fibrje arcutse internse

4. Lemniscus lateralis

5. Lemniscus superior

6. Tract to the substantia nigra

7. Tract to the nucleus hypothalamicus (Corpus Luysi)

8. Tract connecting the nucleus colliculi inferioris with the center

median of Luys

9. Tract connecting the nucleus hypothalamicus of Luys with the

globus pallidus IL Seen from the lateral aspect —

L Tract from the ventro-lateral nucleus of the thalamus to the upper

third of the posterior central gryus 2. Lenticular nucleus to the same zone

B. The fasciculus longitudinalis medialis and commissura posterior

C. Miscellaneous:

\. Fasciculus retroflexus of Meynert

2. Corpus restiforme or inferior cerebellar peduncle, not modeled

3. Brachium conjunctivum or superior cerebellar peduncle

4. Rubro-pontal tract

5. Formatio reticularis, — not modeled

6. Cranial nerves, — not modeled

7. Small bundle at root of optic nerve

8. Optic fibres in relation to the lateral genticulate body

9. Certain fibres associated with the substantia nigra, the nucleus

hypothalamicus of Luys and the globus pallidus of the lenticular nucleus.

The lemniscus medialis is shown best in the three views of the model from the mesial aspect, figs. 1, 3, and 4. The lemniscus begins at the transition between the spinal cord and the medulla, at the point where the decussatio lemniscorum crosses the middle line. In the series studied, the posterior columns of the cord (no. 2, fig. 7), were medullated but were not modeled. The decussatio lemniscorum is a small compact bundle (no. 3, fig .1), which begins in the nucleus funiculi gracilis, curves around the edge of the gray matter bordering the central canal, and crosses the middle


MODEL OF MEDULLATED TRACTS IN BABY's BRAIN 117

line at a point marked 4 in fig. 1. This point is ventral to the central canal, just caudal to the inferior olive. Within the medulla, the bundle from the nucleus funiculi gracilis is joined by the more diffuse mass of internal arcuate fibres from the nucleus funiculi cuneati (no. 5, fig. 1.) These two crossed tracts make the lemniscus in the medulla, which is a broad band as seen from the mesial aspect. It lies close to the raphe, hemmed in between the two inferior olives. In the pons, the lemniscus spreads out to the side making a narrow b'and in the mesial plane. In the lower part of the pons can be seen the position of the superior olive (no. 7, fig. 3), quite far to the side, and the lateral lemniscus emerging from it. Throughout the pons, the medial and lateral lemnisci are parallel. At the cerebral end of the pons, the lateral lemniscus curves still farther lateralward, and turns dorsalward to enter the nucleus of the inferior colliculus (no. 8, fig. 3).

The medial lemniscus, on entering the mid-brain shows a number of interesting points. In the first place, it gives off a small bundle which enters the caudal tip of the substantia nigra (no. 9, figs. 3, and 7). In using the term enters, it is not intended to assume the direction of the fibres since Weigert sections cannot give the relation to the cell body. Secondly, from the dorsal border of the medial lemniscus is given off the superior lemniscus which curves dorsalward parallel to the lateral lemniscus to enter the superior colUcuhis (no. 10, figs. 3 and 4). Thirdly, the entire mass of the medial lemniscus curves lateralward to make room for the red nucleus. This can be seen in fig. 1, but better in fig. 3, and best in fig. 4, from which the red nucleus has been removed. Lateral to the red nucleus, the lemniscus gives off a small bundle to the nucleus hypothalamicus of Luys, exactly similar to the one given off to the substantia nigra (no. 11, figs. 3, 7, and 8).

The relations of the lemniscus to the thalamus are complicated. It has been found that the fibres run to two nuclei, namely the center median of Luys, and the ventro-lateral nucleus. Here it may be well to enumerate the nuclei of the thalamus w^hich can be made out and modeled, though they will not be described until later:


118 FLORENCE R. SARIN

1. The nucleus medialis

2. The nucleus anterior

3. The center median of Luys and the cup-shaped nucleus

4. The nucleus ventro-lateralis —

Medial part receiving the medial lemniscus Lateral part connected with the cortex

5. The nucleus dorso-lateralis not separated from the pulvinar

6. The pulvinar

7. The nucleus corporis geniculati medialis

8. The nucleus corporis geniculati lateralis

In stud3'ing the relations of the lemniscus to the thalamus, I have found the work of ForeP most helpful and shall use his familiar terms. At the very beginning, however, it will be well to be clear about directions in the thalamus. In the first place, the brain stem of the baby is more nearly at right angles to a longitudinal axis of the cerebrum itself than that of the adult. In describing the thalamus the difficult}^ of orientation is familiar to all students of the subject and is due to the fact that the terms dorsal and ventral as applied to the cerebrum itself are not the same as dorsal and ventral for the brain stem, and the thalamus comes at the point of transition. T shall use the terms dorsal" and "ventral" in the thalamus in exactly the same sense as in the brain stem, and hence shall speak of the lenticular nucleus as ventro-lateral to the thalamus. It will be very necessary in following Forel's desciption to note that his sections are frontal to the adult brain, and hence oblique in the baby's brain. If, for example, a line be drawn through the pons, the red nucleus and the center median of Luys in my fig. 4 at about the direction indicated by lines 53, it will correspond in general to Forel's fig. 5, plate 7. Therefore "dorsal" and "ventral," in the sense of Forel's figures, are at an oblique angle to mine.

To return to the relations of the medial lemniscus to the thalamus, as the lemniscus curves around the red nucleus (no. 1 3) , it becomes impossible to separate it from fibres either from the red nucleus itself or from fibres of the brachium conjunc ^ Forel, Untersuchungen ueber die Haubenregion und ihre oberen Verkniipfungen im Gehirne des Menschen und einiger Saugethiere. Archiv fiir Psychiatrie und Nervenkrankheiten. Bd. 7, 1877.


MODEL OF MEDULLATED TRACTS IN BABY's BRAIN 119

tivum passing through the red nucleus. Emerging from the cerebral end of the red nucleus is a great mass of medullated fibres (see fig. 3), which soon divides into two bundles, a dorso-lateral mass which is Forel's BaTh, (no. 12), and a more ventral and medial bundle which is Forel's Feld H. The dorso-lateral bundle which is Forel's BaTh is a composite of possibly three elements, first, a part of the medial lemniscus, secondly, possibly fibres from the red nucleus and thiidly, a small bundle (no. 16), which comes from the nucleus colliculi inferioris and runs to the center median of Luys. This bundle is readily seen in its connections in the section of the baby's brain in fig. 7, no. 16. It is readily identified in sagittal sections of the adult by the angle the bundle makes with the main mass of BaTh (see no. 12, fig. 7). The section of the adult brain (fig. 8), is, however, just too far lateral to show the two bundles as separate tracts. The bundle from the inferior colliculus to the center median of Luys is described by Dejerine,^ under the confusing name of "the arm of the inferior colliculus," a term which should be reserved for the fibres connecting the inferior colliculus with the medial geniculate body. Dejerine, however, makes the distinction, in fact, and notes, that the bundle in question enters the center median of Luys. The true brachium quadrigeminum inferius connecting the inferior colliculus with the medial geniculate body is non-medullated at birth, and lies farther lateralward so that it is seen only from the side (no. 25, fig. 9).

The bundle BaTh (no. 12), then consisting of a part of the leminscus, the tract from the inferior colliculus and possibly of fibres from the red nucleus enters the cup shaped nucleus (no. 15, figs. 8 and 11), and the center median of Luys (no. 14). In connection wdth these nuclei it is interesting to note that Sachs^ found that in degeneration experiments involving the center median of Luys and the cup-shaped nucleus, there were no efferent fibres to the cortex nor to the mesencephalon; that the efferent fibres all ran to the other nuclei of the thalamus,

  • Dejerine, Anatomic d. centres Nerven, Tome 2, 1901, p. 72.
  • Sachs, On the Structure and Functional Relations of the Optic Thalamus.

Brain, August, 1909, vol. 32, page 146.


120 FLOllENCE R. SABIN

those from the cup-shaped nucleus reaching all the thalamic nuclei except the anterior nucleus, those from the center median of Luj^s, reaching all except the anterior and medial nuclei. If these conclusions are correct, we may say that at birth, only the afferent fibres of these two nuclei are medullated and these afferent fibres come from the medial lemniscus, the inferior colliculis and possibly the red nucleus. There are no other medullated bundles associated with these nuclei at birth.

The rest of the fibre mass (no, 17), which emerges from the red nucleus is likewise a composite mass. It lies farther ventral and more medial than BaTh and corresponds to Forel's Feld H. It is a mass of fibres, oval in cross section (see Forel's fig. 11, plate 7), and made up of two parts, as can be seen in fig. 4, a lateral part (no. 18), Forel's Hi, which appears to be a direct continuation of the medial lemniscus, and a medial part (no. 19), Forel's H2 which appears to emerge from the red nucleus, as can be best seen in figs. 1 and 3 of the model and in section in fig 6. The lateral part (no. 18), divides into a forked bundle which enters the ventro-lateral nucleus of the thalamus. The more ventral bundle of the fork enters the external medullated lamina of the thalamus, as can be seen in the section of the adult brain in fig. 8. In the different figures the number is placed at the fork of the Y or on one or both branches. This is an exceedingly im portant bundle since it is a part of the main cortical path, which is now generally thought to consist of these three elements, the posterior columns, the medial lemniscus, and the thalamo-cortical radiation. In the model it will be readily seen that the ending of the main lemniscus bundle in the thalamus agrees with von Monkow's^ description that the lemniscus ends in the caudal ventral part of the great lateral nucleus of the thalamus. This ending of the lemniscus has been confirmed by Mott," more recently by Ramon y Cajal, quoted from Sachs, and by Sachs.' Sachs speaks of the ending of the lemniscus as in the "ventral third and lower (i.e., caudal) half of the middle third of the lateral

5 von Monakovv. Gehirnpathologie; Nothnagel, Specielle Pathologie und Therapie, Wien. Zweite Auflage, 1905, S. 90-91. « Mott, Brain, 1895. ^ Sachs, il)i(l., page 130.


MODEL OF MEDULLATED TRACTS IN BABy's BRAIN 121

nucleus." By comparing the three views of the model from the mesial aspect it will be noted that this forked bundle of the lemniscus enters the medial part and the caudal surface of the ventrolateral nucleus. In fig. 1 is shown the rather thin medial nucleus, especially thin along the ventral border. This has been removed from figs. 3 and 4, showing the large lateral nucleus. The size is better estimated in fig. 4, from which the lateral nucleus itself has been entireh' removed. A form relation of the medial lemniscus can be especially well seen in figs. 9 and 11, namely, that the medial lemniscus, throughout the superior colliculus region, lies close to the lateral surface in the groove between the crus and the colliculi, but the bundle being straight, it does not enter the extreme lateral part of the thalamus which projects so far to the side of the mid-brain.

The brain at birth then shows that the medial lemniscus, as far as it is medullated, enters two nuclei in the thalamus, the center median of Luys and the cup-shaped nucleus considered as one and the medial part of the ventro-lateral nucleus as the the other. Moreover, in connection with the ventro-lateral nucleus, the medullated bundle forms a part of the external medullated lamina. These relations will be clear by comparing figs. 7 and 8.

The relations of the more ventral part of Forel's Feld H, namely, his fasciculus H2, are especially well shown in the model. The bundle (no. 19, and its continuation, no. 20), appears to arise in the red nucleus. It emerges from the red nucleus, at its cerebral end, close to the mid-line, and, as is seen in figs. 3 and 4, takes the following course; it passes forward dorsal to the the nucleus hypothalamicus of Luys to a point just cerebralward to Luys' body, where it curves sharply ventralward, along the medial border of the crus, and then turns first caudalward and then directrectly lateralward along the caudal surface of the first part of the globus pallidus. Within the globus pallidus this bundle is perfectly sharp in sagittal sections of the adult brain as can be seen by comparing no. 20 in figs. 7 and 8. It lies on the caudal surface of the globus pallidus, adjacent to the optic nerve. This bundle was described by Forel as H^, and is called by Dejerine^ Le fais ^ Dejcrine, ibid., vol. 2, page 327.


122 FLORENCE R. SARIN

ceau leniiculaire de Forel. In view of its relations, it would seem to me that it might be termed the fasciculus rubro-lenticularis of Forel. If one bears in mind the oblique plane of Forel's sections as compared with the models it will be seen that his figures 17 and 18, plate 8, which run in the direction of the fibres of bundle 19 in my fig. 4, are in the very best plane to show the course of the lenticular fasciculus around the medial edge of the crus and into the globus pallidus. The same sections must also show, as the}^ do, the lemniscus fibres within the external medullated lamina of the thalamus.

In this connection it will be well to make clear that the rubrolenticular fasciculus of Forel (^2) is not the same as von Monakow's^ dorsal Antheil der Linsenkernschlinge notwithstanding the fact that he identifies it with Forels' H2. From a study of von Monakow's figures 21-30, it is clear that he is dealing with two different bundles both of which he calls H2. The first, which he identifies with Forel's H2, and labels mC] , turns into the globus pallidus around the lateral border of Luys' body while the true Forel's fasciculus as shown in Forel's figures and in the models curves around the mesial border of Luys' body, and is shown in von Monakow's figs. 30-34, labeled H2 and Lisch C. The bundle which von Monakow calls the dorsal Antheil der Linsenkernschlinge is also medullated at birth. It connects Luys' body with the globus pallidus, but cuts directly through the crus at the lateral border of the nucleus hypothalamicus of Luys, instead of curving around the medial edge of the crus. This bundle will be considered in a moment, but it is not the same as the more medial fasciculus rubro-lenticularis which is not connected with the Luys' body. Obersteiner^" identifies the bundle which curves around the lateral border of the nucleus hypothalamicus of Luys as Forel's H2 thereby agreeing with von Monakow. To the bundles which in the adult, curve around the mesial border of the crus and connect with the lenticular nucleus, Obersteiner gives the general

' von Monakow, Experimentclle unci pathologisch-anatomisch. Untersuchnngen uober die Ilaubenrcgion, den Sehhiigel, und die liegio-subtiialamica. Arch, f. Psych, u. Nervenkr., 27, 1875, S. 29.)

'" Obersteiner, Anleitung beim Studium des Baues der Nervosen Centralorgane Vierte Auflage, Leipzig und Wien, Franz Dcuticke, 1901, S. 377 u. 559, figs. 168 — 220.


MODEL OF MEDULLATED TRACTS IN BABY's BRAIN 123

name Him schenkelschlinge. He identifies an ansa peduncularis (fig. 377), an ansa lenticularis and an unterer Thalamus stiel (fig. 220) ; in the brain at birth only one bundle curving around the mesial border of the crus is medullated, namely, the rubro lenticular fasciculus which, I think should be identified with Obersteiner's ansa peduncularis .

In connection with Forel's field, Sachs gets a degeneration of Forel's fasciculus from a lesion anterior to the red nucleus and terms the bundle the tractus rubro-globus pallidus which is probably the same as no. 19. He also finds a still more medial bundle, the inferior peduncle of the thalamus, which is entirely nonmedullated at birth. It may be added here that the ansa lenticularis is also non-medullated at birth.

The bundle just referred to in connection with the nucleus hypothalamicus of Luys, is not well illustrated in any of the views of the model, since it lies lateral to Forel's fasciculus no. 19 and is nearly hidden by that bundle in the drawings. The hypothalamic nucleus is oval or lens-shaped, both in sagittal section (fig. 8), and in the familiar transverse sections such as are shown in Forel's and von Monakow's figures. It has a capsule of fibers, especially marked at the cerebral end (no. 23, fig. 7). At the lateral border of the Luys' body, the fibres of this capsule cut through the crus, to the globus pallidus. Within the globus pallidus, these fibres lie close to the crus (no. 30, fig. 4) and are distinctly separated from the Forel's rubro-lenticular fasciculus (no. 20, figs. 4 and 7). For this bundle connecting the Luys' body and the globus pallidus, transverse sections are far better than sagittal. This is the bundle referred to by von Monakow as the dor sale A^itheil der Linsenkerschlinge. It shows well in Forel's fig. 12 and von Monakow's fig. 26 as the dorsal capsule of Luys' body, lying ventral to the zona incerta and curving through the crus to the globus pallidus. In the model, Luys' body is seen to be connected with the medial lemniscus (no. 11) on the one hand, and with the globus pallidus, on the other hand, by the medullated bundle (nos. 23 and 30) just described.

To sum up, in the course of the afferent paths to the cortex, as far as they are medullated at birth, there are two great relay


124 FLORENCE R. SARIN

stations in the basal ganglia, the venti'o-lateral nucleus of the thalamus and the globus pallidus of the lenticular nucleus. The paths are first the medial lemniscus. The medial lemniscus is the indirect continuation of the posterior columns of the cord. It arises in the medulla from two bundles of crossed fibres, one from the nucleus of the fasciculus gracilis and the other from the nucleus of the fasciculus cuneatus." It passes through the pons without giving off any bundles; in the mid-brain, it gives fibres to the substantia nigra, and the superior colliculus ; in the thalamus it ends in the center median of Luys the cup-shaped nucleus and the ventro-lateral nucleus of the thalamus. Secondly, there are two systems of medullated tracts which connect the inner division of the globus pallidus with lower centers. One of these systems of fibres may be connected with the cerebellum, since the inferior cerebellar peduncle or corpus restiform is partly medullated, connecting the cord and the cerebellum, the superior cerebellar peduncle or brachium conjunctivum is medullated connecting the cerebellum v/ith the red nucleus, and Forel's lenticular fasciculus seems to connect the red nucleus with the globus pallidus . The second connection of the inner part of the globus pallidus is by the tract making a capsule for the hypothalamic nucleus of Luys which is also connected with the medial lemniscus. Thus the model shows two nuclei as relay stations for cortical paths the medial part of the ventro-lateral nucleus of the thalmus, which is known to be a relay station in a sensory path, and the medial part of the first division of the globus pallidus which is, however, not so definitely known as a part of a sensory path. Flechsig regards this lenticular cortical path as ascending. ^^ The cortical radiation as far as it is medullated at birth comes from these same two nuclei, but from their lateral portion. Thus there is a medullated bundle between the lateral part of the lateral nucleus (no. 26) of the thala " The decussating bundle from the sensory nucleus of the trigeminus is not present in this sorios. It was shown in my first model lying in the floor of the fourth ventricle just cerebralward to the nucleus n. absducentis. It was searched for with great care in the present series since Lewandowsky, in the Neurobiologischc Arbeiten, herausgegeben von Oskar Vogt, Erster Band, Zweite Lieferung, 1904, page 93, has shown its importance as making the third bundle to form the lemniscus medialis.

" Flechsig, Archiv. f. Anat. u. Phys., Anat. Ablh.


MODEL OF MEDULLATED TRACTS IN BABY's BRAIN 125

mus, and the cortex, and between the lateral surface of the globus pallidus (no. 27) and the cortex. These two bundles make up the cortical radiation medullated in the specimen (no. 28). The gap between the lemniscus and the cortical bundle is best seen in fig. 4, the lemniscus being numbered 18, and the thalamo-cortical bundle, 26. For the globus pallidus, the fibres to the medial part are seen as 19 in fig. 4, while the cortical radiation no. 28 is far to the side and seen best in fig. 11.

The medullated part of the cortical radiation needs special description. It consists, as has just been said, of two parts, one from the lateral zone of the thalamus (no. 26, figs. 4 and 11), the other from the lateral part of the globus pallidus (no. 27, fig. 11). Beyond the border of the thalamus as seen in fig. 4, these two bundles are absolutely indistinguishable, and the fused mass makes the cortical radiation labeled no. 28 in all the figures. The thalamic part of the radiation is the less extensive at birth. It is shown best in fig. 4 as no. 26. Here it will be seen to be a curved bundle in the form of a figure 6. The caudal part which, from the point of view of the cerebrum proper, is the ventral part, curves around the lateral nucleus of the thalamus ; the stem of the 6 lies in the external medullated lamina of the thalamus, which gradually joins and fuses with the general cortical radiation. In sections this thalamo-cortical bundle which is medullated at birth is well illustrated in the literature, in the new-born, by Barker^^ in his fig. 664 This section is from one of our series and shows the thalamic radiation labeled Th, in its positon just cerebralward to the lateral geniculate body The same figure shows the globus pallidus w^ith its capsule of fibres which makes the lenticular part of the cortical radiation. In this series the cortical radiation extends both to the anterior as well as to the posterior central convolution while in the one from which the model is made, has only the posterior central bundle medullated. In the adult, the first thalamo-cortical bundle to be medullated is illustrated by Dejerine^^ in his second volume (page 310, fig. 282). The caudal curved part of the bundle (see fig. 4), which represents the fibres emerging

1^ Barker, The Nervous System. New York, D. Appleton & Company, 1899. First edition, page 1051, fig. 664. " Dejerine, ibid.


126 FLORENCE R. SARIN

from the nucleus, Dejerine describes as the triangular zone of Wernicke (see page 357).

The medullated cortical radiation occupies the postertior limb of the internal capsule, and the models bring out certain fundamental points about the internal capsule which can be made clear from three figures, namely, 4, 9 and 11. To begin with fig. 4, the internal capsule consists of, two segments, the anterior limb no. 31, entirely non-medullated at birth, and the posterior limb, no. 28, partially medullated at birth. These two sheets of fibres, the anterior and posterior limbs, stand at an angle to one another; the anterior limb is at a slight angle to the median sagittal plane of the brain proper as can be readily seen in figs. 4 and 1 1 . This point would of course be seen better in a view of the model taken from the dorsal surface of the cerebrum, but it is a point well-known in the adult. The entire posterior limb on the other hand in the brain at birth is in a plane exactly parallel to the median sagittal plane of the cerebrum and hence is a perfectly flat sheet in figs. 4 and 11. The anterior limb radiates out betw^een the caudate nucleus and the lenticular nucleus. In fig. 4, if the eye follows the crus or cerebral peduncle (no. 24) forward, it will be seen that it spreads out into a sheet of non-medullated fibres between the lenticular nucleus and the thalamus. This sheet of non-medullated fibres which lies in the primitive groove between the diencephalon and the telecephalon is the knee of the internal capsule and contains the non-medullated pyramidal tract. The posterior limb of the internal capsule extends between the knee of the internal capsule (no. 36) and the tail of the caudate nucleus (no. 32), as it curves around into the roof of the lateral ventricle. These limits are shown best in fig. 9. This point can be readily related to the adult brain in any dissection of the ventricles, for the knee is determined by the foramen of Monroe and the tail of the caudate is readily seen.

The zone in which the posterior limb of the internal capsule reaches the cortex is an especially interesting point. Since the posterior limb is a perfectly flat sheet in the model in a plane exactly parallel to the median sagittal plane, it reaches the cortex in the upper third of the central convolutions. Thus in a trans


MODEL OF MEDULLATED TRACTS IN BABY's BRAIN 127

verse section the posterior limb corresponds to the lateral border of the tail of the caudate nucleus, and the fibres first medullated are projected straight to the cortex in a sagittal plane parallel to the median sagittal plane and exactly midway between the edge of the lateral ventricle and the surface of the island of Reil. In the adult brain the plane of the primitive posterior limb of the internal capsule can be demonstrated by making a sagittal section along the lateral border of the caudate nucleus to the point where it curves around into the roof of the lateral ventricle. In all of the models there is a curious band labeled no. 37" on the cortical radiation at the base of the central sulcus. This band shows both in Barker's and in Flechsig's figures, quoted above, and in all our series, thus it is probably a constant. I think that it is due to the fact that the fibres curve lateralward just at the base of the central sulcus, and from dissections of the fibres in the adult brain I think that this bend in the fibres in due to antero-posterior bundles of fibres near the median plane. With the exception of this slight bend at the base of the central sulcus the primitive posterior limb of the internal capsule consists of straight fibres, I.e., those that take the shortest possible course.

In the baby's brain, the lenticular radiation covers the entire posterior limb (see fig. 4), while the thalamic radiation occupies the middle third (see also fig. 4). In this particular specimen the combined radiation extends only to the posterior central convolution; in the series shown in Dr. Barker's figures, quoted above, also from a new-born, the radiation reaches the anterior central convolution as well, and this is true in brain figured by Flechsig^^ from the a baby born one-half a month prematurely, and which lived three weeks. I conclude, therefore, that the series from which the model was made has a less extensive cortical radiation than the usual brain at birth, and that, to represent the average, the medullated bundle should be extended into the anterior central convolution in the line 53 on fig. 4. From the series which we have, I do not think that any of this radiation, even the part to the anterior central convolution, belongs to the pyramidal tract, the latter

1^ Flechsig, Einige Bemerkungen ueber die Untersuchungsmethoden der Grosshirnrinde, insbesondere des Menschen. Arch. f. Anat. u. Phys., Anat. Abth., 1905. Taf. 16, fig. 8.


128 FLORENCE R. SARIN

being entirely non-medullated. I have, however, not sufficient evidence to determine whether this globus pallidus — cortical radiation is ascending or descending. Flechsig regards the entire radiation both to the anterior and to the posterior central convolutions as ascending. It is, I think, important here to identify the medullated projection bundles shown in the models with those described b}- Flechsig. In Flechsig's summary as given on page 369, his a bundle, consisting of a very few fibres from the globus pallidus (or possibly from the substantia nigra) to the upper third of the central convolutions and distinguished by early medullation, can not be separated from the rest in my series. The |3 bundle from the ventro-lateral nucleus to the upper third of the central convolutions is my thalamo-cortical radiation (no. 26); while the /S bundle, consisting of the large mass of fibres from the globus pallidus to the same zone of the cortex, is the same as my no. 27. The combined thalamo-and lenticular-cortical radiation in one of our series reaches only the post central convolution ; in the others, it reaches the anterior central as well as it does in Flechsig's series. The rest of the medullated bundles of projection fibres described bj^ Flechsig in foetuses up to 50 cm. long, and in the new-born, are not medullated in our series.

There is one more bundle of fibres which, though, non-medullated, has a form relation to the internal capsule, namely, the stria medullaris thalami. In figs. 3 and 4, it shows as a band of fibres (no. 54), extending from the zone of the knee (no. 36) of the internal capsule along the border of the thalamus adjacent to the caudate nucleus. Its position with reference to the anterior nucleus of the thalamus is plain in fig. 1.

Thus, in the central nervous system in one specimen of a newborn babe we have medullated bundles which may be grouped into the following tracts. First, a sensory tract involving three elements, the dorsal columns of the cord to the nuclei of the dorsal columns, the medial lemniscus from the nuclei of the dorsal columns to the ventro-lateral nucleus of the thalamus, and the thalamo-cortical radiation from the ventro-lateral nucleus of the thalamus to the upper third of the posterior (possibly anterior as


MODEL OF MEDULLATED TRACTS IN BABY's BRAIN 129

well) central convolution. Secondly, a tract having as one element the globus pallidus with cortical radiations reaching the same part of the cortex; the other connections of the globus pallidus bring two different medullated bundles the one relating it to the red nucleus and thus to the cerebellar tracts, the other relating it to the nucleus hypothalamicus of Luj^s, which in turn is connected with the medial lemniscus at birth.

This completes the description of the bundles medullated at birth, which make a part of the projection system. It will now be necessary, in order to present a complete record of this specimen, to describe the other bundles which are medullated though the model does not, for the most part, show any point in their form relations not shown in the previous model. The first of these tracts is a long tract, namely, the fasciculus longitudinalis medialis, whose relations as an optic-auditory reflex path are certainly, in part, brought out by the tracts medullated at birth. The rest of the medullated bundles can be regarded only as fragments of tracts in the present state of our knowledge. The fasciculus longitudinalis medialis (no. 1) is essentially the same in this model as in the first model. That it is a continuation of the ventro-lateral funiculus of the cord can be seen in fig. 1. Almost its entire extent is shown also in fig. 2. In entering the medulla, it curves dorsalward with the central canal, to the floor of the fourth ventricle. Here it shows a slight curve corresponding with the pontal flexture, which is still visible. Within the midbrain, it forms a deep trough (no. 39, fig. 1), in which lies the nucleus of the oculomotor nerve. Just beyond this trough it is joined by the posterior commissure (no. 40, fig. 1), which decussates just dorsal to the aqueduct and joins the fasciculus longitudinalis medialis opposite the cerebral end of the red nucleus. The fascicluus longitudinalis medialis curves ventral ward just in front of the red nucleus and comes to an end abruptly.

The fascilulus retroflexus of Meynert is one of the most conspicuous of the short tracts medullated early. It is a small compact bundle (no. 41), seen only in fig. 1, extending from the region of the ganglion of the habenula of the thalamus into the red nucleus. It could not be traced to a ganglion interpedunculare.


130 FLORENCE R. SARIN

The corpus restiforme, or inferior cerebellar peduncle, is medullated and is shown as no. 42 in fig. 7. It was not included in the model, nor were the formatio reticularis fibres, nor the cianial nerves, since they were all shown in the previous model and were not necessarj^ here for orientation. The brachium conjunctivum, or superior cerebellar peduncle is, however, included in fig. 1, (no. 45), because of its relations to a possible descending tract from the red nucleus (no. 46). There is a small bundle of fibres (no. 46), emerging from the caudal end of the red nucleus and extending into the pons just dorsal to the lemniscus medialis, near the median line. It cannot be followed far into the pons. Its relations are best seen by comparing fig. 1 and fig. 5. The sagittal sections are not adequate foi determining whether this is a crossed path or not and the fibres cannot be traced below the upper fourth of the pons.

In connection with the optic nerve (no. 47), there is a tiny bundle of medullated fibres shown at the root of the nerve both in fig. 5 and in fig. 7. Its connections could not be traced — I thought of a Gudden's commissure, but it neither extended to the mid-line nor to the geniculate body. There is, however, a portion of the optic nerve medullated, as it enters the lateral geniculate body. This bundle is shown in the figure already referred to in Barker's Nervous System.

The last of these rather indefinite medullated bundles, is the very small portion of the most medial part of the crus in the hypothalamic region containing a few medullated fibres. These fibres show in fig. 4, just lateral to Forel's fascisulus (no. 19), extending on the one hand toward the substantia nigra and on the other to the little bundle of medullated fibres (no. 48), that separates the two parts of the globus pallidus near the median line. These fibres are most unsatisfactory for study in these sections. All that can be said is that they lie in the crus between three nuclei, the globus pallidus on the one hand and the hypothalamic nucleus of Luys and the substantia nigra on the other.

It now remains to describe the relations the various nuclei shown in the model. The nuclei of the mid brain require no special mention except that the form of the substantia nigra comes out


MODEL OF MEDULLATED TRACTS IN BABY's BRAIN 131

better in these sections than in those of the other model where it was incomplete. The sagittal sections show its length, that it is a thin mass curved to fit into the hollow formed by the crus, and that it extends to the nucleus hypothalamicus of Luys which is a small lens-shaped body lying between the zona inserta and the internal capsule. The cerebral surface of the substantia nigra is curved, indented as it were by the hypothalamic nucleus of Luys. Along the lateral border of the substantia nigra is a tiny medullated bundle, shown only in fig. 9, which appears to connect the substantia nigra with Luys' body.

The nuclei of the thalamus that can be modeled out, are essentially those described by Burdach^^ and more recently by Sachs, and Roussy.^^ In fig. 1, are seen especially the anterior and medial nuclei. Neither of them show any medullated fibres whatever. The medial nucleus (no. 55), covers the floor of the third ventricle; it can not be separated from the pulvinar, nor in sagittal sections at birth is it possible to separate it from the dorso-lateral nucleus along the dorsal border of the thalamus, as is plain in fig. 1. In the sections (figs. 5 and 7), it will be noted that there is a pale band, which represents non-medullated fibres extending forward from the center median of Luys to the anterior nucleus (56) of the thalamus and separating the ventro-lateral nucleus from the dorso-lateral nucleus. In the zone of the median nucleus this non-medullated band entirely disappears. The thickness of the medial nucleus from the middle line toward the side, is only a very small proportion of the width of the thalamus, as can readily be seen in fig. 1, since the caudal part of the nucleus has been cut away to expose the center median of Luys and the ventro-lateral nucleus. In the zone just cerebralward to the center median of Luj^s, the medial nucleus is thicker than it is along the ventral border in fig. 1, and hence it is seen in sagittal sections at the level of Dejerine's fig. 269, vol. 1.

The ventro-lateral nucleus of the thalamus is the largest mass of cells. It is shown best in fig. 3, from which the medial nucleus

" Burdach, Bau und Leben des Gehirus. Leipzig. 1822. Dykschen Buchhandlung. B(l. 2. 1' (Roussy, La Couche Optique, Paris. G. Steinbeil, Edituer, 1907.


132 FLORENCE R. SABIN

has been removed and some of the dorsal-lateral nucleus and pulvinar have been cut away. The extent of the ventro-lateral nucleus is best estimated in fig. 4, from which it has been entirelyremoved, the fibre mass, consisting of the knee of the internal capsule and external medullated lamina on the ventro-lateral aspect, and the posterior limb of the internal capsule on the lateral surface, making a shell in which the nucleus rests. The ventrolateral nucleus is separated from a dorsal nucleus, which is the dorso-lateral nucleus and pulvinar together, by a band of nonmedullated fibres which is plain in the newborn (see no 57, figs. 5 and 7), but not at all sharp in the adult (see fig. 8), on account of the great number of medullated fibres. In fig. 8, the ventrolateral nucleus and the combined dorso-lateral nucleus and pulvinar are imperfectly separated, by the fact that there are more medullated fibres in the ventro-lateral nucleus than in the dorsal mass. As has been described, the ventro-lateral nucleus has two sets of medullated fibres at birth, sfirst the bundle from the medial lemniscus (no. 18), which enters the caudal part of the nucleus, near the median plane just ventral to the center median of Luys. A part of this bundle spreads out in the external medullated lamina. Secondly, the medullated cortical bundle (no. 26) emerges from the extreme lateral surface of the nucleus and helps make up the cortical radiation.

The dorso-lateral nucleus of the thalamus (no. 58), cannot in any way be separated from the pulvinar (no. 51). These two nuclei make a common mass of cells as can be well seen in fig. 5, the term pulvinar being simply used for the caudal part of the mass. Far to the side, as seen in fig. 3, the dorsal mass of the thalamus is less extensive, inasmuch as it lies wholly dorsal to the center median of Luys, than it is near the median plane, as can be seen in fig. 5, in which it lies not only dorsal to the center median of Luys but anterior to it as well. In fig. 3 it will be readily understood that the part labelled 58, and indicated by striation, is the cut surface of the dorso-lateral nucleus, the part which fits over the oblique surface of the ventro-lateral nucleus having been cut away. Thus the ventro-lateral nucleus is bounded on its median surface partly by the median nucleus and partly by the dorso


MODEL OF MEDULLATED TRACTS IN BABY's BRAIN 133

lateral nucleus (fig. 3). The dorso-lateral nucleus is entirely non-medullated at birth.

The pulvinar (no. 51), as has just been said, is the projecting caudal part of the dorso-lateral mass. Its caudal boundary is the posterior commissure (no. 40, fig. l);ventrally it is readily marked off from the center median of Lays (fig. 3) and laterally it extends to the taenia semicircularis (fig. 1), but it has no cerebral border, since it is continuous with the dorso-lateral nucleus (figs. 3 and 5).

On the lateral surface of the thalamus are the two geniculate bodies, the medial and lateral. They are both oval masses of cells whose form is well seen from the surface (nos. 49 and 50, fig. 9). The medial geniculate body has no medullated fibres in this series but the lateral geniculate body has a band from the optic nerve which in part spreads on to the surface, but which for the most part curves around the inner border, just caudal to the medullated thalamo-cortical bundle. This shows in the figure quoted above from Barker. A part of these fibres seem to enter the geniculate body, a part the pulvinar.

The center median of Luys is well shown in figs. 1, 3, 4. It is an oval mass of cells which lies in caudal part of the thalamus near the median line. It is just dorsal in position to the red nucleus, lateral to the fasciculus retroflexus of Meynert and lies between the medial nucleus, the ventro-lateral nucleus and the pulvinar. It fits into the cup-shaped nucleus as can be best seen in fig. 11, and in common with it receives a mass of medullated fibres consisting of a part of the medial lemniscus, possibly of fibres from the red nucleus, and of a small tract from the nucleus of the inferior colliculus. It has no other medullated tracts at birth.

This corpus striatum is of course made of the well known parts, the caudate and the lenticular nucleus. The form of the caudate, with its swollen head and its curved tail is well known and can be readily seen in the models. The best way to obtain an idea of the caudate nucleus is to take an adult brain well hardened in formalin and shell the nucleus out from the bed of fibres, the internal capsule, on which it rests. The form, and the relations of the lenticular nucleus are much harder to make out. Starting from the lateral view of the baby's brain (fig. 10), the


134 FLORENCE R. SABIN

lenticular nucleus underlies the island of Reil and hence it will be seen that it lies for the most part in front of the thalamus, which is its original position. As seen from the side, the lenticular nucleus or its lateral portion, the putamen, is a great oval mass of cells (fig. 9) ; the antero-posterior diameter is slightly greater than the dorso-ventral, speaking in terms of the cerebrum, and close to the optic nerve is a small tongue of the nucleus which extends into the temporal lobe. In the curve of this tongue is the non-medullated anterior commissure.

Since the lenticular nucleus is best known as it appears in horizontal sections of the gross brain, it would be most readily understood by a dorsal view of the model, that is, a view taken from the dorsal surface of the cerebrum. The form of the nucleus in the new-born corresponds with its form in the adult as can be seen by comparing the description with horizontal sections of the adult. As seen from the convex surface of the cerebrum, the lenticular nucleus is a triangular mass, divided into three sections by bands of fibres for the most part non-medullated. The outer division, the putamen, is the largest, it projects farthest toward the dorsal surface of the brain and also extends farthest ventralward. It is crescent in shape with the anterior end much larger than the posterior. Next comes the outer part of the globus pallidus, which is likewise crescent in shape with a swollen anterior end. The inner part of the globus pallidus is rectangular in shape. The medullated fibres are all associated with this inner division of the globus paUidus; starting with the median plane, there are a few medullated fibres between the anterior end of the two parts of the globus pallidus (no. 48) ; secondly, there is a mass of fibres within the medial part of the globus pallidus at its posterior or caudal end (no. 30) , the fibres lying adjacent to the crus as seen in fig. 4, and thirdly, there is the lateral capsule for the inner division of the globus pallidus which is the place of origin of the lenticular cortical radiation previously described (no. 27). If the lenticular nucleus be now viewed again from the mesial aspect it will be seen that the bands of fibres, for the most part non-medullated, which separate the three divisions, all connect (see 48, fig. 4), with the anterior limb of the internal capsule, so that if these bands of fibres be


MODEL OF MEDULLATED TRACTS IN BABY 's BRAIN 135

considered, there are two curvedsheetsof fibres which spread out from the lateral surface of the anterior limbs of the internal capsule and these two shells of fibres separate the parts of the internal capsule. Thus, from the standpoint of the corpus striatium, there are three sheets of fibres, one separating the caudate nucleus from the putamen, the other two separating the three parts of the lenticular nucleus, and these three sheets of fibres meet in the sheet of fibres which makes up the knee of the internal capsule. There is no doubt that the form of the internal capsule can be made much plainer by models of the medullated bundles in series of later stages.


EXPLANATION OF FIGURES


KEY TO REFERENCE NUMERALS


1 Fasciculus longitudinalis medialis

2 Fasciculus cuneatus

3 Decussatio lemniscorum

4 Point of crossing of the decussatio

lemniscorum

5 Fibrse arcuata) internse

6 Lemniscus medialis

7 Nucleus olivaris superior

8 Lateral lemniscus

9 Tract of the medial lemniscus to

the substantia nigra

10 Lemniscus superior

11 Tract of the medial lemniscus

to the nucleus hypothalamicus (Corpus Luysi)

12 Lemniscus medialis at the point

of Forel's BaTh

13 Nucleus ruber

14 Center median of Luys

15 Cup-shaped nucleus

16 Tract connecting the nucleus

colliculi inferioris with the center median of Luys and forming a part of Forel's BaTh

17 Combined mass of fibres from the

lemniscus medialis and from the red nucleus (rubro-lenticular tract) making Forel's Feld II

18 Lemniscus medialis to the ventro

lateral nucleus of the thalamus, Forel's Feld H,

19 Rubro-lenticular tract, Forel's

Feld Hi

20 Rubro-lenticular tract, Forel's

Feld Hi within the first or medial part of the globus pallidus of the lenticular nucleus

21 Ventro-lateral nucleus of the

thalamus


22 Nucleus hypothalamicus (Corpus

Luysi)

23 Medullated bundle following the

cerebral surface of the nucleus hypothalamicus

24 Pedunculus cerebri or crus

25 Brachium quadrigeminum infe rius

26 Thalamo-cortical radiation

27 Lenticular-cortical radiation

28 Cortical radiation made up of

the two elements nos. 26 and 27

29 Small medullated bundle in the

crus with transverse fibres connecting Luys' body with the globus pallidus

30 Medullated fibres within the glo bus pallidus connected with Luys' body

31 Anterior limb of the internal

capsule

32 Nucleus caudatus

33 Inner division of the globus pal lidus of the lenticular nucleus

34 Putamen of the lenticular nucleus

35 Anterior commissure, which in

fig. 4 marks the line between the putamen and the globus pallidus

36 Knee of the internal capsule

37 Band at the base of the central

Sulcus

38 Sulcus centralis (Rolandi)

39 Position of the nucleus n. Ill

40 Posterior commissure

41 Fasciculus retro-flexus of Mey nert

42 Corpus restiforme

43 N. V

44 N. VI


key to keference numekals (continued)


45 Brachium conjunctivum

46 Tract extending between the red

nucleus and the pons

47 Optic nerve

48 MeduUated lamina between the

two divisions of the globus pallidus

49 Corpus geniculate mediale

50 Corpus geniculate laterale

51 Pulvinar

52 Subtantia nigra

53 Approximately the direction of

Forel's fig. 5, plate VII

54 Stria meduUaris thalami

55 Nucleus medialis thalami


56 Nucleus anterior thalami

57 Non-meduUated lamina between

the ventro-lateral and dorsolateral nuclei of the thalamus

58 Nucleus dorso-lateralis thalami.

In fig. 3 this nucleus is striated to indicate that it is a cut surface 5i) Line of the cortex of the anterior central convolution

60 Nucleus coUiculi inferioris

61 Raphe

62 Fornix

63 Mammillnry l)0(ly

64 N. Ill


Pl.AI'K I


TKACTS OF FIBUES MEDULl.ATED JN BABY'S BRAIN

il.ORKNCK H. SABIN



THK AMUIUCAN JOUHNAI, OP ANATOMY, VOL. U, NO. 2.


TRACTS OF FIBRES MKDULLATED IN BABY'S BRAIN

I'-LORENCE R. SABIN


PLATE 2



Fig. 1 Mesial view of the entire model showing the tracts that are meduUated at birth as far as they can be seen from the mesial plane. 2J^X. This view shows especially the fasciculus longitudinalis medialis, the lemniscus medialis in the brain stem, and the anterior and medial nuclei of the thalamus. For the meaning of the numbers see key to numerals. An outline of the cortex of the anterior and posterior central convolutions is given for orientation. Tlie bundles of medullated fibres have a blue tone.

Fig. 2 Mesial view of a baby's brain given to orient figs. 1, 3, and 4. The brain is older than the one from which the models were made.


THE AMEKICAN JOURNAT- OF ANATOMY, VOL. 11, NO. 2.


PLATE 3


TRACTS OF FIBRES MEDULLATED IN BABY'S BRAIN

FLORENCE K. SABIN



Fig. 3 View of tiie mode from the mesial aspect from which the fasciculus longitudinalis medialis, and the anterior, medial and part of tlic dorsolateral nuclei of the thalamus have been removed in order to expose the lemniscus medialis within the thalamus. 23-2X.


THE AMERICAN JOURNAL OP ANATOMY, VOL. 11, NO. 2.


I'l; \(rs OF KIHIM';^ .MEDULLATED in 15ABY'S rkain

FLORENCE R. SABIN


PLATE 4



I'lfi. 4 \ icw of (lu> ni()(l( Iroiii which the nuclcu.s caiKhit us Mild the uucloi of I lie t liahinius except the center uiediaii of Luys (no. 14) have been removed.


THE AMEUICAN JOUU.NAL OF ANATOMY, VOL. 11, NO. 2.


PLATE 5


TRACTS OF FIBRES MEDULLATED IN BABY'S BRAIN

FLOUBNCli n. SADIN


32 - —


■f^*^ ~^^


— —21



Fig. 5 Sagittal section of the brain stem and basal ganglia of the new-born baby's brain from which the model was made taken near the mesial plane and showing especially the lemniscus medialis no. 6, and the fasciculus rubro-lonticularisof Forel (no. 19). 2X.


THE AMEKICAN JOUKNAl. OF ANATOMY, VOL, 11, NO. 2


TRACTS OF FIBRES MEDULLATED IN BABY'S BRAIN

FLORENCE R. SABIN


PLATE 6



■oG


04


-18 -41

-13 -.32

GO


Fig. 6 Sagittal section of the pons, mid-brain and thalamus of an adult brain in about the samo plane as fig. 5, to show especially the relations of the red nucleus to the lemniscus medialis and to the rubro-lenticular tract. About IMX.


THE AMERICAN JOURNAL OF ANATOMY, VOL. II, NO. 2.


PLATE


TRACTS OF FIBRES MEDUi.I- A ri-;D IN liAHY'S BRAIX

Fl.OHENCK U. SAHIN


32_


.5fi


36.


KM


  • ^- —






"



.33.

9'1





„_


_


\



2U


30


20





■^■h \






r>


^







47


11-

, _ .


JJ7


.21


— ^ ■> <


tj


— -2 1


Fif?. 7 S;i};ill;il sect ion of lli(> l)r;un .sLcni and basal Siuigli^i <>i tln' uew-bnni baby's bruin froiii which (ho models weremade, taken farther lateral than fig. 5, lo show the relations of the lemniscus medialis (nos. (>, 12 and IS), to the thalannis. 2X.


TIIK AMIOUIC VV lolIUNM, ol' Wvl'UMV, Vol.. II, \(). 2


THACT« OF FIBRE.S MEDULLATED Ix\ BABY'S BRAIN

FLORENCE R. SABIX


PLATE 8


32-


% '■'i^ -.


%



B5j-?"-4,


"■Vfe.


.5.5 48


20



8


Fig. 8 Sagittal section of the brain stem and basal ganglia, of an adult brain at a level slightly farther to the side than fig. 7. About 13^ X.


THK VMERICAN JOURNAL OK ANATOMY, \ or.. 11, no.


PLATE y


TRACTS OF FIBRES MEDULLATED IN BABY'S BRAIN


FI.OnBNCE U. SA.UIN



THE AMERICAN JOURNAL DF ANATOMT, VOL. 11, NO. 2,


TRACTS OF FIBRES MEDULLATED IN BABY'S BRAIN

FLORENCE R. SABIN


PLATE 10



10


Fig. 9 Lateral view of the entire model showing especially the position of the lemniscus medialis (no. 6) with reference to the surface form of the mid-brain, and the putamen and the cortical radiation meduUated at birth. 2J^X.

Fig. 10 Lateral view of the baby's brain shown in fig. 2, showing a dissection of the pyramidal tract. This is not the cortical radiation meduUated at birth (no. 28), but is, in general, parallel to it. The position of the pyramidal tract is indicated on fig. 9 by the arrow 36.


THE AMERICAN JOURNAt OF ANATOMY, VOL,. 11, NO. 2.


PLATE II


TRACTS OF FIBRES MEDULLATEI) IN HAHYS BHAl.X

FLORENCE R. SAB IN



V


\


4


\-31 \

/

7


21


47


Fig. 11 Lateral viow of tii( model from which tlie piitamen and outer division of the globus pallidus have been removed to show the capsule of medullated fibres of the inner part of the globus pallidus (no. 27), which make the bulk of the cortical radiation at birth. 2HX.


11


THE AMERICAN JOURNAL OF ANATOMY, VOL. 11, NO. 2.


THE BLOOD SUPPLY OF THE PITUITARY BODY

WALTER E. DANDY and EMIL GOETSCH

From the Hunterian Laboratory of Experimental Medicine, The Johns

Hopkins University

FOUR FIGURES

INTRODUCTION

Studies made in the Hunterian Laboratory^ by Reford, Crowe, Homans, Goetsch and Gushing, based upon a series of over two hundred canine and feline hypophysectomies, have supported Paulesco's view that a fragment of the pars anterior is essential to the maintenance of life, a total hypophysectomy being invariably fatal in the course of a few days, the length of survival depending somewhat upon the age of the animal. Durmg the course of these investigations it was found that the gland could be separated from its dural pocket and left dangling by the stalk alone without destroying its vitality. Division or ligation of the stalk, on the other hand, often led to profound histological alterationsi evidently due to an ansemic necrosis confined largely to the

1 (1) Is the pituary gland essential to the maintenance of life? The Johns

Hopkins Hasp. Bull., 1909, 20, no. 217. .. ., • , f

(2) The hypophysis cerebri: clinical aspects of hyperpituitarism and ot

hypopituitarism. J. Am. Med. Ass., 1909, 53, p. 249. , , _„

' (3) Effects of hypophyseal transplantation following total hypophysectomy

in the canine. Quart. J. Exper. Physiol., 1909, 2, p. 389

(4) The functions of the pituitary body. Am. J. Med. bci.l9W.

(5) Experimental hypophysectomy. The Johns Hopkins Hosp. Bull, 1910,

^^' ^e/^ Concerning the secretion of the infundibular lobe of the pituitary body and its presence in the cerebrospinal fluid. Am. J. of Physiol, 1910, 2,, p. 60.


The American Journal of Anatomy, Vol. 11, No.


138 WALTER E. DANDY and EMIL GOETSCH

centre of the anterior lobe. It had been suggested by Paulesco that stalk separation is equivalent to a total hypophysectomy, presumably as a result of such necrosis; and the experiences in this laboratory tended at first to confirm this view. Later observations, however, have shown that although a stalk separation will often lead to a certain degree of central necrosis of the pars anterior, total isolation of the gland, not only from the infundibulum but from its dural attachments as well, is necessary before the procedure becomes equivalent to a total removal; and even in this case fragments of the structure, thus isolated, may under favorable circumstances become revascularized in part, the operation therefore being comparable to a total homo-transplantation of the gland into the tissues elsewhere.

More complete information in regard to the sources of the glandular blood supply has become essential to a better understanding of these operative experiences, and at the suggestion of Dr. Gushing these studies have been made with this end in view.

The present paper deals with the mammalian circulation as observed in the dog — the animal employed for most of the experimental studies. It is presumable that the findings apply as well to man, but we have had no suitable opportunity for a proper injection of the human gland.

PRIOR DESCRIPTION OF THE CIRCULATION

The first mention in the literature of the pituitary circulation is an incidental reference by Duret ('72) in his classical work upon the blood supply of the brain in general. He refers to a small bilateral branch which passes from the posterior communicating artery to the infundibular wall. Heubner, two years later, made a similar reference.

From the time of these casual comments, although it was generally appreciated that the anterior lobe was a very vascular organ and contained peculiar sinusoidal spaces, the subject has been given little attention, until Herring's recent excellent and concise description of the internal circulation of the cat's hypophy


BLOOD SUPPLY OF THE PITUITARY BODY 139

sis.- Herring was the first to show that the anterior and posterior lobes possess independent blood suppHes, the former coming down through the stalk, the latter entering the posterior lobe from behind. No comment was made, however, on the origin of these vessels. Substantially the same results were obtained a few years ago by Dr. G. J. Heuer in unpublished studies carried on independently in this laboratory.

Our observations, as will be seen, entirely confirm the views of Herring upon the general plan of the gland's internal circulation, but especial emphasis is placed upon the grosser circulation, in view of its experimental and surgical importance.

MATERIAL AND METHODS

Immediately after death the animals were injected head ward through both common carotid arteries, during which process the principal veins of the neck were ligated or a tourniquet applied to obstruct the venous return of the injection mass. Of the numerous colored masses, we have derived the best results from a 10 per cent gelatin mass with carmine or vermillion for the veins and capillaries, and Prussian blue or ultramarine for the arteries. Satisfactory injections may be obtained by following a primary carmine injection by one of Prussian blue, a double injection being obtained — the veins red and the arteries blue — since the larger granules of the blue mass are unable to pass through the capillaries.

Carmine gelatin, although the best general injection mass, is a very capricious substance, requiring careful preparation, for if too alkaline it diffuses through the vessel wall, or if too acid it precipitates. We are indebted to Dr. M. J. Burrows of the Rockefeller Institute for the benefit of his experience in the rather elaborate preparation of this mass.

2 Herring, P. T. The histological appearances of the mammalian pituitary body. Quarterly J. Exper. Physiol., 1908, i. p. 154. In this article Herring gives an excellent photograph (fig. 16, p. ]54) of the injected feline hypophysis, which shows unusually well the relatively greater vascularity of the anterior lobe. This condition is rarely and with difficulty brought out by a simple arterial injection, which usually shows little more than in our fig. 4.


140 WALTER E. DANDY and EMIL GOETSCH

To secure an injection of the venous supplj^ alone, Prussian blue was injected head wards into the internal jugular vein, there being no valves in the canine jugular to prevent the reversed stream from reaching the intracranial veins and venous sinuses. After the gelatin had been hardened by coohng, a block of tissue containing the hypophysis and its meningeal envelopes intact . was carefully removed, preserved in gh'cerine or creosote, and either studied by the aid of the binocular after clearing, or fixed and sectioned for microscopical study.

GENERAL CONSIDERATION OF THE CIRCULATION

The blood supply of the hypophysis is so abundant as to seem, were it actually an unimportant gland, more than commensurate with its functional needs and the possibility of harm from vascular disturbances.

Situated in the centre of the circle of Willis, it receives the first installments of blood to the brain, its numerous vessels converging from all parts of the circle like spokes to the hub of a wheel, while the venous outflow is equally abundant and the vessels similarly disposed.

There are three fairly independent circulations to the subdivisions of the gland in correspondence with their structural independence: (1) The supply to the anterior and intermediate lobes; (2) to the posterior lobe; and (3) to a structure which has been discovered during the course of these investigations and which we shall designate the "parahypophysis."

Vessels of the anterior lobe

The anterior lobe receives its blood supply from numerous small arteries which converge toward the stalk of the hypophysis (fig. 2.) The great majority of these vessels arise from the anterior half of the circle of Willis. Thus the anterior communicating artery, which is usually small (cf. figure) in the dog, sends off from eight to ten small branches in its course across the optic chiasm; and each posterior communicating artery sends off from three to five more. Furthermore, a pair of vessels on each side


BLOOD SUPPLY OF THE PITUITAEY BODY


141



^^'^'^^V:£^.AvOV^^^' ¥'



i'i<^. 1 Hypophyseal region viewed from below: dura largely intact except for removal of superficial layer of that part of the membrane which covers the gland. The arterial supply to the posterior lobe shows through the dura, each internal cartoid artery (running in the lateral sinus) giving off a branch the two uniting to form a single trunk (A, p. 1.) The small intradural glandule ("parahypophysis") which normally lies concealed between the two dural layers is exposed and its arterv and vein are shown.


142


WALTER E. DANDY and EMIL GOETSCH



ff^puwT^^"'



Fi{?. 2 Same region us in fig. 1, aftor removal of hypoijhysis from its cerebral attachment, showing infundibular stalk and median vent ricular slit . The arterial supply to the anterior lobe is fully exposed. The vessels radiate from the circle of Willis toward the gland and from the points shown pass down the stalk to the pars anterior.


BLOOD SUPPLY OF THE PITUITARY BODY


143




Fig. 3 Same region as in figs, land 2, after removal of dura and excision of the arteries shown in fig. 2 whicli conceal the venous supply to the anterior lobe. A crescentic segment of anterior lobe has been cut away in order to expose the posterior lobe and bring into view the median vein M.v., which traverses the pars posterior and enters the circular sinus. Great difference in vascuhxrity of two lobes can be seen in gross. Numerous small veins radiate outward from the stalk of the hypophysis to a circle of veins very similar to the arterial circle of Willis. Relation to arteries shown by peripheral stubs of the excised trunks fully shown in fig. 2.


144


AA'ALTEK E. DANDY and EMIL OOETSCH



Fig. 4 Mid sagittal section of canine hypophysis to show posterior lobe artery entering at poiiil of dural attafdimcnt and partially collapsed vein directly above the artery.


BLOOD SUPPLY OF THE PITUITARY BODY 145

is given off b}^ each internal carotid artery, immediately before its trifurcation into the anterior, the middle cerebral and the posterior communicating arteries. Thus, all told, from eighteen to twenty or even more separate small vessels, barely visible to the unaided eye, converge directly toward the infundibular stalk. In addition other vessels from the posterior half of the circle of Willis stream in a fine network over the corpora mamillaria and converge toward the posterior surface of the infundibulum. Upon reaching the stalk these vessels immediately subdivide into capillaries which empty into the dilated channels of the anterior lobe. These channels are lined merely by a single layer of endothelial cells which lie directly in contact with the anterior lobe cells. The sinusoidal spaces are so numerous and so large as to make the anterior lobe one of the most vascular structures of the body, comprising as they do a large part of the volume of the gland. Inasmuch as the anterior lobe contains no arteries or veins, arterial injections, owing to the large size of the granules, tend to stop at the capillaries of the stalk of the h3'pophysis.

The injected hypophysis, view^ed in the gross by the aid of the binocular, shows numerous parallel longitudinal channels uniformly distributed throughout the gland and varying slightly in size. These are evidently the main channels, from which smaller ones are redistributed, although histologically all have the same single endothelial lining. There is, however, no very great difference in their size, although in specimens injected under a relatively high pressure it is often possible to see very large spaces, out of all proportion to the surrounding channels, owing to the irregular disposition of the injection mass.

The venous return from the anterior lobe is very similar in arrangement to the arterial inflow. These channels reform into capillaries and, verj' soon, into numerous small veins in the stalk, from which the collecting vessels radiate to the basilar circle of veins which overlies but has an arrangement very similar to the arterial circle of Willis. There are six or seven of these relatively large collecting trunks (fig. 3), which pass laterally and singly into the basilar veins lying deep under the temporal lobe, and thence upward around the crura cerebri to discharge into the


146 WALTER E. DANDY and EMIL GOETSCH

venae magna? Galeni. There is also a network of venules which emerge from the posterior part of the stalk and pass over the Corpora mamillaria, ending in several veins that empt}- into a transverse connection between the basilar veins which forms the posterior arc of the Willisian venous circle (cf. fig. 3).

Vessels of the pars intermedia

Certain portions of the pars intermedia are also very vascular — much more vascular than the posterior lobe, though less so than the anterior. It derives its circulation from three sources. The thicker, tongue-like portion near the stalk, which is more or less merged with the upper part of the anterior lobe, receives its blood from the vessels of the stalk and from others which cross from the brain substance immediately adjacent. The thin epithelial investment of the pars nervosa, on the other hand, is entirely devoid of vessels, as Herring states. The abundant capillary network, which intervenes between these two histologically different structures constituting the posterior lobe, does not penetrate into the layers of the investing cells. The capillaries of the vascularized tongue of pars intermedia, however, are in immediate contact with the base of the cells and follow their villous formations. From these capillary beds of pars nervosa there are two possible ways of escape for the blood — upward into the veins of the stalk and base of the brain, and backward into the veins of the pars nervosa and thence into the circular sinus by the vascular system which remains to be described.

Vessels of the posterior lobe

The posterior lobe is the least vascular of the anatomical subdivisions of the hypophysis. This is very easily recognized in the fresh uninjected specimen, the anterior lobe appearing a pinkishyellow color in marked contrast to the glistening and whiter posterior lobe. Injected specimens naturally bring out this contrast much more strikingly.


BLOOD SUPPLY OF THE PITUITARY BODY 147

The circulation of the posterior lobe enters the gland at the posterior pole of the pars nervosa and is entirely independent of the systems heretofore described. Each internal carotid, shortly after it enters the cranial chamber and turns forward (fig. 1) in the carotid groove, gives off a small branch : these two vessels unite in front of the posterior clinoid process to form the single median trunk which enters the hypophyseal posterior lobe at the point w^here a small area of firm dural attachment is appreciable in the usual dissection to liberate the gland. It is well to bear in mind that these branches of the carotid lie between the two laj'ers of dura forming the circular 'sinus, and therefore really in a sense lie in this sinus, just as the internal carotid itself lies in the lateral sinus.

The artery enters the posterior lobe near its centre (fig. 4) dividing immediately into numerous large branches which stream outward toward the periphery. There is, however, no marked difference in the inherent vascularity of any part of the posterior lobe until the capillary bed is reached at the extreme periphery under the epithelial investment.

The veins collect the blood and pass to the point of dural attachment in much the same arrangement as the arteries. A single large central vein enters the circular sinus at this point (cf. fig. 3, m. v.) immediately above the entrance of the artery, the two being closely adjacent. In addition, there are other smaller veins which empty independently into the sinus.

Collateral circulation

Although the anterior and posterior lobes have independent blood supplies, the question arises as to the possible vascular communication between structures so intimately related. Is the collateral circulation sufficient to preserve the glandular function in case of occlusion of the blood supply to one or another of the anatomical subdivisions of the gland? Is separation of the vessels of the stalk equivalent to a transplantation of the anterior lobe? There is doubtless some collateral circulation between the anterior and posterior lobes through the pars inter


148 WALTER E. DANDY and EMIL GOETSCH

media as mentioned above — namely, at the relatively narrow area of pars intermedia above the cleft, which elsewhere prevents contiguity of pars anterior and pars posterior. Histologically we should judge that this would be sufficient, especially in view of the small fragment of anterior lobe which apparently will suffice for the preservation of life.

Naturally the final test of the practical efficiency of this collateral must be determined by experimental operative methods. Paulesco records cases of stalk separation which caused consequent degeneration of the anterior lobe cells. One must, however, be certain that this small collateral area of pars intermedia has not also been destroyed by the operative manipulations, thereby preventing the possibility of preservation by collateral.

In a few operations on dogs by one of us (Goetsch), in which the blood supply through the stalk was interrupted by the placement of a silver wire clip" (equivalent to the hgation of the stalk), no evidences of physiological deficiency or histological degeneration of anterior lobe cells were observed

In one of Dr. Cushing's hypophyseal operations for acromegaly by a transphenoidal route it was intended merely to remove the floor of the sefia turcica and to freely incise the dural pocket enclosing the greatly enlarged gland, in the hope of thus reheving the neighborhood symptoms. A fragment of the exposed anterior lobe was removed for examination, and during the necessary manipulations the gland was broken from its stalk and there was a gush of cerebrospinal fluid, much as in an experimental canine hypophysectomy when the gland is detached from its infundibular connections. The patient died in forty-eight hours with symptoms comparable to those seen in animals after a total hypophysectomy ; and post-mortem histological studies showed an anemic necrosis involving practically the entire pars anterior, whereas the pars nervosa and its epithehal covering (pars intermedia) remained normal in appearance, its individual and isolated blood supply having been remote from possible operative injurj^


BLOOD SUPPLY OF THE PITUITARY BODY 149

The circulation of the ^^ parahypophysis'^

As has been stated in discussing methods, after the injections were made a block of tissue containing the gland within its intact meningeal envelopes was removed and cleared in glycerine. On examining the base of the cleared specimen with the binocular dissecting microscope, a minute button-like structure (figs. 1 and 3) was seen lying below the mid-point of the gland and between the two layers of the dura which originally lined the base of the sella turcica. This structure had been previously observed in a number of the serial longitudinal sections which had been made as a routine after all of the experimental pophyhysectomies. No especial importance had been attached to it and it naturally had escaped histological observation in all of the cases in which the dura had not been removed and sectioned together with the gland. It is an epithelial body and appears to be an organ which is invariably present and one which may have some physiological significance. It contains under normal conditions none of the typical anterior lobe (eosinophilic) cells. A minute median pit in which the body rests is usually discernable in the centre of the sella turcica after the removal of its lining dura.

This epithelial body has a separate circulation distinct from the others which have been described (fig. 3). The arterial supply seems to be of two-fold origin. A minute artery enters the gland posteriorly, and by reconstruction of two specimens its origin can be traced by a relatively long intradural course to each posterior lobe artery, the two uniting into a single trunk before reaching the parahypophysis. Moreover the small intradural branch from each internal carotid artery gives an additional bilateral arterial supply. A single small vein cares for the return flow. This vein apparently passes into the bone at the situation of the above mentioned pit, though it is probable that there may also be a connection in the dura with the network of venous channels, which are somewhat radially arranged around the parahj^pophysis. This structure therefore should retain its circulation intact in procedures similar to the experimental hypophysectomies in which

THE AMERICAN JOURNAL OF ANATOMY, VOL. 11, NO. 2


150 AYALTER E. DANDY and EMIL GOETSCH

the gland is removed by an operation conducted from under the temporal lobe. It would be the first to suffer in all operations such as those which have been employed in man in which the gland is approached* from below through the sphenoidal cells.

SUMMARY

The anterior lobe receives its blood supply from about eighteen or twenty small arteries which converge toward the stalk from the various components of the circle of Willis. These vessels immediately break up into numerous large sinusoidal channels, in apposition with the cells and lined only by endothelium. Hence, there are no veins or arteries proper in the anterior lobe. The venous supply is very similar in arrangement to the arterial system; the veins passing from the stalk to a venous circle immediately overlying the circle of Wilhs and draining into the venae magnse Galeni.

The pars intermedia derives its supply from the vessels of the stalk, from the adjacent brain, and from the posterior lobe. A collateral therefore exists at this point between the anterior and posterior lobes, probably sufficient to preserve the function of at least the adjoining portion of either lobe if its individual supply is cut off.

The posterior lobe receives its arterial supply from a small artery formed by the union of a symmetrical branch from each internal carotid. One large vein and other small ones enter the circular sinus immediately above the artery.

The parahypophysis" has an individual blood supply of twofold origin; a posterior vessel from the union of two branches from the pos-terior lobe arteries and a bilateral branch from the internal carotid arteries.

In concluding, it is a pleasure to express our gratitude to Dr. Harvey Gushing for his suggestions during the course of the work, and to Mr. Max Brodel for his instruction and advice in the preparation of the accompanying drawings.


THE ANATOMY OF THE THYROID GLAND OF ELASMOBRANCHS, WITH REMARKS UPON THE HYPOBRANCHIAL CIRCULATION IN THESE FISHES

JEREMIAH S. FERGUSON Assista7it Professor of Histology, Cornell University Medical College

TWENTY FIGURES

Introduction •. . 151

Review of the literature 163

Material and methods 164

The anatomical relations of the thyroid gland 165

The thj^roid vessels 171

The hypobranchial circulation and the origin of the thyroid vessels 175

Veins and lymphatics of the thyroid region 183

The histology of the Elasmobranch thyroid gland 187

Summary 207

Explanation of figures 209

Bibliography 210

INTRODUCTION

Even a casual reference to the literature of the thyroid gland is sufficient to indicate that the organ has been more carefully studied in most all other classes of animals than in the Elasmobranchs.

The organ may be said to make its first appearance in the Ascidians, Amphioxus, and Cyclostomes as a depressed groove, trough, or series of recesses in the ventral fioor of the pharynx, usually known as the endostyle, or the hypobranchial or hypopharyngeal groove, which, as first shown by W. Miiller ('71) who studied Myxine glutii;iosa and Petromyzon, is to be considered the homologue of the median thyroid of the vertebrates. In the Cyclostomes the structure, relations and development of the primitive thyroid have been more recently studied bj^ Guiard ('96), Cole ('05), Schaffer ('06), and Stockard ('06). The structure of the organ is very simple and only partially resembles the thyroid of


152 JEREMIAH S. FERGUSON

higher vertebrates. Possibly the thick tenacious secretion formed by the endostyle, upon the presence of which the function of the organ very largelj' depends, may well be taken to bear a relation to the colloid material which is so characteristic of the mammalian gland. Inasmuch as the retention of an albuminous secretion within the glandular lumina of the animal body, a condition frequently observed by the pathologists and normally present in other glands as well as the thyroid, e.g., mammary gland, kidney, hypophysis cerebri and parathyroid gland, leads to the accumulation of a colloid material bearing a more or less striking resemblance to the colloid material in the follicles of the thyroid gland, the deduction from the phylogenetic standpoint, that the retention within the follicles of the thyroid of a once free mucous secretion would account for the colloid character of the follicular content, would not seem inappropriate. The character of cells which pour forth the free secretion of the endostyle or h}q3obranchial thyroid of the Ascidians, Amphioxus and Cyclostomes is not so very different from the colloid secreting cells of the thyroid follicles of mammals.

In the Teleostei, Wagner ('53) has studied the form and location of the thyroid gland and directed attention to the similarity of its structure to that of mammals. Simon ('44) and Baber ('81) have given extended descriptions of the thyroid gland in several species of bony fishes; Maurer ('86) described the structure and studied fully the development of the thyroid gland of the carp and trout. In the more recent literature the structure of the thyroid in fishes seems not to have received the attention which it apparently deserves.

Extended descriptions of the thyroid gland of reptiles are found in the articles by Simon ('44) and Baber ('81). De Meuron ('86) studied the organ in Lacerta. Van Bemmelen ('87) described the gland in Hatteria and Lacerta as being transversely placed over the trachea near the heart, and as forming a small, thin unpaired organ. Guiard ('96) has also discussed the structure of the organ in reptiles.

In the Amphibia the work of Maurer ('88) and the excellent descriptions by Wiedersheim ('04) apparently leave little to be


THE ANATOMY OF THE THYROID GLAND 153

desired, though the organ has been much studied in this class of animals.

In Aves the thyroid gland has been extensively studied by Simon ('44), Peremischko ('67) who considered the histology as well as the gross anatomy of the organ, Baber ('81) and De Meuron ('86). The literature of the structure and development of the thyroid in the chick is extensive.

In most of the Mammalia the anatomy of the thyroid gland is well known and its literature has acquired voluminous proportions. Its review does not fall within the scope of the present paper.

REVIEW OF THE LITERATURE

A careful study of the available literature has revealed, with the exception of the work of Guiard, only casual references to the anatomy of the thyroid gland in the Elasmobranchs. Simon ('44) studied the Selachian (Squalus) and the skate (Raia). He describes the thyroid gland as "a single organ, situated in the median line, in connection with the anterior surface of the cartilages which bind together the branchial arches of opposite sides of the body," and he states that it may lie in contact with the lingual bone," or may be more or less distant from the mouth, but that it is always at the spot where the great trunk of the branchial aorta distributes its terminal branches. It lies at the angle of this bifurcation . . . ; it is covered by the sterno-hyoid or sterno-maxillary muscle, and also by the myo-hyoid and geniohyoid, w^hen these are present." His description I find to hold good for Raia, but it does not entirely correspond to the position of the thyroid gland of Squalus, Mustelus, or Carcharias. As to its vascularization, Simon states that the gland receives its blood supply by means of a recurrent branch given off by the first branchial vein, while yet within the gill, and that "it never receives the smallest share of supply from the branchial artery with which it is in contact." The last portion of this statement is precisely correct for all the species which I have examined, though apparently at variance with the observations of some authors : the first portion, as to the origin of the thyroid artery,


154 JEREMIAH S. FERGUSON

would appear to be not very accurately expressed, for it never arises from the afferent branchial vessel which leaves the dorsal end of the branchial arch, but, on the contrary, arises from the ventral end of the nutrient or efferent loop. The relation of the thyroid gland to the bifurcation of the ventral aorta is so intimate as to readil}^ suggest the error of other observers who have presumed that the organ received some blood directly from the first pair of branchial arteries. Moreover, the thyroid artery passing from its origin lies directly under, and in contact with the first pair of afferent branchial arteries so that until these latter vessels have been carefully dissected out of their sheath it is impossible to determine with certainty that they have no connection with the thyroid vessels. In the skate the gland lies directly upon the aortic bifurcation and the pulsating blood-vessels are readily seen through the transparent organ as soon as it is exposed.

Miiller ('71) speaks of the thyroid gland of Raia clavata as a flattened brownish-red body, lying at the point of division of the branchial artery. It possesses a connective tissue capsule with trabecula which divide the organ into a small number of lobes, within which the connective tissue penetrates between the lobules. The follicles possess a thin membrana propria and a cylindrical epithelium; they contain a homogeneous, gelatinous yellowish mass. The epithelium possesses a shiny cuticular border and appears to send processes into the lumen. The description given by Miiller holds good for Raia, the genus which he studied, but it does not correspond to the condition of the thyroid gland of Mustelus, Squalus or Carcharias, the difference being chiefly due to the fact that in the Batoidei the connective tissue forming the thyroid trabecula and interfollicular septa is apparently much more abundant than in the Selachii.

Balfour ('78) discusses very briefly the early development of the thyroid gland of Elasmobranchs prior to the appearance of a lumen within its follicles. He does not consider the anatomy of the organ in the adult.

