American Journal of Anatomy 8 (1908)

From Embryology
Embryology - 16 Sep 2019    Facebook link Pinterest link Twitter link  Expand to Translate  
Google Translate - select your language from the list shown below (this will open a new external page)

العربية | català | 中文 | 中國傳統的 | français | Deutsche | עִברִית | हिंदी | bahasa Indonesia | italiano | 日本語 | 한국어 | မြန်မာ | Pilipino | Polskie | português | ਪੰਜਾਬੀ ਦੇ | Română | русский | Español | Swahili | Svensk | ไทย | Türkçe | اردو | ייִדיש | Tiếng Việt    These external translations are automated and may not be accurate. (More? About Translations)

Historic Disclaimer - information about historic embryology pages 
Mark Hill.jpg
Pages where the terms "Historic Textbook" and "Historic Embryology" appear on this site, and sections within pages where this disclaimer appears, indicate that the content and scientific understanding are specific to the time of publication. This means that while some scientific descriptions are still accurate, the terminology and interpretation of the developmental mechanisms reflect the understanding at the time of original publication and those of the preceding periods, these terms and interpretations may not reflect our current scientific understanding.     (More? Embryology History | Historic Embryology Papers)
Amer. J Anat.: 1 - 1901-02 | 2 - 1902-03 | 3 - 1904 | 4 - 1905 | 5 - 1906 | 6 - 1906-07 | 7 - 1907-08 | 8 - 1908 | 9 - 1909 | 10 - 1910 | 11 - 1910-11 | 12 - 1911-12 | 13 - 1912 | 14 - 1912-13 | 15 - 1913-14 | 16 - 1914 | 17 - 1914-15 | 18 - 1915 | 19 - 1916 | 20 - 1916 | 21 - 1917 | 22 - 1917 | 23 - 1918 | 25 - 1919 | 26 - 1919-20 | 27 - 1920 | 28 - 1920-21 | 29 - 1921 | 30 - 1922
Other Historic Journals: Amer. J Anat. | Anat. Rec. | J Morphol.
Links: Historic Embryology Papers | Association of American Anatomists | Wistar Institute of Anatomy and Biology | Wliey - Developmental Dynamics | Internet Archive | biodiversity library

THE AMERICAN JOURNAL OF ANATOMY

EDITORIAL BOARD

CHARLES R. BAKDEEN,

University of Wisconsin.

HENRY H. DONALDSON,

The Wistar Institute.

THOMAS DWIGHT,

Harvard Universitif.

SIMON H. GAGE,

Cornell University.

O. CARL HUBER,

University of Michiyan.


GEORGE S. HUNTINGTON,

Columbia University.

FRANKLIN P. MALL,

JoJins IlopJcins University.

J. PLAYFAIR McMURRICH.

University of Toronto.

CHARLES S. MINOT,

Harvard University.

GEORGE A. PIERSOL.

University of Pennsylvania.

HENRY McE. KNOWER, Secretary.

Johns Hopkins University.

VOLUME VIII

1908

PUBLISHED QUARTERLY BY THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY

36th street AND WOODLAND AVENUE

PHILADELPHIA, PA.

CONTENTS

I. Howard Ayers and Julia Worthington. The Finer Anatomy of the Brain of Bdellostoma Dombejd. 1. The Acustico-Lateral S3'stem 1 With 8 Phites.

II. Effie a. Eead. a Contribution to the Knowledge of the Olfactory Apparatus in Dog, Cat and Man 17 With 17 Plates and 1 Text Figure.

III. William F. Allen. Distribution of the Subcutaneous Vessels in the Tail Eegion of Lepisosteus 49 With 25 Figures.

IV. Fred J. Taussig. The Development of the Hymen 89 With 14 Figures.

V. Irvixg Hardesty. On the Nature of the Tectorial Membrane and its Probable Role in the Anatomy of Hearing. 109 With 13 Figures.

VI. Charles Russel Bardeex. Early Development of the Cervical Vertebra? and the Base of the Occipital Bone in Man 181 With 3 Figures.

VII. Henry Fox. The Pharyngeal Pouches and their Derivates in the Mammalia 187 With 73 Figures.

VIII. Leonard W. Williams. The Later Development of the Notochord in Mammals 251 With 20 Figures.

IX. George L. Stueeter. The Peripheral Nervous System in the Hnnian Embryo at the End of the First Month 285 with 3 Plates and 1 Text Figure.

X. Walter A. Baetjer. The Origin of the Mesenteric Sac and Thoracic Duet in the Embryo Pig 303 With 10 Figures.

XL William E. Kellicott. The Growtli of the Brain and Viscera in the Smootli Dogfisli (Mustelus canis, Mitchill). 319 With 7 Plates.

XII. Harris H. Wilder. The Morphology of Cosmobia. Speculations concerning the Significance of certain Types of Monsters 355 Witli 4 Plates and 3? Text Figures.

XIII. Mabel Bisiior. Heart and Anterior Arteries in Monsters of the Piceplialus Group : A Comparative Study of Cosmobia 441 With r Plates and 5 Text Figures.

The Finer Anatomy Of The Brain Of Bdellostoma Dombeyi 1. The Acustico-Lateral System

By

Howard Ayers And Julia Worthington. With Eight Plates.

In keeping with its position near the lower end of the vertebrate phylum, Bdellostoma possesses a brain relatively simple in its structure, yet much more complex than that of Amphioxus. It is much less complicated than that of even the lowest fishes, and represents an early and most important stage in the phylogeny of the vertebrate brain. The vertebrate brain plan, already clearly and^ firmly outlined in Amphioxus, is developed so much farther that all the fundamental brain organs found in higher vertebrates are present in Bdellostoma, and the fundamental divisions and fiber tracts, persistent in all other vertebrates, including man, are clearly defined. The connections between different parts of the brain are, as we would expect to find them, simpler and more direct than in higher forms. ~0f especial importance is the fact that the fundamental relations are not obscured by the intrusion of the many secondary tracts and centers which are present in the higher representatives of the phylum. Hence, it seemed to us that a careful study of the brain of Bdellostoma, a survey of its paths, and a charting of its relay stations, would throw much light not only on the many intricate questions of brain anatomy, but on that still more intricate and baffling problem, the origin and development of the vertebrate head.

The gross anatomy of the brain of Bdellostoma, together with the peripheral distribution of the cranial nerves, has been worked out in a previous paper (Worthington, '05), and reference will be made thereto for macroscopic relations.

We recognize two main divisions of the brain of Bdellostoma, the Hindbrain, including the medulla and cerebellum, and the Forebrain, including the midbrain, "tweenbrain, and forebrain. This division is fundamental and is present in Amphioxus. The Hindbrain will be described first in this and succeeding papers, each one of its component systems being studied both as a separate organ and in its relation to the other brain organs. The present paper is devoted to a study of the Acustico-lateral system, in both its internal and external relations.

Methods.

The material on which this study is based consists of embryos, the young and adult hagfish — species Bdellostoma dombeyi, Lac, — collected by the writers at various times from the Bay of Monterey. The hagfish were caught in traps and kept in healthy condition in a large tank supplied with running salt water. They were taken out as needed and the living tissue treated according to the chosen method. This course was found to be absolutely necessary, as the brain tissue deteriorates very rapidly. When the hagfish Avere caught by hook and line and died in consequence, brains taken from them within two hours after death uniformly failed to give the Golgi reaction. The methods found most satisfactory for the study of the relations of the fiber tracts in the adult brains were the rapid Golgi method and Cajal's absolute alcohol and ammonium-absolute alcohol silver nitrate methods, though good results along certain lines were sometimes obtained by intra-vitam methylene blue staining. The results obtained in these ways have been controlled by brains hardened and cut in situ and stained with carmine, haematoxyline, etc. We have not been able to obtain the characteristic reaction from experiments with the Weigert methods.

We wish here to express our thanks to Miss Elizabeth Worthington for her kindness in making the wash drawings for Tigs. 47, 48 and 49.

The Acustico-latekal System. The name acustico-lateral system is given to the group of nerves comprising the two ear nerves, the two lateral line nerves, anterior and posterior, and the acusticus nucleus with its connections to other parts of the brain. The nucleus itself is very different in Bdellostoma from what it is in higher forms, even in its nearest relative Petromyzon. In this latter form (Johnston, '02) the tuberculum acusticum consists of three distinct parts, a ventro-lateral nucleus, a dorso-median nucleus lying close upon it, and, dorsad to both, separated from them by an area of fine fibers, a third nucleus, the lohus linece lateralis. In Bdellostoma, while the lohus linear lateralis is fairly well marked out, it is not entirely distinct, as in Petromyzon, and there is no differentiation of the other two parts.

The Acusticus Nucleics. — The acusticus nucleus (Figs. 1, 2, 3, 4, 26 and 30) is a long and narrow club-shaped structure, much thicker in its cephalic than in its caudal half. It is about 2.15 mm. long, .675 mm. broad just before it tapers to its cephalic end, about .6 mm. deep at its deepest part about midway of its length, and from .3 mm. to .4 mm. deep through its caudal half. This difference in depth is caused by the arrangement within the nucleus of the entering fibers. It lies in the dorso-mesial part of the dorsal column of the medulla, separated from the dorsal surface of this body by the fasciculus communis. It appears somewhat wedge-shaped as it lies in the medulla, its mesoventral angle being its most ventral part, the dorso-mesial and ventrolateral surfaces both sloping dorsad. The ventro-lateral surface is comparatively fiat, but the dorso-mesial surface, although quite flat in the caudal third of the nucleus, is very convex in the cephalic part. The long axis of the nucleus runs fore and aft in its hind part, but about three-quarters of the length of the nucleus from its cephalic end it bends laterad, at an angle of about 120°, following the curvature of the dorsal column of the medulla.

There are five sets of fibers that enter the acusticus nucleus, those of r. acusticus utricularis, of r. acusticus saccularis, of N. lateralis posterior, and two sets for N. lateralis anterior. Of these entering fibers those of r. acusticus saccularis are the most caudal set (Figs. 26 and 30). They enter along the lateral face of the nucleus from about 1.1 mm. to 1.3 mm. from the caudal end. Immediately cephalad of the fibers of r. saccularis, and with only a break of about .03 mm. between the two groups, the fibers of r. acusticus utricularis enter.

The Auditory Fibers. — The auditory fibers of the r. utricularis and those of the r. saccularis will bp considered together because their internal distribution is the same. The fibers are straight and of varying diameters when they enter the brain, some measuring 2.2 microns, while some are no more than .9 microns in thickness. They run cephalo-mesad, at an angle of about thirty degrees with the long axis of the brain until they reach the lateral border of the acusticus nucleus. After entering the nucleus the fibers divide into two branches, one running cephalad and one caudad within the nucleus (Figs. 5, 6, 26


4 Howard Avers and Julia Worthington

and 30). Each of these branches may, and generally does, divide again, and sometimes a third time. After the secondary or tertiary division one branch will usually continue in the same general direction as before, the other frequently turns mesad. In many cases it could be followed no further, but in some, after reaching a more mesial part of the nucleus, it turned again and continued in the same direction as its fellow. Figs. 6 and 29 are good illustrations of the branching of these fibers. After each branching the fibers are smaller in diameter, and after the secondary branching they also become beaded; the final product in Golgi sections is a very fine fiber bearing round or oval beads (Fig. 29). These ascending and descending fibers form a large central core, extending almost the entire length of the nucleus, and we have not been able to find any direct connection between these fibers and the cells of the nucleus, i. e., no end plates resting against the body of the cell. When, however, the nucleus is studied in methylene blue and Cajal preparations, it is seen to possess, in addition to cells and fiber:^, an exceedingly fine intercellular network, extending all through it, and composed of the finest fibers thickly beaded with tiny beads. In Cajai sections a few of the finer fibers of the entering acusticus fibers have been found to be connected with this network, and it is probable that this is the destination of all of them.

The Lateralis Posterior. — This lateral line nerve (Acusticus h, Worthington, '05) enters the medulla on the dorsal surface at the level of the saccularis fibers, entering the acusticus nucleus directly (Figs. 3, 8 and 10). It runs cephalo-meso-ventrad until it reaches a point about one-third of the way between the dorsal and ventral surfaces of the nucleus; here it turns latero-cephalad, and runs as a distinct bundle of fibers along the mesial edge of the nucleus (Figs. 5, 6 and 11). As it nears the cephalic end of the nucleus, the fibers separate somewhat (Fig. 32), some of them turning dorsad, some ventrad (Fig. 13), to •distribute themselves in the dorso-cephalo-mesial angle of the nucleus (Figs. 8 and 14). The fibers measure between 1.8 microns and 2.7 microns in diameter before they turn cephalad, and are of quite uniform size. Afterwards, during their course cephalad, some of them become irregular in outline (Figs. 13 and 32), the thickened parts being ovoid in shape, and sometimes double the diameter of the fiber, — occasionally the irregularities are almost as sharply defined as the beads of the regular acusticus fibers, Init this appearance is very seldom seen. Most of the fibers, however, maintain approximately the same thickness as far as they can be traced. jSTot many of tliem branch, but occasionally one is found, after it has turned cephalad, that divides into two branches as in Fig. 32, both branches continuing cephalad. One Golgi slide also showed a few fibers of this nerve which, after division, sent a branch caudad, and one fiber that turned caudad without division. Branching fibers are, however, infrequent.

The Lateralis Anterior. — This nerve is not a separate and distinct trunk like the lateralis posterior; instead its fibers join the trunks of such nerves as give most convenient passage to their destination. The largest and most conspicuous bundle of lateralis anterior fibers joins the posterior sensory trunk of the trigeminus and runs with it to the skin of the side of the head where these lateralis fibers supply some if not all of the anterior group of lateral line canals. The fibers of this combined trigeminus and lateralis trunk pass over the utricular ganglion in close juxtaposition to it, and enter the brain surrounded by the fibers of the auditory utricular root (Figs. 15, 26 and 30). So closely are these fibers intertwined and so difficult are they to follow that, judging from hematoxylin sections, it was thought they all entered the acusticus nucleus, and the entire trunk was called lateralis anterior {Acusticus a, Worthington, '05). With Golgi and Cajal sections, however, their true relations are made plain. Figs. 16 and 17 show the greater part of the fibers of this trunk running meso-caudad in the general cutaneous nucleus, while a smaller part turns directly mesad, entering the acusticus nucleus with a bundle of utricular fibers (Fig. 16). We have not yet succeeded in disentangling this middle bundle of lateralis anterior fibers from its accompanying utricular fibers inside of the nucleus. This particular bundle cannot be identified in the Golgi sections at our disposal ; the Cajal sections show it to run mesad and slightly caudad to the center of the nucleus, but as the fibers subdivide they are lost in the maze.

The anterior bundle of the lateralis anterior is smaller than the posterior bundle just described. It accompanies the anterior sensory trunk of the trigeminus. After this trunk has passed through the cranial capsule on its way to the brain, several small bundles of fibers separate themselves slightly from the others, and instead of entering the general cutaneous nucleus directly, like the rest of the trunk, enter the acusticum at its ventro-cephalic angle (Fig. 18). These bundles run caudad for a short distance through the ventral part of the acusticum. and then those of them that belonor to the general cutaneous system curve laterad into the general cutaneous nucleus (Fig, 18), while the lateralis fibers remain behind. As this extreme meso-cephalie part of the nucleus is also the part in which the fibers of the lateralis posterior distribute themselves, as previously stated, it may be called the representative of the lobus Imece lateralis of higher forms. This anterior division of the lateralis anterior supplies certain neuromasts in the tentacles.

There is a probability that there is still a third division of the lateralis anterior which accompanies the facialis. After the latter nerve has passed through the cranial capsule on its way to the brain a very small bundle of fibers is given off that runs dorsad. Having penetrated about to the middle level of the utricular ganglion, the bundle appears to divide, sending some of its fibers dorsad with the utricular auditory root, and the rest mesad with the utricular general cutaneous fibers. This bundle is difficult to trace; it is almost impossible to follow it from the facialis unless the angle at which the sections are cut is favorable. We have not been able to trace its fibers with absolute surety to their entry into the brain; nor have we succeeded as yet in tracing the facialis to its endings in the skin, and consequently cannot state that any of its fil)ers innervate lateralis neuromasts. Lacking this confirmation, we do not care to state positively that this small bundle contains lateralis fibers, but can only say that the probabilities are all in favor of this conclusion. If these probabilities should prove to be actualities, then all three of the divisions of the lateralis anterior as found in the higher forms, the two that compose the "Dorsal VII"' of the Amphibian facial nerve, and the hyomandibular branch, are represented in Bdellostoma.

The Cells of the Nucleus. — "When the acusticus nucleus is studied in sections, it is found that the entering fillers form a central core, running fore and aft throughout its entire extent. Cells are interspersed among these fibers, but the great majority of the cells lie outside of the core as a surrounding cortex, and they are massed in particular dorsad and ventrad of the core. The cells may be divided roughly into two classes, large and small.

The Small Cells. — The small cells are in overwhelming majority, and are found all through the nucleus. They measure from 5.9 microns to 8.5 microns in diameter, and the cell body may be round, pear-shaped, or spindle-shaped. They have large round nuclei occupying nearly the entire body of the cell, and one or more prominent nucleoli. In some Cajal sections fine fibers running towards these cells divide, on reaching them, into two or more fibrils which apply themselves closely to the surface of the cell.

We have not been able as yet to identify these fibers. In the sections where they appear they are impregnated in a somewhat different manner from the ends of the entering fibers, and as yet we have not succeeded in tracing them into definite bundles.

When the small cells are studied in Golgi sections they are seen to belong to both the spindle and multipolar types of cells. Fig. 31 shows two small cells lying in the dorso-cephalic part of the nucleus. The left cell is the more mesial of the two and belongs to the spindle-shaped variety. The fiber running mesad is cut off close to the cell; the other, running latero-ventrad, forks a little beyond the cell, giving rise to three others, all of which run more or less laterad. The one that runs farthest forks again, one of its subdivisions being a fine-beaded fiber. None of the prolongations of this cell run beyond the acusticus nucleus. The other cell is of the multipolar variety, giving off two very fine and two heavy prolongations. The two fine fibers run laterad and mesad. Of the two heavy ones the shorter runs ventrad, the longer dorsad; this latter, about half way of its lengih from the cell, becomes very fine and beaded with round or oval beads. It penetrates among the fibers of the fasciculus communis that overlie the acusHcum. Figs. 33 and 35 show small cells lying in the ventro-mesial part of the nucleus about midway of its length, close to the fibers of the great ventral commissure. Fig. 33, which likewise sends its axone into the ventral motor column of its own side, is a tri-polar cell, having, beside the axone, one short fine fiber, and one stout one that subdivides later into two fine ones. Fig. 35 is a spindle cell whose mesial prolongation bifurcates, sending one branch dorsad, the other ventrad. The dorsal fiber is cut off close to its starting point, the ventral one divides again, giving off a finebeaded fiber running mesad, and the axone that runs ventrad, and pierces through the ventral commissure to end in the ventral motor column of its own side.

The Large Cells. — The large cells are few in number compared to the small ones. Occasional ones are found all through the nucleus and a large group of them occurs in the ventral part caudad of the middle. These large cells are generally elongated in shape with large nuclei. Sometimes they are spindle-shaped, as in Fig. 34, sometimes tri- or multipolar (Figs. 36 and 37). These cells, whatever their form, usually give out very thick protoplasmic processes. Sometimes the axone is found to arise from one of these processes at a little distance from the body of the cell; in other cases, however, it is impossible to distinguish the axone from the other processes.

The connections established by these large cells are quite various. Some of them send their axones ventrad into the ventral motor column of the same side of the brain; some send them across the brain through the ventral raphe. Sometimes they apparently serve to establish connection between different parts of the nucleus. The cell in Fig. 31 lies about in the center of the nucleus and sends one process dorsad and the other ventrad, where it divides in the extreme ventral part. It is possible that the cell processes, though not to be followed in this section, really penetrate beyond the acusticum into the general cutaneous nucleus. In some horizontal sections large cells lying in the dorsal part of the nucleus near the middle send axones caudad into the dorsocaudal part. In one section a fiber was seen that, though it could not be traced to a cell, resembled in every particular the axones of the large cells in the same locality; this fiber, first seen in the dorsal part of the acusticum of the left side of the brain, slightly cephalad of the middle, crossed to the dorso-caudal part of the nucleus of the right side.

Many of these cells lie on or near the borders of the nucleus and send their processes into adjacent parts of the brain. In Fig. 37 the cell lies on the ventral border of the acusticus nucleus; processes 1, 2 and 3 go into the upper layers of the acusticum, while process 4 divides into branches that penetrate into the general cutaneous nucleus. Fig. 38 shows a cell lying in the lateral part of the nucleus that sends process a into the nucleus, and processes d^ and d' into the general cutaneous nucleus. In this cell a appears to be the axone. Fig. 43 is a similar cell in the cephalic end of the acusticus nucleus which sends its axone into the nucleus, processes rf^ an'd d- into the general cutaneous nucleus, and d^ among the fibers of the fasciculus communis. On the other hand, cells are sometimes found that lie in the general cutaneous nucleus close to the acusticum that send some of their processes into the acusticum while their axones remain in the general cutaneous nucleus (Fig. 40). Thus the connection is established in both directions between these two important centers of the medulla.

The Secondary Fiber Tracts of the Acusticum. — Thus far we have considered the connection of the acusticus nucleus with other parts of the brain by means of individual cells located in various parts of the nucleus. We will now describe the outgoing fiber tracts. First are the fibers that cross in the ventral raphe. These fibers come from all parts of the nucleus, run ventro-mesad to its ventro-mesial angle, and cross to the other side of the brain in the raphe without forming a distinct bundle (Figs. 26 and 30). All of the fibers of the raphe distribute themselves in the ventral part of the brain of the opposite side, and it is impossible to differentiate those derived from the acusticum from the fibers of a different origin. All that can be said is that by this means a motor connection is established with the opposite side of the brain. A motor connection is also established between the acusticus nucleus and its own side of the brain by means of a large tract that forms about the level of the entering auditory fibers, and crosses the raphe to run directly to the ventral motor tract of its own side of the brain (Figs. 1, 2 and 26). Arrived there, many of its fibers turn caudad, running with the other fibers of this tract. How far they continue cannot at present be stated. Individual cells, both large and small, taking part in the formation of this tract have already been mentioned, and are shown in Figs. 33, 35, 36 and 39.

Another important outgoing tract may be called the tr. acusticofuniculi. This tract is forked in the acusticum. The dorsal fork arises at the level of the entering auditory fibers and lies on the lateral border of the acusticum. It receives many fibers from the interior of the nucleus, fibers that run laterad or caudo-laterad until they reach this tract, but which, upon reaching it, turn caudad and form part of it. Toward the hind part of the acusticum this division joins the ventral fork of the tr. acustico- funiculi. This ventral division arises in the extreme cephalic part of the nucleus and runs caudad along its lateral border imtil it nears its hind end. Here, turning dorso-meso-caudad, it rises to the dorsal part of the nucleus, receiving on its way the dorsal division of the tract (Fig. 26). The combined tract leaves the nucleus at its dorso-caudal angle and runs caudad, immediately ventrad of the nucleus fasciculus communis, some of whose cells send processes into it, into the nucleus funiculi (Figs. 1, 2, 3, 4, 26 and 30). There is also a small tract us acustico-cerehellaris. This tract arises near the cephalic end of the acusticum along the mesial border, near the mid dorsoventral plane. It runs caudo-mesad close to the mesial surface for about 1.2 mm.; then it turns mesad into the cerebellum, runs slightly dorso-mesad, and crosses to the other side. After crossing its fibers separate and turn cephalad. They run dorso-latero-cephalad in small bundles for varying distances, and are eventually distributed in the cerebellum (Figs. 26 and 30, tr. a. c). In addition to this definite cerebellar tract an occasional fiber may be seen that leaves the central part of the acusticum and runs cephalad into the cerebellum. There is no indication that these fibers cross.