Baber ('81) says that in the skate the gland is single (with the exception of a few detached vesicles) and forms a j^ellow, flattened, lobulated body, occupying the median line at the bifurcation of


THE ANATOMY OF THE THYROID GLAND 155

the branchial artery. Anteriorly it sometimes presents a narrow process of gland-tissue running forward, and behind it is limited by the bifurcation of the branchial artery." The contents of its vesicles consisted of coarsely granular masses or globules of various sizes which correspond with the 'colloid substance' of authors." He surmises the non-existence of lymphatic vessels and says that in both the skate and the conger-eel "an extensive system of vessels lined with epithelium becomes injected by the method of puncture" ; he considers that these are blood-vessels. The narrow process of glandular tissue which Baber says is occasionally present in Raia is more frequently seen in the Selachii; it is constantly present in all the examples of Carcharias which I have dissected. It extends forward until it comes into contact with the anterior margin of the basi-hyal cartilage (lingual bone) which presents a depression, frequently amounting to a complete foramen, for the reception of the anterior extremity of the glandular process with the connective tissue by which it is heavily invested. This process is obviously analogous to the pyramidal lobe of the mammalian thyroid, and as it extends all the way to the pharyngeal submucosa in many instances it may well be considered as indicative of a phylogenetic connection of the gland with the cavity of the pharynx, a condition which is also indicated by the ontogeny of the organ in all the orders, and which appears to be permanent in the Ascidians, Amphioxus, and the Cyclostomes.

Baber's suspicion of the non-existence of lymphatics within the thyroid gland appeared to the writer to be such a remarkable observation and so out of harmony with the known anatomy and physiology of the organ in the higher animal orders as to require further study. This study was pursued by means of a considerable series of careful dissections with many injection experiments and did not appear to confirm Baber's opinion. Baber's observation that the blood-vessels in these animals could be injected by the method of puncture is quite accurate, but it does not by any means disprove the existence of lymphatics. His results were apparently dependent upon the fact that the smallest veins and capillaries are of very considerable caliber and readily admit of injection, while the lymphatics are very minute and are entered


150 JEREMIAH S. FERGUSON

only with difficulty and not frequently when the needle is thrust into the substance of the gland.

Balfour ('81) referring to the development of the organ in Scyllium and Torpedo says that at first it is solid and attached to the esophagus. "Eventually its connection with the throat becomes lost, and the lobules develop a lumen."

Dohrn ('84) in his plate XI, fig. 5, indicates by outline the thyroid gland of Ammocetes, but does not illustrate or describe the thyroid gland of the Selachii, though in his text he includes an extended description of the thymus of the latter animals. His outline of the thyroid of Ammocetes does not conform to that the the gland in the Selachii.

De Meuron ('86) says that in Scyllium the thj'^roid is elongated, in Galeus and Acanthias, much flattened, in Raia pyramidal or rounded. It lies just above the terminal bifurcation of the branchial artery. In Myelobates it lies behind the os hyoideus, beneath the sterno-mandibularis muscle, and is triangular in shape, short, flattened, transversely elongated, and has a length of 2 cm. In Acanthias the thyroid gland presents an irregular contour, certain groups of follicles being even completely detached, and placed around the principal group. The observations of De Meuron would appear to be accurate as far as they go but are possibly founded upon the examination of too few individuals. Thus in Squalus acanthias I found the thyroid frequently broken as described by De Meuron for Acanthias but other individuals presented a thyroid which was perfect, not the least broken up or irregular in contour, and in the closely related ^Nlustelus canis irregularity of contour is certainly the exception, not the rule. In Scyllium he says the thyroid is elongated and I find that superficial examination of the related species Carcharias, would indicate a similar condition, but if the semi-opaque white mass of connective tissue, in which the thyroid gland of Carcharias is heavily clothed and so closely invested that it seems to form paart of the gland, be dissected out and held, stretched in its normal form, between the bright sun and the eye of the observer there is readily seen within the reddish-white connective tissue mass the outline of the yellowish-orange thyroid gland, which instead of having theelon


THE ANATOMY OF THE THYROID GLAND 157

gated form of the outward mass is flattened, transversely elongated, and of the same peculiar triangular or shield-like shape which is characteristic of the organ in the dogfish and closely simulated by that of Raia. It seems possible that the elongated gland observed by De Meuron in Scyllium might be susceptible to a similar analysis.

Guiard ('96) studied six species of the Selachii and five of the Batiodei. In Scyllium he found the thyroid gland of pyriform shape, the anterior extremity being prolonged forward as far as the anterior margin of the lingual cartilage ("copule"), where it passes between the two lateral halves of the coraco-hyoid muscle. This description, as given by Guiard, corresponds with the position and form of the gland which I find in Carcharias and which, as regards the anterior prolongation, appears to be analogous to the pyramidal process of mammals. But Guiard's fig. 1, in the absence of specific contradiction in his text, might be taken to indicate that the thyroid gland had been found beneath the coraco-hyoid muscle; this is not the case in any of the species which I have examined and I presume it is not the case in Scyllium catulus, from which species his figure was drawn. In each species I have found the gland lying, without exception on the ventral surface of the coraco-hyoideus, between it and the coraco-mandibularis, except that at the anterior portion the gland lies between the coracohyoid muscles of the two sides, the divergence of the two muscles exposing the ventral surface of the cartilage at this point. In the Batoideithe coraco-hyoidei are so widely separated that the whole thyroid gland may come to lie directly upon the basi-hyal cartilage, the aortic bifurcation and the coraco-branchialis muscles, which are successively exposed from before backwards by the separation of the coraco-hyoids, but in this case the fascia which covers the ventral surface of the coraco-hyoids dips beneath the dorsal surface of the thyroid gland.

In Acanthias vulgaris and Mustelus loevis Guiard as did De Meuron, notes the tendency of the thyroid to present detached vesicles, its contour being very irregular. In Galeus canis the thyroid gland lies rather farther forward and is partially covered by a fold of the buccal mucosa. In Carcharias glaucus the


158 JEREMIAH S. FERGUSON

coraco-mandibularis muscle is relatively very broad and completely hides the thj^roid gland ; the gland is described as reniform and voluminous. With the exceptions recorded the position of the thyroid in the various species of Selachii corresponds fairly well with that described for Scyllium.

Of the Batoidei, in Raia alba the coraco-hyoidei are small and widely separated, and between these muscles the first pair of branchial arteries emerge. The thyroid gland is described as lying between the bifurcation of the aorta and the hyoid arch; it is a very large globular organ and its deeper surface is slightly prolonged as far as the arterial bifurcation. In Raia oxyrhynchus the thyroid gland in transversely elongated. In Raia pastinaca the coraco-hyoidei approach one another and the gland is longitudinally elongated ; in this particular it corresponds to the Selachian type. In this last species it is a large pyriform organ with its broad end in relation with the hyoid cartilage, and its point extending nearly to the bifurcation of the branchial artery; at its point the gland presents a prolongation "which descends between the branchial sacs to a depth of about 0.5 cm." I desire to call attention to the fact that in Carcharias a posterior prolongation apparently also exists and is constantly present, but so far as I am able to observe it consists solely of connective tissue and contains no glandular substance; it can not, therefore, be in any way analogous to the anterior prolongation of the processus pyramidalis. I think it is to be connected with the fascia of the thyroid sinus, which will be discussed later on, rather than with the gland itself.

Guiard sums up his work on the morphology of the thyroid gland in the rays by saying that the organ lies beneath the coracomandibularis, between the coraco-hyoids, is always globous, of more of less pyramidal form, and with a prolongation backward to the bifurcation of the branchial artery." This corresponds very well with the condition which I find in Raia Erinacea except that in this species, at least, the gland constantly overlies the bifurcation of the ventral aorta (branchial artery), and that it is always somewhat flattened, its ventro-dorsal axis being shortened. The organ is relatively thicker than in the Selachians because of the


THE ANATOMY OF THE THYROID GLAND 159

presence of an increased amount of connective tissue between its vesicles.

On page 26 Guiard says that the thyroid gland "of fish" is always unpaired; it is quite obvious that this remark should apply only to the Elasmobraiichs, the only order of fishes which Guiard appears to have studied.

Bridge ('04) passes the thyroid gland with the brief statement that '4n adult Elasmobranchs the thyroid is represented by a moderately large compact organ, situated near the anterior end of the ventral aorta." Although he describes the gland as one of the blood glands" in connection with the vascular system, he does not mention, nor indicate in any way, the source of its blood supply. The statement of its intimate relation with the aortic bifurcation might well lead one to erroneously suspect a supply from this source. In quite another place (page 332) he speaks of "a remarkable system of arteries for the supply of nutrient blood to the gills and heart," which takes origin from the ventral ends of the loops about the gill slits, the commissural vessels forming by their union the "median longitudinal hypobranchial artery which lies beneath the ventral aorta." He fails to mention the ultimate ramifications of this system of vessels or its relation to the thyroid gland, falling into the same error in this particular as T. J. Parker, from whose plates Bridge takes his figures, and upon whose description he appears to have largely based his text.

The literature upon the blood supply of the Elasmobranch thyroid begins with Hyrtl ('58) who first described the hypobranchial arterial system in the Batoidei, if we except the very incomplete description by Monro (1787). Hyrtl described the thyroid artery as the "Ramus thyreoideus seu submentalis" which takes origin from the 'Vein" of the second gill sac, and which gives off muscular branches to that part of the oral mucosa which lies between the inferior maxilla and the tongue bone as well as the Glandula thyreoidea." Hyrtl did not at this time describe a median hypobranchial artery, this vessel being represented in his description by two anastomosing vessels on either side of the median line which arise from the second and third arches and which


160 JEREMIAH S. FERGUSON

pass backward to supply the anterior coronary vessels. Hyrtl very clearly pointed out that the posterior coronary vessels arise from the subclavian artery in the Batoidea and in 1872 he showed that these vessels (posterior coronaries) were absent in the Selachii, a point emphasized at considerable length some years later by G. H. Parker and Davis ('99) . In 1872 Hyrtl extended his description of the hypobranchial arterial system to the Selachii. He found the thyroid gland to be supplied by the Arteria thyreo-maxillaris seu submentalis" which supplied the floor of the mouth and thyroid gland as in the Batoidei but which took origin from the veins" of the first gill sac, rather than from the second as he had previously described for the Batoidei. He also described the Arteria cardio-cardiaca," called later the "commissural" and "longitudinal commissural" (T. J. Parker) and the commissural and lateral hypobranchial" (G. H. Parker and Davis), which fused in the median line to form a median vessel (median hypobranchial) and from which the coronary vessels were derived. At this time Hyrtl emphasized the absence of the posterior coronary branches of the subclavian in the sharks and called attention to an anastomosis from the subclavian forward to the median vessel from which the coronary arteries arose. This anastomotic vessel has since been called by T. J. Parker the "hypobranchial artery."

Turner ('74) injected the conus arteriosus and studied the course of the afferent and efferent branchial vessels; the course of these vessels is now well known. He neither mentioned nor excluded any relation to the thyroid gland, apparently not recognizing this organ, nor did he work out the ultimate connections of any of the smaller cervical vessels.

T. J. Parker (*80) described the venous system of Raia nasuta and called attention to "the extraordinary number of transverse anastomoses it [the venous system of the skatej presents, the results being to produce numerous 'venous circles, 'comparable to the circle of Willis in the arteries of the mammalian brain, and the circulus cephalicus in the arterial system of bony fishes. There is also a direct passage from the sinus venosus and back again, in four different ways, namely: (1) by the hepatic sinus; (2) by the anterior part of the cardinal vein and the cardinal sinus ; (3) by the whole length of the cardinal veins and their posterior anastomosis ;


THE ANATOMY OF THE THYROID GLAND 161

(4) by the lateral veins and the prolongation of the cardinal sinus into which they open."

I would like to direct attention to the presence of similar venous anastomoses in the Selachii as well as in the Batoidei. These anastomoses in the cervical region are frequent and voluminous. A complete circular anastomosis surroundsthe mouth close to the maxillary and mandibular cartilages. Though not so readily seen in the Selachii, it follows the same course as in the skate in which fish it is visible through the skin and oral mucosa; it ends in a maxillary sinus at either angle of the mouth, which is connected with the orbital sinus and with the jugular vein. The hyoid sinuses are similarly connected across the median line near the ventral surface, two anastomotic vessels, the anterior the larger, connecting the opposite sides. This anastomosis bears a most important relation to the thyroid gland. The anterior vessel is so large as frequently to almost envelop the gland as in a capsule, the vessel is subdivided by fibrous partitions, or consists, rather, of a mass or series of vessels within a common sheath, and from its relation to the thyroid gland, in its more or less dilated condition it is more truly a sinus than a vein ; it is conveniently designated the thyroid sinus. It fills and empties with each movement of the mouth and gills as water is forced through the branchial clefts, thus functioning with the aid of extrinsic muscles after the manner of a venous heart. When the fish is examined out of t)ie water the violent movement of the gills so distends the sinus as often to wholly obscure the thyroid gland by the volume of its contained blood. It is almost impossible to reach the gland by dissection from the ventral surface without cutting the sinus or some of its numerous tributaries. The thyroid sinus receives the veins from the thyroid gland, most of these vessels leaving the dorsal surface or posterior margin of the organ.

T. J. Parker ('86) offers a description of the larger blood-vessels of Mustelusantarcticus, which is, however, deficient as regards the ultimate distribution of the smaller arteries. Exceptions may also be taken to his statement of the distribution of the arteries which constitute the rather remarkable hypobranchial arterial system, which as already mentioned, bears an important relation to the thyroid gland. Parker's description of the venous system is quite


162 JEREMIAH S. FERGUSON

accurate. The hj-oid sinus is shown to empty into the jugular vein with a valve at the orifice. The anastomosis between the hyoid sinuses is considered, but its relation to the thyroid gland is not discussed; since no mention of the thyroid gland is made it would appear that this important organ was either overlooked or ignored. Unless one is specially looking for the gland, in the effort to separate the muscles without injury to the venous channels the organ is easily broken up, and once disintegrated its particles are readily lost amongst the mass of muscular tissue.

The tributaries of the hyoid sinuses are stated by Parker to include the submental, posterior facial, internal jugular, and the nutrient veins from the first hemibranch.

Parker's description of the hypobranchial system follows that of the ventral and dorsal aorta and begins with the subclavian artery, which, he says, gives off the branchial and hypobranchial arteries. Apparently he omits to mention the large lateral or epigastric artery, whose course parallels that of the lateral vein, though the beginning of the vessel is indicated but not named in some of his figures. The hypobranchial artery described and figured as a continuation of the subclavian, after giving off the anterolateral arter}^ — which I find to be distributed to the pericardium and adjacent muscles — unites with its fellow of the opposite side, passes forward 2 cm. in front of the conus arteriosus, and forms a plexus from which are given off the coronary arteries posteriorly, and anteriorly the median hypobranchial artery. The plexus communicates laterally by two commissural arteries on either side with the longitudinal commissural vessels uniting the ventral ends of the efferent branchial loops. One gathers from the description that the course of the circulation is from the subclavian artery through the hypobranchial to the efferent branchial loops, a direction which may be thus tabulated :

branchial hypobranchial

coronary (paired)

median hypobranchial (azygos hypobranchial)

commissural (two pairs) antero-lateral (paired)


THE ANATOMY OF THE THYROID GLAND 163

Mention is not made of the lateral nor of all the coronary arteries. Parker's observations were made on Mustelus antarcticus. I have dissected three species of the Selachii and one of the Batoidei, I have not only been unable to confirm the course of the circulation as indicated but I find that beyond the so-called hypobranchial artery the course of the circulation is in the opposite direction, viz., from the efferent branchial loops to the coronary vessels and systemic capillaries, and the hypobranchial artery serves as a relatively unimportant anastomosis which, in these species is not even constantly present. In addition to the pair of coronary arteries distributed to the ventricle I have in my specimens observed a dorsal artery which ramifies largely in the wall of the auricle. The mandibular artery as described and figured by Parker, is the one from which in my preparations the thyroid artery is sometimes, though not constantly, derived. His description leaves one somewhat in doubt as to the origin of this vessel, but he has figured it correctly as coming from the first efferent branchial loop. His coraco-mandibular artery, derived from the mandibular, is a vessel which apparently corresponds with that which distributes its main branches, in my preparations, within the thyroid gland and only incidentally gives small branches to the coraco-mandibular and coraco-hyoid muscles; T have therefore called this vessel the thyroid artery.

G. H. Parker and Davis ('99) in an article on the blood-vessels of the heart in Carcharias, Raia, and Amia" repeated the work of Hyrtl ('58 and '72) so far as it immediately concerned the origin of the coronary vessels, but being concerned only with the cardiac vessels they made no mention of the thyroid artery or other derivatives of the first hemibranch, nor of the gastric and pharyngeal branches which arise in close relation to the anterior coronaries. They described "the irregular longitudinal artery by which the ventral ends of some or all of the efferent branchial arteries of a given side are brought into communication, "hitherto referred to as "longitudinal commissural" vessels (T. J. Parker) or as part of the Arteria cardio-cardiaca (Hyrtl), and called the vessels the "lateral hypobranchial artery, "reserving for the name commissural "those arteries which leave the lateral hypobranchials on their median


164 JEREMIAH S. FERGUSON

sides and, after more or less tortuous courses, unite with one another in the median plane" to produce by their union the median hypobranchial artery. The ventral continuation of the subclavian artery they call the "coracoid artery." Concerning the anastomosis of this vessel with the median hypobranchial formed by the hypobranchial artery of T. J. Parker they speak as follows: "Moreover neither of these vessels [median and lateral hypobranchial] can be properly considered a dependency of the subclavian, for the branch which leaves that artery, and which T. J. Parker regarded as their root, may be connected with them, as Hyrtl ('58, p. 17, Taf. 2) has shown, by only a relatively small vessel. The union, then, is not in the nature of a continuous trunk, but an anastomosis, and the vessel posterior to this union must be considered in the light of an independent artery. This we have called the coracoid artery."

MATERIAL AND METHODS

For the purposes of the present study I have dissected 32 specimens of Mustelus canis, 10 of Carcharias litoralis, 3 ofSqualus acanthias, and 14 of Raia erinacea. In addition to these I have had access to a number of sections from various Elasmobranchs prepared by my late assistant. Dr. Guy D. Lombard. The most of these animals were dissected through the courtesy of The Wistar Institute of Anatomy at the Marine Biological Laboratory at Wood's Hole, Massachusetts. My thanks are due these institutions for the opportunity afforded.

The form and position of the thyroid gland was carefully observed in each instance and its vascular connections determined bothby dissection and by various methods of injection. The injections were made chiefly with a hypodermic needle of very fine caliber, though finely drawn-out glass tubes were used with some success. For pressure an aspirating syringe was used for routine work and served very well; air pressure was also used at times. For tracing the lymphatics, injections were frequently made into the substance of the thyroid gland and into the connective tissue about the thyroid blood sinus and the other cervical blood vessels. For tracing


THE ANATOMY OF THE THYROID GLAND 165

the blood-vessels injections were made into both remote and nearby vessels, the points selected including the hyoid and thyroid sinuses, the thyroid artery, the efferent branchial loops and commissural arteries, the median hypobranchial artery, the coronary arteries, the ventral aorta, conus arteriosus, cardiac ventricle and auricle, the caudal artery and vein, the mesenteric artery and the dorsal aorta. Injections from these various points were made not only because of expediency in a given species but for the special purpose of determining the direction of flow and the relation of the vessels to the thyroid circulation; hence, injections were made from both sides of the branchial circulation, in the direction of the flow in the veins and the arteries while other injections were made in a direction opposite to the usual course of the circulation on the arterial side, though this was, of course, impossible in the veins because of the presence of valves.

Many of the thyroid glands were cleared and mounted in toto. This was best accomplished with those from Mustelus, in which species the gland is very thin. Some of the others were cut freehand into thick sections. These preparations gave very good pictures of the lymphatics and blood-vessels except in the case of the very thick glands. Still other thyroid glands were sectioned for histological study.

THE ANATOMICAL RELATIONS OF THE THYROID GLAND»

The thyroid gland is more or less closely related to most of the structures of the ventral cervical region, a region included between the mandible in front, the coracoid arch or shoulder girdle behind, and the branchial clefts on either side. This region forms the ven 1 The fact that the thyroid ghxnd may be readily overlooked in the Selachii is amply demonstrated by the frequency with which this region has been studied and the almost entire absence of any adequate description of the gland. A brief description of the methods of dissection which may be relied upon to locate and expose the gland is offered in the hope that it may materially aid future investigation of this organ.

The thyroid gland of Elasmobranchs can be readily reached from either the oral or the cutaneous surface. By the cutaneous route two methods are especially serviceable, the one by longitudinal, the other by transverse incision.

THE AMERICAN JODRNAl. OP ANATOMY, VOL. 11, NO. 2


166 JEREMIAH S. FERGUSON

tral wall of the pharynx and contains the whole course of the ventral aorta and its immediate branches. In Raia the heart is contained within this area at its posterior margin, Ijang in the median line just in front of the cartilaginous arch, but in Mustelus, Squalus and Carcharias the heart has been pushed backward and lies just beneath the coracoid arch.

The skin of the ventro-cervical region is thin but very tough; laterally it is folded upon itself at the branchial clefts, on the inner surface of which it becomes continuous with the mucosa covering the gills. The fibers of the mylohyoid and geniohyoid muscles are attached to the derma over the greater part of the ventral cervical area. These muscles form a thin sheet, thickest and most prominent in Carcharias, thinnest and frequently almost wanting

The first method is the more applicable in Raia, where the skin is loosely attached and the coraco-mandibular muscle is small, thin and easily lifted. With some variations I have followed the method outlined by Lombard (*09). A longitudinal incision is made in the median line through the skin and membranous constrictor pharyngis. The skin and adherent muscle are dissected away and retracted laterally, exposing the coraco-mandibularis. (Fig. 1 B."* A probe is •passed beneath the muscle which, after being well freed, is divided midway between the mandible and the coracoid arch. The anterior flap is grasped and reflected forward, at the same time dissecting away from its dorsal surface the deep cervical fascia in which the thyroid gland is embedded. The gland is easily recognized by its deep yellowish orange color and its peculiar rounded or triangular form. In those exceptional instances when the thyroid gland is displaced forward in Raia the anterior division of the coraco-mandibularis will have to be reflected forward all the way to its mandibular insertion before the gland is fully exposed; ordinarily the organ will be found directly over the aortic bifurcation about midway between the mandible and the point at which the muscle was bisected.

The above method is less easily applied to the Selachii for the reason that the skin and the constrictor pharyngis are much more firmly adherent to the underlying structures than in Raia; moreover, in reflecting forward the anterior division of the coraco-mandibularis one is almost certain to injure the thyroid sinus, deluging the part with blood, before the gland can be exposed. In Carcharias one encounters the added disadvantage that the thyroid gland is quite firmly united to the coraco-hyoideus and the surface of the basi-hyal cartilage, and the gland is buried in a mass of connective tissue by which it usually is entirely obscured even after the coraco-mandibularis has been completely reflected away from its surface. The second method is, therefore, the more applicable in Mustelus and Squalus and is very much more certain in Carcharias. One blade of a blunt scissors is inserted into the first branchial cleft on the right and then on the left side and the clefts lengthened to their extreme ventral limits, the ends


THE ANATOMY OF THE THYROID GLAND 167

in Raia, and subject in all species to great individual variation in volume. The muscles take origin posteriorly from the coracoid arch, anteriorly from the hyoid arch and mandible, and laterally from the outer surfaces of the branchial arches. Acting from these fixed points" upon the more movable, but tough and inelastic skin, these muscles form a very powerful constrictor of the pharynx and collectively are very properly termed the "constrictor pharyngis" (fig. 1,A). In addition to this constriction the muscular contraction at the same time tends to draw open the branchial clefts, thus permitting the more ready passage of water during the rhythmic pharyngeal contraction or respiratory movement. The muscular fibers of the constrictor pharyngis are intimately adherent to the derma.

of the incisions exposing the margins of the coraco-hyoideus muscle. One [)lade of the scissors is then pushed beneath the skin where it readily passes between the coraco-hyoideus and coraco-mandibularis muscles (fig. 1); the incision is continued across the median line from side to side. This divides the coraco-mandibularis; its anterior portion is grasped with the forceps, lifted, and a longitudinal incision through the skin and fascia carried forward along either margin of the muscle.' In the dogfish the divided muscle with the attached skin is easily raised and the loosely attached deep cervical fascia dissected away from its dorsal surface, exposing the thyroid gland. In Carcharias it is better to dissect the deep cervical fascia away from the ventral surface of the coraco-hyoideus muscle, rather than from the coraco-mandibularis; the thyroid gland is then raised with the latter muscle and dissected out from the mass of connective tissue which envelopes it. Finally, the gland must be dissected away from its anterior attachment to the margin of the basi-hyal cartilage, or, in Carcharias, to a median depression, in the ventral surface of this cartilage, which corresponds to the foramen caecum linguae of mammals; occasionally this depression is a true foramen, in which case the thyroid process becomes obviously analogous to the lobulus pyramidalisof the mammalian thyroid gland. This lobule is represented in Mustelus by a short triangular projection, not constantly present, which overlies a shallow median groove in the anterior margin of the cartilage. In Raia a similar condition is much less frequently present.

The thyroid gland is readily accessible from the oral cavity. A needle passed through the oral mucosa just in front of the basi-hyal cartilage— "lingual bone" —enters the substance of the thyroid gland if directed backward in Mustelus and Squalus, well backward and close to the cartilaginous surface in Carcharias, or backward and slightly ventralward in Raia. A transverse incision through the oral mucosa, parallel to and just in front of the basi-hyal cartilage, exposes the anterior margin of the thyroid gland and it may then be readily dissected out from Mustelus or Raia, though with greater difficulty from Carcharias or Squalus.


168 JEREMIAH S. FERGUSON

Removal of the integument with the adherent constrictor muscles exposes the coraco-mandibularis (fig. 1, B), a slender paired muscle, its two sides intimately fused in the median line, which takes origin by a tendinous fascia from the ventral surface and anterior margin of the coracoid arch. The paired muscle passes forward to its insertion, ending in short, rounded and slightly divergent tendons which are attached to the posterior margin of the inferior mandible. The muscle is inclosed within the folds of a superficial cervical fascia, which forms its aponeurosis and extends laterally to the surface of the branchial arches, but on either side of the muscle, the aponeurosis fuses with the deep cervical fascia with which it is in more or less close contact.

On lifting the coraco-mandibularis with its superficial cervical fascia the coraco-hyoid muscle (fig. 1, C) is exposed; it is similar in shape and appearance to the coraco-mandibular, but is much broader, its lateral margin projecting from beneath the coraco-mandibularis, and in Mustelus, Squalus and Carcharias extending laterally almost to the ventral ends of the branchial clefts, or even overlapping them somewhat. In Raia the gills are more widely separated, leaving a broad portion of the floor of the pharynx exposed at the side of the coraco-hyoideus in the anterior portion of the cervical region. Posteriorly the several divisions of the coraco-branchialis muscle cross this exposed portion of the pharyngeal floor to be inserted into the branchial arches.

The ventral surface of the coraco-hyoideus is smooth, its dorsal surface separates into several muscular processes, the musculus coraco-branchialis {M. c. br., fig. 5), to be inserted into the membranous floor of the pharynx by a tendinous fascia overspreading and firmly adherent to the fibrous pharyngeal submucosa and the surfaces of the cartilaginous branchial arches. The divergent portion of the coraco-hyoidei, on either side of the median line, are similarly inserted into the movable basi-hyal cartilage (lingual bone), so that the combined coraco-hyoid and coraco-branchial muscles, arising from the anterior border of the coracoid arch, form a very powerful dilator of the mouth and pharynx.

As the coraco-hyoideus does not extend forward beyond the hyoid arch it exposes the membranous floor of the oro-pharynx


THE ANATOMY OF THE THYROID GLAND


169



Fig. 1. Dissection of the thyroid gland of Mustelus canis. In A the skm has been reflected to show the superficial "constrictor pharyngis' muscle. In B the "constrictor pharyngis" has been removed and the coraco-mandibulans exposed. In C the coraco-mandibularis has been divided, exposmg the thyroid gland lying upon the coraco-hyoideus. (See page 209 for explanation of abbreviations used in all figures.)


170 JEREMIAH S. FERGUSON

between this point and the inferior mandible. The cartilages of the hyoid arch except only the dorsal surface of the basi-hyal, are loosely adherent to this membranous floor •so that when the mouth has been closed and the pharynx contracted the tongue-like basi-hyal cartilage is pushed forwaid, producing a deep fold in the oral mucosa. A needle thrust through this fold in the median line, passing dorsal to the cartilage, penetrates directly into the thyroid gland.

Not all of the ventral surface of the basi-hyal cartilage is covered by the insertion of the coraco-hyoideus muscle ; the portion of the cartilage thus exposed varies in different species and to some extent in individuals. In Carcharias and Raia the muscle is inserted into only a small portion of the cartilaginous surface, while in Mustelus all but a narrow anterior margin is covered by the muscle fibers. The thyroid gland typically lies upon this exposed cartilaginous surface, extending backward for a greater or less distance upon the ventral surface of the coraco-hyoideus (fig. 1, C). In Mustelus and Squalus the ventral surface of the basi-hyal cartilage has a raised arciform anterior margin with a very slight medial depression, in Raia it is nearly flat, and in all these species the thyroid gland overspreads the cartilage like a thin membrane whose convex anterior border nearly corresponds with the outhne of the cartilage; in Carcharias the cartilage presents a deep median groove or furrow into which the thyroid gland sinks, lying there in a gelatinous mass of connective tissue so voluminous that the gland is partially, sometimes wholly, obscured.

In the dogfish and shark the thyroid gland is rarely pushed forward beyond the margin of the basi-hyal cartilage; in Raia the organ may extend farther forward so that it rests in part upon the membranous floor of the oro-pharynx. In one of the skates I found the gland carried so far forward that it lay wholly in front of the cartilage. In Mustelus and Carcharias individual variations are much less frequent than in Raia, but in Raia in the majority of individuals the gland lies directly upon the bifurcation of the ventral aorta (fig. 7).


THE ANATOMY OF THE THYROID GLAND 171

THE THYROID VESSELS

The anterior margin of the thyroid gland is firmly attached by a dense fibrous fold of fascia to the antero-ventral margin of the basi-hyal cartilage. Its ventral surface is in contact with the coraco-mandibular muscle, to which it is firmly united by a fascia. Its dorsal surface rests upon the basihyal cartilage and the insertion of the coraco-hyoid muscle as described in the preceding section. The lateral angles and posterior margin of the thyroid gland are continuous with a fold of the deep cervical fascia which is placed between the coraco-mandibular and coracohyoid muscles, loosely attached to the opposed surfaces of these muscles, but more firmly fixed to their lateral margins, so as to form an aponeurosis for each. This fold of the fascia is much thickened anteriorly where it approaches the margin of the thyroid gland: at this point it incloses a transverse anastomosing vein which connects laterally with the hyoid sinuses.

At the posterior margin of the gland the fascia splits, a thin layer passing dorsally between the gland and the coraco-hyoid muscle, a thicker portion extending forward over the ventral surface of the organ, between it and the coraco-mandibular muscle. This ventral division is of special importance; it contains the largo thyroid sinus consisting of an intricate net of veins and lymphatics, which connects laterally with the hyoid sinus and in the median line spreads over the ventral surface of the thyroid gland so that when distended with blood it entirely obscures the organ.

The thyroid sinus is surrounded with connective tissue containing a network of lymphatic vessels. Ink injected in the living animal into the space between the lateral margin of the thyroid sinus and the ventral end of the first branchial cleft will after a few minutes be found filling many of the lymphatic vessels of the thyroid sinus as well as many other perivascular lymphatics in relation with most of the cervical veins and the arteries of thehypobranchial system (fig. 8). The lymphatics of the thyroid plexus are so mumerous and anastomose so freely that when filled with ink they form a sac-like investment entirely obscuring the sinus and the thyroid gland (fig. 2). An attempt to inject with


172 JEEEMIAH S. FERGUSON

a fine needle directly into the thyroid sinus may force the fluid into either the venous or the lymphatic plexus, according as the one or the other sj'^stemof vessels happens to be entered. I am convinced, as a result, of my injection experiments, that the lymphatics open freely into the veins of this part, after the manner of the vasa lymphatica" of Favaro {vide infra). These observations are of interest in connection with Baber's inability to find lymphatics in the thyroid gland as indicating the relation between the vascular systems. I find, however, that it is not possible for fluids injected into the thyroid sinus or its plexus of lymphatics to pass in any quantity into the vessels within the thyroid gland; this is presumably because of the presence of valves.

The alternate expansion and contraction of the mouth and pharynx, forcing the stream of water through the branchial clefts, alternately fills and empties the thyroid sinus, so that by means of these respiratory movements the sinus acts somewhat after the manner of a venous heart. In this connection it is interesting to consider the observation of Favaro ('06), as quoted by Sabin ('09), that the relation of the veins and lymphatics in fishes is much more primitive than in mammals and that both lymphhearts and vein-hearts may be present in these animals. The emptying and filling of the veins can readily be seen in the Selachii or Raia on removing the skin, or even through the integument in the living skate, the colored blood showing readily through the vascular walls. The relation of the lymphatics to the blood-sinus is so intimate that they must also be emptied and filled in the same way, though since they contain a colorless fluid they can not be so readily observed. I have, however, demonstrated that a colored fluid injected into these lymphatic vessels will overspread the thyroid region in ten to fifteen minutes and will almost entirely disappear within the next fifteen minutes ; the lymphatic circulation must therefore proceed with considerable rapidity.

The thyroid artery approaches the organ from either side; in Mustelus and Squalus it enters at the extreme lateral angle of the triangular gland (fig. 19). In Raia, where the organ is of a more rounded form it enters near the middle of the lateral border. The branches of the artery ramify upon the surface of the organ send


THE ANATOMY OF THE THYROID GLAND


173


. ' -mouth



■ 'I ihynsn.





-M.C.hy.


^W'"^^'"'^


\M


});■*!



hy.sn.


Fig. 2. Showing the form and relations of the thyroid sinus when injected with ink. A, in Raia. B, in Carcharias.


174


JEREMIAH S. FERGUSON


ing finer twigs into the interior. In this particular they offer an interesting analogy to the condition found by Major ('09) in man. Major says (page 484), "in the human the branching of the large arteries takes place mostly upon the surface of the gland, and having by their branching obtained their approximate distribution, the smaller branches are sent in." In the Elasmo


Fig. 3. Outline of the arterial supply of the thyroid gland of Mustelus canis as usually found. The vertical shading indicates the area supplied by the right thyroid artery, the horizontal shading that supplied by the left.

Fig. 4. Showing the less usual distribution of the thyroid arteries; the shading as in fig. 3.

branchs the condition might well be described in the same words; this is the more remarkable inasmuch as Major states that this is not the condition in other mammals, e.g., the dog and cat. The Elasmobranchs, therefore, seem to harmonize with the human rather than the lower mammalian condition as regards the distribution of the main branches of the thyroid arteries.


THE ANATOMY OF THE THYROID GLAND 175

The left thyroid artery usually supplies a greater portion of the gland than the right (fig. 3), though the relative area is subject to extreme variation and in occasional instances the ratio may be reversed (fig. 4). The area of distribution in the great majority of individuals is approximately as indicated in fig. 3.

THE HYPOBRANCHIAL CIRCULATION AND THE ORIGIN OF THE

THYROID VESSELS

In attempting to trace the circulation of the thyroid gland by means of injection experiments I was at once struck with the difficulty of reaching the gland by means of injections into the gill arteries, the ventral aorta or the heart. It was obvious that the thyroid artery has no direct connection with the ventral aorta or the afferent branchial vessels, a fact which seems to have been first observed by Simon ('44) who stated, without further explanation orany outline of his reasons therefor, that the thyroid gland in Raia never receives the smallest share of supply from the branchial artery with which it is in contact." This fact seems to have been seldom recognized and never sufficiently emphasized by later writers.

The thyroid artery arises either from the mandibular artery, or by a common trunk with this artery, from an arterial sinus at the ventral extremity of the first branchial cleft (figs. 5 and 6) ; this sinus forms the ventral portion or connecting vessel of the efferent vascular loop contained in the hyoidean hemibranch and first holobranch. From the dorsal extremity of this same loop the first efferent branchial artery passes to the dorsal aorta. It is therefore necessary for fluid injected into the ventral aorta or its immediate branches to pass through the gill capillaries before it can enter the thyroid arteries, and few injection fluids readily pass through capillary vessels. On the other hand, fluid injected into the thyroid artery or, as I later found, into any portion of the hypobranchial system passes readily into the vessels of the thyroid gland; such injections I repeatedly made, into the sinus at the ventral end of the first efferent branchial loop, into the thyroid artery, the median hypobranchial artery, and even into the coronary artery


176 JEEEMIAH S. FERGUSON

taking care to prevent the escape of the fluid through the coronary vessels into the sinus venosus.

The hypobranchial system of vessels is so important for the thyroid gland as to deserve more than passing mention. T. J. Parker ('86) has described this system in connection with his much quoted work on the circulation in Mustelus antarcticus, but he makes no mention of its relation to the thyroid gland, in fact, he mentions neither the gland nor the thyroid artery; the gland was apparently not observed.

The main trunk of the hypobranchial system, in the species which I have examined is the median hypobranchial aitery ; it is formed by the commissural arteries coming from the ventral ends of the loops formed by the efferent branchial vessels which receive blood from the gill capillaries. These loops surround each branchial cleft, and within the gill they lie parallel to the afferent branchial arteries; they are just antero-internal to the afferent vessels. The ventral efferent vessels are very much smaller than either the dorsal efferent or the afferent (fig. 5). Opposite the ventral end of the second and third branchial clefts (sometimes only the second or the third) each loop gives off a commissural branch which passes inward and somewhat backward to the ventral surface of the ventral aorta where these vessels, with consider^ able variations, unite with their fellows of the opposite side to form a median hypobranchial artery which is frequently double so as to form a sort of elongated arterial circle. Sometimes the vessels fail to unite in front so that instead of a median hypobranchial there is a right and left hypobranchial artery, one on either side of the ventral aorta (fig. 5). Frequently the vessels so unite as to form an annular anastomosis which encircles the aorta, but portions of the ring may be absent. Some of these variations are indicated in figs. 5 and 6. The varied arrangements of these vessels are all indications of a more or less complete fusion of the commissural vessels to form a median hypobranchial artery. This vessel terminates posteriorly in a small sinus-like dilatation, single or double as the case may be. From this sinus the coronary vessels arise either as a median vessel which promptly divides, or as two or three independent vessels. From this same sinus a small paired


THE ANATOMY OF THE THYROID GLAND


177




Fig. 5. Diagram of the hypobrancliial arterial circulation in Mustelus canis; the insertions of the ventral divisions of the coraco-branchialis muscle are also shown. Ventral view. Anastomosis with the subclavian artery was wanting in this specimen.


178 ' JEREMIAH S. FERGUSON

artery passes backward on either side of the median line beneath the dorsal portion of the pericardium at the lateral margin of the cartilaginous floor of the pharynx formed by the basi-branchial cartilage ; after anastomosing with its fellow of the opposite side beneath the apex of the cardiac ventricle it distributes its terminal branches to the wall of the esophagus and stomach near the cardia (figs. 5 and 6).

From the loop at the ventral end of the fourth branchial arch a very small anastomotic branch (less frequently arising as in fig. 6, hypohr') passes backward along the lateral wall of the pericardium and penetrating between the precaval sinus and the coracoid arch anastomoses with the subclavian artery just prior to its division into the brachial (axillary) and the lateral (or hypogastric, a large artery lying parallel to the lateral vein. This anastomotic branch is undoubtedly that which T. J. Parker ('86) describes as the hypobranchial, which, according to his description receives blood from the subclavian and supplies the coronary arteries and whole hypobranchial system. Such is not the case, however, in any of the species I have studied and Hyrtl ('72) in his careful study of various species of the Selachii did not so find it, nor did Parker and Davis ('99). The hypobranchial is a very small artery, so small that its connection with the median hypobranchial is scarcely traceable, and in some individuals is entirely wanting (fig. 5), there being in these cases a small branch from the subclavian and a similar vessel from the median hypobranchial which follow the usual course but never unite, the subclavian branch distributing its blood to the muscles while the anterior division supplies the lateral pericardial wall and the adjacent muscles in front of the coracoid arch. Certainly where the vessel is wanting the flow of blood can not be in the direction indicated by T. J. Parker. Parker and Davis ('99), as already quoted, found the hypobranchial artery insignificant, though they did not record its absence.

If the median hypobranchial artery of Mustelus be injected the major portion of the fluid passes into the coronary arteries and thence through the coronary veins to the sinus venosus and auricle, while at the same time very little passes through the hypobran


THE ANATOMY OF THE THRYOID GLAND


179


chial artery; the fluid pours into the sinus venosus and auricle very freely before it has even reached the subclavian by way of the hypobranchial artery. This is the case whether the fish has been previously bled or not. This experiment would certainly show



COTXiCOld


A.lat


Fig. 6. Diagram of the hypobranchial arterial circulation in Squalus acanthias. Lateral view.


that the hypobranchial is of too small a caliber to supply the blood necessary to fill the coronary arteries, and if it can not supply this much it certainly is still less competent to supply blood for the whole hypobranchial system, which T. J. Parker's description


180 JEREMIAH S. FERGUSON

would seem to indicate was the case. The whole system may be readily injected from any one of the ^ih-loops with which it is connected. The true direction of flow is therefore from the efferent gill-loops through the commissural arteries to the median hypobranchial and from it to the muscular, pericardial, gastric, esophageal, and coronary branches, with only a relatively insignificant and inconstant anastomotic supply from the subclavian artery.

Anastomoses in both the arterial and venous systems, forming "circles" about the body wall and the viscera are of very frequent occurrence in this class of fishes as was pointed out for the venous system by T. J. Parker in 1880 {vide supra). The subclavian artery forms such a circle beneath the coracoid arch and several similar circles" are formed by anastomosis between the two sides in the hypobranchial system as well as in other parts of the body with which we are not now specially concerned. The arrangement in the arterial system is therefore very similar to that which Parker found in the venous.

At the ventral extremity of each efferent gill-loop, at the point where the hypobranchial commissural arteries arise, is a small sinus-like dilatation (figs. 5 and 6, s. a. v.) which obviously serves as a reservoir where the blood coming from the two sides of the loop, which are in adjacent branchial arches, will intermingle, and from this sinus blood is distributed through the commissural arteries (lateral hypobranchial" of Parker and Davis) anteriorly, posteriorly, or to the median hypobranchial in such proportion as the caliber of the several vessels and the course of the circulation dictate. The arterial sinus at the ventral end of the first gill-loop (first ventral sinus) is usually a trifle larger than the others. The thyroid artery arises from the anterior end of this sinus or from the adjacent portion of its anterior limb in the hyoidean hemibranch. It arises either as a separate and independent vessel or as a conjoined trunk with the mandibular artery; more frequently, in the specimens I have dissected, it was independent. The artery passes directly forward and inward to the extreme lateral border or angle of the thyroid gland. It continues its path along the surface of the thyroid gland (figs. 3, 4 and 19) near


THE ANATOMY OF THE THYROID GLAND 181

its anterior border, distributing its main branches to the substance of the gland and small collateral branches to the floor of the pharynx in front of the hyoid arch and to the anterior third of the coraco-hyoideus and coraco-mandibularis muscles. A small median unpaired vessel arising from the left thyroid artery (less frequently from the right), penetrates the thyroid gland, divides, and enters the coraco-hyoideus to supply the antero-median portion of this muscle. This is very probably homologous with the anterior portion of the arteria thyroidea impar, derived from the median hypobranchial as described by Hyrtl (72).

The left thyroid artery is usually larger, longer, and more extensive as to its area of distribution than the right. Lombard ('09) dissected a number of specimens of Mustelus and Raia and found that the left thyroid artery more frequently entered the dorsal surface, and the right the ventral surface, of the thyroid gland. I have found a somewhat similar condition, though I very frequently find both vessels coursing upon the ventral surface and sending their branches dorsally into the substance of the gland. Occasionally the right thyroid artery enters the dorsal surface of the gland and the left the ventral (fig. 3 and 19). The position and distribution of the thyroid arteries is, however, subject to considerable variation and, as I have already pointed out, the right may even supplj^ a greater portion of the gland than the left thyroid artery (figs. 3 and 4).

The thyroid artery as described very probably in part corresponds to the vessel which was recognized by T. J. Parker ('86) as the coraco-mandibular artery. This latter is an obviously inaccurate designation, for the coraco-mandibular branches are insignificant as compared with the other ramifications of the artery. Hyrtl ('72) recognized and more accurately described the thyroid artery as arising from the "veins of the first gill-arch" in conjunction with the submental artery; his description appears to be accurate with the exception that the two arteries in my dissections appear more frequently to arise independently.

The observations which I have recorded concerning the origin

THE AMERICAN JOURNAL OP ANATOMY. VOL. 11, No. 2


182


JEREMIAH S, FERGUSON




Fig. 7. The position and relations of the thyroid gland in Raia.

Fig. 8. Injected lymphatics in the tunica adventitia of the right commissural artery at its junction with the median hypobranchial. The specimen is from Carcharias, the fish having been injected with ink at a point just ventral to the right hyoidean hemibranch. The injection followed the lymphatics and was traced as far as the median hypobranchial artery in one direction and into the thyroid gland in the other.


THE ANATOMY OF THE THYROID GLAND 183

and course of the thyroid artery and the hypobranchial system were carefully worked out with specimens of Mustelus and Squalus and verified in all thetr important particulars in Carcharias and Raia.

VEINS AND LYMPHATICS OF THE THYROID REGION

The numerous small veins of the thyroid gland discharge into the thyroid sinus," which connects together the hyoid sinuses of the two sides. The conformation of the hyoid sinuses and their tributaries and connections have been well described by T. J. Parker ('86). In addition to the transverse anastomosis formed by the thyroid sinus, the hyoid sinus receives a submental vein from the region of the mandible and numerous small muscular branches from the neighboring muscles. This sinus and its connecting vessels can be most readily observed in Raia. The submental vein is seen to begin as a double transverse anastomosis; the larger, anterior, tributary lies close behind the cartilage of the inferior mandible; the smaller, posterior vessel arches across the floor of the mouth just in front of the hyoid arch and the anterior border of the thyroid gland. At the angle of the jaw these vessels unite in a small sinus which also receives a transverse anastomosis from in front of the maxilla, so that the mouth is thus encircled by an annular venous sinus. The thyroid sinus similarly forms a double transverse anastomosis, rather more deeply placed, behind the hyoid arch at the posterior border of the gland. These vessels convey the blood from the ventral cervical region to the hyoid sinus.

The thyroid veins open into the thyroid sinus as several small branches, the largest of which are a median vein, leaving the organ near the middle of its dorsal surface, and two anterior veins which leave the same surface near the anterior margin of the organ, but a little to either side of the median line. Other smaller veins leave the lateral margins of the organ passing either to the thyroid or the hyoid sinus. The thyroid veins must contain valves, for although the vessels can be readily traced with the dissecting microscope and even with the naked eye, it is with difficulty that


184 JEREMIAH S. FERGUSON

fluid injected into the thyroid or hyoid sinus can be forced back into the venous channels of the thyroid j^land; an extreme pressure will accomplish this result to a limited extent onl3\ I have been able to find some traces of valves in microscopical sections.

The hyoid sinus passes around the base of the hj^oidean hemibranch to connect with the jugular vein, through which a portion of its blood is transmitted to the precaval sinus beneath the coracoid arch, and thence to the sinus venosus and auricle. The flow through the hyoid sinus in this direction is quite intermittent, and, as already indicated, it is chiefly dependent upon the muscular force of the pharynx as it alternately relaxes and contracts to force water through the gill-openings.

Blood is also transmitted from the hyoid sinus to the heart by the more ventral and direct path through the mferior jugular (anterior cardinal) vein. This vessel maintains a more constant flow, receiving blood from the ventral cervical region and the branchial arches, along the ventral ends of which it courses to terminate in the precaval sinus.

In Raia the thyroid sinus is thin, and its investment of connective tissue containing the lymphatic plexus is less pronounced than in the other species studied, so that, except when the sinus is fully distended with blood, the gland in Raia is not much obscured. In Mustelus the sinus is larger and the fascia about it is more voluminous so that the gland is usually more or less obscured, though there is much individual variation: the same is true of Squalus. In Carcharias the vascular walls in the sinus are so thick, and the connective tissue about it so abundant that in most of the animals examined the outline of the thyroid gland contained within this mass could only be discerned on holding up the stretched membranous mass between the eye and the bright sun so that the intense transmitted light showed the yellowish orange gland contained within the connective tissue mass.

The thyroid lymphatic plexus forms an extensive group ot vascular channels surrounding the gland and the vessels of the blood sinus. It is contained in a fold of the deep cervical fascia which stretches across from side to side between the ventral ends of the first branchial clefts: it is broad in the mid-portion,


THE ANATOMY OF THE THYEOID GLAND 185

but tapers from the postero-lateral angle of the thyroid gland outward to the tissue surrounding the hyoid sinus. The vessels form perivascular lymphatics about the venous sinuses. Ink or a colored fluid injected into the connective tissue about the hyoid and thyroid sinuses readily fills the anastomosing vessels forming a sheetlike mass of peculiar form (fig. 2, thyr. sn.) . Ink thus injected can also be traced into the perivascular lymphatics of the hypobranchial arterial vessels (fig. 8) as far backward as the walls of the coronary arteries; it can likewise be found in small perivascular lymphatics in the walls of the thyroid arteries and to some extent in the broad venous spaces between the vesicles of thethja'oid gland, indicating that the lymphatic vessels to some extent may open into the veins of the thj^roid. The vessels of the lymphatic plexus in the cervical fascia are apparently connected with the blood-vessels of the thyroid sinus, for excessive pharyngeal contraction in the living fish forces blood into areas which otherwise appear to be occupied only by lymphatic vessels. The blood-vessels may without doubt be classed as venae lymphaticae" and the lymphatics as vasa lymphatica" after the terminology of Favaro ('06), who says that the same vessel may in fishes carry either blood or lymph at the same or different times so that these vessels may in this sense be either vasa or venae lymphaticae. Fluid injected into the lymphatics spreads so rapidly over so great an area that it seems almost impossible to trace a connection with the blood sinus by means of injections; the fluid enters the bloodvessels so readily that one is unable to exclude the possibility of an intra- venous injection.

The statement by Baber ('81) that he was able to demonstrate no lymphatics in the thyroid gland of Elasmobranchs led me to pay special attention to the study of these vessels by injection methods. As I have already pointed out, Baber states that in both the skate and the Conger-eel an extensive system of vessels lined with epithelium becomes injected by the method of puncture." He then injected the blood vessels of a Conger-eel with Berlin blue through the "efferent branchial vein" and"dorsal aorta and thereupon states that "in the Conger-eel at least, there is no evidence of any system of lymphatic vessels," emphasising


186


JEREMIAH S. FERGUSON


his statement b}^ the use of itahcs. I can confirm that portion of Baber's statement which says that an extensive system of vessels within the thyroid gland can be readily injected bj^ the method of puncture, but I would maintain that neither that procedure, nor the injection of the dorsal aorta or efferent branchial vessels with Berlin blue, would demonstrate the absence of lymphatics; that they may still be present, I have demonstrated both in microscopical sections and by injection (figs. 9, 10 and 11).



Fig. 9. Lymphatics, "vasa lymphatica," and veins, "venae lymphaticae," of the thyroid gland. The "vasa lymphatica" have been injected with ink and the thyroid gland cleared and mounted in toto; the lumen of the follicle and the follicular epithelium are only indistinctly seen. At X in the specimen the two sets of vessels anastomose.


Injection by puncture does not always fill the extensive system of vessels observed by Baber. If one takes care to use only a very gentle pressure, this system, which completely surrounds each vesicle, is only filled near the point of injection, while at the margins of the injected area the fluid spreads through more minute vessels which lie in closer contact with the vesicular epithelium (fig. 9, L) . I believe these last are true vasa lymphatica in the sense


THE ANATOMY OF THE THYROID GLAND . 187

of Favaro and I find the contents of the vesicles apparently secreted into them as in the mammahan thyroid (fig. 10). The larger vascular channels then are venae lymphaticae, readily injected by puncture if the pressure is excessive, the veins being easily entered because of their large caliber and extremely thin walls; they transmit only lymph when the intravenous bloodpressure is low within the gland, but fill with blood when from any cause the pressure is raised. I have invariably found some blood cells in the venae lymphaticae; I have never found them filled with blood in all the three score animals I have examined except in one case in which as a result of injury the thyroid gland was greatly congested. In this case they were filled to distension. In microscopical sections I have been able to trace the connection of the vasa with the venae lymphaticae (fig. 10). I have been unable to demonstrate positively the presence of any valves at the orifices of these vessels, but the extreme obhquity of the anastomosis considered in conjunction with the very thin vascular walls might well serve a valvular function when the blood pressure is low, though with increased pressure and venous distention some blood would be forced back into the vasa lymphaticae and even into the vesicles. The frequent occurrence of red blood corpuscles within the vesicles of all animals is well known and in the Elasmobranchs it is thus accounted for. The intimate relation between the venous and lymphatic systems pointed out by Sabin ('09) would possibly suggest that an homologous vascular relation may account for the presence of red blood corpuscles within the vesicles of the mammahan thyroid gland.

THE HISTOLOGY OF THE ELASMOBRANCH THYROID GLAND

The thyroid gland in Elasmobranchs consists of a mass of vesicular follicles (figs. 10, 11 and 13 to 18) which very closely resemble those of the mammahan gland. The vesicles are fined by epithelium of a low columnar type, contain more or less colloid material, and are loosely bound together by a connective tissue framework which is very richly supphed with blood-vessels.

The shape of the gland in Mustelus canis(fig 1,C) is sufficiently







^f^ -'-■-^'^, y^^^j^mr <:^-^-\ ^f^^ - ^,:..M0^ \ V 11


"■-iii


..•;^:^:S«Si|


v..


. Fig. 10. Section of the thyroid gland of Mustelus eaiiis sliowing the anastomosis at the point X of the vasa and venae lymphaticae.

Fig. 11. Typical section of the thyroid gland of Mustelus canis; the intimate relation of the epithelium of the thyroid follicles to the lymphatics and blood vesvels is accurately shown.


THE ANATOMY OF THE THYROID GLAND 189

peculiar to deserve passing mention. It may be described as consisting of two triangles whose bases are fused in the median line, the apices directed outward, the anterior borders convex and conforming to the anterior margin of the basi-hyal cartilage, the posterior borders concave and free, except for their attachment to the deep cervical fascia. In the median line the conjoined bases are prolonged backward to form a short median projection; anteriorly a shallow notch separates the two lateral triangular halves. The gland is approximately bilaterally symmetrical (figs. 1, 3 and 4).

The thin, almost membranous character of the gland in Mustelus canis offers an excellent opportunity for the recognition of a lobar or lobular structure if such exists, for the whole gland is frequently no more than four or five follicles in thickness and may be stained, cleared and mounted in toto, giving very excellent microscopical pictures of the entire organ. I have not been able to find any indication of definite lobes or lobules. Portions of the thyroid substance are here and there wanting, as observed by Lombard ('09), and these deficient areas occur more frequently in the posterior than in the anterior half of the gland. In one of the thirty-two fishes of this species the deficiencies were so great that the gland was only represented by a few specks which were positively identified as portions of the thyroid only after microscopical examination. A similar case was found in Squalus, and one gland from Carcharias consisted of three small pieces.

Occasionally the posterior border presents a notched deficiency in or near the median line. There may be one, two or three such notches, either symmetrically or asymmetrically disposed. Deficiencies of the thyroid tissue also occur within the gland and may, or may not, be connected with the notches in the posterior border. These deficiencies are all of inconstant occurrence, irregular location, and could scarcely be taken to indicate any suggestion of definite lobes. They seem rather to be due to the extreme thinness of the gland and in many of the thicker specimens they are in no way indicated. When present they are occupied by connective tissue continuous with the glandular capsule. Frequently they transmit the larger thyroid vessels.


190 JEREMIAH S. FERGUSON

Bits of thyroid tissue of inconstant form or location are occasionally separated from the body of the gland by narrow partitions of connective tissue; they are most frequently found near the border of the gland or adjoining an area in which the thyroid substance is deficient. Since they possess no constant relation to the vascular supply, the detached masses can not correspond in any sense to true anatomical lobules. The arteries branch irregularly, for the most part after a somewhat dicotymous fashion (fig. 12), the arterial twigs passing off at acute angles. Partially injected specimens in which the injection fluid has passed through the arteries but has not penetrated in quantity into the veins



Fig. 12. Terminal divisions of the thyroid artery. The area occupied by injected capillaries, "venae lymphaticae," directly connected with each terminal arteriole is roughly indicated by the dotted lines.

show areas of injected capillaries surrounding the terminal arterioles (fig. 12), but the extent of these injected areas and their relation to the artery seems to be dependent rather on the pressure of the injection than on any constant or characteristic relation to the vascular system. I can not recognize any probable vascular or anatomical unit which might in any sense serve as an anatomical lobule or structural unit, as described for various other glands by Born, Mall and others.

In Raia the occurrence of partially detached groups of thyroid follicles is more frequent than in the other species, but the number of such groups present in a gland varies from two or three to a score or more. The groups are outlined by connective tissue in


THE ANATOMY OF THE THYROID GLAND 191

which broad venous spaces to a certain extent encircle the quasilobiile. The veins thus lie at the periphery while the artery on reaching the group promptly breaks up into a plexus of broad capillary spaces — venae lymphaticae — which surround the follicles within the quasi-lobule. The number of follicles in the group varies from four or five to several score.

In Carcharias the condition is similar to that in Mustelus, there being no indication of lobular groups except about the occasional irregular deficiencies in the thyroid mass. Except for the anatomical disintegration of the gland in one fish there was similarly no indication of lobulation in Squalus, but as none of my specimens from this species were prepared as total mounts I can not speak with the same certainty as in the other species.

The form of the thyroid folhcles is subject to considerable variation, but, in general, they may be said to be of ovoid shape, and, as pointed out for the mammalian thyroid by Streiff ('97), they present frequent diverticula. The Elasmobranch thyroid differs from those described by Streiff in that they show very little tendency to branch and no indication of a tubular character when the whole follicles are examined in total mounts of the gland (fig. 13) . In cut sections diverticula are of frequent occurrence and are apparently the result of pronounced infoldings of the follicular wall rather than of any protuberance, or of any tendency of the follicle to branch. Fig. 13 shows characteristic follicles from all four species; the figures are of whole folhcles and differ from the cut sections in that only the largest infoldings of their wall are visible. As already indicated, diverticula are more apparent in sections than in the preparations (total mounts) from which the drawings have been made. The particular follicles drawn from Carcharias present rather greater infoldings than those from the other species. I have not, however, observed that this is characteristic of Carcharias. In the figure the magnification is the same for the several Selachian species but less by one half for Raia; the follicles of Raia are, therefore, relatively about twice as large as shown. The size of the folhcles is subject to considerable variation as regards the individual follicles, the different thyroids, and the


192


JEREMIAH S. FERGUSON



O




o ^^



B



Fig 13. Outline of the follicles of thyroid glands as seen in total mounts A from Mustelus canis, X 152. B, from Squalus acanthias, X 152 C from Carcharias litoralis, X 152. D, from Raia erinacea, X 80


various species. I have tabulated the results of some of the measurements.


SPECIES


Mustelus. . Squalus. . . , Carcharias Raia


MAXIMUM

DIAMETER OF

FOLLICLE


MINIMUM AVERAGE

DIAMETER OFDIAMETER OF

FOLLICLE FOLLICLE


.160mm. .079 .273 .340


.017mm. .013 .023 .053


.067mm. .047 ,100 .167


AVERAGE

LENGTH OF

FISH


67.8cm. 53.7 117.2 46.2


RATIO

FOLLICLE TO

FISH


101.7

115.

117.2

27.7


AVERAGE

DIAMETER

THYROID

GLAND


13.6mm. no data 13.7 6.5


THE ANATOMY OF THE THYROID GLAND 193

The very large relative size of the follicles of Raia is at once apparent. They are approximately four times as large, relatively to the length of the fish, as in the case of any other species. When compared with the diameter of the gland the ratio is again increased, but this difference is in part compensated for by the increased thickness of the gland in Raia as compared with the other species. The thyroid gland of Raia is 1,5 times as thick as that of Carcharias, and 2 to 2,5 times the thickness of the gland in Mustelus.

The column of ratios in the above table would indicate that in the Selachians the size of the follicle is in approximate proportion to the size of the fish, but that in Raia the relative size of the follicle is many times as great; the actual number of follicles in the thyroid gland of Raia is only a small fraction of those in the gland of any of the other species. It is readily susceptible of mathematical proof that the combined circumference of large follicles contained in a given area is less than the combined circumference of smaller follicles in the same area; hence the gland cf Raia with its larger follicles will contain proportionately less epithelium than the glands of the other species. I estimate that the difference is just about sufficient to render the volume of secretory epithelium in the gland of Raia relative to the size of the fish equal to the volume of secretory epithelium in each of the other species.

But it is equally susceptible of mathematical proof that the cubical contents of the combined follicles is greater in the gland having the larger folhcles; hence there is in Raia a greater volume of intrafollicular space than in the other species. It is scarcely susceptible of proof but entirely reasonable to suppose that the epithelium of different individuals of the same or different species so closely allied and the Batoidei and the Selachii is approximately equally active as regards its secretory function. There is ample evidence that the fluid secreted by the thyroid epithelium into the cavity of the follicle finds its way through the wall of the follicle to the neighboring vascular spaces so that the direction of flow must, in part at least, be from the epithelium into the follicle and thence through the follicular wall to the vessels. In view of these facts the rate of this secretory flow in Raia, with its relatively


194 JEREMIAH S. FERGUSON

large intrafollicular space must be slower than in the other species with their relatively small intrafollicular cavities, or, to express it differently, there is relative stagnation in intrafollicular secretorj^ flow in the case of the thj^-oid follicles of Raia. It is well known that an albuminous secretion which is rendered relatively stagnant within the epithelial cavities of the body tends to produce colloid masses whose microchemical reactions more or less closely resemble those of the colloid material of the thyroid gland; this occurs, e.g., in the ducts and tubules of the resting mammary gland and in dilated cystic tubules in the kidney. We would therefore expect that in the thyroid follicles of Raia with their relatively stagnant secretory flow we should find an increased amount of colloid material. This I find to be the case, the proportionate volume of colloid present in the follicles of Raia being decidedly greater than in the other species. Similarly I find it the rule that the larger follicles contain relatively more colloid than the small follicles in the same gland. The volume of colloid contained in the thyroid follicles, therefore, can not be regarded as an index of the activity of the secretory epithelium; it would rather appear as a sort of by-product whose volume was dependent upon the rate of flow in the fluid from which it was formed. This view harmonizes the appearance of colloid material in the thyroid gland with the occurrence of similar material in the other glandular portions of the body, and with those theories of thyroid secretion which regard the colloid as a byproduct rather than as the secretion. Moreover the great variations in the amount of colloid in the thyroid follicles are then explicable upon the basis of variations in the rate of secretory flow which, in turn, is dependent upon the physiological factors of blood and nerve supply as well as upon the anatomical factors. It is also interesting to observe that the volume of secretory epithelium in the several species examined remains in each case approximately proportionate to the size of the thyroid gland and to the size of the fish. The relation of the epithelium to the follicular content and to the blood vessels and lymphatics seems to me to indicate most clearly that the secretion is poured out from the epithelial cells so as to find its waj^, on the one hand directly into


THE ANATOMY OF THE THYROID GLAND 195

the vessels, and on the other hand into the folhcular cavity, whence it eventually passes through the follicular wall to reach the vessel. It is in the course of the latter flow that the colloid appears and its volume is dependent upon the rate of flow.

The question arises as to whether or not the colloid may serve for the storage of secreted materials, somewhat after the manner in which the hepatic glycogen may be considered as stored carbohydrate to be delivered as the needs of the economy necessitates ; if so the colloid material should show further evidences of change, at least, under certain conditions. That the colloid does undergo changes is evidenced by the appearance within its otherwise homogeneous mass of such structural alterations as vacuolation, basophile degeneration, and disintegration into granules of greater or less size, changes which are frequently observed (figs. 14, 15 and 18). As to the physiological nature of these changes in colloid, and their possible connection with a storage function I can offer no conclusive proof, but it seems to me quite possible that such a relation exists.

The thyroid follicles are lined by a simple columnar type of epithelium (fig. 11) whose cells show considerable variation in height. In the same gland the epithelium lining certain vesicles measured as much as .010 mm., others only .006 mm. The epithelium of occasional vesicles was even lower, but was possibly open to the criticism of mechanical distortion since the colloid was often crowded against one side instead of lying in the middle of the follicular lumen, even though the tissues had been prepared with the greatest care. Being anxious to avoid any possible distortion of the tissue, I removed nearly all of the glands studied without allowing them to be touched by either instrument or fingers, the knife or scissors being passed through the muscle beneath, and the gland, supported on a thin layer of muscle, dropped bodily into the killing fluid.

As a rule those follicles which were well filled with colloid possessed low epithelium, in those with taller epithelium the reverse was the case. In making this comparison the surface area of the sections of colloid mass was compared with that of the containing follicle. The average height of the epithelium of a number of fol


196 ^ JEREMIAH S. FERGUSON

licles in Mustelus which were well filled with colloid was .077 mm., w^hile in a similar number of follicles which were either devoid of colloid, or nearh^ so, the height of the epithelium averaged .087 mm.

Each epithelial cell possesses a fairly distinct cell-wall. Many cells appear to have a well marked cuticular border which appears to be more highly refractive than the endoplasm, but wherever the colloid lies in contact with the surface of the cell the cuticular border is obscured. I have also noticed that it is less pronounced in the thinner sections so that I am inchned to regard it as an optical diffraction line rather than a true cuticular membrane.. The conformation of the free ends of the epithelial cells tends to confirm this opinion. These cells project slightly into the lumen of the follicle by means of a somewhat convex free border so that the height of a cell is greater in its axis than at its margin. Thus the lower portions of a cell will, in the thicker sections (.010 mm. or more), show through the taller central or axial portions and so account, at least, for a portion of the cuticular appearance.

The exoplasmic membrane is specially distinct in the epithelium of the Elasmobranch thyroid gland. Baber ('81) called attention to this fact, and desciibed it as an intercellular network enveloping the cells and connecting the lumen of the follicle with the surrounding tissue spaces. If the lining epithelium of the follicle be cut parallel to the surface the resulting sections will show the membrane as a distinct mosaic within whose meshes the cells are apparently contained. It appears to me that this mosaic, which is distinct from the intercellular colloid observed by Langendorf (see page 200), is rather to be regarded as a cell membrane than as an intercellular substance, for there are many portions where in thin sections a narrow intercellular space is distinctly apparent and is bounded on either side by the exoplasmic membrane of adjacent cells. Occasionally the cells are separated by wider intervals through which the follicular lumen is placed in direct communication with the surrounding tissue spaces.

There appears to be no distinct basement membrane upon which the follicular epithelium may rest. At intervals the cells are invested with a very small amount of loose connective tissue (fig. 11), but in large part the epithelium rests directly upon the


THE ANATOMY OF THE THYROID GLAND 197

walls of the venous channels and lymphatic vessels (fig. 14). Thus the relation of the epithelium to the vascular lumen is a very intimate one.

The cytoplasm of the chief" cells is relatively clear, but contains a coarsely granular eosinophile reticulum. Some cells appear much more granular than others. In such cells as are filled with colloid, colloid cells," the granular reticulum is entirely obscured (fig. 14, A).

The nuclei of the chief cells are spheroidal, vesicular, and are placed near the base of the cell. From the apices of many of these cells threads of secretion extend to the central colloid mass. The apices of many of the chief cells appear ragged, frayed, and often shrunken, so that the height of the cell is decreased- Such cells present an appearance suggestive of an advanced stage of secretion. Other cells contain granules at the distal ends which are arranged in vertical rows, giving this portion of the cell a somewhat rodded appearance; such cells are usually well filled with granules. Occasionally a similarly ragged and rodded appearance is seen at the base of the cell and it suggests that secretion may also be discharged at that point. Such a possibility is rendered more probable by the absence of basement membrane and the intimate relation to the lymphatics and blood vessels, these cells often resting directly upon the vascular endothelium. Laterally the epithelial cells frequently are separated from one another, leaving considerable spaces or channels through which secretion may find its way from the follicular lumen to the neighboring vessels; such channels are often occupied in part by colloid and in a few cases I have traced the colloid in a continuous line from the intrafollicular mass to the interior of the vasa and venae lymphaticae (fig. 10).

The above observations suggest that secretion may either be discharged from the chief cells into the lumen of the follicle and thence find its way through the follicular wall to the blood and lymphatic vessels, or that it may be discharged from the cells directly into the vessels; this is in harmony with the conditions indicated in the thyroid gland of mammals.

THE AMERICAN JOURNAL OF ANATOMY, VOL. 11, NO. 2.


198


JEREMIAH S. FERGUSON



Fig. 14. A, section of the thyroid gland of Mustelus showing a follicle completely surrounded by "colloid cells." At y a carmine granule, derived from the injection mass, lies in the vena lymphatica. B, section of the thyroid gland of Mustelus showing follicles lined entirely by "chief cells."


THE ANATOMY OF THE THYROID GLAND 199

The colloid cells described by Langendorf ('89) are remarkably distinct in most of the sections of Elasmobranch thyroids and constitute one of the most characteristic features of the thyroid gland of these fishes. The colloid cells are distinctly acidophiie and are easily recognized in specimens stained with hematoxylin and eosin if the eosin is used in dilute solution and allowed to act for one-half hour or more. They present a glistening, highly refractive colloid appearance, which is in marked contrast to the granular chief cells. The colloid cells occasionally occur singly, but are more frequently disposed in groups along one side of the follicle. One such group (fig. 14, A), more extensive than the others, was seen to include fulh^ three-fourths of all the epithelium in its follicle. The groups are often in contact with the central colloid mass, and the colloid within the cell may then appear continuous with that within the follicle. Occasionally a group of such epithelium appears to have been completely engulfed by the colloid mass, the epithelial nuclei then appearing well within the colloid. Such appearances might suggest mechanical distortion, but as the surrounding follicles show no evidences of injury, and, as already stated, the tissues were very carefully prepared, I am more inclined to agree with Bozzi ('95) that these appearances are the result of vital phenomena.

The nuclei of the colloid cells are small and deeply stained, so deeply, in fact, that in the usual preparations they frequently show neither nuclear wall nor karyosomes. Unlike the nuclei of the chief cells thej^ are usually situated near the inner extremity of the cell rather than at its base. The greater the cell is distended with colloid the farther its nucleus is pushed toward the cell's apex; in the most distended cells there was frequently some distortion and even fragmentation of the nucleus. A further continuation of this process would account for at least a portion of the extruded and disintegrating nuclei found within the intrafollicular colloid masses (figs. 11 and 18).

The intrafolhcular colloid closely resembles that of the mammalian thyroid gland. It is stronglj^ acidophiie and is usually homogeneous or very finely granular in appearance. Frequently a minor portion of the mass, e.g., one side, is finely granular while


200 JEREMIAH S. FERGUSON

the major portion is clearly homogeneous. The well-known filaments pass at frequent intervals from the colloid mass to the epithelial surface. Some of these filaments can be traced to the free surface of the epithelial cell while others quite clearly enter the intercellular spaces, where, in tangential sections of the follicle, they form intercellular masses simulating the net-work described by Baber ('81) and interpreted by Langendorf ('89) as the ramification of colloid cells.

Occasionally the colloid mass appears to have been disintegrated into small spherules .007 to .008 mm. in diameter (fig. 15). The size of these spheres is suggestive of the red blood cells of mammals, but the red cells of Elasmobranchs are ovoid and larger. Of the spherules some are distinctly acidophile but many are slightly basophile, none, or very few, are strongly basophile. All the spherules are homogeneous, and I have observed in the more basophile no tendency to chromatolysis such as one might expect to find if the spherules of this type were thought to represent degenerating nuclei of the red cells, nor have I been able to trace stages of transition from the nucleus to the basophile spherule. Since all the blood cells of the species studied are nucleated one could not well infer that the acidophile spherules could represent any stage in the disintegration of red blood cells, for none of these spherules contain even traces of chromatin. On the other hand, both red and white blood cells can occasionally be found within the colloid quite independently of the spherules I have described; in this particular the Elasmobranch thyroid is in accord with the well known structure in other vertebrate orders. The appearance, location, disposition and reactions of the spherules indicate their origin from the solid colloid masses, from which they would appear to be formed by disintegration with progressively increasing basic reaction. That the reverse process occurs, viz., that the spherules may represent intermediate stages in the formation of the colloid masses, is contraindicated by the fact that only very few follicles contain spherules, nor does there appear to be any indication of a tendency of the spherule to fuse. On the other hand, a tendency to further disintegration is quite apparent, and the possibility is suggested that the colloid in this way may be trans


THE ANATOMY OF THE THYROID GLAND


201



Fig. 15. Section of the thyroid gland of Raia showing the colloid within a follicle disintegrated into spheroidal masses of varying size and depth of stain.



flBC


^-fol. thyr.


Fig. 16. Section through the ventral margin of the thyroid gland of Mustelus with sections of the very broad venae lymphaticae of the thyroid sinus simulating an endothelial capsule about the gland.


202 JEREMIAH S. FERGUSON

formed to such a state that it may be secreted through the wall of the follicle, in which case the intrafollicular colloid would presumably assume the nature of stored secretion which is first poured out from the cells as a fluid, is then condensed through retention within the folhcle, to form colloid, and is later disintegrated, passing out of the follicle with the secretory flow. The chemical analyses of the thyroid gland, showing the common relation of iodin to the colloid and to the active principle of thyroid secretion would hannonize with such an hypothesis.