The Auditory Nerves.

In Bdellostoma there are two distinct auditory nerves, each with its &wTt ganglion, one for each of the two main divisions of the auditory 5ac. These nerves are the ramus acusticus utricularis, and the ramus acusiicus ^accularis.

The Auditory Ganglia. — These two ganglia lie laterad of the medulla between it and the cranial wall (Figs. 15, 21, 22, 23 and 24) ; they are closely apposed to each other, and the two are wrapped together in a common connective tissue sheath, so that when dissected out they appear as one body, and it requires microscopic study to show their histological separateness. The utricular ganglion (Fig. 47) lies the more cephalad, and is also the larger of the two. It is cone shaped, with its apex directed dorsally. The long axis of its ovoid base runs parallel to the long axis of the lateral face of the medulla. The plane of the base is inclined at an angle of about 15° to the plane of the long axis of the brain. The extreme length of the base of the cone is about 1.37 mm. and its height about 1.25 mm. The caudal face of the cone is not uninterrupted in its slope from apex to base like the other faces. About .9 mm. ventrad of the apex it leaves the saccularis ganglion, to which it has been applied, and its face runs ventro-cephalad from this point of division for about .4 mm. It then makes an angle of 18° and runs ventro-caudad for about .4 mm. At this point there is a space of about .2 mm. between the dorsal and ventral limits of the ganglion, and across this entire face the ramus utricularis is given off (Fig. 47, VIIIu). Through the indentation mentioned above the facialis passes between the cranial wall and the brain (Figs. 21 and 47).

The utricular fibers pass from the ganglion to the brain in two distinct roots (Fig. 47), a dorsal root, whose emergence is confined to the dorsal third of the ganglion (this root consists exclusively of acusticus fibers), and a ventral root emerging along the mesial face from the indentation dorsad about .3 mm., consisting of general cutaneous fibers.

In histolooieal structure the utricular sfanalion is found to consist of two distinct sizes of nerve cells, large ones measuring from 46 microns to 66 microns and small ones measuring from 16 microns to 26 microns. The large cells are found in all parts of the utricular ganglion, but the small cells are largely confined to its ventral portion, although a fairly large number are found in its dorso-caudal part (Fig. 15). They are particularly numerous in its ventro-caudal part (Fig. 21 and 23), but comparatively few are found in the ventro-cephalic part of the ganglion. The large cells are the ganglion cells of the general cutaneous fibers, which innervate the lining epithelium of the ear apart from the sense organs (Fig. 19) ; the small cells belong to the acusticus fibers. Fig. 23 shows how these two sets of fibers cross in the ganglion to reach their proper paths of entry to the brain.

The general cutaneous fibers enter the brain in numerous bundles arranged in a series along the lateral surface of the medulla and penetrate directly into the general cutaneous nucleus. In Worthington, '05, this is described as an acusticus root, as it was thought at that time that its fibers penetrated to the acusticus nucleus. Golgi sections, which had not been obtained at that time, disprove this completely, as they show that these ventral fibers turn and run cephalo-caudad in the general cutaneous nucleus (Fig. 24). This is not the ventral root mentioned by Sanders, '94, the group of fibers that he calls the ventral root of the acusticus being in reality the motor root of the ti^igeminus. Holm, '01, probably saw the root, but did not attach any significance to it, for he speaks of it as a few fibers that leave the ganglion here and there and enter the medulla."

The saccularis ganglion, when stripped of its nerves, may also be considered as a cone, this time an inverted one, its apex directed ventrad, and having the base cut away at its caudal end (Fig. 47). It is about .37 mm. from the base to the apex of the cone, and its greatest cephalocaudal length is also about .37 mm. The fibers of ramus saccularis leave the ganglion in two distinct sets, those of ramus saccularis anterior (Fig. 47, VIIIs. a.), leaving from the apex of the cone, and those of ramus saccularis posterior (Fig. 47, VIIIs. p.), leaving from the middle portion of the caudal surface. The cells of the saccularis ganglion are similar in size and character to the acusticus cells of the utricular ganglion.

The ramus saccularis has but one root, a dorsal one that leaves the ganglion at its dorso-cephalic end and enters the brain caudad and dorsad of the dorsal root of the ramus uiricularis. The saccular nerve does not, apparently, carry an) general cutaneous fibers.


The Lateralis Nerves.

The lateralis system is but slightly developed in Bdellostonia when compared with its condition as found in higher fishes and the Amphibia. Its peripheral nerves are distributed exclusively to the head, but both anterior and posterior lateralis nerves are represented, and tl:e sense organ canals and isolated lateral line organs, neuromasts, are also present. The N. lateralis posterior leaves the brain as a distinct cranial nerve, while the fibers of the lateralis anterior are bound up with and accompany both trunks of the trigeminus. There are indications that a few of them may -also accompany the facialis.

There are two groups of lateral line canals in the head of Bdellostonia, an anterior group, innervated by the lateralis anterior, and a posterior group, innervated by the lateralis posterior. The anterior group is composed of four, occasionally three, or five short canals, nearly equidistant from each other, and located on the side of the head in front of the eye of its side of the body (Fig. 48). The posterior group lie? on the dorsal surface of the head and consists of two divisions ; the three (occasionally two) inner canals run meso-laterad, and the outer ones run at a slight angle to the long axis of the body (Fig. 49). These canals are almost impossible to find in the full grown adult, but in young hagfish, about eleven inches long, they are easily seen on heads that have been hardened in chromic acid. A description of their structure and of the effect of different hardening agents upon the underlying tissue has already been given in a previous paper (Ayers and Worthington, '07).

The lateralis components in the nerves of Bdellostonia, owing probably to their inferior bulk and the intimate association of the anterior fibers with those of other systems, have been overlooked by most of the previous workers in Myxinoid anatomy, not only by tho^e who were working principally by dissection and dealing with the entire head, but also by those who were working with sections and making a special study of the brain. Sanders, '94, makes no mention of them whatever, not having detected their presence even in the second trigeminus trunk, and apparently not having seen that part of the lateralis posterior that lies between the cranial capsule and the brain. He seems to have included a description of the peripheral part of the lateralis posterior in his "upper division of the trigeminus/' but his account of the distribution is so brief and vague that it is difficult to tell what he really does mean. Max Fiirbringer, '97, calls the lateralis posterior the first spino-occipital nerve. Holm, '01, who has done very good work on the internal structure of the brain of Myxine, and who has produced some beautiful plates, unfortunately follows Sanders in this matter, as in several others of equal importance, and considers the second trunk of the trigeminus to be composed entirely of general cutaneous fibers, though he recognizes that its root is distinct from the rest of that nerve. The lateralis posterior he has apparently failed to find. None of these writers, even those who have dealt with the internal anatomy of the brain, have shown a true conception of the complex nature of these trigeminus trunks; Allis, '03, comes nearest, when he calls the second trigeminus trunk r. huccalis lateralis n. trigeminus II. He has not found the larger trigeminus component of this trunk, but considers it to be solely lateralis in character, and declares it to be homologous with the huccalis facialis of the higher fishes. In addition to this, Allis finds lateralis fibers in the first trunk of the trigeminus, and finds also the lateralis posterior, which he interprets correctly as a lateral nerve, and, while not feeling sure of its homology, is inclined to call it the linecv lateralis vagi.

Summary.

We find that the several parts of the acusticus nucleus are connected with each other by individual cells and their processes. The acusticum is connected with its fellow of the opposite side of the brain and with the general cutaneous nucleus of its own side by individual cells. The connections between the acusticus and general cutaneous nuclei are both numerous and intimate, as shown by the very considerable number of cells lying on either side of the boundary that send out interpenetrating fibers between the two. This close connection tends to support Johnson's theory that the tuherculum acusticum and the general cutaneous nucleus have developed concomitantly from the same fundamental part, the dorsal horn of the spinal cord, — the morphological differentiation being due to a division of labor with a specialization of function. Certain cells of the acusticus nucleus send fibers into the tractus fasciculus communis just as they do into the general cutaneous nucleus. Whether this is reciprocated by the cells of the communis nucleus has not yet been ascertained. The acusticus nucleus is connected with the lobe of the cerebellum of its own side of the brain by individual fibers.


t is connected with the cerebellar lobe of the opposite side by a fiber tract. It is also connected by a fiber tract with the ventral motor column of its own side of the brain, and with the ventral motor column of the opposite side by the great number of fibers that cross in the ventral raphe. It is connected with the nucleus funiculi, and hence with the spinal cord, by a tract into which cells of the communis nucleus also send processes.

ISTeither the trigeminus, the facialis, nor the auditory nerves of Bdellostoma are mono-functional nerves. Each carries at least two, and probably more, sets of functionally distinct fibers. The auditory nerve consists in reality of two distinct nerves, N. utricularis and N. sacculai'is. N. utricularis carries a large general cutaneous component that innervates the lining membrane of the ear apart from the sense organs. The N. lateralis anterior runs with both trunks of. the trigeminus and probably with the facialis also. The N. lateralis posterior, instead of being associated with the NN. glosso-pharyngeus and vagus, has a trunk to itself. We have found no trace of lateral nerves or organs in the region of the trunk, consequently a lateral line and nerve sensu siriciu is not formed in Bdellostoma.

CiNCixNATi, June, 1907.


LITERATURE CONSULTED.

1903. Allis, Edward P., Jr. Ou Certain Features of the Cranial Anatomy of Bdellostoma dombeyi. Anatomischer Anzeiger, Bd. XXIII, pp. 260-281, 322-339.

1892. Ayers, Howard. A Contribution to tbe Morphology of the A'ertebrate

Ear, with a Reconsideration of its Functions. Journal of Morphology, Vol. YI, Nos. 1 and 2.

1907. Ayers, Howard, and Worthington, Julia. The Skin End-Organs of the Trl(ieminus and Lateralis Nerves of Bdellostoma dombeyi. American Journal of Anatomy, Vol. VIT, No. 2.

1899. Clapp, Cornelia Maria. The Lateral Line System of Batrachus Tau. Ginn & Co., Boston.

1902. CoGHiLL^ G. E. The Cranial Nerves of Amblystoma tigrinum. Journal of Comparatii:e Neurology, Vol. XII, No. 3.

1896. Cole, FRA^-K J. The Cranial Nerves of Chimfera monstrosa. Proceedings of the Royal Society of Edinburgh.

1893. DoGiEL, A. S. Zur Frage liber das Verhalten der Nervenzellen zu einander. Archiv fur Anatomic und Physiologic.

DoGiEL, A. S. Zur Frage iiber den Bau der Nervenzellen und iiber das VerMltniss ihres Axencylinder-(Nerven)-Fortsatzes zu den Protoplasmafortsiltzen (Dendriten). Archiv fiir milcrosTcop. Anatomic, Bd. XLI.

1906. Edingek, L. Ueber das Gehirn von Myxine glutlnosa. Sitzungsher. d. P7-eHSS. Akad. Wiss., Anhang. Sep. Berlin, Reimei*.

1897. FiJRBRiNGEK, Max. Ueber die Spino-Occipitalen Nerven der Selachier und Holocephalen, und ihre vergleichende Morphologie. Festsch. TOsten Geburtstage von Carl Gegenbaur, Dritter Band.

1875. FtJRBBEiNGER, PAUL. Muskelatur des Kopfskelets. Jenaische Zeitschrift f. Natu7-it\, Bd. IX, iNO. 1.

189G. GoEONOwiTSCH, N. Der Trigemino-Facialis Komplex von Lota vulgaris. Festsch. 70sten Geburtstage von Carl Gegenbaur, Dritter Band.

1897. Herrick. C. Judson. The Cranial Nerve Components of Teleost,3.

AnatomiscJier Anzeiger, Bd. XIII, No. 16. 1899. Herrick, C. Judson. The Cranial and First Spinal Nerves of Menidia.

Journal of Comparative Neurology, Vol. IX, No. 3 and 4. 1901. Herrick, C. Judson. The Cranial Nerves and Cutaneous Sense Organs

of the North American Siluroid Fishes. Journal of Comparative

Neurology, Vol. XI, No. 3. 1901. Holm, John F. The Finer Anatomy of the Nervous System of Myxine

glutinosa. MorpJiologisclies Jahrljuch, Bd. XXIX, Heft 3. 1901. HouSER, Gilbert L. The Neurones and Supporting Elements of the

Brain of a Selachian. Journal of Comparative Neurology, Vol. XI,

No. 2.

1898. Johnston, J. B. The Hind Brain and Cranial Nerves of Acipenser.

AnatomiscJier Anzeiger, Bd. XVI, Nos. 22 and 28.

1901. Johnston, J. B. The Brain of Acipenser. Zoologlsches Jahrljuch,

Ahtli. f. Anatomie und Ontogenie, Bd. XV, Heft 1 und 2.

1902. Johnston, J. B. The Brain of Petromyzon. Journal of Comparative

Neurology, Vol. XIII, No. 1.

1905. Johnston, J. B. The Cranial Nerve Components of Petromyzon.

MorpJiologisclies JaJtrbucJi, Bd. XXXIV, Heft 2.

1903. JoRis, Hermann. Nouvelles Recherches sur les rapports anatomiques

des neurones. Hayez, imprimeur de I'Academie Royale de Medecine de Belgique, Bruxelles.

1906. Kappers, C. U. Ariens. The Structure of the Teleostean and Selachian

Brain. Journal of Comparative Neurology and PsycJiology, Vol. XVI, No. 1. 1895. Kingsbury, B. F. On the Brain of Necturus maculatus. Journal of Comparative Neurology, Vol. V, December.

1897. Kingsbury, B. F. The Structure and Morphology of the Oblongata in Fishes. Journal of Comparative Neurology, Vol. VII, No. 1.


1902. KiNGSLEY, J. S. The Cranial Nerves of Amphiuma. Tufts College

Studies, No. 7. 1900. VON KuPFFER, Carl. Zur Kopfentwickelung von Bdellostoma. Studien

zur vergleichonden Entwicklungsgescbichte des Kopfes der Kra nioten, Heft 4. 1881. Mayseb, p. VergleicbenJe Auatomisclie Studieu iiber das Geliirn

der Knoclienfische. Zeit. f. wiss. ZooL, Bd. XXXVI. 183T. MuLLER, Johannes. Anatomie der Myxinoiden.

1894. Sanders, Alfred. Researches in the Nervous System of Myxine glu tinosa. Williams & Norgate, London.

1895. Strong, Oliver S. The Cranial Nerves of Amphibia. Journal of

Morphology, Vol. X, No. 1. 1905. Worthington, Julia. The Descriptive Anatomy of the Brain and Cranial Nerves of Bdellostoma dombeyi. Quarterly Journal of Microscopical Science^ Vol. XLIX, Part 1, October.

ABBREVIATIONS. a. =: axone.

ac. c. = acusticus cells in the utricular ganglion. ac. f. = acusticus fibers. ac. n. = acusticus nucleus, c. =: cerebellum. (ZS d^, (Z^ = processes of nerve cells.

e. =: cavity of the ear.

f. 0.=- fasciculus communis.

g. c. c.=: general cutaneous cells in the utricular ganglion. g. c. /".=: general cutaneous fibers of the ear.

g, c. ».=: general cutaneous nucleus.

m. =: medulla.

s. g. ==: saccular ganglion.

tr. a. c. = tractus acustico-ccrebellaris.

tr. a. f. = tractus acustico-funicuU.

tr. a. t\^ tract from the acusticus nucleus to the ventral motor colunin at

the same side. u. fir. = utricular ganglion. V. c. = ventral motor column. V. r. =: ventral raphe.

X. = support on which the model of the acusticus nucleus rests. Vi^ fibers of anterior root of N. trigeminus. Vj=: fibers of posterior root of V. trigeminu!<. F7/=:V. facialis.

VIII I. a. a = Anterior branch of N. lateralis anterior. VIII I. a. p. = posterior branch of V. lateralis anterior. VIII I. p. = V. lateralis posterior. VIII s. = r. saccularis N. acustici. VIII u. = r. utricularis V. acustici.


PLATES.


PLATE I.

Fig. 1. Model of acusticus nucleus, dorso-lateral angle, x 18% Fig. 2. Model of acusticus nucleus, cephalo-ventral angle, x 18-2 Fig. 3. Model of acusticus nucleus, mesial surface, x 14i4 Fig. 4. Model of acusticus nucleus, dorsal surface. X 18-2 Fig. 5. Ascending and descending acusticus fibers in acusticus nucleus. X 41%.

Fig. 6. Branching acusticus fibers in acusticus nucleus. X 41%.


ANATOMY OF THE BRAIN OF BDELLOSTOMA DOMBEYI

HOWARD AYERS AND JULIA WORTHINGTON


PLATE II.

Fig. 7. Descending ncusticns fibers in aeusticus nucleus. X 41%.

Fig. 8. Entrance of .A. latcraJis posterior into aeusticus nucleus (sagittal section). X 66%.

Fig. 9. Fibers of ,Y. httcnilh posterior aiul y. aeusticus, rr. utricuJaris et saeciilaris. x 66%.

Fig. 10. Entrance of N. lateralis posterior into aeusticus nucleus (cross section). X 66%.

Fig. 11. Aeusticus nucleus through entering fibers. X 41%.

Fig. 12. Aeusticus nucleus through entering fibers. X 66%.


PLATE III.

Fig. 13. Fibers of N. lateralis posterior. X I'^O.

Fig. 14. Ci'oss section of dorsal columns of medulla, near the cephalic end. X 41%.

Fig. 15. Utricular and saccular ganglia and I'oot of N. trigeminus II (horizontal section). X 66%.

Fig. 16. Entrance of posterior branch of N. lateralis anterior with JSi. trigeminus II into the medulla, x 108%.

Fig. 17. N. trigeviinus II within the medulla. X 41%.

Fig. 18. Entrance of anterior branch of A", lateralis anterior with N. trigeminus I into the medulla. X 66%.


PLATE lllB

PLATE IV.

Fig. 19. General cutaneous fibers in the ear. x GG%.

Fig. 20. Cells lying in acusticus nucleus that send processes into the general cutaneous nucleus. X G6%.

Fig. 21. Utricular ganglion and y. facialis (horizontal section). X 66%.

Fig. 22. Utricular and saccular ganglia near their dorsal siu-face (horizontal section). X 41%.

Fig. 23. Cross section of the utricular ganglion, x 66%.

Fig. 24. Entry of general cutaneous fibers from the utricular ganglion into the medulla (horizontal section). X 66%.

Fig. 25. Entry of general cutaneous fibers from the utricular ganglion into the medulla (cross section), x 66%.



Fiti. H-i. I Entry of general cutsineon iiit'i the/u»?dul la) (horizontal section)

liG. 25. Entri of general cutaneous the medulla (croes section). X 66%


PLATE IVb

PLATE V.

Fig. 26. Diagram of acusticus nucleus (sagittal), x 24.

Fig. 27. Large cell of acusticus nucleus that sends axone into the ventral raphe.

Fig. 28. Cell of acusticus nucleus, x 250.

Fig. 29. Branching acusticus fibers, x 125.

Fig. 30. Diagram of acusticus nucleus (horizontal). X 24.

Fig. 31. Small cells of acusticus nucleus. X 250.

Fig. 32. Fibers of N. lateralis posterior where they begin to separate. X250.


PLATE VI.

Fig. 33. Small cell in acusticus nucleus, a runs to ventral motor column. X 250.

Fig. 34. Large cell in acusticus nucleus, d ramifies in ventral part, near the general cutaneous nucleus, x 458.

Fig. 35. Small cell in acusticus nucleus, a runs to ventral motor column. X 250.

Fig. 36. Large cell in caudal end of acusticus nucleus, a runs to ventral motor column, x 125.

Fig. 37. Large cell on ventral border of acusticus nucleus. Processes 1, 2 and 3 go into acusticus nucleus, process 4 into the general cutaneous nucleus. X458.

Fig. 38. Cell lying near the lateral border of acusticus nucleus, a runs into the nucleus, d^ and d^ into the general cutaneous nucleus, x 250.


PLATE VII.

Fig. 39. Cell in aciisticus nucleus.

Fig. 40. Cell in general cutaneous nucleus. 0} and (Z^ branch In the acusticus nucleus, x 458.

Fig. 41. Large cell in the ventro-lateral part of the acusticus nucleus. (1^ and d^ reach just outside the nucleus, x 333.

Fig. 42. Spindle cell in acusticus nucleus. It lies at the level of the entry of r. sacciilaris, and its axis is parallel to the entering fibers. X 250.

Fig. 43. Large cell in the cephalic end of acusticus nucleus, a is cut off, f^ runs to the fasciculus coiuniunis, d" and d^ stop among the trigeminus tracts. X 125.

Fig. 44. Group of large cells in the cephalic half of acusticus nucleus. Cell 3 lies just within the lateral border, d of cell 3 extends into the general cutaneous nucleus, a of- cell 4 runs into the ventral raphe, x 125.


PLATE VIII. Fig. 45. Large cell in acusticus nucleus.

Fig. 4G. Two large cells in acusticus nucleus, d- of cell 2 reaches the general cutaneous nucleus.

Fig. 47. The utricular and saccular ganglia, x 24.16.

Fig. 48. Head of a young hagfish eleven inches long, lateral view. X 1% Fig. 49. Head of a young hagfish eleven inches long, dorsal view. X !%•


A Contribution To The Knowledge Of The Olfactory Apparatus In Dog, Cat And Man

EFFIE A. READ, Ph.D.

From the Lahoratory of Histologj/ and Emhryology, Cornell University. With 17 Plates and 1 Figure.

CONTENTS.

PAGE

Introduction and statement of the problem 17

Historical summary 19

Gross anatomy of the olfactory nerves 24

Relation of the olfactory fibers and bundles to the olfactory mucosa .... 24

Comparative anatomy 26

Methods for gross dissection 27

Histological methods 28

Gross anatomy of the nose 29

Histological structure of the olfactory epithelium 31

The olfactory bulb 33

Distribution of the 5th nerve to the nose 315

Free terminations of the 5th nerve within the nasal mucosa 36

Organon vomeronasale 38

Gross anatomy of the organon vomeronasale 39

Histology of the organon vomeronasale 41

Results 41

Conclusions 42

Statement of the Problem.

1. Position, extent and character of the olfactory epithelium in (a) dog;

(&) cat ; (p) man.

2. The position and nature of the olfactoi-y cells.

'This paper was submitted to the Faculty of Cornell University as a thesis for the degree of Doctor of Philosophy, .Tune, 1907. I wish to express my grateful appreciation to Professor S. H. Gage, whose aid and encouragement made this work possible, and also to acknowledge the abundant material put at my disposal by the Departments of Physiology and Anatomy.

American .Touenal of Anatomy. — Vol. VIII.


18 Effie A. Eead

3. The olfactory nerve fibers as central prolongations of the olfactory cells

and the character of their termination in the olfactory bulb.

4. The relations of the olfactory nerve fibers in their passage from the

sensory epithelium to the olfactory bulb.

5. The position of the vomeronasal or Jacobson's organ and its sensory

epithelium and nerves.