Vacuolation of the colloid mass (fig. 14) is of frequent occurrence; it may result either from the inclusion within the intrafollicular colloid of portions of the peripheral cup-like impressions which Langendorf has surmised result from the secretion pouring out fiom the surface of the epithelial cells, or it maj^ be further evidence of disintegration of the colloid with formation within its substance of fluid droplets, rather than of solid spherules. The vacuoles are filled with a clear fluid and occasionally contain basophile granules. A colloid mass may contain many small vacuoles mostly at or near the periphery of the mass and containing few, if any, basophile granules, or it may contain one or more vacuoles of relatively large size which occupy the interior of the mass and may be more or less completely filled with the granules. These granules stain deeply with hematoxylin and similar dyes, and they are either amorphous or somewhat crystalline in form. The origin of the chromatic material within the vacuoles may be from chromatolysis of the nuclei of either the disintegrating follicular epithelium or of blood cells included within the colloid. Evidences of disintegration of epithelium and extrusion of the nuclei as well as of the penetration of the nuclei together with blood cells (fig. 20) into the colloid mass are frequently seen. But the disintegration of such cells and nuclei can scarcely account for the much more numerous vacuoles in which no basophile chromatic substance is found.

The follicles of the thyroid are supported within the meshes of a connective tissue stroma in which the blood-vessels lie. The volume of connective tissue is never great, much less than in the mammalian gland. There is more connective tissue in the gland


THE ANATOMY OF THE THYROID GLAND 203

of Raia than in the other species examined ; in Mustelus and Squalus there is so Httle that one wonders at the relative compactness of the organ. In these Selachii the epithelium rests directly upon the walls of the blood-vessels, and they in turn consist of little else than endothelium and form broad sinuses rather than capillaries or venules.

The gland is inclosed by a very thin connective tissue capsule and its tissue is thus always sharply defined from the surrounding structures. In Mustelus and Squalus, and to some extent in the other species, the broad vascular channels of the thyroid sinus are in direct contact with the capsule, so that in sections the ventral surface of the gland often appears clothed with an endothelial coat derived from these vessels (figs. 16 and 17) ; a similar disposition of the vascular endothelium of collapsed blood-vessels is also occasionally seen on the margins and dorsal surface of the gland.

The blood vessels have been in each case carefully studied by dissection, injection, sections, and transparent total mounts of the gland. Both blood vessels and lymphatics were demonstrated beyond doubt, though lymphatics have not hitherto been observed in these fishes and their existence was denied by Baber ('81). Fig. 9 shows the lymphatics filled with injection mass, lying between the blood channels and the follicular epithelium; they appear as perivascular lymphatics in the wall of the venae lymphaticae. Similar vessels, perivascular lymphatics, are found in the walls of the arteries and veins of the thyroid, the thyroid sinus^ and the arteries of the hypo-branchial system (fig. 8).

The course of the larger blood-vessels was readily followed in injected specimens in which the whole gland was examined under the microscope. The arteries course upon the surface of the gland, the major portion of them being always on the ventral surface. Fig. 19 shows the distribution of the arteries in the thyroid of Mustelus, and figs. 3 and 4 indicate the relative area of the gland supplied by the arteries of the right and left side, the left thyroid artery, as in figs. 3 and 19 usually supplying the greater part of the organ, though occasionally the major part, as in fig. 4, is supplied from the right side. Twigs from the superfi3ial branches here and there penetrate the gland, break into arterioles, and


204


JEREMIAH S. FERGUSON


ITBC^



Fig. 17. Section through the ventral margin of the thyroid gland of Squalus showing peripheral venae and vasa lymphatica.


rh


- -<^ol. thyr.


, Fig. 18. Section of the thyroid gland of Carcharias showing the disintegrating ^nuclei of leucocytes or red blood cells within the intrafollicular colloid.


THE ANATOMY OF THE THYROID GLAND 205

promptly empty into groups of broad interfollicular, blood capillaries, the venae lymphaticae (figs. 9, 11, 12), which envelop the follicles on all sides. From the venae lymphaticae the veins pass out of the gland at its posterior and lateral borders and dorsal surface to enter the thyroid sinus ; a few veins from the lateral border of the gland pass directly to the hyoid sinus.

The course of the lymphatics was much .less easily determined than that of the blood vessels. Stick injections," as ordinarily made, spread so rapidly through the loose connective tissue of the gland and so easily entered and filled the venae lymphaticae that they entirely obscured the smaller vasa lymphatica. The venae lymphaticae thus injected form a dense almost opaque mass, showing that the thyroid may well occupy the place in these fishes assigned to it by Tscheuwsky ('03) as the most vascular of mammalian glands. After several futile attempts to inject the lymphatics in the ordinary way the method was so altered as to inject only minute areas under a very low pressure. In this way it was found that at the margins of the injected area the fluid which had entered the vessels traveled farther than that in the connective tissue spaces, and in many cases the vasa lymphatica were filled beyond the limits of the injected venae lymphaticae, so that in the outermost zone of the injected area the true lymphatics could be readily studied, the venae lymphaticae in this zone being either empty or only partially filled.

The vasa lymphatica are, for the most part, perivascular channels (fig. 9), but they are also in direct contact with the epithelial walls of the follicles (fig. 10). The vasa lymphatica could not be followed for any great distance through the injected zone, for, on the one hand, they entered the area of opaque injection mass, and in the other direction they ended abruptly, often with a small knob-like dilatation. By means of serial sections I was able to determine in uninjected specimens that the vasa lymphatica opened at the points of terminal dilatation directly into the venae lymphaticae (fig. 10, X). Having demonstrated the connection between the two sets of vessels in uninjected specimens, showing the true relation of vasa and venae lymphaticae, many points in the injected specimens could be readily found at which it seemed quite certain that the injection mass was


206


JEREMIAH S. FERGUSON



art. thymdextra


art. thyrsinestm

19



Fig. 19. Drawn from a total mount of the thyroid gland of Mustelus canis, showing the course and distribution of the thyroid arteries and the origin of some of the veins of exit. The vessels on or near the ventral surface are indicated by the solid black lines, those on or near the dorsal surface of the gland by dotted lines.

Fig. 20. A section from the thyroid gland of Carcharias, showing invasion of epithelium and colloid by leucocytes.


THE ANATOMY OF THE THYROID GLAND 207

passing from the vasa lymphatica directly into the venae. The relatively intimate relation between the veins and lymphatics in fishes is well known; Wiedersheim ('07) and Favaro ('06) have recently emphasized the fact so far as the tail vessels of fishes were concerned. This intimate relationship seems to be quite as obvious in the thyroid vessels and in those of the region occupied by the thyroid sinus {vide supra). .

That the vasa lymphatica are true lymphatics and not bloodvessels is shown by the fact that in many cases they are of altogether too small caliber to transmit the large red blood-cells of the Elasmobranch fishes. Moreover, the quasi valvular nature of their anastomosis with the veins, as already described, renders highly improbable the regurgitation of blood-cells into the vasa lymphatica even in the larger vessels.

In mammalian thyroid glands, especially in dogs, one now and then observes instances where the colloid has accumulated beyond the bounds of the follicles, giving to sections of the organ the appearance of a tissue completely infiltrated by the waxy colloid substance. No such appearance was found in the Elasmobranch thyroids which were studied, unless the follicles lined chiefly by colloid cells could be so interpreted.

SUMMARY

1. The Elasmobranch thyroid gland closely simulates the human, both in the form and structure of its follicles and the distribution of its blood-vessels.

2. The gland rests upon the basi-hyal cartilage whose anterior margin forms an excellent guide to its location.

3. The pyramidal lobe of mammals is often represented in Elasmobranchs by a process passing forward and reaching the floor of the pharynx through a notch in the anterior margin of the basihyal cartilage; this notch is sometimes converted into a foramen.

4. Baber's opinion that lymphatics are not present in the thyroid of Elasmobranch fishes was founded on insufficient evidence and is incorrect.

5. Lymphatics are present in considerable numbers both in and


208 JEREMIAH S. FERGUSON

about the thyroid gland and can be demonstrated by injection and in sections; they are true "vasa lymphatica."

6. The blood-vessels within the thyroid gland terminate in a network of venae lymphaticae" which invest the follicles, receive the vasa lymphatica, and transmit either or both blood and lymph, under varying conditions of blood-pressure.

7. The thyroid artery arises from the ventral end of the efferent hypobranchial arterial loop contained in the hyoidean hemibranch and the adjacent half of the first holobranch by an independent origin or by a common stem with the mandibular or submental artery.

8. The thyroid veins in these fishes for the most part enter the thyroid sinus," a mass of veins and lymphatic vessels which pour their blood into the hyoid sinuses.

9. The rhythmic respiratory movements of the pharyngeal wall cause the thyroid sinus to act somewhat after the manner of a "vein-heart" or "lymph-heart."

10. The hypobranchial arterial system is formed as described by Hyrtl in 1858 and 1872, and the direction of the flow of its blood is from the gill vessels toward the coronary and other terminal arteries, as indicated by Hyrtl and again by Parker and Davis in 1899, and not from the subclavian artery toward the coronaries, as described by T. J. Parker in 1886. The hypobranchial artery of T. J. Parker, forming an anastomosis between the subclavian and median hypobranchial arteries, is of insignificant importance and is frequently wanting.

11. The relative volume and distribution of the "colloid" in the glands of different species indicates that this substance is a retention product, formed from the albuminous secretion of the follicular epithelium.

12. The further changes occurring in the "colloid" indicate the possibility of its usefulness as a sort of stored-up secretion.

13. The follicular epithelium contains both "chief" and "colloid" cells, the latter being even more numerous and characteristic than in the mammalian thyroid.

14. The parenchymal epithelium is in intimate relation with the vasa and venae l^^mphaticae; it rests directly upon or in close proximity to the endothelial wall of these vessels.


THE ANATOMY OF THE THYROID GLAND


209


EXPLANATION OF FIGURES


Abbreviations


I — V, first to fifth branchial clefts

A. lat., lateral artery

A. thry., thyroid artery

A. thyr. imp., arteria thyreoidea impar

Art., artery

b. hy., basi-hyal cartilage

cap., fibrous capsule of the thyroid gland

ch. ep., chief cells of the follicular epithelium

col., colloid

com., commissural artery

cor., coronary arteries

coraco-mdb., coraco-mandibular artery

coracoid, coracoid artery

DA., dorsal aorta

en., vascular endothelium

ep., epithelium of the thyroid follicles

epibr., epibranchial artery

fol. thyr., follicles of the thyroid gland

gastric, gastric arteries

hy. sn., hyoid sinus

hypobr., hypobranchial artery

hypobr.', its anastomosis with the commissural artery.


L., vasa lymphatica

I. c, lateral commissural arteries

Lai. hypobr., lateral hypobranchial artery

M.c.br., coraco-branchial muscle

M. c. hy., coraco-hyoideus muscle

M. c. vidb., coraco-mandibularis muscle

Med. hypobr., median hypobranchial artery

p. c, pericardial arteries

pharyn., pharyngeal arteries

R B C, red blood cells.

s. a. v., ventral arterial sinus of the first hypobranchial loop.

sbcl., subclavian artery

thyr., thyroid gland

thyr. sn., thyroid sinus

V. vein ■/

V A, ventral aorta

V L, venae lymphaticae W B C, white blood cells

X., point of anastomosis of vasa and

venae lymphaticae y., injected carmine granules in the

venae lymphaticae


210 JEREMIAH S. FERGUSON

BIBLIOGRAPHY

Baber, E. C. 1881 Phil. Trans., Roy. Soc. London, 172, 577.

Balfour, F. M. 1881 Treatise on Comp. Anat., London 2, 626.

Bridge, T. W. 1901 Cambridge Nat. Hist., London, 7.

Cole, F. J. 1905 Anat. Anz., 27, 323.

De Mexjron, p. 1886 Rec. zool. Suisse, 3, 517.

DoHRN, A. 1884 Mitth. a. d. zool. Station z. Neapel, 5, 102.

EcKER tr. WiEDERSHEiM. 1904 Anat. des Frosches, Braunschweig, 205.

Favaro, 1906 Quoted by Eisler, Schwalbe's Jahresb., 12, 323.

GuiARD, J. 1896 These Paris.

Hyrtl, J. 1858 Denks. d. k. Akad. Wissen., Wien, Math. Nat. CI., 15, 1. 1872 Denks. d. k. Akad. Wissen., Wien, Math. Nat. CI., 32, 263.

Lagendorf, O. 1889 Arch. f. Physiol., Suppl. Bd., 219.

Lombard, G. D. 1909 Biol. Bull., 18, 39. Major, R. H. 1909 Am. J. Anat., 9, 475. Maurer, F. 1886 Morph. Jahrb., 11, 129.

1888 Morph. Jahrb., 13, 296. MtJLLER, W. 1871 Jena. Zeitschr., 6, 428.

Parker, G. H. and Davis, F. K. 1899 Proc. Bost. Soc. Nat. Hist., 29, 163. Parker, T. J. 1880 Trans, and Proc. New Zealand Institute, 13, 413.

1886 Phil. Trans., Roy. Soc. London, 177, 685. Peremischko 1867 Zeitschr. f. wis. Zool., 17, 279. Sabin, F. R. 1909 Am. J. Anat., 9, 43. ScHAFFER 1906 Anat. Anz., 28, 65. Simon 1844 Phil. Trans., Roy. Soc. London, pt. 1, 295. Stockard, C. R. 1906 Anat. Anz., 29, 91. TscHEuwsKY 1903 Arch. f. d. ges. Physiol., 97, 210. Turner 1874 J. Anat. and Physiol., 8, 285. Van Bemmelen, J. F. 1887 Zool. Anz., 10, 88.

Wagner, R. 1853 Handworterbuch d. Physiol., Braunschweig, 4, 111. Wibdersheim, R. 1907 Comp. Anat. of the Vertebrates, p. 432.


ON THE MUSCULAR ARCHITECTURE OF THE VENTRICLES OF THE HUMAN HEART

FRANKLIN P. MALL

Professor of Anatomy, Johns Hopkins University, Baltimore, Md.

TWENTY-TWO FIGURES

The present study is to be considered as a continuation of John Bruce MacCallum's, whose ill health prevented him from continuing his work. MacCallum was a brilliant student, a man of marked artistic temperament, whose untimely death has been a very great loss to scientific anatomy. At the beginning of his medical career, while he was yet a student of histology, he made important observations on the histogenesis of the heart muscle cell' which he believed to show that the main growth of the wall of the ventricle takes place immediately under the endocardium, — possibly in the Purkinje fibers. In order to give this question a fuller test, he made a study of the growth of the sartorius muscle," for it was thought that in this simple organ a key to the growth of the muscle walls of the heart might be found. Although the hypothesis regarding Purkinje fibres has proved to be erroneous and although the study of the sartorius has been found to be of little value in the study of the heart muscle, he did succeed in unrolling the wall of the left ventricle into a single sheet or scroll of muscle fibers.^ His presentation of the archi ^ J. B. MacCallum, On the histology and histogenesis of the heart muscle cell. Anatom. Anzeiger, 13, 1897.

2 J. B. MacCallum, On the histogenesis of the striated muscle fiber, and the growth of the human sartorius muscle. Johns Hopkins Hospital Bulletin, 1898.

^ J. B. MacCallum, On the muscular architecture and growth of the ventricles of the heart. Welch Festschrift, Johns Hopkins Hospital Reports, vol. 9, 1900.

THE AMERICAN JOURNAL OF ANATOMY, VOL. 11, NO. 3 MARCH, 1911

211


212 FRANKLIN P. MALL

tecture and growth of the ventricles of th« heart marks a milestone in this study, the like of which is found only in Gerdy's some seventy-five years before. Both Gerdy' and MacCallum studied the heart muscle as a whole and did not deal with it in fragments. MacCallum, as Gerdy, did his work while still a medical student, but unlike him, presented his work in a masterly way. His paper is comprehensive. When we recall that MacCallum unraveled the heart musculature of the foetal pig in the brief period of a week, conceived his illustrations in a second week, and wrote his beautiful paper in a third week, we realize that he was possessed with genius of a very high order. Ill health checked his studies in this direction and his untimely death brought them to an end. However, the problem and his spirit of work have lingered with us, and it was first Knower^ who showed that what MacCallum had found in the foetal pig's heart could be confirmed in the human adult. This has made it possible to round out the work of MacCallum in order to make it of use to anatomists and physiologists. The desire to do this MacCallum had often expressed to me, and I consider it a privilege and a duty to a friend and to our science, to carry out, in a measure at least, a plan which he was compelled to abandon.

The first good analysis of the musculature of the heart was given by Winslow about two hundred years ago and we see some progress in this line of study in Paris through his pupils and successors until the brilliant work of Gerdy published in his doctor's thesis about a century later. The connection of the fiber bundles at the base of the heart, which turn upon themselves at its apex, was well known to Lower,*' and a good description of the arrangement of the external fibers was given by Winslow," and later by C. F. Wolff. « 

^ Gerdy, Recherches, discussions et propositions, etc., These, Paris, 1823. ' Knower, Demonstration of the intorvenlricular muscle bands of the adult human heart. Anatom. Record, 2, 190S.

^ Lower, Tractatus de cordc. London, 1669.

' Winslow, Mcmoires de I'Academie Roy. des Sciences, Paris, 171L

« Wolff, Acta, Acad. Sci. Lnp. Petropal., Vols. 2-10, 1780-92.


MUSCULAE ARCHITECTURE OF THE HUMAN HEART 213

The heart muscle problem comes up anew in the great edition of Hildebrand's anatomy by E. H. Weber nearly a century ago.® He also considered the organ as a whole and did not neglect the study of its function, a distinction which has characterized the work of the Leipzig anatomists and physiologists since his time, as the publications of C. Ludwig,io Krehl,ii His, Sr.,12 ^nd His, Jr.^^ bear witness. All this illustrates that progress in anatomy is most likely to occur when its problems include the study of growth and function, as well as of structure.

Although MacCallum found that he could unroll the foetal pig's heart by macerating it in a modified Krehl's mixture of nitric acid, this is by no means necessary. Hearts that have been boiled in water slightly acidulated with acetic acid are of very great value for the study of the course of the fibers. Unfortunately this method causes the hearts to shrink — puts them into the systolic state — and softens the tendons at their base. For these reasons I boil only two hours or less or until the outer connective tissue and fat can be easily removed without softening the tendons too much. For careful dissection, however, it is well to have the hearts distended and somewhat tougher than the boiled hearts are. Such specimens can be made by fixing either distended or contracted hearts in a 3 per cent solution of carbolic acid. Specimens prepared in this way may be kept in stock, but their dissection is slow and tedious. The outer connective tissue must be stripped off before the muscle bundles may be separated with the forceps and fingers. The disadvantage of the carbolic acid is apparent, but I have found that the acid can be washed out in flowing water in the course of several days. Weak alcohol and other macerating fluids are also of value, but since most of the hearts used came from cadavers which had

9 E. H. Weber, Hildebrand's Handbuch der Anatomie des Menschen. Braunenschweig, 1831, Bd. 3.

19 C. Ludwig, Zeit. fiir rat. Med., Bd. 7, lS-19.

" Krehl, Abhandl. d. K. S. Ges. d. Wiss. Math.-phys. CI., Bd. 17, 1891.

12 His, Sr., Anatomie mensch. Embryonen, Bd. 3, Leipzig, 1885, and Beitragezur Anatomis des mensch. Herzens, Leipzig, 1886.

15 His, Jr., Abhandl. d. K. S. Ges. d. Wiss. math.-phys. Classe Bd. 18, 1891. Arbeiten aus der med. Klinik zu Leipzig, 1894.


214 FRANKLIN P. MALL

been embalmed with carbolic acid, there was little opportunity to use other methods with the human heart. However, carbolic acid specimens may be further prepared nearly as well as fresh hearts by boiling and such secondary treatment was also employed. Either boiled or fresh hearts that are to be dissected subsequently or preserved permanently may be kept perfectly well in a 3 per cent solution of carbolic acid or in formalin.

Before discussing the architecture of the heart musculature it is necessary to define a fiber. It is now well-known that heart muscle cells form a syncytium in which may be found the primitive fibrils of the cells. The bundles in general are parallel in direction with numerous lateral branches which pass out of a single group of cells at very acute angles. On account of the numerous large clefts between the cells, to make room for blood vessels as well as strands of connective tissue, groups of cells can be separated into larger bundles or fasciculi which are clearly recognizable to the naked eye. These are the so-called fiber bundles which are not entirely free but anastomose constantly with adjacent fiber bundles. It thus happens that no bundles are single but they are a portion of a conthmous network which is a repetition on a larger scale of the primitive fibrils seen under the microscope. The fasciculi have a general parallel direction which, however, are constantly shifting in direction as they penetrate the heart wall, so that ultimately the fibers on the outside of the heart lie at right angles to those under the endocardium.

The direction of the fibers is also of physiological significance ; the fibers always shorten and widen in contracting, so that a square centhneter of surface becomes shorter and not wider in the direction of its fibers, but thicker in the direction of the thickness of the heart wall. In the change of shape from diastole to systole the external surface of the heart becomes smaller and the thickness of wall becomes greater. This is all well known.

In stripping off the fibres it is found that successive bundles are constantly passing under one another so that they overlap much as do the shingles of a roof. The outer fibers which arise at the base send small bands into the depth which have a tendency to turn upon themselves to return to the base. This is


MUSCULAR ARCHITECTURE OF THE HUMAN HEART 215

more marked on the right side of the heart than on the left and in front than behind. The most sharply defined parallel fibers are found crossing the posterior longitudinal sulcus.

I shall speak of a group of fibers as a fasciculus and a number of them side by side, sufficiently collected to be pulled off together, as a sheet. I wish to repeat that the fasciculi and sheets are never fully separated from adjacent fasciculi or sheets, but are in constant communication with them. The fasciculus marks the chief direction of the fibrils which can be stripped off with considerable ease, and pass in the direction in which the muscle shortens in contracting.



Schema A. The outer and inner bundles are continuous at the apex as V-shaped loops with very acute angles. Somewhat deeper the angle at the point of turning is a right angle and in the middle layer the bend forms an obtuse angle.

No simple schema can be given which applies equally well to all portions of the heart wall. In general the wall is composed of V-shaped loops lying within one another, the outer forming very acute angles at the heart apex, with one stem of the V on the outside of the heart and the other on the inside (Schema A). Passing towards the middle of the ventricle wall the V-shaped loops do not reach to the apex, and the angle is less acute. Finally as they come to lie in the middle of the wall the V's form quite obtuse angles. This change may be said to be due to the lateral anastomoses of fasciculi becoming greater than the main


216 FRANKLIN P. MALL

bundles. This simple schema becomes distorted because the \'-shaped loops are not limited to a relatively flat portion of the ventricle wall but they encircle the whole ventricle and are also tucked up" into the septum, especially at its anterior border. The fasciculi and sheets which I shall describe are marked by their points of origin, their ending, and especially by their relation to the vortex of the left ventricle as well as by the muscular septum of the ventricles.

There is an agreement among authors regarding the course of the muscular fibers on the external surface of the heart. However, it is well, in studying their course, to prepare several specimens of adult hearts as well as those of children, in order to determine variations in case they exist. This is best done by cleaning the muscle of the ventricles of hearts which have been well fixed in carbolic acid or in alcohol, and by removing the atria entirely. The superficial blcod vessels of the heart should be removed also. By comparing a number of such specimens it will easily be seen that in general the fibers over the right ventricle are in a transverse direction and over the left in a perpendicular direction. All this is clear when it is remembered that the apex belongs entirely to the left ventricle and towards it most of the muscle bundles stream to form the great vortex of the left ventricle which surmounts the apex. Fibers on the right side must cross both anterior and posterior longitudinal sulci to reach the great vortex upon the apex, thus giving these a transverse direction while those on the left side simply stream downward to the apex. The transverse direction of the fibers over the right ventricle is maintained somewhat more on the posterior surface of the heart than on the anterior, because the posterior fibers stream towards the small vortex of the right ventricle which is higher up, while the anterior stream toward the great vortex of the left ventricle which is lower down. What has been said may be seen easily by superficial observation of any heart which has had all of the epicardium removed. '^

1 This is well shown in Mac(' nil urn's dinsram, fig. 15, which is reproducod in Piersol's Anatomy, fig. 664.


MUSCULAR ARCHITECTURE OF THE HUMAN HEART 217

All of the superficial fibers may be described as arising around the tendinous rings at the base of the heart to which the valves are also attached. However, these rings are intimately related to the aorta and pulmonary artery through the membranous septum, which marks the inter-ventricular opening in the embryo. In fact it may be considered to represent much more than this, for it extends upward to include the aorta and pulmonary artery, the tendon between these having been well described by Krehl. The tendinous rings inclose both atrio-ventricular openings and extend to include the membranes which close the openings between the right and left hearts in the embryo. The extent of this tendinous band is well shown in fig.l, in which the interventricular membrane is marked X and its extension to the pulmonary artery X' . It is further seen that the course of the muscle fibers from these tendons is not uniform in all directions showing marked differences in each portion of the heart. The fibers from the left side of the heart. A, pass downward towards the apex, those around the right side, B', B" , are transverse, while those around the pulmonary artery. A' , are circular. Those marked A pass directly to the great vortex forming its posterior horn^^ (fig. 2, A,) while those marked B cross the posterior longitudinal sulcus to the vortex of the right ventricle and finally across the anterior longitudinal sulcus to the anterior horn of the great vortex, (fig. 2, B).

In general, then, the superficial fibers of the wall arise from the tendinous structures at its base and converge toward the apex to form the great vortex of the left ventricle. Those arising from the conus, the left side of the aorta and the left side of the left ring,( fig. 1, A, A' ), pass to the posterior horn of the vortex, and ultimately to the septum, while those arising mostly from the right fibrous ring posteriorly, (fig. 1, B, BO,pass around to the anterior side of the heart to form the anterior horn of the vortex,

1^ In Haller's Physiology, London, 1764, vol. 1, p. 7.5, we find the vortex described as being composed of two horns to correspond with the two main bands of muscle which penetrate the heart here. According to Haller one group is inserted into the septum and the other penetrates the left ventricle and returns in a contrary direction to the base. This is substantially correct.


218 FRANKLIN P. MALL

(fig. 2, B), and ultimately enter the papillary muscles of the left ventricle. It follows, then, that there is one group of superficial fibers of the heart which belongs to the left ventricle and one to the right. Since these two muscle groups have been recognized for a very long time and since I have been able to define them with even greater precision through the horns of the vortex, it is well for the sake of easy description to give them specific names. The bundle from the conus and the root of the aorta, — the aortic bulb, — takes a spiral course to the vortex and then enters the septum. It may be termed the bulbo-spiral. The. other group arises from the venous (sinus) end of the embryonic heart and takes a complementary course. It may be termed the sino-spiral bundle. Each group falls into two chief layers, a superficial and a deep, so we have superficial and deep bulbospiral bands, and superficial and deep sino-spiral bands. When the modifying term is not used, the superficial band, the band to the vortex, is meant.

E. H. Weber^ states expressly that the vortex of the heart is limited exclusively to the left ventricle, while the superficial fibers of the right ventricle do not form a vortex at its apex, but pass into the heart along the anterior longitudinal sulcus as well as at the apex of the right ventricle. They also pass over the posterior longitudinal sulcus to blend with the superficial fibers of the left ventricle. The fibers which pass into the depth are shown in fig. 7, A. A similar figure has been published by C. F. Wolff. 1^ In general this description is correct, but we notice from time to time a vortex of the right ventricle spoken of in the literature. ^^ The examination of a number of well preserved human specimens from which the pericardium and connective tissue have been carefully removed will show definitely that there is a vortex at the tip of the right ventricle as well as at that of the left. This is shown, giving also the course of the main muscle bundles, in fig. 2. This figure has been draWn with care and through its interpretation we learn the course of the main muscle

16 Weber, I.e., vol. 3, p. 146.

•^ Wolff, I.e., 17S0, Fig.3. TliisfigurewascopiedbyLodor,Anatom.TafeIn, Weimar, 1794, Bd. 4, Taf. 114, Fig. 2. 1* For example, Krehl, I.e., page 351.


MUSCULAR ARCHITECTURE OF THE HUMAN HEART


219


A'



B'


Fig. 1 Base of a well developed heart showing the course of the superficial muscle fibers. A, A' {BS), origin of the superficial bulbo-spiral muscle; B, B', (SS) origin of the superficial sino-spiral muscle. From X to A around the front of the aorta indicates the coi^ise of t lie aortic septum. TR, posterior triangular field. With the exception of ngs. 13 and 15 the figures arc tiaUiral size.


220 FRANKLIN P. MALL

bundles of the ventricles of the heart. Throughout this paper I shall speak of a vortex of the right ventricle, and the vortex, or vortex of the left ventricle, to designate the vortex of the B N A.

That the main superficial muscle bundles enter the heart at its apex to spread out on the inside of the ventricle has been known to anatomists since the time of Borelli, whose account of the contracting mechanism of the heart muscle is the one mjstudy strives to revive. E. H. Weber has given us the most satisfactory account of the arrangement of the muscle bundles of the ventricle and his description should be studied by all who wish to become familiar with the subject. It is one of the most satisfactory accounts that has yet appeared. i* Weber pointed out clearly that the superficial and inner muscle bundles radiated spirally from the apex towards the base of the heart, but in opposite directions. In viewing the heart as an object it is found that they pass from left to right towards the apex on the outside of the heart, and from right to left on the inside. Between these two layers there is a middle layer which according to Wolff can be broken into an outer sheet with fibers more nearly parallel to the outer layer of muscle bundles and an inner sheet in which the direction of muscle bundles corresponds more nearlj^ with those of the inner layer. So the statement made by Ludwig that any cube of heart wall extending from the pericardium to the endocardium is composed of fibers on the outside which are at right angles to those on the inside, rests upon a sound anatomical basis. When such a block is torn to pieces it is further found that the fibers on one side gradually rotate in position as they are followed to the other.

Weber's studies of the musculature of the heart, especially that of the left ventricle, do not deal with the course of the fibers in the septum in a satisfactory manner, and he expressly states that this portion of the heart must be unraveled before a comprehensive view of the whole system of muscle bundles will be ol)tained. In this region the many studies of Casper Friedrich

'^ 'llic larffc inodiTii text Ijooks with t lie t'X('t'])t ioii of Qiiain's Anatomy ignore the musculature uf the lieart altogether or give but a meager aecount oi' this subje(!t.


MUSCULAR ARCHITECTURE OF THE HUMAN HEART 221


Vortex



AP'


Fig. 2 Apex of the heart to show the two vortices. A, superficial bulbo-spiral bundle forming the posterior horn of the vortex, which then enters the septum; B, superficial sino-spiral bundle forming the anterior horn of the vortex. This bundle encircles the right ventricle, including its vortex, in its course to the anterior horn of the left vortex, SLA and SLB, anterior and posterior longitudinal sulci.



Fifi;. 3 Heart sliown in fi^. 2 broken open from Ix'liind aerording to MacCalhun's mctliod. SS, superficial au(Uleej).sino-spiral sheet (Uit transversely; LRV, the muscle band from the membranous septum to the right ventricle cut transversely. The tendon of the conus, comis ])art of the aortic septum, is shown on I he medial side of the (ronus, X'. On the ])()sterior side of the aortal he a1 rio-vent ricular bundle is seen. TR, posterior triangular field extending downward into the raphe.


MUSCULAR ARCHITECTURE OF THE HUMAN HEART 223

Wolff fail, as well as those of the earlier anatomists. However, it appears that Gerdy's study interpreted this portion of the heart wall in a very satisfactory manner, although his work was not accepted by Weber. Weber states expressly that Gerdy's work was so poorly presented that it was impossible for him to separate the theoretical from the observations in it. Although Weber could not confirm the fleshy fibers which connect the papillary muscles of the right and left ventricles they really do exist, as has been shown by MacCallum and as I have also frequently observed. Furthermore, Gerdy described and pictured nearly correctly the strand of muscle fibers which I have called the superficial bulbo-spiral bundle. 2° This he describes as a bundle which crosses upon itself to form a figure 8 a comparison which has been applied quite differently by later investigators. Subsequently the ventricle wall including the septum was carefully investigated by Ludwig and than by Krehl, whose work comes from Ludwig's laboratory. They constantly kept before their minds the heart as a whole, and the function it has to perform in contracting.

A good account of the middle muscular layer of the left ventricle is given by Krehl who describes it as a cylindrical band of fibers which form loops and do not end at the atrio- ventricular ring. Although this circular muscle was recognized by Weber^^ it is here described anew and is known in the literature as Krehl's Triebwerk. That the fibers of the Triebwerk arise also from the atrio-ventricular ring my own studies show. Below a bundle extends from it to the apex, as was noted by Krehl. ^2

The cylindrical muscular band described by E. H. Weber and by Krehl is nothing but the transverse fibers of the left ventricle repeatedly described by anatomists during the past two centuries. ^^ That this layer of fibers can be shelled out of the wall

2" Although this bundle is not quite correctly given in Gerdy's fig. 12, it is easily seen that he observed it in his specimens.

^1 Weber, I.e., p. 148. AlsoRied, Todd's Cyclopaedia of Anatomy and Physiology, London, 1836, vol. 2, page 592.

22 Krehl, I.e., p. 348.

2' These are the spiral fibers of the middle layer according to Borelli and Lower. Haller, I.e., states that the transverse fibers of the middle layer go to form the septum.


224 FRANKLIN P. MALL

of the left ventricle as a basket, or better as a cylinder, for it is open at both ends, was first emphasized by Krehl, and just this point is what has caused so much difficulty in our laboratory, for according to MacCallum the cylinder easily resolves itself into a single sheet or scroll in the foetal pig as well as in the adult heart. To explain this apparent contradiction is one of the objects of this communication. Before taking up this main point it will be necessary to consider briefly the arrangement of the muscle fibers at the base of the heart, at the apex, the membranous septum and, incidentally, the atrio-ventricular bundle.

Fig. 2 is given to show the arrangement of the superficial bundles at the apex of the heart, which shows a vortex under each ventricle. A satisfactory analysis of the vortex of the left ventricle, showing its two horns, is given by Pettigrew.^^ He showed that it is easy to separate the two bundles of muscle entering the apex, into anterior and posterior horns, as Haller did, and judging by his illustration, one passes to the septum and the other to the interior of the left ventricle. This I have been able to confirm fully in numerous specimens of human hearts. It is not so easily confirmed in the pig's heart.

An excellent illustration of the arrangement of the muscle fibers at the base of the heart is given by Bonamy, Broca and Beau.-^ My fig. 1 differs from theirs inasmuch as it includes the tendinous connection between the pulmonary artery and the aorta with the fibers arising from it. Also on the posterior side I show the fibers coming out of the septum and passing under those that arise from the left fibrous ring ; these two sheets are pictured as a single sheet in their figure.

Since the chief muscle bundles of the heart are connected either directly or indirectly with the fibrous bands at its base it is necessary to have a clear understanding of them. These include the membranous septum of the ventricles, which is continued into the aortic septum. E. H. Weber made it clear that the architecture

" Pettigrew, Phil. Trans. London, 1864. This figure is copied in Quain'.s Anatomy, Tenth Edition, 1892, vol. 2, Fig. 322.

" Bonamy, Broca and Beau, Atlas d' Anatomic descriptive, Paris, gives an excellent illustration of the apex of the human heart. Toldt (Atlas, Berlin und Wien, 1901, Fig. 931) also gives a satisfactory figure.


MUSCULAR ARCHITECTURE OF THE HUMAN HEART 225

of the musculature of the heart would not be fully understood until the muscle bundles of the septum are included in the problem. Since Weber's time the problem has shifted from the study of the muscular septem to that of the membraneous septum. The English anatomists have been familiar with the membranous septum for some time but it was not known to anatomists generally until it was rediscovered in Hyrtl's laboratory.-" This membrane marks the place at which abnormal openings between the two ventricles are most likely to occur, that is, persistence of the interventricular foramen of the embryo. However, for a proper understanding of the membranous septum it must be extended to include the aortic septum.

That the muscle bundles of the conus form relatively simple riugs which attach themselves to the root of the aorta has been long known. In fact, they arise and end in a raphe or tendon at the point of juncture of the pulmonary artery and the aorta, as is well shown in figs. 3, 8 and 13." When the conus is carefully separated from the aorta in a heart which has been boiled for several hours in dilute acetic acid or in a heart (preserved in carbolic acid) from which the connective tissue and fat have been removed, it is found that these two vessels are firmly blended along this tendinous line which when followed enters the membranous septum. By replacing the right ventricle in figs. 3 and 8 it is easy to see the connection between the tendon of the conus and the membranous septum. It takes but little imagination to realize that these two structures are derived from the aortic septum of the embiyo as has been clearly pointed out by His.^^ Attention may be called at this point to the muscle bundles which end in the membranous septum in figs. 3 and 8. Undoubtedly they belong to the atrio-ventricular bundle which passes through the membranous septum, for in early stages of development this bundle passes through the interventricular foramen and becomes

^* See Hope, A treatise on diseases of the heart, 1849, p. 302; Huschka, Wiener med. Wochenschrift, 1855; Reinhard, Virchow's Archiv, Bd. 12, 1857; and Virchow, Ibid, Bd. 13, 1858.

" Krehl, I.e., Fig. 1. He has also given a good description of this tendon as the tendon of the conus. See also MacCallum.

-* His, Beitrage zur Anatomie des mensch. Herzens, Leipzig, 1886, p. 8.


226 FRANKLIN P. MALL

tied up in the membranous septum as the aortic septum closes the foramen.

A clear understanding of this region of the heart was not obtained until the subject was investigated embryologically. Then the relation of the tendon of the conus became clear, and the membranous septum received a new meaning, in the study of anomalies in the understanding of the atrio-ventricula bundle, as well as of the muscle bundles of the ventricles in general. This septum is in the center of the heart to which important structures pass; the aorta is firmly tied to it and the contracting muscle of the ventricle acts towards it.

A brief review of the history of our knowledge of the development of the membranous septum is given by His.-^ In order to explain an anomalous heart Lindes-^^ studied the development of the chick's heart and discovered that in its separation into right and left hearts the single heart tube was divided by three independent septa, namely the septum of the atrium, the septum of the ventricle, and the septum of the aorta. Subsequently these united as is now well known. The work of Lindes was extended by His in a study of the human heart and he found that the aortic septum grows from above downward to reach the septum of the ventricle and finally closes the interventricular foramen through the formation of membranous septum. As Lindes pointed out correctly, the aorta arises in the embryo from the right ventricle and through the formation of the membranous septum its communication with the right side is cut off. All this may be seen easily in fig. 3 when it is recalled that the points marked X, X.' were in apposition. The figures of His are needed to make this point clear.31

According to His,^- the heart muscle, which extends over the bulb of the aorta in the embryo, must degenerate in part for it does not extend correspondingly high in the adult. He divides the septum aorticum into three parts: (1) The inter-arterial region or

=» His, Anatomie mensch. Embryonen, Pt. 3, 1885, S. 178.

'" Lindes, Inaug. Diss., Dorpat, 1865.

" His, Anat. mensch. Embryonen, Figs. 101, sa; 106, sa; 111, sa; and 118, sm.

  • 2 His, Beitriige, etc.


MUSCULAR ARCHITECTURE OF THE HUMAN HEART 227

the septum aorticum superius; it consists chiefly of two layers of elastic plates with connective tissue between them. (2) A region between the aorta and the right ventricle or the septum aorticum inferius; it consists of the elastic wall of the aorta, a thin layer of the muscle of the conus and a layer of connective tissue. (3) The region between the ventricles or the septum membranaceum.^^

Before considering the arrangement of the deeper layers of the muscle bundles of the left ventricle it is necessary to describe briefly a few of the peculiarities of the superficial bundles. These, as has been stated, arise at the base and converge spirally toward the apex; those over the right ventricle are in a more transverse direction, while those over the left ventricle are more perpendicular This is well shown in figs. 5 and 6, and has been well illustrated by Wolff in his various papers. ^^ But in addition to the spiral fibers which cross the anterior longitudinal sulcus quite transversely there is often seen in the posterior longitudinal sulcus a bundle which runs perpendicularly towards the apex and appears to be better marked in the heart of the new-born child than in the adult. This bundle is well pictured in Wolff's paper ^^ as well as in Henle's Anatomy,^^ and as it is constant, it need not be considered a variation but should be included in the description of the superficial bundles of the posterior side of the heart, which here have a downward tendency often forming bundles as they approach the apex."

In addition to the bundle just described a small thin superficial sheet is occasionally seen over the middle of the right ventricle near the base of the heart. This is also better seen in the hearts of

33 Figures illustrating the membranous septum may be found in His, I.e. ; Quain's Anatomy, figs. 310 and 317; Toldt's Atlas, fig. 929; Spalteholz's Atlas, fig. 420; and Piersol's Anatomy, fig. 660.

'^ See also Bourgery, Traite Complet de I'Anatomie de I'Homme, Paris, 1836, Tome 4, PI. 10; Bonamy, Broca et Beau, I.e., Tome 2, PI. 4; Quain's Anatomy, I.e., vol. 2, fig. 320; Toldt's Atlas, I.e., figs. 929 and 930; Henle, I.e., vol. 3, fig. 39; as well as elsewhere.

35 Wolff, I.e., Tome2, Tab. 6.

3« Henle, I.e., fig. 39. Also Toldt, fig. 930.

3' Weber, I.e., p. 145, denies the existence of a fasciculus in the posterior longitudinal sulcus.


228 FRANKLIN P. MALL

young children as is pictured by Henle.^*^ I mention this as a variation in the human heart, but it is certainly constant in the pig's heart and caused MacCallum much difficulty in studying the heart musculature. It may be that more careful examination of this sheet will show that it is present on all hearts, for our present method of cleaning off the epicardium might easily destroy it.

The longitudinal bundle of Wolff lying in the posterior longitudinal sulcus is certainly present in most young hearts and it does not interfere materially with the study of the outer spiral muscle bundles for it lies upon them. However, Weber observed that although superifical bundles entered the septum both through the anterior and the posterior longitudinal sulci, the penetrating fibers were much more marked along the former than the latter. It is quite easy to strip off the superficial bundles over the posterior sulcus but not over the anterior, for here the sperficial bundles enter the septum while behind they pass over it. This arrangement was encountered by MacCallum in dissecting the macerated heart of the foetal pig when he attempted to strip off the superficial bundles with a blunt probe. He then found that it was easy to lift off the superficial layer of muscle bundles over the back of the heart but not over the front. These bundles having been cut, he further found that the deeper fibers all entered the septum which when broken open permitted him to unroll the musculature of the left ventricle as a scroll. Through MacCallum's method of dissection, which Bourgery^^ had almost invented, it is possible to unravel the musculature of both ventricles in a satisfactory manner, that is, the same result is obtained in all specimens.*"

The chief bundles of the superficial fibers all arise from the tendinous rings and membranes at the base of the heart, especially those around the aorta, as may be seen in fig. 1 . After the superficial layers have been removed it is seen that the deeper bundles stream towards the aorta more markedly than the super ssHenle, I.e., fig. 39, B.*

3» Bourgery, I.e., Tome 4, PL 10 b, fig. 5.

° Searle, Todd's Cyclopaedia of Anatomy, vol. 2, describes a similar unrolling and gives excellent illustrations of specimens made in this way in his figs. 278 to 282. The chief circular band he calls the rope, for when unrolled it appears like a twisted rope.


MUSCULAR ARCHITECTURE OF THE HUMAN HEART 229

ficial. The ligaments that attach themselves to the aorta are the aortic septum, the valves as well as two fibrous rings, one of which encircles the right ostium venosum and the other the left. The bicuspid and tricuspid valves are attached partly to the fibrous rings and partly to the septum aorticum. Through them the papillary muscles are attached to the tendinous structure in the base of the heart. These all tend to the tie firmly the muscle bundles to the aorta, towards which they force the blood in contracting, and were it not so the force of contraction might '^ shoot the aorta out of the heart." The two rings encircling the ostia venosa unite over the septum of the ventricle into a single band and continue into the membranous septum. Here they form the posterior fibrous triangle, which is very pronounced in the pig. So it may be seen that the aorta is held in place by three ligaments all of which are extensions of the septum aorticum. They correspond to the three semilunar valves of the aorta and also mark the trigona fibrosa of the B N A.

I shall describe them (1) as the posterior one which is an extension of the membranous septum backward between the venous ostia."*^ It forms the posterior fibrous triangle and to it are attached the medial cusp of the tricuspid valve and the anterior cusp of the bicuspid valve.^2 At the point of attachment of this ligament to the aorta the atrio-ventricular bundle of His perforates the membranous septum to enter the left ventricle. This ligament expands into that portion of the wall of the aorta which lies opposite the posterior semilunar valve.^^ (2) The left ligament^^ marks the fibrous triangle to the left of the aorta opposite the left leaflet of the semilunar valve immediately below the left coronoary artery. The left ligament appears to have nothing to do with the aortic septum. The left fibrous ring to which the bicuspid valve is attached arises from left and the posterior ligaments of the aorta. (3) The right ligament encircles the right side of the aorta opposite the right cusp of the aortic valve below the right coronary artery.

" Poirier and Charpy, Tome 2, fig. 360, nodule droit. « See Spalteholz's Atlas, 1896, fig. 419.

« Note that the B N A as used for this valve, improperly called the right semilunar value by English anatomists. ^■^ Nodule gauche.


230 FRANKLIN P. MALL

It is formed largely by His' septum aorticum superius or Krehl's tendon of the conus. It is shown in fig. 3, X, X', and fig. 8. The muscle bundles of the conus as well as the bulbo-spiral band arise from this ligament. The right ligament which is much better marked in the pig and dog than in man sends a delicate lateral branch around the right venous ostium to form the anterior segment of the right fibrous ring.*^ To sum up, the aorta is tied to the heart muscle by three ligaments, as well as through the valves to the papillary muscles. Two of these ligaments, the right and, the posterior, are derived directly from the membranous and aortic septa, while the third, the left, seems to be independent of them, encircles the heart to the left, and marks the line of separation between the origin of the bulbo-spiral and sino-spiral muscular bands.

The superficial fibers of the heart, all of which arise from the tendinous structures at the base (septum aorticum and its extensions), pass spirally towards the apex of the heart to form the great vortex there. Above they form a thinner layer than below; as they approach the apex the fibers must either diminish in number or they must form a thicker layer. Probably both conditions prevail for fiber bundles are constantly leaving the superficial layer to pass into the depth, but transverse sections as well as dissections show that the main superficial fiber bundles gradually become piled upon one another as the apex is approached until they finallj^ form the whole thickness of the heart wall at this point In other words there are no circular fibers in the apex.

At the vortex all of the fibers penetrate the heart to form the inner layer of the heart muscle of the left ventricle, as is well known. Here they spread upward and finally are attached again at the fibrous bands from which they arise. However, in general a bundle which arises at a given portion of the heart on the outside returns to the opposite side of the ventricle on the inside. Thus, bundles arising from the septum aorticum in front of the heart on the outside, the bulbo-spiral bundle, and posterior to the aorta on

■•^ The tendinous bands just described are well illustrated in Toldt's Atlas, fig. 932.


MUSCULAR ARCHITECTURE OF THE HUMAN HEART


231



l-i


O


M


o



a

  • |-^


-|J


0)


U-i


CO

• 1— t


O


U



<A


bO


-i-i


.S




^4-1


13


o


o



_o



m


03



PM


rt



•^


T3



fl


rfi


03


•+J


-Q


73


<D


r-i



c5



a>


3


"o


a







c3


r-*


S-H




S


'S,


>


T


-ti


1



-Q





3



-a


+->


a>



-a


o


-Hi



o




a


bj)

.s



r-<


r^



QJ


.S ^


o


^3



^-1



-^J O


f:


t; 'C


o


O +0


■n

O



+-*


+J +=


o


3 -a,



u bC


o


.1— <


r

"^ n-,



i (U


P^


.£3


S


32


-f-

P*^ i:-!-!


c3


G °


s


03 a?


a,


^ ^


rj2


oj rt



"3 S

03 rQ


>>


"c


3 _


o


fl c3



o3 T3


bi)


^ "^


yn


'a.t^



tp bfl

O 9


a


9 « 


o3


^ J3


Tt"\


-tJ


Ul


1-^


^


§ .t^




W


-H O




THE AMERICAN JOURNAL OF ANATOMY, VOL. 11, NO. 3


232 FRANKLIN P. MALL

the inside. The sino-spiral bundle arises on the posterior surface of the heart on the outside forms the anterior horn of the vortex and ends largely on the anterior side of the heart on the inside. The two horns of the vortex are constant in human hearts and through them it is possible to divide all of the superficial fibers into two distinct groups, for all of the fibers stream either to one horn or to the other. By cutting the superficial fibers which cross the posterior longitudinal sulcus it is seen that the right ventricle is easily broken from the left. It is clear, by comparing figs. 2 and 3, that all of the fibers that encircle the right ventricle help to make its vortex and then enter the anterior horn of the vortex of the left ventricle. ^^ In separating the right heart from the left it is well to keep the break as near to the right ventricle as possible and in so doing it is soon observed that a large bundle crosses the septum passing from the aorta to the median wall of the right ventricle. This bundle is shown in fig. S, L R V, and is the one described and pictured by MacCallum as a band of muscle from the right ventricle ending in the left atrio-ventricular ring." According to MacCallum it must be cut in order to unroll completely the left ventricular wall of the foetal pig. This bundle has also been observed by MacCallum in the heart of a child and repeatedly by Knower '*^ in the adult heart. No doubt it is this bundle that Senac^ described a century and a half ago and which has been identified repeatedly since. According to E. H. Weber"** the medial wall of the right ventricle is thinner than that of the left ventricle at the same point. The fibers are parallel to the long axis of the heart, and immediately below them are the circular fibers of the

  • ^ Although this point is brought out by MacCallum's method of dissection,

MacCallum did not demonstrate it because he studied the hearts of foetal pigs which had been well macerated in a nitric acid and glycerine solution. This treatment injures the delicate arrangement of the muscle bundles and only the coarser bands remain. Moreover in the pig's heart the muscle bundles at the apex are more intimately blended than in man, which explains why my description of the apex of the human heart docs not correspond with MacCallum's. See MacCallum's diagrams.

^' Knower, Anatom. Record, vol. 2, p. 204.

  • ^ Senac, Tratio du Coeur, Paris, 1849.

" Weber, I.e., p. 150.


MUSCULAE ARCHITECTURE OF THE HUMAN HEART


233


B5


D C


Fig. 5 Right side of a large hypertrophied heart. Figs. 5 to 12 are from the same specimen. BS, superficial bulbo-spiral band; SS, superficial sino-spiral band. Corresponding bundles in the figures of this heart are marked A, B, C, D, and E, respectively. For the sake of clearness bundles A, B, and C, are drawn with a smooth surface.


234 FRANKLIN P. MALL

left ventricle.^" It is this bundle which Ludwig '^ describes as arising from the border between the septum and aorta and from the free border of the papillary muscle of the right ventricle. It continues forward and downward to the anterior surface of the apex of the heart where it enters the septum along the anterior longitudinal sulcus by the way of the vortex. Probably the best description is by MacCallum, who isolated this bundle by his method of dissection. It is clearly shown in fig. 3 and its distribution over the inner wall of the right ventricle is shown in fig. 15. Here it is seen to blend with the sino-spiral bundle in the anterior horn of the vortex and is ultimately lost in the papillary muscles of the left ventricle. It therefore belongs to the interventricular bands of Gerdy and MacCallum.

By cutting this bundle the septum may be split completely as fig. 3 shows. The conus is next separated from the aorta and its tendon, to demonstrate the septum aorticum. From here superficial fibers encircle the heart; these are indicated by the cut surface in fig. 4, A. In fig. 3 the fibers of the horns of the vortex have been separated which can easily be done with the fingers or with the handle of a scalpel. Those from the anterior horn of the vortex have an inward tendency, while those from the posterior horn course towards the septum and come to cross the others at right angles as shown in fig. 4. In separating the two horns of the vortex it is found that the cavity of the left ventricle is soon reached and in order to reach it most easily it is well to have a probe passed into the heart to its apex where it can easily be felt from the outside. At this point the walls of the left ventricle are thinnest, in proportion to the diameter of the lumen as Weber has pointed

OUt.^2

So far it is quite easy to unroll human hearts with constant results, but when it comes to a study of the deeper layers of the

  • " This bundle is shown in Gerdy's diagram, fig. 12.

51 Ludwig, I.e., p. 199.

5- Weber, I.e., p. 144, says that this thin area is not remarkable because in different animals the heart thickness is in proportion to the diameter of the lumen, so as the lumen varies in different portions of the heart the wall varies correspondingly. Now we say "functional adaptation."


MUSCULAR ARCHITECTURE OF THE HUMAN HEART


235



Fig. 6 Anterior surface of the heart, fig. 5, to show the arrangement of the fibers over the right ventricle. BS, bulbo-spiral band. X, the line across bundle C shows the position of the cut surface X in fig. 8.


236 FRANKLIN P. MALL

left ventricle it is not so easy to convince one's self regarding a constant arrangement. Furthermore, it is difficult to describe the arrangement of the muscle bundles clearly, which in turn makes the literature difficult to understand.

The immediate connections of the two horns of the vortex of the left ventricle can be given with precision and in order to make them clear I have had a number of figures (5 to 12), drawn from the same heart. The views given in figs. 2, 3 and 4 have been omitted but in general they correspond so closely with them that it, is unnecessary. In figs. 5, 6 and 7 the superficial muscle bundles which form the superficial bulbo-spiral band which enters the apex of the heart through the posterior horn of the vortex are drawn with a smooth surface and the several bundles in different drawings are marked with the same letters, A , B and C, in order that it is possible to identify them in the different figures. The most posterior sheet, that which arises from the tendinous ring around the left venous ostium, is marked A, that which arises from the side of the aorta, B, and that which arises from the septum aorticum (or the conus) C. Those marked D and E belong to the superficial sino-spiral band and enter the anterior horn of the vortex. It is seen by comparing figs. 2 and 7 that the apex is formed entirely by the anterior horn of the vortex and that the posterior horn encircles the apex. This is shown in an exaggerated way in MacCallum's account of the foetal pig's heart. ^'^ The locking of the two horns of the vortex is shown at C and D in fig. 7. Here they are blended completely but they may easily be separated by running a probe into the left ventricle. Specimens made in this way show a general tendency of all of the fibers of one horn to run at right angles to those of the other. The bundle marked C arises from the septum aorticum superior •'^■* which is much more pronounced in the dog and pig than in man. The line drawn on this bundle in fig. 6, X, shows the position of the cut surface, X in fig. 8.

  • ^ MacCallum, I.e., fig. 16. Also in Piersol's Anatomy, fig. 665 and in MorrisMcMurrich's Anatomy, fig. 383. Most of MacCallum's figures are reproduced in

the latter work.

^* Tendon of the conus, MacCallum; tendinous band between the pulmonary artery and aorta, Krehl.


MUSCULAK ARCHITECTURE OF THE HUMAN HEART 237


/CLV



Fig. 7 Posterior view, somewhat to the left, after the superficial sino-spiral band has been removed to the posterior longitudinal sulcus. The deep bulbospiral band, BS', enters the septum, above the superficial bulbo-spiral band, BS; CLV, circular muscle band of the left venous ostium; TR, posterior triangular field; IV, interventricular bands.


238 FRANKLIN P. MALL

The remainder of the surface fibers, the sino-spiral band, arises from the posterior borders of the venous ostia, converge towards the apex to form the anterior horn of the vortex. In general, the fibers that arise from the arterial side of the heart, the bulbus of the embryo, pass to the posterior horn of the vortex, while those from the venous side of the heart form the anterior horn. It may be that this arrangement of the fibers is due to the bending of the heart in development ; His thinks that it may be associated with the formation of the inferior septum. At any rate since all of the fibers that arise on one side of the heart enter the apex at the other, and vice versa, their course must be spirally around the heart which proves to be the case.

To follow the bundles which arise from the septum aorticum around the heart through the posterior limb of the vortex into the apex and through the septum to their ends is not altogether easy. This has been the great difficulty in the study of the architecture of the musculature of the heart, and it will not be solved to the satisfaction of all anatomists until the development of the bundles is known. At present it is difficult to rest what I have found upon an embryological basis for the main part of the story is run through by the time the embryo is 12 to 15 mm. long, and such hearts can not be dissected. Nor can the course of the muscle fibers be followed with any degree of certainty through the study of serial sections.'^

It is apparent that the bulbo-spiral was well known to Gerdy who attempted to picture it.'^^ Gerdy also observed that the walls of both ventricles are formed by a series of loops which are attached to the tendinous rings at the base of the heart; they become larger as they extend towards the apex of the heart, and successively fall into one another. The loop which reaches to the apex, the bulbo-spiral, forms a double loop in passing through the septum, and thereby is converted from a single loop into an 8-like figure. Although Gerdy's illustrations are crude and his description is

^^ His says, I.e., p. 177, that the fibers must develop from left to right along the anterior longitudinal sulcus and from right to left along the posterior longitudinal sulcus.

«• Gerdy, I.e., fig. 12.



0^


I

%^ CD r


OQ O .^ 03

1^

.S 03

o a

Xi 03

-^ .2

03 S 03

^ .2

=^ a

^ 03 03 >

'S ^j

-^:^


cq


o

03


ft


oj


C3 03

■73 TS t|_.

'> '^ o

>> 4) C3


03 -^


03


a

03 0)


bJO Pi o3


00


be O

bO P2 03

S ft "ra

'^ 03 O

ft ft


240 FRANKLIN P. MALL

by no means clear, it is evident, that he recognized the main groups of muscle bundles of the heart, for he studied the muscle bundles of the whole heart, including those of the septum.

The long muscle bundles around the left ventricle, those forming the figure 8 according to Gerdy, were better described and illustrated by Ludwig in his well-known study. However, Ludwig's figure 8 is not open above like Gerdy's, who connected one limb of the double loop with the tendinous structures at the base of the left ventricle and the other to the base of the right. According to Ludwig," the loops forming a figure 8 are confined to one ventricle. The course of the fibers which arise externally at the base of the ventricle encircles the apex and returns on the inside of the heart to be attached at the base again. Those that arise deeper encircle the heart midway between the apex and the base and end again at the base; they form a complete figure 8. In general, I find this description of the main muscle bundles of the walls of the ventricles the most nearly correct, although his schema is not accepted by Krehl.^"* In Krehl's publication a band of muscle is described which is not attached to the tendinous structure at the base of the heart, but instead makes a complete circle and ends in itself. That this cannot be correct is evident from Krehl's own description^ when he says that the direction of the course of the fibers of this layer is practically the same as the rest of the fibers of the wall of the ventricle. In the circular layers (Triebwerk) the fibers on the outside run from above downward from left to right and on the inside in the opposite direction. While many may see in Krehl's description a contradiction to Ludwig's original description, I see in it only a confirmation. Where Krehl erred was in saying that the fibers in his circular layer do not arise nor end in tendons as do the outer and inner muscle layers of the heart. Since MacCallum succeeded in unrolling the wall of the left ventricle into a single sheet of muscle showing that it must form a scroll, it has been difficult for us to account for Krehl's statement regarding a middle muscle band in the wall of the left

" Ludwig, I.e., fig. 8.

68 The work of Krehl was done in Ludwig's laboratory and received his approval.

"9 Krehl, I.e., pp. 347, 349, and figs. 9, 10.


MUSCULAR ARCHITECTURE OF THE HUMAN HEART


241


^-LRV



Fig. 9 Deep bulbo-spiral band drawn in the outline of fig. 8. 0, origin of the fibers on the opposite side of the ventricle.


242 FRANKLIN P. MALL

ventricle, which ends in itself. My own dissections of the muscle wall of the left ventricle, as well as MacCallum's, show that all of the muscle bands arise in tendons either at the base of the heart or in the papillary muscles which in turn are attached to the tendon of the base through the valves.

Returning now to the two main muscle bands which form the horns of the vortex, it is found by separating them that the cavity of the left ventricle is opened, as shown in fig. 4. This figure shows that where the two main bundles lock there is a third bundle — the so-called circular fibers of the heart. In order to separate the bundles at this point it is necessary to remove this third layer which is shown in fig. 7, BS\ as arising from the left side of the left ostium and then passes into the septum through the posterior longitudinal sulcus. It forms the deep bulbo-spiral band. In fig. 8 a piece of the deep bulbo-spiral band, BS", is cut out to show the course of the superficial bulbo-spiral band, BS, as it passes through the septum. In studying this figure it is to be remembered that the cut edge of the bundle marked X does not belong to the deep bulbo-spiral bundle but to the superficial bulbo-spiral which passes around the heart and enters at the apex. The position of this cut is shown by the straight line X in fig. 6. It is also to be observed that the deep bulbo-spiral band passes under the longitudinal bundle of the right ventricle (figs. 8 and 9, LRV, which arises from the aortic septum just below the medial cusp of the tricuspid valve.

The superficial layer of the deep bulbo-spiral band is shown in fig. 9. The view is exactly the same as in fig. 8 with part of the origin of the longitudinal bundle, LRV, from the aorta to the right ventricle removed. It is also noticed in fig. 9 that the bundle arises on the opposite side of the aorta marked as a cut end, 0, in the figure (See also fig. 10, BS'). The course of the fibers is also given which reminds us somewhat of Krehl's figures.

In fig. 10 the large flap of muscle which is turned back is the portion of the bulbo-spiral band marked A and B in figs. 5-7. It encircles the apex in fig. 10, enters the septum, fig. 8, and passes below the deep bulbo-spiral band, figs. 8, 10 and 12.



Fig. 10 Same view as fig. 5 with the superficial bulbo-spiral band thrown back and the deep bulbospiral band removed entirely. The origin of these two bundles is shown, BS, and BS', at the base of the heart; C, a strand of the bulbo-spiral band not turned back; LRV, longitudinal bundle of the right ventricle. The course of the bulbo-spiral band is marked, BS; it ends at X.


244 FRANKLIN P. MALL

The deep bulbo-spiral band having been removed the continuation of the superficial band below is revealed. This now encircles the heart and ends below and behind the aorta in the aortic septum as shown in figs. 8, 10, 12, X. Fig. 11 shows the deep bulbospiral band drawn upon the outline of fig. 10.'^" The older authors recognized this cylinder of circular fibers around which the superficial and deep fibers are wrapped to form the vortex. This arrangement may be seen by superposing figs. 10 and 11 as well as in fig. 12. Krehl also recognized that the deep bulbo-spiral band was intimatel}^ connected with fibers from the apex, for he states expressly that not only fibers from the outer and inner layers but also numerous fibers from the circular layer pass to the apex of the heart. His description of the circular bands is by no means consistent for he describes it as a cylinder, as a cylinder with fibers going to the apex, and pictures it as a basket entirely closed below. These variations are due entirely to the thickness he gives to this layer.

Coming back now to the superficial bulbo-spiral band, it is seen that it enters the apex and passes below the deep bulbo-spiral band, blends with it to encircle the left ventricle and with a general upward tendency ends in the septum aorticum below and behind the aorta. That is these fibers make nearly a double circle around the heart to form the figure 8 of Gerdy and Ludwig. The lower loop of the 8 encircles the apex and the upper loop lies within the deep bulbo-spiral band. It is also clear that my figure 8 is not closed above, just as Gerdy described it, and that its two free ends are attached to the septum aorticum on the two sides of the aorta.

In my description I have included with the deep bulbo-spiral band those circular fibers which pass through the septum as a single bundle, fig. 12, BS', between the superficial bulbo-spiral bundle BS, the sino-spiral bundle, SS, and the longitudinal bundle of the right ventricle, LRV. Two views of the deep bulbo-spiral band are shown in figs. 9 and 11. It is noticed at once that the course of the

•' This cylinder of fibers was well known to Winslow, Wolff and Weber (Weber, I.e., pp. 151, 152) and had recently been described anew by Krehl.


MUSCULAR ARCHITECTURE OF THE HUMAN HEART


245



Fig. 11 The deep bulbo-spiral band which was removed from the specimen shown in fig, 10. The view is the same. 0, origin of the fibers which are shown out in fig. 10, BS'.


246 FRANKLIN P. MALL

fibers is not circular but in general inclines externally towards the superficial fibers of the heart and internally towards those under the endocardium. ^1 Within the septum they have a transverse direction on the outside and within they have a general upward tendency as shown in figs. 9 and 11. It is further observed that the fibers of the deep bulbo-spiral band arise at the base of the heart immediately below the origin of the superficial band which passes from the same point to the vortex. This is shown in fig. 10, BS, and BS', which gives this band as an extension of the longitudinal band of the right ventricle, LRV. The entire course of the deep bulbo-spiral band is shown in fig. 11. The band as a whole shows the fibers turning upon themselves on the apical side, then becoming circular and blending with the superficial bulbospiral band as it comes up from the septum as shown in figs. 10 and 11. The point of separation between these two bands marks the place w^here the muscle fibers of the left ventricle are most nearly circular. Within the septum, fig. 12, X, this is not the case, as here the end of the superficial bulbo-spiral band passes toward the root of the aorta to which it is attached.

It is quite easy to see that Krehl's Triebwerk, as pictured in his figs. 9 and 10, includes not only the deep bulbo-spiral band but also many of the fibers of the superficial which enter the heart through its apex. His figures show that the fibers arise evenly all around the tendon of the venous ostium, a condition which I can not verify. They arise only on the left side of the aorta and the ostium venosum. fig. 10, and as they pass the septum are entirely free from tendinous connections with the base of the heart as fig. 11 shows. In fact this separation from the base is well marked in the embryo and can easily be seen in serial sections.

As figs. 8 and 9 show, the circular bands form quite a prominence below the opening of the aorta just above the septum. This prominence is much better marked in the heart of the pig and ox, so much so that it is very noticeable below the right semilunar

  • ' The gradual change of the course of the fibers in passing through the heart

wall is from longitudinal to transverse, then to longitudinal again. This is well shown in Gerdy's diagram which may be seen in Todd's C'ycloi)aedia, vol. 2, fig. 271. See also my schema A.


MUSCULAR ARCHITECTURE OF THE HUMAN HEART 247

valve of the aorta. In fact the floor of this valve is fleshy, unlike that of the human.

It is also noticed in the pig's heart that many of the circular bands do not pass through the fleshy septum ])ut are attached to the posterior ligament of the aorta and therefore do not make a complete circle. This bundle is shown cut off under the thumb of the left hand in fig. 19. In fact, there is a kind of raphe at this point in the pig's heart which reaches through the heart wall to the base of the posterior papillary muscle, and in a measure forms a break in the main circular bands. This is noted to call especial attention to a similar but much smaller group of muscle bundles around the left ostium in the human heart. Its attachment is to the posterior ligament of the aorta as shown in figs. 3, TR, and 8, TR. The circular bundles extending partly around the left ostium have been well pictured by MacCallum for the pig where they are very pronounced.^ The extent of the deep Indlw-spiral l)and varies very much iji different hearts. In the new-born and in young children this band is very insignificant which indicates that during growth it must enlarge faster than the other heart muscle l)undles. It also varies in size in the adult heart. Figs. 5 to 12 are taken from an hypertrophied heart which shows the circular bands markedly thickened. On the other hand in a dilated heart with thin walls it is barely present as fig. 13 shows. To show further variations I add an illustration of a well-developed small heart in fig. 14. Here the deep bulbo-spiral band is unusually well-developed, in fact as well as in the hypertrophied heart shown in figs. 5 to 12.

To show the course of the chief nmscle strands of the ventricles of the heart, Schema B. is given. The bulbo-spiral bands are in red and the sino-spiral in blue. The superficial l)ulbo-spiral band roaches to the apex of the heart and there enters the septiun; a layer somewhat deeper encircles the apex, while the middle layer, the doej) bulbo-spiral, encircles the ventricle and ends with

«2 MacCallum, I.e., fig. 20. MacCallum also fiiuis that the tice]) bull)o-si)ir;il band arises not, only from the aortic septum but also from the tendon of the right ostium venosum. Such an extension of the origin of this bundle beyond the aortic septum does not exist either in the pig or in man.

THE AMERICAN JOURNAL OF ANATOMY, VOL. 11, NO. 3.


248


FRANKLIN P. MALL



Schema B. The chief muscle buudlcb which have been described arc ri von. Tlie bulbo-spiral group of fibers are in red and the yino-spiral in blue. The bundles immediately around the left ostium and tlie conns form single loops which attacli themselves to the aortic septum. All other bundles may be considered modihciations of these two simple loops.


MUSCULAR ARCHITECTURE OF THE HUMAN HEART 249


-CO



Fig. 12 Dissection of the septum viewed from the right side. CO, conus; BS and BS', superficial and deep bulbo-spiral bands which end at X; LRV, longitudinal band of the right ventricle; TR, posterior triangular field; *S<S, sino-spiral band passing under fasciculus C of the superficial bulbo-spiral band; D, layer D of fig. 5 which is blended with the interpapillary layer, IV.


THE AMEBICAN JOURNAL OF ANATOMY, VOL. 11, NO. 3


250


FRANKLIN P. MALL



^ 2

i=l ^

•- -a

>^ -s

> ^

cq _2

« 2

-^ "S

, , -(J

!-( O

CO 4)

I -t^

O to

_o O


a; .2

a; . >

O 'in


° - 5

2 :3 d

^ t; o

t^ fl a>

=3 ° S

S -o '^

<P =5 03

bC-2 ^3

  • 5 — I «-i

rt o3 o3

" .J: &

O bO

CO ^

  • "* ^ "^

• 2 "o

f^ ^ t^


MUSCULAR ARCHIIECTURE OF THE HUMAN HEART


251



<s



u





"Sh



m



O



rO





3



^



1— <



a





o



y=l



^




g


s


^


02


• fH



+3


q;


m


^


O


H


CD



S



o


03


PI >


Xi


+3



0)


S3


F-H





lU



r£3



-(J


^ ■


5:



^


cq


r-i



r/2


j3



•+J



.f-(


OJ


03


.S


'O


'co


rt


(B


JS


bC


3


lA



Sh


-a


Ol


fl


t>


o3


03




4J


C4-H


H



o3


-tJ


(D


^


A


03



r-]



,s3



-|j


-^


%-H


OJ



a,




<D



CO


a;


o3


>


^



A


^


-(-»


q;



^


4^


<1




CQ


•*


tc


T-H


03



ft


bC


t3


s


o3



J2


252 FRANKLIN P. MALL

the superficial band on the dorsal side of the aorta. The small circular band ends in the posterior triangular field. The sinospiral bundle encircles the right ventricle and blends with the longitudinal band of the right ventricle to form the anterior horn of the vortex before entering the left ventricle. Here they end partly in the papillary muscles. Other single loops encircle the conus and are attached to the aortic septum, or tendon of the conus.

While most anatomists state that all of the muscle bundles of the left ventricle form loops or V-shaped bundles which are attached to tendons at the base of the heart, this has been denied from time to time. Occasionally I have also found a bundle which seems to turn upon itself at the base of the heart as shown in fig. 14. It is noticed in this figure that a large bundle lies midway between the superficial bulbo-spiral and the deep bulbo-spiral band, passes diagonally through the septum to the base and then turns downward. It then encircles the base, blends with the deep bulbospiral band and ends at the dorsal side of the aorta in its posterior ligament. No such bundle is seen in figs. 8 nor 13. But it is quite easy by comparing figs. 8 and 10 to imagine a portion of the superficial bulbo-spiral band separated during part of its course to form the variation seen in fig. 14. Beside this one example I have not found any bundles which might be considered as forming Vshaped loops pointing away from the apex and towards the base in the left ventricle. But since the loops entering through the apex, as well as the main circular bands, pass nearly twice around the heart before they end, many variations must occur, and by a stretch of the imagination they may be found in every specimen.

The two muscle strands which cover almost the entire exterior of the heart and enter at the vortex are finally distributed over the interior of the left ventricle as well as to the papillary muscles. All ultimately end at the atrio-ventricular ring. A complete separation of these bands under the endocardium as they are separated under the epicardium is hardly possible. However, with some reserve a statement may be made regarding their distribution within the left ventricle.


MUSCULAR ARCHITECTURE OF THE HUMAN HEART


253



254 FRANKLIN P. MALL

The superficial bulbo-spiral band, arising from the conus and entering through the posterior horn of the vortex, passes up the septum and blends with the deep bulbo-spiral band. In some specimens they attach themselves at once to the aorta, as shown in fig. 13. In this specimen the circular bundle around the left venous ostium is unusually poorly developed. In other hearts, as in figs. 8 and 14, practically no attachment to the aorta is made at once but all of the bands encircle the ventricle again and are finally attached to the posterior ligament of the aorta. This is shown in figs. 8 and 10 as well as in fig. 14. From this description it follows that the bulbo-spiral band distributes itself within the left ventricle chiefly on its posterior side. That is, it arises on the anterior side of the heart, externally passes once-and-a-half times around the heart and ends on its posterior side internally.

The sino-spiral bundles, arising at the ostium behind and passing over the right ventricle to enter the left vortex through its anterior horn, are reenforced by a thick band of muscles from the interior of the right ventricle and pass at once to both papillary muscles of the left ventricle which seem to be composed almost entirely by them. Many fibers pass the bases of these muscles and extend upward to end in the anterior side of the fibrous ring of the left venous ostium. In general, they line the anterior side of the left ventricle and also end in both papillary muscles. '^■^

This brings us to the interventricular bands. They were already known to Gerdy whose description of the heart musculature has stood the test of time. Since E. H. Weber denied absolutely the existence of these bundles'^^ they were not taken seriously by anatomists until they were rediscovered and well described by MacCallum.

Fig. 2 shows a strand of muscle fibers, C, passing from the vortex of the right ventricle to the apex of the left ventricle where

  • 8 In the pig MacCallum associated the sino-spiral bundle exclusively with the

anterior papillary muscles and the bulbo-spiral with the posterior papillary muscle. However, this is correct only to a certain extent, for in man the posterior papillary muscle is associated with both spiral bands. This arrangement is necessary in order to unroll the ventricle of the pig into a sheet, as MacCallum did.

" Weber, I.e., p. 153.


MUSCULAR ARCHITECTURE OF THE HUMAN HEART 255



Fig. 16 Pig's heart cut open on its left side and tlien Ijoiled after which the papillary muscles were torn out. A and P, anterior and posterior papillary muscles. A muscle band from the right ventricle passes between them. The bundle X belongs to the anterior papillary muscle.


it enters at the locking point of the two horns of its vortex. If now the strands of the anterior horn of the vortex of the left ventricle are cut as they pass over the posterior longitudinal sulcus, SLP, the interventricular strands are brought to light deep in the septum as is well shown in figs. 3 and 8. By cutting through all the muscle which connects the two ventricles in fig. 8, fig. 12 is obtained. Here it is clearly seen that the interventricular bands, IV, lie nearest the lumen of the ventricle, while the bulbo-spiral bundle, C, and that forming the anterior horn of the vortex, D, lie more superficially. The connection of the interventricular strands is shown in fig. 15; this is from a distended heart which has been macerated and fully dissected. It is seen that the interpapillary bands also extend upward in the right ventricle to end in the membranous septum.

Although the muscle strands are more intimately blended with one another at the apex of the heart of the pig than in man, it has been easier for me to follow the attachments of the papillary


256 FRANKLIN P. MALL

muscles in the former than in the latter. Possibly this has been the case on account of an abundance of pig's hearts. It is seen by unrolling the left ventricle of the pig's heart that the chief circular bands which arise from the bulbus fall into three distinct bands which are easily separated from one another. The first is the single circular band around the left ostium, and the second and third are the spiral bands which enter the septum. The superficial bulbo-spiral band instead of encircling the left ventricle as it does in man divides into two bands (figs. 19 and 20) ; one is attached to the anterior tendon of the aorta, BS, Post, while the other encircles the anterior papillary muscle and is ultimately attached to the posterior ligament of the aorta, BS, Ant, In doing this the anterior superficial bulbo-spiral band, BS, Ant, passes within the walls of the ventricle, between the anterior and the posterior papillary muscles. In order to show this relation better fig. 16 is given. In the specimen from which this figure was made the left ventricle was cut open on its left side before the heart was boiled in dilute acetic acid. Then the papillary muscles were broken apart; the specimen shows clearly that the posterior muscle sends its fibers over the right ventricle, and the fibers from the anterior papillary muscles pass out of the apex of the heart on the medial side of the right ventricle. The strand marked x in fig 16 belongs to the anterior papillary muscle. Between the strands which enter the papillary muscles there is a third which passes from the right ventricle to the left and blends with the anterior bundle of the superficial bulbo-spiral band. By this arrangement it is easily seen that the anterior papillary muscle is intimately connected with the sino-spiral band and the posterior partly with the bulbo-spiral butmostly with the muscle bundles from the right ventricle both on its lateral and medial sides. This arrangement is well shown in figs. 16 and 18. In fig. 18 it is further seen that the roots of the posterior papillary muscle not only reach outside of the heart through the bulbospiral band to the aorta but also on the medial side of the right ventricle through the longitudinal bundles of the right ventricle to the membranous septum. On its posterior side it is intimately attached to the circular bands by the way of a raphe which is not



05 3


o

S3

o

CO

d

a;

s

03


O

a

02 to O O 4)


03


H)


O

a. S '^

03 02

■^ a

o3 (O 03 Qq

O 03

bC 00

^ OJ

o «  o 02

030

=; o3

o —

cS ft

h

O


3


o3 .

a; u -Q 0)

CI

O

•- a>

a

oj


258 FRANKLIN P. MALL

well marked in the human heart, but is indicated in figs. 3 and 8 (tip of TR) and is very pronounced in the pig. At this point the fibers from the posterior papillary muscles are so intimately blended with the deep bulbo-spiral band that often in peeling off this band it breaks at the raphe and leaves the large ckcular band around the left ostium as described by MacCallum and pictured in figs. 19 and 20, BS'. If this bundle, as shown in fig. 20, is pushed downward its cut end will come in contact with the base of the posterior papillary muscle to which it was intimately attached.

What has just been said shows that the papillary muscles are in direct continuation with all of the chief muscle bands of the heart and the attachment of the atrio-ventricular system to them gives meaning to this. An impulse or a wave coming through the bundle of His is at once communicated to the entire musculature of the ventricles.

In studying the musculature of the ventricles of the heart one's attention is directed mostly towards that of the left ventricle, and there is a tendency to neglect the right ventricle because it appears to be of simple construction. On the right side the inner muscle bundles are directed towards the conus as they are towards the aorta on the left side.

In tearing ofif the muscle bundles from any portion of the ventricle it is at once observed that the fibers always tend to pass upwards, towards the base, as they are stripped off. That is, they are constantly passing into the depth. Over the right ventricle this is so marked that when bundle after bundle is lifted up they are found to lie upon one another like the shingles of a roof with the deeper longitudinal muscle bands below serving as rafters (compare figs. 6 and 15).

The superficial muscle bundles, the superficial sino-spiral, come mostly from the left side of the heart, pass over the right ventricle quite freely and then back to the left ventricle. Just below this is a sheet which arises from the posterior part of the left ostium and enters the right ventricle to end there. This, the deep sino-spiral, is the sheet so well described by MacCallum. When these two sheets are examined together it is seen that fiber


MUSCULAR ARCHITECTURE OF THE HUMAN HEART 259



Fig. 18 Outer half of specimen shown in fig. 17 cut off and superposed upon the inner half to show proper relation of the papillary muscles. Letters as in fig. 17.

bundles are constantly passing into the depth whence they pass upward towards the conus. Especially is this true along the anterior longitudinal sulcus, as shown in fig. 6. This observation was first made by Wolff and has been repeatedly confirmed. ^^ It is not difficult to see in this arrangement an extension upward of the vortex of the right ventricle. When these bundles once reach the inside of the right ventricle they all have an upward tendency, towards the conus. On the septal side the inner fibers pass towards the membranous septum, that is towards the ostium. These fibers communicate freely with the papillary muscles of

«* Parchappe, Du Couer, Rouen, 1846, PI. 5, fig. 1.


260 FRANKLIN P. MALL

both ventricles as shown in fig. 15. The fibers from the conus cross the anterior longitudinal sulcus and pass to the apex as the bulbo-spiral band (fig. 6). Below these the fibers encircle the right ventricle (fig. 1) and dip into the anterior longitudinal sulcus as indicated in fig. 6. At the apex of the right ventricle the fibers pierce the septum to pass into the left ventricle as shown in fig. 12.

At the base of the heart the fibers encircle largely the right venous ostium (fig. 1) to reach the conus where they blend intimately with the tendon of the conus (figs. 1, 3, 4, 8 and 13). At this point we also recognize the circular band of the conus as described by Ludwig, and subsequently by Krehl. This MacCallum could not verify in the pig. However it is present but by no means as marked as one would believe by reading Krehl 's paper. It is seen in figs. 6 and 8 that a large loop encircles the conus which is broken by a tendon, the superior aortic septum. However, careful dissections of this region in man and in the pig reveal a small circular band of fibers which encircles the conus and is attached to the aortic septum only. So, as the aorta has its own simple circular band which includes the left venous ostium, there is a corresponding band in the right ventricle which includes the conus only. This arrangement is fundamental and may be described as the figure 8 of the base, one loop around the conus and the other around the aorta and the left ostium, with the cross piece as the septum aorticum. All the other loops of muscle bands may be considered as modifications of the two loops of the transverse figure 8, but they must all come back to the cross piece of the 8, or aortic septum. One loop encircles the main part of the bulbus and the other the main part of the sinus. (See Schema B).

The longitudinal bundles of the medial side of the right ventricle connects with the papillary muscles of the left ventricle and with the membranous septum on the right side of the heart (figs. 3, 8 and 15). A portion of it passes over the front of the heart to blend with the superficial bulbo-spiral band. This is shown in fig. 13 and in fig. 15 where its attachment to the membranous septum has been cut off, and in figs. 6 and 8 with the


MUSCULAR ARCHITECTURE OF THE HUMAN HEART 261

corresponding points marked by an X. In fact this bundle arises in common with the superficial bulbo-spiral band from the membranous septum and could be well considered with it (fig. 10, LRV and BS). If described with it it would be said to extend on the outside of the heart to the apex and thence through the septum, and on the inside of the right ventricle over the septum, finally ending in the papillary muscles of the left ventricle. It also ends in the large papillary muscles of the right ventricle.

MacCallum found that he could unroll foetal pigs' hearts which had been macerated in a solution of glycerine, water and nitric acid, into a single sheet or scroll of fibers. He was also able to unroll a number of foetal hearts as well as the heart of a child three or four years old. Subsequently Knower showed that hearts from cadavers which had been embalmed with carbolic acid could be unrolled by MacCallum's method with considerable ease. Since, however, my aim now is to interpret MacCallum's scroll, I shall use the adult pig's heart in my descriptions. I have extended his work by leaving the tendons at the base of the heart intact in order to show better the attachment of the fiber bundles at this point. The specimens were prepared by distending them with a three per cent solution of carbolic acid in water as already described. ^^ They were then dissected, as shown in figs. 3 and 4. In doing so it is well to let the split be natural and not forced either to one side or to the other.^^ Soon the longitudinal muscle band from the membranous septum to the right ventricle comes into view and after it has been well isolated it is to be cut squarely. Next the aorta is to be torn from its root which takes with it the muscle bundles, superficial and deep bulbo-spiral, arising at this point. This may be understood by a glance at figs. 10 and 11. The splitting

'^ By this method hearts may be prepared in either the dilated or the contracted form. In order to make specimens rapidly and also quite satisfactorily hearts may be unrolled after boiling them in dilute acetic acid for half an hour. The specimen can then be cleaned easily, the muscle is not shrunken and the tendons at the base are still intact.

" Searle, Todd's Cyclopaedia of Anatomy, 1836, vol. 2, evidently had this sheet before him when he described the "rope" of the heart. His fig. 278 is similar to my fig. 17.




2


u




a


p^


^


-^j








'S.


'-^


0)


o




"rt


"o




O


cC


"5


"-3



'u



"2


o



o


32


5

^




^-\




JS



^— 1


OJ


^^




■ c3


■^



CO

c3




(4-1


CO


_« 



-<


o


1




O



O


"o





O


-fj



"sZ


/^



CO



<a


o


.rt


cj



-f-*






c


-^

F—


■5



rt


o


"3


_C



(C



'S


03



^^



=^




a


~


Oj




a


s





-<



Z3 03


JH



S-'


^

oq


o


-23



c3


. -V



H





o





— '


a


.



0)


CO


.2


03



-1^


o


^


O



S



o


03




-f^


&(


4)



-4^


(4-H




Q)


^


pC




1— 1


Hi-i


-tJ



02


o;


CO

O f5« 


o



-S


.^



-l-J



3


CO

o



c

s



^


o



e3



o


^


bO



>



o


(-1

.2



o


rC


OJ


'in



r^


o


<-— 4


+->



02



-f^


03



o


'1


«4-l

o


o



73


-^3


o


03



>


C


p


-fj



O


c3



c



g


^


o




S


,_^


rC


-o



^


oi


-t^


a;



a:


i^


o


-M

L^




r>


-fj


« 



tn


>>


CO



«3 4^


O



^


CQ


.f^


^


_,



+5


3 O



^ ^


c3 U3


o3


-C


+J


t>


02


a


-♦^



-C


■^


• r^


^


tt-H


o


73


X


-^^


o


rt


1=1


02



^


-►^


o3


P<


>



+^



03


»^


o


rt


+J


-fj




a


a;


a


o


CO

o


o


-£3

bC

C3


CO


So


IS


-t-3


2


E


cq


•a


c3

o



C71



c3


j3


•T3


t-H


■A


-— ■■


o


C3


M



"rt


.£3

•4J


o3


s


CO


-i


02


(4-1



s



"3


a;


MUSCULAR ARCHITECTUEE OF THE HUMAN HEART


263



03 V


o o ;^


o aJ

-G C -*^ C

a « 

.2 'S o


a,

o3 .S

«-£

^ =^

03 • G "^

03 -G

-^^ 03

^ "S

O C

O O

X "

tD

T— I _i. . o

bC 'C -ki 'Ai G °*

^ 1^ G ^

G « ^ o3 -G +=

"^ -^ ^

o S -^


be a; "C •-< O O p!^ 03 O


264 FRANKLIN P. MALL

of these bands is then extended around the heart to the posterior side much as is shown in fig. 8. As the aorta has been taken out (or macerated out in MacCallum's specimens) the first break into the cavity of the ventricle is immediately under the aorta, in fig. 8 which is also shown in fig. 18. Here it is clearly seen that the anterior papillary muscle is attached to the tip of the scroll as shown in fig. 17. That this split is natural and is easily made is shown by fig. 16, in which the two papillary muscles are shown with a strand of muscle tissue between them. This strand extends to the tip of the scroll and ends in the posterior ligament of the aorta (fig. 17, PL). This figure also shows the entire left ventricle unrolled with the two spiral strands hanging loosely below. The one associated with the posterior papillary muscle coming from the aorta is the bulbo-spiral band and the other which is attached to the anterior papillary muscle is the sino-spiral. Fig. 18 is made from the specimen shown in fig. 17, by cutting it in half and replacing the distal end within the proximal. In general it represents the left ventricle cut open from behind, and expresses clearly the course of the break when the wall of the left ventricle is unrolled.

What is here given only apparently contradicts what Krehl found when he isolated the Triebwerk of the left ventricle. My illustrations (figs. 10 and 11) show that this may also be unrolled if its point of origin is detached from the aortic septum. In so doing the entire circular bands which pass the septum, including the anterior papillary muscle, hold together, as shown in fig. 18. So Krehl's Triebwerk is not composed of muscle bundles which form complete rings but bundles which both arise and end in the tendinous structures at the base of the heart. In so doing the bundles turn upon themselves to form V-shaped loops just as do the bulbo-spiral and sino-spiral bundles as they enter the vortex. MacCallum's scroll again shows that all of the muscle fibers form V-shaped bundles which encircle the left ventricle. They are fitted into one another, those forming the circular bands making obtuse angles while the outer and inner bundles form acute angles at the apex. Between these two systems there are all intermediate gradations. (See Schema A.)


MUSCULAR ARCHITECTURE OF THE HUMAN HEART 265

It is apparent that the aiTangement of the superficial fibers of the heart is such that their contraction will cause the heart to rotate, as is well known to physiologists. In so doing the transverse diameter of the heart must diminish but it is not necessary that the long diameter should change as was pointed out by Hesse. ^ Since the rotation of the apex in contraction is accompanied by a straightening of the superficial spiral fibers, it must cause the inner spiral fibers to curve upon themselves for they run at right angles to the outer fibers. This is well illustrated in a diagram by Nicolai,^* which shows that in contraction the outer fibers of the ventricle become straighter, while the inner ones become more spiral. According to my description of the heart muscle fasciculi such a contraction need not change the length of the heart, but it exaggerates to the utmost the folds which are formed within the heart during systole; in fact the lumen of the left ventricle is nearly obliterated. In fig. 19 the heart, which is held in both hands, is represented as it is in diastole. In order to imitate contraction of the heart as it takes place in systole it is necessary to rotate the apex as shown in fig. 20. In this change of position the inner bundles are twisted as one wrings out a wet rag. This was Borelli's^*^ conception of the heart contraction which I do not believe can be improved on very much. Borelli gives an account of the arrangement of the muscle bundles of the heart in which it is pointed out that there is a general downward course of the fibers from the base to the apex where they form the vortex. He also gives a figure of two hands twisting a rope to illustrate the way the rotation of the contracting heart presses the blood out during systole. My dis ^^ The statement to this effect by Hesse, His and Braunc's Archiv, 1880, reUxtes to the dog's heart. Krehl, I.e., p. 349, thinks that it is equally applicable to the human heart, although evidence is wanting.

^^ Nicolai, Nagel's Handbuch der Physiologic, Braunenschweig, 1909, Bd. 1, fig. 74. Nicolai's description is entirely theoretical for he states that the architecture of the heart is l>y no means clear, and that the longer this subject is worked upon the more confused it becomes. Evidently Nicolai has neither studied suitable specimens, nor the literature upon the subject. It is becaue of the importance of NageFs Handbuch that I call attention to Nicolai's dilemma.

■'o Borelli. De motu animalium, Romae, 1681.


2G6 FRANKLIN P. MALL

sections su])])oit Boicllis cunception of the lucchaiiisin of the heart beat. In order to make it clear to the reader two illustrations of the same heart, in diastole and in systole, are given. The curved arrows in figs. 19 and 20 indicate the necessary rotation of the heart at its apex to convert the one figure into the other. It is to be noted that this specimen is from a i^ig's heart in which the circular band at the base is much more pronounced and that of the septum much less marked than in man. Also the ending of the bull)o-spiral band within the heart divides into two distinct bands, one of whicli unites with the front side of the aorta and the other encircles the heart as in man and ends in the posterior ligament of the aorta. The external spiral bundles have been removed. In the dilated heart, fig. 19, the inner bundles are unwrapped and the outer ones, which have been cut off, were lengthened. In fig. 20 the opposite is the case, the inner bundles including the papillary muscles are wrapped upon themselves and fill the lumen of the ventricle. Specimens like the one from which figs. 19 and 20 are made are not so difii(;ult to prepare and they show the mechanism of contraction nnich better than the illustrations do.

What has been said about the pig's heart can easily be read into the human, from the descrijition I have given of it. Fig. 7 shows that contraction of the bulbo-spiral band will rotate the apex and wind up, or put under great stress, the chief deeper bundles of the left ventricle, as shown in fig. 8. Contraction of the deep bulbo-spiral band, the Triebwerk of Krehl, will then complete the contraction of the ventricle.

To obtain a proper understanding of the architecture of tlu; heart the organ must be considered as a whole with a conception of the function it has to perform kept uppermost, as was done \)y Borelli and was emphasized by Weber and by Ludwig. Only from this standpoint is it possible to get a clear understanding of (his hitricate network of nuiscle bundles. Not only is this the case for tin; adult heart, l)ut without it we cannot hope to unravel its development, for the arrangement of the fibers nuist be due to functional adaptation from the time the heart begins to beat.


HERMAPHRODITISMUS VERUS IN MAN

J. F. GUDERNATSCH

From the Department of Embryology, Cornell University Medical College,

New York City

SEVEN FIGURES THREE PLATES

Hermaphroditismus verus is of such rare occurrence and so eminently important in our knowledge of the development of the genital organs, that it would seem worth while to add a new ease to the few so far recorded. Numerous instances of supposedly true hermaphroditism have been described, but only in rare instances have they stood a critical consideration.

The microscopic diagnosis in many cases has been incorrect, particularly in cases where the normal structure of the tissues had been altered by neoplasms so that their identification was almost impossible. Some instances of true hermaphroditism have been reported in which histological examination was neglected ; yet without a microscopic investigation the correct interpretation of malformations of this kind is at least doubtful.

The 'ovotestis' to be described in this article was taken from an individual forty years old who came to the hospital to be operated upon for tumor of the right inguinal region. In the left inguinal canal a similar, but somewhat smaller nodule was detectable.

The external genitals were of the female type; labia majora and minora were well developed and an introitus vaginae was present. The noticeably peculiar feature was the extremely enlarged clitoris with the opening of the urethra on its ventral

THE AMERICAN JOURNAL OF ANATOMY, VOL. 11, NO. 3

2G7


268 J. F. GUDERNATSCH

side, so that the organ offered rather the aspect of a hypospadia penis than that of a cUtoris.

No uterus was present and the vagina ended blindly (atresia vaginae). During the course of embryonic development, therefore, the greater part of the Miillerian ducts had been lost and the formation of the genital apparatus must have inclined toward the male type. A prostate-like organ was felt attached to the urethra, yet its real nature remains doubtful, since no microscopic examination could be made. In other cases in which both a vagina and prostate have been found they were seen to be connected with one another, but in the present instance no detectable communication existed between the vagina and the supposed prostate.

The distribution of the hair on the body was that tj-pical of the female. The pelvis was wide, but mammary glands were not developed, and the larynx was externally that of the male. The secondary sexual characters were not so decidedly of the male type that there arose any doubt about the sex of the individual, although the knowledge of the malformation existed. The individual was believed by herself and her associates to be a woman.

The individual has never menstruated, sexual intercourse has never taken place, and libido sexualis is not present. Her psychic disposition is that of a woman and she earns a living as a cook.

The tumor in the left inguinal region was left in place since it did not cause inconvenience and that in the right channel was removed. The nodule extirpated had about the form of a testicle with attached epididymis. It measured 6 cm. in length and 5 cm. in width and thickness. Histologically it proved to be male genital tissue, with the exception of a small nodule which exhibits a structure very similar to that of an ovary. Dr. James Ewing diagnosed this region of the specimen as true ovarian tissue. Others who were asked to examine the sections agreed with Dr. Ewing in considering the tissue ovarian.'

1 I wisli to thank Prof. A. Kohn, the well known histologist in the German University of Prague, for his careful examination of the sections and for the many suggestions he made in regard to their interpretation. The preparations were also demonstrated before the Eighth International Zoological Congress in Graz, and the structure was interpreted by all as ovarian.


HERMAPHRODITISMUS VERUS IN MAN 269

THE TESTICULAR STRUCTURE

The mass of the testicle is surrounded by an extremely broad tunica albuginea, the elements of which are arranged in a somewhat undulating manner and contain elastic fibres with but few nuclei and vessel. As might be expected the testicular tissue proper (fig. 2) is not of the appearance of the normal male sexual gland since the organ developed under very abnormal conditions. It resembles the well known pathological condition of a degenerating testicle, and the hyaline type of degeneration is typical of the kryptorchic mammalian testicle retained in the inguinal canal. The sections through the contorted tubules vary much in size, the smaller ones are circular in shape, the larger ones more or less oval and some a little bent or even S-shaped owing to the plane of section. The average diameter of the tubuli contorti is much smaller than normal. The epithelium which lines the tubules shows only one row of basal cells, and these are propably all of the Sertoli's cell type being rather large, triangular or cubical in shape with clear cytoplasm and large nuclei (fig. 2, ct). From their surfaces protoplasmic processes project into the lumen. Germ cells seem to be entirely absent since there are no indications of spermatogonia, spermatids or spermatozoa. It is not improbable that in the younger years of the individual anlagen of germ cells were present in the tubules, but failed to undergo further development on account of the abnormal physiological conditions. It must be assumed that formerly germ cells existed, for without them the development of male sexual glands is hardly conceivable.

The cells of Sertoli show different stages of degeneration as is indicated by the different densities and staining abilities of the nuclei. In some cells the nuclei are shrunken and lie in a nuclear cavity. This degeneration of the epithelial cells is probably due to the above mentioned hyaline degeneration of the wall of the seminal tubules. The inner layers of the tunica propria upon which the epithelial cells lie are swollen, while the constituent cell nuclei disappear entirely, forming a hyaline mass which lies as a band between the follicular wall and the epithelium (fig. 2, ct). This swollen band of cells pushes the epithelium towards the


270 J. F. GUDERNATSCH

lumen which thus becomes more and more occluded and often with gradual dissolution of the cells becomes entirely obliterated. The hyaline band exhibits a somewhat fibre-like structure with processes from it extending between the epithelial cells. It varies in thickness and in some places is entirely missing, and along with this variation in thickness the degeneration of the epithelial cells presents a regional distribution. In some regions the epithelial cell nuclei stain in a normal manner and are situated towards the periphery, in other regions the cell limits are indistinct and the nuclei lie near the lumen. This accords with former observations, and Finotti states that in the kryptorchic testicle, even when the individual is not a hermaphrodite and the germinal epithelium in places develops spermatozoa, the degeneration is not of uniform degree in all regions.

Wherever there is a membrana propria left in the form of a thm layer of spindle-like connective tissue cells, elastic fibres are usually present.

The testicle under consideration offers still another peculiarity. The interstitial tissue is enormously increased so that the seminal tubules are in places pushed far apart (fig. 2, i). Connective tissue fibres are comparatively scarce in this tissue.

The interstitial cells are normal in appearance, the majority being of a triangular or polygonal shape. The large nucleus (there are occasionally two nuclei) possesses one or more nucleoli and is usually excentrically situated. The cytoplasm is somewhat denser than that of ordinary connective tissue cells, it is finely granular and often contains in some regions a very fine brown pigment. Of the so-called Reinke's crystals nothing could be detected.

The interstitial cells are usually arranged either in small irregular groups or narrow streaks, often, however, they are united in large compact nests and grow so excessively that they actually invade the tunica albuginea. There is no relationship between these cells and the blood vessels as is often claimed.

This striking increase in the number of interstitial cells is a feature well known to the human embryologist as well as to the pathologist. During the fourth month of embryonal develop


HERMAPHRODITISMUS VERUS IN MAN 271

ment these cells constitute about two-thirds of the parenchyma of the testicle. Their number is also very much increased in the kryptorchic testicle. Whether the large amount of interstitial tissue in the present case is an embryonal condition due to arrested development, or simply a secondary pathological feature, is difficult to say. The latter, however, seems to be more probable since there are no indications, except perhaps the entire lack of germinal epithehum, that the gland did not develop normally. It probably degenerated later on account of its unusual position. Finotti claims that the gland does not degenerate for such a reason, but on account of an early predisposition to do so.

The entire accessory system of the male genital gland, rete testis, ductuli efferentes, ductus epididymis and vas deferens, are present. The globus major is, as far as the arrangement of the efferent ducts is concerned, rather well developed. The epithelium, however, is very degenerate in places (fig. 4), though towards the duct of the epididymis it approaches a normal condition. The epithelium in some tubules is of the low cubical type without foldings (fig. 3), while in others it shows the normal projections into the lumen with alternating columnar and cubical cells.

The structure of the epididymis resembles in parts that of the epoophoron and since both organs are derived from the Wolffian body, it is not impossible that in a true hermaphrodite we might find them somewhat mixed.

The ductus epididymidis is normally developed. The muscular coat of the efferent ducts increases in thickness as they approach the epididymis. There are numerous elastic fibres among the muscle cells ; if these, as Stohr states, do not appear until puberty is reached we must conclude that the entire efferent system reached a mature state independently of the testicle. This is also emphasized by the fact that the better developed parts are those farthest removed from the testicle. The duct of the epididymis, for instance, has a normal, highly columnar, ciliated epithelium, which shows no signs of degeneration. In the lumen is seen cell detritus, a finely granular mass and numerous concrements. The muscular coat of the spermatic cord is much increased.


272 J. F. GUDERNATSCH

The sclerotic blood vessels, as every^vhere seen in the sections, are typical of the kryptorchic testicle.

THE OVARIAN STRUCTURE

The female portion of the genital gland is rather small, the rudimentary ovary being a little nodule only 3 mm. in length and 2 mm. in width and thickness. It is enclosed within a cyst in the tunica between the testicle and the head of the epididymis (fig. 1, o). The typical ovarian stroma is easily recognized by the arrangement of the spindle-shaped connective tissue cells (fig. 5). This structure is nowhere else to be found in the human body. Cortical and medullary portions are distinguishable. The former consists of dense connective tissue rich in cells, and traversed by small blood vessels. The slender cells sometimes resembling smooth muscle fibres, are arranged either in strands or twirls. In the central portion of the ovarian body the connective tissue contains fewer nuclei and its elements are arranged in broader streaks. The blood vessels are large and bent.

The entire nodule is surrounded by a single-layered, cubical or cylindrical epithelium (fig. 5) which although rather primitive shows here and there slight cellular differentiation. Some cells are larger and broader and their nuclei are large and more circular and contain less chromatin than the neighboring cylindrical cells (figs. 6, 7, y). These cells are very probably primordial ova, yet a definite diagnosis cannot be made. However the decision that the body is ovarian in structure is sufficiently warranted by the typical stroma with its surface epithelium.

The ovary remained in an early stage of development as is indicated by the rather high columnar cells in . certain regions (fig. 6). A migration of primordial ova into the stroma and the formation of Graafian follicles has not taken place, probably due to the abnormal conditions of development. This accords with earlier reports on the subject which state that in all cases of hermaphroditism, whether true or spurious feminine, the epithelial part of the ovary is below the normal in development. Various transitions have been described from almost matui'e


HERMAPHRODITISMUS VERUS IN MAN 273

ovaries to those containing only primordial follicles or even empty follicles.

The primitive and rather small female portion found in this hermaphroditic genital gland indicates perhaps that in many cases of spm-ious hermaphroditism traces of ovarian tissue might be found, provided the entire testicular tissue be thoroughly searched, so far, however, this has never been done.

That both types of germinal tissue are in a hypoplastic condition is explained by Halban in the following way, the impulse for development, which normally is concentrated upon one system, in cases of hermaphroditism is called to act upon two systems and thus is insufficient to force either to the normal degree of development. •

In the present case male and female tissues are found in close contact in one gland, but an intermixing of the two kinds of tissue does not exist, and therefore the term "ovotestis," as is often used for this kind of gland, does not seem to be appropriate. In the true ovotestis of invertebrates male and female sexual cells are produced by one glandular structure.

The male part of an ' ovotestis' is as a rule considerably larger and further developed than the female. Kopsch and Szymonovicz have observed this in all hermaphroditic vertebrates and it likewise holds for the present case. This individual, however, shows many more external female characters than one should expect when considering the large male part of the genital gland. If the interstitial cells of the testis are really responsible for the accessory sexual characters then the person in question should show a typical male condition. The body of the individual is not a perfect woman, yet the male characters are not so outspoken that she could be called a man-woman. This fact is surprising from the study of the genital gland tissue, as far as this could be investigated, but it must be remembered that the actual amount of ovarian tissue the individual really carried in its body is not known. As has been mentioned above, a nodule existed in the left inguinal canal similar to that described from the right, this in all probability may have also been genital gland tissue and it might have contained a preponderance of ovarian material.


274 J. F. GUDERNATSCH

Such is not improbable since in malformations of this kind a great difference between left and right genital glands has often been observed, Salen, for instance, describes a case of true hermaphroditism, in which the right side contained an "ovotestis," while on the left a perfect ovarium was found. Lilienfeld states that in all cases of disturbed development of the genital region the female type is predominant on the left side. In this individual a large ovary may have been present on one or both sides higher up in the abdominal cavity, or there may have been more ovarian tissue present during an earlier period of life than can now bedetected. This latter suggestion would account for the development of the strong female characters of the individual. Kermauner states in Schwalbe's "Die Missbildungen des Menschen und der Tiere" that "whether the entire defectrmay be regarded as primary aplasia, or later involution, cannot always be decided. In no case can it be entirely denied that microscopic remnants of ovarian tissue, perhaps transformed beyond recognition, may be located somewhere in the abdominal cavity."

Whatever circumstances were responsible for the strong female characters of this person, it is interesting that along with them the male sexual apparatus developed to the perfection here described. The cavities and the concrements of the epididymis would indicate that a secretory function was performed by the epithelium.

This case of true hermaphroditism recalls the old theory of Waldeyer according to which there is a bisexual anlage of the genital gland, as opposed to Lenhossek's idea of an indifferent anlage. Instances in which male and female genital tissue are found next to each other speak at least for Waldeyer's view that the ovary develops from a different region of the genital ridge from that of the testicle even though they may not entirely support his theory of hermaphroditism. In all the cases of true hermaphroditism the ovary occupies the same relative position to the testicle. It seems strange that there should always be a sharp distinction between the two kinds of tissue and never an undefined mixing of both elements (true ovotestis) as might be expected, if all cells of the germinal epithelium could produce either male or female tissue. *■


PLATE I

EXPLANATION OF FIGURES

1 General view of the relative iiositions oi" testicle, I, ovary, o; nnd epididymis,

e. Dia. 1:17.

2 Te.stioulai- tissue, showing I lie hyaline wall of the eonvoluLed tuhuh^s, cf, and tiie larffe masses of interstitial crells, /. Dia. 1:90.

PLATE 2

EXPLANATION OF FIOURES

3 and 4 Sections through dirf<'r(Mit narts of the e])ididyinis. The tissue in fio;. 3 reseml)les somewiiat a pM.rovarinii siruelure. Dia. 1:00.

PLATE 3

EXrLAN.\TION OF FIGDRES

5 Ovarian tissue. Dia. 1:50.

6 and 7 Germinal epithelium of the ovary; p, somewhat differentiated cells, prol)al>ly inimoi-dial ova. Dia. 1:200.


HERMAPIIRODITISMUS VERUS IN MAN

J. F. CilIDEKNATSCH


PLATE 1




y^^ -->.»• -r-^o, * :^\ ^r^ xj^tr ••-V^i: «<»




" ^»»<;-«. -J ■•\: > '■ ■« \



'A




C t w j




'4« 









>^^ •



^'


THE AMERICAN JOURNAL OF ANATOMY, VOL. 11, NO. 3


HERMAPHRODITISMUS VERUS IN MAN

J. F. GUDERN^ATSCH


PLATE 2





THE AMERICAN JODRNAI. OF ANATOMY, VOL. 11, XO. 3


HERMAPHllODITLSMUS VERUS IN MAN

J. F. GUUKHNATSCII


PLATE 3


•W^'^'*'^,


)


JT


</'



5. • -.^, « .^.•^t:^.*^.





THE AMEllICAN JOURNAL OF ANATOMY, VOL. II, NO. 3


HERMAPHRODITISMUS VERUS IN MAN 275

Every embryo has the anlagen of the efferent ducts for the expulsion of both male and female products of the genital glands, which indicates that the male and female sexual apparatus are rather distinct from one another, thus it does not seem impossible that two distinct regions of the germinal epithelium might exist next to one another, one giving off the male, the other the female primordial cells. Wliy is the Mlillerian duct always laid down in the male if there are no female tendencies in the undifferentiated embryo? It seems that in every embryo there is a trace of a female tendency. Some authors, Benda for instance, go so far as to claim "that the primary anlage of the entire sexual system of the vertebrates must be regarded as female." Waldeyer's view, however, may be correct, that the Miillerian duct alone is the primary efferent duct for the genital glands in both sexes. He believes that in its function it corresponds to the primitive opening for the products of the gonads in the lowest vertebrates, the porus abdominalis. This scarcely seems possible, since in some fish adbominal pores and efferent genital ducts exist side by side and thus the second do not replace the first. Benda on the basis of Waldeyer's view may be justified in his conclusion that at no time do both ducts, Wolffian and Miillerian, exist as parallel genital ducts. Yet this is not entirely true, for the Wolffian duct, even in early periods, when it certainly serves as the mesonephric duct, must possess the potential faculty of developing into a male genital duct. This faculty is possessed by the duct in all embryos, whether the further development is male or female.

In the resulting female the possession of the quality to develop the mesonephric duct into a male sperm duct seems likely, if not proven, by the fact that the remnants of the Wolffian body, the Epoophoron and Paroophoron, closely resembling the structure of some parts of the epididymis, are found to exist.

The present case is interesting in still another du-ection. The sister of this hermaphrodite also shows irregularities in the formation of the external genital organs. Unfortunately only the testimony of laymen could be secured regarding this. In man a hereditary tendency towards hermaphroditism has never been scientifically proven, though several cases have been supported


276 J. r, GUDERNATSCH

by laymen. Reuter, however, found among three pigs of one Utter one true and two spurious hermaphrodites, and in a later litter from the same sow a pseudo-hermaphrodite occurred.

From this investigation the sex of the individual remains undetermined. According to Virchow it is an ' individuum utriusqua generis.' According to Klebs the condition is a typical hermaphroditismus verus. Klebs regards an individual as a true hermaphrodite, when the genital glands of both sexes are united in it. The physiological state is of no importance, simply the anatomical fact. An anatomical hermaphroditism seems to be all we can expect to find in vertebrates. The physiological hermaphroditism, as is normally the case in invertebrates, may hardly be looked for in man. In the higher vertebrates the persistence of both genital glands, is looked upon as an imperfect development and under such circumstances it is only natural that their physiological faculty should be reduced.

Our knowledge of the etiology of these malformations is almost nil, since in general our conceptions of the principles involved in the development of sex are still rather vague. It remains for the anatomist and embryologist, perhaps for the experimenter, to bring about a deeper understanding of these abnormalities. So long as the interest in them is reserved for pathologists and clinicians, and the malformations of the external genitals remain the only thing of interest, the literature regarding such anomalies will be as Benda states overcrowded with sensational reports," with no exact investigation of the anatomical and histological features.

Hermaphroditism in the sense that separate testicles and ovaries are found has not been demonstrated in man, nor even in other mammals beyond doubt. Yet there are four cases of the socalled ovotestis on record, two of these with neoplastic changes in the male portion. The present is therefore the fifth recorded case of true hermaphroditism in man.


HERMAPHRODITISMUS VERUS IN MAN 277

LITERATURE CITED

Benda, C. 1895 Hermaphroditismus und Missbildiingen mit Verwischung des Geschlechtscharakters. Ergebn, d. allg. Path., vol. 2, p. 627.

Born, G. 1894 Die Entwickelung der Geschlechtsdriisen. Erg. An. u. Entw., vol. 4, p. 592.

Corby, H. 1905 Removal of a tumor from a hermaphrodite. Brit. Med. J., vol. 2, p. 710.

FiBiGER, J. 1905 Beitriige zur Kenntnis des weiblichen Scheinzwittertums. Virchow's Arch. f. path. An., vol. 181, p. 1.

FiNOTTi, E. 1897 Zur Pathologic und Therapie des Leistenhodens nebst einigen Bemerkungen iiber die grossen Zwischenzellen des Hodens. Arch. f. klin. Chir., vol. 55, p. 120.

Halban, J. 1903 Die Entstehung der Geschlechtscharaktere. Arch. f. Gynaek. vol. 70, p. 205.

Hansemann, D. 1895 Uber die grossen Zwischenzellen des Hodens. Virchow's Arch. f. path. Anat., vol. 142, p. 538.

HiRSCHFELD, M. 1905 Ein Fall von irrtiimlicher Geschlechtsbestimmung. Monatschr. f. Harnkr. u. sex. Hyg., vo.. 2, p. 53.

1905 Ein seltener Fall von Hermaphroditismus. Monatsschr. f . Harnkr. &. sex. Hyg., vol. 2, p. 202.

HoFMEiSTER 1872 Untersuchuugen liber die Zwischensubstanz im Hoden der Saugetiere. Sitzungsb. Akad. Wiss. Wien, vol. 65.

Janosik, J. 1887 Bemerkungen iiber die Entwicklung des Genitalsystems. Sitzungsber. Akad. Wiss. Wien, vol. 99, 3. Abt., p. 260.

KoPSCH, Fr. u. Szymonowiez, L. 1896 Ein Fall von Hermaphroditismus verus bilateralis beim Schweine, nebst Bemerkungen iiber die Entstehung der Geschlechtsdriisen aus dem Keimepithel. An. Anz., vol. 12, p. 129.

LuKSCH, F. 1900 tjber einen neuen Fall von weit entwickeltem Hermaphroditismus spurius masculinus internus. Ztschr. f. Heilk., Abt. f. Path., vol. 21, p. 215.

Meixner, K. 1905 Zur Frage des Hermaphroditismus verus. Ztschr. f. Heilk., Abt. f. prakt. Anat., vol. 26, p. 318.

MiHALKOWicz, G. 1885 Untersuchungen iiber die Entwicklung des Harn-und Geschlechtsapparates bei Amnioten. Intern. Monatsschr. f. An. u. Phys.; vol. 2, p. 1.

Neugebauer, F. 1908 Hermaphroditismus beim Menschen. Leipzig.

Philipps, J. 1887 Four cases of spurious hermaphroditism in one family. Transact. Obst. Soc. London, vol 28, p. 158.


278 J. F. GUDERNATSCH

Pick, L. 1905 Uber Adenome der mannlichen und weiblichen Keimdriise bei Hermaphroditismus verus und spurius. Berl. klin. Woch., vol. 43, p. 502.

1905 XJber Neubildungen am Genitale bei Zwittern. Arch. f. Gynaek., vol. 76, p. 191.

Plato, J. 1897 Die interstitiellen Zellen des Hodens und ihre physiologische Bedeutung. Arch. f. mikr. Anat., vol. 48, p. 281.

Reinke, Fr, 1896 Beitrage zur Histologie des Menschen. Arch. f. mikr. Anat., 1896, vol. 47, p. 34.

Reizenstein, a 1905 tJber Pseudohermaphroditismus masculinus. Miinchn. med. Woch., vol. 52, p. 1517.

Salen, E. 1899 Ein Fall von Hermaphroditismus verus unilateralis beim Menschen. Verb. Deutsch. Path. Ges., vol. 2, p. 241.

ScHiCKELE, G. 1906 Adenoma tubulare ovarii (testiculare). Hegar's Beitr. z. Geburtsh. u. Gynaek., vol. 11, p. 263.

ScHWALBE, E. 1906 Die Morphologic der Missbildungen des Menschen und der Tiere, /3. Jena.

Simon, W. Hermaphroditismus verus. Virchows' Arch. f. path. An., vol. 172, p. 1.

Spangaro, S. 1900 tJber die histologischen Veranderungen des Hodens und des Samenleiters von Geburt an bis zum Greisenalter. Anat. Hefte, vol. 18.

TouRNEUX, F. 1904 Hermaphroditisme de la glande g6nitale chex la taupe femelle adulte et localisation des cellules interstitielles dans le segment spermatique. Comp. rend, de I'assoc. des anat., Toulouse, p. 49.

Unger, E. 1905 Beitrag zur Lehre vom Hermaphroditismus. Berl. klin. Woch., vol. 42, p. 499.

Waldeter, W. 1870 Eierstock und Ei. Leipzig,


THE DEVELOPMENT OF THE SYMPATHETIC NERVOUS SYSTEM IN TURTLES

ALBERT KUNTZ

From the Laboratories of Animal Biology of the State University of Iowa

THIRTEEN FIGURES

CONTENTS

Introduction 279

Observations 281

Sympathetic trunks 281

a Development 281

b Histogenetic relationships ' 294

Prevertebral plexuses 295

Genital plexuses 299

Vagal sympathetic plexuses 299

a Introductory 299

b Myenteric and submucous plexuses 299

c Pulmonary plexuses 302

d Cardiac plexus 302

e Histogenetic relationships 303

Discussion 305

Summary • 309

Bibliography 312

INTRODUCTION

The present series of observations on the development of the sympathetic nervous system in turtles represents a continuation of the writer's investigation of the development of the sympathetic nervous system in vertebrates. The observations set forth in the present paper are based on embryos of Thalassochelys caretta (loggerhead turtle) and of Chelydra serpentina (snapping turtle) ,

Embryos of Thalassochelys caretta develop more slowly and are comparatively larger than embryos of Chelydra serpentina.

279


280 ALBERT KUNTZ

Preparations of the former are, therefore, somewhat more satisfactory for the study of the development of the sympathetic nervous system than preparations of the latter. The present study is based primarily on embryos of Thalassochelys caretta, observations on embryos of Chelydra serpentina being introduced wherever such introduction has seemed desirable.

It is a real pleasure to express my obligations to Prof. G. L. Houser for many helpful suggestions and for valuable criticism during the progress of this investigation. I also desire to express my indebtedness to Prof, F. A. Stromsten for the use of a large number of embryos of Thalassochelys caretta which were collected by him at the Dry Tortugas, Florida, during the summer of 1910.

As indicated in my earlier papers,^ the most satisfactory preparations for the study of the development of the sympathetic nervous system were obtained from embryos which were fixed in chrom-aceto-formaldehyde, sectioned to a thickness of 10 micra, and stained by the iron-hsematoxylin method. This method was employed almost exclusively in this investigation.

The literature bearing on the development of the sympathetic nervous system has been reviewed by the writer in an earlier paper f therefore, only such references will be made to the literature in this paper as seem to be necessary.

Observations on the development of the sympathetic nervous system in reptiles are few and incomplete. According to C. K. Hoffman ('90), the ganglia of the sympathetic trunks in reptiles arise from cells which are derived directly from the spinal ganglia according to the view first advanced by Balfour and later elaborated by Onodi. Neumayer ('06) described the earliest anlagen of the sympathetic nervous system in Lacerta (Spec ?) as slender cellstrands growing ventro-mesially from the mixed spinal nerves. As development advances these cell-strands become more conspicuous and terminate distally in distinct cell-aggregates which constitute the anlagen of the ganglia of the sympathetic trunks.

1 See bibliography.

2 Tho (ievelopment of the sympathetic nervous system in mammals. Jour. Comp. Neur. and Psych., vol. 20, no. 3, pp. 211-258.


SYMPATHETIC SYSTEM IN TURTLES 281

Neumayer's observations on embryos of Lacerta reveal nothing which is characteristic of reptiles. His conception, according to which the sympathetic anlagen arise from cells which are to be regarded as the offspring of the dorsal and the ventral nerve-roots and are differentiated in situ, is based primarily on other classes of vertebrates.

According to Held's ('09) observations on embryos of Emys europea, the sympathetic trunks arise as cell-aggregates lying along the lateral surfaces of the aorta at the distal ends of cellchains which grow mesially from the spinal nerves. Furthermore, Held has observed that in this species more than one such cell-chain may grow from the spinal nerve toward the aorta. According to his observations, the cell-aggregates constituting the anlagen of the sympathetic trunks soon become broken up into small cell-groups which become more or less scattered in the mesenchyme. From these cell-groups cells migrate ventrally, some of which enter the mesentery and give rise to the sympathetic plexuses in the walls of the digestive tube. Held agrees with Neumayer that the sympathetic anlagen arise as outgrowths from the spinal nerves. He, however, traces the origin of the cells giving rise to the sympathetic nervous system to the spinal ganglia.

OBSERVATIONS

Sympathetic trunks

a. Development. The earliest traces of the anlagen of the sympathetic trunks appear in the thoracic and the dorsal region of embryos of Thalassochelys caretta during the eighth day of incubation, as loose cell-aggregates lying along the lateral surfaces of the aorta just below its dorsal level. At this stage the spinal ganglia are not yet completely differentiated. Traces of a few short fibers constituting the earhest fibers of the dorsal nerve-roots appear in the distal ends of the spinal ganglia, while in a few sections the rudiments of the ventral nerve-roots appear as inconspicuous fiberbundles growing out from the neural tube. The spinal ganglia are not sharply limited distally. Tracts of cells may be traced


282


ALBERT KUNTZ


from the distal ends of the spinal ganglia diagonally through the mesenchyme toward the lateral surfaces of the aorta, where they terminate in loose cell-aggregates which constitute the anlagen of the ganglia of the sympathetic trunks (fig. 1).

The cells which thus become separated from the cerebro-spinal nervous system and migrate peripherally before the spinal nerves



Fig. 1 Transverse section through the trunk region of an eight-day embryo of Thalassochelys caretta. X 210. Ao., aorta; M.C., path of cells migrating into sympathetic anlage; M.P., muscle plate ; Nc, notochord; Sp.G., spinal ganglion; Sy., anlage of symphathetic trunk.


arise may be distinguished from the cells of the mesenchyme by their slightly deeper stain and by the characteristic chromatin structure of their nuclei. The cells in these tracts leading from the spinal ganglia into the cell-aggregates constituting the anlagen of the sympathetic trunks are not closely aggregated nor is there any evidence that they are embraced in syncytia. Although mitotic figures in their various phases occur frequently along


SYMPATHETIC SYSTEM IN TURTLES 283

these paths, the peripheral advancement of the cells cannot be accounted for by the pressure due to mitotic division. These elements are too loosely aggregated to permit of being pushed forward by the pressure which might be exerted by the mere crowding due to the multiplication of cells. Neither is there any apparent line of weakness in the mesenchyme which might determine the path of these migrant nervous elements. The course taken by them is approximately the most direct course from the distal ends of the spinal ganglia into the regions in which the anlagen of the sympathetic trunks arise (fig. 1, M. C).

At the close of the ninth day of incubation the anlagen of the sympathetic trunks are present from the cervical to the sacral region. The cell-aggregates constituting the anlagen of the sympathetic ganglia appear in tranverse sections as oval or elongated cell-masses lying along the lateral surfaces of the aorta and along the dorsal surfaces of the carotid arteries. These cell-aggregates are best developed in the thoracic and in the dorsal region, i. e., in the regions in which the sympathetic anlagen first arise. The spinal ganglia are now well differentiated and the fibers of the spinal nerves may be traced peripherally for some distance beyond the level of the aorta. The anlagen of the ganglia of the sympathetic trunks are connected with their respective nerves by cellular tracts. In some sections fibers appear in the proximal parts of these cellular tracts, constituting the earliest fibers of the communicating rami. The cellular tracts extending from the distal ends of the spinal ganglia into the anlagen of the sympathetic trunks, which were so conspicuous in the previous stage, no longer appear. The cells which become separated from the cerebro-spinal nervous system in the region of the trunk now migrate peripherally along the paths of the spinal nerves and of the communicating rami (fig. 2) .

At this stage, cells apparently become separated from the spinal ganglia in considerable numbers. At the same time cells may be traced from the mantle layer in the ventral part of the neural tube, across the marginal veil, into the proximal parts of the ventral nerve-roots. That cells migrate from the neural tube into the ventral nerve-roots cannot be doubted. In many sections

THE AMERICAN JOURNAL OP ANATOMY, VOL. 11, NO. 3


284


ALBERT KUNTZ



Fig. 2 Transverse section through the trunk region of a nine-day embryo of Thalassochelys caretta. X 210. Ao., aorta;' M. P., muscle phxte; Nc, notochord; Sp.G., spinal ganglion; Sp.N., spinal nerve; Sy., anlage of symphathetic trunk; V.N.R., ventral nerve-root.


continuous lines of cells may be traced from the mantle layer into the proximal parts of the ventral nerve-roots, and occasionally one of these cells may be observed half in and half out of the neural tube. These cells obviously advance peripherally along the fibers of the ventral nerve-roots. After they have advanced beyond the point of union of the dorsal and the ventral nerveroots it is no longer possible to distinguish between the cells which have their origin in the ventral part of the neural tube and wander


SYMPATHETIC SYSIEM IN TURTLES 285

out along the ventral nerve-roots and those which wander down from the spinal ganglia. As the cells advancing peripherally along the paths of the spinal nerves reach the origin of the communicating rami many of them deviate from the course of the spinal nerves and enter the anlagen of the sympathetic trunks (fig. 2, Sy.). The cells which wander into the anlagen of the sympathetic trunks along the paths of the communicating rami are, doubtless, derived in part from the spinal ganglia and in part from the ventral part of the neural tube. The great majority of the cells present in the spinal nerves are still undifferentiated and there is no reason to suppose that cells advancing peripherally from the spinal ganglia enter the sympathetic anlagen while those advancing peripherally from the ventral part of the neural tube do not. Furthermore, the fibers growing peripherally into the communicating rami are primarily ventral root fibers. If the growing nerve-fibers exert any guiding influence on the peripherally advancing cells, as, doubtless, they do, it is highly probable that many of the cells which wander out along the fibers of the ventral nerve-roots wander along these fibers into the anlagen of the sympathetic trunks.

In his work on the development of the head of Gymnophion, Marcus ('10) described light areas in contact with and partly surrounding the anlagen of the sympathetic trunks in the early stages. These areas he interpreted as lymph spaces and expressed the opinion that the anlagen of the ganglia of the sympathetic trunks grow mesially from the spinal nerves in these lymph spaces or at least are surrounded by them. To quote: "Der Zellstrang der Sympathicusanlage wachst medianwarts vor, gegen die Dorsalseite der Aorta zustrebend. Dabei kann ich mich des Eindrucks nicht erwehren, dass der Sympathicus in einem Lymphraum seitlich der Aorta hineinwachst oder dochjedenfalls, dass er von Lymphgefaszen umspielt ist."

In the early stages in embryos of the turtle, narrow spaces similar to those described by Marcus in embryos of Gymnophion may be observed in contact with and in some cases almost surrounding the anlagen of the sympathetic trunks as they appear in transverse sections. Such spaces, however, are not present


286 ALBERT KUNTZ

in all sections, while in some sections they appear on the lateral sides of the sympathetic anlagen and are not apparent on the mesial sides.

It may be noted at this point that while in the eight-day stage such light areas sometimes appear in contact with the loose cellaggregates constituting the anlagen of the sympathetic trunks, there is no evidence of such areas of weakness in the mesenchyme along the paths of migration of the cells which wander from the distal ends of the spinal ganglia into the anlagen of the sympathetic trunks before the spinal nerves may be traced peripherally. In view of this fact and in view of the fact that the connections of the anlagen of the sympathetic trunks with the spinal nerves along the paths of the communicating rami arise after the anlagen of the sympathetic trunks are already present, it does not seem probable that in embryos of the turtle the anlagen of the sympathetic trunks grow peripherally in lymph spaces, but rather that lymph spaces are formed in contact with the sympathetic anlagen. My observations on embryos of Thalassochelys and on embryos of Chelydra are in full agreement on this point. Lymph spaces in contact with the anlagen of the sympathetic trunks cannot be satisfactorily traced in the later stages in embryos of the turtle because, as will be shown presently, the anlagen of the sympathetic trunks soon become more or less scattered and no longer appear as compact cell-aggregates.

At the close of the eleventh day of incubation the anlagen of the sympathetic trunks have become somewhat larger and more conspicuous. They now lie in closer proximity with the aorta and are connected with the spinal nerves by comparatively thick fibrous communicating rami (fig. 3, C. R.). The anlagen of the sympathetic trunks no longer appear as definitely limited cellaggregates, but are becoming somewhat scattered. In transverse sections they appear to be made up of several irregular cellgroups more or less closely associated with each other (fig. 3, Sy.). Such scattering of the anlagen of the sympathetic trunks is less apparent in the anterior than in the posterior region of the body. In the anterior region the anlagen of the sympathetic trunks appear in transverse sections as somewhat transversely


SYMPATHETIC SYSTEM IN TURTLES 287

elongated cell-aggregates lying along the dorsal surfaces of the carotid arteries. Farther posteriorly the sympathetic anlagen lie close to the dorso-lateral aspects of the aorta. From the region of the suprarenals posteriorly scattered cell-groups may be found as far ventrally as the ventral level of the aorta.

The spinal nerves are at this stage large and conspicuous fiber-tracts containing numerous accompanying cells which are apparently migrating peripherally. Migration of medullary cells into the ventral nerve-roots has probably reached its maximum at about this stage. In nearly every section which passes through




\ ,^^



CK

Fig. 3 Transverse section through the anlage of the sympathetic trunk in an eleven-day embryo of Thalassochelys caretta. X 210. Ao., aorta; CjB., communicating ramus; Sp.N., spinal nerve; Sy. anlage of sympathetic trunk.

the origin of a ventral nerve-root cells may be observed in the marginal veil, while in many sections complete lines of cells may be traced from the mantle layer into the proximal part of the ventral nerve-root.

It may be of interest to note that in embryos of Thalassochelys caretta cells may be traced from the ventral part of the neural tube into the ventral nerve-roots more readily than in embryos of Chelydra serpentina. This is probably due to the fact that the former develop more slowly than the latter and that, therefore, the migrant medullary cells linger for a longer period at the point of exit from the neural tube.

Fibers are now present in the dorsal nerve-roots connecting the spinal ganglia with the neural tube. In a few sections cells


288 ALBEKT KUNTZ

could be observed apparently on their way from the dorsal part of the neural tube into the dorsal nerve-roots. It is not probable, however, that many cells wander out from the dorsal region of the neural tube along the fibers of the dorsal nerve-roots. Migration of cells from this region of the neural tube is probably only a transient process which may play some part in the development of the spinal ganglia.

At the close of the thirteenth day of incubation the anlagen of the sympathetic trunks appear to be somewhat more widely scattered than in the preceding stage. The communicating rami have assumed no greater prominence and their fibers cannot yet be traced beyond the anlagen of the sympathetic trunks. In many sections cell-strands may be observed pushing out from the spinal nerves proximal to the origin of the communicating rami toward the anlagen of the sympathetic trunks. These cell-strands are similar to those described by Held ('09) in embryo of Emj^s europea.

From the region of the suprarenals posteriorly cells move ventrally from the anlagen of the sympathetic trunks and build up cell-aggregates along the ventro-lateral aspects of the aorta which, as will be shown later, constitute the anlagen of the prevertebral plexuses.

Cells still migrate peripherally both from the spinal ganglia and from the ventral part of the neural tube along the paths of the spinal nerves. The marginal veil in the neural tube, however, is becoming wider and cells are apparently migrating into the ventral nerve-roots less rapidly than in the preceding stages.

During the seventeenth day of incubation the anlagen of the sympathetic trunks are still more widely scattered than in the preceding stage (fig. 4, Sy.). The communicating rami have not increased in size appreciably nor is there much evidence that cells still continue to wander along their paths into the sympathetic anlagen. The cell-strands which at the close of the thirteenth day were observed pushing out from the spinal nerves proximal to the origin of the communicating rami now appear as irregular cellular tracts extending diagonally through the mesenchyme toward the dorso-lateral aspects of the aorta where they become


SYMPATHETIC SYSTEM IN TURTLES


289


merged with the anlagenof the sympathetic trunks (fig. 4,>S. C. R.). In a few sections these cellular tracts were observed to unite with the communicating rami just proximal to the point at which the latter enter the sympathetic anlagen.

The cellular tracts above described apparently have their origin in the spinal nerves. In a few instances at this stage fibers were


^#






i-SpR


/^ ^Sy.


Fig. 4 Transverse section through the anlage of the sympathetic trunk in a sixteen-day embryo of Thalassochelys caretta. X 210. Ao., aorta; Ci2./ communicating ramus; S.C.R., secondary cellular tract extending from the spinal nerve into the anlage of the sympathetic trunk; Sp.N., spinal nerve; Sy., anlage of sympathetic trunk.


observed to deviate from the spinal nerves along these cellular tracts. These fibers may usually be recognized as ventral root fibers. They deviate from the spinal nerves at a point so near the union of the dorsal and the ventral nerve-roots that they may readily be traced proximally as fibers of the ventral roots.


290 ALBERT KUNTZ

As development advances these cellular tracts assume greater prominence and gradually become more fibrous. The interval along the spinal nerve separating the origin of the primarj'- communicating ramus from the origin of this secondary tract gradually decreases until the two tracts come into close proximity and finally fuse with each other. The secondary tract is not carried forward along the spinal nerve-trunk, but the primary communicating ramus is crowded backward somewhat. This is probably due to the formation of the coelom and of the Wolffian bodies. The angle between the primary communicating ramus and the proximal part of the spinal nerve, which in the early stages is an obtuse angle, has now become an acute angle. At the close of the twentieth day of incubation the primary communicating ramus and this secondary tract have come to lie in such close proximity with each other that in some instances they can no longer be distinguished as separate tracts.

That the origin of the communicating ramus is actually shifted proximally along the spinal nerve-trunk is shown by the curve in the accompanying figure (fig. 5). This curve is based on actual measurements of the interval between the origin of the ventral nerve-root and the origin of the primary communicating ramus in successive stages ranging from the thirteenth to the twentyeighth day of incubation. The curve is based on the averages of five independent measurements. The figures in the horizontal line indicate the number of days of incubation of the embryos; the figures in the vertical line indicate the relative length of the intervals between the origin of the ventral nerve-roots and the origin of the primary communicating rami in embryos at successive stages of incubation. This curve shows that the interval between the origin of the ventral nerve-root and the origin of the primary communicating ramus increases until the close of the twentieth day of incubation; then decreases until the close of the twenty-second day of incubation, after which it again increases. The somewhat abrupt descent in this curve which occurs between the twentieth and the twenty-third day of incubation, doubtless, indicates an actual proximal displacement of the origin of the primary communicating ramus along the spinal nerve-trunk.


SYMPATHETIC SYSTEM IN TURTLES


291


The accompanying diagrams (figs. 6 and 7) have been introduced to illustrate successive stages in the process by which the primary communicating rami are shifted proximally along the spinal nerve-trunks until they fuse with the secondary tracts growing mesially from the proximal parts of the spinal nerves.

After the twenty-fourth day of incubation but a single tract may be recognized connecting the sympathetic anlage with the



u IS 3,6 n n


Fig. 5 Curve designed to indicate the relative length of the intervals between the origin of the ventral nerve-root and the origin of the primary communicating ramus in embryos of Thalassochelys caretta in successive stages of development. For explanation see text.

spinal nerve. This tract is distinctly fibrous, but still contains numerous accompanying cells. Most of the cells present in the communicating rami in these later stages, doubtless, have wandered out from the spinal ganglia. There is no longer any evidence of the peripheral migration of cells from the ventral part of the neural tube, except as an individual cell occasionally passes through the external limiting membrane.

The development of the sympathetic trunks in embryos of the turtle proceeds comparatively slowly. After cells cease to


292


ALBERT KUNTZ


migrate peripherally from the cerebro-spinal nervous system the anlagen of the sympathetic trunks assume more definite propor


Fig. 6 Diagrammatic transverse section through the trunk region of a twentyday embryo of Thalassochelys caretta. Ao., aorta; C.R., communicating ramus; A'^c, notochord; S.C.R., secondary cellular tract from spinal nerve to sympathetic anlage; Sp.G., spinal ganglion; Sp.N., spinal nerve; Sy., anlage of sympathetic trunk; V.N.R., ventral nerve-root.

Fig. 7 Diagrammatic transverse section through the trunk region of a twentyfour day embryo of Thalassochelys caretta. Ao., aorta; C.R., communicating ramus; Nc, notochord; S.C.R., secondary cellular tract from spinal nerve to sympathetic anlage; Sp. G., spinal ganglion; Sy., anlage of sympathetic trunk; V.N.R., ventral nerve-root.


tions, but the sympathetic ganglia do not appear as compact and definitely limited cell-aggregates until development has advanced for a considerable period after this stage is reached.


SYMPATHETIC SYSTEM IN TURTLES 293

The breaking up of the anlagen of the sympathetic trunks, the formation of secondary cellular tracts extending from the proximal parts of the spinal nerves into the sympathetic anlagen, and the proximal shifting of the origin of the communicating rami along the spinal nerve-trunks above described, which are so conspicuous in embryos of the turtle, in all probability, represent a condition which is characteristic of reptiles and which has phylogenetic significance.

His, Jr., ('97) called attention to the fact that in the chick two pairs of sympathetic trunks arise in the course of ontogeny. These he has designated as the 'primary' and the 'secondary' sympathetic trunks. In my work on the development of the sympathetic nervous system in birds,^ I was able to substantiate this observation of His, Jr. The primary sympathetic trunks in the chick arise about the beginning of the fourth day of incubation, as a pair of cell-columns lying along the sides of the aorta and along the dorsal surfaces of the carotid arteries. The anlagen of the secondary sympathetic trunks arise about the beginning of the sixth day of incubation, as ganglionic enlargements on the median sides of the spinal nerves. The cells giving rise to both the primary and the secondary sympathetic trunks migrate peripherally from the spinal ganglia and from the ventral part of the neural tube along the paths of the spinal nerves. In the early stages some of the cells advancing peripherally along the paths of the spinal nerves deviate from the course of the latter and wander toward the lateral surfaces of the aorta where they become aggregated to give rise to the primary sympathetic trunks. A little later the cells advancing peripherally no longer wander toward the aorta, but become aggregated at the median sides of the spinal nerves to give rise to the anlagen of the secondary sympathetic trunks. As development advances and the communicating rami grow mesially the entire cell-aggregates constituting the anlagen of the secondary sympathetic trunks are displaced toward the aorta at the distal ends of the growing communicating rami. As the second ' The development of the sympathetic nervous system in birds. Jour. Comp. Neur. Psych., vol. 20, no. 4, pp. 283-308.


294 ALBERT KUNTZ

ary sympathetic trunks increase in size and prominence the primary sympathetic trunks decrease until they finally disappear.

The phenomena above described in embryos of the turtle are of peculiar interest in view of the phenomena involved in the development of the sympathetic trunks in the chick. In embryos of the turtle, as in the chick, the earliest traces of the sympathetic trunks are found along the lateral surfaces of the aorta. In embryos of the turtle these formations do not give way completely to secondary formations, as is the case in the chick, but early break up to become aggregated once more during the later stages of development. Again, in embryos of the turtle, cells do not become aggregated at the median sides of the spinal nerves to form ganglionic enlargements, as is the case in the chick, but deviate from the course of the spinal nerves proximal to the origin of the communicating rami and advance in irregular cellular tracts toward the anlagen of the sympathetic trunks. In short, the phenomena involved in the development of the sympathetic trunks in the turtle seem to represent a generalized prototype of what has become a highly specialized condition in birds. This does not mean, however, that the sympathetic nervous system in turtles is the direct ancestral type of the highly speciahzed sympathetic nervous system in birds. The sympathetic nervous system in turtles is, doubtless, a speciaUzation of a still more generalized type in the ancient reptiles. The points of correspondence which have been pointed out, however, seem to warrant the conclusion that the sympathetic nervous system in birds bears a more or less direct phylogenetic relationship to the sympathetic nervous system in the ancestral type of reptiles.

b. Histogenetic relationships. In my earlier papers I have shown that the cells which migrate peripherally from the cerebro-spinal nervous system in embryos of mammals and of birds are the descendants of the 'germinal' cells (Keimzellen) of His; viz, the 'indifferent' cells and the 'neuroblasts' of Schaper. They have the same genetic relationships, therefore, as the cells which give rise to the neurones and to the neuroglia cells in the central nervous system. The great majority of the cells which migrate peripherally from the spinal ganglia and from the ventral


SYMPATHETIC SYSTEM IN TURTLES 295

part of the neural tube in embryos of the turtle also answer to the description of the 'indifferent' cells of Schaper. They are characterized by very little cytoplasm and by large rounded or elongated nuclei showing a delicate chromatin structure. In the sympathetic anlagen some of these cells early develop protoplasmic processes and may, therefore, be recognized as neuroblasts. Although neuroblasts have not infrequently been observed in the spinal and in certain of the cranial nerves in vertebrate embryos, I was not able to observe cells with distinct protoplasmic processes along the paths of migration of the cells giving rise to the sympathetic nervous system in embryos of the turtle.

The great majority of the cells present in the mantle layer in the neural tube in embryos of the turtle answer to the description given above for the cells which migrate peripherally. There can be no doubt, therefore, that the cells which take part in the development of the sympathetic trunks have the same genetic relationships as the cells which give rise to the neurones and to the neurolgia cells in the central nervous system. Mitotic figures in their various phases occur frequently all along the paths of migration and in the sympathetic anlagen. We are not to suppose, therefore, that ail the cells taking part in the development of the sympathetic trunks actually migrate as such from their sources in the cerebro-spinal nervous system. Doubtless, many arise by " the mitotic division of ' indifferent ' cells along the course of migration as well as in the sympathetic anlagen.

Prevertebral plexuses

At the close of the eleventh day of incubation of embryos of Thalassochelys caretta, after the anlagen of the sympathetic trunks have begun to break up and to become somewhat scattered, cells may be observed wandering ventrally from the anlagen of the sympathetic trunks in the entire region from the suprarenals posteriorly. At the close of the thirteenth day, these cells have become aggregated into groups of considerable size lying along the ventro-lateral aspects of the aorta. These cell-groups constitute the anlagen of the prevertebral plexuses.


296 ALBERT KUNTZ

At the close of the sixteenth day of incubation the anlagen of the prevertebral plexuses have become more conspicuous, but, like the anlagen of the sympathetic trunks, they are composed of cell-aggregates which are more or less scattered (fig. 8, P.V.). The limits of the anlagen of the several prevertebral plexuses cannot be determined at this stage. The cell-aggregates composing them are scattered to such an extent that traces of one or the other of these plexuses are not wanting in any transverse section in the entire region from the suprarenals to the posterior limits of the hypogastric plexus.

From the seventeenth to the twentieth day of incubation the anlagen of the prevertebral plexuses assume more definite proportions, but many lesser cell-groups still remain more or less scattered. In the region of the origin of the iliac arteries, cells wander mesially from the anlagen of the sympathetic trunks and descend between the iliac arteries to give rise to a wedge shaped cell-aggregate lying in the median plane of the body just dorsal to the mesentery (fig. 9, S. C. G. ). In some sections the ventral angle of this cell-aggregate projects slightly into the mesentery, while a few cells apparently become separated from it and wander ventrally toward the rectum. Continuous lines of sympathetic cells could not be traced ventrally in the mesentery, but in many sections in this region groups of nervous elements may be observed in the tissues associated with the walls of the rectum (fig. 9, S. C. R.). This process of migration of cells from the anlagen of the prevertebral plexuses in the posterior region of the body toward the rectum, doubtless, begins at an earlier stage. I was not able, however, to observe distinct groups of nervous elements associated with the rectum until about the nineteenth or the twentieth day of incubation.

In the earlier paper referred to above the writer traced the origin of the ganglion of Remak in the chick to cells which migrate ventrally from the anlagen of the hypogastric plexus. The ganglion of Remak in the chick is a conspicuous cell-column more or less circular in transverse section lying in the mesentery just dorsal to the rectum. Furthermore, the suggestion was offered that the ganglion of Remak may have its prototype in


SYMPATHETIC SYSTEM IN TURTLES


297




Fig. 8 Transverse section through the anlage of the sympathetic trunk and


the genital ridge in an embryo of Chelydra serpentina 10 mm. in length. X 210. Ao., aorta; G.P., anlage of genital plexus; G.R., genital ridge; P.V., anlage of prevertebral plexus; Sy., anlage of sympathetic trunk.


298


ALBERT KUNTZ


the aggregates of sympathetic cells associated with the rectum in embrj'os of reptiles. In view of the probable phylogenetic relationship which has already been pointed out between the


'•Xf *•-• "^" •• «•» •• ._ « « •••


.-i*r5




Fig. 9 Diagrammatic transverse section through the iliac arteries and the rectum in a wenty-day embryo of Thalassochelys caretta. I. A., iliac arteries; Mes., mesentery; R., rectum; S.C.G., sympathetic cell-aggregate dorsal to the mesentery; S.C.R., sympathetic cells associated with the rectum. Sy., anlage of sympathetic trunks; P.V., anlage of prevertebral plexus.

sympathetic nervous system in reptiles and in birds, it is highly probable that the sympathetic cell-aggregates associated with the rectum in turtles are not far removed from the ances


SYMPATHETIC SYSTEM IN TURTLES 299

tral type of the ganglion of Remak which is so enormously developed in the avian branch of the vertebrate series.

Genital plexuses

In embryos of Thalassochelys caretta from the sixteenth to the twentieth day of incubation or in embryos of Chelydra serpentina about 10 mm. in length, cells may be traced ventrally along the median sides of the Wolffian bodies to the lateral surfaces of the genital ridges (fig. 8, G. P.), where they become aggregated to give rise to the genital plexuses. The conditions here described agree with the conditions described by Held in embryos of Emys europea.

Vagal sympathetic plexuses

a. Introductory . In my earlier papers I have shown that in mammals and in birds the sympathetic plexuses related to the vagi; viz., the cardiac plexus and the sympathetic plexuses in the walls of the visceral organs, do not arise from cells which migrate ventrally from the sympathetic trunks and from the prevertebral plexuses, as the earlier investigators believed, but have their origin in cells which migrate from the vagus ganglia and from the walls of the hind-brain along the paths of the vagi. Because of the genetic relationship of these plexuses to the vagi, I have designated them as the 'vagal sympathetic' plexuses.* More recently 1 have traced the cells giving rise to the sympathetic plexuses in the walls of the digestive tube in fishes to the same sources. The present series of observations will show that in the turtles also the cardiac plexus and the sympathetic plexuses in the walls of the visceral organs arise from cells which have their origin in the vagus ganglia and in the walls of the hind-brain and migrate peripherally along the paths of the vagi.

h. Myenteric and submucous plexuses. In transverse sections through the anterior region of the oesophagus in embryos of Thal ^ The development of the sympathetic nervous system in mammals, p. 235. Jour. Comp. Neur. Psych., vol. 20, no. 3, pp. 211-258.

THE AMERICAN JOURNAL, OF ANATOMY, VOL. II, NO. 3


300


ALBERT KUNTZ


assochelys caretta during the twelfth or the thirteenth day of incubation, the vagus trunks appear as conspicuous fiber-bundles lying a little above the level of the trachea. In some sections short fibrous branches may be traced from the vagus trunks. Cells become separated from the vagus trunks and wandering out along these fibrous branches escape from their growing tips and become arranged in a broken ring encircling the oesophagus (fig. 10, M.S. P.). Cells maybe observed apparently wandering out from the vagus trunks also in sections in which fibrous branches


\6.


AW. '■»




  • Jf «Si»; V.





^ t


i!>C-i


x:/ r::rfel^


■i^;A^


> ^


■ ;^ • X'




i«'.