6. The relations and terminations of the branches of the trigeminus in the

nasal mucosa.


As -will be more fully stated in the body of the paper, I have employed in this work the methods necessary for showing clearly the gross anatomy, and for the fine anatomy the three standard histological methods: (1) Gold chloride; (2) chrome silver or Golgi method, and (3) the methylene blue method.

Except where otherwise stated, the results given depend upon repeated gross dissections and upon clear demonstrations by each of the histological methods. That is, no statement has been made which has not been abundantly verified.

ISTaturally the quality of the human material available did not make all the histological demonstrations and verifications so extensive as for the dog and cat.

1-3. I have demonstrated in the clearest possible manner that the olfactory sensory cells are present in the slightly pigmented mucosa on the conchas and septum usually designated as olfactor}^ and that the sensory cells are true nerve cells and their central prolongations are the olfactory nerves, which extend to the olfactory glomeruli in the olfactory bulb. This work is confirmatory of the published results briefly summarized in the historical part of this paper.

My results as to the extent of the Olfactory Eegion in man differ considerably from those of von Brunn. the latest and most quoted authority npon this point. He shows only a small area upon the septum and superior concha as olfactory (Figs. 28, 29). My dissection shows that the olfactory nerves extend over a much larger area, al)out one-third of the septum and nearly the whole of the superior concha (Figs. 30, 31).

4. With regard to the relation of the olfactory nerves in their passage from the olfactory cells to the olfactory bulb, the results of my work are strikingly different from the published statements of human and comparative anatomists from the time of Scarpa to the present.

Xaturally the conditions are more completely described for man than for the lower animals. The figures of Leveille have been and still are


Olfactory Apparatus in Dog, Cat and Man 19

frequently copied into text-books (E. G. Barker's Laboratory Manual of Anatomy, 1904; Quain's Organs of the Senses, 1906) and represeiit pictorially the opinion of anatomists as to the true relation.

Instead of a plexus of the olfactory nerves I have found that the nerves extend in non-anastomosing bundles to the olfactory bulb. All appearance of anastomosis being due (a) to a crossing of the bundle of nerves or (6) to a net-like arrangement of the connective tissue or blood vessels.

5-6. The position of the vomeronasal organ and its innervation by the olfactory and 5th nerves have been shown in gross preparations of dog and cat. In histological preparations the sensory cells of this organ with their nerves have been demonstrated in the cat and mouse. Branches of the anterior ethmoidal nerve have been traced among the olfactory nerves of the conchfe in dog and cat ; their terminations among the folds were not found in gross specimens.

The naso-palatine nerve was found on the septum in all gross preparations. Kerves with free terminations were seen in histological specimens, both on the nasal septum and in the concha; it is thought that these are the endings of the 5th nerve.

These facts are in agreement with the results of other workers.


Historical Summary.

The olfactory region has been a subject of special investigation for many years. Various opinions concerning the endings of the olfactory nerves have been published. Some views have been disproven, but as early as 1856 Max Schultze had established with considerable certainty the true conditions of the endings of the olfactory nerves in the nasal mucosa. A review of the literature will give the present standpoint.

Eclvhard, 1855, found that the olfactory epithelium of the frog was formed of two kinds of cells, a cylindrical cell and a fusiform cell. These were morphologically and physiologically distinct. The cylindrical cell, an epithelial cell, had a central bifurcating end which terminated in the subjacent layer. The fusiform cell was entirely different from the epithelial cell. He suspected a difference in function, and thought, without doubt, that one was the true termination of a nerve fibril; but he did not say which one.

Eclcer, 1855, published his observations on the olfactory mucosa of man and some mammals. He saw two kinds of cells, a cylindrical cell


20 Effie A. Read

and a fusiform cell. The cj'lindrical cells reached the free surface and were connected, according to Ecker, by the central prolongations with the olfactory fibrils. These he called the true olfactory cells. The fusiform cells, replacement cells, situated at the base of the epithelium never reached the free surface. These replacement cells were simply stages in the development of the olfactory cells. Thus there was, according to Ecker, only one kind of cell, hut in difEerent developmental stages.

Scliultzc, 1856, worked on man, mammals, birds and amphibia. He found three kinds of cells, olfactory cells, epithelial cells and stellate cells. The epithelial cell was long, with a prismatic ])eripheral end. The central end was a short process and was connected with neighboring epithelial cells through side processes. These cells were pigmented but not ciliated. Between the epithelial cells he found cells of a peculiar chemical reaction. The cell bodies were round and had two processes, one reaching the free surface of the epithelial cells and the other passing to the connective tissue. This central process was the finer and could be recognized by the enlargements. The peripheral process was wide, at first, but tapered quickly and was then the same width to the surface. It bore at the end six to ten long brush-like hairs which were free in the air current of the nose. He describes each epithelial cell as surrounded by at least four to six of these hair cells.

In a comparison of these peculiar fiber cells of the olfactory region with other known cell forms, he first emphasizes the fact that in no other epithelial layer, either in the nose outward from the olfactory region, or back in the air tubes, is a trace of sucli varicose fiber cell found. The stellate cells, which lie under and between the surface cells, do not have the form, length or nature of the other cells of the olfactory region.

Pie believes the nerve cells of the retina to be the most favorable for comparison with these cells. By a comparison of these with the bipolar cells and by a comparison of the chemical reaction of the two cells, it is highly probable, according to Schultze, that these cells are also ganglion cells. He adds that comparative researches have made it as good as certain that the varicose fiber cells of the olfactory region are nerve cells. It, however, lacks proof of a direct connection with the fiber of the olfactory nerve. He concludes by saying that it is highly probable that the varicose fiber cells are the peripheral ends of the olfactory nerves. It is to these cells, and not to the epithelial cells, as Ecker thought, that the name olfactory cell should be given. These cilia


Olfactory Apparatus in Dog, Cat and Man 21

bearing cells serve both to collect the molecules of odorous substance and to serve directly in their perception.

In 1862 he still had never seen the olfactory fibers connected with the bipolar cells, but believed there was no ground to doubt this, and says: "The future will prove this view through observation."

Exner, 1872, alone disagreed. In his work on amphibia, birds and mammals, he could find all intermediate stages between epithelial and olfactory cells; the epithelial cells had all the characters attributed to olfactory cells. He believed that the olfactory nerve fibers reached the superficial connective tissue and terminated in a special greatly reticulated layer,' the subepithelial network. From this network pass two kinds of fibers, one of epithelial cells, one of olfactory cells. This network forms, with the two kinds of cells, the terminal apparatus of the olfactory nerve. Exner says : "It would be difficult to say whether all parts of this apparatus serve in the same degree in the olfactory perception."

Cisoff, 1874- By use of isolation methods, Cisoff claims to have seen the nerve cell with a long central varicose process, and also to have seen the connection of these cells with nerve bundles. His work, however, seems not to be credited.

von Brunn, 1875, worked on cat, dog, rabbit and sheep. He found the epithelial cells and the olfactory cells. The olfactory cells were pear-shaped with a round nucleus. Beneath the epithelium the central process joined with other processes to form a network in which stellate cells were found. The olfactory nerves were broken up in the same manner in the upper part of the mucous membrane. He did not see the direct connection of these fibers with the central processes of the olfactory cells, and says, "I can only declare such a connection as possible." Both these and the retinal cells are, according to von Brunn, bipolar sense cells with similar function. For mammals, he describes a membrana limitans olfactoria, which covers the epithelial cells as a whole. The peripheral processes of the olfactory cells project through pore-like openings in the membrane.

In 1880 von Brunn modifies his views concerning this membrana limitans, and thinks it lies underneath the "rudimentary cilia" of the epithelial cells.

Ehrlich, 1886, by methylene blue established with certainty the direct connection of the olfactory fibers and the bipolar cells of the mucosa. This stain is very transitory and lasts only a short time, so the work was not credited until confirmed by the Golgi method.


22 Effie A. Eead

Arnstein, 1887, confirms Elirlich's work. He saw the olfactory cells with the central thread-like processes passing into the nerve bundle of the submucosa. He claims also to have seen the same thing in the gold chloride preparations of Cisoff and in the isolated osmium preparations of Dogiel. He, like Ehrlich, used methylene blue.

Ranvier, 1889, found three kinds of cells in batrachians. His general descriptions of these do not differ from those of other investigators. In the frog, salamander, triton, dog and rabbit, Eanvier found a plexus formed from the olfactory fibrils. The central prolongations of the olfactory cells appeared to connect with this plexus. Eanvier claims that the subepithelial plexus described by Exner was above the basal membrane, while the one he found was beyond the basal membrane and hence in the connective tissue. Eanvier does not believe that the fibrils of the olfactory nerve continue directly with the central prolongations of the olfactory cells. He adds that all histologists who pretend to have seen this are victims of a delusion.

Grassi and Castronovo, 1889, worked on dogs from two to six years old. They demonstrated by the Golgi method an olfactory cell with the peripheral process and with the central end connected with a varicose nerve fiber. This fiber is shown as dividing and subdividing in the connective tissue. In one figure two neighboring cells joined. They were undecided whether the supporting cells were such or whether they were also connected with the nerve, but they state "the connection of these cells with the nerve fiber has never been seen," nor have they seen a connection between this supporting cell and the olfactory cell. In the "limiting zone," at the boundary between the respiratory and olfactory epithelium, they find many varicose nerve fibers which are described as ramifying in the deeper and middle layers of the epithelium. From the many horizontal branches there are some which pass up close to the surface of the epithelium and some which end in the cylindrical olfactory cells. The former may end free, but "this is still not determined." They consider it probable that these fibers are olfactory fibers, Ijut they cannot prove it.

They also describe for the cylindrical cells of this zone a varicose central process which appears like the nerve fiber. In some cases these have unmistakable signs of nerve fibers, and one figure shows cylindrical cells joined by these branching processes.

Van GehucMen, 1890, by his work on the rabbit, has confirmed Cajal's work, and says that Cajal's figures are an almost exact representation


Olfactory Apparatus in Dog, Cat and Man 33

of his preparations. The olfactory fibers unite into thick bundles, but, according to Van Gehuchten, the individual fibers do not vary in size during their entire length.

They are rarely varicose; the varicosity is probably due to an incomplete reduction of the stain. At the base of the epithelium the fibers nuiy turn abruptly or may pass to the olfactory cell directly in a more or less undulating course. The olfactory cell is bipolar, its peripheral end is the longer and reaches the free surface, in some cases where there is no deposit of silver, ending by a cilia-like projection as described by Eanvier for the frog. The central process may be followed for some distance in the connective tissue. Van Gehuchten concludes thus: by methylene blue and by the Golgi method it has been proven that there is a direct continuity of the olfactory fiber and the bipolar cell. There is no plexus, as thought by Exner and Eanvier, no free intraepithelial terminations, nor a connection of the nerve fibers with the cylindrical cells in the limiting zone, as described by Grassi and Castronovo.

von Brunn, 1892, finds the membrana limitans and the olfactory hairs; these are on a bud-like swelling of the olfactory cell. He is not certain whether or not the enlargements are due to reagents. The olfactory hairs come out of holes in the membrana limitans; this limitans is comparable to the homogeneous border which is penetrated by cilia or ciliated cells, and he considers it comparable to a cuticular border.

He has seen the nerves join the olfactory cells and seen them join with other threads, but has never seen free endings which were olfactory fibers. He has seen fibers on the border of the olfactory and ciliated epithelium which pass up into the epithelium, but these did not join with any cell and were therefore free ending fibers.