\ <» A-"»i'




J. ■«* ■ I / * ■ ■" ■-■I



Fig. 10 Transverse section through the oesophagus of a thirteen-day embryo of Thalassochelys caretta. X 140. M.S. P., anlagen of myenteric and submucous plexuses; Oe., oesophagus; 7'., trachea; Vag., vagus nerves.


are not apparent. At this stage small groups of sympathetic cells may be traced completely round the oesophagus dorsally, while on the ventral side similar cell-groups may be traced mesially from the vagus trunks into the area between the oesophagus and the trachea where they become lost in the deeply staining tissues in that area. The vagus trunks may be traced posteriorly nearly to the stomach, and as far as they may be traced


SYMPATHETIC SYSTEM IN TURTLES


301


cells wander out from them and become arranged in a broken ring encircling the oesophagus.

At the close of the seventeenth day of incubation the cell-aggregates in the walls of the oesophagus have become more conspicuous and the cells composing them more numerous. They are no






e ^ ^




>:^i* r " ® ^


      • ^


«> a





Fig. 11 Transverse section through the region of the heart in a seventeenday embryo of Thalassoehelys caretta. X HO. B., bronchi; C.P., anlage of cardiac plexus; D.PF. A., dorsal wall of atria; Mes., mesentery; M.S. P., anlagen of myenteric and submucous plexuses; Oe., oesophagus; Vag., vagus nerves.


longer arranged in a single broken ring, but the anlagen of the myenteric and the submucous plexuses are becoming more distinct (fig. 11, M. S. P.). Branches of the vagi may now be traced along the walls of the stomach. From these branches cells may be traced laterally in the walls of the stomach until the latter is completely encircled by groups of sympathetic cells.


302 ALBERT KUNTZ

The exact course of the development of the myenteric and the submucous plexuses in the walls of the intestine is not easily determined. Cells apparently migrate posteriorly in the walls of the intestine from the anlagen of the myenteric and the submucous plexuses in the anterior region of the digestive tube. On the other hand, as already indicated, cells migrate ventrally from the anlagen of the prevertebral plexuses in the posterior region of the body and become aggregated along the walls of the rectum. There is no evidence of the migration of cells from the anlagen of the prevertebral plexuses into the walls of the digestive tube farther anteriorly until the anlagen of the myenteric and the submucous plexuses have become well established. The evidence at hand seems to indicate that in the anterior region of the intestine the myenteric and the submucous plexuses arise from cells which migrate peripherally along the paths of the vagi, while in the posterior region at least some of the cells taking part in the development of the sympathetic plexuses in the walls of the intestine wander down from anlagen of the prevertebral plexuses.

c. Pulmonary plexuses. In transverse sections through the region of the lungs of embryos of Thalassochelys caretta at about the fifteenth day of incubation branches of the vagi appear to be closely associated with the bronchi as the latter enter the tissues of the lungs. Cells wander in along these branches and give rise to the anlagen of the pulmonary plexuses.

d. Cardiac plexus. In transverse sections through the anterior region of the heart of embryos at about the seventeenth day of incubation the mesocardium lies far to the left. In this region a fibrous branch accompanied by numerous cells may be traced ventrally from the left vagus trunk into the dorsal wall of the left atrium where many of the accompanying cells become aggregated to form the anlagen of the cardiac plexus in this region (fig. 12, C.P.). In transverse sections a little farther posteriorly in embryos of the same stage the mesocardium lies ventral to the oesophagus. In this region fibrous branches accompanied by numerous cells may be traced from both the right and the left vagus trunks into the dorsal wall of the heart where cells become aggregated to give rise to the anlagen of the cardiac plexus in this region (fig. 11, C.P.).


SYMPATHETIC SYSTEM IN TURTLES


303


e. Histogenetic relationships. The above observations on the development of the myenteric and the submucous plexuses, the pulmonary plexuses, and the cardiac plexus in embryos of the turtle agree essentially with the writer's observations on the development of these plexuses in embryos of mammals and of birds. These plexuses arise from cells which have their origin in the hindbrain and in the vagus ganglia and migrate peripherally along the paths of the vagi. In sections passing through the vagus rootlets in




^aU ^



Fig. 12 Transverse section through the anterior region of the heart in the same embryo as fig. 11. X 140. B., left bronchus; C.P., anlage of cardiac plexus; L.A., left atrium; L.Vag., left vagus nerve.


embryos at about the tenth or the eleventh day of incubation medullary cells may be traced from the walls of the hind-brain into the rootlets of the vagi (fig. 13, Vag. R.). That such cells wander into the vagus rootlets in considerable numbers cannot be doubted. In many sections medullary cells are drawn out into cone-shaped heaps in the vagus rootlets as they traverse the marginal veil. Occasionally one of these cells may be observed half in and half out of the external limiting membrane, while numerous cells are present in the vagus rootlets just outside the external


304


ALBERT KUNTZ


limiting membrane. The vagus ganglia at this stage appear as elongated cell-masses which are not sharply limited distally. Cells apparently become separated from their distal ends and migrate peripherally along the paths of the vagi. As far as the latter may be traced peripherally they are accompanied by numerous cells many of which become separated from the vagus trunks and become distributed in the walls of the digestive tube to give rise to the myenteric and the submucous plexuses, or wander into the anlagen of the other vagal sympathetic plexuses.



ElM. V^^.R


M. \4y.R.


Fig. 13 Sections through vagus rootlets in an eleven-day embryo taken at different levels. X 350. E.L.M., external limiting membrane; Vag.R., vagus rootlets.


It may be of interest to note at this point that embryos of the turtle afford exceedingly satisfactory preparations for the study of the development of the vagal sympathetic plexuses. The cells constituting the anlagen of these plexuses are exceedingly numerous and respond readily to differential stains. In well stained preparations the cells giving rise to these plexuses can be traced from the vagus trunks with such certainty that there can be no doubt as to their genetic relationship. Furthermore, cells cannot be traced into the anlagen of these plexuses, except in the posterior region of the intestine, from any other source until they have become well established. These observations do not preclude the possibility that a few cells may be transferred from


SYMPATHETIC SYSTEM IN TURTLES 305

the sympathetic trunks into the anlagen of these plexuses after the latter have become connected wih the former by sympathetic nerves. Such connections of the vagal sympathetic plexuses with the sympathetic trunks must, however, be looked upon as of only secondary importance in their development.

The cells which migrate peripherally from the walls of the hindbrain and from the vagus ganglia along the paths of the vagi in embryos of the turtle, like the cells which migrate peripherally from the cerebro-spinal nervous system in the trunk region, are characterized by very little cytoplasm and by large rounded or elongated nuclei showing a delicate chromatin structure. They are, therefore, cells of the same character; viz., the 'indifferent' cells of Schaper. Inasmuch as thses cells give rise to the vagal sympathetic plexuses, these plexuses also bear a direct genetic relationship to the cerebro-spinal nervous system. Mitotic figures occur frequently along the paths of the vagi and in the anlagen of the vagal sympathetic plexuses. We are not to suppose, therefore, that all the cells which take part in the development of the vagal sympathetic plexuses actually migrate as such from their sources in the hind-brain and in the vagus ganglia. As in the case of the sympathetic trunks, doubtless, many of these cells arise by the mitotic division of 'indifferent' cells along the course of migration.

DISCUSSION

The observations set forth in the preceding pages have shown that in turtles the sympathetic nervous system bears a direct genetic relationship to the central nervous system. The cells giving rise to the anlagen of the sympathetic trunks and of the prevertebral plexuses have their origin in the spinal ganglia or the neural crest and in the ventral part of the neural tube and migrate peripherally either through the mesenchyme or along the paths of the spinal nerves and of the communicating rami. The vagal sympathetic plexuses; viz., the cardiac plexus and the sympathetic plexuses in the walls of the visceral organs, are not derived from the same sources, but arise from cells which have their


306 ALBERT KUNTZ

origin in the hind-brain and in the vagus gangha and migrate peripherally along the paths of the vagi. These findings agree essentially, in regard to the sources of the cells giving rise to the sympathetic nervous system, with the writer's observations on the histogenesis of the sympathetic nervous system in mammals, birds, and fishes. They disagree widely with all the observations hitherto recorded on the development of the sympathetic nervous system in reptiles. They disagree also with the observations of the earlier investigators on the development of the sympathetic nervous system in the other classes of vertebrates primarily in two particulars: (1) some of the cells which enter the anlagen of the sympathetic trunks are found to have their origin in the ventral part of the neural tube and to wander out along the paths of the motor nerve-roots; (this fact was observed by Froriep ('07) in embryos of Torpedo and of the rabbit) ; (2) the vagal sympathetic plexuses are found to bear no direct genetic relationship to the sympathetic trunks, as the earlier investigators supposed, but arise from cells which have their origin in the hind-brain and in the vagus ganglia and migrate peripherally along the paths of the vagi.

The phenomena presented in embryos of the turtle may throw some new light on the problems involved in the peripheral migration of embryonic nervous elements.

Among the more recent investigators, Froriep ('07) has suggested that the nerve-fibers constitute the vehicles by means of which nervous elements are carried peripherally. According to his view the peripheral displacement of these cells is accomplished either by the growth of the axones alone or by the growth of the axones coupled with the active migration of the cells along the fibers. Held ('09) advanced the theory that the 'so called' wandering of the sympathetic anlagen is brought about by the pressure which is exerted by the mitotic division of proliferating elements in an elongating cell-column, coupled with the formation of peripheral protoplasmic processes which are endowed with the property of contractility and must, therefore, exert a pull in a longitudinal direction as soon as osmotic influences act upon their growing substance.


SYMPATHETIC SYSTEM IN TURTLES 307

In a recent paper on the development of the sympathetic nervous system in certain fishes,"* I have shown that neither of these theories are adequate to account for the phenomena observed. Furthermore, I have presented evidence in support of the view that the peripheral migration of cells from the cerebro-spinal nervous system into the sympathetic anlagen is probably determined by the influence of substances, hormones, which are produced by the cells in the richly nourished regions which are to become their ultimate destination.

As already indicated, before the spinal nerves may be traced peripherally in embryos of the turtle cells become separated from the distal ends of the spinal ganglia and migrate diagonally through the mesenchyme toward the lateral surfaces of the aorta. Fibers are not present along these paths nor are the cells closely aggregated . There seems to be no purely mechanical means, therefore, by which these cells could be carried forward or by which they could be guided in their course. During the later stages of development, as has also been shown, cells deviate from the course of the spinal nerves before reaching the origin of the communicating rami and advance by a more direct course toward the anlagen of the sympathetic trunks . In this case the cells are usually more or less closely aggregated, thus forming distinct cellular tracts which advance mesially from the spinal nerves. Mitotic figures occur along these tracts, but they are not sufficiently numerous to account either for the rapid increase in the number of cells present or for their advancement toward the sympathetic anlagen. We must conclude, therefore, that cells are actually displaced from the spinal nerves toward the sympathetic anlagen along these cellular tracts. In these later stages, as in the stages in which the spinal nerves can not yet be traced peripherally, there is apparently rtothing in the structure of the mesenchyme which might determine the course of the peripherally advancing nervous elements.

The phenomena above described seem to support the view that the peripherally advancing nervous elements are attracted toward the regions in which the sympathetic anlagen arise by the influ ^ Jour. Comp. Neur. Psych., vol. 21.


308 ALBERT KUNTZ

ence of substances, hormones, which are produced in those regions. Furthermore, the form-changes of the nuclei of the migrant cells indicate that these elements play more than a passive role in their peripheral advancement. All along the paths of migration, both in the nerve-trunks and in the cellular tracts passing through the mesenchyme, many of the nuclei are distinctly elongated, while in the sympathetic anlagen they resume a more rounded outline. Not infrequently these migrant cells present evidence of amoeboid movement. In some instances the nuclei are irregular in outline, while in still others they are distinctly pyriform with the broader end directed peripherally. These variations in the form of the nuclei of these migrant nervous elements, doubtless, indicate the presence of processes going on within the cells which play a part in their peripheral advancement and which are probably stimulated by the influence of the same agents which determine the direction of migration.

The phenomena observed in the development of the vagal sympathetic plexuses indicate that the peripheral migration of the cells giving rise to these plexuses is determined by the same influences which determine the peripheral migration of the cells giving rise to the sympathetic trunks. The cells which become distributed in the walls of the digestive tube to give rise to the anlagen of the myenteric and the submucous plexuses wander out from the vagus trunks and become aggregated into small cellgroups which are more or less closely associated with each other, but many of which, in the early stages, are quite free from nervefibers. The distribution of these cells cannot be explained by the purely mechanical processes involved in growth or by osmotic influences. If, however, we assume that the location of these cell-groups is determined by the influence of hormones which are produced by the cells in the walls of the digestive tube the problem becomes very simple. Likewise, the cells which wander into the walls of the heart to give rise to the cardiac plexus are not compactly aggregated in the early stages, nor are they always found in contact with nerve-fibers. Here again the problem becomes simple if we assume that sympathetic cells are attracted toward the heart by the influence of hormones which are produced in that region.


SYMPATHETIC SYSTEM IN TURTLES 309

It may be pointed out, furthermore, that the theory advanced above maybe applied to the peripheral migration of the cells giving rise to the vagal sympathetic plexuses as well as of the cells giving rise to the sympathetic trunks. In either case the direction of migration is toward a region in which there is an abundant food supply and which is the seat of primary vegetative processes. Inasmuch as the sympathetic nervous system is concerned primarily with the control of the purely vegetative functions we may suppose that the sympathetic elements respond primarily to the influence of hormones which are produced in these regions.

As has already been pointed out, the cells which migrate peripherally from the cerebro-spinal nervous system have the same genetic relationships as the cells which give rise to the neurones and to the neurogla cells in the central nervous system. The sympathetic nervous system is, therefore, homologous with the other functional divisions of the peripheral nervous system and the sympathetic neurones are homologous with their afferent and their efferent components.

SUMMARY

1. In embryos of the turtle the anlagen of the sympathetic trunks arise as cell-aggregates lying along the lateral surfaces of the aorta and along the dorsal surfaces of the carotid arteries. The cells which give rise to the anlagen of the sympathetic trunks have their origin (a) in the spinal ganglia or in the neural crest and (b) in the neural tube. Before the spinal nerves may be traced peripherally, cells advance from the distal ends of the spinal ganglia, directly through the mesenchyme, into the anlagen of the sympathetic trunks. After the spinal nerves have grown peripherally, cells migrate from the spinal ganglia and from the ventral part of the neural tube along the paths of the spinal nerves and of the communicating rami into the anlagen of the sympathetic trunks. These findings agree, in regard to the sources of the cells giving rise to the sympathetic trunks, with the writer's observations on the histogenesis of the sympathetic trunks in manamals, birds, and fishes. They disagree with the findings of the earlier investigators, except those of Froriep, primarily in the fact that


310 ALBERT KUNTZ

cells which wander out from the ventral part of the neural tube take part in the development of the sympathetic trunks.

2. About the eleventh day of incubation the anlagen of the sympathetic trunks begin to break up and to become more or less scattered. This scattering continues for a considerable period until the cell-groups again become aggregated into compact ganglia.

3. About the thirteenth day of incubation cell-strands push out from the spinal nerves proximal to the origin of the communicating rami and advance toward the aorta. These cell-strands increase in size and advance mesially until at the close of the sixteenth day they appear as irregular cellular tracts extending from the spinal nerves into the anlagen of the sympathetic trunks.

4. As development advances, the primary communicating rami are shifted proximally along the spinal nerve-trunks until they fuse with the cellular tracts extending from the proximal part of the spinal nerves into the anlagen of the sympathetic trunks.

5. A comparative study of the development of the sympathetic trunks in embryos of the turtle and in the chick strongly suggests a more or less direct phylogenetic relationship between the sympathetic nervous system in birds and in the ancestral type of reptiles.

6. The prevertebral plexuses arise as cell-aggregates lying along the ventro-lateral aspects of the aorta. They are derived from cells which migrate ventrally from the anlagen of the sympathetic trunks.

7. In the sacral region, cells may be traced ventrally from the anlagen of the prevertebral plexuses into the mesentery where they become aggregated into small cell-groups associated with the rectum. These sympathetic cell-groups probably represent the prototype of the ganglion of Remak in birds.

8. In the region of the genital ridges cells migrate ventrally from the anlagen of the prevertebral plexuses and become aggregated at the lateral surfaces of the former to give rise to the genital plexuses.

9. The vagal sympathetic plexuses; viz., the cardiac plexus and the sympathetic plexuses in the walls of the visceral organs, arise,


SYMPATHETIC SYSTEM IN TURTLES 311

not from cell which migrate ventrally from the sympathetic trunks, as earlier workers supposed, but from cells which have their origin in the hind-brain and in the vagus ganglia and migrate peripherally along the paths of the vagi. The results here recorded agree with the writer's observations on embryos of mammals, birds, and fishes.

10. The phenomena presented in embryos of the turtle afford evidence in favor of the view advanced by the writer in an earlier paper, according to which the peripheral displacement of the cells taking part in the development of the sympathetic nervous system is probably determined by the influence of hormones.

11. In turtles, as in the higher vertebrates, the cells which migrate peripherally from the cerebro-spinal nervous system into the sympathetic anlagen have the same genetic relationships as the cells which give rise to the neurones and to the the neuroglia cells in the central nervous system. The sympathetic nervous system is, therefore, homologous with the other functional divisions of the peripheral nervous system, and the sympathetic neurones are homologous with their afferent and their efferent components.


312 ALBERT KUNTZ

BIBLIOGRAPHY

Froriep, a. 1907 Die Entwickelung unci Bau des autonomen Nervensystems. Med. naturwiss. Archiv, vol. 1, pp. 301-321.

Held, H. 1909 Die Entwickelung des Nervengewebes bei den Wirbeltieren. Leipzig. Die Entstehung der sympathischen Nerven, pp. 212-242.

Hoffmann, C. K. 1890 Syinpathisches Nervensystem. Bronn's Klassen und Ordnungen des Thier-Reichs, pp. 1961-1962.

KuNTZ,'A. 1909 A contribution to the histogenesis of the sympathetic nervous system. Anat. Rec, vol. 3, pp. 158-165.

1909 The role of the vagi in the development of the sympathetic nervous system. Anatomischer Anzeiger, vol. 35, pp. 381-390.

1910 The development of the sympathetic nervous system in mammals. Jour. Comp. Neur. Psych., vol. 20, no. 3, pp. 211-258.

■ 1910 The development of the sympathetic nervous system in birds. Jour. Comp. Neur. Psych., vol. 20, no. 4, pp. 284-308.

1911 The development of the sympathetic nervous system in certain fishes. Jour. Comp. Neur. Psych., vol. 21.

Marcus, H. 1910 Beitrage zur Kenntnis der Gymnophionen. IV. Zur Entwickelungsgeschichte des Kopfes. II. Teil. Sympathicus, pp. 419424. Festschrift fiir R. Hertwig, Bd. 2 .1910.

Neumayer, L. 1906 Histogenese und Morphogenese des peripheren Nervensystems, der Spinaiganglien und des Nervus Sympathicus. Handbuch der vergl. und exper. Entwickelungslehre der wirbeltiere, pp. 513-626.

His, Jr. 1897 Ueber die Entwickelung des Bauchsympathicus beim Hiihnchen und Menschen. Archiv Anat. u. Entwg. Supplement.


THE DEVELOPMENT OF THE PARAPHYSIS AND PINEAL REGION IN REPTILIA

JOHN WARREN

From the Anatomical Department, Harvard Medical School

THIRTY-NINE FIGURES THIRTEEN PLATES

CONTENTS

Introduction 313

Description 314

Lacertilia 314

Chrysemys marginata 326

Discussion 332

Subdivisions of fore brain 332

Paraphysal arch and paraphysis 347

Velum and post velar arch 351

Choroid plexuses 357

Commissures 360

Superior commissure 362

Posterior commissure 363

Epiphysis and pineal eye 365

Summary and conclusion 368

Bibliography 371

INTRODUCTION

The term 'pineal region' is used here in the same sense in which it was introduced by Minot in the morphology of this region in Acanthias, and refers to those structures arising from the roof of the prosencephalon and diencephalon, particularly the paraphysis, velum transversum, epiphysis, superior and posterior commissures and the choroid plexuses. This paper is the fifth one on this region based on studies from material in the Harvard Embryological Collection and the observations were made

THE AMERICAN JOURNAL OP ANATOMY, VOL. 11, NO. 4 MAY, 1911

313


314 JOHN WARREN

on various forms of lizards, Lacerta muralis, agilis and viridis, and also on the turtle, Chrysemys marginata. The other papers were the following: 1. C. S. Minot. The Morphology of the Pineal Region based on its Development in Acanthias, 1901. 2. F. Dexter. The Development of the Paraphysis in the Common Fowl, 1902. 3. J.Warren. The Development of the Paraphysis and Pineal Region in Necturus maculatus, 1905. 4. R. J. Terry. The Morphology of the Pineal Region in Teleosts, 1910. As a great deal has been written on the details of this region, especially in the lacertilia, it seemed desirable to consider the growth of the structures arising from this part of the brain rather from the standpoint of topographical development, and with this end in view wax reconstructions of the fore and mid brain of each stage have been made. Owing to the amount of material available in the Harvard Embryological Collection, a more complete series of stages of the complicated development of this region can be shown than has been described heretofore. The reconstructions display the structure of this region better than ordinary sections and also demonstrate more clearly the topography of this part of the brain.

DESCRIPTION

Lacertilia

The observations were made on specimens of Lacerta muralis agihs, and viridis. The models are magnified 110 diameters.

Fig. 1 is a model of a part of the brain of an embiyo of Lacerta muralis of 1.8 mm. A gi'oove marks the limit between the prosencephalon, P., and the mesencephalon, M. The stalk of the optic vesicle appears on the lateral aspect of the prosencephalon and there is as yet no sign of any further subdivision of the fore brain.

Fig. 2 is a model of the left half of the brain of Lacerta muralis of 2 mm. Here a slight angle, V, is seen in the roof of the fore brain which is the anlage of the velum and marks the sub-division of the fore brain into telencephalon, T, and diencephalon, /. D and II. D. There is a sHght ridge passing downward from the


PARAPHYSIS AND PINEAL REGION IN REPTILIA 315

velum and ending behind the optic stalk, 0. S. That part of the brain cephalad to this ridge is the first segment of the fore brain or the telencephalic neuromere of Kupffer. The roof of this segment forms a slight arch in front of the velum. This is the beginning of the paraphysal arch, P. A., from which the paraphysis will develop at a later stage. The optic vesicle grows out from the ventro-lateral wall of this segment. Behind the velum the roof of the diencephalon forms a sHght curve ending at a small but very distinct arch, *S. This division of the roof, belongs to the second segment of the fore brain or the first diencephahc segment, /. D. The segment is limited cephalad by the ridge passing from the velum to the optic stalk, and caudad by a slight ridge passing from the cephalic end of the small arch mentioned above to the habenular angle, or flexure, H. F., at the base of the brain. This segment is Kupffer's (57) parencephalic neuromere, or parencephalon, and from its roof will develop later the post velar arch, epiphysal arch, and the superior commissure. Between this segment of the diencephalon and the anterior end of the mid brain we have a short segment, the roof of which forms the short, well-defined arch S. This subdivision lies between the ridge forming the caudal limit of the first diencephahc segment and the dorsal groove and lateral ridge marking the cephahc Hmit of the mid brain, and is the second diencephalic segment, Kupffer's (57) ' Synencephales Neuromer des Diencephalon,' or synencephalon, II.D. Its roof will form that part of the diencephalon lying between the epiphysis and the mid brain. The posterior commissure wih develop in the caudal end of this region, and eventually occupy all of it. The ridges which separate the three segments of the fore brain appear as grooves on the outer surface. At the present stage, therefore, we have the three sub-divisions fairly well defined, but of the structures that develop from the roof of each division we recognize only the anlage of the velum. It should be noted, however, that the roof of the second diencephalic segment at this stage is very well defined, and is therefore one of the earliest portions of the roof of the fore brain to be <?learly differentiated.


316 JOHN AVARREN

Fig. 3 is from a similar model of the brain of Laeerta muralis of 2.4 mm. The velum forms a more striking angle in the roof and is now prolonged into a very distinct ridge passing behind the optic stalk to a thickening in the floor of the diencephalon. The paraphysal arch, P. A., is now distinct. It curves forward and upward in front of the velum to form the roof of the telencephalon and passes into the upper end of the lamina terminalis, L. T., but there is no sign yet of the paraphysis. Behind the velum is the post velar arch, P. V. A., ending at another low arch or thickening in the roof, E. A., which is the beginning of the epiphysal arch. Both arches therefore develop from the roof of the first diencephalic or parencephalic segment, /. D. Behind the epiphj'sal arch is the last segment of the diencephalic roof, forming a distinct arch, S. The lateral ridges bounding this second diencephalic segment (synencephalon), //.Z)., are now more clearly shown than in the preceding stage. We have here a series of primary arches in the roof of the fore brain, each of which is developed from one of the three subdivisions of the fore brain; — the paraphysal arch from the telencephalic, the post velar arch and epiphyseal arch from the first diencephalic, and the intercalated or synencephalic arch from the second diencephalic segment.

Fig. 4 is a lateral view of fig. 3 and shows clearly the subdivisions of the fore brain and the grooves separating them that correspond to the internal ridges seen in fig. 3. At the ventro-lateral part of the telencephalon, T, is the optic stalk lying cephalad to the groove V, which corresponds to the velum and separates the telencephalon from the parencephalon of first diencephalic segment, I. D. A well defined groove separates the first diencephalic segment from the synencephalon or second diencephalic segment, //. D., which in turn is bounded caudad by another groove separating it from the mid brain. r^'

Fig. 5 shows the brain of Laeerta agilis of 2.4 mm. Here we have all the details of the previous stages rather more clearly marked, but the beginning of the pineal organ, E., is now seen as a slight diverticulum growing from the epiphysal arch, E. A., fig. 3. ' ,.

A .. /


PARAPHYSIS AND PINEAL REGION IN REPTILIA 317

Fig. 6 shows the brain of Lacerta muralis of 3.2 mm. In the paraphysal arch three small evaginations, P, have made their appearance, which represent the anlagen of the paraphysis. The velum is seen at the sharp angle immediately behind the third evagination, and extending backward from that point is the post velar arch, forming a well marked curve in the diencephalic roof. There are now two pineal evaginations with a sort of common opening into the diencephalon. The larger or cephalic outgrowth corresponds to that seen in fig. 5 and is the anlage of the future pineal eye, P. E. This will break away from the other vesicle and migrate eventually to the parietal foramen. The smaller or caudal outgrowth is really secondary and has developed immediately behind the former. This will be the future epiphysis, E., and will always be attached to the roof of the diencephalon. We have here, I think, without doubt two distinct vesicles, but they lie so close together that as they develop and enlarge the wall between them is pushed dorsally so that we have a common opening for the two into the diencephalon. Behind the epiphysis in this specimen the intercalated arch has been flattened out and is not well marked.

Fig. 7 shows the brain of Lacerta agilis of 3.6 mm. In this case there is only one paraphysal outgrowth and there is probably a considerable variation in the number of the primary paraphysal outgrowths in different forms and stages of lacerta. The velum appears as a sharp angle in the roof and is continued laterally into a well marked ridge projecting into the brain cavity. Anterior to this ridge the hemisphere is now very well developed, with the deep opening of the optic stalk just inferior to it. The post velar arch is essentially the same as in the preceding figure. The two outgrowths, the future pineal eye and the epiphysis, are seen from the outside and of the two the cephalic is much larger than the caudal. Their internal arrangement and opening into the diencephalon is practically the same as in fig. 6, the separation between the two vesicles being more marked, however.

Fig. 8 shows the brain of Lacerta muralis of 4.5 mm. Here again there are three well marked primary paraphysal outgrowths in essentially the saniQ position as in fig. 6. The velum appears


318 JOHN WARREN

as an angle in the roof, and behind it the post velar arch forms a very striking dome in the diencephalic roof. The separation between the two pineal outgrowths is more marked than in the preceding stages, and a median section would show that the future pineal eye is now nearly separated from the epiphysis. Caudad to it is the pars intercalaris (synencephalic arch), S, which has now been invaded in its caudal part by the beginning of the posterior commissure, P. C. It can be seen that this commissure lies almost entirely in this part of the diencephalic roof and does noi extend backward beyond the dividing line between diencephalon and mesencephalon. It would therefore appear as if this commissure belongs — at this early stage at least — to the diencephalon, and that it invades the mesencephalon only at a later period of development. At this stage we have several new structures developing from the primary arches in the roof of the three subdivisions of the fore brain: the paraphysis arising from the paraphysal arch in the roof of the telencephalic ; the epiphysis and pineal eye from the epiphysal arch in the roof of the first diencephalic; and the posterior commissure in a part of the roof of the second diencephalic segment. Further changes will consist in the appearance of the telencephalic or lateral choroid plexus from the telencephalic segment and the diencephalic plexus and superior commissure from the first diencephalic segment. The posterior commissure will eventually occupy all the roof of the second diencephalic segment.

Fig. 9 shows the brain of Lacerta muralis of 5 mm. A very marked increase in the development of all parts is apparent. The paraphysis is now a large vesicle with a wide mouth and at its apex are two distinct outgrowths while near its base, on the posterior side, is a third. Its general inclination is upward and backward, tending to follow the curve of the post velar arch. The velum is not as sharply marked as in the earlier stages and forms rather a wide angle. The post velar arch occupies the greater part of the roof of the diencephalon, forming a high, wide dome, the brain roof here being extremely thin and formed only by a single layer of cells. The pineal eye, P. E., is now a rounded vesicle, completely separated from the epiphysis, Ep., but is, how


PARAPHYSIS AND PINEAL REGION IN REPTILIA 319

ever, still in close contact with it and with the brain roof. The epiphysis itself is of oblong shape and its opening still communicates with the cavity of the diencephalon. The posterior commissure does not occupy all of the roof of the synencephalon and there is still a distinct interval between it and the epiphysis. The greater part of this tract, however, lies in the diencephalon. but its hinder end is beginning to overlap onto the mesencephalon and blend with the outer layers of its wall. In fact, from this stage on, the posterior limits of the commissure become more and more diffuse and are indicated arbitrarily in the following models. Fig. 10, Lacerta muralis of 17.5 mm. The growth of the paraphysis is very striking. It is now a large rounded tube ending in three distinct tubules and following the direction of the post velar arch, against which it is closely moulded. The opening of the organ is much smaller than in the previous stage and lies immediately anterior to the velum. It seems probable that all the primary anlagen — one, two or three, as the case may be — are eventually taken up and absorbed into one large tube. It is conceivable that the three terminal tubules here may represent three primary outgrowths. This must naturally be merely a matter of conjecture, and probably individual cases vary too much to make any such comparisons of real value. One does not see the velum becatise it is covered up by the anlage of the lateral choroid plexus, L. C. P. If the plexus were removed from the model, the velum would appear about as it does in fig, 9. This is the earliest stage in which I have been able to identify the plexuses of the lateral ventricles. They develop from the roof of the telencephalon anterior and lateral to the paraphysis and invaginate the dorso-mesial wall of the hemisphere, see fig. 25, Chrysemys marginata, where the relations are essentially the same. They show already a tendency to form little villus-like projections and are the first of the plexuses to make their appearance. The post velar arch has increased in height as well as antero-posteriorly, forming now a high dome-like roof to the diencephalon, with a deep recess at its postero-superior part. The shape and relative extent of the post velar portion of the diencephalon at this stage should be carefully noted. Until now


320 JOHN WARREN

this part of the roof has grown steadily upward and backward so as to form a large sac-like enlargement, and the distance between the velum and the opening of the epiphysis has continually increased. After this stage we shall find a tendency for these two points to approch each other, causing very striking alterations in the shape and relations of the various parts of the roof of the diencephalon. As an apparent result of this excessive enlargement of the hinder part of the post velar arch, the epiphysis has been pushed somewhat backward. It forms a long, hollow tube with still a very narrow communication with the cavity of the diencephalon. The pineal eye has shifted dorsally and is still almost in contact with the tip of the epiphysis, but is separated by a considerable interval from the top of the post velar arch. There is present a new structure — the superior commissure, S. C. — which appears considerably later than the posterior. It lies in its characteristic position immediately anterior to the opening of the epiphysis, which separates it from the posterior commissure. The latter is greatly increased in size and is in contact with the posterior aspect of the epiphysis, occupying now all of that segment of the diencephalic roof caudad to the epiphysis. It has also extended backward well into the mid brain and its hinder end is so blended with the outer layers of the wall that one cannot well give it any definite limit. Noteworthy is the marked increase in the size of the mid brain, the roof of which now tends to roll somewhat forward toward the epiphysis. As a result of this, and also of the backward development of the post velar arch, the posterior commissure is somewhat compressed, and an angle is formed at about its middle, dividing it in a general way, into an anterior part in relation to the diencephalon and a posterior part in relation to the mid brain.

Fig. 11 shows the brain of Lacerta muralis of 26 mm. Here the paraphysis is larger, its outline is rather irregular and it is growing farther backward toward the epiphysis and pineal eye. The relation of the organ to the surrounding veins is well shown in the model. A large vessel — the superior sagittal sinus, S. S. S. — lies between the two hemispheres. The paraphysis as it develops grows into the small veins which are the anlagen of this big


PARAPHYSIS AND PINEAL REGION IN REPTILIA 321

vessel and is at this stage almost completely surrounded by the sinus which has expanded and been more or less broken up to enclose it. This shows a very intimate relation between the paraphysis and these veins, the walls of the paraphysis lying in close contact with the cells in the walls of the veins. This big sinus is still further broken up into a very complicated net-work of smaller vessels over the top of the post velar arch which were so small and irregular that no attempt was made to model them. These vessels at a later stage share in the formation of the choroid plexus which develops by the infolding of the top of the arch. On reaching the epiphysis the vascular net-work surrounds that structure closely and behind it becomes concentrated into a large single median vessel overlying the mid brain. The epiphysis and pineal eye are practically the same as in figure 10, but their increased caudal inclination is probably due to some distortion of the model at this point. The recess in the upper dorsal aspect of the post velar arch is very marked above the superior commissure and is, I think, somewhat exaggerated by the distortion mentioned above. The posterior commissure is steadily increasing in size and the angle between its two parts is becoming more marked. The position of its anterior part and that of the superior commissure with reference to the habenular flexure at the base of the brain and also the position of the hinder end of the mid brain, M., should be noted and compared with the previous figure, fig. 10. There the two commissures and the opening of the epiphysis lie directly over this flexure, and the caudal limit of the mid brain is approximately on a plane with the optic commissure, 0. C. In fig. 11 the walls of the mid brain are greatly thickened, the cavity much reduced in size, and the whole of this part of the brain has shifted upward and forward so that its caudal limit is now on a plane much above the optic commissure. The anterior end of the posterior commissure and the superior commissure, together with the stalk of the epiphysis, are carried forward so as to lie over the hypophysis, H. This change in position of the mid brain has tended to compress the roof of the diencephalon, as shown by the appearance of a marked angle in the post velar arch. The choroid plexus of the lateral ventricle, L. C. P., is represented


322 JOHN WARREN

by a large irregular ingrowth from the mesial wall of the hemisphere, its long axis extending vertically. The direction of the hemisphere has also changed. In the previous figure it was approximately horizontal, while here it is tipped distinctly forward and downward.

Fig. 12. This model is of Lacerta viridis of about 37 mm. We notice at once a very great thickening in the brain walls and a corresponding reduction in size of the cavities. The most striking feature is the change in shape and size of the dorsal part of the diencephalon and in the relations between paraphysis, epiphysis and pineal eye. The paraphysis has elongated, but its diameter is much less than in the two previous stages and a number of small tubules have appeared especially at the tip which touches the tip of the epiphysis and has been thrust apparently in between the epiphysis and the pineal eve, the latter now resting on the upper third of the paraphysis. The change in relation between paraphysis, epiphysis and pineal eye is a result of the approximation of the anterior and posterior parts of the post velar arch following the continued shifting upward and forward for the mid brain. This portion of the brain has altered its position still further from that in the preceding stage. The caudal end of the mid brain is now about on a plane with the velum, V., and a line drawn between those points touches the lowest part of the posterior commissure and is practically horizontal. If this is compared with a similar line drawn from the caudal limit of the mid brain to the posterior commissure in fig. 10 one gets a good idea of the change in position of the mid brain. In fig. 10 such a line forms an angle of about 80° with the horizon, while in fig. 12 the line is about horizontal. Therefore the mid brain has shifted in a dorsal as well as in a cephalic direction, its walls at the same time have become greatly thickened and its cavity proportionately reduced. This forward development is indicated by the new position of the superior commissure and anterior end of the posterior commissure. These now lie directly over the optic commissure and have moved forward from the habenular flexure to this point during the last three stages. This process has approximated the anterior and posterior portions of the post


PARAPHYSIS AND PINEAL REGION IN REPTILIA 323

velar arch so that the arch itself has been reduced to a narrow pocket while its roof has been folded transversely. The transverse folds have been invaded by the overlying mass of vessels and we have now the first appearance of a choroid plexus developing from the diencephalon, D. C. P. This diencephalic plexus is therefore of later origin than the telencephalic, which first appeared at the stage shown in fig. 10, 17.5 mm. The epiphysis, like the paraphysis, has elongated considerably and lies close against the wall of the post velar arch. It is still hollow, but its cavity no longer communicates with that of the diencephalon. It is, however, attached to the roof of the brain by a solid constricted stalk which separates the two commissures. The superior commissure is rather larger than in the previous stage, while the posterior now consists of two very distinct parts, — the posterior being bent forward over the anterior at a very acute angle. This is also due to the forward development of the roof of the hind brain. In the lateral wall of the diencephalon is seen in section the so-called middle commissure, which contains a distinct cavity. There is also a deep pocket extending down into the floor of the brain immediately above the optic commissure. The velum transversum is clearly seen just behind the lower end of the paraphysis. It is much thickened and contains a band of commissural fibers passing between the mesial walls of each hemisphere which is the aberrant commissure described by Elliot Smith. The plexus of the lateral ventricles has become more convoluted, its greatest extent being still in the vertical diameter. The inclination of the cavity of the hemispheres downward and forward is much more marked than in earlier stages. This ventral bending, together with the peculiar development of the mid brain contributes to the closing up of the post velar arch and the reduction of the upper part of the diencephalon to a cavity very limited in its antero-posterior diameter but relatively broad transversely.

Fig. 13 is from a model of the pineal region of a brain of an adult Lacerta muralis. The lower outline of the model is very irregular as it was impossible to get all the sections into the field. The paraphysis is very striking, being a long tube rather narrow in its lower half but increasing in size in its upper half from which


324 JOHN WARREN

arise a large number of tubules that are very closely applied against the anterior wall of the post velar arch. The long axis of the tube has changed its direction and now curves upward and slightly forward, whereas in previous stages it curved upward and backward and had a very sharp angle at about the middle. The change in direction seems to be due to a backward growth of the hemispheres, which in turn have forced the paraphysis and the anterior wall of the post velar arch backward toward the posterior wall and have thus cooperated with the forward growth of the mid brain in reducing this post velar region of the diencephalon to a mere transverse slit. The upper part of the slit is filled by the folds of the diencephalic plexus, which protrude downward for a considerable extent into the cavity, fig. 15. This compression of the upper part of the diencephalon with the subsequent changes in the shape and relations of the parts is one of the most striking features of the development of this part of the lizard's brain. The pineal eye has now moved to its permanent place in the parietal foramen. While the tip of the paraphysis lies immediately below the vault of the skull the pineal eye is situated in its foramen at a considerable distance anterior to it. The migration of the eye covers quite an extensive field. It is formed at first from an outgrowth close to that of the epiphysis, both hollow and communicating by a sort of common opening with the diencephalic cavity. It then separates from the epiphysis forming a rounded hollow structure which rests against the epiphysis and on the roof of the diencephalon. It next moves away from the brain wall but remains in contact with the tip of the epiphysis until it is finally separated from it by the extremity of the paraphysis which is thrust in between them. Finally, after lying on the dorsal aspect of the paraphysis it is carried forward to its final resting place in the parietal foramen. During these moves its shape gradually changes from a rounded vesicle to a circular disk flattened from above downward. In only one stage have I been able to see anything of a nerve for the eye; that was in the brain of the 17.5 mm. embryo when a narrow band of tissue could be seen passing from the eye downward, anterior to the epiphysis. The staining of the specimens did not permit my following it to


PARAPHYSIS AND PINEAL REGION IN REPTILIA 325

its termination. The shape of the epiphysis has also materially changed. Its upper end has become much enlarged, is somewhat hammer-shaped and quite broad in the transverse diameter. It is barely separated from the tip of the paraphysis by the highest part of the roof of the diencephalon. The enlarged extremity is hollow, with relatively thick walls but the cavity soon terminates in a very narrow solid stalk which represents fully threefourths of its total length. The stalk is closely applied to the posterior wall of the post velar arch and becomes more and more attenuated as it approaches the brain roof. As the specimen was somewhat torn at this point, I am unable to state with certainty whether it is still attached between the two commissures. The wall of the epiphysis is rather thick and consists of several layers of cells with large round nuclei. The preservation of the specimen did not permit of making a drawing to show the characteristics of the cells here or in the wall of the paraphysis. The superior commissure is now a large rounded tract and is separated by a slight interval only from the anterior wall of the post velar arch lying about opposite its middle point. As regards its relation to the floor of the diencephalon it lies about over the optic commissure as in the preceding stage or if anything a trifle farther back. The increasing size of the hemispheres seems to have counteracted the forward push of the mid brain and the upper part of the diencephalon seems to be inclined relatively further back than in the last stage. This observation was made, however, on a specimen cut in the sagittal plane, which was not perfect and this statement possibly is open to some doubt. Elliot Smith's (88) aberrant commissure, A. C, is seen as a well rounded fiber tract lying in the original velum transversum. The tract probably forms a more prominent projection into the cavity, but again this part of the specimen was slightly damaged and the model here is not absolutely perfect. The choroid plexus of the lateral ventricle forms a very striking tuft in the ventricular cavity. From its origin in the roof of the telencephalon, immediately in front of and lateral to the opening of the paraphysis, the plexus extends in the form of a round band of tissue through the foramen of Munro into the lateral ventricle.


326 JOHN WARREN

The drawing has been shghtl}^ distorted in order to show this. Once in the lateral ventricle this narrow rounded mass expands into a flattened mass of tissue terminating in a complicated, tufted extremity. Its shape is quite characteristic, being longer from above downward as was the case in all its previous stages. This plexus together with that at the top of the diencephalon are the only forms of plexus present in the specimen studied. Fig. 14 is a transverse section along the line A-B, fig. 13, to show the tubules of the paraphysis. It is somewhat torn in the region of the superior commissure, the position of which only is indicated. Fig. 15 is a corresponding section along the line C-D to show a section of the epiphysis and the diencephalic plexus. It is also somewhat damaged in the same region as in the previous figure.

Chrysemys marginata

The models are enlarged only 80 diameters as the older stages were too large to draw conveniently at a greater magnification.

Fig. 16 shows the fore and mid brain of Chrysemys marginata of 3.2 mm. The line of demarcation between fore and mid brain is indicated by a groove on the dorsal, and a slight ridge on the inner side of the brain, similar to that in fig. 2. The roof of the fore brain forms a simple arch, there being no sign as yet of the velum. At its caudal end immediately anterior to the mid brain is seen a slight arch, >S, corresponding to a similar arch in fig. 2.

This marks the roof of the second diencephalic segment. In the lizard this appeared at about the same time as the velum and consequently all three subdivisions of the fore brain were developed together. Here there is no sign of any demarcation between the telencephalic and first diencephalic segment. Consequently the second diencephalic segment seems to antedate the appearance of the other two. Fig. 17 is from an embryo of 4.8 mm. The velum, V, is well marked and continued into a ridge passing down behind the optic stalk, thus separating the telencephalic, T., from the first diencephalic segment, /. D. Two well defined ridges also mark the boundaries of the second diencephalic seg


PARAPHYSIS AND PINEAL REGION IN REPTILIA 327

ment, //. D., at its dorsal side. It should be noted however that the second diencephalic segment is distinctly wedge shaped. There is a well defined arch in its roof with a cavity just below, but this cavity soon fades away into a thick ridge separating diencephalon from mesencephalon. This segment is not as well marked here as it was in the lizard, but all three segments are defined at this stage. Furthermore all the primary arches are present in the roof of the fore brain; a well marked paraphysal arch in the roof of the telencephalic segment ; a short post velar arch and a rather poorly developed epiphysal arch or thickening in the roof of the first diencephalic; and a low arch, the synencephalic, in the roof of the second diencephalic segment immediately anterior to the mid brain. This stage therefore corresponds closely to that in fig. 3, although the parts are not quite so sharply defined as in the lizard of 2,4 mm. The mid brain here seems to be divided into two segments, I.M., and II.M.

Fig. 18 shows a brain of an embryo of 5 mm. Here the velum, the paraphysal and post velar arches are all well marked. A small outgrowth extending upward and backward from the region of the rudimentary epiphysal arch, is the anlage of the epiphysis, E. Behind this is a well marked intercalated arch, S, with very definite boundaries below at the habenular flexure, although its anterior boundary is not very distinct in the lateral wall of the brain. This stage corresponds very closely to that of Lacerta muralis of 3.2 mm., fig. 5.

Fig. 19 shows the brain of an embryo of 7.6 mm. Immediately in front of the velum which forms a well marked angle in the roof is a single small outgrowth, the beginning of the paraphysis, P. Among all the specimens at my disposal I was unable to find any case of multiple paraphysal outgrowths which were quite common in the specimens of the lacertilia. A large vein is seen lying close to the paraphysis. The branches of this vessel form a sort of network about the paraphysis shov/ing that at the earliest possible stage the paraphysis enters into a close relation with the veins overlying this part of the brain which later concentrate to form the superior longitudinal sinus. Practically the same relations are seen in lacerta embrvos of a


328 JOHN WARREN

corresponding degree of development, namely from 3 to 4.5 mm. The post velar arch now forms a well arched roof to the diencephalon ending at the epiphysis which is a short oblong body opening by a very short hollow stalk into the diencephalon. It extends forward and lies directly on the top of the post velar arch. There is no sign now or at any later stage of any differentation of this body into a pineal eye and an epiphysis. Behind it is seen the last segment of the diencephalon, the intercalated arch, S, which has been invaded by the first traces of the posterior commissure, P.C., which is confined at this stage wholly to the diencephalon as was the case in the corresponding stage of Lacerta muralis, 4.5 mm., fig. 8. This stage corresponds to a combination of the stages of Lacerta agilis of 3.6 mm., fig. 6, and Lacerta muralis, 4.5 mm., fig. 8.

Fig. 20 shows the brain of an embryo of 9.5 mm. The paraphysis has now become a fairly long, round straight tube extending vertically upwards parallel to the anterior wall of the post velar arch. The close relations of the veins to it are clearly shown. The post velar arch has developed very rapidly forming a high vaulted roof to the diencephalon. Its walls, as was the case in lacerta, are very thin and are covered over by a very complicated net-work of small vessels, which will form later the diencephalic choroid plexus. The epiphysis has elongated and is much the same shape as the paraphysis, though smaller. It no longer opens into the diencephalon but is attached to it by a solid stalk and is closely surrounded by the venous plexus, Ve., overlying the post velar arch. The blood from this plexus, as shown in fig. 19 is carried off by several large veins passing downward on the lateral aspect of the diencephalon which empty into the primitive jugular vein. A similar arrangement is seen in Lacerta muralis of 28 mm. and also in earlier stages. On either side of the stalk of the epiphysis are seen the superior and posterior commissures. The former appears first at this stage and in fact can only just be seen in the specimen. It has been slightly exaggerated in the model. The posterior commissure is increased in size, and overlaps now slightly onto the wall of the mid brain. Its posterior limit is here quite clear and definitely placed, but later, as was the case


PARAPHYSIS AND PINEAL REGION IN REPTILIA 329

in lacerta, it becomes very diffusely expanded over the roof of the mid brain. The choroid plexus of the lateral ventricles, L. C. P., here makes its first appearance and consists of a few folds pushed into the ventricle from the mesial wall of the telencephalon. As was the case in lacerta this plexus antedates any of the others. The velum cannot be seen in the figure as the plexus is in the way. In both the lizard and the turtle the beginning of the plexus and the first trace of the superior commissure appear practically at the same time. This stage corresponds fairly closely to Lacerta muralis of 17.5 mm., fig. 10, in as much as both the lateral choroid plexuses and the superior commissure appear first in both. The shape of the post velar arches is essentially the same, though that of the turtle is more highly and uniformly arched.

Fig. 21 shows the brain of an embryo of 16.5 mm. The paraphysis forms a long narrow tube curving backward and resting closely against the post velar arch. There are no signs yet of any tubules and it is simply a hollow rounded tube with rather i hin walls. The velum is represented by merely a rounded angle concealed by the plexus of the lateral ventricle. The lateral plexus has become very complicated consisting of a mass of villuslike tufts forming a quadrilateral shaped mass of tissue. The post velar arch is more highly developed than before. Its outlines are very regular, the anterior and posterior walls being almost parallel with each other. It forms a very extensive space at the top of the diencephalon, in the upper part of which small ingrowths of the wall can be seen. These mark the beginning of the diencephalic plexus, D. C. P., which at a later stage wdll fill all this part of the diencephalon. The epiphysis like the paraphysis is a narrow elongated tube and is closely moulded to the posterior wall of the post velar arch. Its tip is much expanded laterally and the extremities of both these structures are gradually approaching each other. Beyond an increase in size there is nothing remarkable about the superior commissure. A slight recess can be seen in the top of the diencephalon just above it. The posterior commissure is still nearly flat. It forms a broad band of fibers reaching to the stalk of the epiphysis anteriorly and

THE AMERICAN JOTIRNAt, OF ANATOMY, VOL. 11, NO. 4


330 JOHN WARREN

mingling with the wall of the mid brain behind. Its posterior limit is here arbitrarily shown.The general shape of the cavity of the diencephalon has undergone a marked alteration, although the antero-posterior distance between epiphysis and paraphysis is essentially the same as in the previous stage. The lower or infundibular part of the diencephalic cavity has developed in a ventral direction forming a deep pocket, while the upper part of the cavity, bounded by the post velar arch, has developed in a dorsal direction forming a striking dome-like space at the top of this part of the brain. The hypophysis is not shown in this model.

Fig. 22 shows the brain of a turtle of 26 mm. In this model the upper half of the lateral choroid plexus, L. C. P., has been dissected away in order to give a clearer view of the paraphysis. The organ consists of a large central canal which gives off a complicated mass of tubules and is much larger than in the previous stage (see fig. 24). It curves backward over the post velar arch and its apex almost touches the tip of the epiphysis. Lying c^ver the top of both organs is the large superior longitudinal sinus. From this vessel small branches, Ve., pass down on all sides of the paraphysis and intermingle intimately with its tubules. In addition to these veins draining the organ from above another set passes downward from the lower part of the organ and disappears in the deep groove between diencephalon and telencephalon on both sides. The association between the walls of the tubules and of these veins is most intimate and we have here really a sort of sinusoidal circulation. The relations are better marked than in the lacerta where the development of tubules from the paraphysis is not so complex and approaches more nearly the conditions seen in amphibia. Fig. 27 shows the structure of the paraphysis and the relations of the vessels in a specimen 32 mm. in length. The drawing is slightly diagrammatic. The plexus of the lateral ventricles has increased greatly in size and complexity and its long axis is about horizontal. At the hinder border of the mass is seen a tuft of plexus, extending backward into the diencephalon with a curious prolongation stickin up from it (see also fig. 26, T. C. P.). This tuft springs from the ori


PARAPHYSIS AND PINEAL REGION IN REPTILIA 331

gin of the lateral plexus on each side of the opening of theparaphysis. This paired plexus is presumably homologous with the single plexus of amphibians which, arising from the telencephalon immediately anterior to the paraphysis, fills up the lower part of the diencephalon and is known as the telencephalic choroid plexus or plexus inferioris. Here it does not extend so far into the diencephalon as is the case in Necturus for example, Warren (97, fig. 16).

That part of the cavity of the diencephalon the roof of which is formed by the post velar arch is now filled up by a large mass of choroid plexus, D. C. P. The plexus begins close to the opening of the paraphysis but reaches its greatest development at the top of the diencephalon. In the model one sees a large rounded mass of plexus just below the roof. This is really the plexus of the right side of the brain and a corresponding mass is hidden from view on the opposite side. The diencephalic plexus therefore tends to be paired or double as was the case with the telencephalic plexus mentioned above, fig. 26, D.C.P. This is also in striking contrast, as was the case with the telencephalic plexus just mentioned, to the amphibian type where its homologue is represented by a single median mass, Warren (97, figs. 11, 13, 14). The epiphysis has increased in length as well as in breadth especially at the tip where it is quite expanded laterally. It is still a hollow tube, though attached to the brain by a solid stalk. Like the paraphysis it is closely moulded over the roof of the diencephalon. There is not much to add about the commissures. They are naturally larger especially the posterior which now shows a slight angle on its dorsal surface. Its posterior limit, as here marked, is purely arbitrary, since the fibers are spread out diffusely over the outer layers of the mid brain wall. The outline of this model is not quite correct in as much as the posterior half of the model is carried too high above the anterior. Figs. 24, 25 and 26 are three sections of this model to illustrate the relations of the parts which cannot be seen clearly in the model. They correspond to the lines A-B, C-D, E-F, fig. 22. Figs. 25 and 26 show how the cephalic wall of the dorsal sack has been carried forward on either side of the median line so as to form two recesses,


332 JOHN WARREN

the walls of which are convoluted and enclose the paraphysis between them.

Fig. 23 shows the pineal region only of the brain of a turtle 32 mm., at a magnification of 110 diameters, in order to compare this region more exactly with the larger lizards, as it is from the largest turtle at my disposal. The paraphysis has a narrow central cavity from which many lateral tubules are given off, the whole organ now appearing as a very complicated glandular structure. The walls of these tubules are closely related to vessels as mentioned above, fig. 27. The walls consist of a single laj^er of cells with large rounded or oval nuclei. The cell boundaries are indistinct. The endothelial cells in the walls of the vessels lie directly against the cells in the wall of the paraphysis. We have here a sinusoidal arrangement very similar to that of amphibia, Warren (97, fig. 20), but not so clearly developed. The diencephalic plexus is essentially the same as in the previous stage, though considerably more developed. The right half has been mostly removed to give a better view of the left half. The same is also true for the telencephalic and lateral plexuses which however were not modeled. The epiphysis is much expanded towards the apex and its outline a good deal distorted. It still contains a cavity partially interrupted at intervals by incomplete septa and has a solid hollow stalk. In structure it resembles the epiphysis in the lizard. The commissures which are not shown are the same as in the previous stages. In no specimen was there any trace of Elliot Smith's commissure which was so striking in Lacerta muralis. These last two stages are rather difficult to compare with those of the older lizards. The older stage of Chrysemys is probably very close to the final adult type. They probably would fit in between the two oldest lizards, though as regards the paraphysis and especially the plexuses the oldest turtles studied show a development which seems more advanced and complex than the adult lizard of fig. 13.


PARAPHYSIS AND PINEAL REGION IN REPTILIA 333

DISCUSSION

Subdivision of the fore brain

The question of segmentation of the neural tube has been one of great interest for many investigators and much has been written on the segmentation both of the medullary groove and of the neural tube after its closure. The earlier writers confined their attention chiefly to the hind brain and the medulla where the neural segments appear to best advantage. Von Baer ('28) observed folds on the lateral wall of the medulla in the chick. Bischoff C45), Remak C50-'55), Dursy ('69), Dohrn ('75) saw these folds in the medulla of the dog, chick, cow and bony fish embryos. Beraneck ('84) studied those segments with reference to their relations to the cranial nerves in lacerta, and he (7) and Prenant (83) in the chick also. Kupffer (57) found five pairs of segments in the medulla and three in the mid brain of trout embryos. He stated that he could then find no segments in front of the mid brain. He found later in Salamandra atra a regular segmentation of the wide open neural plate and counted eight pairs of neuromeres. He showed therefore that there was present at this early stage an ' ontogenetisch-primare neuromerie' extending throughout the brain plate. He was unable to state definitely how many segments belonged to the fore and mid brain as the line of demarcation between these regions was ill-defined at this early stage.

Orr (77) studied the segmentation in the lizard on the closed brain tube and applied the term neuromere to the ' ' replis medullaires" of Beraneck. In his earliest stages the three brain vesicles fore, mid and hind brain were formed. He found that the optic vesicles developed from the lateral wall of the extreme anterior part of the primary fore brain so that the anterior walls of the optic stalks were in the same plane with the anterior surface of the fore brain. The hind brain neuromeres were well marked and their histological structure quite characteristic. Each one was separated from its neighbor by an internal ridge and an external groove and each pair were placed exactly opposite each other. The cells were elongated and placed radially to the inner curved


334 JOHN WARREN

surface of the neuromere while the nuclei lay nearer the outer surface and approached the inner surface only at each end. The cells of one did not pass over into an adjacent neuromere and the cells between each pair were so crowded together that they formed a sort of septum between each neuromere. The five hind brain neuromeres all showed these features. He regarded the mid brain as one neuromere and between this and the secondary fore brain were two whose structure however was not quite the same as those in the mid brain and therefore he did not regard them as true neuromeres. His fig. 6, pi. 12, and fig. 40, pi. 15, show the secondary fore brain or telencephalon; then follow two well marked neuromeres and behind them the mid brain. These figures are essentially the same as figs. 29 and 31 and 37. Orr's fig. 63, pi. 16, shows the fore brain in sagittal section and corresponds to fig. 8, Lacerta muralis, 4.5 mm. Here the posterior commissure is seen occupying a part of the intercalated portion of the diencephalic roof. Hoffmann (50) as a result of observations on reptilian embryos agrees essentially with Orr,

McClure (66, 67) studied Amblystoma, lizard and chick. He divided neuromeres into myelomeres or constrictions of the myelon and encephalomeres or constrictions of the encephalon. He found in the hind brain five encephalomeres in Amblystoma and six in the chick. In the mid brain there were two. To these he gave the name of oculomotor and trochlear neuromeres. In the fore brain he saw two, which he named the olfactory and optic neuromeres and also traces of a third. He observed a similar condition in the newt and chick, and states that these forebrain neuromeres are true neuromeres as far as their external character and structure are concerned. As regards the third fore brain neuromere this appeared just in front of the mid brain It was smaller than the others and McClure expressed doubts about its neuromeric value. This segment is doubtless the synencephalic neuromere of KupfTer. McClure's (67) fig. 9 shows the condition in the chick and corresponds to figs. 29, 31, 35 and 37. His II N. M. is the second diencephalic, and the I A^. M. the first diencephalic neuromere as shown in my models figs. 3, 4, 17, and 18, and in the figures just mentioned.


PARAPHYSIS AND PINEAL REGION IN REPTILIA 335

In McClure's fig. 8a there is seen the trace of a tMrd diencecephahc neuromere on one side only as the section is quite oblique. I have found nothing to compare with it, my results corresponding with his fig. 9. The optic neuromere has no nerve connected with it as the optic nerve is secondary in character and is not reckoned as one of the segmental nerves. He suggests that the primitive segmental nerve which belongs here has degenerated. The olfactory nerve of course belongs to the first neuromere. He concludes that the encephalomeres are not only remnants of neural segments similar to the myelomeres, but that they were originally continuous."

Waters (98) made his investigations on the cod and Amblystoma. His results in the latter were more satisfactory than in the former. He found here three neuromeres in the fore brain and two in the mid brain. Zimmermann (101) studied Mustelus, chick and rabbit embryos. He found eight neuromeres at the time of closure of the neural tube. The three anterior corresponded to the fore, mid and hind brain vesicles. The five posterior belonged to the medulla. The fore brain divided secondarily into two, the mid brain into three and the hind brain into three neuromeres. He gave to these secondary subdivisions full metameric value. His observations on the fore brain and mid brain therefore gave five neuromeres to these regions which corresponds to McClure, Waters and von Kupffer.

Herrick (39) studied the brain of the snake, Eutaenia. His fig. 8, pi. 19, shows a horizontal section through telencephalon, two diencephalic neuromeres and a part of the mid brain. Compare this with McClure, fig. 9, Orr, fig. 6 and 40 and my figs. 29, 31, 33, 35 and 37, Herrick's figs. 6 and 7 sagittal show the three fore brain subdivisions as they appear in my figs. 7 and 18. He criticizes Waters, Zimmermann and McClure for having tried to homologize dorsal expansions in the fore brain with ventral expansions in the mid and hind brains. If neuromeres once existed in the fore brain they would be only visible at an early stage and would be obscured by altered conditions. The so-called fore brain neuromeres differ from those in the medulla and cord in involving only


336 JOHN WARREN

dorsal structures. They are wholly illusory from a morphological standpoint."

Froriep (31) in mole embryos of 5.5 mm., which corresponded about to a four weeks human embryo, found two neuromeresin the ' Zwischenhirn' and three in the mid brain. In Salamandra maculosa and Triton cristatus he observed the segments seen earlier by Kupffer but differed from that author as to their number. This author is rather sceptical about the morphological value of neuromeres and is inclined to regard them as results of mechanical pressure from the mesoderm.

Locy (63, 65) in Acanthias observed a very early segmentation of the primitive neural plate upon closure of the neural tube, which appeared along the edges of the plate. He stated that these segments could be traced without interruption through all stages of the open neural plate into those structures that in later periods were called neuromeres. This segmentation extended through all of the brain. Locy found two mid brain segments and three fore brain segments. The fore brain neuromeres were first the olfactory, and second the optic. The optic vesicles appeared ver}^ early on the neural plate and after the neural tube closed they seemed to belong especially to the parencephalic segment of the diencephalon. As regards the third neuromere he thought that its nerve was possibly the nerve to the pineal sense organ as sense organs and cranial nerves undoubtedly at first had definite segmental relations in neural segments. This condition persists in the cord but in the brain the primitive relations are greatly modified or obliterated. He concluded that neuromeric segmentation appears long before there is any segmentation in the mesoderm and therefore it is more primitive than mesodermic segmentation. The various segments are serially homologous and related to cranial and spinal nerves. The segments undergo most modification in the fore brain and mid brain where they disappear first.

Neal (73, 74) made his observations on embryos of Acanthias. He did not agree with Locy that the segments found on the edge of the open neural plate were true neuromeres, because of their irregularity and variation both in size and number. He believed


PARAPH YSIS AND PINEAL REGION IN REPTILIA 337

hat they were due to prohferation of cells along the edges of the plate. He found however in early stages a continuous primitive segmentation of the nervous sj^stem serially homologous throughout head and trunk — the neuromeric segmentation. In later stages there appears in the encephalon a secondary (in time) segmentation resulting in the so-called vesicles, which are not serially homologous with the segments of the myelon, but give rise to an anterior cephalic tract which is a region sui generis." He gives two tables showing the number of segments determined by previous investigators and their relation to the vesicles of the brain and to the nerves. He agrees with Orr as to the structure and number of hind brain neuromeres, finding five. As regards the neuromeres in the mid brain and fore brain there is a considerable difference of opinion and Neal thinks that most authors have counted dorsal expansions here which are really secondary subdivisions. He says that Morphologically different structures have been described by them as neuromeres or encephalomeres and that the divergence in their results does not seem to justify this assumption." As hind brain neuromeres involve dorsal, lateral and ventral zones, fore brain neuromes should do the same if they are morphologically equivalent. If they do not then one should be able to explain how that condition has been lost or modified. At an early stage, (74) fig. 45, pi. 7, he finds six vesicles in a parasagittal section of the cephalic plate. I = fore brain in region of the optic vesicles. 11= the mid brain. Ill = 'Hinterhirn' and IV, V, and VI are hind brain neuromeres.

Neal's fig. 47, a parasagittal section of a slightly older stage, shows five expansions in the fore and mid brain. I = prosencephalon and II = Kupffer's parencephalon and supports the epiphysis. The mid brain shows three exjDansions. I = that part of brain which later carries the posterior commissure. It seems to me that Neal should classify this first mid brain segment as the second diencephalic segment or synencephalon of Kupffer. This picture then would correspond to my figs. 28, 30, 32, 34 and 36. Fig. 52, a frontal section of fig. 47, shows these divisions well and corresponds to my figs. 29, 31 33, 35 and 37. According to my interpretation we would have here then three fore brain and two mid


338 JOHN WARREN

brain segments. Neal says his first three vesicles correspond to Zimmermann's fore, mid and hind brain. However, Zimmerman's seventh and eighth hind brain segments were not at first developed and so Neal concludes that there are six primary 'encephalomeres/ not eight. The divisions of the primary fore brain into 'Secundare Vorderhirn' and ' Zwischenhirn ' Neal does not consider as morphologically equivalent to neuromeres because as Herrick stated they are really dorsal expansions and should not be compared with ventral expansions. Neal agrees with Zimmermann that the primary mid brain divides into three segments of which the anterior lies in front of the posterior commissure. But according to my figures this should go with the fore brain to form its synencephalic segment (Kupffer). Neal concludes that structures of different morphological value have been described as neuromeres in the brain in front of the cerebellum. These structures are really secondary subdivisions which differ from typical neuromeres in shape, size, time of appearance and relation to dorsal and ventral zones. He thinks it best to regard each of the primary fore brain and mid brain vesicles, neuromeres I and II, as being serially homologous with the hind brain neuromeres III-IV. He feels however that on basis of structure and relation to other segmentally arranged organs that the primary vesicles, fore brain and mid brain give evidence — as do the primary expansions of the hind brain — ^of the primitive segmentation of the vertebrate head." Hill (42, 43) worked under Locy on Salmo purpuratus studying both dead and living specimens. He found like Locy eleven segments separated by grooves running around the whole brain. Of these five were found in the mid brain and fore brain. These segments he called the primary neuromeres and they antedated the three fore, mid and hind brain vesicles. They all had the typical characteristics described by Orr and were the same as those in the medulla. Essentially the same results were seen in young chicks. Three segments in the fore and two in the mid brain could be seen on the open and later closed tube. Later after the vesicles appeared the outlines of the primary neuromeres disappeared in the anterior part of the brain first and secondary subdivisions appeared in the fore brain, but Hill did not think


PARAPHYSIS AND PINEAL REGION IN REPTILIA 339

that these had the same morphological value as the early segments owing to their late appearance and their dorsal lines of division.

Hill's (42) fig. 3 shows the line between the first and second neuromeres splitting to surround the optic vesicle. In fig. 10 the optic vesicle is clearly assigned to the second neuromere. Figs. 40 and 41 show the three secondary subdivisions of the fore brain similar to my model, fig. 3. In fig. 40 the optic vesicle seems to belong to the first segment or the telencephalon. Fig. 42 is a parasagittal section of Salmo purpuratus showing five segments to fore brain and very similar to my figs. 28, 30, 32, 34 and 36. The third segment is wedge shaped, broad above, but narrow below at the habenular angle.

Weber (99) studied the segmentation of the brain in the pheasant shortly before and after the closure of the neural tube and made models of various stages. In his younger stages there were four neuromeres in the fore brain. The first is the lobus olfactorius impar, the second is the telencephalon. On its roof are the hemispheres, on its lateral walls the optic vesicles and in the floor the saccus vasculosus. The third is Kupffer's parencephalon with which the eye communicates and is broad dorsally and narrow below at the tuberculum posterius as in Hill's fig. 42. This one has the epiphysis on its roof and the recessus mammillaris in the floor. The fourth he calls the diencephalon" bounded below by the interpeduncular eminence.

Weber finds that these segments are clearly marked off by dorsal and ventral folds and he assigns to them a metameric value. His figs. 4, 5, 10, and 11 show these four neuromeres in the pheasant. In my models Weber's I and II form the first neuromere with the optic vesicle arising from it. I have been unable to detect any sign of subdivision in this first segment such as he shows. In the chick he finds five neuromeres in the fore brain (see text figs. 5 and 6).

Mrs. Gage (33) in a study of a three weeks old human embryo describes certain total folds in the fore brain radiating downward from the membranous roof along the lateral wall. These do not seem to correspond to any of the folds seen in the turtle or lizard.


340 JOHN WARREN

Johnson (52) discusses the question of neuromeres agreeing with the results of Hill and Locy and (53) he considers at length the morphology of the fore brain. He lays great stress on the shifting of the optic stalk from the primitive optic groove or infundibular recess to the preoptic recess. He assigns the optic vesicles to the diencephalon and states that the boundary between diencephalon and telencephalon is the velum and a ridge extending from the velum to the floor of the brain and passing in front of the neuromere to which the optic vesicle belongs.

Johnson's (53) fig. 42 shows parasagittal sections of the brain of a 7 mm. pig, fig. 34, a model of the brain of a 5 mm., and fig. 35 parasagittal sections of the brain of a 6 mm. pig. He finds three neuromeres in the fore brain and two in the mid brain. The first segment is the telencephalon, and the second is the optic vesicle, and the third lies behind the vesicle and supports later the paraphysis. I am unable to observe in my models this shifting of the optic vesicle and from the earliest stage I have studied it seems always to be in front of the point where the optic chiasma appears. Furthermore, the ridge which continues the velum is seen in figs. 2 and 17 appearing thus very early and passes clearly behind the optic stalk to end in the brain floor at the chiasma. This relation is also shown in Johnson's fig. 34, the opening of the optic stalk apparently lying in front of this ridge. I should therefore include the optic vesicles in the telencephalon and count his first two neuromeres as one segment. The second segment would lie behind this ridge. In a pig of 7.5 mm. one can find a trace of a third subdivision of the fore brain ; but this is seen much better at a more advanced stage. Fig. 34, a pig of 10 mm. shows a parasagittal section with three distinct segments in the fore brain and two in the mid brain. This third segment is the synencephalon, while the second is the parencephalon and produces the epiphysis. Fig. 35 is a horizontal section of the same stage and clearly shows these three subdivisions. They are also shown, I think, in Johnson's fig. 41 (central section) of a 15 mm. pig. The first segment in this figure is the telencephalon. Then comes a rather long first diencephalic segment followed by a short second diencephalic segment in front of the mid brain. This


PARAPH YSIS AND PINEAL REGION IN REPTILIA 341

third fore brain segment appears here relatively much later than in the lizard and turtle where it was one of the first structures to appear in the roof of the fore brain.

In my earliest models the fore and mid brain vesicles are clearly shown but there is no trace of any neuromeres in this region. In the hind brain of these specimens not shown in the model, the neuromeres were well marked. The first sign of subdivision of the fore brain of the lizard is shown in fig. 2, where there is just the beginning of the velum separating telencephalon from diencephalon and in the roof of the latter division is seen the short arch, S, lying just anterior to the mid brain and forming the roof of the synencephalic segment. In the turtle fig. 16 the velum has not appeared and the arch forming the roof of the synencephalic segment alone is seen. The features are shown more clearly in the next models, figs. 2, 3, 4, 17 and 18. Here the internal ridge continuous with the velum marks the caudal limit of the telencephalon and ends in the floor just behind the optic stalk. In fact it forms a distinct dorsal boundary to the optic stalk. This seems to me to assign the optic vesicle clearly to the telencephalon and not to the diencephalon as stated by others (see Hill (42), Kupffer (57), and Johnson (53)).

In more advanced stages, figs. 5, 7, 8, 18 and 19, the hemispheres are seen gradually bulging outward above the optic stalk. There appears now another ridge above and in front of the optic stalk, and apparently continuous above the velum. This might seem to mean that the optic vesicles did not belong to the telencephalon but lay behind its caudal limit in the front part of the diencephalon. This new ridge is, however, secondary and is due to the deepening of the hemisphere and also of the optic stalk. The ridge first mentioned and shown in figs. 2, 3, and 17 is really the primary line of separation between telencephalon and diencephalon and the optic stalk lies clearly in front of it. This first subdivision of the primary fore brain is the telencephalon and contains the eye vesicles. In its roof appears the paraphysal arch and later the paraphysis and the telencephalic plexuses. Its caudal limit is the velum and the internal ridge continuing the velum, which ends in the floor just behind the optic recess at the


342 JOHN WARREN

point where the optic commissure develops. The diencephalon in early stages shows signs of divisions into two segments. In figs. 2 and 17, the differentiation is best marked in the roof. The first segment is limited in front by the velum and the ridge continuous with the velum. In the floor appears the anlage of the optic chiasma, the infundibular recess and the mammillary region. Its caudal limit is the tuberculum posterius. The roof in these early stages forms a low arch. From this will develop the post velar arch, the diencephalic plexus, the supra commissure and the epiphysis, the latter forming the caudal limit in the roof. This segment is the parencephalon the first diencephalic neuromere or segment. The second segment is narrower and somewhat wedge shaped. Its roof forms a short arch ending behind at a groove marking the cephalic limit of the mid brain. This part of the roof becomes the pars intercalaris or synencephalon and in it appears the posterior commissure. Later the hinder part of the commissure extends backwards into the mid brain, but at first it seems to be wholly confined to this segment, figs. 7 and 19. Below, the second fore brain segment is narrow and its lower boundary extends from the tuberculum posterius to the highest part of the habenular arch, figs. 2, 3, and 4. In the turtle the segment at first has a cavity in its upper part only, which ends below in the thickened ridge separating diencephalon from mesencephalon, (see fig. 17) and is poorly marked at this early stage, but later, however, figs. 18 and 19, it becomes broader both dorsally and ventrally. In both lizard and turtle its floor lies between the tuberculum posterius and the apex of the habenular flexure. Its caudal and cephalic limits are formed by slight ridges, but its cephalic limit on the inner aspect of the brain is not very distinct.