Betzius, 1892, worked on mouse, cat, dog and rabbit, using the rapid Golgi method. He found two kinds of cells: the supporting cells and the olfactory cells. The supporting cells had a nucleus in the outer third of the cell body. The inner part of the cells had two, three or more wing-like processes which reached to the inner surface of the epithelium. These did not form a fiber. Between these supporting cells were found the olfactory cells. They were bipolar; the cell body was oval or spindle shape with two processes. The outer, thicker process passed to the surface between the supporting cells and bore cilia-like hairs. The inner one was much finer and often varicose. There were several layers of cells, so that the processes were of varying


24: Effie A. Head

lengths. The central process had a straight or undulating course. It often passed just under the epithelial layer for some distance and then entered the mucosa to join the olfactory bundle which passed through the foramina of the cribriform plate to the olfactory bulb. The fiber remained often the same width from the olfactory cell to the olfactory bulb; it did not anastomose or divide, at least not before its entrance into the olfactory bulb.

In mouse at the transition point between respiratory and olfactory epithelium, Eetzius has seen free nerve endings reaching nearly to the surface of the epithelium. He describes them as very fine and varicose, only here and there were small end knots seen, and these did not differ from the varicosities found on the nerve fiber and were not true end knots. He suspects that they are the endings of the 5th nerve, l)ut is not willing to give this verdict.

Cajal, ISOJf, speaks of his results thus : Our observations prove not only the continuation of a fiber of the olfactory nerves with a bipolar cell of the mucosa, but also the unity and independence of this fiber in all its course as far as the bulb, where it ends by means of a free arborization. The network and the ramification described in the intra or extra epithelial course of these nerves he has not confirmed by the new methods of coloration.

Morrill, 1898, investigated the olfactory organ of dog-fish, using Ehrlich's method. He found continuity of the nerve fiber and cell, and also found free nerve endings. He describes three types of olfactory cells, cylindrical, spindle-shaped and conical; whether the difference in shape is due to function or to mechanical causes has not been determined.

Wiih Reference to the Gross Anatomy of the Olfactory Nerves. — Up to a comparatively short period the olfactory tracts were called olfactory nerves; and further, in speaking of the filaments in the nasal mucosa it was always assumed that they extended from the olfactory bulb. In the newer literature, the nerves are described as extending fi-om tlie olfactory epithelium to the olfactory bulli. They are so considered in this paper.

Relation of the Olfactory Fibers and Bundles to the Olfactory Mucosa. — In the newest and most reliable works on anatomy of tlie present time the authors describe, in their explanations of the olfactory regions of man both for the nasal septum and lateral wall of the nose, a plexus of the large nerve bundles before they pass through the cribriform plate of the ethmoid bone. In many cases the figures of Leveille


Olfactory Apparatus in Dog, Cat and Man 35

have been used to represent this condition. In eases where the figures used are original, they lack distinctness, which is, no doubt, due to the uncertainty of knowledge of this region.

The idea of a plexus of the nerve bundles is of earlier origin than the work of Leveille, as will be shown by Figs. 10 and 13. These are copies of the figures of Scarpa, 1785. The ideas of Leveille did not differ esentially from those of Scarpa, and were, no doubt, strongly influenced by them. Our knowledge at the present day concerning the plexiform arrangement of the olfactory nerve bundles is practically that of Scarpa. Much credit is due Scarpa for his excellent work, the facts of which have formed a basis for the knowledge of that region to the present time, as will be seen by the following resume of a part of the second book of his Anatomicarum Annotationum.

A series of nerve bundles varying in number with the subject come from the apex of the bulb. These, covered by the meninges, pass through the foramina of the cribriform plate and are spread out as nerves of olfaction within the nose. The principal branches are arranged in an internal and an external series. The internal send out filaments to the nasal septum. When the nasal membrane is turned back from the septum it is found to be filled with filaments of nerves running doAvn in series. They difi^er in length, some often so long as to reach the lowest base of the septimi and almost touch the floor of the nasal cavity. Others descend only half way. Some pass perpendicularly, while others are arched, as the posterior ones (Fig. 10).

The external series is distributed far and wide through the upper turbinal bones. The longest branches reach from the upper nares to the lowest edge of the middle turbinated bones. These are perpendicular at first and then recurved to the posterior. The posterior ones are arched (Fig. 12). These nerves in their course from the cribriform plate to the pituitary membrane form anastomosing plexiform connections. The plexiform nerve bundles are found in canals of the turbinated bones of the nose, as is admirably shown in the figures of Scarpa (Figs. 11, 12). Not many olfactory filaments go to the lower turbinals, and he questions whether they are of much importance.

There are no olfactory nerves to the membranes of the pituitary sinuses, and hence these are not olfactory in function.

His descriptions of the 5th nerve to the nose are practically those of to-day.

The following are Scarpa's own words concerning the olfactory nerves :


26 Effie A. Eead

"Eami porro isti copiosiora mox emittunt filamenta, quorum magna pars nudo oculo conspicua, inter membranam pituitariam, & periosteum septi narium a summo ad imum septum decurrunt. . . . Maiores vero rami non intermisso per cribriformem laminam itinera continues canaliculos superiorum turbinatorum ingrediuntur, intra quos iterum, ac saepe divisi, & ramosi porro pergunt late per turbinata ossa superiora distribuendi. Quo in itinere, utpote canaliculorum quaniplures communicationem inter se alunt, crebrisque orificiis ad narium cavitatem hiant; ita nervorum, de quibus loquimur, rami intra hos canaliculos adhuc reconditi anastomosim, & plexuosas copulationes (t) inter descendendum in vicem constituunt, frequentesque propagines extus per patula canaliculorum orificia membranae pituitariae turbinata ossa superiora vestienti largiuntur. . . . ]\Iedio modo se liabent, qui per median! turbinatorum superiorum regionem f eruntur : nempe quo ad numerum, crassitiem, & incessus rationem; in eo autem discrepant, quod omnium huius provinciae longissimi sunt (x) quippe a summis naribus ad imam usque oram turbinati medii pertingunt. . . . Sed neque ad turbinatum inferius paris primi filamenta deduci plura sunt, quae sin minus suadent, saltern dubitationi locum praebent vehementer. . . . Keque enim ad organi olfactus sedem adscribendi sunt finus pituitarii, quoniam olfactilis nervus membranae eas caveas vestienti filamenta nullatenus tribuit."

For Comparative Anatomy the statements of Milne-Edwards, Clieveau and Owen agree very closely with those in the works on human anatomy.

The following is a generalized statement by Owen:

"The nerves are grouped in all Mammals into a set for the septum and a second for the upper or ethmo-turbinals, a third or middle short set being, in some, distinguished for the labyrinth or roof of the nasal chamber. The branches of the second set, after expanding on the ethmo-turbinals, usually converge to become connected with the lateral nasal branch of the 'fifth.' Their mode of distribution is best seen on the ethmo-turbinal : here they divide, subdivide, expand and anastomose with each other, forming a reticular nervous expanse, with long and narrow meshes, and becoming impacted in the central, or inner, layer of the olfactory membrane."

For the true relation of these nerve bundles see the body of this paper, page 33 and Figs. 24-27.


Olfactory Apparatus in Dog, Cat and Man 37


Methods for Gross Dissection.

The nitric acid method was used for gross dissection. The head was placed in 20 per cent nitric acid for 6-12 hours, depending upon the size; the decalcification had then proceeded so far that the bone could be easily cut.

The bones were removed from the nose and orbit, thus exposing the olfactory bulb, the nasal mucosa and the lining of the maxillary, frontal and sphenoidal sinuses.

As the bone was removed from the mucosa the deepest or attached surface of mucosa was exposed (Figs. 5, 6). It is this surface which must be exposed to view them.

In dog and cat the ethmo-turbinal bones were easily removed, as they are not perforated by the nerve bundles. In man, however, this is not so easily accomplished. The turbinated bones are filled with small canals through which the nerve bundles pass (Fig. 11). There is, therefore, an interweaving of bone and nerve. Much care is necessary to free these bundles without injury. If the specimen is favorable there is a marked contrast between the white nerves and the darker mucosa. This differentiation is destroyed if the material is left too long in nitric acid. The olfactory nerves are very prominent and are spread out in a fanshaped manner upon the olfactory folds. They stand out with remarkable sharpness as white cords against the darker background of the nasal mucosa (Figs. 5, 6). This is also true of the branches of the 5th nerve which innervate the nose. It is this differentiation and the fact that the nerves lie in the deeper layers of the mucosa next to the bone which made this dissection of the fine terminal branches of the 5th nerve possible. Even under these favorable circumstances it was necessary to dissect under water and in brilliant light (sunlight or electric light) with a magnifier giving 8-13 diameters.

Material prepared by the nitric acid method may be preserved during the dissection in 3 per cent formalin without markedly changing color. This does not hinder the dissection and material will not deteriorate in it. Five per cent formalin is recommended for permanent preservation.

If material preserved in formalin is used, further decalcification may not be necessaiy. There will, however, be no differentiation in color between the mucosa and the nerves, and the material, therefore, does not give as satisfactory results.


28 Effie A. Eead


Histological Methods.


Four methods have been used : the rapid Golgi, the mixed Golgi, gold chloride, methylene blue and dissociation methods.

The Rapid Golgi Method. — Fresh tissue was put into osmium-bichromate mixture for 3-4 days and kept in the dark.

3 per cent potassium bichromate 2 parts.

1 per cent osmic acid 1 part.

This was changed at least once. The material was then placed in % per cent silver nitrate for 3 to 4 days, being changed several times in the first half hour until no precipitate formed. Dehydration was as rapid as possible, li/o per cent, 3 per cent and 8 per cent collodion was used for infiltration. Tissue was left in 8 per cent collodion i/o day without harm and was imbedded in 8 per cent collodion. It was hardened in chloroform vapor for 2 to 12 hours. The knife and block were fiooded with 95 per cent alcohol during the cutting; sections were 60 to 80 microns.

The results were very good both in the dog and in the cat. Olfactory cells, with their axones, peripheral processes and the olfactory hairs, could be seen. Sensory cells were found in the vomeronasal organ of the cat. In man the results were less satisfactory, due to the lack of fresh material, but positive verification was obtained.

The Mixed Golgi Method. — Good results were obtained in the dog and the mouse from the mixed Golgi method. (The tissue was treated as for the rapid method, except that it had been previously fixed in Miiller's fluid.) The nasal conchge and the septum of this dog were still cartilagenous, so it was possible to make sections through the entire nose and olfactory Imlb. Nerves could be traced for a long distance even through the cribriform plate to the olfactory bulb. Olfactory cells were obtained and also sensory cells in the vomeronasal organ of the mouse.

The Gold Chloride Method. — Both Eanvier's formic acid method and Hardesty's modification of the gold chloride method were used. The difficulty in the use of the former method is due to the fact that the epithelium is very easily exfoliated in fresh material. Good results, however, were obtained from human material by this method. Hardesty's modification of the method- gave good results with dog and cat. The dog material had been in 10 per cent formalin for eight years, the cat only a few weeks. Sections were made from 1 to 20 microns. The sustentacular cells were stained as well as the olfactory cells; in fact, the

^Hardesty, Neurological Technique.


Olfactory Apparatus in Dog, Cat and Man 29

whole mucosa was stained. The thicker sections proved valueless for the olfactory cell. Sections 1 to 3 microns showed the olfactory cells and in some cases a very small part of the axone. Its course is undulating and can be followed only in thick sections. The peripheral process was easily found.

The Methylene Blue Method. — Huber's modification method was used.' Olfactory cells with their two processes were found in dog and cat. The same difficulty was encountered here as with the gold chloride material, much of the epithelium had been exfoliated.

Dissociation Method. — The gold chloride material and fresh tissue were placed in formaldehyde dissociator (2 cc. formaldehyde and 1 liter of normal salt solution) for forty minutes. Olfactory cells with their two processes were obtained in dog and cat.

Gross Anatomy of the Nose.*

The cavity of the nose (cavum nasi) is divided into two lateral halves by the nasal septum (septum nasi) (Figs. 15-23). This septum is formed of two parts, the septum cartilagineum or cephalic part and the septum nasi osseum which joins the cribriform plate (Lamina cribrosa). In the dog and cat the septum is extended dorsally by the median parts of the os frontale and os nasale.

In this paper the term septum does not include this area. When referred to, it is designated as the turbinated part of the septum.

The lateral halves of the nose consist of the turbinated bones (conchffi nasales) (Figs. 15-23). In the dog and cat these conchge may be divided into two parts. The ethmo-turbinals (Figs. 16, 17, 19, 20) and the maxillo-turbinals or concha nasalis inferior (Figs. 15, 18). The ethmo-turbinals are thin plicated bones which are attached to the cribriform plate. In the dog these extend about ^/^ and in the cat about Y2 the length of the nose. Figs. 1 and 3 show the mucosa of the ethmo-turbinals (mucosa nasi), but the bones have been removed.

The maxillo-turbinal is also a plicated bone situated cephalad of the ethmo-turbinals. This is a larger bone and much more plicated in dog than in cat.

In man the condition is much different. There are three turbinal bones (concliEe nasales), concha nasalis superior, media and inferior

^Journal of Applied Microscopy, April, 1898, p. 64. The Methylene Blue Method for staining Nerve Tissue, G. Carl Huber.

The B. N. A. terms are introduced as far as possible.


30 Effie A. Eead

(Fig. 23). These are plate-like bones and are roughened and perforated (Fig. 11), but not plicated, as in dog and cat (Figs. 17, 20).

The superior turbinated bone is attached to the cribriform plate and is more or less united to the median one which lies ventrad to it (Figs. 11, 23). The inferior is just dorsad of the palate (Figs. 11, 23). The extent of the turbinated bones is relatively much less in man.

The nasal cavity is divided into three regions according to the nature of the epithelial lining.

The vestibule or cephalic part of the nose is lined with stratified epithelium which is continuous with the epidermis. In the respiratory region (regio respiratoria) the epithelium is replaced by the columnar ciliated type (Fig. 43),

The olfactory region (regio olfactoria), with which this paper deals, is adjacent to the cribriform plate. In fresh material the mucosa is slightly yellow, due to the pigment in the sustentacular cells. The extent of this area is relatively much greater in dog and cat than in man. In dog and cat it comprises about I/2 of the numerous ethmoturbinals (Figs. 1, 3, 5), and from ^/g to I/2 of the nasal septum (Figs. 2, 4, 7, 8). With reference to the three sinuses opening into the nasal cavity, viz., the sphenoidal, the maxillary and the frontal, only branches of the 5th nerve could be traced to the mucosa of the sphenoidal and maxillary. This is in agreement with previous workers. In works on human anatomy (Quain, Piersol) only the 5th is given as innervating the mucosa of the frontal sinus.

In the dog and cat there is one scroll (Jayne) of the ethmo-turbinal extending for a short distance into the funnel-like opening of the frontal sinus. This may be in the form of a somewhat curved leaf, the free margin dividing the funnel-like outlet in part or the scroll may be rolled up more completely so that the free end in the frontal sinus is curved and looks like the open mouth of a snail shell. Olfactory nerves ramify in this scroll. In the dog they extend also for some distance into the mesal mucosa covering the bony wall of the sinus opposite the cribriform plate. In the cat the scroll-like projection is more lateral and the mucosa lining the sinus opposite the orbit has the greater number of olfactory nerves. That is, in the dog the olfactory nerves of the mouth of the frontal sinus are toward the middle line, while those in the cat are lateral in position. The brown coloration of the epithelium in the olfactory part of the sinus is marked. From the position of the olfactory nerves in the cephalic part of the sinus and its


Olfactory x4.pparatiis in Dog, Cat and Man 31

opening into the nose^ any movement of the air back and forth through the narrow outlet would be likely to bring the odorous particles in contact with the olfactory epithelium. There is a variation of opinion concerning the extent of the olfactory area in man. According to Scarpa, this is very extended. It includes the entire area of the upper turbinated zones (a few filaments going to the inferior turbinal). Some of the nerves of the septum are pictured as reaching the floor of the nasal cavity (Fig. 10). Sappey's pictures show a less extended distribution of these nerves a little less than lA of the septum; the superior i^ of the middle turbinated bones. According to von Brunn only a small portion of the superior turbinals and a corresponding area of the septum are olfactory in function (Figs. 28, 29). My results are midway between those of Sappey and von Brunn. Figs. 30, 31 were made from dissections and show that the olfactory nerves reach nearly to the free edge of the superior turbinated bone and about % the width of the lateral wall and occupy about ^/g of the septum.

Tpie Histological Structure of the Olfactory Epithelium.

The epithelium of the olfactory region consists of three kinds of cells : the supporting or sustentacula! cells, the olfactory cells and the small stellate basal cells (Figs. 39, 42). In the sub mucosa serous glands are found; these are known as Bowman's glands and are well pictured in all the books. The ducts of these glands are stained by the Golgi method and pictured in Fig. 41.

The supporting cells are elongated and cylindrical; they have an oval nucleus and a thin cuticular border (Figs. 40, 42). The central end has wing-like processes, often irregular in outline, which project toward the basement meml)rane between the olfactory cells. This cell was distinguished from the olfactory cell by Eckhard in 1855; but he and other early writers were doubtful as to its true nature. These cells occupy the superficial border of the epithelium and contain pigment. Stellate cells lie near the basement membrane among the processes of the epithelial cells.

The olfactory cells have been studied by four different methods : the Golgi, gold chloride, methylene blue, and dissociation (Figs. 32-38). In all, their position appeared the same. They lie in the middle and deeper layers of the epithelium and send their process between the supporting cells. They are fusiform in shape, with a spherical nucleus


33 Effie A. Eead

in the central end. The peripheral process is often irregular and reaches the surface of the epithelium. Its outer edge is bulbous and has numerous cilia-like appendages, the olfactory hairs (Fig. 34). These extend beyond the outer border of the epithelium, free in the nasal cavity. The central process is the axone or olfactory nerve fiber. It is very fine and extends in an undulating course into the underlying connective tissue. Only in thick sections could this be followed. These were best seen in the Golgi preparations and in methylene blue material (Figs. 32-34). The sections of gold chloride material showed the axone for a slight distance (Figs. 35, 36). In the dissociated material the axone was generally broken off, but in some preparations axones were found (Figs. 37, 38).

My work agrees with the results of Van Gehuchten as to the shape of these olfactory cells. He believes the varicosities are due to imperfect impregnation. I found both varicose fibers and those which were uniform in outline.

With Max Schultze, I consider these the true olfactory cells. The peripheral process bears the olfactory hairs. The central process is the axone. Early writers described a network for these olfactory axones directly beneath the epithelium as they enter the connective tissue. Eecent work has disproven this, and it is now believed that the axone or olfactory fiber "keeps its unity and independence from the olfactory cell to the olfactory bulb," branching only when it reaches the glomerulus of the bulb. In none of my work was the branching or anastomosis of an olfactory fiber seen except at this place. Upon reaching the deepest layers of the connective tissue next to the bone these axones or fibers collect into bundles of various sizes and as olfactory nerve bundles extend to and pass through the cribriform plate to the olfactory bulb.

As has been stated in an earlier part of this paper, almost all authors describe a nerve plexus for these olfactory bundles. This has nothing to do with the network just mentioned, as it concerns only the large nerve trunks and not the individual axones. From the time of Scarpa, 1785, to that of Barker, 1904, and Quain, 1906, the olfactory bundles are pictured and described as forming a plexus on the septum and lateral wall of the nose of man.

If the bone is removed from the orbit and side of the nose (Figs. 13, 14), there is certainly a plexiform appearance of the tissue in which the nerve bundles extend. With a consideration of the gross specimen


Olfactory Apparatus in Dog, Cat and ^lan 33

only, Scarpa and subsequent authors were justitied in their conclusions that the nerve bundles form a plexus in this region. But upon a microscopical examination after diiferential staining, it is found that the nerve bundles do not anastomose.

This plexiform appearance is due, not to a joining of nerve bundles, but rather to the ramification of the blood vessels and to the arrangement and abundance of the connective tissue which surrounds these vessels and nerves (Figs. 24, 27). The nerves have been traced in these cords of connective tissue. As shown in the drawing and photograph, the}^ pass almost vertically through this to the foramina of the cribriform plate without anastomosis or the formation of a plexus (Figs. 24, 27).

There is but little appearance of a plexus upon the nasal septum (Figs. 25, 26), and the picture of Scarpa (Fig. 10) is much more accurate than are those of Leveille. Figs. 13, 14, 24 to 27 show strikingly that there is a marked difference in the plexiform appearance of the lateral wall and septum. In both cases, especially upon the septum, there is a crossing and recrossing of nerves, but focusing shows that these do not join. There is, however, some slight joining of the smaller nerve bundles near their origin (Figs. 25, 26).

All authorities on comparative anatomy, wherever the subject is discussed, speak of a plexus of the olfactory bundles. But there is no such marked appearance of this in dog and cat as that found in man. It looks as if the conditions in man had been interpreted for mammals without adequate investigation. Whenever there is an appearance of a plexus, it has been found to be merely a crossing of nerve bundles.

The Olfactory Bulb.

The olfactory bulb has been descriljed Ijy various workers as consisting of from two to seven layers, according to the subdivisions made l)y these investigators.

Golgi, 1875, describes three layers, olfactory fibers, mitral cells and nerve bundles of the olfactory tract. Van Gehuchten and Martin, 1891, also describe three main layers. In this paper we are concerned only with the olfactory fibers, the glomeruli and mitral cells ; we will not enter into the discussion beyond this.

Van Gehuchten and Martin, 1891, worked on the dog and the cat, both adult and young animals, also the rabbit, rat, and mouse. The


34 Effie A. Bead

rapid Golgi method was used, with results as follows: The olfactory fibrils collect into bundles which go to the glomeruli; these fibers form the outermost layer of the bulb.

Betzius, 1S92, says that the nerve fibers divide either at a short distance from the glomerulus or oftener near it. After a repeated and profuse dichotomous branching the fibers weave through the glomerulus, but do not form a network.

Van Gehucliten and Martin, 1891, have seen these fibers bifurcate in the cat and form fibrils of equal thickness, which pass to a single glomerulus, or each may pass to a different glomerulus. Some fibers bifurcate more than once. Thus a single olfactory cell would be connected with two or more glomeruli. "This bifurcation cannot be said to be constant but it is frequent."

The olfactory glomerulus is formed by an interlacing of the terminations of the olfactory cells and the dendrites of the mitral cells. These are independent of each other, that is, there is no anastomosis as was thought by Golgi, 1875. Olfactory fibrils were free in the glomerulus of the cat, the dog, the rabbit, the rat and the mouse, and a number of olfactory fibrils go to each glomerulus.

In the dog they believe the glomeruli to receive dendrites from a great number of mitral cells. In all mammals studied, each mitral cell is connected with a great numl)er of bipolar cells, but each olfactory cell of the mucosa is connected with one, rarely two, mitral cells; at the glomerulus each olfactory fibril terminates generally with only one mitral cell.

In all animals where the olfactory sense is greatest, each bipolar cell may be in contact with several mitral cells, not because the fiber bifurcates and goes to different cells, but because in the same glomerulus mav be found the dendrites of several mitral cells.


Personal Observations.

The following are the results which Avere obtained from the olfactory bulb of the dog and cat. The olfactory bullj Avas studied in gross preparations and in sections; in the gross dissection the olfactory nerves were traced from the mucosa through the foramina of the cribriform plate to the olfactory bulb. They could be plainly seen lying irregularlv upon the bulb (Fig. 4). This was also seen in the transections and saaittal sections of the olfactorv bull) and mucosa. Individual


Olfactory Apparatus in Dog, Cat and Man 35

fibers could be traced for a considerable distance, and in some cases fibers were traced nearly through the cribriform plate. The nerves were not seen to bifurcate in the layer as described by Van Gehuchten and Martin, but remained as individual fibers until near the glomerulus. At their entrance into the glomerulus they divide and subdivide to form many branches which interlace but do not anastomose with the other fibers found there. In some cases four or five of these axones were traced into the same glomerulus (Fig. 47).