Seen from the external side, fig. 4, the ridge separating telencephalon from diencephalon appears as a groove passing behind the optic stalk. Behind this is a swelling best marked towards the dorsal aspect of the brain which is the first diencephalic segment or parencephalon, /. D. Behind this comes a smaller swelling separated by a slight groove from the first. This is the second diencephalic segment or synencephalon, //. D. Compare


PARAPHYSIS AND PINEAL REGION IN REPTILIA 343

Kupffer (57), fig. 185; Nectums, 244, Anguis fragilis; and Hill (42) figs. 40 and 41, chick. It is best marked near the dorsal side of the brain while below its limits become less distinct. Its caudal limit is the groove separating it from the mid brain, M. This segment has been described by von Kupffer as the 'Schalthirn,' synencephalon, synencephalic neuromere and pars intercalaris. It forms a narrow though a well marked segment immediately cephalad to the mid brain and above the habenular flexure. It is a sort of isthmus between mesencephalon and diencephalon but belongs to the latter. Its roof is occupied by the anlage of the posterior commissure which appears here before spreading back into the mid brain and finally covers in all this part. As the mid brain enlarges on its dorsal aspect this part becomes much compressed and in later stages its identity seems to become in many cases practically lost. It is very constant in the vertebrate series (Burckhardt), but is not found in Accipenser according to Kupffer. Its early appearance in the lizard and turtle is very striking, being here differentiated almost before the velum or any of the other arches in the roof of the fore brain. In the chick it also makes its appearance at an early stage with the velum, Kupffer (57), fig. 277, chick of 30 somites. In the pig I have seen it as a trace only at 7.5 mm., but well marked from 8-10 mm., figs. 34 and 35. This is true also of the sheep, it being seen first here at about 8-9 mm., figs 36 and 37. In these mammalian forms it appears relatively late after the velum, paraphysal and post velar arches are well formed and the boundary between telencephalon and diencephalon clearly defined. In amphibia, according to Kupffer, it is best marked, persists in the adult brain as a distinct interval and is not wholly taken up by the posterior commissure. Burckhardt (13) gives diagrams of various forms of vertebrate brains. These diagrams represent sagittal sections of Amphioxus, Petromyzon, sturgeon, trout, Notidanus, Protopterus, Ichthyophis, Anguis, crow and man. In all of these the roof of the ' Schaltsttick' or synencephalon is clearly shown though varying in length in different cases (see also Terry (102) for teleosts).

Figs. 28 and 30 are parasagittal sections of the lizard of the same stage as figs. 3 and 4, and of the turtle of the same stage as


344 JOHN WARREN

fig. 18. Here one sees clearly these three fore brain segments or neuromeres, T, I. D, II. D. Figs. 32, 34 and 36 show a similar condition in the snake, pig and sheep and are similar to KupfTer (57) fig. 89, Neal (74) fig. 47, Acanthias, and Hill (42) fig. 42, Salmo (see also Neumayer (75) figs. 6 and 7 for the sheep).

Figs. 29, 31, 33, 35 and 37 are horizontal sections along the lines A-B shown in the above parasagittal sections to show these fore and mid brain segments in this plane. Compare with these Orr (77) figs. 6 and 40, Uzard; McClure (67) fig. 9, chick; Herrick (39) fig. 8, Eutaenia; Johnson (53) fig. 41, pig.

The models, figs. 2-7 and 17-19 give a sagittal view of these structures and may be compared with the following figures in sagittal planes, Neal (74) figs. 19, 20, Acanthias; Kupffer (57) figs. 178-181, Salamandra, 215 Rana fusca, 239 Lacerta viridis, 240 Anguis fragilis, 246 Lacerta vivipara, 277, 278, 286, chick; Herrick (39) figs. 6 and 7, Eutaenia; Ziehen (100) figs. 47, rabbit, and 66, sheep.

As regards the number of mid brain segments, these are rather apart from the subject of this paper which concerns chiefly the fore brain. In the early models the mid brain has appeared as a single simple segment. In fig. 17 turtle there seems to be a line of subdivision into two segments. In the reptilia Orr found only one mid brain neuromere while McClure found two. I should think that if enough early stages were examined there might be two segments or neuromeres here. Herrick in Eutaenia shows the mid brain as a single vesicle. This seems to be the case in figs. 32 and 33, though there seems to be a hint here of a subdivision in the mid brain. Figs. 34 and 35 of a pig show two distinct segments in the mid brain, as Johnson (53) has shown in his figs. 35, 36, 42. The same is seen in the sheep figs. 36 and 37. Zimmermann in the rabbit showed three mid brain segments and Froriep three in the mole. Zimmermann finds three, McClure two and Beraneck one in the chick's brain and McClure and Waters each found two in Amblystoma. Kupffer describes five in Salamandra, three inteleosts, Accipenser and Ammocoetes. It is evident that there is a considerable difference of opinion as to the number. I have given this brief resume of results because I


PARAPHYSIS AND PINEAL REGION IN REPTILIA 345

have had to touch on the mid brain slightly in this paper and the drawings and models take in this region. I have not gone into it deeply enough to express SiUj definite opinion though it seems to me probably that there are two subdivisions in this region. Where three have been mentioned it is possible that the most cephalic segment really belongs to the diencephalon and is that portion known as the 'Schalthirn' or synencephalon. As the cephalic end of the posterior commissure develops here that might explain the reason for adding it to the mid brain.

The three subdivisions of the fore brain have made their appearance after the formation of the fore brain vesicles. The question now arises, Are these true neuromeres equivalent to those in the hind brain, which appeared before the three primary central vesicles were developed, or are they merely secondary subdivisions of the fore brain appearing later than true neuromeres and having a different morphological value? Kupffer in Hertwig's Handbuch, 1903, pp. 19 and 152-166, gives a careful review of the work done on neuromeres. He distinguished between primary neuromeres seen on the open neural plate, which extend throughout its whole length, and secondary neuromeres, which appear later in the closed tube and are especially well marked in the hind brain, but much less distinct in the fore and mid brain. The expression primary is to be understood in an "ontogenetic not in a phylogenetic sense." These secondary neuromeres are better marked on the roof and lateral walls than on the floor of the brain.

Orr did not consider that the fore brain neuromeres were true neuromeres like those in the hind brain as their structure did not conform to that of hind brain neuromeres. McClure states that these fore brain neuromeres are true as regards their structure and character and he considers that they are continuous with and equivalent to those in the hind brain, but he suggests that there may be a rudimentary neuromere in the fore brain. Waters and Weber give the fore brain neuromeres full metameric value and Locy considers that his early segments can be traced into those structures that later become neuromeres. Herrick feels however that dorsal structures have been homologized with ventral and that if fore brain neuromeres were ever present they would be

THE AMERICAN JOURNAL OF ANATOMY, VOL. 11, NO. 4


34G JOHN WARREN

seen only at early stages. He does not believe in the ones described by the above writers. Neal agrees with him and states that the so-called fore brain and mid brain neuromeres are not true neuromeres owing to their shape, structure and late time of appearance. He prefers to homologize the fore and mid brain vesicles with hind brain neuromeres and to regard these segments as secondary subdivisions. Hill found that the five neuromeres of the mid and fore brains disappeared early before the primary fore and mid brain vesicles were developed. The future subdivisions of these he does not consider true neuromeres on account of their time of appearance and dorsal position.

As regards the structure of the segments shown in the models and figures I agree with Orr that they do not conform to regular hind brain neuromeres, though there is some resemblance. The diencephalic segments are certainly much better marked in the dorsal than in the ventral zone, especially the synencephalic segment and they appear much later than the neuromeres of the hind brain. From the evidence it seems clear that there is a primary set of neuromeres of which the majority of writers assign three to the fore brain and two to the mid brain. These apparently disappear when the mid and fore brain vesicles develop. The fore brain vesicle then divides later into the telencephalic segment and two diencephalic segments. This is clearly shown by the models. It is conceivable that, if the fore brain was primarily subdivided in three parts and that the lines of demarcation were later obliterated by modifications in the shape of the neural tube, that a secondary subdivision occurring here would tend to follow the lines of the primary subdivision, as the limits between neuromeres might be considered in a way as lines of least resistance. The relative shape and size of the secondary subdivisions or segments would naturally be modified and distorted by the unequal growth of the parts. This of course is merely a supposition and I do not think that a direct connection between the primary and secondary subdivisions has j^et been clearly proved. I am inclined to agree with von Kupffer's conclusions (57), p. 166,when he says that there are six neuromeres behind and five neuromeres in front of the fissura rhombo-mesencephalica in the region of fore and


PARAPHYSIS AND PINEAL REGION IN REPTILIA 347

mid brain. Everything else seems uncertain, especially the derivation of each of the secondary subdivisions from the primary neuromeres. These later segments or neuromeres should be distinguished as secondary from those which are primary. Therefore in the present state of our knowledge it seems to me that these later segments or subdivisions should be distinguished from those which appear earlier and are essentially primary neuromeres.

Paraphysal arch and paraphysis

Minot (71) showed three primary arches in the roof of the fore brain of Acanthias and gave the name paraphysal arch to that part of the roof of the telencephalon which lies immediately cephalad to the velum and passes into the dorsal end of the lamina terminalis. This section of the fore brain roof is early differentiated, figs. 2, 3, 17 and 18, and forms a striking arch up to the time of the appearance of the paraphysis and the telencephalic choroid plexuses. The paraphysal arch has been called by Burckhardt (12) the lamina supra-neuroporica and the plexus inferioris and plexus hemisphericum (telencephalic plexus) develop from it immediately in front of and lateral to the paraphysis. The shape and extent of the arch with its plexus development can be followed in Burckhardt's diagrams (13). The increasing size of the hemispheres soon tends to cover up this portion of the roof of the telencephalon. The differentiation of the arch in the lizard and the turtle can be also seen in figs. 2-10 and 17-21. Minot (71), Dexter (20) and Warren (97) show the development of this arch in Acanthias, chick and Necturus, and Johnson (53) in a series of pig embryos. It is well marked at first in the sheep, Neumayer (75). The arch appears distinctly as such in early stages only. The paraphysis and telencephalic plexuses tend to absorb this segment and when they are well developed no traces of it remain, but the opening of the paraphysis and the origin of the telencephalic plexus both still lie between the velum and the dorsal end of the lamina terminalis and together occupy all the space formerly occupied by the paraphysal arch.


348 JOHN WARREN

The paraphysis was first recognized as a distinct organ by Selenka (86) in the opossum. Kupffer and Burckhardt traced its history in cyclostomes. In Petromyzon and Ammocoetes it forms rather a small sack-like outgrowth, Kupffer (57) figs. 47 and 57, Burckhardt (12). Minot has traced it in elasmobranchs, especially in Acanthias (71), figs. 6-10, where it is a very simple evagination. Kupffer found it in ganoids and described it in Accipenser as a small vesicle, slightly folded at first, fig. 117, which later gives off tubules. Hill, Eycleshymer and Davis have observed it in Amia. It has been described by Terry (102) in Opsanus and by Burckhardt in other teleosts where it is lacking or else appears in a very rudimentary condition. In Protopterus, Burckhardt (13), pi. 8, records a wide sack-like paraphysis the walls of which are much folded. In amphibia it reaches the highest degree of development becoming a large complicated glandular organ, with a central lumen from which a very complicated set of anastomosing tubules are given off. It extends forward above and between the hemispheres and has a well marked sinusoidal type of circulation, Warren (97) figs. 17 and 20. Osborn has shown it in Siredon, Necturus, Proteus and Siren, Burckhardt in Ichthyophis and Triton. In all these forms it closely resembles the condition seen in Necturus as is the case also with the paraphysis of Rana, Minot; Amblystoma, Eycleshymer; and Diemyctylus, Mrs. Gage.

Francotte made the earliest observations on the paraphysis in the lacertilia. At first he confused it with choroid plexus but later recognized it as a distinct organ. In his thesis (28) he presented a series of photographs of embryos of Anguis fragilis. In his fig. 8 the paraphysis, velum, post velar arch and epiphysis are similar to my models of Lacerta muralis, 5 mm., fig. 9, while his figs. 12 and 13 are sections of the paraphysis of an embryo rather older which corresponds to my fig. 10. His fig. 15 would correspond to that of Lacerta viridis of 37 mm., fig. 12. Francotte does not give, however, the size of his embryos. His fig. 31, a lizard embryo, is also about this stage, though there is no plexus in the diencephalon. His figs. 27 and 28 show the relations in


PARAPHYSIS AND PINEAL REGION IN REPTILIA 349

frontal sections of paraphj'-sis, epiphysis and diencephalic plexus. The same author (29) shows a Lacerta muralis similar to my fig. 10, and his fig. 11 of Anguis is an excellent sagittal section about similar to my fig. 12. The epiphysis in this figure, however, is wrapped over the head of the paraphysis lying between it and the pineal eye, and the upper part of the diencephalon is not so compressed as in my models. In fig. 23 he shows a double paraphysal outgrowth similar to the three in the model in fig. 8. Francotte also gives three sections of the paraphysis in a young human embryo. My results therefore are essentially similar to his. The paraphysis in lacertilia develops from one, two or three outgrowths and as it grows curves backward over the post velar arch in the opposite direction to that in Necturus. It very early comes into close relation with the veins which overlie the roof of the diencephalon, but its relations to them are not so intricate as they were in amphibia and we do not find here such a perfect type of sinusoidal circulation as in Necturus. The final stage, fig. 13, gives a good view of the adult condition, which I have not seen elsewhere.

In the turtle we have a paraphysis which approaches rather the amphibian type. Its early stages are like those of the lizard, but as it grows larger the development of lateral tubules is much more marked, and therefore it has a very close relation to the veins about it, so that we have here a circulation of the sinusoidal type more perfect than in the lizard though not so marked as in amphibia. In fig. 23 the paraphysis is a well marked glandular structure with an elaborate set of lateral tubules which, however, at this stage are quite short and show no signs of anastomosing with each other. For relation to vessels, see fig. 27. In the lizard the tubule formation is slight except at the apex where the organ is expanded and a number of tubules appear, figs. 13 and 14, which show the same relations to veins as in the turtle.

In the crocodilia Voeltzkow (96) shows a paraphysis which follows closely the development in Lacerta except that at first it is inclined slightly forward. In the caiman the development of lateral tubules is much more marked. In Chelone imbricata the


350 JOHN WARREN

paraphysis is more like that of Chrysemys but seems to reach its highest development before the final stages and then to retrograde in size distinctly.

In Chelone mydas, Humphrey (51), fig. 7, the paraphysis is long and narrow as in Lacerta muralis while in Chelydra serpentina, fig. 30, it is a large wide tube with some diverticuli and resembles rather the paraphysis of serpents. Leydig (61) has described the paraphysis in Tropidonotus natrix and Vivipara urcini, Ssobolew (91) in Tropidonotus and Vipera berus, and Herrick (39) has shown it in Eutaenia. In all these it is well developed and resembles in general the paraphysis of Chrysemys and Lacerta.

In birds the paraphysis was described in detail by Dexter (20). It is greatly reduced in size and at its final stage is a very small outgrowth with a narrow lumen and very thick walls. It is much obscured by the diencephalic plexus, see his figs. 1-7 and Kupffer's (57) 278, 288-289.

But few observations have been made on the paraphysis in mammals since Selenka's article on the opossum in '90. I have observed it in a sheep embryo of 26 mm., fig. 38, as a short outgrowth rather narrower and more elongated than the paraphysis of the oldest chick mentioned above. Francotte has shown it in a twelve weeks human embryo and it has been modeled by Ewing Taylor in a human embryo of 22.8 mm., H. E. C. No. 871. I have also seen it in a human embryo, H. E. C. No. 1598, 28.8 nmi., fig. 39.

We see therefore that at the upper and lower end of the vertebrate series the paraphysis is much reduced while in the middle especially in the amphibia it is most highly developed.

The paraphysis is a distinct organ of relatively insignificant size in lower forms where it appears as a simple sack-like evagination, later giving off more or less side branches. In amphibia its glandular character is most highly marked and this is true to a lesser degree in reptilia. It is a mistake to say as some writers have that in advanced stages the paraphysis becomes changed into a sort of choroid plexus. This error is due to the fact that the structure gives off many complicated tubules which are in relation to a correspondingly complicated mass of vessels form


PARAPHYSIS AND PINEAL REGION IN REPTILIA 351

ing more or less of a sinusoidal circulation. This complex in sections looks somewhat like a plexus, but when modeled it can be clearly seen that the organ is absolutely distinct from any plexus. Its wall consists of a single layer of cells as is the case with the plexus, but its cells are always larger than the epithelial cells of the plexus, velum and post velar arch.

Velum and post velar arch

The velum has been recognized in practically all classes of vertebrates. It makes its first appearance as a slight infolding or angle in the roof of the fore brain, which grows downward into the brain cavity to a greater or lesser extent, and has a cephalic and caudal wall with connective tissue between. It is continued laterally into an internal ridge corresponding to an external groove, which ends in the floor just behind the optic stalk at the optic chiasma, figs. 2, 3, 5, 7, 8, and 17-19. As development goes on the shape and size of the velum become modified by the growth of the paraphysis and the telencephalic plexus in front of it, which absorb a large part of the paraphysal arch and at the same time the cephalic wall of the velum. This corresponds to Burckhardt's description (13). When these structures are well developed the velum appears in the sagittal plane as a rather insignificant angle, and becomes still more obscured by the hemispheres which grow dorsally on either side of it so that it is much reduced towards the median line. In more advanced stages it becomes still less evident and forms merely a lip bounding the caudal edge of the opening of the paraphysis. These steps can be traced in figs. 8, 9, and 12. In those forms which have a well developed diencephalic plexus the velum is still further modified. In early stages the velum forms a sharp angle with a cephalic and caudal wall. The former as I have mentioned is taken up by the paraphysis and telencephalic plexuses, while the caudal wall forms the cephalic portion of the post velar arch and is taken up more or less in the formation of the diencephalic plexus, fig. 22. The term post velar arch was given by Minot (71) to that part of the roof of the diencephalion extending from the velum to the epiphysal


352 JOHN WARREN

arch. This segment of the diencephahc roof varies greatly in its development. In almost all forms except some elasmobranchs it develops dorsally to a greater or lesser degree into a domelike expansion which is very striking. The roof is always a thin membrane, consisting of a single layer of low cells. In manycases it becomes very convoluted and forms vascular tufts which represent the diencephalic plexus. This region has been called by various names, 'Dorsal sack,' Goronowitsch, '88, 'Zirbelpolster,' Burckhardt (11), 'Parencephalon,' Kupffer (57), 'Postparaphysis,' Sorenson (90). I have used chiefly the term post velar arch though in the later stages 'dorsal sack' might be more appropriate. In Petromyzon Burckhardt (13), pi. 8, there is practically no velum. The hinder border of the mouth of the paraphysis is a mere lip and extending caudad from it is a short, low post velar arch. In Acanthias the velum is seen early as a sharp angle separating the paraphysal arch from the post velar arch Minot (71) figs. 1 and 2. It develops rapidly in a ventral direction absorbing in this process all the post velar arch so that the superior commissure and the opening of the epiphysis come to lie close against its base. Both walls of the velum especially the cephalic become convoluted later to form the diencephalic plexus which hangs down like a curtain across the cavity of the diencephalon, Minot (71), figs. 1-12. D'Erchia (24) shows in Pristiurus much the same condition, figs. 1-6, and also in Torpedo, figs. 7-15. Here, however, the post velar arch persists to some extent.

In Accipeliser, Kupffer (57), figs. 117, Burckhardt (13), pi. 8, there is a long velum the cephalic wall of which is convoluted, the caudal being perfectly flat. It resembles closely the velum of Acanthias, but the post velar arch forms a high vaulted dorsal sack so that the superior commissure is here separated by a long interval from the velum. In Protopterus Burckhardt (9) finds essentially the same condition as in Accipenser. In Amia and Lepidosteus Kingsbury (54) describes lateral expansions from the dorsal sack. On either side diverticuli pass out just behind the velum. These divide into a cephalic limb extending forward to the olfactory lobes and a caudal limb extending backwards to


PARAPHYSIS AND PINEAL REGION IN REPTILIA 353

the cerebellum. In teleosts, the trout, Kupffer and Burckhardt find again a long velum not so extensive as in previous forms, however, and showing no signs of plexus formation. The post velar arch is rather high and narrow. Terry (102) in Opsanus shows a greatly expanded velum taking on the character of a true choroid plexus.

In amphibia, Ichthyophis, Burckhardt (9) Necturus, Kingsbury (55), Warren (97) Diemyctylus, Mrs. Gage (32), the velum is well marked at an early stage and makes a deep angle in the roof. The paraphysis grows as a long narrow tube which takes up all of the cephalic wall of the velum, and causes a deep angle to appear between the prosencephalon and diencephalon the hinder limit of which is made by the caudal wall of the velum, the anterior limit by the remains of the paraphysal arch, Warren (97) figs. 3-8. The caudal wall of the velum soon begins to bulge into the diencephalon to form the diencephalic plexus, and continues until all of the velum and nearly all of the post velar arch are absorbed in the plexus formation. Only a slight extent of the post velar arch persists in front of the superior commissure. In Ichthyophis, Burckhardt (9), this process is not so extensive, a larger part of the post velar arch remaining intact. See also Kupffer (57), fig. 179-181, Salamandra and Eycleshymer (26), Amblystoma. In all these cases the velum is reduced to a mere lip, a very different condition from that in fishes and lower vertebrates where it makes a deep partition in the fore brain. In reptiles the velum is not so well marked as in amphibia. It forms only a relatively low angle in the roof, but is continued laterally into a well marked ridge, figs. 2-7, 17-20. The early stages shown in the models compare with Francotte's figures of about the same stages, Kupffer (57) figs. 239, Lacerta viridis; 240, Anguis fragilis; and Herrick (39) Eutaenia. The angle between the caudal wall of the velum behind and the paraphysal arch in front is much less than in Necturus owing to the fact that the velum originally does not grow so far ventrally and the paraphysis by folding backwards over the post velar arch does not seem to force the paraphysal arch so far downwards, figs. 9 and 10. This is certainly the case in the lizard, though in the turtle the condition is nearer


354 JOHN WARREN

that in Necturus, fig. 20. In lizards the post velar arch is at first low but soon develops up and back to form a large dome like cavity in the upper part of the diencephalon, figs. 3-10. This is even more marked in the turtle, figs. 17-21, Coincident with this process in both cases is the growth of the paraphysis, and the velum becomes reduced here to a mere lip or border bounding the mouth of the paraphysis. In the lizards after the stage shown in fig. 10 the cephalic and caudal walls of the extensive post velar become steadily forced towards each other by the mid brain encroaching from behind and the hemispheres from in front. The end result is seen in fig. 13. The velum at this stage becomes more prominent and rounded by the ingrowth of the fibers of Elliott Smith's aberrant commissure, figs. 12 and 13; Kupffer (57), fig. 248, Anguis fragilis; fig. 265, Lacerta vivipara. In the lizard the plexus formation is confined to the apex of the diencephalon, which has now been reduced to a slit very narrow antero-posteriorly, the superior commissure being separated by a slight interval only from the velum. The velum, however, is quite distinct and also a large part of the post velar arch, which has not entered into the formation of the diencephalic plexus. This closing up of the post velar arch or dorsal sack is one of the most striking features in the development of the fore brain of the lizard. If we compare this condition with my oldest turtle, which represents probably quite closely the final stages, we find a very different picture. The velum does not contain a commissure as was the case in the lizard and remains an insignificant, narrow lip, which becomes still further obscured by the development of the diencephalic plexus, figs. 20-23. There is no reduction in the size of the dorsal sack, which remains an extensive vaulted structure, but it is almost entirely filled up with an elaborate plexus extending from the mouth of the paraphysis to the superior commissure, figs. 22, 23, 25, 26. Voeltzkow (96), fig. 14, shows an embryo of Caiman niger that resembles the lizard of 37 mm. fig. 12. The velum a mere lip, with no commissure, the dorsal sack deep and fairly narrow with a diencephalic plexus at its apex only. The superior commissure lies very near the velum as in the lizard.


PARAPHYSIS AND PINEAL REGION IN REPTILIA 355

The same author gives a series of sagittal and transverse sections of crocodilus madagascarensis, figs. 1-11, which correspond still more closely with the development of this region in the Lacerta. The post velar arch, however, is rather more highly developed at first. The oldest stage is almost identical with fig. 13, though the epiphysis is absent and the paraphysis not so large. The velum is a thin lip, then comes a portion of the post velar arch, flat and unconvoluted. Towards the apex of the dorsal sack the diencephalic plexus is seen extending almost to the superior commissure, and the dorsal sack itself forms a high narrow slit, the superior commissure being almost in contact with the velum.

Voeltzkow (96) figs. 21 and 22 of Chelone imbricata show a picture quite similar to my fig. 23. The dorsal sack is spacious and filled by an elaborate diencephalic plexus which extends to the mouth of the paraphysis and obscures the velum. In Chelydra serpentina Humphrey (51) described a well marked post velar arch the greater part of which is occupied by folds of the diencephalic plexus, fig. 30. The velum is reduced to a mere lip as in my models. He also gives a sagittal view of Chelone midas, where the dorsal sack, however, is more compressed antero-posteriorly and the superior commissure is nearer the velum than in Chrysemys. In pi. 2, fig. 7, he shows the brain of Chelone midas. This approaches nearer to the lizards in having a very deep and narrow dorsal sack, as in fig. 12. In ophidians Leydig shows figure of Tropidonotus natrix and Ringelnatter which approach the conditions seen in turtles. The velum is insignificant, the post velar arch fairly high and expanded and the diencephalic plexus well developed. In Vivipara urcini the conditions resemble rather those seen in Lacerta viridis, fig. 13. In Eutaenia the velum becomes very indistinct being reduced to the narrowest possible lip. The dorsal sack is well arched and completely filled by a large diencephalic plexus extending from the paraphysis to the supra commissure. Ssobolew (91) has described this region in Tropidonotus natrix and Vipera berus, where the velum and dorsal sack are the same as shown by Leydig.


356 JOHN WARREN

In birds the velum has been described by Burckhardt (13), fig. 3, crow, and by Dexter (20) in the chick, fig. 1, where it forms a distinct angle with a well marked post velar arch behind it. See also KiipfTer (57) fig. 278 for a similar stage, and Burckhardt (13) crow embryo. The post velar arch forms a high tent-like structure the cephalic wall of which becomes convoluted to form the diencephalic plexus. The two walls are finally quite closely approximated and the resulting condition resembles somewhat that in the lizard. The plexus formation extends to the velum, which is again reduced to a mere lip, and becomes so extensive that the velum is practically covered up by the folds of the plexus, Dexter (20), fig. 3, and Kupffer (57), fig. 289. In mammals Johnson (53) shows the velum in a series of pig embryos from 5-17 mm., figs. 34-39. It forms at first an angle in the roof continued around the lateral walls as a distinct fold, but steadily becomes smaller; when at 17 mm. it is a mere notch in the roof. Behind it there is a well marked post velar arch.

Neumayer (75) gives a series of models of the brain of sheep embryos from 2.5-16 mm. There is a well marked velum with a low paraphysal arch behind it in embryos of 10 mm. In the sheep, H. E. C. as 1334, 6.6 mm., the velum is distinct with a well developed post velar arch. Later it greatly diminishes in size and at 26 mm., H. E. C. no. 1112, fig. 38, becomes covered up by the diencephalic plexus which develops from it and a large part of the post velar arch.

Ziehen (100), fig. 47, shows a sagittal section of a rabbit taken from Neumayer. Here there is a slight fold in the fore brain which Neumayer called the velum, but Ziehen has doubts whether the velum of lower forms appears in mammals, p. 279. He gives the boundary between diencephalon and telencephalon as the fossa praediencephalica, but is uncertain whether this should be homologized with the velum. His fig. 23, hedgehog embryo after Groenburg, seems to show a distinct velum with a well marked post velar arch behind and there is a distinct diencephalic plexus in the arch.

W. His (47), figs. 33, 35, and 38, shows a distinct velum in models of human embryos. In the lower forms, except in Petro


PARAPHYSIS AND PINEAL REGION IN REPTILIA 357

myzon, the velum is well developed and forms a curtain-like fold between telencephalon and diencephalon. In amphibia it is well developed at first but becomes absorbed by the excessive growth of the paraphysis and plexuses. In reptiles, birds and mammals it becomes proportionately less well marked especially toward the median line.

Choroid plexus

The choroid plexuses are closely associated with the paraphysal and post velar arches. The plexus choroideus lateralis springs from the paraphysal arch immediately in front of and lateral to the mouth of the paraphysis, fig. 25, and invaginates the dorsomesial wall of the hemisphere. In certain forms especially amphibia there is a median plexus arising from the paraphysal arch immediately in front of the paraphysis and developing downwards and backwards into the lower part of the third ventricle to a greater or less extent, Warren (97), Necturus, figs. 9 and 11. This is the plexus inferioris, anterior, or telencephalic plexus. When this plexus is present the lateral plexus springs from its base, Warren (97), fig. 12. In that article I used the term telencephalic for this plexus inferioris in order to distinguish it from the diencephalic plexus arising from the post velar arch. This term would apply equally well to the lateral plexus as they are both of telencephalic origin. I have however restricted it to this lower median plexus in order to differentiate it from the upper median or diencephalic plexus. This inferior or telencephalic plexus does not develop in Amphioxus and appears in a very reduced form in Petromyzon and Teleosts, Burckhardt (13). In Acanthias the velum projects deeply into the ventricular cavity and both its surfaces become much folded to form the choroid plexus. This is also the case in Accipenser and Amia and also Opsanus, Terry (102). It is present in Dipnoi, Protopterus, Burckhardt

(13).

It is a question whether these folds of the velum represent the diencephalic or telencephalic plexus of amphibia. The telencephalic plexus, plexus inferioris, of amphibia arises from the paraphysal arch in front of the paraphysis. The folds on the


358 JOHN WARREN

cephalic surface of the velum in the lower forms mentioned above really come from the telencephalon if we use the tip of the velum as the boundary between the two regions. However, they lie behind the opening of the paraphj^sis instead of in front of it. Kupffer (57), fig. 227, Triton cristatus, shows the plexus inferioris arising behind the paraphysis and forming an unmistakable mass in the lower part of the ventricular cavity. Strictly speaking in spite of the difference in relation to the paraphysis I suppose that the folds on the cephalic surface of the velum correspond to the telencephahc plexus or plexus inferioris of amphibia, while the folds on the caudal surface of the velum would correspond to the diencephalic plexus. In this case there would be in advanced stages practically no line of demarcation between the two plexuses as one could not be sure just what point corresponded to the original apex of the velum.

The telencephahc plexus reaches its highest development in amphibia forming a very striking median mass in the lower part of the diencephalon. See Necturus, Warren (97), Diemyctylus, Mrs. Gage (32), figs. 6 and 20-22, Ichthyophis, Burckhardt (13). Burckhardt (13) states that it is present in reptiles though in a greatly reduced form. I have found no traces of it in the lizard and have seen no signs of it in the figures of Francotte (28, 29, 30), Kupffer (57), Voelzkow (95), or Herrick (38, 39). In the turtle there are two paired masses growing backward from the origin of the lateral plexus into the diencephalon, figs. 22 and 26. These I think may possibly be regarded as homologous to the single median plexus of amphibia. Humphrey (51), fig. 30, found the plexus well developed in Chelydra serpentina, but in birds and mammals it is absent, Burckhardt (13).

The lateral plexus according to Burckhardt (13) and Edinger (23) «is lacking in teleosts. In Acanthias the lateral plexus is well developed, Kupffer (57), p. 84, and is also seen in Pristiurus, fig. 95. In ganoids it is absent, Burckhardt (13).

In amphibia it is very well developed and springs from the base of the telencephahc plexus to enter the lateral ventricle, Mrs. Gage (32), fig. 18, Studnicka (93), pi. 7, figs. 8, 13, Warren (97), figs. 12 and 14. In lizards it appears at 17 mm., fig. 10,


PARAPHYSIS AND PINEAL REGION IN REPTILIA 359

at the same stage in which the superior commissure first appears. It is a curious fact that this is also the case in Necturus and in the turtle, fig. 10. It forms here at first a rather solid mass, which soon, however, takes on the characteristic form of the choroid plexus. In the adult lizard it has rather a peculiar shape. It begins as a long narrow rounded mass just anterior and lateral to the paraphysis, which broadens out in the lateral ventricle. Its long axis is in the dorso-ventral direction, it being relatively narrow in the antero-posterior plane. Figs. 10-13 show the steps in its development. In the turtle this plexus is more developed and appears at earlier stages. It forms a large irregular mass roughly oval or quadrilateral and much more extensive than in the lizard. Its development is shown in figs. 20, 21, 22, and in section in figs. 25 and 26. In birds and mammals, especially in the latter, it is highly developed. See Ziehen, figs. 13, hedgehog embryo, 32 Echidna, 52-56 rabbit, 82 cat; Minot (71), fig. 390, sheep.

The diencephalic plexus develops from the caudal wall of the velum and from the post velar arch. This plexus is not well developed in lower forms and is absent in Petromyzon. It is probably represented by the thin membranous roof of the diencephalon in this and other lower vertebrates, Studnicka (93). In Acanthias, Accipenser and Amia it is seen in those folds that appear on the caudal wall of the velum (see p. 357). Burckhardt (13) describes it in Protopterus. In amphibia its development is excessive. It takes in all the caudal wall of the velum and the greater part of the post velar arch and extends backwards as a broad median mass, wide from above downward, but narrow from side to side, to end in a tufted extremity in the fourth ventricle. Warren (97), figs. 13, 16, and 23, shows this in Necturus, Burkhardt (13) in Ichthyophis, Mrs. Gage (32) in Diemyctylus, figs. 6, 23-25. In the lizard this plexus is represented by a series of transverse folds in the roof of the dorsal sack, fig. 12, and the plexus here is relatively small except at the adult stage, fig. 13. See also Francotte (29, 30), Kupffer (57), fig. 248, Anguis fragilis, 265 Lacerta vivipara, 258 Ornithorhyncus (G. E. Smith), Voeltzkow (96), Crocodilia, figs. 10, 11, 14, 15.


360 JOHN WARREN

In the turtle this plexus is much more extensive. It appears at a relatively early stage, 16 mm., fig. 21, and at 26 mm., fig. 22, is highly developed. Here we have a distinct bilateral arrangement as is shown in figs. 25 and 26. This is a striking feature of the turtle's plexus and the rudiments of the telencephalic plexus are also paired as was mentioned above. All the post velar arch and dorsal sack from the opening of the epiphysis up to the opening of the paraphysis is taken up in this plexus formation, figs. 22 and 23, and there is a sort of median ridge thrust in between the paired masses. This plexus is well marked in serpents, Leydig, fig. 5, and in Chelone imbricata, Voeltzkow (96), figs. 21-22. Herrick (38) shows a sagittal drawing from Sorensen of the brain of Cistudo where the diencephalic as well as the telencephalic plexuses are well developed. In Chrysemys I could find no trace of the latter plexus. In birds the plexus of the chick is described by Dexter (20), figs. 2-5, and by Kupffer (57), fig. 289. The diencephalic plexus is well marked and involves apparently the cephalic portion of the post velar arch only, the caudal part remaining smooth. It completely covers up the velum, and approaches more the condition seen in mammals.

In mammals the diencephalic plexus is strongly developed, forming large masses of plexus in the roof of the third ventricle which extend from the velum back to the epiphysis.

Commissures

In the velum of the lizard of 37 mm., fig. 12, and in the adult, fig. 13, is seen a well marked commissural band. This passes from the median wall of one hemisphere through the lowest part of the velum just caudal to the opening of the paraphysis to reach the median wall of the opposite hemisphere. I have also seen it in an earlier stage between figs. 11 and 12. This commissure was first mentioned by Rabl-Rlickhard ('81) in Psammosaurus and called by him Tornix-rudiment,' Meji^er ('93) named it the 'Hintere Mantel-kommissur,' Kupffer and Edinger have called it Commissura pallii posterior.


PAEAPHYSIS AND PINEAL REGION IN REPTILIA 361

Kupffer (57) states that it enters a swelling in the median wall of the hemisphere which represents the embryonic hippocampus of mammals. Kupffer (57) shows it in figs. 248 and 265, sagittal sections of Anguis, and in 259, a transverse section of Anguis. The latter figure illustrates its course clearly and as it crosses the median line it is bounded cephalad by the wall of the paraphysis and caudad by the single layer of cells forming the caudal wall of the velum. Kupffer states that laterally a part of the fibers enters the Eminentia medialis" in the medial wall of the hemisphere, while another part probably enters the stria medullaris, but owing to the presence of other fiber tracts he thinks the exact composition of the commissure is not clear. I have seen a similar section to fig. 259 in H. E. C, no. 1604, Lacerta muralis. G. Elliot Smith (88) described very thoroughly this commissure in Sphenodon and called it an aberrant commissure, ' Commissura aberrans.' Dendy (18) mentioned this commissure also in Sphenodon and called it Commissura fornicis. Elliot Smith's figs. 1 and 2 give a good view in the sagittal plane and show how the commissure makes a striking rounded swelling in the velum, which forms a marked projection between the opening of the paraphysis and the post velar arch. He describes two masses of gray matter in the mesial wall of the hemisphere as the paraterminal bodies which are connected by the lamina terminalis (88), figs. 4, 5 and 6. Above these structures in the mesial wall of the hemisphere is the region of the hippocampus, below them the corpus striatum. The figures show the relations of the commissura dorsalis and ventralis to these structures and to the ventricles In fig. 10 the aberrant commissure appears at a later stage and the writer states that it crosses over the ventricle in the bridge of gray matter formed by the fusion of the caudal extremities of the paraterminal bodies. After passing through the paraterminal body the fibers end in the hippocampus in a manner exactly analogous to that which characterizes the commissura dorsalis in the most cephalic region of the hippocampus. He concludes that the commissure is derived from the caudal end of the hippocampus and crosses by the more direct route via the velum rather than by the longer one via the lamina terminalis

THE AMERICAN JOUKNAl. OF ANATOMY, VOL. 11, NO. 4


362 JOHN WARREN

as do the commissural fibers in mammalia. It is foi-med, however, as a distinct commissure only in Hphenodon and Lacertilia. He thinks it more than probable that its fibers are represented in amphibia by fibers which cross the roof of the diencephalon in the superior commissure,

Superior commissure

The superior commissure is a very. constant tract in the vertebrate brain and has a very definite position. It lies immediately cephalad to the stalk of the epiphysis and below the pineal recess in the upper and hinder part of the diencephalon. Osborn (78) first gave it the name superior commissure and traced its homologies. He showed it in the brain of Menopoma, fig. 8. It was also described at about the same time by Bellonci. Burkhardt (18) described its homologies in lower vertebrates, reptiles and birds. Cameron (15) gives an excellent account of this structure in teleosts, amphibia, reptiles, birds and mammals. In teleosts the tract appears early and he found fibers arising from the cells in one ganglion habenula and passing to the other, while other fibers passed to cells in the epiphysis of the same and also opposite sides, forming a real decussation.

In amphibia he studied the frog, toad and newt. In these forms the commissure was found to lie some little distance cephalad to the stalk of the epiphysis with quite an interval between. He emphasizes this as an amphibian characteristic. This is not always the case, as in Necturus, Warren (97), it lies close to the stalk of the epiphysis. This is shown also by Osborn (78) in Menopoma, fig. 8, Mrs. Gage (28), Diemyctylus, and Kupffer (57), figs. 180, Salamandra, However Kupffer, fig. 228, Rana esculenta, shows the condition described by Cameron. In Cameron's amphibian type, he finds that later a second set of fibers appears in the anterior part of the commissure and does not enter the ganglia habenulae. He was unable to follow their further course. It is possible that these fibers may correspond to Elliot Smith's commissura aberrans of reptilia which according to him


PARAPHYSIS AND PINEAL REGION IN REPTILIA 363

is homologous with fibers crossing the third ventricle in the supra commissure in amphibia.

In Lacerta the commissure is well marked and appears first in an embryo of 17 mm., fig. 10. It becomes a well marked tract in higher stages, figs. 11-13. In Chrysemys it appears first in an embryo of 9.5 mm., fig. 20. As I mentioned above it is a curious fact that in these two forms and also in Necturus the commissure appears at the same stage in which the first traces of the lateral choroid plexus are seen. Cameron also describes a well marked commissure in reptiles. See also Burckhardt, Francotte, Dendy, Beraneck and others.

In birds the commissure was first described by Dexter ('02) in the chick, figs. 2, 3 and 9. It appeared in a chick of 19.5 mm. Cameron finds it first in chicks of ten days incubation close in front of the epiphysis. Cameron (15) calls attention to the fact that in Cunningham's Text-book of Anatomy, p. 506, mention is made of fibers from the stria medullaris that pass across the median line in front of the stalk of the epiphysis to reach the ganghon habenularum. Cameron considers these fibers homologous with the superior commissure of lower forms. He found in studying sections of this region stained by the Pal-Weigert method that this commissure was well defined and divided into an anterior and posterior part by a cleft. He noticed also that some of these fibers passed into the epiphysis at its base on the opposite side, figs. 8 and 9. This show^s that there is a distinct decussation of fibers here as was seen in amphibia. He observed furthermore that the fibers belonging to the striae medullaris lay in the anterior portion, while the posterior part was occupied by fibers arising in one ganglion habenula and passing to the other or to the epiphysis. This hinder part alone he thinks to be homologous with the commissure of lower forms.

The commissure has been described in pigs by Minot (72) , who states that it is very constant in all classes of vertebrates, and Neumayer (75) has shown it in the sheep and rabbit (101), fig. 47. It alw^ays appears after the posterior commissure is well developed except in Ammocoetes, where it appears before the posterior, Kupffer (57).


304 JOHN WARKEN

Posterior commissure

The posterior commissure is constant in all vertebrates and always precedes the superior commissure except in Ammocoetes. This tract is usually regarded morphologically as a part of the mid brain and as the boundary between mesencephalon and diencephalon. However, if we regard the 'Schalthirn' or synencephalon as a part of the fore brain as it really is, a portion of the commissure must be assigned to the diencephalon. The first appearance of the commissure in the lizard is seen in fig. 8, Lacerta muralis, 4.5 mm. It forms a well marked flat band of fibers in the outer part of the wall of the brain and is confined entirely to the synencephalon. This is still the case in fig. 9, Lacerta muralis, 5 mm. Kupffer (57), fig. 286, states that the posterior commissure always appears first in the posterior part of the synen-. cephalon. At first there is a section of the synencephalon between this commissure and the epiphysis where there are no commissural fibers, but later in most cases this part is invaded by fibers and absorbed by the commissure. This is not the case, however, in Necturus where there is a portion of the synencephalon which always remains unoccupied by commissural fibers. After this stage it grows backward into the mid brain and there it spreads out and blends with the outer layer of the wall. It becomes very diffuse and has no definite caudal limits. In later stages as the mid brain encroaches on the fore brain the commissure becomes doubled over on itself and an anterior and posterior part is formed separated by a sort of septum which really consists of the outer cells of the brain wall. These points have been emphasized by Terry (102) in Batrachus. This encroachment of the mid brain and the doubling up of the commissure compresses the roof of the synencephalon to such a degree that this segment of the fore brain becomes virtually suppressed. It is a question whether we have to deal here with two commissures, or two distinct parts of one commissure, one forming first in the roof of the synencephalon and a second of later appearance joining this and spreading into the mid brain. I have not been able as yet to determine this point.


PARAPHYSIS AND PINEAL REGION IN REPTILIA 365

In the turtle the same condition is seen, fig. 20. Here the commissure is confined to the synencephalon and later, figs. 2122 spreads backward into the mid brain. In the turtle the commissure is not doubled over on itself as was the case in the lizard.

Epiphysis and pineal eye

Of all the structures that develop from the roof of the fore brain the epiphysis and especially the pineal eye have been of the greatest interest to investigators. Consequently the literature on this subject is very voluminous and the region has been very thoroughly studied in all the more important vertebrate forms. It is intended to discuss here a few points only concerning the epiphysis and pineal eye as nothing especially new has been brought out in this paper on these structures. The models however give a good view of the topographical relations of this region.

The epiphysis always arises from the caudal part of the diencephalic roof and appears in every case before the paraphysis. Its anlage is the epiphysal arch, Minot (71), from which it appears first as a simple outgrowth. The superior commissure develops immediately in front of the epiphysis and usually lies close against its stalk, except in certain amphibia, where there is a distinct interval between it and the epiphysis, Cameron (15, 16), Kupffer (57), fig. 228. The epiphysis marks the caudal boundary of the parencephalon and behind is the roof of the synencephalon on pars intercalaris. Here develops first the posterior commissure which later usually takes in all that segment and comes to lie close against the hinder side of the stalk of the epiphysis, Terry (102). In some cases, for example Necturus, the commissure is separated from the epiphysis by a distinct interval. The epiphysis always remains attached to the brain by some sort of stalk which is usually very thin in adult forms. This stalk contains a cavity up to certain stages which communicates with the diencephalon, but the cavity is almost always obliterated before the final stage in development is reached. The changes in shape and size of the epiphysis in both lizard and turtle can be followed in the models. In both forms the organ is ex


366 JOHN WARREN


s


panded at its distal end and tapers to a rounded solid stalk at its proximal end which lies between the two commissures. That of the lizard is very attenuated in the adult stage. . The epiphysis lies close against the caudal wall of the dorsal sack and is even embedded somewhat in it, especially so in the turtle. , In most of the stages the epiphysis is surrounded by vessels, figs. 11 and 20. and later lies immediately beneath the superior sagittal sinus, fig. 22. In both the turtle and lizard the distal end of epiphysis and paraphysis come almost in contact. In the latter the proximal ends are separated by the wide dorsal sack and both structures develop over the top of this region until their distal ends are closely approximated. In the lizard the compression of the dorsal sack brings not only the distal ends but also the proximal ends of the two structures very close together. Francotte shows a section of this region in Anguis where the tip of the epiphysis is actually wrapped over the tip of the paraphysis. The epiphysis is present in all forms except myxinoids, Torpedo and crocodilia, Studnicka (90), Voeltzkow (96).

In certain reptilia there is also present the parietal or pineal eye which lies at some distance from the epiphysis in the parietal foramen, fig. 13. The pineal eye or pineal organ resembles in many respects the paired eyes and has a nerve which connects it with one of the ganglia habenulae. The staining of my specimens was not favorable for the study of this nerve, but is is generally accepted that such a nerve exists. (Francotte, Beraneck, Burckhardt, de Klinckowstrom and Dendy.) Cameron (15, 16) has also shown nerve fibers passing from the ganglion habenulae of one side to epiphysal elements on the opposite side forming a true decussation in the superior commissure. He observed these fibers in teleosts, amphibia, birds and man.

One of the most vexed questions about the development of these pineal organs concerns the origin of this so-called i ye. One set of writers considered that it was formed from the distal end of the epiphysis from which it was simply constricted off. Another set maintained that the pineal eye arose by an independent outgrowth from the brain and consequently there were two evaginations, one immediately in front of the other, the more


PARAPHYSIS AND PINEAL REGION IN REPTILIA 367

cephalic being the future eye, the more caudal the future epiphysis. Francotte suggested the following explanation, namely that the first vesicle to appear was the eye and the second coming behind it was the epiphysis. This latter outgrowth developed so close behind the former that it appeared almost to spring from its posterior wall, and as the two grew larger they became combined into what was apparently one vesicle. He shows (30), fig. A, two outgrowths, one immediately in front of the other. In figs. 7, 8 and 9 he shows a large one and then a smaller one close behind the base of the larger. I am inclined to agree with this explanation of Francotte. Fig. 6 shows a large cephalic vesicle, then comes a very narrow lip and then a smaller caudal vesicle directed somewhat backward from the former. The first is the eye, the second the epiphysis. They have a common opening into the diencephalonfrom which the cavities of each slant caudad and cephalad. The lip between the two, which looks like a sort of septum partly dividing a single vesicle, is really the line of division between the two primary outgrowths which arose individually from the roof of the diencephalon. Therefore, the single outgrowth E seen in fig. 5 would be realty the pineal eye and the smaller caudal outgrowth in fig. 6 has probably appeared secondarily and will form the epiphysis. This latter, therefore, corresponds to the single evagination in the turtle, £',fig. 18, where there is no pineal eye. This view of the individuality of the pineal eye is I believe now generally accepted although certain writers still claim that it is a differentiation of the distal end of the epiphysis. Another matter of interest is the question of the bilateral origin of the pineal outgrowths. According to the results of Beraneck, Dendy, Hill, Locy and Cameron it seems that there are in very early stages two bilateral outgrowths, one of which is always smaller than the other and soon disappears. Cameron (14, 16) shows these clearly in the chick and in amphibia and Dendy in Sphenodon. Both these writers found that the left outgrowth always persisted. Locy in elasmobranchs and Hill in teleosts and Amia also observed bilateral evaginations. In Amia Hill found that the left one disappeared while the right one developed. Cameron argues that the fact that the nerve to the pineal


3G8 JOHN WARREN

eye probablj^ crosses from one side to the other as observed by Dendy and de Klinckowstrom goes to prove that the pineal outgrowths are really bilateral. My specimens of Lacerta and Chrysemys did not show this bilateral arrangement. The region of the pineal organs is very thoroughly considered by Gaupp (34), and Studnicka (95) in Oppel's Lehrbuch, V, 1904.

SUMMARY AND CONCLUSIONS

1. After the appearance of the primary fore brain vesicle the prosencephalon is subdivided into telencephalon and diencephalon. The diencephalon undergoes a further subdivision into two segments, the cephalic one being the parencephalon and the caudal one the synencephalon or pars intercalaris. There are, therefore, three subdivisions to the fore brain.

2. The first segment or subdivision is the telencephalon, which is bounded caudad by the velum and the ridge passing from the velum to the optic commissure. From its roof develop the paraphysal arch, paraphysis and telencephalic choroid plexuses. From its lateral walls arise the hemispheres and ventrad to them the optic vesicles. In the floor is the optic recess and opening of the optic stalk.

3. The • second subdivision is the parencephalon, which is bounded caudad by the hinder wall of the epiphysis dorsally, and by the tuberculum posterius ventrally. Between these points there is a slight ridge in early stages. From the roof arise the post velar arch, diencephalic plexus, epiphysal arch, both pineal organs, epiphysis and pineal eye, and the supra commissure. In the floor are the infundibular and mammillary regions.

4. The second subdivision of the diencephalon is the synencephalon, or pars intercalaris. It is bounded caudad by a dorsal groove and a ridge which ends below at the highest part of the habenular flexure. Its floor is limited and the segment is often wedge shaped. In the roof a portion of the posterior commissure develops.

5. The mid brain probably is subdivided into two segments after the primary mid brain vesicle is formed.


PARAPHYSIS AND PINEAL REGION IN REPTILIA 369

6. The three subdivisions of the fore brain and the two subdivisions of the niid brain presumably are secondary segments. They do not seem to correspond exactly in structure with the typical hind brain neuromeres and should therefore be distinguished from those segments which appear earlier and are essentially primary neuromeres.

7. In the lizard the paraphysis develops from one, two or three primary outgrowths from the paraphysal arch. These appear first in embryos of 3.2-3.6 mm. The organ forms a long tube with well developed tubules in its distal part and has a sort of sinusoidal circulation. It is closely moulded over the dorsal sack.

In the turtle it develops from one outgrowth seen first in embryos of about 6-7 mm. It becomes a relatively complicated structure with many lateral tubules and a sinusoidal circulation approaching that of amphibia. As in the lizard it grows backward in close contact with the dorsal sack.

8. The velum in both lizards and turtles forms only a slight angle in early stages and is later much reduced towards the median line where it becomes a mere lip, which forms the caudal boundary of the opening of the paraphysis.

9. The post velar arch in the lizard forms a wide dome-like dorsal sack which later becomes compressed to a high, deep transverse slit. In the turtle the post velar arch forms an extensive, vaulted dorsal sack which does not undergo that compression so striking in the lizard.

10. The pineal eye and the epiphysis of the lizard arise as two outgrowths from the epiphysal arch, that for the eye being immediately in front of that for the epiphysis. The outgrowths appear in embryos of 2.4-3.6 mm. The pineal eye is separated from the epiphysis in an embryo of 5 mm. and migrates gradually from the region of the epiphysis to reach in final stages the parietal foramen. The epiphysis always remains attached to the brain bj'^ an attenuated solid stalk and becomes much expanded distally. It lies close against the caudal wall of the dorsal sack. In the turtle there is no pineal eye. The epiphysis appears at 5 mm. and becomes an elongated body with an expanded tip and rounded stalk. It curves forward and lies over the dorsal sack.


370 JOHN WARREN

11. The superior commissure is well developed in both lizard and turtle. It appears in the former in embryos of about 17 mm. and in the latter in embryos of about 8.9 ram. In both cases it is coincident with the first appearance of the anlage of the lateral plexuses.

12. The posterior commissure develops first in the roof of the synencephalon or second subdivision of the diencephalon. It invades the mid brain later.

13. The plexus choroideus lateralis is much better developed in the turtle than in the lizard. It appears in the former in embryos of about 8.9 mm., in the latter in embryos of about 17 mm.

The diencephalic plexus is very much more highly developed in the turtle than in the lizard. It forms two lateral masses that occupy the whole length of the dorsal sack from the velum to the supra commissure. Its anlagen appear first in embryos of about 16 mm. In the lizard it appears later, 37 mm., and occupies the apex only of the narrow dorsal sack.

The telencephalic plexus or plexus inferioris in the turtle is possibly represented by short bilateral prolongations growing caudad from the lateral plexus. There is no sign of them in the lizard.

In conclusion I wish to acknowledge the valuable advice and assistance kindly given me by Professor Minot in the preparation of this article.

ADDENDA

I regret that I was unable to consult the two following papers before the main part of my article was written and sent in for publication. The first paper by Tandler and Kantor (105) gives a most admirable series of pictures of models of the brain of the gecko. The general development of the pineal region corresponds closely to that shown in my models. The paraphysis, however, is carried backward until its tip overlaps the epiphysis and both of these structures are crowded against the wall of the mid brain. In the oldest stage (fig. 17) there is no diencephalic choroid plexus in the dorsal sack and no signs at any stage of the pineal eye. The synencephalic segment of the


PARAPHYSIS AND PINEAL REGION IN REPTILIA 371

diencephalon is well marked in the early stages. The second paper by Dendy (103) gives a very interesting and thorough description of this region in sphenodon. Especially interesting, is Dendy's statement of the shifting of the opening of the paraphysis and the formation of the supra commissural canal, text fig. 3. I was unable to observe this in Lacerta muralis. The circulation and topographical relations of the pineal complex are well shown in text figs. 13-16. The formation and histological structure of the pineal eye has been treated very thoroughly and with great care.


372 JOHN WARREN

BIBLIOGRAPHY

1 Ahlborn, F. 1884 Ueber die Bedeutung der Zirbeldriise. Zeit. f. wiss"

Zool., Bd. 40.

2 1884 Ueber die Segmentation der Wirbelthierkopfes. Zeit. f. wiss. Zool., Bd. 40.

3 B6raneck, E. 1887 Ueber das Parietalauge der Reptilien. Jena. Zeit.,

Bd. 21.

4 1892 Sur le nerf parietal et la morphologie du troisieme ceil des vertebras. Anat. Anz. 7.

5 1893 L'individualite de I'oeil parietal. Anat., Anz. 6.

6 1884 Recherches sur le developpement des nerfs craniaux cliez les Lizards. Recueil Zool. Suisse.

7 1887 Etudes sur les replies meduUaires du ponlet. Ebendas.

8 Broman, I. 1895 Beschreibung eines menschlichen Embryo. Morpholog.

Arbeiten von Schwalbe, Bd. 5.

9 BuRCKHARDT, R. 1891 Die Zirbel von Ichthyophis glutinosus und Protop terus annectens. Anz. Anat., 6.

10 1891 Untersuchungen am Hirn und Geruchsorgan von Triton und Ichthyophis. Zeit. f. wiss. Zool., 52.

11 1893 Die Homologien des Zwischenhirndaches und ihre Bedeutung fiir die Morphologie des Hirns bei niederen Vertebraten. Anat. Anz., 9.

12 1894 Die Homologien des Zwischenhirndaches bei Reptilien und Vogeln. A.nat. Anz., 9.

13 1894 Der Bauplan des Wirbeltiergehirns. Morpholog. Arbeiten von G. Schwalbe, 4.

14 Cameron, J. 1903 On the origin of the pineal body as an amesial structure

deduced from the study of its development in Amphibia. Anat. Anz., Bd. 23.

15 1904 On the presence and significance of the superior commissure throughout the vertebrata. Jour. Anat. Physiol., vol. 38.

16 1904 On the origin of the epiphysis cerebri as a bilateral structure in the chick. Proc. Royal Soc. Edinburgh, vol. 25.

17 CARRii}RE, J. 1889 Neuere Untersuchungen liber das Parietalorgan. Biol.

Centralbl., Bd. 9.

18 Dendy, a. 1899 On the development of the pineal eye and adjacent organs

in Sphenodon (Hatteria). Quart. Jour, of Micr. Science., vol. 42.


PARAPHYSIS AND PINEAL REGION IN REPTILIA 373

19 Dean, B. 1895 The early development of Amia. Quart. Jour. Micr.

Science, vol. 38.

20 Dexter, F. 1902 The development of the paraphysis in the common fowl.

Am. Jour. Anat., vol. 2.

21 DoHRN, A. 1875 Ursprung der Wirbelthiere und das Prinzip des Fonctions wechsels. Leipzig.

22 DuRSEY, E. 1869 Entwicklungsgeschichte des Kopfes.

23 Edinger, L. 1904 Vorlesungen liber den Bau der nervosen Zentralorgane.

7. Aufl., Leipzig.

24 D'Erchia, F. 1896 Contribute alio studio della volta del cervello inter medio e della regione parafisaria in embrioni di pesci e di mammiferi. Monit. Zool. Ital. Ann., 7.

25 Eycleshymer, A. C. 1890 The development of the optic vesicles in

Amphibia. Jour. Morph., vol. 8.

26 1892 Paraphysis and epiphysis in Amblystoma. Anat. Anz., 7.

27 1895 The early development of Ainblystoma. Jour. Morph., 5.

28 Francotte, p. 1888 Recherches sur le developpement de I'epiphyse.

Arch, de bioL, vol. 8.

29 1894 Note sur rceil parietal, I'epiphyse, la paraphyse et les plexus choroides du troisieme ventricule. Bull, de I'Acad. roy. de Belg., no. 1.

30 1892 Contribution a I'etude de I'epiphyse et de la paraphyse chez les Lacertiliens. Bruxelles, 1896.

31 Froriep, a. 1892. Zur Frage der sogennanten Neuromerie. Verhandl.

Anat. Ges.

32 Gage, S. P. 1893 The brain of Diemyctylus viridescens.

33 1905 The Wilder quart, century book. Ithaca, N. Y. Serial order of segments in the fore brain of three and four week human embryos; comparisons with lower forms. Abstract. Science, n. s. vol. 17. pp. 486.

34 Gaupp, E. 1898 Zirbel, Parietalorgan und Paraphysis. Merkelu. Bonnet's

Ergeb. d. Anat. u. Entwicklungsgeschichte, Bd. 7.

35 Goette, a. 1875 Die Entwicklungsgeschichte der Unke. Leipzig.

36 Graaf, H. W. de 1886 Zur Anatomie und Entwicklung der Epiphyse bei

Amphibien und Reptilien. Zoolog. Anzeiger, Jahrg. 9.

37 Herrick C. L. 1891 Contributions to the morphology of the brain of

bony fishes. Jour. Comp. Neur. Psych., 1.


374 JOHN WARREN

38 1891 Topography and histologj' of the brain of certain reptiles. Jour. Comp. Neur. Psych., 1, pp. 25 and 162, vol. 3, pp. 92 and 131, 1893.

39 1892 Embryological notes on the brain of the snake. Jour. Comp Neur. Psych.

40 Hill, Ch. 1891 Development of the epiphysis in Corregonus albus. Jour.

Morph., 5.

41 1894 The epiphysis of Teleosts and Amia. Jour. Morph., 9.

42 1900 Developmental history of primary segments of the vertebrate head. Zool. Jahrb., Abt. f. Anat. u. Ontog., Bd. 13.

43 1899 Primary segments of the vertebrate head, Anat. Anz., Bd. 16, 353-369.

44 His, W. 1889 Die Formentwicklung des menschlichen Vorderhirns vom

Ende des ersten bis zum Beginn des dritten Monats. Abt. Mat. phj's. CI. kon. -Sachs. Ges. Wiss. Bd. 15, Leipzig.

45 1892 Zur allgemeinen Morphologie des Gehirns. Archiv f. Anati u. Phys., Anat. Abt.

46 1893 Ueber das frontal Ende des Gehirnrohres, Archiv f. Anat. u. Entw., 157-171.

47 1904 Die Entwicklung des menschlichen Gehirns., Leipzig.

48 Hoffmann, C. K. 1885 Weitere Untersuchungen zur. Entwicklungsges chichte der Reptilien. Morphologisches Jahrb., 11.

49 1888 Epiphyse und Parietalauge. Bronn's Klassen und Ordnungen des Thierreichs, Bd. 6, Abt. 3.

50 1888 Reptilien. In Bronn's Klassen und Ordnungen des Thierreichs, pp. 1996-1971, 1989.

51 Humphrey, O. D. 1894 On the brain of the snapping turtle. Jour. Comp.

Neur. Psych., vol. 4.

52 Johnson, J. B. 1905 The morphology of the vertebrate head from the view

point of the functional divisions of the nervous system. Jour. Comp. Neur. Psych., 15.

53 1909 The morphology of the forebrain vesicle in vertebrates. Jour. Comp. Neur. Psych., 19.

54 Kingsbury, B. J. 1897 The encephalic evaginations in Ganoids. Jour.

Comp. Neur. Psych., vol. 7.

65 1895 The brain of Necturus maculatus. Jour. Comp. Neur. P.sych.,

vol. 5.

56 Klinckowstrom, A. de 1893 Le premier d(5veIoppement de I'oeil parietal, I'i^piphyse et le nerf pari(5tal chez Iguana tuberculota. Anat. Anz. 8.


PARAPHYSIS AND PINEAL REGION IN REPTILIA 375

57 KuPFFER, C. VON 1903 Die Morphogenie des Nervensystems. In Oskar

Hertwig's Handbuch der vergleichenden u. experimentellen Entwicklungslehre der Wirbelthiore. Jena.

58 Lbydiq, F. 1890 Das Parietalorgan. Biol. Centralbl., 10.

59 1891 Das Parietalorgan der Amphibien und Reptilien. Abhandl. der Senckenberg. naturf. Gesellsch., 6.

60 1896 Zur Kenntnis der Zirbel u. Parietalorgane. Abhandl. der Senckenberg. naturf. Gesellsch., Bd. 19.

61 1897 Zirbel und Jacobsonsche Organe einiger Reptilien. Archiv f. mikr. Anatomie, Bd. 50.

62 LocY, W, A. 1893 The derivation of the pineal eye. Anat. Anz., 9.

63 1894 Metameric segmentation in the medullary folds and embryonic rim. Anat. Anz., 9.

64 1894 The mid brain and accessory optic vesicles. Anat. Anz., 9.

65 1895 A contribution to the structure and development of the vertebrate head. Jour. Morph., vol. 11.

66 McClure, Charles I. W. 1889 The primitive segmentation of the verte brate brain. Zool. Anz., no. 314.

67 1890 The segmentation of the primitive vertebrate brain. Jour. Morph., 4.

68 Meek, A. 1907 The segments of the vertebrate brain and head. Anat.

Anz., vol. 31.

69 Meek, W. G. 1907 A study of the choroid plexus. Jour. Comp. Neur.

Psych., vol. 17.

70 MiNOT, C. S. 1892 Human embryology. New York.

71 1901 On the morphology of the pineal region based upon its development in Acanthias. Am. Jour. Anat., vol. 1.

72 1910 Laboratory text book of embryology, 2nd ed.

73 Neal, H. V. 1896 A summary of studies on the nervous system of Squalus

Acanthias. Anat. Anz., Bd. 12.

74 1898 The segmentation of the nervous system in Squalus Acanthias; a contribution to the morphology of the vertebrate head. Bull. Mus. Comp. Zool., Harvard Coll., 31.

75 Neumayer, L. 1899 Studien zur Entwicklungsgeschichte des Gehirns der

Saugetiere. Festschr. z. 70, Geburtstag von C. von Kupffer. Jena, Fischer.


376 JOHN WARREN

76 1900 Zur Morphogenie des Gehirns der Siiugetiere. Sitzungsber. d. Gesellsch. f. Morph. u. Phy. inMunchen, Bd. 15.

77 Ore, H. 1887 A contribution to theembryology of the lizard. Jour. Morph.,

vol. 1.

78 OsBORN, H. F. 18S4 Preliminary observations on the brain of Menopoma.

Proc. Acad. Nat. Sci. Phila.

79 1887 The relation of the dorsal commissures of the brain to the formation of the encephalic vesicles. Am. Nat., vol. 21.

80 1888 A contribution to the internal structure of the amphibian brain. Jour., Morph. 2, pp. 51-96.

81 1888 Origin of the corpus callosum, etc. Morph. Jarbhuch, 8.

82 Pettitt et Gerard. 1902-03 Sur la fonction secretoire et la morphologic

des plexus choroides. Archiv. d'anat. micros., 5.

83 Prenant, A. 1889 Replismedullaires chei- I'embryon du pore. Bull, de la

soc. d. sc. de Nancy.

84 1893 Sur I'ceil parietal accessoire. Anat. Anz., 9.

85 Scott, W. B. 1887 Notes on the development of Petromyzon. Jour.

Morph., 1.

86 Selenka, E. 1890 Das Stirnorgan der Wirbeltiere. Biol. Centralbl., 10.

87 Smith, G. Elliot 1897 Brain of a foetal ornithorynchus. Quart. Jour.

Micros. Sci,. n. s. vol. 39.

88 1903 On the morphology of cerebral commissures in the Vertebrata with special reference to an aberrant commissure formed in the fore brain of certain reptiles. Trans. Linnean Soc. London, 8 (2nd Ser. Zool.).

89 SoRENSoN, A. D. 1893 The pineal and parietal organ in Phrynosoma coro nata. Jour. Comp. Neur. Psych., 3.

90 1894 Comparative study of the epiphysis and roof of the diencephalon. Jour. Comp. Neur. Psych., 4.

91 SsoBOLEw, L. W. 1907 Zur Lehre von Paraphysis und Epiphysis bei

Schlangen. Arch. f. mikros. Anat., vol. 70.

92 Straul, H., AND Martin, E. 1888 Die Entwicklung des Parietalauges bei

Anguis fragilis und Lacerta vivipara. Archiv f. Anat. u. Entw., 146-161.

93 Studnicka, F. K. 1893 Zur Losung einiger Fragen aus der Morphologic

des Vorderhirns der Cranioten. Anat. Anz., 9.

94 1895 Zur Anatomie dersogenanntenParaphysedesWirbeltiergehirns. Sitzungsber. d. konig. Bohm. Gesellsch. d. Wissensch. in Prag.


PAKAPHYSIS AND PINEAL REGION IN REPTILIA 377

95 ■ 1905 Die parietalorgane. Oppel's Lehrbuch tier vergleichenden

mikroskopischen Anatomie der Wirbeltiere. 5., Jena.

96 VoELTZKOW, A. 1903 Epiphysis uud Paraphysis bei Krokodilien und Schikl kroten. Abhand. d. Senckenberg. naturf. Gesellsch., Bd. 27.

97 Warren, J. 1905 The development of the paraphysis and pineal region in

Xecturus maculatus. Am. Jour. Anat., vol. 5.

98 Waters, B. H. 1S92 Primitive segmentation of the vertebrate brain.

Quart. Jour. Micros. Sci., 33.

99 Weber, A. 1900 Contribution a I'etude de la metamerie du cerveau ante rieur chez quelques oiseaux. Arch, d'anatomie microscop.

100 Ziehen, Th. 1906 Morphogenie des Centralnervensystems der Saugetiere.

In O. Hertwig's Handbuch der vergleichenden u. experimentellen Entwicklungslehre der Wirbeltiere. Jena, Fischer.

101 ZiMMERMANN, K. W. 1891 Ueber die Metamerie des Wirbeltierkopfcs.

Verh. Anat. Gesell.

102 Terry, R. J. 1910 The morphology of the pineal region in teleosts. Jour.

Morph., vol. 21.

103 Dendy, a. 1910 On the structure, development and morphological inter pretation of the pineal organs and adjacent parts of the brain in the Tuatara (Sphenodon punctatus). Philos. Trans. Royal Soc, London, Series B., Vol. 201.

104 Herrick, C. J. 1910 The morphology of the fore brain in Amphibia and

Reptilia. Jour. Comp. Neur. Psych., 20.

105 Tandler, J. u. Kantor, H. 1907 Die Entwickelungsgeschichte des Gicko gehirns. Merkel u. Bonnet's Anatomischen Heften, Bd. 33.


THE AMERICAN JOURNAL OP ANATOMY, VOT.. 11, NO. 4


378


JOHN WARREN


Abbreviations


A. C. Aberrant commissure

D., Diencephalon

I. D., First diencephalic segment

//. D., Second diencephalic segment

D. C. P., Diencephalic choroid plexus E., Epiphysis

E. A., Epiphysal arch

F. B., Fore brain

F. M., Foramen of Munro

H., Hypophysis

H. F., Habenular flexure

L. C. P., Lateral choroid plexus

L. T., Lamina terminalis

L. v., Lateral ventricle

M., Mesencephalon


/. M., First mid-brain segment

//. M., Second mid-brain segment

M. B., Mid-brain

0. C, Optic commissure

P., Paraphysis

P. A., Paraphysal arch

P. E., Pineal eye

P. V. A., Post velar arch

P. C, Posterior commissure

S., Synencephalon

S. C, Superior commissure

S. S. S., Superior sagittal sinus

T., Telencephalon

T. P., Tuberculum postcrius

v., Velum transversum

Ve., Vein


PARAPHYSIS AND PINEAL REGION IN REPTILIA

JOHN ■WARREX


PLATE 1


M. p


M


§ IJ.Q LD


^^!x^^


'IDS. E.A.PVAV PA



T V.


'■D ll.D M


  1. t#


A


S E PYA.V. PA



— T


EXPLANATION OF FIGURES

1 Lacerta muralis, 1.8 mm. Harvard Embryological Collection, Sagittal series, no. 895. X 110 diams.

2 Lacerta muralis, 2 mm. H. E. C, Sagittal series, no. 903. X HO diams.

3 Lacerta muralis, 2.4 mm. H. E. C, Sagittal series, no. 906. X 110 diams.

4 Lateral view of fig. 3.

5 Lacerta agilis, 2.4 mm. H. E. C, Sagittal series, no. 590. X 110 diams.

(379)


PLATE 2


PARAPH YSIS AM) PIXEAL I{EGIOX IX REPTILIA

JOHN W.VnUEN



3 £ PE PVA V. R



EXPLANATION OF FIGURES


G Lacerla iiiuralis, 3.2 mm. II. E. C, Sagittal series, no. 865. X 110 diams.

7 Lacerta agilis, 3.6 mm. H. E. C, Sagittal serie.s, no. (iOO. X 110 diams.

8 Lacerta muralis, 4.5 mm. II. E. C, Sagittal scries, no. S.'iO. XllOdianis.

(380)


PARAPHYSIS AND PINEAL REGION IN REPTILIA


JOHN WARKEN


PLATE 3



10


EXPLANATION OF FIGURES


9 Lacerta muralis, 5 mm. H. E. C, Transverse series, no. 726. X 110 diams. 10 Lacerta muralis, 17.5 mm. H. E. C, Transverse series, no. 817. X HO diams.

(3S1)


PLATE 4


I'ARAPIIYSIS AND PIXEAL KEGIOX IX P.KPTrLIA

JOHN WARREN


PC PE. E S.C


P s.s,s.



12


EXPLANATION OF FIGURES


11 Lacertamuralis, 28mni. H. E. C, Transverse series, no. 809. X 110 diams.

12 Lacertaviridis, 37 mm. H. E. C, Transverse series, no. 604. XI 10 diams.

(382)


PARAPHYSIS AND PINEAL REGION IN REPTILIA

JOHN WARREN


PLATE 5



D B


EXPLANATION OF FIGURE

13 Lacerta muralis, adult, Transverse series. X HO diams.

(383)


PLATE 6


PARAPHYSIS AND PINEAL KEGION IN REPTILIA


JOHN WARRKX



EXPLANATION OF FIGURES


14 Same series as fig. 13, section C-58, line A-B, fig. 13.

If) Same series as fig. 13, soot ion C-72 and 73, lino C-D, fig. 13.

(384)


PARAPH Y.SIS AND PINEAL REGION IN REPTILIA

JOHN WARREN


PLATE 7



16


I.M. II.D. S.E.A. RVA. PA.



"\ ^..^T/'T



PC. S Ve. E. P.VA. V. P.



EXPLANATION OF FIGURES

16 Chrysemysmarginata,3.7mm. H. E.G., Sagittalsenes,no. 1023. XSOdiams.

17 Chrysemysmarginata, 4.8 mm. H. E. C, Sagittal series,no. 1048. XSOdiams.

18 Chrysemys marginata, 5 mm. H. E. C, Sagittal series, no. 1056. X 80 diams.

19 Chrysemysmarginata, 7.6mm. H. E. C, Transverse series, no. 1061. X 80

diams.

(385)


PLATE 8


PARAPPIYSIS ANT) JMXKAL REGION IN' REPTILIA

JOHN WARREX


PC SC E. ID v« 



L.C.P


H.


20


S.C E


DCP



21


EXPLANATION OF FIGURES


20 Chrysemys marginata, 8.8 mm. Sagittal series, no. 1433. X SO diams.

21 Chrysemys marginata, 16.7 mm. Frontal series, no. 1092. X SO diams.

(380)


PARAPH YS IS AXD PINEAL REGION IN REPTILIA

JOHX WARRKN


PLATE 9


D.C.P



L.C. P



EXPLANATION OF FIGURES

22 Chrysemys marginata, 27 mm. Transverse series, no. 1090. j_X 80 diams.

23 Chrysemys marginata, 31.7 mm. Sagittal series, no. 1101. X 110 diams.

(387)


PLATE 10


PARAPHYSIS AND PINEAL REGION IN REPTILIA


JOHN WARRKN


L.C.RL.V.


L.C.P


^^'



EXPLANATION OF FIGURES


24 Same Scries as fig. 22, section 157. X SO diains. Line A-B, fig. 22.

25 Same Series as fig. 22, section 168. X 80 dianis. Line C-D, fig. 22.

(388)


PARAPHYSIS AND PINEAL REGION IN REPTILIA


JOHN WARREN


PLATE 11



L.C.R



S.S.S.


27


EXPLANATION OF FIGURES

26 Same Series as fig. 22, section 176. X 80 diams. Line E-F, fig. 22.

27 Chrysemys marginata, .32 mm. H. E. C, Transverse series, no. 1653, section 229. X 500 diams.

(389)


EXPLANATON OF FIGURES

28 Lacerta muralis, 2.4 mm. H. E. C, Sagittal series, no. 906, section no. 49. X 110 diams.

29 Lacerta muralis, 2.8 mm. H. E. C, Frontal series, no. 866, section 30. X 110 diams. Line A-B, fig. 28.

30 Chrysemys marginata, 5 mm. H. E. C, Sagittal series, no. 1056, section 56. X 80 diams.

31 Chrysemys marginata, 5.5 mm. H. E. C, Frontal series, section 26. X 80 diams. Line A-B, fig. 30.

32 Eutaenia sirtalis, 3.7 mm. H. E. C, Sagittal series, no. 1558, section 32. X 110 diams.

33 Eutaenia sirtalis, 4 mm. H. E. C, Frontal series, no. 1422, section 46. X 110 diams. Line A-B, fig. 32.

34 Pig, 10 mm. H. E. C, Sagittal series, no. 414, section 101. X 40 diams.

35 Pig, 10 mm. H. E. C, Frontal series, no. 401, section 129. X 40 diams. Line A-B, fig. 34.

36 Sheep, 8.7 mm. H. E. C, Sagittal series, no. 1337, section 113. X 40 diams.

37 Sheep, 9 mm. H. E. C, Transverse series, no. 1231, section 265. X 40 diams. Line A-B, fig. 36.


(390)


PARAPHYSIS AND PINEAL REGION IN REPTILIA

JOHN WARREN


PLATE 12


M IID. ID. T



(391)


l'J>ATE 1:5


PAUAPIIVSIS A\D PI.VEAL REGIOX IX IJEI'TIl.IA

JOHN WAKREN


PC E SC


PVA



EXPLANATION OF FIGURES


38 Sheep, 26.1 mm. H. E. C, Sagittal series, no. 1112, section 222. X26 diams.

39 Human embryo, 28.S mm. H. E. C, Sagittal series, no. 1598, section 289. X 26 fliams.

(392)


MORPHOLOGY OF THE TUBULES OF THE HUMAN TESTIS AND EPIDIDYMIS

JOHN LEWIS BREMER Harvard Medical School, Boston, Mass.

TWELVE FIGURES

The intention of this paper is to show accurately the form of the seminiferous tubules of the testis and of the tubules of the epididymis, and to trace their development in man, especially in the late embryonic and fetal stages. With this study the blood vessels are so intimately associated that a description of them is added.

TESTIS

Many attempts have been made heretofore to decide whether the tubules are single with blind ends, or anastomosing, or merely branching, but the methods used gave contradictory results, and were unsatisfactory. Teasing methods are not convincing, because, in spite of the particularly tough reticular tissue described by Hill as encircling them, the tubules are easily broken ; and injections are never complete, as the injection mass is forced through the walls of the tubules before the resistance of the many convolutions is overcome. The method employed as a basis of this paper is the study of serial sections, from which usually wax reconstructions have been made. The material is human, though frequently the many embryos of pig, sheep, cat, rabbit, etc., in the Harvard Embryological Collection have given valuable assistance in interpreting the human material.

Allen has given us an account of the origin of the seminiferous tubules, of the rete testis, and of the connections of this latter with the tubules of the Wolffian body on the one hand and the

THE AMERICAN JOURNAL OP ANATOMY, VOL. 11, NO. 4

393


394 JOHN LEWIS BREMER

testis tubules on the other. His results, obtained by studying pig and rabbit, I have confirmed, with slight variations in man. Briefly stated, Allen's facts are these; the testis tubules originate as cords of epithelial cells containing germ cells, which grow inward from the peritoneal epithelium covering the middle third of the genital ridge; the rete is formed of similar cords growing inward from the anterior third of the same ridge. Both sets soon lose their attachments to the peritoneum, so as to lie free within the ridge. The rete cords, forming a network, grow into the mediastinum, extending caudally to reach the inner or central ends of the testis cords, with which they become joined. On their way, the rete cords unite with the glomeruli of the Wolffian body, thus completing the passages by which the products of the seminiferous tubules are later carried away from the gland. The testis cords Allen described as anastomosing and branching, and occasionally growing parallel to the surface.

Further detailed study of these testis cords gives rather surprising results. Instead of branching and anastomosing irregularly, as suggested by Allen's description, the cords form a complete network, every cord anastomosing with others, leaving no free ends except those at the periphery and those near the mediastinum. The bases of the cords, at the periphery of the gland, form free ends when they have lost their connection with the peritoneal epithelium from which they grew; and the distal ends of the cords, which, since growth is centripetal, are found near the mediastinum, are also usually free ends, though occasionally two may join at their tips. Otherwise all the cords are joined by anastomosing branches. As a whole this network is crescentic in cross section, occupying a peripheral zone of the genital ridge, of which the mediastinum is the eccentric core; this brings the central ends of the cords nearer together, and accounts for the occasional anastomosis of their tips.

Although this network (fig. 5) seems at first sight to be quite irregular, a more critical study shows that each of the cords has three (occasionally four) sets of branches, so that there are three sets of cross connections joining the radially disposed cords. One set of branches is given off a little distance from the peritoneal


HUMAN TESTIS AND EPIDIDYMIS


395


epithelium, and the branches run more or less parallel to the surface, as described by Allen; the second and third sets are given off respectivel}' nearer the mediastinum. Beyond the third set of branches the cords grow centripetally without further branching. The figure or pattern thus produced, which is given in a very much simplified and idealized form in figure 1, results apparently from the fact that the testis cords possess a normal rate of branching, and are moreover limited in length by the thickness of the genital ridge. Three, or possibly four, sets of branches, a certain distance apart, are all that each cord produces.



Fig. 1 Diagram of testis network of human embryo of 20 mm. Outer dotted line represents germinal epithelium, solid lines represent testis cords.


The cords forming this network vary in diameter, and, though usually approximately round, in certain places they are flattened, forming plate-like structures, often where three or four cords join. Sometimes these plates are pierced, making larger or smaller rings.

This network is completed shortly after the cords have become detached from the peritoneal epithelium, and before the central tips have joined the rete cords; in the human embryo this corresponds to a length of about 20 mm. to 22 mm. Further growth takes place in two ways; by the increase in diameter of the cords, and by the increase in thickness of the genital ridge, caused by the lengthening of the radially disposed cords, not at their ends, but throughout their whole extent, so that the cross connections are


396 JOHN LEWIS BREMER

more widely separated. No more branches are produced. With this there occurs, apparently, an absorption of the peripheral free ends into the network, leaving the outer set of cross connections as a series of arches, joining the ends of the radial cords (fig. 6). Also, since the cross connections do not lengthen so much as the radial cords, the network assumes a distinctly radial appearance.

There now occurs a partial destruction of the network. Although there is a general increase in the diameter of the cords, certain ones, usually, but not exclusively, those forming cross connections, remain of their original size or even become smaller. Many of these attenuated connections soon become severed, strand after strand, and the loose ends are absorbed into the network. This partial destruction of the network goes on for a long time, in the human embryo certainly from 22.8 mm. to 9.1 cm., perhaps longer; in the later stages, when the cords have become more established, the loose ends are not usually retracted or absorbed, but remain as short knobs or as long branches with blind ends.

The results of this process may be seen by comparing the models of tubules from embryos of 37.0 mm. and 9.1 cm. (figs. 6, 7, 8 and 9), a full description of which will be given later. At a glance, the destruction of the network and the consequent isolation of certain cords can be easily traced, as the figures are still uncomplicated by convolutions.

During this time the inner or central ends of the radial cords have come into contact with the rete cords, which also have formed a network. The rete network, the 'Keimdriisennetz' of Mihalkowicz and other older writers, is quite irregular, of small mesh, and persists throughout life. It occupies the mediastinum testis, and in the lower two-thirds of this spreads out in a fanshape, filling the space enclosed by the mass of the testis cords, which is itself cresentic in section. A single testis cord may come directly in contact and join with the rete network, or peripheral rete cords may extend far into the area of the testis network, so that the boundary line between rete network and testis network is irregular and wavy (fig. 6). These extensions probably indicate the position of the septa of the adult testis, along which the tubuli


HUMAN TESTIS AND EPIDIDYMIS 397

recti often run for some distance before joining the seminiferous tubules. Frequently two testis cords anastomose just before joining a rete cord; on the other hand, one testis cord maybe connected with several rete cords.

In regard to the age at which the testis and rete cords join, I find such differences between my findings in man and Allen's in pig and rabbit that they seem worthy of note. Allen gives the time of junction as about 13.0 cm. in the pig, and 21 days in the rabbit ; in both cases the rete cords were already hollow before joining. In man the development of this connection is much more rapid, though there seem to be quite wide individual variations. At 16.0 mm. there is no extension of the rete cords downward, while at 23.0 mm. the cords have already grown past the upper glomeruli into the mediastinum , and in embryos of 32.0 mm. have already united with the testis cords. (In one embryo, H. E. C, no. 819, of 19.0 mm., this union has taken place.) For some time after joining, the rete cords in man remain solid, without lumen. This precocious development of the rete cords in man may be correlated with the small size and rapid degeneration of the mesonephros, as compared with that of pig and rabbit. In the sheep, another embryo with large mesonephros, the rete again develops late; whereas in the cat, whose mesonephros is small, the rete cords and testis cords have nearly joined at 24.0 mm. (H. E. C, no. 467).

By the end of the third month or the middle of the fourth the rapid destruction of the testis network probably ceases, though many connections may be severed much later; my preparations give no information on this point. The cords become so long that they are forced into convolutions, which increase progressively till puberty; on the other hand the cords decrease in diameter, becoming more and more slender until at seven months they are of about one-half as great diameter as at three months. From seven months on there is a gradual increase in calibre. This reduction in size may be due to a rearrangement of the cells to allow for the rapid increase in length. The convolutions are in short, stiff curves, which remain within a small area, condensing the connective tissue around them. It thus happens that different parts of


398


JOHN LEWIS BREMER


the same tubule are isolated from each other, and lie in compartments, which can be easily recognized in the adult testis, marking subdivisions of the parenchyma between the septa. Jj^

The greatest complexity of convolutions is in the peripheral part of the gland, and apparently in the cross connections, not in the radial cords, which latter may frequently be seen in the adult running from the rete with only a slightly wavy course to a con


Fig. 2 Diagram of the course of" several tubules in testis of seven months fetus, made by noting their connections and the positions of the various branches in the gland. The original network is represented by fine lines, the permanent portions of the tubules by heavy lines; rete connections at bottom of figure.


voluted portion situated near the outer surface of the gland. The diagram (fig. 2) gives the approximate course of several tubules in a fetus of seven months and suggests from what part of the original network they have been derived, while the model (fig. 11) shows the actual form of some of the tubules at the same age. In the adult I have been unable to follow completely any single tubule but from the portions studied I feel convinced that, except for


HUMAN TESTIS AND EPIDIDYMIS 399

the greater number of convolutions, the conditions are similar to those found at seven months.

From the diagram we see that there are to be found tubules with no connections whatever, except that with the rete, ending blindly; others with several branches which end blindly; others anastomosing with their neighbors. In one a short blunt knob, x, was seen (found also in the adult testis) such as has been described as commonly present by some authors. But to me the most interesting part of the diagram is the preservation, at seven months and probably in the adult, of the course, connections, and position of each tubule as determined for it by the original network of the testis cords.

THE BLOOD VESSELS

Hill has described the development of the blood vessels of the testis in the pig, and in a later work has apparently taken for granted that, although the adult arrangement in pig and man differ widely, the early development is similar in each case. In the pig the spermatic artery arises as a separate vessel from the dorsal aorta, caudal to the last mesonephric artery, or rarely as a branch of the latter. It makes its way horizontally to the mesial border of the Wolffian body, and then turns upward, toward the head, passing mesial to the mesonephric arteries, and crossing over the last five or six of them, to reach the genital gland ; it thus approaches the testis from the caudal end. In man there is no such vessel formed. The arteries to the Wolffian glomeruli in the region of the future testis send branches to supply the gland directly, so that at first there are many spermatic arteries, which enter all along the attachment of the testis to the Wolffian body.

On examining embryos of other mammals in regard to the origin of their spermatic arteries, I find that in sheep, rabbit, cow, and deer (cervus capreolus) a separate vessel is formed to supply blood to the genital gland, whether it be testis or ovary; while in the cat the arteries to the glomeruli send branches directly to the gland. In sheep, the new artery is more apt to arise as a branch of the last mesonephric artery than as a direct outgrowth from the aorta, and in one deer embryo of 19.6 mm. (H. E. C, no 1230)


400 JOHN LEWIS BREMER

one ovarian artery is a branch of the mesenteric artery. I find here, then the same grouping of animals that occurred when the time of junction of rete cords and testis cords was under consideration; species with large Wolffian bodies (so far as my limited studies show) provide a new vessel for the genital glands, while those with small Wolffian bodies utilize branches from the nearest arteries.

In man and cat the mesonephric arteries anastomose freely with each other before entering the glomeruli, and with the early degeneration of the glomeruli the number of mesonephric arteries diminishes gradually, until one only is left; this one is, however, connected with all the arteries of the testis, and so becomes the single spermatic artery. The factors in this decrease in number of the Wolffian arteries seem to be the descent of the testis, which stretches them into long parallel vessels, and the ingrowth of the cords which are to form the cortex of the suprarenal body, which occurs directly in the course of these vessels, and presses upon them. Incidentally it may be mentioned that small pieces of this suprarenal tissue are often carried down with the lengthening arteries, and left as the small aberrant glands not infrequently found on the posterior wall of the abdominal cavity, along the course of the spermatic artery. In a human embryo of 37.0 mm. (H. E. C, no. 820) there are two complete spermatic arteries in each side, and four or five arterial stems parallel to the main arteries either joining them or ending blindly, evidently recently obliterated. In an embryo of 44.3 mm. (H. E. C, no. 293) one artery on each side remains, with two or three obliterated pieces beside it. In both embryos suprarenal tissue appears along the course of the arteries. This method of arriving at a single spermatic artery on each side in man accounts for the wide range in its point of origin in the adult, as described in the text-books of anatomy.

The veins of the testis in all the mammals examined arise as simple offshoots from the sinusoids of the Wolffian body in the neighborhood of the genital gland.

The blood vessels of the testis in man, then, arise as two capillary networks, one from the branches of the efferent arteries of the


HUMAN TESTIS AND EPIDIDYMIS 401

glomeruli of the mesonephros, the other from the sinusoids. Both sets interdigitate with the network of cords, and extend beyond the the outer cross connections of the network so as to lie just underneath the peritoneal epithelium. The flow of blood is in two directions from the hilus or mediastinum testis toward the periphery, and from the periphery toward the center; the veins also return the blood in both directions. There are at first, then, no terminal arteries or veins. As certain cords become destroyed, the capillaries lying near them are allowed to assume a straighter course; they then become the more favored vessels, grow larger and are established as main arteries or veins. Since the cross connections of the network are those most frequently severed, the radial vessels are the most favored, and hence the main arteries and veins of the testis run radially. But not infrequently, as we have seen, the radial cords are severed, and this fact accounts for the few main vessels which, though not figured by Hill, are commonly present in the adult testis, running diagonally or even for some distance parallel to the surface of the gland, quite deep within the substance. Another curious arrangement of vessels not mentioned by Hill, is found in the testis ; three or four arteries run parallel to one another for long distances to supply an area which would usually be served by a single artery with short branches. To explain this it is only necessary to imagine that certain cross connections of the cord network which may at first have separated the different vessels quite widely, were destroyed late, after each vessel was well established, and that subsequent radial growth drew the vessels together.

The terminal arteries are at their first appearance probably portions of the capillary network not favored by a direct course. In the fourth month three sets of terminal arteries can be made out, one set situated between the outer and second sets of cross connections of the testis cords, another between the second and inner sets of cross connections, and a third set nearer the rete. At seven months new arteries have grown from these, and also apparently from the main stems of the radial and peripheral arteries, so that the picture is much complicated; yet even in the adult, the embryonic arrangement of three main branches from each radial artery


402 JOHN LEWIS BREMER

can be traced in some places. (Hill, fig. 12.) In both the fourth and the seventh months a vascular unit can be made out, as the veins are arranged in a network around the terminal arteries in the usual manner. Hill mentions "vascular units which correspond to units of structure and which repeat themselves similarly throughout the organ," but does not state what these units of structure are; I have found in the testis of seven months and in the adult that the structural unit corresponding to the vascular unit consists of a number of coils of a single tubule enclosed in a compartment, as described earlier in this paper. The border veins of the unit lie in the connective tissue which surrounds the convolutions, the terminal artery pierces the compartment. It is probable that the terminal artery is the causative factor in this unit ; the portion of a tubule situated nearest to the artery would be more favorably placed for growth and consequent convolution, than the portions of the tubule further from the blood supply; the convolutions would therefore form around the arteries, one set for each terminal artery. These units are quite large, readily visible to the naked eye; the capillaries surrounding the tubules are therefore of considerable length. As structural units they are not typical, like the unit of the lung or of the salivary glands, since they only very occasionally represent the terminal, blind ends of the secreting or active portion of the gland, and since there is no constant relation between the terminal artery and the channel through which the secretion leaves the unit. One tubule passes through several units, but all the convolutions within a unit belong to the same tubule.

The peripheral layers of the original capillary networks, both arterial and venous, interdigitate with the peripheral portions of the cords, while these portions are still attached to the peritoneal epithelium from which they grew. When this connection is lost and the peripheral ends of the cords have been absorbed into the network, the peripheral vascular networks remain, and become incorporated in the vascular layer of the tunica albuginea ; the main vessels are given their prominence by their direct connections with the main radial vessels within the testis. In the ovary, this


HUMAN TESTIS AND EPIDIDYMIS 403

same peripheral layer of the vascular network is buried within the organ, and becomes the system of arched vessels found between the medulla and cortex. For in the ovary, the original or medullary cords degenerate, and new cords (Pfliiger's cords) grow from the surface, carrying before them the vessels which lay just beneath the surface. From these new vessels grow peripherally, while the network, which supphed the medullary cords for the most part is lost. The terminal arteries of the ovary, then, like the real sexual cords, are secondary affairs when compared with similar structures in the testis.

EPIDIDYMIS

The precursors of the epididymis are the anterior mesonephric tubules, which form the ductuh efferentes, and the anterior portion of the mesonephric duct, whose convolutions form a large part of the head and the tail of the epididymis, while the posterior part remains unconvoluted and forms the ductus deferens. As in the case of the testis tubules, this paper deals with the more detailed morphology of the epididymal tubules in man, especially in the late embryonic and fetal stages.

Perhaps the most accurate accounts of the Wolffian tubules are those of MacCallum, who studied pig, and, less thoroughly, man, and Grafe, who worked with chick material. According to MacCallum the tubule of a fully formed Wolffian body, in the pig, is a long affair, running from the glomerulus in sweeping curves, from mesial to lateral border of the organ (see his text-figure no. 8). Some of these tubules were seen to branch soon after leaving the Wolffian duct, others just before entering the glomeruU. Evidences of anastomosis and the formation of networks of tubules were also made out," particularly in the region of the dorsal border. MacCallum made models from serial sections and also injected fresh material; the anastomoses and networks were found by the latter method. Grafe also found braiiches of the tubules, and affirms that they indicated new tubules which have grown by budding. He also made a point of the fact that some of the tu


404


JOHN LEWIS BREMER


bules enter the duct on its dorsal aspect, some on its ventral. In the chick the tubules are not so long nor so convoluted as in the pig. In man also the tubules seem to be of two more or less distinct groups, one entering the mesonephric duct on the ventral, one on the dorsal side, and these two groups run with few convolutions along the ventral and dorsal borders of the gland respectively to the glomeruli, which lie far dorsally. The tubules of the ventral set thus approach nearer to the genital anlage than their respective glomeruli, but this is not true of the dorsal set of tubules. None of the tubules in man ever attain the extreme length found in pig, sheep, rabbit, and other animals which retain functioning Wolffian bodies to an advanced embryonic age. MacCallum found the full formation of the Wolffian body in the pig at between 40 mm. and 95 mm., while in man he noted a reduction of the number of tubules after 12 mm. This latter statement, however, I cannot reconcile with what I have found in the embryos of man in the Harvard Embryoligical Collection, unless very wide individual differences occur. MacCallum counted 27 tubules in one Wolffian body in a human embryo of 12 mm., 20 in one of 14 mm., and 9 in one of 20 mm. ; while, as the following table shows, I have found no constant reduction in the number of tubules up to 44 mm.


Number oj tubules in one Wolffian body


LENGTH OF EMBRYO


H.E.C. NO.


TUBULES


4.0


714


23


7.5


256


34


8.0


817


28


9.1


734


35


9.4


529


37


10.0


1000


34


11.5


189


30


12.0


816


27


16.0


1322


30


19.0


819


25


22.8


871


31


37.0


820


33


44.3


293


32


HUMAN TESTIS AND EPIDIDYMIS 405

These numbers are not affected to any appreciable extent by the formation of new branches, as in no case have I found more than five or six of these in one Wolffian body. That the tubules must remain as functioning and useful parts of the embryo longer than is suggested by MacCallum will be evident when we consider that the kindey in man is only beginning to be provided with glomeruli and convoluted tubules in an embryo of 20 mm. ; but there seems to be no constant relation between the size of the Wolffian body and the growth of glomeruli in the kidney, since the kidney in pig is fully as far advanced at 20 mm. and in later stages as in man, in spite of the much larger and longer lasting Wolffian body. Hill agrees with Pohlman, that the vascularization of the human kidney takes place between 25 and 30 mm., a little later than the presence of glomeruli would indicate, and gives the size for pig embryos as 28 mm. The cause of the continued growth and activity of the mesonephros in the pig and several other animals, even after the kidney is apparently able to act as an excretory organ, is a subject which I shall have to leave for future investigation.

Of these mesonephric tubules the 6th to the 20th in pig, the 12th to the 20th in rabbit lie, according to Allen, opposite the rete region, and presumably (though it is not so stated by him) join with the rete cords. In man the rete region is more cephalad, opposite the first eight or nine glomeruli ; but the rete cords anastomose not only with these but with many others in their course down the mediastinum. Occasionally the first one or two glomeruli do not join the rete, and remain as small cysts, losing their tubules (fig. 3). Such isolated glomeruU would give rise to the appendix epididymis, described by Toldt as being present in 27 per cent of cases examined. The disconnected tubules of such glomeruli would end blindly, as shown in fig. 3 and 4, and in fig. 12, and would ordinarily lie inconspicuously among the convolutions of the epididymal duct ; if the upper glomerulus and its tubule were separated by a considerable distance from the others, as I have seen it in two of the embryos studied, this blind tubule might form the infrequent lower paradidymis of Toldt, a single tubule in a connective tissue sheath lying behind the head of the epididymis.


406


JOHN LEWIS BREMER


The number of mesonephric tubules which join with the rete, as described by Allen and others, varies in the specimens examined from eleven to nineteen or twenty. The rete cords do not always meet the glomeruli, but in man not infrequently connect with the proximal part of the mesonephric tubule. This is the result of the course taken by the ventral tubules in the human mesonephrost as described above, for the rete seems to join with the neares, part of the tubule. Tubules thus tapped along their course are,



Fig. 3 Diagram of epididymal tubules of a fetus of 10 cm. To show slight branching and disconnected upper glomerulus. Many glomeruli have entirely disappeared; rete not indicated.


by a rearrangement of their parts, brought to seem like two tubules each arising from the rete, one running to the duct, the other ending blindly, usually with an expanded end. Such a blind tubule may be seen in fig. 4, 2, and in a fetus of seven months I have traced five or six similar tubules. It is probable that these are the tubules of the appendages of the rete testis, as described by Roth and Poirier, and also the upper, shorter ductulus aberrans,


HUMAN TESTIS AND EPIDIDYMIS


407


described by other writers as opening into the rete and ending bhndly. Roth and Poirier regarded them as tubules which, after acquiring a union with the rete, lost their connection with the Wolffian duct; but the presence of the blind tubules I have de


V


— 6

— 7

— 8

— 9


10 11

-12 13 14

15

-16


Fig. 4 Diagram of epididymal tubules of a seven months' fetus. Junctions with rete indicated by fine lines.

scribed renders unnecessary such an unlikely supposition as the separation of the tubule and the duct after a new channel has been formed. These tubules may lie inconspicuously among their neighbors or, as in the seven months fetus, be grouped together in a separate sheath of connective tissue.


408 JOHN LEWIS BREMER

In several embryos and in a fetus of three months and another of seven months I have found tubules lymg in the rete region and yet unconnected with the rete cords (fig. 4, 4 and i 5). In the older fetus such tubules are of smaller diameter than the ductus epididymis, but have a similar epithelium. If separated from their neighbors, these tubules would form the lower ductulus aberrans. The upper paradidymis of Toldt (organ ofGiraldes) is probably correctly described by him as mesonephric tubules lying below the rete region but maintaining their connections with the duct.

While not dealing in this paper with the histological differentiation of the epithelium lining these tracts, I may mention that this differentiation seems to depend not on the portion of the mesonephric tubules or duct which gives rise to any tubule in the adult, but on the connections which are permanently established. At three months the epithelium of the tubules and duct is similar, all trace of former differentiation having disappeared; at about seven months a new differentiation takes place, but this time all the tubules connected with the rete show a similar epithelium, all tubules connected with the duct have duct epithelium. Thus tubules of like origin may at seven months and later be lined by different kinds of epithelium. The similarity of the epithelium in the duct and the blind tubules emptying into it seems to me to point distinctly toward a secretory function of this coiled tube.

As will be seen from the diagrams, (which have been compiled after carefully following each tubule in serial sections of the organs, and which have been offered instead of models or actual reconstructions, as showing more clearly the courses and connections of the different tubules), there are several cases of branching in each epididymis, as found by MacCallum and Grafe in the Wolffian body. In only one fetus was an anastomosis found (fig. 4, 1 and 2), and the formation of a more considerable network was nowhere seen. It is probable that tubule 5, in the same diagram, originally anastomosed with tubule 4 or some other, since in this way the position of the persistent glomerulus may be explained, by imagining two tubules running to the same glomerulus, only


HUMAN TESTIS AND EPIDIDYMIS 409

one of which joined with the rete. But anastomoses among the vasa efferentia of man must be considered as of rare occurrence. Convolutions of the vasa efferentia and the ductus epididymis Dogin to appear at about the fourth month of fetal life, as was found to be the case with the testis tubules. Here also the convolutions are in short, stiff curves, and here also certain portions of the tubules form groups of coils, joined by unconvoluted portions, each group ultimately developing a vascular unit of its own. In the case of the coni vasculosi, single vasa efferentia are usually separated by connective tissue, though occasionally two or more may be intertwined; each conus contains several units. In position the coni are of two distinct groups, lying mesial and lateral respectively to the convolutions of the ductus epididymis. This arrangement, not describea in the text-books, seems to be due to the two sets of mesonephric tubules which enter the duct on its ventral and dorsal aspect respectively, as described above; the ventral tubules form the lateral group of coni vasculosi. The head of the epididymis thus shows three main lobes, more or less distinct, the middle lobe containing the ductus epididymis.

CONCLUSIONS

1. The testis cords, growing from the germinal epithelium of the genital ridge, form a network with three sets of anastomosing branches. After completion, this network breaks down partially, leaving certain cords as persistent stems. The tubules of the adult show, in their course, connection, and position in the testis, traces of this network. Testis tubules may be single, ending blindly, may branch, or may anastomose.

2. The unit of the testis is a considerable number of coils of one tubule, enclosed within a sheath; there are many units for each tubule, connected by less convoluted portions.

3. The spermatic artery is not a special vessel, as in the pig, etc., but the survivor of the mesonephric arteries in the genital region. The others were obliterated by stretching and by the

THE AMERICAN JOURNAL OP ANATOMY, VOL. 11 , NO. 4.


410 JOHN LEWIS BEEMER

growth of the cortex of the suprarenal gland. Pieces of this latter are common along the course of the spermatic artery.

4. The mesonephric tubules in man join the duct on either its dorsal or ventral side. The dorsal ones run dorsally, so that the rete tubules join their glomeruli; the ventral ones take a more ventral course, so that the tubules before reaching the glomeruli pass by the mediastinum testis, and are joined by the rete tubules. The glomerular ends of the ventral tubules form the appendages of the rete testis, (Roth and Poirier), and the upper ductulus aberrans.

5. The rete tubules in man develop opposite the first eight or nine mesonephric glomeruli, but are connected with many more in their course downward in the mediastinum. The first one or two may remain unattached, forming the appendix epididymis, their tubules making the lower paradidymis (Toldt). Tubules below the junction of the rete form the lower ductulus aberrans and the organ of Giraldes.

6. The small percentage of cases in which these appendages are found is due to the fact that the tubules involved frequently lie inconspicuously among the convolutions of the normal ducts.

7. The epithelium lining these appendages depends upon their final connections, not upon their origin.


LITERATURE CITED

Hill, E. C. 1909 The vascularization of the human testis, Am. Jour. Anat. vol. 9, no. 4.

1907 On the gross development and vascularization of the testis, Am. Jour. Anat. vol. 6, no. 4,

Allen, B. M. 1904 The embryonic development of the ovary and testis in mammals. Am. Jour. Anat., vol 3, no. 2,

MacCallum, J. B. 1902 Notes on the Wolffian body of higher mammals, Am. Jour. Anat., vol 1, no. 3,.


HUMAN TESTIS AND EPIDIDYMIS 411

Grafe, E. 1905 Beitrage zur Entwickelung der Urniere und ihrer Gefasse beim Hunchen. Arch. f. Mikr. Anat. Bd. 67.

Hill, E. C. 1905 On the first appearance of the renal artery, etc., in pig embryos, Johns Hopkins Hosp. Bui. vol 16, no. 167.

PoHLMAN, A. G. Concerning the embryology of kidney anomalies, Amer. Medicine, vol. 7, no. 25.

ToLDT, in von Langer and Toldt's "Anatomy," p. 384 et al.

Roth de Basle 1876 His u. Braune's Zeitschrift, p. 125,

PoiRiER, P. 1890 Congres Internat. de Med. Berlin, references given in Poirier and Charpy, Anat. Humaine, tom. 5, 2nd edit.


412


JOHN LEWIS BREMER



Fig. 5 Model; testis of human embryo of 22.8 mm. (H. E. C, no. 871). A segment of a transverse slice is shown; the limit of the genital ridge and the outer border of the mediastinum testis are indicated by dotted lines. The proximal or peripheral ends of the cords have already lost their attachment to the peritoneum; the distal or central ends are seen reaching toward the mediastinum, in one case uniting at their tips. Except for these two sets there are no free ends, each branch forming an anastomosis with others; the cut surfaces represent connections beyond the extent of the model. The arrangement of three cross connections shows best at the two cut edges of the model; plates and ring formation arc also to be seen. X 180 diam.


HUMAN TESTIS AND EPIDIDYMIS


413




Alf^c


/


/



Fig. 6 Model; testis of human embryo of 37.0 mm. (H. E. C, no. 820). Orientation same as in fig. 4. Peripheral ends of cords have been absorbed, leaving a series of arches as the outer border of the figure. The rete network has already joined the cords, and rete cords can be recognized by their small diameter; the lower dotted line indicates their irregular extension. Other cut ends represent, as before, anastomoses beyond the limit of the model. Fewer cross connections, more radial disposition of cords. X 180 diam.


414


JOHN LEWIS BREMER


a



r^ \\\\


8



4


J


Figs. 7, 8 and 9 Models; human fetus of 9.1 cm., age given as three months. Tubules from different parts of same testis; figs. 8 and 9 two views of same model. The rete cords are slender with cut ends. The large cut ends represent anastomoses with tubules not modeled. Loops formed by radial tubules and cross connections are seen, some including the peripheral set of cross connections, some the second set; while shorter connections of the central set can be made out nearer the mediastinum. In fig. 8, at «, a peripheral loop is just breaking apart; in fig. 7, at X, another has just been severed. At h and c, in figs. 8 and 9, the tubules are very small and will probably part in a short while. Tubule A is unconnected except near the rete, and consists of a radial cord with the greater part of a peripheral loop. Tubule B has a short anastomosing branch representing the inner set (at e) and a looped end consisting of the outer two sets of connections and the part of the radial tubule between them, the rest of which has been lost. Tubule C has all three sets of connections represented. Ring formation can be seen at ?•. X 90 diam.

Fig. 10 Model; human fetus of about 10.0 cm., age given as lOG days. Convolutions have begun, chiefly in the cross connections ; the tubules have become of nearly even diameter. The only blind ends are at x and ij. Tubule A is connected with three rete tubules, and extends only to (he inner cross ooimection, which can be traced through a few convolutions to another radial limb, also without branches till near the rete, a separate short loop is thus made. C'ross connections belonging to the other two sets are recognizable, and can be traced easily in the actual model. X 90 diam.


HUMAN TESTIS AND EPIDIDYMIS


415



10


416


JOHN LEWIS BREMER


PLATE 1

Fig. 11 Model; human fetus of seven months. Two tubules of the testis with their connections. The tubules are colored red and yellow, (tubules A and B). Cross connections between them, of which there are two, are colored two shades of orange and represent the outer and middle sets. From the outer cross connection come two branches, one anastomosing with other tubules, not modeled, the other ending blindly at x. From the other cross connection there is also an anastomosing branch, a; while other branches from tubule B, which should be considered as belonging to the middle set, are seen at b and c. The inner set of cross connections is represented by branches y and z, of which y ends blindlj-. Tubule B joins another before meeting the rete tubule (fig. 5) while two rete tubules connect with tubule A. The group of tubules in yellow, between the middle set of cross connections and the periphery, form a unit; a single artei-y supplies them, and a network of veins surrounds them, lying partly between them and the tubules in orange. At r a ring is seen in the course of tubule .1. X 90 diam.



12

Fig. 12 Reconstruction; epididymis of human fetus of about 10 cm., age given at 106 days. The Wolffian duct is shown with fifteen Wolffian tubules ojjcning into it one of which is traced to the rete (shown by the fine line), the others represented as cut short. The upper end of the duct is probably the first tubule which has failed to unite with the rete. The connections of all the tubules is shown in fig. 4. X 40 diam.


HUMAN TESTIS AND EPIDIDYMIS

JOHN LEWIS BREMER


PLATE 1



THE AMERICAN JOURNAL OF ANATOMY, VOL. 1 1 , NO. 4


FURTHER STUDIES ON THE VARIATION IN SUSCEPTIBILITY OF AMPHIBIAN OVA TO THE X-RAYS AT DIFFERENT STAGES OF DEVELOPMENT

CHARLES RUSSELL BARDEEN Fro7n the Anatomical Laboratory, University of Wisconsin

EIGHTEEN FIGURES

CONTENTS

Introduction 420

1 Effects of exposure of spermatozoa to the x-rays 435

1 Fertilization 436

2 Period of cleavage 437

3 Gastrulation 438

4 Period of larval differentiation 439

General effects 440

External form 441

Internal structure 441

5 Tadpole stage 445

Conclusions 446

2 Effects of the exposure of mature ova to the x-ray 450

1 Fertilization 451

2 Period of cleavage 452

3 Period of gastrulation 452

4 Period of larval differentiation 453

5 Tadpole period 454

Comparison of the results of exposure of spermatozoa and of

mature ova 454

3 Action of the x-rays on fertilized ova and on larvae : 457

Introduction 457

1 Period of fertilization 458

2 Period of relative immunity 462

3 Early cleavage stages 464

Effects of temperature 472

4 Advanced cleavage and gastrulation stages up to the closure of

the blastopore (12 to 36 hours after fertilization) 474

5 Period from the closure of the blastopore to the period of hatching 481

6 From the period of hatching to the period of metamorphosis 482

7 Period of metamorphosis 485

4 Summary of experiments 486

Bibliography 492

419

THE AMERICAN JOURNAL OF ANATOMY, VOL. 11, NO. 4.


420 CHARLES RUSSELL BARDEEN

INTRODUCTION

In a brief communication presented at the 1908 meeting of the Association of American Anatomists and published in the Anatomical Record, April, 1909. I have summarized the results of experiments which showed that spermatozoa, ripe ova, and newly fertilized ova of the frog and toad are very susceptible to the X rays, that the susceptibility decreases early in the second hour after fertihzation, then rapidly increases to a maximum during the earlier cleavage stages, and finally greatly decreases during the period of gastiulation. In 1909 and 1910 further experiments have served to confiim these results and to throw some new light on the nature of the effects of X rays on living tissues. In order that the more general bearings of these experiments may be the better appreciated it seems advisable, before describing the experiments in detail, first brieflj^ to review work of various investigators which seems to give insight into the fundamental features of the action of the X rays on protoplasmic activity.

The activities of protoplasm may be divided into three correlated groups, sensory-motor, metabolic, and morphogenic.

Exposure to the X rays apparently does not directly disturb the sensory-motor activities. Motile unicellular organisms and the spermatozoa of the higher organisms seem to move about as freely when exposed to the rays as when not thus exposed. ^

Joseph and Prowazek have described in Paramoecia and Daphnia a negative tropism toward the Roentgen rays. This is certainly not well marked in the paramoecia with which I have experimented. Paramoecia exposed for twelve hours to the rays showed no disturbance in freedom of movement either during exposure or subsequently. Muscular and ciliary activity in planarians exposed to the X rays for considerable periods was apparently not directly affected by this exposure. Specimens

iBohn, Comptes rendus de I'Acad, de Sciences T. 136 1903, p. 1085 states that 171 vitro radium rapidly causes cessation of motion in the spermatozoa of the sea urchin, llertwig, 1910, on (he contrary finds that IG to 23 iiours exposure to radium does not affect the motility or the fertilizing capacity of the spermatozoa of soa urchins.


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 421

repeatedly exposed to the X rays continued to react normally to light, to mechanical and chemical (food) stimuli. Tadpoles repeatedly exposed show no direct impairment of movement. On the other hand Barratt ('10) has recently found that under a disc of radium bromide the skin of the rabbit after two and a half hours exposure becomes pale while about the margin of the disc a circle of pigment is formed. This probably indicates that pigment carrying cells are stimulated to movement by irradiation. Casemir ('10) described a cessation of nuclear and cell division in germinating Vicia faba after two and a half hours exposure to the Roentgen rays.

The metabolic activities of protoplasm cannot well be directly followed. Since active life cannot persist without constant metabolism and since exposure to the rays does not as a rule make its effects visible for some time after exposure, it seems fair to conclude that the simpler metabolic functions are not immediately affected by the X rays. In paramccea exposed for twelve hours to the rays I have found no apparent disturbance of metabolism either during or subsequent to the exposure. M. Zuelzer found much variation in susceptibility to radium in different species of protozoa, when observed under the microscope during exposure. In those showing direct injury the nucleus appeared affected before the cytoplasm. Planarians repeatedly exposed to the X rays subsequently die (usually about a month after the first severe exposure). The first effects are seen in the region of the head in which there is a gradual degeneration, apparently in part parasitic in nature. It is uncertain to what extent these effects are to be attributed to interference of the simpler metabolic activities but it seems likely that they are to be attributed chief y to interference with those morphogenic activities which have to do with the restitution of worn or injured parts. In mammals Lepine and Boulud have described an increased amylasis after exposure of the pancreas to the X rays, and in the liver and blcod after moderate exposure an increased glycogenesis and glycolysis. After prolonged exposure both are diminished. Bearmann and Linser have described an increased elimination of nitrogen after severe exposure to the X rays. Irradiation has been used sue


422 CHARLES RUSSELL BARDEEN

cessfully in reducing secretion of the sweat glands in mjin (Pusey). The mode of action in these cases is uncertain.

It is undoubtedly the morphogenic protoplasmic activities which show the chief effects of exposure of living things to the X rays or to radium. These morphogenic activities may be subdivided into reparative, reproductive and differential or evolutionary activities. The reparative activities have to do with the restitution of worn or injured structures; the reproductive, with the multiplication of like individuals; the differential, with the organization of daughter individuals varying in structure to a greater or less extent from that of the parents. Unicellar organisms multiply largely by simple reproductive morphogenesis, while the cells in the bodies of multi-cellular organisms undergo extensive, though specifically determined differentiation. In some tissues, as in the nervous system, cell differentiation may lead finally to a loss of reproductive power, although not to a loss of reparative potentiality. In other tissues, as in the epithelium, the bone-marrow and the generative epithelium of the testicles, cell multiplication accompanied by specific differentiation of certain of the daughter cells continues through life.

The various types of morphogenesis, the reparative, reproductive and differential, do not seem to be equally susceptible to the X rays.

I know of no experiments made to test the effects of exposure on reparative activity unaccompanied by a reproductive or differential cellular morphogenesis. The test could most easily be made by removing a part of the body of a unicellar organism and then exposing it to the X rays. Since most unicellar organisms, even after prolonged exposure, readily multiply by fission and the daughter cells have no apparent difficulty in assuming the parent form it would seem probable that such organisms would be able to regenerate lost parts after exposure to the rays. In multicellar organisms, the power of regeneration may be inhibited by exposure to the rays. Thus Bardeen and Baetjer ('04) have shown that exposure to the X rays inhibits the power of regeneration in fresh-water planarians; and Schaper ('04), that exposure to radium produces similar effects. Since regeneration in planar


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 423

ians is due not onlj' to multiplication and differentiation of cells but also, apparently, to tissue redifferentiation (morpholaxisMorgan) it would appear that the power of differential morphogenesis, even when not accompained by cell division, may be disturbed by irradiation. In the experiments mentioned, however, the chief effects of the rays seemed to be the inhibition of the formation and differentiation of new tissue. By exposing triton larvae to radium Schaper inhibited the regeneration of the tail and limbs. The wounds healed and a mass of cells accumulated in the region of the lost parts but no specific regeneration took place. Here the power of simple cell reproduction was not completeh' inhibited, although the power of differentiating new tissues was destroyed. In the human skin exposed to the X rays one sometimes finds the Malpighian layer greatly thickened. Here again there appears to be an interference with differential morphogenesis, although not at first of cell multiplication. In the skin of mice exposed to radium G.Guyot,('09)describes at first an increased activity in the multiplication of the cells of the Malpighian layer. Here, apparently, all the daughter cells undergo involution instead of only part, as in the normal skin. The reserve supply of genetic cells is thus used up and with the completion of involution in these cells the skin becomes denuded of epithelium.

The effects of irradiation on simple reproductive morphogenesis may be most conveniently studied in unicellular organisms. Experiments with these show that as a rule they are relatively very resistant to the X rays or to radium. Thus neither Paramoecia aurelia nor Paramoecia caudatum showed any difference in form or in rate of division when exposed twelve hours to Roentgen rays, although 15 minutes exposure of toad eggs undergoing cleavage to these rays inhibited gastrulation. Schaudinn '99, has shown, however, that some species are decidedly susceptible. Zuelzer ('05) has likewise described injurious effects which long exposure to radium causes in some protozoa. Bacteria are relatively resistant to theXrays although Rieder, and others, have described some inhibition of growth when the exposure was very intense. Similar effects have been described when bacteria are exposed to radium (Koerniche). The relatively great resis


424 CHARLES RUSSELL BARDEEN

tance to the X rays exhibited by unicellular organisms seems to indicate that simple reproductive morphogenesis is not readily disturbed by irradiation. Even conjugating paramoecia seem to be in no way disturbed by such exposure and the subsequent offspring appear to be perfectly normal. If morphogenic determinants in these forms are injured by the rays they can apparently be, in most instances, repaired. The relatively stable condition of the cytoplasm, which in large part is carried over nearly unchanged from the parent to daughter cells doubtless plays a part in maintaining the general morphogenic stability.

Differential morphogenesis, cell multiplication accompanied by specific change of organization, is readily influenced by irradiation both in plants and animals. Thus Koernicke ('04-05'), and others have found that sufficient exposures to X rays or radium may check the growth of germinating seeds. Tuilleminot ('09), has shown that while there may be a slight germination of seeds exposed to X rays or radium the latter cannot take the place of sun-rays. Immediately after exposure development may be quickened but this is followed later by retardation and abnormality or even inhibition of development. Fertilized eggs in the early stages of development are, in all animal species studied, very susceptible to the X rays, although the eggs of some species are apparently somewhat more susceptible than those .of other species. Thus Perthes on exposing the eggs of Ascaris megalocephala to the X rays found they give rise to irregular masses of cells or to abnormal embryos. Oilman and Baetjer showed that exposure to the X rays causes abnormal development in amphibian and avian eggs and Schafer and G. Bohn obtained similar results by exposing amphibian and reptilian eggs to radium. Tur has found that exposure of the eggs of the snail Philine aperta to radium gives rise to very abnormal larvae. The experiments of Bergonie and Tribondeau indicate that mammalian eggs are susceptible to the X rays.

In the adult mammal those tissues in which differential morphogenesis is constant are those which are most susceptible to the rays. Thus the epidermis, the bone marrow and the generative epithelium of the testicle are all highly susceptible. The


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 425

tissues in which differential morphogenesis is not constant appear in large part, at least, to be merely secondarily affected through alterations produced in the general metabolism or blood supply by the injuries produced in the actively reproductive tissues. Toxic effects thus produced have been described in man and the lower vertebrates by a large number of investigators. Several of the smaller mammals have been killed by exposure to the rays.^ Changes in the interior of the smaller blood vessels are frequently described after exposure to the X rays. Whether these are primarily due to the X raj^^s or are the secondary effect of injuries produced in rapidly reproductive tissues is uncertain, Scholz and many others believe that the X rays and radium may produce primary changes in the intima. CI. Regaud, on the other hand has been able to produce extensive alterations in the seminal epithelium without visible alterations in the blood vessels of the testicles.

The effects of the X rays on the testicles have been most carefully studied by CI. Regaud and his pupils to whom we likewise owe an excellent review of the literature of the effects of the X rays and of radium on the sex glands. (C. Regaud, '08). Most of the work has been done on the testicles of the rat, guinea pig, and rabbit, but enough has been done on other forms to indicate that throughout the animal kingdom the seminal epithelium is exceedingly sensitive to irradiation. In small mammals, such as the rat, the effects are more rapid and complete than in larger animals owing, apparently, to the smaller amount of filtration of the rays by overlying tissues. The seminal epithelium is far more sensitive than the epidermis so that in the smaller mammals sterility may be produced without marked injury to the skin (Bergonie et Tribondeau, '04).

In the rat lesions in the seminal epithelium begin to become manifest two or three days after irradiation and from this time on they become more and more marked until at the end of the 3rd or 4th week the generative elements may completely disappear. Regaud distinguishes two kinds of effects of irradiation,

^For th(> literature on this subject see A. S. Warthin, 1906.


426 CHARLES RUSSELL BARDEEN »

direct and cy to-hereditary. The former represent direct cell destruction, the latter alterations invisible in the exposed cells but which make themselves manifest in abnormalities and degeneration in the daughter cells or cells of more remote descent. Since the lesions do not appear for several days after irradiation it would seem difficult to distinguish these two kinds of effects from one another. In experiments on fertilized amphibian ova in which the action of the rays can be more directly followed the effects seem to be always of the cyto-hereditary type and hence we should be inclined to believe it probable that such is also the case in the testicles, although here doubtless the effects may sometimes become manifest at once in the daughter cells, at other times not until several generations if cells have been produced bj'^ division of the irradiated cells.

Of the elements of the seminal epithelium the basal spermatogonia appear to be the most sensitive. The mitotic figures in these cells and in those of the spermatocytes of the first order arising from them become in large part abnormal and the cells degenerate. In the experiments of Guyot we have seen that irradiation with radium apparently stimulates the cells of the Malpighian layer of the skin to hyperactive reproductive power accompanied by involution of all the daughter cells instead of only a portion of the daughter cells. It is not improbable that a similar condition is produced in the generative epithelium of the testicles. All of the daughter cells of the spermatogonia may undergo involution which, however, is abnormal and abortive in many of the daughter cells of the first generation. If the irradiation has not been too severe a few spermatogonia may remain for a time in a state of suspended activity and subsequently may begin to divide again and give rise to new generations of generative cells. At first, at least in the rabbit, man}" of the cells of the new generations are abnormal in form. (Regaut.) If the irradiation has been very severe the spermatogonia may all disappear so that permanent sterility is produced.

While Regaud attributes to the spermatocytes of the first order a sensitiveness almost equal to that of the spermatogonia it seems not improbable that many of the abnormal cells belonging


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 427

to this generation have arisen from exposed spermatogonia. It is also possible that young spermatocytes of the first order are so affected by irradiation that they degenerate instead of expanding in the normal manner. Complex morphogenic processes are doubtless active during the growth of the spermatocytes and the substances governing these processes may be very sensitive. Ovarian ova are similarly sensitive during the period of growth.

The spermatocytes of the second order show few abnormalities of form. Those exposed to the rays give rise to abnormal spermatids and spermatozoa but do not themselves manifest, as a rule, marked alterations. One would expect that the exposure of the spermatocytes of the first order would cause these to give rise to abnormal spermatocytes of the second order but apparently, as a rule, the spermatocytes of the first order are either destroyed by irradiation or affected in such a way that the effects first become manifest in the granddaughter cells (spermatids and spermatozoa.)

Irradiation of the spermatids and spermatozoa usually does not affect the external form and development of these cells; it however has been shown that irradiated amphibian (Bardeen '07) and mammalian spermatozoa (Regaud and Dubreuil '08) may give rise to monstrocities in the ova which they fertilize.

The nutritive syncytium of the seminal glands may presist after the complete disappearance of the generative cells but according to Regaud it is itself somewhat sensitive to the rays. The effects noted in the nutritive syncytium may perhaps be considered rather as secondary to injuries to the generative cells than as primary effects of irradiation. Regaud, however, believes that irradiation may produce in the syncytium a necrosis which secondarily causes the death of generative cells situated within its sphere of influence.

After severe irradiation the interstitial cells of the testicle may be injured although they are far more resistant than the generative cells. While Regaud believes the effects of irradiation on the interstitial cells are probably direct it would seem possible that they might be due to toxines produced by the necrosis of the generative cells. The connective tissues, vessels and nerves of the


428 CHARLES RUSSELL BARDEEN

testicle are, however, less affected by irradiation than are the interstitial cells. The modifications produced in the epididymis appear to be secondary to the aspermic condition of the testicles.

In the ovaries the follicles are far more susceptible than the other tissues to the X rays or radium. The primordial follicles are more susceptible than the older follicles, (Specht, '06; Bergonie et Tribondeau, '07).

In the primordial follicles the first modifications are seen in the nucleus of the ovule the chromatin! of which becomes massed together. The protoplasm retracts and then apparently the epithelial cells act as phagocytes and absorb the ovule and then themselves disappear. In older follicles the effects are similar. The zona pellucida is more resistant than the other parts. (Bergonie et Tribondeau.) When female toads with uterine ova are sufficiently exposed to the X rays the ova do not complete the process of maturation and cannot be fertilized (Bardeen, '09).

The various experiments to test the action of the X rays and radium on the generative cells thus show that irradiation maj' not only produce marked disturbance in the normal process of multiple differential morphogenesis of the sex cells but also may cause retrograde metamorphosis in the sex cells during the period of expansive differential growth, (spermatogonia of the first order, ovarian ova) and may prevent normal maturation.

Clinically the X rays and radium are utilized for the following physiological purposes: (1) to cause atroph}^ in the apendages of the skin, (glands, hair;) (2) to destroy parasitic organisms in the tissues; (3) to stimulate tissue metabolism; (4) to destroy pathological tissues; and, (5) for their anodyne effect. (Pusey and Caldwell.)

The cells of the sebaceous glands and the cells of the hair follicles seem to be somewhat more susceptible to the rays than are the cells of the deep layers of the epidermis so that atrophy of the sebaceous glands and hair frequently may be produced without serious injury to the epidermis. It is not certain whether or not the increased susceptibility of the cells of the sebaceous glands and hair follicles is due to the greater specific differential morphogenesis which characterizes these cells, but this seems not improbable.


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 429

Sweat glands are much more insusceptible to the rays than the sebaceous glands, but improvement in some cases of hyperidrosis has been reported after exposure to the X rays. (Pusey ) The action of the rays in these cases is not clear since it seems improbable that atrophy of the sweat glands could be caused without resorting to exposures severe enough to injure seriously the epidermis.

Beneficial treatment of malignant growths seems to be due chiefly to interference with cell multiplication and a consequent retrograde metamorphosis. In some carcinomata the cancer cells are evidently more susceptible to the X rays than the normal epidermis is. This susceptibility is apparently to be ascribed to interference with the anabolism of genetic, probably nuclear material. The question as to whether cancers exhibiting a considerable degree of differential morphogenesis are more susceptible than those characterized by a comparatively simple multiplication of like cells has not, so far as I am aware, been carefully studied. There is usually, however, some degree of differential morphogenesis in cancers. In sarcomata this is usually less marked and sarcomata seen to be frequently less susceptible to the rays than carcinomata. It is now well established that the differential morphogenesis of the normal epithelium may be occasionally so distributed by the repeated exposure to the X rays as to cause the production of carcinomata.

The beneficial effect of X ray treatment in other than malignant growths may be due in part to a tonic stimulative irritation which moderate exposure to the rays may give rise and in part to inhibition of abnormal reproductive activity in the inflamed tissues, or to the destruction of tissues of low resistance. It does not seem to be due to any considerable extent to direct action of the rays on the organisms which give rise to the inflammation. The nature of the anodyne effects of the rays is not understood. The clinical use of the X rays in the treatment of diseases of the blood seems to depend upon the action of these rays on the production of blood corpuscles.

It is possible that both X rays and radium might have a favorable effect on development if carefully regulated in intensity.


430 CHARLES RUSSELL BARDEEN

Wiiitrebert, ('06) states that radium emanation within certain Umits, the maximum of which is higher than that in any natural radio-active waters, stimulates the development of amphibian eggs and larvae. Eggs, however, require weaker solutions than larvae and will die in solutions favorable to the latter.

Having thus briefly' reviewed the physiological action of the X rays on various tissues we may proceed to examine a little more closely into the nature of the cytological disturbances produced by the rays.

Various experiments on ennucleated unicellar organisms have proved that non-nucleated protoplasm retains for some time the power of sensory-motor response and of simple metabolic activity. The power to digest substances, is however, impaired; the power to secrete substances and the power to repair wounds are reduced ; the power to form new chlorophyl granules is lost and the power of cellular reproduction is destroyed.^

The effects of exposure to the X rays resemble so closely the effects produced by removing nuclei from cells that one is led at once to infer that the action of these rays is primarily exerted on the cell nuclei. The alterations exhibited in the cytoplasm may as a rule be referred to disturbances, visible or invisible, produced in the nuclei. In the main, the nuclear and cytoplasmic disturbances produced by exposure to the rays are exhibited not in the cells exposed but in daughter or granddaughter cells or in cells of more remote generations. This leads us to believe that, as a rule, it is not so much the dynamically active substances in the nucleus as it is the reserve determinants which are injured by the exposure. On the other hand, the sensitiveness of ovarian ova and of the spermatocytes of the first order during the period of rapid growth seems to show that the powers of constructive metabolism exerted by the nuclei are quite susceptible to disturbance by exposure to the X rays. It is highly probable that the most complex of organic substances are the substances which go\^rn differential morphogenesis and that of

'For ;i review of the literature on this subject see Wilson — The Cell; Verworn AUgenieine Physiologie.


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 431

the activities which these substances are called upon to excite or direct the most complex are those which have to do with their own anabolism. On this anabojism depends the reproduction of like individuals through sex cells.

The belief that substances which govern differential morphogenesis reside chiefly in the nucleus is based, in the main, on the fact that the spermatozoon plays a part equal to that played by the egg in transmitting inheritable characteristics. Since the chief part of the spermatozoon is the chromatic material in the head and since this chromatic material gives rise directly to half of the nuclear material of the fertilized egg the belief seems well grounded. Those who, like Meves, contend that there are specific substances in the cytoplasm which play an equally or nearly equally important part in the transmission of heritable characters, have as yet offered no convincing proofs. That certain cytoplasmic substances or structures, centrosomes, chondriosomes, and the like, may be passed on from parent to daughter cells through spermatozoa as well as through ova does not in any way prove that these structures are important determinants of the subsequent morphogenic differentiation. Until further proof is brought forward in support of the importance of the influence of the cytoplasm in determining differential morphogenic activity it seems safe to follow Hertwig in assuming that the chief determinants lie in the nuclei. The nuclei reside in and work on the cytoplasm, so that the power of the nucleus to determine differential morphogenesis is determined in turn to some extent by the cytoplasm, much as the activity of any organism is determined by its environment.

It is well known that the most characteristic chemical substances in the nucleus are the nucleins, complex organic compounds containing phosphorus. It seems quite probable that the more complex of these compounds are the cell elements most easily disturbed by the X rays. The X rays are known to have the power of ionizing certain chemical substances. It seems quite probable that they may break down the more complex nuclear substances in such a way that the latter cannot be perfectly restored. The loss of these substances may be noted


432 CHAKLES RUSSELL BARDEEN

either within a few cell generations or not until a cytologically remote period. The destruction of ova, spermatocytes of the first order and spermatogonia is an illustration of the former, the failure of a limb to develop after exposure of a spermatozoon is an example of the latter.

Apparently the most sensitive substances in the ovum are those destined to determine the final stages in somatic morphogenesis. The more severe the exposure the earlier the cell generations in which the effects of exposure appear.

The sensitiveness of tissues to the rays depends upon the amount of differential morphogenesis which they are undergoing, or are destined to undergo, upon the rapidity of the production of nuclear material and upon peculiarities either individual or specific. The most susceptible tissues are those undergoing differential morphogenesis accompanied by a rapid production of determinative nuclear material (germinating seeds, ova during the earl} cleavage stages, germinative epithelium of the testicles.) On the other hand, the least susceptible of organisms seem to be some of the unicellular plants and animals (bacteria, paramoecia.) In these, in spite of rapid production of nuclear material, resistance to the X rays is marked. This resistance may be attributed in part to specific characteristics (unicellular organisms are known to vary greatly in their sensitiveness to light) and in part it may, as pointed out above, be due to the relative lack of disturbance of organic stability in reproduction through simple cell division.

In the higher organisms much idiosyncracy is shown in the sensitiveness of tissues to the X rays. The human skin, for instance, of some individuals is relatively resistant, in other individuals it is "burned" by relatively slight exposures, while in rare cases the epithelium is altered so that it gives rise to epitheliomata. If frog eggs be fertilized by exposed sperm most of the ova develop into abnormal larvae. A few may pass through apparently normal tadpole stages and then show abnormalities during metamorphosis (failure of one or more legs to develop) and some may become apparently normal frogs. These differences must depend either upon individual differences in susceptibility of the exposed spermatozoa or upon differences in susceptibility of


SUSCEPTIBILIIY OF AMPHIBIAN OVA lO X-RAYS 433

the ova to alterations in the fertihzing spermatozoa or to both factors. Since not all of the molecules of a gas exposed to the rays are ionized, we find in inorganic as well as in organic compounds variations in susceptibility.

To some extent, at least, idiosyncracy depends upon the general health of the exposed organism. When fertilized by exposed sperm ova which are overripe are much more prone to show marked deformities early in development that normal ova do. The susceptibility of cancer cells to the rays may be due in part to a weakening of the cancer cells produced by the reaction of the healthy tissues against the cancer tissue. In most instances, however, it is at present impossible to determine just what internal conditions make one organism at a given stage more susceptible to the X rays than a sister organism of the same stage ip.

In the following study of variations in susceptibility to irradiation of amphibian sex cells and larvae at different stages of development, individual as contrasted with specific sensitiveness has to be taken into account. When a given lot of organisms has been exposed some individuals will show the effects far more than others. Those most affected will show the effects first, those' least affected will show the effects late in developments or not at all. In a given group the greater the percentage of organisms severely affected the greater we may assume the susceptibility of that group. Thus we may compare the susceptibility of organisms at different stages of development by comparing the percentage of severely affected organisms in the groups exposed. In the various experiments the developing eggs and larvae were kept in large shallow glass dishes. The water was either frequently changed or was kept pure by a constant small stream of aeriated water. A small amount of various kinds of lake vegetation was kept in the dishes and the older larvae and tadpoles were fed with various kinds of food. Control experiments were carried on in all cases and every effort was made to keep the control and experiment specimens under equivalent conditions. By the methods used it is easy to carry the developing organisms up to the stage of well developed tadpoles, when there are not too many organisms in the dish. Where not otherwise stated, 100


434 CHARLES RUSSELL BARDEEN

per cent of the control specimens were carried to this stage. Late in the season fertilized eggs obtained under natural conditions frequently show some abnormalities of development and at this period a certain percentage of the control specimens show abnormalities even during the earlier stages of development. Note is made wherever this occurred.

To carry tadpoles through the later stages of development and metamorphosis special precautions are necessary. A large amount of well aeriated water and an abundance of food supply is necessary and even then a considerable number of apparently normal tadpoles usually fail to complete a normal metamorphosis.

In the tables where the percentage of 'normal specimens' is given it is to be understood that by this term is meant the percentage of tadpoles developing into large well developed tadpoles a few of which were isolated and followed through metamorphosis. When the percentage of isolated specimens which metamorphosed was approximately equivalent to the percentage of a similar set of control specimens undergoing metamorphosis the whole group from which the experiment specimens were isolated was considered 'normal.' With greater facilities at hand it might have been possible to determine the percentage of the whole group capable of undergoing metamorphosis as compared with the control in each experiment. It was, however, possible to do this merely in the few experiments in which it has been noted.

The effects of exposure may be roughly subdivided into the following subdivisions, although it is difficult to draw sharp lines of division between them.

1. Development stopped during cleavage. Cleavage more or less abnormal figs. 1 and 6.

2. • Gastrulation abortive or abnormal. May stop early fig. 7, or lead to :

a. spina bifida specimens (figs. 2 and 3,) and specimens with large anus (figs. 8, 9 and 10).

b. hemi-embryos (figs. 4 and 5).

3. Gastrulation complete though- more or less abnormal No distinct larval differentiation (fig. 11).


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS


435


4. Larvas markedly abnormal very early In development (figs. 12—13).

5. Abnormalities in larvae become well marked as the period for hatching approaches (figs. 14, 15 and 16).

6. Failure of larvae after hatching to grow normally into active healthy tadpoles.

7. Failm-e of normal metamorphosis.

In the following subsections this classification will be used in studying results of exposure at different stages of development.

1. EFFECTS OF EXPOSURE OF SPERMATOZOA TO THE X-RAYS

The following table gives a summary of the more successful experiments made to test the effects of exposure of the sperm of the toad and frog to the X rays. The details of the effects of this exposure on fertilization, cleavage, gastrulation, embryonic differentiation and subsequent development are discussed in the paragraphs which follow the table.

TABLE 1 Experiments on X-ray exposure of sperm






a m


Results



s


GASTRtTLATION



GASTRULATION



p<




o


INCOMPLETE



COMPLETE



X



p


m









f



2

H

o


< > o


1


1

"3



>


1 1 1


■o a

— c o



o3



n

- a

7-, <u


Remarks


<

o



1 « 


m

B

s

p


U

o


O



%


J2 J OS

a ft


o ij


> o

J2


as 5^


> ■S.'o


ll



a


■«!


>J


z


'^


'A Per


Per


w

Per


Per


Per


<!


<


P








Per


Per


Per


Per


Per







cent


cent


cent


cent


cent


cent


cent


cent


cent


cent



1


Toad


15 m.


75


l49l







9.5


64.3


26.



Experiment discontinued


2 Frog


30 m.


156


[2]


5.4


9.3


0.7


12.4


11.9


21.2


35


4



two weeks after irradia

3


Toad


37 m.


150


150]







98.7



1.3



tion, at this time 11.9 per


4


Frog


30 m.


650


[.8]




0.5


0.1



3.7


77.4


IS.


3


cent appeared normal.


5 Frog


2hrs.


28


[501








96.5


3.5




6 1 Frog


20 111.


267


13.5]







0.7


82.8


11.6


4.9



7 Frog


12 m.


370


[25.31


0.8






7.8


76


15.


4



8 1 Frog


40 m.


60


[661



1.7






76


13.3


8.3



9


Toad


70 m.


250









97.1


2.9



All appeared abnormal five days after fertilization.


Figures in brackets indicate percentage of eggs discarded because of lack of fertilization.


THE AMERICAN JOURNAL OF ANATOMY, VOL. 11, NO. 4


436 CHARLES RUSSELL BARDEEN

Fertilization

There is no definite evidence that exposure to the X rays affects the power of spermatozoa to fertiUze ova. The percentage of ova fertiUzed by spermatozoa in different experiments varied greatly; in some experiments nearly all, in others few or no ova were fertilized. The percentage of the ova fertilized seems to depend chiefly on the maturity of the sperm and the ova at the time of making the experiment. If the ova are mature nearly all are fertilized by spermatozoa obtained from males caught not too long after emerging from hibernation. Males, however, before being caught and brought to the laboratory may have discharged so large a proportion of their mature spermatozoa that an emulsion made from the testicles can fertilize comparatively few eggs. If the ova are in the right condition those fertilized seem to develop in essentially the same manner whatever be the proportion of the fertilized to unfertilized eggs. Apparently only those spermatozoa can fertilize which can fertilize normally. Uterine ova early in the breeding season fertilized by ripe sperm practically all develop normally. Late in the season the uterine ova are frequently over-mature. The capacity of these overmature eggs for fertilization seems frequently to be reduced and some of them, even when fertilized by normal sperm develop abnormally. In studying the effects of the X rays or other agents on spermatozoa and ova it is therefore necessary to take the physiological conditions of the sex cells into consideration and to make careful control experiments. When careful control experiments are made it is usuall}^ found that the percentage of eggs fertilized by the control sperm corresponds closely with the percentage of those fertilized by the sperm exposed to the X rays. Thus for instance, in one experiment where frog sperm was exposed for twelve minutes to the X rays the percentage of eggs fertilized by the X ray sperm was 25.3 and that of those fertilized by the control sperm 26. In some of my first experiments a greater proportion of the eggs were fertilized by the control than by the exposed sperm, but all subsequent experiments have led me to


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 437

believe that this difference must have been due to other influences than those exerted by the X rays.

The period of exposure of the sperm to the X rays has varied in my experiments from twelve minutes to two hours. The relation between the percentage of ova fertilized and the length of time intervening between the removal of the spermatozoa from the animal and the fertilization of the ova is less obvious than one might expect. Thus in one experiment in which the sperm was exposed an hour and ten minutes apparently every egg out of two hundred and fifty was fertilized while in another experiment after fifteen minutes exposure but 66 per cent of the ova were fertilized and in one after forty minutes exposure but 44 per cent were fertilized. Within certain limits power of fertilization decreases with the length of time elapsing between the removal of the spermatozoa and the fertilization of the eggs; up to an hour this factor has a less marked influence than the physiological development of the sex cells at the period of fertilization.

Boveri and others have suggested that the spermatozoon centrosome (middle piece) contains those chemical substances which excite cleavage in the ovum. If this be true it is evident that centrosomes are relatively resistant to the X rays.

2. Period of cleavage

In all experiments, except in Experiment 2, cleavage in most of the eggs fertilized by exposed sperm seemed to be normal. In several of the experiments it appeared to be slightly more rapid than in the control eggs. In Experiment 2 a considerable number of eggs ceased development in the late cleavage stages, but this was found to occur in but one other experiment, (Expt. 7,) and here in but three out of three hundred and seventy eggs. In Experiment 2 the effects of the exposure of the spermatozoa on gastrulation and the early stages of differentiation of the larva were much greater than in any of the other experiments, even where the spermatozoa were exposed much longer to the X rays (table 1.) It, therefore, seems probable that the ova of this lot must have been slightly abnormal so that although most of them


438 CHARLES RUSSELL BARDEEN

were fertilized they were more deeply affected by the X ray sperm than are normal ova. Unfortunately my notes on the control of this experiment were misplaced before the marked difference between this lot of eggs and other lots fertilized by exposed sperm was noted. In cases in which development ceased in the late cleavage stages the cleavage cavity was, in the eggs examined, of at least normal size. Hertwig, 1910, has found that long exposure of the sperm of sea urchins and amphibians to radium gives rise to disturbances manifest in the early cleavage stages.

3. Gastrulation

In most of the experiments all of the eggs externally appeared normal during the period of gastrulation. In three experiments marked abnormalities appeared during this period in some of the eggs. The percentage of eggs thus affected was much the greatest in the atypical experiment mentioned above, (Expt. 2.) In this experiment in 9 . 1 per cent of the eggs gastrulation was abortive and development ceased after the production of a large blastopore through which a large yolk mass protruded, (figs. 1, 6, and 7, plate 1). In 0.7 per cent of the eggs a hemi-embryo was formed similar to those artificially produced by Roux and others by injuring one of the blastomeres in the two cell stage, (figs. 4 and 5). In 12 .2 per cent of the eggs more or less typical spina bifida specimens were produced similar to those produced by Hertwig and others in eggs placed in NaCl solutions, (figs. 2 and 3). Marked abnormalities of this kind were much rarer in the other two experiments mentioned. In one of these. Experiment 4,0.5 per cent of the eggs produced hemi-embryos and . 1 per cent of the specimens were of the spina bifida type. In the other experiment (Expt. 8,) 1.7 per cent of the specimens showed an irregular cap of cells surmounting a protruding yolk mass.

In a considerable number of instances in all of the experiments eggs which on external appearance seemed normal during the period of gastrulation, internally suffered more or less well marked abnormalities of structure. In the atypical experiment, (Expt. 2,) 11.7 per cent of the eggs ceased development soon


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 439

after the blastopore closed. In some of these specimens the archenteron was very abnormally distended, in others it was rudimentary in character. A neural plate was well marked in some of the specimens. In some rudimentary traces of a notochord were found. In all the other experiments the specimens in which the gastrulation was internally abnormal went on to somewhat further development so that specimens with a more or less distinctly larval form were produced.

4. Period of larval differentiation

(a) Early stages, (up to the appearance of the rudiments of the head and tail). In most of the experiments some of the specimens ceased to develop so early in the larval period or were so abnormal that the rudiments of the head and tail were not very distinct. In those experiments in which definite data were obtained concerning the proportionate number of forms of this character the percentage varied from 3 . 7 to 9 . 5 except in the atypical Experiment 2, in which the percentage reached 20 . 8. Examination of these rudimentary larvae revealed the fact that in most instances the abnormalities must have begun to appear within the body during the gastrulation period. The extent of structural differentiation varied greatly. As a rule the neural tube rudiment was a somewhat irregular mass of cells without a trace of lumen. The alimentary canal in some specimens was fairly well developed, in others it was either greatly dilated or very rudimentary. In some specimens in which other structures were but little differentiated some of the myotomes were fairly distinct. The notochord was at times somewhat differentiated, at other times not distinctly to be made out. Irregular outgrowths in some specimens indicated a head or tail (fig. 13). In one specimen a large ' cleavage ' cavity appeared at one side of the body.

(b) Later stages, (from the period when the rudiments of the head and tail are distinct up to the time when normally the definitive tadpole form is assumed.) Specimens which at the beginning of the larval differentiation appear nearly normal frequently begin to appear abnormal during the formation of


440 CHARLES RUSSELL BARDEEN

the neural groove and still more so during the formation of the neural canal. Thus, for instance, in Experiment 4, in the neural groove stage about 25 per cent appeared decidedly abnormal. As the canal closed and the anlages of the head and tail appeared about 80 per cent became abnormal in from and a quarter of these appeared extremely distorted. In all of the experiments the greater number of the larvae appeared abnormal during the latter part of larval differentiation (table 1). The variety of abnormal forms was very great. As might be expected, the earlier abnormalities of development make themselves manifest the more profound are the deviations from normal structural form and, as a rule, the earlier the larvae die. The alterations in structural development manifest themselves in quite varied ways in different larvse from a given lot of eggs fertilized by a given lot of sperm. In one larva it is chiefly the cranial end that is affected, in another chiefly the caudal end, in a third the trunk may be relatively more affected than the head and tail. Internally the effects may be seen chiefly in the central nervous system, in the organs of special sense, in the vascular system or in the alimentary canal. In the more extreme types the external form and all of the internal organs are profoundly affected.

In a previous paper, ('07), I have given a description of several of these abnormal larvae and have summarized the effects noted. I give here a brief review of this summary together with additional data derived from subsequent experiments.

GENERAL EFFECTS

In all specimens growth is decidedly retarded during larval differentiation. If the larvae are hatched or are shaken from the investing jelly they expand but shghtly while the normal larvae increase very rapidly in size. Since this rapid expansion of the normal larvae is known to be due largely to inhibition of water it is fair to assume that this inhibition of water and the subsequent secretion into the cavaties and into the connective tissues of the body is in large part inhibited or altered in the experiment specimens.


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 441

EXTERNAL FORM

Head. In most, but not in all specimens, the head is decidedlyabnormal in form. As a rule the distortion of the head, is due in the main to a failure of normal formation and expansion of the cerebral vesicles, but it may also be due in large part to the failure of normal differentiation of mouth, pharynx and gill slits. Frequently a sucker is formed when the head is but slightly developed but differentiation of a sucker may be relatively behind that of the head, and not infrequently the sucker is abnormal.

Tail. In most specimens the tail is but a short deformed budlike process, sharply deflected in a dorsal direction. It may be exceedingly rudimentary, but on the other hand in a few specimens it is relatively well developed.

Trunk. In a large proportion of the specimens the abdomen protrudes markedly from the body. This protrusion is due to an abnormal dilatation of the body cavity. In most specimens there is a well marked dorsal flexion in the dorsal region of the trunk. This seems to be associated chiefly with defects in the development of the central nervous system. Irregular outgrowths from the trunk, especially in the ventral portion, are not infrequent. These seem to be chiefly finger like projections from the ectoderm.

INTERNAL STRUCTURES

Central nervous system. In those instances in which the effects are most profound in the neurogenic tissue, the neural plate may thicken without folding to form a distinct neural groove or if the neural groove is formed it may persist as an open groove while the larvae develop far beyond the stage in which the groove normally has become converted into a tube. In many specimens the neurogenic tissue then, instead of forming a neural tube with a clean cut lumen, gives rise to a more or less clearly defined rodlike mass of cells in which only here and there are to be seen evidences of a lumen. Rudiments of a lumen are usually to be seen in the head but the lumen even here fails in the extreme forms to expand so as to give rise to well marked cerebral vesicles.


442 CHARLES RUSSELL BARDEEN

If a well defined neural tube is differentiated, as a rule the tissue in the walls of the tube either in local regions, as for instance generally in the brain, or throughout its entire length, undergoes retrograde metamorphosis as development proceeds, and masses of protoplasm with more or less definite cell boundaries and containing more or less clearly degenerate nuclei and pigment are cast into the central canal. In most of these specimens the neural tube in places, especially in the cerebral region, becomes abnormally dilated and, in places, thin walled. Not infrequently the abnormalities are unilateral rather than bilateral. There may be an absence of development of the hind brain or of a portion of the spinal cord on one side, while it is fairly well differentiated on the other side. Sometimes the central nervous system is abnormally dilated at both extremities at an early period (fig. 16). Peripheral nerves. Peripheral nerves are developed onl}^ in those specimens in which the central nervous system is relatively slightly affected and larval differentiation is relatively advanced. Development of the ganglia of the cerebral and spinal nerves seems to be more or less closely associated with the development of a neural tube and in those instances in which a definite neural tube is not found it is usually difficult or impossible to distinguish sensory ganglia.

Eye. In most specimens the eye is more or less profoundly affected. In the more extreme forms there is no development of an optic vesicle on either side and apparently no traces of an eye are to be found. In some specimens the optic vesicles may project toward but not reach the ectoderm and no lens formation is apparent. In the better developed larva? the optic vesicles reach the ectoderm and a rudimentary lens is formed but in most of these specimens the optic stalk and optic cup become abnormally dilated and structural differentiation becomes quite abnormal. A narrow optic stalk containing nerve fibres is found only infrequently, but occasionally the eyes are relatively well differentiated.

Nose. In the more extreme forms there is no distinct differentiation of nasal organs but in those specimens in which olfactory lobes are developed in the brain well marked olfactory pits, as a


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 443

rule, are differentiated. In some specimens nasal fossae may be traced from the exterior to the pharynx by a column of cells containing an imperfect lumen or the column of cells may not reach the pharynx. Development on one side of the body may be much more advanced or less abnormal than that on the other side.

Ear. Larvae with fairly distinct heads are rarely found in which the auditory vesicles are not formed. In most cases the early stages of development in the vesicle seem fairly normal, but after the vesicle has been cut off from the ectoderm and the dorsal diverticulum has been given off differentiation, as a rule, ceases, unless the larva is relatively little affected. Extreme types of abnormality in the vesicles have not been found in the specimens studied.

Alimentary canal and its appendages. During the earlier stages of larval differentiation the archenteron may become abnormally dilated or it may be abnormally contracted. As the alimentar}^ canal differentiates the abnormalities may become especially marked at the anterior or posterior end or may be quite general in character. In most of those larvae which become fairly well differentiated there is formed a mouth with an opening into the pharynx. The lips and jaws are usually rudimentary but the abnormalities which affect them are quite varied.

In one instance the epithelium was entirely missing on one side of the pharynx. Patent gill slits may be formed but external gills are rarely well formed and internal gills are differentiated only in those forms which are comparatively slightly affected. A well marked operculum is seldom formed. The oesophagus becomes patent in some specimens but remains closed in others. The stomach in the most advanced specimens curves to the right but is, as a rule, rudimentary. The intestines generally show only a slight coiling and may be represented by a single straight tube.

Vascular system. In nearly all of the specimens showing marked abnormalities in body form during the latter part of larval differentiation the vascular system is profoundly affected. In the more extreme forms it is represented by abnormal rudiments of the heart and larger vessels. In some instances not even these may be distinguished. The heart, as a rule, is differ


444 CHARLES RUSSELL BARDEEN

entiated in the form of a rudimentary S shaped tube which may contain no continuous lumen. The pericardial cavit}' is not infrequently' abnormally dilated. Rudiments of some of the larger vessels in the head and trunk may usually be distinguished in the more advanced specimens but these rudiments frequently seem to be discontinuous. Since no artificially injected specimens have been studied it is impossible to decide definitely to what extent the vascular system is differentiated in discontinuous parts in these specimens. In the liver, capillaries may be distinguished in the better developed specimens, but elsewhere in the body they are usuallj^ difficult to trace, owing in part to the anaemic condition of most of the specimens. In some specimens no blood corpuscles can be found in the blood vessels. In most specimens they are far fewer in number than in normal tadpoles of a corresponding stage.

Lymphatic system. Abnormally dilated 'lymph' spaces are very frequent in the more advanced abnormal larvae but beyond this little concerning the development of the lymphatic system can be learned from a study of the specimens.

Gentio-urinary system. It is only in the more advanced larvsB that much can be made out concerning the genito-urinary organs. In these specimens the pronephric tubules are generally abnormal in form and considerably dilated. The Wolffian ducts may not extend to the cloaca. Frequently they are abnormally dilated. Mesonephric tubules are seldom differentiated in specimens which show marked abnormalities during the latter part of larval differentiation. The sex cells in some specimens appeared reduced in numbers but the data on this point are inconclusive.

Muscular system. The myotomes are sometimes well differentiated in places even in specimens in which there are profound abnormalities in other organ systems. On the other hand, they may sometimes be distinguished with difficulty in larvae in which these abnormalities are less profound. Sometimes the myotomes are represented by scattered muscle cells. In specimens in which there is a unilateral defect in the spinal cord there is usually an absence of myotomes in the region of the defect. In the head in the more advanced abnormal larvse the musculature is usually


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 445

relatively well developed except in those forms in which the whole head is very profoundly affected.

Skeletal tissues. Except in those specimens in which larval differentiation is interrupted at a very early period a distinet mesoblast with visceral and parietal layers is usually differentiated in the region of the trunk. In very many specimens the peritoneal, like the pericardial cavity, subsequently becomes abnormally dilated. In those specimens which reach a relatively advanced stage of development the loose connective tissue of the head and body wall frequently appears much distended with fluid and abnormally great in extent compared with the more definitive organs. The chorda dorsalis in those specimens which reach an advanced stage of larval development appears to be relatively normal in form, although frequently it may be in places assymetrical, dilated, or small and defective.

Cartilages and other definitive skeletal structures although generally somewhat abnormal in form, are differentiated in a relatively normal manner in those specimens which reach the stage of skeletal differentiation. Most specimens die before auditory capsules are produced.

5. Tadpole stage

In all of the successful experiments with irradiated sperm some of the fertilized ova developed into tadpoles fairly normal in external form, with operculum, rounded belly, and membranous tail. Most of these tadpoles, however, failed to grow and expand like normal tadpoles, and very few reached and successfully passed through the period of metamorphosis. The majority soon exhibited externally visible defects of one sort or another, and within two or three weeks after the fertilization of the ova, died off. In table 1 these tadpoles are classed as 'defective'. From this table it will be seen that they constitute from 1.3 per cent to 26 per cent of the total number of eggs fertilized. The number of specimens undergoing apparently normal development, although not in all cases followed to metamorphosis was 8.3 per cent in one experiment in which the sperm of the frog was exposed


446 CHARLES RUSSELL BARDEEN

forty minutes and 4.9 per cent in an experiment in which the exposure was twenty minutes. In four experiments in which the exposures were respectively thirty minutes, thirty-seven minutes, one hour and ten minutes, and two hours, no normal tadpoles were developed. In the other experiments the number of tadpoles finally undergoing normal development was not definitely determined. In those experiments in which some eggs developed into apparently normal tadpoles, a few of which metamorphosed, it is by no means certain that the resulting toads or frogs would have been perfectly normal individuals. In one instance a metamorphosed toad lacked a hind leg. It is possible, if facilities had existed for carrying a large number of tadpoles through metamorphosis and up to the adult stage, many more defects might have been discovered as attributable to the exposure of the sperm to. the X rays.

In the tadpoles which failed to develop normally, abnormalities were especially frequently found in the central nervous system. The tissue of the neural tube seemed to lack capacity for higher differentiation and degenerate cells and protoplasmic masses were discharged into the ventricles of the brain and into the central canal of the spinal cord. Abnormal dilatation of the body cavity and of the pronephric and Wolffian tubules are other phenomena frequent in these specimens. Occasionally the gut exhibits degeneration or marked abnormality of form. Sometimes the nasal fossae and the eyes are defective. The optic stalk may be dilated. Occasionally the defects are unilateral.

Conclusions

The results of fertilizing amphibian ova with sperm exposed to the X rays indicate that the alterations produced by the exposure in paternal determinants do not as a rule make themselves manifest until larval differentiation or later in development except in those instances in which there is reason to believe that the ova were somewhat defective (over-ripe, etc.,) at the time of fertilization. The paternal determinants become relatively more and more important as development proceeds. They may be


SUSCEPTIBILIIY OF AMPHIBIAN OVA TO X-RAYS


447


injured in such a way that localized lesions are produced in the larva. In rare instances the entire side of the body may fail to develop (hemi-embryo.) This suggests the possibility that at times the injured paternal chromosomes may be localized in one of the first two blastomeres, while the normal maternal chromosomes are localized in the other blastomere and are capable of inducing differentiation up to the larval stage.

In this connection the various experiments made on crossing different species and genera of amphibians are of interest. Ziegler ('02), has summarized the results of Pfliiger, Born, Gebhardt and others as follows:


Rana esculenta, 9 \

Rana fusca, cf / The eggs develop to the blastula stage.

Owing possibly to the shape of the head of the spermatozoon of R. esculenta crossing in the reverse direction does not lead to fertilization.


\ The eggs develop into larvse, some of which meta/ morphose into frogs.

No fertilization in the reverse direction. Development continues up to gastrulation; in favorable cases up to the origin of the medullary plate. The cleavage stages appear normal but during gastrulation development becomes very abnormal. \ The eggs segment and develop to the morula / stage.

As a rule no fertilizations take place in the reverse direction, but one of 100 eggs one segmented regularly and two irregularly. \ The eggs develop to larvse, and these metaJ morphose into toads. Bataillon [Comptes rendus de I'Acad. des Sciences, 1908,] has crossed several species of amphibia which have a varying number of chromosomes in reduction division. The species used together with the estimated number of chromosomes were as follows: Pelodytes punctatus, 6; Bufo calamita, 12; Bufo vulgaris, 8-9; Rana fusca, 12.

1. Some not fertilized.

2. Some underwent parthenogenetic segmentation.

3. Some ceased developing in the blastula stage. Blastula halves characterized by 2 kinds of cells; (1) Large with large nuclei (12 chromosomes), and (2) small with small nuclei (6 chromosomes)

J 4. About 10% developed into larvse.


Rana arvalis, 9 Rana fusca, cf

Rana arvalis, 9 Rana esculenta, cf

or Rana arvalis, cf Rana esculenta, 9 Bufo vulgaris, 9 Rana fusca, cf


Bufo cinerius, 9 , Bufo variabilis, cf


Pylodytes punctatus, <f Bufo vulgaris, 9


448 CHARLES RUSSELL BARDEEN

Bufo calamita, cf ]

Bufo vulgaris, 9 [ Developed into larvie.

Bufo vulgaris, cf

Bufo calamita, 9

Rana fusca, d' \ Developed to the "Stereo-blastula" stage but

Bufo calamita, 9 / did not undergo gastrulation.

Parthenogenetic development stimulated by spermatozoon ceases in early cleavage stages. In the P. puntatus cross the female pronucleus acted as the segmentation nucleus. In the Bufo calamita cross the second maturation division is not completed but instead a cleavage nucleus is formed.


Triton alpestris, cf Pelodytes puntatus, 9 Triton alpestris, cf Bufo calamita, 9


From these experiments it will be seen that when the spermatozoon of one species enters the egg of a distant species the egg may undergo cleavage, but as a rule development becomes abnormal, before or during gastrulation and larval formation, and only exceptionally proceeds to metamorphosis. The foreign spermatozoon has the power of exciting cleavage, but as development proceeds it either fails to furnish determinents requisite for development or inhibits the action of maternal determinents which might otherwise prove effective. By action of alterations in temperature and of sugar solutions parthenogenic development has been excited in the frog's egg. (Bataillon, '04). In these experiments development extended to the blastula stage. Cleavage was incomplete so that the roof of the blastula alone showed cytological segmentation. Bataillon follows Boveri in recognizing two stages in early embryonic development, a 'promorphological' in which cell division is not followed by complex morphological differentiation, and a ' metamorphological ' in which complex differentiation takes place. The nuclei appear to play a much more specific part in the latter than in the former stage. It is the former stage that is initiated in the experiments on parthenogenesis in amphibia.

While the maternal nuclei may govern the cleavage stages, the experiments suggest that beyond these stages the maternal determinants are of themselves incapable of stimulating development. If the paternal determinents are of foreign origin or have been injured, as by exposure to the X rays, development may proceed


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 449

beyond the cleavage stages but sooner or later, as a rule, becomes abnormal. Occasionally adjustments take place of such a nature that development may proceed at least to metamorphosis. Such adjustments are most frequent when the paternal determinants come from species closely related to the maternal, or when the injurj^ to the paternal determinants has been slight.

In crossing Bufo calamita and Pelodytes puntatus, Bataillon found that apparently in the first division one cell received chromosomes from the female pronucleus and the male pronucleus, the other merely from the male pronucleus. In considering the fertilization of amphibian eggs with exposed spermatozoa we have seen reason to believe that at times the injured paternal determinants become localized in certain regions so as to lead to regional defects. Of these the most extreme and rarest examples are hemi-embryos. Unilateral defects in parts of the central nervous system are far more common.

In fertilized eggs of Nereis liillie found that by centrifugalization he could destroy the activity of the male pronucleus, although as a merely chemical complex it remained within the egg. In those eggs which had formed the fertilization membrane and had started development, differentiation proceeded to the formation of the second polar body but no complete cleavage spindle was formed and the eggs remained unsegmented. Here we must assume the female pronucleus incapable of alone initiating cleavage unless we assume that it, too, was injured although less obviously than the male pronucleus.

In an interesting study of heredity in fundulus hybrids H. H. Newman, ('10) finds that the developmental rhythum of the young embryo is distinctly influenced by the foreign spermatozoon as early as the second cleavage stage and comes to the conclusion that the nuclear material is the chief factor in determining the character of early development. Fischel has shown that in a number of Echinoderm hybrids the male influence is expressed structurally in the early blastula stages, not only in the general size of the embryos, but also in the size and shape of the cells and has given evidence that 'Hhe role of the spermatozoon is from the beginning formative in character in that it is able to place the


450 CHARLES RUSSELL BARDEEN

stamp of its own specific characters upon the early developmental stages of the organism, while the egg cytoplasm furnishes only the material for the formative operation of the combined nuclear material of the two parents" (Newman). This conclusion is strongly supported by the action of irradiated spermatozoa on the eggs they fertilize.

2. EFFECTS OF THE EXPOSURE OF MATURE OVA TO THE X-RAYS

Owing to a rapid loss of capacity for fertilization by amphibian eggs after being laid in w^ater it is necessary to expose the eggs within the body of the female. Here the eggs are to some extent protected from the rays by the overlying tissues. This perhaps accounts for their, relative insusceptibility to short exposures or weak rays. On the other hand, thus exposed to powerful rays for an hour or more the eggs are very susceptible. In the following table the results of seven experiments are recorded. Several other similar experiments were tried but proved unsuccessful owing, usually, to lack of maturity of the eggs or to the fact that all the ripe eggs had previously been discharged. Uterine ova are prevented from maturing by irradiation.

In each of the experiments recorded the body was opened after the irradiation and a considerable number of eggs were fertilized with an emulsion prepared from the testicles of one or more normal males. In Experiment 1, table 2, twelve minutes exposure had but slight effect. When so large a number of eggs are artificially fertlized and kept in glass dishes a similar percentage frequently shows some defects in late larval and early tadpole stages, owing to the overcrowding. In Experiment 2, thirty minutes exposure gave decided results. The effects of exposure began to be noticed in many specimens early in larval differentiation and over 50 per cent of the eggs showed marked abnormalities before the larvae reached the period of hatching. After forty-five minutes exposure, Experiment 3, the results are still more marked while after an exposure of an hour or more. Experiments 4-7, many specimens showed abnormalities during gastrulation and few developed into normal tadpoles.


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS


451


TABLE 2 Exposure of ovum before fertilization


n


5


o

g

o


m

o o

pq


is


Rbsttlts



S

3 PI


GASTRTJLATION INCOMPLETE


GASTRULATION COMPLETE



o

%


i


■d

1 1


d o

is

o


Pi


8i >

a


0]

■a

1


'1

11

og

Per

cen<


1

Per

cent


"3

as <


1


1

s

"3 g

1^


Remarks


1


Toad


12 m.


503


Per

cent [.061



Per

cent


Per

cen<


Per

cent


Per

cent 10.7


Per

ceret

5


Per

ceni 84.3


Control gave about the


2


Toad


30 m.


703


]0.15]







50.


9


14


35


same percentages. Five specimens followed


3


Toad


45 m.


679


[0.44]



1


.8



7.1


24.7


36.4


10.9


19


to metamorphosis. But one specimen could be


4


Toad


Uhr.


216


[0.44]



3


1



5.59


26.8


31


1,4


4.2


preserved through metamorphosis. Eggs slightly injured, X


5


Toad


Ifhr.


428


[0.56]



3




3.7


38.3


40.2


12.4


5.1


ray current weak. Severel specimens followed


6


Toad


1 hr.


846


[22.0]



5.7


1.3


17


2.7


35.6


34.9


2.3


0.4


to metamorphosis. X-ray currentstronger than


7


Frog


Uhr.


177


ll]



9


2.3


6.2


11.3


21.5


48


1.7



in the above experiment.


Figures in brackets indicate percentage of eggs discarded because of lack of fertilization.


The internal effects have been studied chiefly in the toad eggs of Experiment 6 and the following account is based mainly upon that study, but the results reported have been confirmed by a study of the frog eggs of Experiment 7.

1 . Fertilization

In all of the experiments there were a few eggs not fertilized but the percentage of unfertilized eggs was large merely in Experiment 6 where it reached about twenty per cent. In this experiment the large number of eggs used and the difficulty of mixing quickly so large a number of eggs with the sperm emulsion probably account for the unfertilized eggs. Experiment 7, in which but one out of one hundred and seventy-eight eggs was not fer THE AMERICAN JOURNAL OF ANATOMY, VOL. 11, NO. 4.


452 CHARLES RUSSELL BARDEEN

tilized although the subsequent effects of irradiation were marked, indicates quite clearly that irradiation of mature eggs has little or no effect on their capacity for fertilization.

2. Period of cleavage

The early cleavage stages appeared to be normal. No specimens were seen to stop segmenting early, as in the case of the two atypical experiments with exposed sperm. (See Experiments 2 and 7, table 1.) We have, however, seen reason to believe that in the latter the effects were due to other causes than irradiation.

3. Period of gastrulation

In a considerable number of instances, 5 . 7 per cent in Experiment 6, and 9 per cent in Experiment 7, gastrulation did not advance far and left a mass of yolk protruding through a large blastopore. Some of these specimens were very abnormal. Similar forms were found in two of the experiments on the exposure of sperm, but in considerable numbers merely in the atypical Experiment 2, table 1, in which they formed 9.1 per cent of the fertilized eggs.

In Experiment 6, table 2,1.3 per cent of the ova gave rise to hemi-larvse, and in Experiment 7, 2 . 3 per cent. This percentage is larger than that found in either of the two experiments with the exposed sperm which resulted in the production of forms of this character.

In Experiment 6, table 2, 17 per cent of the ova developed into more or less highly differentiated types of spina bifida, and in Experiment 7, 6.2 per cent. Spina bifida specimens were found in two of the experiments on exposure of the sperm, but in considerable numbers only in the more atypical one. Experiment 2, table 2.

In Experiments 1 and 2, table 2, no eggs, and in Experiments 3 and 5, but few eggs ceased development before the completion of gastrulation. In Experiment 4 a very large percentage of eggs ceased development during gastrulation or showed marked


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 453

abnormalities in the process. In this lot the female was pithed, the abdominal cavity opened and some eggs were removed several hours before exposure so that the eggs remaining in the abdominal cavity at the time of exposure were probably somewhat injured. The current used was weaker than in Experiments 6 and 7.

Many .specimens in which externally gastrulation appeared normal, and in which the yolk was completely covered by pigmented cells, undoubtedly internally deviated markedly from the normal course of development. Of the specimens of this character a considerable number (see Experiments 3-7, table 2), ceased development as soon as the blastopore was closed, while others made abortive attempts at larval differentiation before dying. In the experiments with exposed sperm specimens ceasing to live as soon as the blastopore was closed were found only in the atypical Experiment 2, table 1, where they formed 11.7 per cent of the total number of eggs fertilized. For a description of the internal changes in eggs of this tj^pe see the description of this latter experiment.

4. Period of larval differentiation

(a) Early stages. In Experiment 6, 35.6 per cent of the fertilized eggs ceased developing after giving rise to abnormal forms without distinct heads or tails; in Experiment 7, 21.5 per cent; in Experiments 3, 4 and 5, 24 . 7 per cent, 26 . 8 per cent and 38 . 3 per cent respectively. In the experiments with the exposed sperm in which the proportion of abnormal forms of this type was determined the number was found to be less than 10 per cent of the fertilized ova, except in the atypical Experiment 2, table 1, in which the number reached 20 per cent. The types of abnormalities corresponded with those described in connection with the sperm experiments.

(6) Later Stages. In Experiment 6, 34.9 per cent, and in Experiment 9, 48 per cent of the fertilized eggs developed into monstrous forms in which rudiments of the heads and tails were fairly distinct. In Experiments 3, 4 and 5, 36.4 per cent, 31 per cent and 40 . 2 per cent respectively. In Experiment 2, 50.9


454 CHARLES RUSSELL BARDEEN

per cent became abnormal during larval differentiation. Comparatively few of the abnormal larvse in these experiments, however, exhibited the relatively advanced degree of differentiation attained by the majority of the abnormal forms in the sperm experiments. The types of abnormalities appear to be not essentially different in the later larval stages in the two sets of experiments.

5. Tadpole period

In Experiment 6, 2.7 per cent reached the definite tadpole stage and of these only a ninth (0 . 3 per cent of the total number of eggs fertilized) developed as normal tadpoles. In Experiment 7, 1 . 7 per cent of the ova fertilized became definite tadpoles but none of these lived long. In Experiments 4 and 5, 4.2 per cent and 5 . 1 per cent respectively became apparently normal tadpoles. None of those in Experiment 4, however, lived more than a few weeks while some of those in Experiment 5 were followed to metamorphosis. In Experiments 1-3 a considerable percentage of organisms reached the tadpole stage and developed normally.

COMPARISON OF THE RESULTS OF EXPOSURE OF SPERMATOZOA

AND OF MATURE OVA

A comparison of Experiments 1, 2 and 3, table 2, with Experiments 1, 2, 3, 4, 7, and 8, table 1, shows quite clearly that the effects of exposing spermatozoa in water for from twelve to fortyfive minutes are more marked than when ova within the toad or frog are exposed for similar periods of time to rays of similar intensity. On the other hstnd, exposure of male frogs to weak rays seems to have less effect on the spermatozoa than exposure of female frogs has on the ova. (Experiments described on page 456.)

The number of fertilized ova which become tadpoles is approximately equal when sperm or ova are exposed for an hour or more to intense rays, but the effects are noticeable earlier in development and are more profound in the experiments with ova han in those with the sperm. The injurious effects of exposed sperm


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 455

do not manifest themselves except oecasionalh' before the later stages of larval differentiation.^ The injurious effects of exposing the ova before fertilization make themselves manifest early in development in a large number of the specimens. In both sets of experiments hemi-larvse are occasional!}^ produced (figs. 4 and 5), while in specimens exposed to the X ravs after fertilization no abnormal forms of this type were found. In one instance in a lot of eggs exposed for three-quarters of an hour after the first cleavage stage had begun one specimen exhibited bilateral assymetry, but it did not on careful examination appear to be a true hemi-larva. In spina bifida specimens produced by salt solutions the abnormal forms are nearly always symmetrical or nearly symmetrical. In an examination of a large number of specimens of this kind I have found no hemi-larvse. It would, therefore, appear that occasionally one germ cell nucleus or the other may at least predominate in influence in one of the first two cleavage cells destined to form a lateral half of the body, and that if this sex cell nucleus is sufficiently injured no development of the blastomere takes place.

If it be true that a sex cell nucleus can be so injured by exposure to the rays that it becomes incapable of initiating or properly governing development even in the earliest stages, we have an explanation of the greater number of early abnormalities seen after exposing the unfertilized ova. In such eggs we may assume that if the nucleus has been sufficiently injured development must be guided largely by the male sex elements. These probably are at first not intimately adjusted to the demands of the protoplasm of the egg and hence development appears abnormal at an early stage. In the later stages it appears that the male and female sex elements are by no means evenly distributed in the differentiating tissues. The tissues in which the injured germ plasm predominates are those which first show abnormalities and these differ in different specimens. In the few specimens which survive it is probable that every tissue gets some normal germinal elements derived from the uninjured germ cell.

^The experiments of Hertwig, 1910, however, show that severe prolonged radium irradiation can give rise to disturbances manifest earlier.


456 CHARLES RUSSELL BARDEEN

McGregor ('08) after exposing both parents to the X rays found that the results were not markedl} different from those obtained after exposing the male before fertilization. He found more striking results after exposing the male than after exposing the female. I endeavored in the spring of 1910 to repeat experiments along these lines on a more extensive scale, but, unfortunately, I used too weak a current to get very positive results.

"in the control experiment out of 110 eggs 11 per cent showed abnormalities during the period of larval and early tadpole differentiation. This unusual percentage of abnormalities makes the results of the experiments somewhat uncertain since it indicates that other factors than the X rays played a part in causing abnormalities in development in the exposed specimens.

After exposing a male to the X rays for an hour and ten minutes and then fertilizing eggs taken from the toad used for the control mentioned above, out of 410 eggs all but 8 per cent developed into normal tadpoles. Of the abnormal specimens one died before completing gastrulation.

After exposing a female to the X rays for an hour and ten minutes and then fertilizing the eggs with sperm derived from the control mentioned above, out of 190 fertilized eggs, 14 per cent showed abnormalities. In three eggs gastrulation was not completed, in one a condition of spina bifida appeared, in the others Ihe abnormalities appeared during larval differentiation.

Out of 376 eggs taken from the exposed female and fertilized by spermatozoa taken from the exposed male 21 per cent showed abnormalities of development. Of the abnormal specimens 27 showed marked abnormalities during gastrulation and 52 during larval differentiation.

So far, therefore, as the experiments go they indicate that exposure of both sex cells before fertilization gives rise to more severe effects than the exposure of either alone and that weak X rays filtered by the tissues affect the female sex cells more than the male sex cells. The experiment must, however, be repeated and extended before definite conclusions are warranted.


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 457

3. ACTION OF THE X-RAYS ON FERTILIZED OVA AND ON LARVAE

Introduction

My studies of the effects of exposure of developing organisms to the X rays have been most systematically carried out on the ova and larvae of the toad. The results of these experiments have been summarized in the accompanying tables. Numerous, although less systematic, experiments on frogs' eggs indicate a variation in susceptibility in these eggs which corresponds closely with that found in toads' eggs. No attempt has been made to summarize these results in the tables.

The more important experiments are designated in the accompanying tables as follows:

Experiment A. Spring of 1909. Successive groups from a batch of fertilized toad eggs were given forty-five minutes exposure at varied intervals from the time of fertilization up to the tadpole stage. The different groups are designated by the number of hours intervening between the beginning of fertilization and the period of exposure, i.e., Experiment A, 14^: batch of eggs exposed from 14^ to 15| hours after the beginning of fertilization.

Experiment B. Spring of 1909. Successive groups were given forty-five minutes exposure from a period begininng 15f hours after fertilization and extending into the tadpole stages. Group designation as in Experiment A.

Experiment C. Spring 1908. Successive groups from a batch of toad eggs were given thirty minutes exposures from the time of fertilization up to the larval stage, with an interruption during the later cleavage stages. Group designation as in Experiment A.

Experiment D. Spring of 1908. Successive groups from a batch of toad eggs were given thirty minutes exposures from the period of fertilization up to the early cleavage stages. Group designation as in Experiment A.

For comparison some other experiments have been included in the tables but since these were less systematic they need not be described here.


458


CHAELES RUSSELL BARDEEN


1 Period of fertilization

According to Helen Dean King ('01) the second polar body is given off about ten minutes after the entrance of the spermatozoon into the toad's egg and the two pronuclei fuse about thirty-five minutes later. In table 3 we have summarized the results of exposure of toad's eggs during the first hour after fertilization.

In Experiments AO a string of fifty-four eggs was exposed for forty-five minutes beginning immediately after fertilizing fresh ova with fresh sperm. Practically all of the batch of eggs from which this string was derived were fertilized. It is, therefore, possible although not certain, that the exposure to the rays prevented development in the two eggs in the string which failed to show any signs of cleavage. Forty-nine eggs failed to develop past the early gastrulation stages. Of these forty-one ceased


TABLE 3 Irradiation of eggs during fertilization


P.






Results



m


Si


H


IS





s

H


GASTRULATION


GASTRULATION COMPLETE



ft


«5




INCOMPLETE




Ui












^


^^


&<


E3



a




i.


8





Remarks


X


IS


H


ss


0}


5



d


&


s



i


%



o


^a

B K


O


(4

o



•3


^1


•3


— o



•— « 


o


0.



I? a



W


PQ

a


o


o




o <a


>

.2


a ^

J3 —


> 5 A


ag ga



Q


Eh

Hr.


»


iz;


iz;


Per


Per


'A


«0 Per


<J


« 


!4




Min.



-per


Per


Per


Per


Per







cent


cent


cent


ceni



cent


cent


cent


cent



A



45


54


3.7


75.9


14.8




1.9


3.7




No specimen lived beyond the second day after fertilization.


C



30


317


[1]


9



1



1


60.8


28


.1


20 per cent "normal" at end of two weeks when expt. was discontinued.


C


4hr.


30


148



3.4





2.4


74.6


19


.6


8 per cent of the specimens appeared normal after 2 weeks when expt. was discontinued.


D


Jhr.


30


160


[29.51


1.3



1.3



69.9


7


25.6


1.9



D


Jhr.


30


30


[2.5]



20.




76.7



3.3



All specimens abnormal five days after irradiation.


Figures in brackets indicate percentage of eggs discarded because of lack of fertilization.


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 459

developing immediately before or immediately after the first stages of gastrulation, while eight showed evidences of irregular gastrulation and gave rise to forms with a large blastopore and a protruding yolk mass. In three specimens the blastopore was closed. In one of these development stopped very early in larval development ; in the other two larvae with abnormal heads and tails were produced. The effects mentioned indicate that an exposure of three-fourths hour immediately after the beginning of fertilization has a more profound effect on development than exposure of either of the germ cells for an even longer period before fertilization. (Compare tables 1, 2, and 3.)

On the other hand, an exposure of half an hour immediately after the beginning of fertilization in Experiment CO produced effects apparently somewhat less severe than those in Experiment 1, table 1. In the latter experiment, in which the spermatozoa were irradiated for fifteen minutes 11.9 per cent of the tadpoles appeared fairly normal two weeks after fertilization. In Experiment CO, 20 per cent appeared fairly normal at this period. Since neither experiment was carried beyond two weeks after fertilization the percentage of specimens capable of undergoing ultimate normal development was undetermined in each case. In Experiment 2, table 1, in which frog sperm was exposed thirty minutes to the rays, and Experiment 3, table 1, in which toad sperm was exposed thirty-seven minutes to the rays, the effects were more severe than in Experiment CO, table 3. We have, however, seen reason to believe that Experiment 2, table 1, gave results quite atypical in nature and not to be attributed wholly to the action of the X rays; while the longer exposure (thirty-seven minutes) in Experiment 3 would perhaps suffice to account for the greater severity of the results. In Experiment 4, table 1, (exposure of frog sperm for twenty minutes), Experiment 7, table 1, (exposure of frog sperm for twelve minutes, and Experiment 8, table 1, (exposure of frog sperm for forty minutes) the results were similar to, although slightly more marked than those found in Experiment CO, table 3. We may, therefore, conclude that exposure of eggs for one-half hour after the beginning of fertilization gives results approximately equivalent


460 CHARLES RUSSELL BARDEEN

to, although perhaps less marked than the exposure of the sperm for from one-quarter to one-half hour before fertilization. Experiments 2,-3, table 2, likewise indicate that exposure of the ovum within the female for thirty to forty-five minutes gives rise to less severe effects than exposure of ova for half an hour during fertilization.

Experiment C^ indicates that during the second half-hour after the beginning of fertilization the fertilized ova are somewhat more susceptible than during the first half hour. Compare Experiments CO and C^, table 3; in the former experiment 20 per cent, in the latter but 8 per cent of the specimens appeared normal at the end of two weeks. Experiment Dj indicates a possible greater susceptibility of the eggs during the middle twofourths than during the first or last half of the first hour after the beginning of fertilization, but the differences may be due to individual peculiarities in the different batches of eggs used. In Experiment D| but 1 . 9 per cent of the eggs were normal two weeks after fertilization.

Summary. From the above experiments we conclude that the ova immediately after fertilization are somewhat less susceptible than the spermatozoa or unfertilized mature ova to short or weak exposures but are more susceptible to long exposures. During the conjugation of the pronuclei the susceptibility is greater than during the preceding period. There probably exist in the protoplasm of the egg substances capable of protecting the pronuclei against moderate injury or of restoring them after such injury, but these substances, if such exist, cannot overcome the effects of exposures to powerful X rays for three-quarters of an hour and they become less potent during the fusion of the pronuclei than in the period preceding. The fusion of the pronuclei seems to start a protoplasmic reorganization during w^hich the susceptibility of the cell is increased.

The effects of the X rays during fertilization become manifest after half-hour exposures, chieflj^ in the later larval periods, after the rudiments of the head and tail appear and while the larva is becoming transformed into a free swimming tadpole. During the tadpole stage many specimens which at first appear normal


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 461

prove incapable of growth and further differentiation and die off. Experiment CO and C-§ were not carried far enough to determine whether or not any of the eggs exposed for half an hour wdthin the first hour after fertilization were capable of metamorphosis. In Experiment D-j one specimen was followed through metamorphosis.

In both Experiments CO and C-^ a considerable number of eggs failed to undergo cleavage but since in the control for these experiments 6 . 9 per cent of the eggs were not fertilized it is probable that the ova in the experiments which did not undergo cleavage failed to develop not because of the exposure to the rays but because of lack of fertilization. In Experiments D-j and D-f a considerable number of the eggs became mouldy and decaA'ed before it was determined whether or not they had been fertilized. The percentage of unfertilized eggs given is, therefore, not exact, since under the eggs so classed there may be some which stopped development during cleavage.

When the exposure is more prolonged and severe the results, as we have seen in Experiment A-0, are manifest just before or early in gastrulation. Most eggs stopped developing at this period but a few went on to form abnormal larvae. In none of the eggs in which cleavage began did development stop until the approach of the gastrulation period. If the cleavage once commences at all it is evidently capable of continuing until the critical period of gastrulation is approached. It is, of course, by no means certain, how^ever, that cleavage is perfectly normal during this time. I have, however, been able to detect in sections no very obvious abnormalities, except a tendency for the vegetative pole to divide much more slowly than normal so that the yolk may be but slightly divided into cells while the cells of the animal pole are approximately of normal size. The abnormalities which appear during gastrulation and larval differentiation are essentially similar to those previously described as characterizing specimens of which one of the parent sex cells had been irradiated previous to fertilization.


462


CHARLES RUSSELL BARDEEN


2. Period of relative immunity

After the fusion of the two pronuclei there appears to be a passive period during which the ovum prepares for the period of great activity which characterizes cleavage. During the passive period the ovum becomes less susceptible to the rays than during the preceding period of fertilization or during the subsequent periods of cleavage. This is indicated in the three experiments summarized in the following table.

TABLE 4 Irradiation following fusion of pronuclei



1

•<

3s:




Results



s









K


PS




GASTRULATION



GASTRULATION




s


«H


B


S:


INCOMPLETE




COMPLETE





P 00


B









M











§


5« 


O


§



j




Si



1



Remarks


B a


K


CL.


13

•4-3

o




.2


>



•a




O % o


MB BETWE TION AND I


O U

o


B

n S



1



-5 2

ci

d

0. 02


o « 


J2



S 1

4^ tn

•So

(0 Q.


0)

en



H


>J


2


12;


2;


i,


'Z,


<


<!


Q


55




Rr.


Min.



Per



Per



Per


Per


Per


Per


Per







cent



cent



cent


cent


cent


cent


cent



A


1


45


98


2



37



10


37


14




All specimens died two days after irradiation.


C


1


30


107


4.7






4.7


19.8


11.2


59.6


Experiment discontinued two weeks after irradiation.


D


u


30


30


[13. 3l



3.8





23.1


23.1


50


Experiment discontinued two weeks after irradiation.


The unfertilized specimens in lot D were not counted in figuring the percentage of eggs variously affected by the rays.

In Experiment A-f , a string of eggs was exposed for three quarters of an hour from a period beginning three-quarters of an hour after the mixture of the sperm and eggs. The effects of exposure during this period were less severe than those found in Experiment A-0, Table 3. Two per cent of the eggs did not undergo cleavage but it is uncertain whether or not this failure was due to the exposure. In the control practically all the eggs were fertilized. Of the eggs undergoing cleavage 37 per cent showed irregularities early in gastrulation and ceased develop


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 463

merit. In these eggs a yolk plug projected from a large blastopore. In the rest of the eggs the blastopore was completely or nearly closed. In 10 per cent of the total number development ceased with the closure of the blastopore; in 37 per cent it ceased after the formation of abnormal larvae without definite head or tail rudiments; in 14 per cent abnormal larvae with irregular heads and tails were produced. The abnormal specimens in the various lots resembled those previously described in the sections on the effects of exposure of the sperm and of the ova to the rays.

The relative immunity during this period of rest is still more strikingly shown in Experiment C-1, in which eggs were exposed for a period extending from one hour to an hour and a half after fertilization. Of these eggs 4 . 7 per cent failed to undergo cleavage but this probably means little since 6 . 9 per cent of the control specimens were unfertilized. The rest of the eggs passed through the period of gastrulation and over one-half of these appeared normal at the end of two weeks when the experiment was discontinued. Of the total number of eggs 4.7 per cent appeared abnormal early in larval differentiation, 19.8 per cent late in larval differentiation, and 11.2 per cent early during the tadpole period.

In a series of experiments with another lot of eggs the period of ' rest ' as measured by relative immunity to the rays came later than in the two experiments above mentioned. The results of the first two exposures in this series have been described above, Experiments D-j and D-f, table 3. Nearly all the specimens exposed were affected. The third exposure of eggs from this lot was for one-half hour, beginning one and one-fourth hours after the eggs were fertilized. Fifty per cent of these eggs appeared to develop normally and several were followed to metamorphosis. It is probable that in the batch of eggs used in this series of exposures the periods of fusion of the nuclei and of rest were later than in the other experiments, so that the period of relative immunity was also later.


464 CHARLES RUSSELL BARDEEN

3. Early cleavage stages

In the toad's egg a fissure marking the first cleavage plane usuall}^ appears at ordinarj^ room temperatures in from two hours and a quarter to two hours and three-quarters after fertilization. When it is warm cleavage appears earlier; when it is cold, somewhat later. There is considerable individual variability shown by different eggs of the same lot. The fissure of the second cleavage plane varies in the time of its appearance from three and one-quarter to six hours; the third varies from four to nine hours; the fourth from five to ten hours. In the later cleavage stages the variability becomes still more marked but by the twelfth hour rapid cell division is usually well under way. It is during the earher period of cleavage that the susceptibility of the ovum to the rays becomes greatest. This is illustrated in table 5a.

In this table are summarized the results of exposing groups from two batches of eggs at successive forty-five minute intervals, A-lf, A-2|, A-3i A-4, A-5, A-6, A-6f, A-7i A-Qf, A-11, and E-4 and E-12, groups from three other batches for thirty minutes, C-1^, C-2, D-lf, D-6, and All, and groups from a sixth batch for fifteen and thirty-five minutes, (G-6).

The eggs exposed for thirty minutes show from a period one and one-half hours after fertilization onwards an increasing susceptibility to the rays. In lot C-1^, 10.2 per cent of the fertilized eggs developed into tadpoles and nearly half of these lived for two weeks after fertilization. Discontinuance of the experiment at this period makes it impossible to state how many of these might have developed normally and undergone metamorphosis. Comparison of this experiment with experiments C-1 and D-l|, table 4, shows that after the period of 'rest' after fertilization, (p. 462), susceptibility rapidly increases as the period of cleavage approaches. This increased susceptibility is still more marked in Experiment D-lf in which the eggs were exposed between one and three-quarters and two and one-quarter hours after fertilization and from which no free swimming tadpoles developed. It is slightly less marked in Experiment C-2, exposure two to two and one-half hours after fertilization, than in Experiment D-lf,


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS


465


but this is probably due to an individual difference in susceptibility in the two lots of eggs. In Experiment C-2 but 1.6 per cent developed into free swimming tadpoles and these died within ten days after fertilization of the ova.^


TABLE 5a Irradiation during cleavage


1


k

as?

K <!

H M

< •Z. PS

^S

h2


PS

p

m O P(

O

o


m

» S u 1^ P. m


pg P


RBStn^TS




GASTRULATION INCOMPLETE


GASTRULATION COMPLETE



O

o

%

2 n


■n 1

U

1


k "3

1^


Cm


CO


o S 12;


<


Ea


T3

ID

>

O CD

p


1

o

s Is


Remarks



Hr.


Aftn.



Per

cent


Per

cent


Per

cent


Per

cent


Per

cent


Per

cent


Per\ Per\ Per cent cent cent



C


u


30


156



6.7





5.8


77.3


10.2


4.8 per cent normal two weeks after















irradiation.


D


If


30


41


[8.91



12.2




58.5


29.3





C


2


30


121



23.1





9.1


66.1


1.6




A


If ?

2}


45 30 45


103 58 41


7


89

4


11.7 14.6


6.7


8.3

4.9






All


73.3



A


43.936.6



A


3i


45


161



93.6 6.4









A


4


45


50



72 28








2d. cleavage fissure visible in some ova.


E


4


45


29


[541


75.824,2








2-8 cell stage. None survive. 62 hrs.















after irradiation.


A


5


45


41



93


7








2-4 cell stage.


D


6


30


28


10.7


23.453.6


1.5


10.8






4-8 cell stage.


A


6


45


38



90 10









A


6f 6

6


45 15

35

45


30 72

52 65


[21 [41


49 51


20.8 1.5




15.3




4-8-16 cell stage


G


8.3


55..5


8-16 cell stage. Only one spec, lived




10.8



to 6th day. Died 7th.


G


100


16-64 cell stage.


A


1.5 86


16-32 cell stage. None alive 2 days














after irradiation.


A


^


45


69


1.4


56.5


21.7


4.4


8.7


7.2




32-64 cell stage. None alive 4 days















after irradiation.


A


11


45


39


2.5



12.5



17.5


10


30


27.5



Advanced cleavage.


E


12


45


110


[401



15.7




78.9


5.4




None alive 5 days after irradiation.


^In experiments C-l| and C-2 mould and degeneration made it impossible to subdivide the specimens which did not develop intolarvse. Most of these, however, were probably unfertilized eggs. In Experiment D-lj, 8.9 per cent of the total number of eggs were apparently not fertilized. The lack of development in these eggs is not, however, to be attributed to exposure to the rays, since 9 per cent of the control eggs were not fertilized.


466 CHARLES RUSSELL BARDEEN

The increasing susceptibility of the ova as the period of cleavage is approached may likewise be seen by comparing Experiment A-lf, exposure from one and three-quarters to two and one-half hours after fertilization, with Experiment A-f, exposure from three-quarter to one and one-half hours after fertilization. In Experiment A-f, table 4, 61 per cent of the ova completed gastrulation and 14 per cent did not show marked deformities until after the rudiments of the head and tail were fairly well developed. In Experiment A-lf gastrulation was not completed in a single individual. In 7 per cent there was no cleavage. Since in the control practically all eggs were fertilized there is a possibility that in these eggs showing no cleavage the process was inhibited by the exposure to the rays. In 89 per cent of the eggs development stopped very early in gastrulation. In 4 per cent there was marked abnormality of gastrulation with a yolk mass projecting through a large blastopore. It is the period of greatest susceptibility.

Following the first cleavage there seems to be a second period of lessened susceptibility. In Experiment All a lot of eggs which had been recently laid in the lake and which were fertilized naturally showed practically all of the eggs to be in the two cell stage. Exposure of this lot of eggs to the X rays for one-half hour caused 73.3 per cent of them to cease developing early in the blastula stage, 11.7 per cent to become spina bifida specimens in which organ differentiation was slight, 6.7 per cent to stop developing as soon as the blastopore was closed, and 8.3 per cent to develop into very abnormal larvae without well marked heads or tails.

In Experiment A-2§ eggs exposed from two and one-half to three and one-quarter hours after fertilization likewise showed less marked effects than Experiment A-lf. ,While in the latter experiment no eggs developed beyond the early gastrulation stage in the former 14.6 per cent developed into spina bifida specimens and 4 . 9 per cent into abnormal larvse without definite heads and tails.

On the other hand, immediately preceding the appearance of the fissure of the second cleavage plane the susceptibility once


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 467

more becomes very high. (Experiments A-3j and A-4.) In these experiments most of the ova stopped developing very early in the period of gastrulation. A few show irregular gastrulation with a large protruding yolk plug. In Experiment E-4 eggs in about the same stage of development as those in Experiment Al-4 (2-8 cell stage but chiefly 2 cell) were exposed for forty-five minutes with a result closely similar to that found in the latter experiment.

During the eight and sixteen cell stages the susceptibility remains very high, as may be seen by comparing Experiments A-5 D-6, A-6 and G-6 with the experiments just described. It will be noted that even an half-hour exposure is sufficient to inhibit development beyond mere gastrulation and to stop the process of gastrulation at a very early period except in about 10 per cent of specimens in one experiment, D-6. After a fifteen minute exposure, G-6, but 15.3 per cent developed into definite larvae and these were abnormal and died early.

After the sixteen cell stage the susceptibility gradually becomes less as shown in Experiments A-6f, A-7|, A-9|, and A-11. In Experiment A-6f , (exposure from six and three-quarters to seven and one-half hours after fertilization,) 5 per cent of the eggs exhibited an irregular gastrulation with protruding yolk; in Experiment A-7^, (exposure from seven and one-half to eight and one-quarter hours after fertilization,) 10.8 per cent acquired a closed blastopore before ceasing to develop; in Experiment A-9|, (exposure from nine and one-half to ten and one-fourth hours after fertilization) 15.9 per cent developed into pigment covered larvae of which nearly half showed heads and tails; in Experiment All, (exposure from eleven to eleven and three-quarter hours after fertilization,) 57.9 per cent developed into pigment covered larvae half of which showed heads and tails. In Experiment A-95, 21 . 7 per cent of the specimens were of spina bifida type. After the period represented by this experiment exposure to the rays did not cause spina bifida although in one instance it caused an egg to cease development early in gastrulation. The eggs in Experiment E-12 were unusually susceptible compared with those in Experiment A-11.

THE AMERICAN JOURNAL OF ANATOMT, VOL. 11, NO. 4


468 CHARLES RUSSELL BARDEEN

The frog's egg during cleavage shows a susceptibiUty at least equal to that of the toad's egg. Frog's eggs exposed for forty minutes immediately preceding the first cleavage division stopped development either just before or at the time of the appearance of the dorsal blastopore lip. Frogs' eggs in the four to eight cell stage exposed for forty minutes all stopped developing before the dorsal blastopore lip became clearly marked. Exposed in the sixteen to sixty-four cell stages they stopped developing as soon as the dorsal blastopore lip became well marked. Exposed in the more advanced cleavage stages the blastopore in many specimens became small but all specimens died before assuming definite larval form.

Summary and analysis. During the early cleavage period, from the latter part of the second to the twelfth hour after fertilization, the ova of the toad are exceedingly susceptible to the X rays. The susceptibility increases from the period of 'rest' following the fusion of the pronuclei, up to the first cleavage division, there is then a slight decrease in susceptibility followed by a second increase up to the second cleavage. After the formation of four blastomeres there is again a slight decrease in susceptibility followed by a third increase which lasts during the eight and sixteen cell stages and then steadily declines during the subsequent cleavage stages.

It is thus apparent that preparation for cleavage in the ovum brings about some alteration in the organism which renders it especially susceptible to the rays and that corresponding alterations are brought about in the first eight blastomeres. The greater irregularity in cell division which takes place after the first eight to sixteen blastomeres are formed may account for the decrease in susceptibility after this period. During the half or three quarters of an hour during which the organisms are exposed cells about to divide will be in a stage of heightened susceptibility, cells in the resting stage will be in a state of relatively reduced sensibility. As the latter increase in relative number we should expect a decrease in susceptibility in the organism as a whole. It is also not improbable that when relatively few cells


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 469

are severely affected the less affected cells may not only recover themselves but also be enabled to overcome to some extent the injurious effects of the more affected cells of the organism. Whether this explanation alone would account for the decrease in susceptibility in the later cleavage stages is open to question. It is not improbable that as differentiation proceeds the individual cells become relatively more stable as they decrease in potentiality of differentiation.

The susceptibility during the early cleavage stages is greater than that exhibited by either of the sex cells before fertilization or by the fertilized ovum during fertilization. The effects of exposure are not, however, exhibited immediately, even during the periods of greatest susceptibility. Cleavage usually goes on until the period of gastrulation is approached. As a rule, cleavage is more or less incomplete in the yolk laden protoplasm of the vegetable poll. After forty-five minutes exposure to powerful X rays during the early cleavage period gastrulation is either inhibited in its earliest stages or begins in an irregular manner and then stops with a mass of yolk projecting through a very large blastopore. After thirty minutes exposure the blastopore may close but development then stops. During cleavage after exposure the cleavage cavity is not infrequently abnormally small. On the other hand, it may be abnormally dilated. Spina bifida specimens seem to arise usually from forms in which the cleavage cavity has been abnormally small. When the cleavage cavity is well developed but gastrulation incomplete in that the yolk is not completely covered, the yolk mass in the larval stage protrudes through the anal region instead of the back.

In frogs' eggs the susceptibility is likewise greatest during the early cleavage stages but it appears to decline less rapidly than in toads' eggs." The cause of this is obscure.

Hertwig and others^ have shown that eggs vary in susceptibility to changes in temperature and other external conditions

^For the literature see Handbuch der ver^leichen und experimentellen Entwicklungsgeschichte.


470 CHARLES RUSSELL BARDEEN

at different stages of development. Fertilized eggs of the frog kept at a temperature abov.e a maximum which varies for different species, develop abnormally. At the animal pole cleavage is very rapid; at the vegetable it is very slow and spina bifida larvae are frequently produced. Freshly fertilized eggs placed in water at a temperature of 0° Centigrade and kept there twenty-four hours develop abnormally, (Hertwig) while in the gastrulation stages egg may be kept at a low temperature for fourteen days and still develop normally when brought to the room temperature. (Schultze)

E. Godlewsky (1908), found in sea urchin eggs no increase in the amount of nuclear material in the first two blastomeres. Each nucleus was approximately half the size of the nucleus of the fertilized ovum. The proportion of nuclear to cytoplasmic material was estimated as 550:1 in the unfertilized egg; as 275:1 in each of the first two blastomeres. During cleavage from the two to the sixty-four cell stages there is a very rapid production of chromatin. The nuclei increase in number but remain about the same size. The proportion of nuclear cytoplasmic material increases from 275:1 to 12:1. From the sixty-four cell stage to the termination of the blastula stage the nuclei increase rapidly in number but decrease so much in size that the total amount of nuclear material is relatively slightly increased. The proportion of nuclear to cytoplasmic material changes from 12:1 to 6:1. With the decrease in size of the nuclei a relatively greater amount of nuclear surface is presented to the cytoplasm. During the gastrula and pluteus stages as the cells multiply there is a gradual increase in the amount of chromatin in the organism. The nuclei do not change much in size.

Similar studies on amphibian eggs have not, so far as I am aware, been made. It is highly probable, however, that the relations of nuclear to cytoplasmic material during cleavage are homologous. The period of greatest susceptibility of amphibian ova to the X rays, therefore, probably corresponds with the period of greatest relative activity in the production of nuclear at the expense of cytoplasmic material. If this be true we should expect that in


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 471

the toad's egg there is a rapid production of chromatin preceding the first cleavage.^

In this connection I may refer to an interesting experiment on the action of concentrated sugar solution on frogs' eggs. I find that if freshly fertilized eggs frogs' are placed in a six to eight per cent sugar solution and are left there from two to eight hours and then transferred to water all die within a few days. Very few eggs develop past the neural groove stage. On the other hand, if left in the same solution from twenty to thirty hours, and then transferred to water, many lived to pass through a normal development. This experiment indicates that a sudden change of condition during the period when the chromatin is being most rapidly manufactured much more seriously interferes with the normal course of development than if development during this period is allowed to proceed under conditions far from normal.

For refined cytological studies the eggs and larvse of anura are not well adapted and for this reason no prolonged and detailed examination of the cells in the specimens obtained from the experiments have been attempted. In the abnormal larvse many evidences of direct nuclear division are to be found, especially in the epidermis. On the other hand, occasional apparently normal mitotic figures are to be seen in all of the tissues. Giant nuclei are not infrequent, especially in the mesodomic tissues. In the cells of the central nervous system nuclear abnormalities are especially apt to be well marked. An irregular clumping of large masses of chromatin within the nucleus is a common occurrence. In specimens exposed during early cleavage and preserved immediately after exposure a nucleus within a definite membrane within the cleavage spindle may sometimes be seen. It would appear as if the nuclei in mitosis were forced into a resting stage within the spindle by the X rays.

^Hertwig, 1910, in ascribing the great susceptibility of fertilized ova to the irradiation of both the male and the female pronucleus has failed to give sufficient weight to the increase of susceptibility during the early cleavage stages.


472


CHAKLES RUSSELL BARDEEN


EFFECT OF TEMPERATURE


Several experiments were made to test the effect of temperature on the action of the X rays. Since at a not too high temperature the eggs divide more rapidly than at a low temperature, and in consequence in such eggs the anabolism of nuclear material is more rapid, we should expect the former to be more susceptible than the latter. That this is the case is clearly indicated by the following table :

TABLE 5b. Effects of variations in temperature on eggs daring the early cleavage stages



5


S H

^.§

B Eh


S m

H H

m &

(7

9 « 

[0


TIME BETWEEN FERTILIZATION AND IRRADIATION


fa

K

p o

b O

W

g

z


w

n



Results



O

El H

O «  IS

a


-(J

1


CO



a ft

02


^1

Its


8a >

>

u o

<


'2

Sffi


1

13 en

Q


12;


Remarks


6A

a

b c

6B a

b

c

6C a

b c UA


Deg

66

66 66

50 50 50

50

50 50


60°

70°

70° 70°

57° 40°

42 hr.

42 hr.

42 hr.

57°

40°

18 hr.

18 hr.

18 hr.

77°


Hours

4

12 3i

5i 13 4i

5i

13

4i


Min

30

30 30

35 35 35

35

35 35


440

81

72

?

34 ?

86

61 65


4 cell

32-64 cell 2-4 cell

No cleavage

32-64 cell No cleavage

No cleavage

32-64 cell No cleavage


(4)


Per

cent

98 100 100

100 50

88


Per

cent

2

12

2


Per cent

100

59

26

3

8


Per cent

41

12

87 3


Per

cent

10


Per

ceni


Per cera<


Natural fertilization all died within 8 days after Irradiation.

Natural fertilization.

After 42 h. transferred to room temperature.

All died wlthm 6 days after fertilization.

After 18 hrs. transferred to room temp. All dead within 6 days after fertilization.

All dead within ten days after fertilization


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS


473


This table includes Experiments 6 and 11. In Experiment 6, eggs from three batches 'A/ 'B,' and 'C were exposed for about half an hour to the X rays. One set was exposed at room temperature, about 66 degrees F. ; two other sets were exposed in water at about 50 degrees F. One of these sets was kept for 18 hours, the other for 42 hours out of doors. The out-of-doors tempera


TABLE 5b— Continued



o ■< a a

b

o m

a

s ^


m s ■

H

z P ^


TIME BETWEEN FERTILIZATION AND IRRADIATION


a


O

g


ED

z m s 3

m n

S


g P

H a

> 6,

go oS

H H

r


Results



designation op

periment



i

m


a « 


3 ft

02


^1

1^


8i >■

> u

o

<


1^


la

Q



Remarks



Deg

77


80°


Hours 3


Mm 15


101


2-4 cell



Per cent

33


Per

cent

1


Per

cen<

37


Per cent

25


Per cent


Per cent

3



al


2


Natural fertilization.
















One spec, followed to


a2


77


80°


21


11


160


2 cell



6



10


82


1


1


metamorphosis. Artificial fertilization. One followed to meta

bl


77


80°


3


35


178


4-8 cell



63



32


5




morphosis. Natural fertilization. All dead by 4th day


b2


77


80°


2i


35


92


2-4 cell



77



16


7




after Irradiation. Artificial fertilization. Last spec, died 6 days after irradiation.


IIB



65°














al


67


70°


4


15


179


2-4 cell



11


15


32


39


3



Natural fertilization. Last spec, died 10 days


a2


67


70°


3i


15


139


2 cell


(1)


5


9


33


47


6



after irradiation. Artificial fertilization. Last spec, died 17 days





bl


67


70°


4


32


109


2-4 cell



69


4


28






after irradiation.


b2


67


70°


3i


32


106


2-4 cell


(I)


67


5


28






Natural fertilization. Artificial fertilization.


lie


50


65°















al


57


70°


5


15


100


1-2 cell



16


14


21


30



3


16


Natural fertilization.
















3 followed to meta



a2


57


70°


4i


15


151


1-2 cell


(1)


3



14


8


7


68


morphosis. Artificial fertilization. 4 followed to metamor

bl


57


70°


5


35


93


2 cell



5


10


43


37


5



phosis.


b2


57


70°


4}


35


193 144


2 cell Control


(1) (6)


7


12


15


53 15


9 1


4 84


Natural fertilization.

Artificial fertilization. One followed to metamorphosis.


474 CHARLES RUSSELL BARDEEN

ture at this time varied from 40 degrees - 57 degrees and averaged below 50 degrees ; this proved to put a severe strain on the eggs kept outside 42 hours although of these about half of those eggs in the sixteenth to thirty-two cell stage ('C') developed further than anj^ of those of the same lot exposed at room temperature, A large part of the eggs kept outside for 18 hours after exposure at 50° developed better than any of those exposed at room temperature.

In Experiment 11 eggs from two batches, one fertilized naturally, one artificially, were exposed for 15 and for 30 minutes,, some at 77 degrees, some at 67 degrees, and some at 50 degrees57 degrees. The eggs exposed at 77 degrees were subsequently kept at that temperature on the top of an incubator. The other lots were kept at room temperature after exposure. A greater per cent of the eggs exposed at 77 degrees than those exposed at 66 degrees showed injury early in development but, on the other hand, a greater number recovered. A much greater percentage of the eggs exposed at 50 degrees than those of the two later mentioned, developed normally showing that a moderately low temperature partially protects the eggs against the action of the rays.

4. Advanced cleavage and gastrulation stages up to the closure of the blastopore {twelve to thirty-six hours after fertilization).

During this period there is at first a rapid and then a much more gradual decline in susceptibility.

In table 6 are summarized the results of experiments with five different batches of fertilized eggs. In Experiments A, A-1 and B, forty-five minute exposures were given. In Experiment A-14^, exposure from fourteen and a half to fifteen and a quarter hours after fertilization, the susceptibility was decidedly less than in Experiment A-11, table 5. In a third of the eggs in the latter gastrulation was incomplete or very abnormal and only a fourth of the eggs developed into larvse with distinct heads and tails. None of these larvse developed into tadpoles. In Experiment A-14^ the blastopore was closed in all specimens and all reached the larval stage. Over half of the larva? became free


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 475

swimming tadpoles but all of the latter died early, none surviving beyond nineteen days.^

In subsequent periods a greater and greater number of larvse reached the tadpole stage and an increasing number of these developed normally up to metamorphosis. In Experiment 15f , exposure from fifteen and three-quarters to sixteen and one-half hours after fertilization 19.2 per cent developed normally; in Experiment A- 17^ exposure from seventeen and one-half to eighteen and one-quarter hours after fertilization, 23.2 percent; in Experiment A-25^, exposure from twenty-five and one-half to twenty-six and one-third hours after fertilization (early in the gastrulation period) 73 . 7 per cent ; in Experiment A-29 exposure from twenty-nine to twenty-nine and three quarter hours after fertilization, (blastopore moderate or small) 84.6 percent; in Experiment A-30^, exposure from thirty and onefourth to thirty-one hours after fertilization, (small blastopore) 90.9 per cent; and in Experiment A- 1, exposure from thirty-one and one-half to thirty-two and one-quarter hours after fertilization, (blastopore small or closed) approximately 100 per cent. Exposure to the X rays for two successive periods of forty-five minutes each during this period caused all the eggs to develop abnormally and none reached the tadpole stage (Experiment A-20).

In Experiment B on another batch of fertilized eggs, successive groups of which were likewise exposed for forty-five minutes, the susceptibility proved at first to be much greater than in the batch just described. The causes for this greater susceptibility are uncertain. In the various groups exposed before the twentieth hour after fertilization in only one instance did a larva reach the free swimming tadpole stage, but this went on to metamorphosis. (Experiment B-18|.) In Experiment B-15|, exposure from fifteen and three-quarters to sixteen and one-half hours after fertilization, one egg failed to complete the process of gastrulation, 3 . 3 per cent stopped developing as soon as the blastopore was closed,

^Hertwig, 1910, has found that exposure to radium for from one-half to one hour causes the blastulae of the oxolotl and the frog to die before gastrulation is complete.


476


CHARLES RUSSELL BARDEEN


65 per cent after abnormal development died early in the period of larval differentiation, and 30 per cent died in the later periods of larval differentiation. None developed into free swimming tadpoles and none lived beyond the eighth day after fertilization. In the groups subsequently exposed all specimens showed some larval differentiation and a successively greater number developed apparently normally until the latter part of larval differentiation. In Experiment B-21|, exposure from twenty-one and one-half to twenty-two and one-quarter hours after fertilization, 6 per cent of the eggs developed into free swimming tadpoles but none of those lived longer than two weeks; in Experiment B-22|, exposure from twenty-two and one-half to twenty-three and onefourth hours after fertilization, 16.7 per cent developed into free swimming tadpoles; in Experiment B-26, exposure from twentysix to twenty-six and three quarter hours after fertilization, 35 . 3


TABLE 6.

Irradiation at stages from advanced cleavage to closure of blastopore.

Experiment A


I t

k:5


g


1



Results




l« 


00


1 '







h 2


o


S








» ^


8



•B


Jt





fa





s



Remarks


« H O


a



1


"08


>

4^ m




MB

LIZA

lATI


O

z


n

S


.a


a g


<o

<1> o>


as



a


s


a


<!


<!


Q


z



Hours


Min.



Per cent


Per

cent


Per

cent


Per

cent



14J


45


31



58


42



At end of 3rd day after Irradiation 33 per cent were very abnormal. None lived after 19th day.


151


45


26



15.4


65.4


19.2


Retarded development but some metamorphosed.


17§


45


31



19.4


57.4


23.2


Retarded development.


20


45\



48.5


51.5




During 2d. exposure the beginning of gastrulatlon was


22}


45)


68




evident;? of spec, were dead by the 11th day and all by









the 11th day.


25}


50


38




26.3


73.7


At time of exposure gastrulatlon was well started . Development at first delayed but many spec, were followed through metamorphosis.


29


46


28




15.4


84.6


At time of exposure blastopore was small or moderate In size. At first development was retarded.


30}


45


22




9.1


90.9


Small blastopore at time of exposure. Development normal In most specimens.


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS


477


TABLE 6— Continued Experiment Al


Hours


Min.






Per cent


3H


45


36





about 100


Small blastopore at time of exposure. All appeared to develop normally although not all were followed through metamorphosis.


z a m %


Hours

161 171 18}

20f


21} 45

22} 45


45


23i

26

33}


45 45 45


Min


45 60 45 45 45


Experiment B


Results


-3


■a o


13


a 8 Q r


0)

>


I o

03


j3 .n>^ (u ft o H


Per^ Per' Per\ Per Per Per cent cent cent cent cent cent


1.7


.3 65 30 36.2 63.81


|36.7 39.4

12.5

18.2


87.5


5.9


75.8 0.6 83.3 16.7 87.5 12.5 58.8 35.3


1.5


100


Remarks


Development retarded. None lived beyond 8th day. Development retarded. None lived beyond 8th day. Development retarded. None lived beyond 8th day.

Only one specimen lived beyond the 9th day. This specimen metamorphosed .

Backward in development. All dead one week after irradiation.

Backward In development.

Backward In development.

Backward In development.

Backward in development.

Development apparently normal in all specimens not all were followed through metamorphosis.


None lived more than two weeks. None lived more than two weeks. None lived more than two weeks. None lived more than two weeks, though


Experiment C


Hours


Min.





Per cent


Per cent


13


30


71




25.4


74.6


17


30


72




36.1


63.9


23


30


75




29.3


70.7


Con


trol


72




18.1


81.9


The percentage In the last column Indic ates the number alive two weeks after fertilization. Gastruiation was well advanced 23 hours after fertilization.


Experiment D


Hours


Min





Per cent


Per

cent


15


30


68





20.6


79.4


19


30


75





30.6


69.4


25


30


87





31


69


Cont


rol


425





17.6


82.4


Percentage in last column as in Experiment C. Gastruiation visible In 19th hour and nearly complete in 25th hour in most specimens.


478 CHARLES RUSSELL BARDEEN

per cent. None of these tadpoles lived more than two weeks after exposure. In experiment B-33|, however, exposure thirtythree and one-half to thirty-four and one-quarter hours after fertilization, all the eggs developed into normal tadpoles and a considerable number of these were followed through metamorphosis. At the time of exposure this group of eggs had moderate sized or very small blastopores.

The results summarized in table 6, Experiments C and D followed the exposure of successive groups from two batches of eggs in which the control specimens showed a considerable percentage of defective and abnormal larvae, (about 18 per cent in each case.) The causes of the defects are uncertain. The eggs in each case were laid while the female was in captivity and were fertilized by a male present in the same jar. There was a small amount of water in the jar and possibly the eggs were in some way poisoned. It was late in the season so that there is also a possibility that the eggs were over-ripe. The exposure of each of the groups of eggs was for thirty minutes. The exposed eggs showed a greater percentage of larvae which failed to develop into free swimming tadpoles than the control. In both groups the eggs from the thirteenth to the fifteenth hour, before gastrulation was apparent, showed less susceptibility than in the seventeen and nineteen hours when the process of gastrulation was commencing. A similar decreased susceptibility preceding gastrulation was not clearly marked in the groups previously described. In Experiments C and D the development of the tadpoles was not followed more than two weeks after fertilization so that the number capable of ultimate normal development was undetermined.

The internal changes found in abnormal larvae and tadpoles derived from eggs exposed during the period of gastrulation seem essentially similar to those found in larvae derived from exposed sex cells. In general, however, the alterations produced seem to be somewhat more diffused in the former than in the latter. After exposure of the spermatozoa and to a less extent after exposure of the ova before fertilization one not frequently finds the abnormalities confined mainly to a restricted part of the body or to a single organ system, but when specimens are exposed dur


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 479

ing and after gastrulation the alterations are not regionally localized although much more marked in some organs than in others.

Of the various organ systems the central nervous system, including the optic stalk and retina, is the most regularly affected. The neural tube is usually dilated and thin walled. It seems quite evident that the neuroblasts lose the power of normal differentiation and, failing to differentiate in a normal manner, they finally undergo retrograde metamorphosis giving rise to irregular protopoasmic masses frequently pigmented. The earlier the time or the more severe the exposure the earlier do the disturbances in the neuroblasts become manifest.

The vascular system is likewise usually markedly affected in specimens exposed during gastrulation. The heart and blood vessels are, as a rule, very rudimentary and the blood corpuscles seem to be much reduced in number or missing. External gills usually appear in these specimens although not as a rule in specimens severely exposed during the earlier stages.

The alimentary canal in the specimens severely exposed during gastrulation seldom shows much differentiation of coils and the pancreas and liver are rudimentary. The mouth in most specimens is patent. The pharynx is poorly differentiated.

The pericardial and peritoneal cavaties are usually greatly dilated.

The pronephric tubules are sometimes abnormally dilated. The sex cells seem fairly normal.

The chorda is usually relatively normal but may appear distended.

The myotomes and the muscles of the head are as a rule poorly differentiated.

Frog eggs show, at times at least, a greater susceptibility to the X rays than toad eggs during the gastrulation period. In one experiment in which frogs' eggs with small blastopores were exposed fifty minutes to the rays practically all the larvae at the time of hatching showed marked abnormalities and none survived this period. The chief abnormalities were seen in the alimentary canal, the central nervous system and the vascular system. In a considerable number of specimens both the gut and


480 CHARLES RUSSELL BARDEEN

the neural canal were abnormally dilated. In another experiment out of 111 fertilized eggs exposed at the beginning of gastrulation for two and one-half hours 32 . 4 died early in the period of larval differentiation and the rest before hatching. In a third experiment all of a lot of seventy-five frogs' eggs exposed at the time of closure of the blastopore were abnormal in form by the period of hatching and died just before or after this period.

Godlewski's interesting studies on the relative proportions of nuclear to cytoplasmic material during the early development of the eggs of sea urchins have already been cited, (p. 470.) During the blastula period he finds that the nuclei increase rapidly in number but at the same time decrease in size so that the proportion of nuclear to cytoplasmic material is not greatly increased.

It is essentially a period of distribution of nuclear material, of increase of nuclear surface but of slow increase of nuclear substance. If similar conditions prevail in amphibian eggs the greater susceptibility during the earlier cleavage stages as compared with the blastula stages may probably be ascribed to a greater susceptibility of cells when the production of nuclear material is rapid than when it is slow. This likewise would apply to the gastrula and early larval stages. In the sea urchin Godlewski finds during the gastrula and pluteus stages that the nuclei gradually multiply in number but remain of about the same size. The production of new nuclei material is far less rapid than in the early cleavage stages.

Ruffini ('08) has shown that in amphibian eggs gastrulation depends largely on cell migration and on cell secretion. The neural tube is likewise formed by processes of cell migration and cell secretion, the latter due to the periectoderm cells which come to line the medullary tube. Osmosis likewise plays a part in the swelling of closed cavaties. There is little evidence that exposure to the rays affects these various processes directly. Exposure during the early cleavage stages may prevent gastrulation or render gastrulation very abnormal by preventing the formation of the normal cells on which the process depends. But if these cells are once formed gastrulation usually goes on well in spite of severe exposure. The effects


SUSCEPTIBILITY OF AMPHIBIAN OVA TO X-RAYS 481

come out later when larval formation calls for the differentiation of