The glomerulus of the olfactory bulb is formed by the interlacing of branches from the axones of olfactory cells and the dendrites of the mitral cells of the olfactory bulb (Figs. 48-53). (For clearness these have been shown in separate drawings, that is, axones of nerve cells and dendrites of mitral cells are not shown in the same figure.) A glomerulus may be formed by the interlacing fibers from one axone (Fig. 46), and from one dendrite (Fig. 50), or from several axones (Fig. 47), and several dendrites (Fig. 52). While each axone comes from an individual olfactory cell, the dendrites may come from a single or several mitral cells.

In the cat three dendrites from difl^erent mitral cells were found in one glomerulus (Fig. 53). Fig. 51 shows .three dendrites from at least two difi^erent cells. Fig. 53, two dendrites from the same mitral cell. The branching of a single dendrite to different glomeruli was not seen.

In the dog, dendrites from several different mitral cells were traced to a glomerulus (Fig. 49), and a single dendrite was seen to branch to three different glomeruli (Fig. 48).

In man the olfactory bulb has been studied only in gross preparations. The olfactory nerves were traced through the cribriform plate to the outer layer of the Inilli. The histology of the olfactory bulb was not studied, but is given by all authors. The glomeruli are formed by an interlacing of the axones of the olfactory cells and the dendrites of the mitral cells, as in lower animals.


DlSTEIBUTION OF THE OTH ISTeRVE TO THE NoSE.

The nose is innervated by branches of two divisions of the 5th nerve. The anterior ethmoidal (nervus ethmoidalis) of the ophthalmic aiid the spheno-palatine (nervii spheno-palatini) of the maxillary division.

In the orbit the anterior ethmoidal nerve passes between the muscles of the eve and enters the cranial cavity through the anterior ethmoidal


36 Effie A. Eead

ioranien (foramen etlimoidale) into the cranial cavity. It passes along the olfactory bulb (bulbus olfactorius) (Fig. 3) cephalad through an opening on the cribriform plate and passes along the upper part of the nasal septum (septum nasi) (Figs. 2, 4, 7, 13, 14), where it divides into the external nasal nerve (nervus nasalis externus) and the internal nasal nerves (nervii nasales internii). The external nasal nerve passes along the sulcus ethmoidalis of the nasal bone (os nasale) and passes out to innervate the skin of the nose (Figs. 13, 14). The internal nasal nerve divides into the median nasal (ramus nasalis medialis), which supplies the septum, and the lateral nasal nerve (ramus nasalis lateralis), which innervates the mucosa of the lateral wall.

The remaining part of tlie mucosa is innervated by the sphenopalatine nerves (Figs. 1, 3). The naso-palatine branch of this nerve (n. palatinus) was traced along the septum to the canal of the incisors (canalis incisivus). It sends several branches into the middle of the septum (Figs. 2, 4). In dog and cat this .was traced into the vomeronasal organ (Figs. 2, 4). This nerve was also dissected in man and was traced almost into the organ. The terminal branches were so fine that their complete dissection was not successful.

Free Terminations of the 5th Nerve Within the Nasal Mucosa.

von Brunn, 1892, saw free nerve terminations within the nasal epithelium at the border of the respiratory region. According to him these fibers could not be the olfactory axones, as they were much thicker than those. He, therefore, concludes that they are the endings of the Trigeminus. He quotes Cajal as supporting his decision.

von Lenliossek, 1892, has seen the fibers described by von Brunn, but instead of being thick, as described by that author, those seen by him are finer than the olfactory fibers, varicose and with terminal endings; these did not always reach the free surface of the epithelium. The nerves which are pictured and described by von Lenhossek are like those pictured by Cajal and not of the ordinary much branched appearance of a sensory nerve in epithelium, von Lenhossek did not commit himself as to the origin of these fibers.

RetziuSj, 1892, pictures in the nasal epithelium of the mouse and cat, both in the respiratory and olfactory regions, fine, much-branched nerve fibers, which end free in the epithelium like other sensory nerves. These are varicose, but not always with an end Ivnot. Eetzius wishes to confirm


Olfactory Apparatus in Dog, Cat and Man 37

the appearance of these' nerve fibers within tlie nasal epithelium, but does not wish to give his verdict as to their origin, he adds that it is plausible that these are of a sensory nature. In his work on Fishes he does not find any structures comparable with the "Geruchsknospen" of Blaue. Eetzius considers as false the theory of Blaue that there are such structures which have sense cells in direct connection with the olfactory nerve.

Cajal, 189Jf, in his Systeme Nerveux denies having committed himself upon the character of these nerves, but ascribes their discovery to von Brunn. According to his work, the endings of the 5th nerve are found only in the submucosa and do not extend into the epithelium. He finds in man fibers which end free at the surface of the epithelium, but these are nearly vertical and end in a conical projection at the top, as is shown by von Lenhossek. He withholds his verdict as to the origin of the fibers thus ending until work then in progress was complete. He has seen them only in the embryo, but never in new-born animals or those several days old.

Disse, 1896, found in the nasal mucosa of some mammals "Epithelknospen" which resemble the taste buds in appearance. These buds are of two kinds, the large buds in the olfactory epithelium and the small buds in the respiratory epithelium. These consist of supporting cells and sense cells^ (the sense cells are not ganglion cells). By the Golgi method he traced nerve fibers into the large buds. He considers these fibers as belonging to the 5th nerve. Disse does not credit Blaue's theory that these buds are in connection with the olfactory nerves, but thinks that they have to do especially with the sweet and sour sense of taste in the nose.

Kallius, 1905, has seen the free endings of the 5th nerve in the respiratory and olfactory epithelium of calf. He finds nothing in his preparations, except possibly nests of mucous cells, which in any way resemble the "Epithelknospen" of Disse, nor have any such structures been found in the nasal epithelium of man.

Personal Observations.

I have seen in the nasal epithelium of the kitten a few days old, both in the respiratory and olfactory regions, many much-branched nerve fibers. These were varicose and often ended with a varicosity (Figs. 44, 45). From the gross dissection, fibers from the 5th nerve


38 Effie A. Eead

pass to the olfactory folds (Fig. 3), and to the lateral wall and septum (Figs. 1-4). It would, therefore, seem probable that the nerves described above are the free terminations of the 5th nerve. My preparations agree very closely in appearance with those of Eetzius for the mouse and cat. I find no structure in the nasal epithelium of dog, cat or man which resembles the "Geruchsknospen" of Blaue or the "Epithelknospen" described by Disse.

Organon Vomeronasale.

This organ has been the subject of various investigations; a detailed account is given by Kolliker, 1877, 1883, and Harvey, 1883 ; von Brunn, 1892; von Lenhossek, 1893; Merkel, 1893; Mihalkovics, 1898. Klein worked on the guinea pig, the rabbit and the dog; von Lenhossek on the rabbit ; Harvey on the mouse and the cat ; von Brunn on the sheep ; Kolliker, Merkel and Mihalkovics on man.

The gross structure, briefly stated, is as follows :

The vomeronasal organ of the dog and the cat is a bilateral tubular organ situated in the ventral part of the septum in the region of the pre-maxillary and maxillary bones. It is either entirely or partially surrounded by a capsule of hyaline cartilage (Figs. 15, 18, 54, 55). At the cephalic end of the nose there are two prominent folds on each side of the nasal septum. The dorsal one is due to a solid fold of the mucosa and to the presence of glands. This is the smaller and passes dorsad of the incisors. The cartilaginous capsule is complete in the cephalic part of the vomeronasal organ of the cat. In the remaining portion in the cat and through its entire extent in the dog this capsule is only partial. As stated above, the vomeronasal organ is tubular and is flattened laterally. It is blind at the caudal end, but opens at the cephalic end into the ductus nasopalatinus. In man the vomeronasal organ is much less developed than in dog and cat. It is a bilateral organ situated in the mucous membrane of the ventral part of the nasal septum (Fig. 31). It is a short blind tube only a few millimeters in length which opens anteriorly into the nasal cavity by a small pore-like opening just above the incisors. This opening was seen both in child and adult. The cartilage of this organ is much reduced and lies entirely below the organ (Fig. 21). The shape of the tube in the dog and the cat varies in the different regions ; near the cephalic opening it is circular in transection and lined with stratified epithelium; in the


Olfactory Apparatus in Dog, Cat and Man 39

median and caudal parts it is kidne3'-shaped and the epithelium is columnar. The median and lateral epithelia differ in thiclaiess; the median is sometimes two or three times thicker than the lateral. In the human which were examined the vomeronasal organ of a 6-7 weeks embryo was flattened as descril)ed for the dog and the cat and man, and the epithelium of the median wall was the thicker. In a four months human fetus it seemed to be circular in outline for its entire length, with a uniform thickness of epithelium.

The epithelium, like that of the nasal cavity, consists of sustentacular cells and are longer and narrower than those of the nasal mucosa; the sensory cells found in cat had a process which passed to the surface of the ej)ithelium (Figs. 54a, 55a). These cells have not been found in man, according to Mihalkovics, 1898, and Quain, 1906. According to Klein, the sensory cells are found only in the thick median epithelium, von Lenhossek found olfactory cells in the median and lateral epithelium of an embryo kitten. The central process undivided and imbranched passes into the submucosa as a fine varicose nerve fiber. No olfactory hairs were found by von Lenhossek, 1882, as a precipitate was present. He saw in the deeper layers of the epithelium of the vomeronasal organ free nerve endings. An end knot was always present, but a little rod often projected beyond this; according to him, these were either free endings of the 5tli nerve or of olfactory nerves whose cells were somewhere in the olfactory course.

von Brunn in Golgi preparations of the vomeronasal organ of the sheep saw the connection of the olfactory cell and nerve. He also found olfactory hairs.

Personal Observations.

Gross Anatomy.

The gross anatomy of the organon vomeronasale, or Jacobson's organ, has been carefully worked out. As has been previously stated, the large nerves of the nose lie in the deepest layers of the mucosa next to the bone. In order to see these, it is necessary to remove the bone and thus to expose the back or deepest parts of this mucosa. The nitric acid method described above made this possible. The mucosa was freed from the cartilaginous septum, being careful not to tear the nerves which lie almost on the bone. Figs. 2, 4, 8, 9 show such a dissection. 'In the doo; and the cat the vomeronasal organ was also intimately con


40 Effie A. Eead

nected with the palate, and this was divided in the median suture. The most successful dissection of this organ in those cases was obtained by sawing the entire head in two from front to back, including both the nose and the brain; the entire septum being on one side. The cartilage and the bone were then removed from the mucosa. The vomeronasal organ was found as an elongated flattened fold at the cephalic end of the nose, just above the palate (Figs. 2, 4, 8, 9). Its small cephalic end passes ventrad to the incisors and opens into the ductus naso-palatinus, which leads from the oral to the nasal cavity. In man the position of this organ is somewhat difi^erent (Fig. 21). It is found some distance above the palate and not in intimate relation with it as in the dog and the cat. It opens directly into the nasal cavity. The vomeronasal cartilage is represented by only a small piece of cartilage which lies some distance ventrad of the organ and not enclosing it as in the lower animals (Fig. 21).

I wish to emphasize what has been stated before. It is the deepest layers of the mucosa next to the bone and not the nasal side with which we are at present concerned. There are many nerves in this septal mucosa. These nerves are from two distinct sources : the olfactory nerves, which are connected with the olfactory cells and which can be seen to pass through the foramina of the cribriform plate, and the anterior ethmoidal and spheno-palatine branch of the 5th nerve. The olfactory nerves are found near the cribriform plate (Figs. 2, 4, 7, 9) ; the branches of the 5th nerve innervate the middle and cephalic parts of •the nose (Figs. 2, 4, 7).

There are still several prominent nerves which we have not described (Figs. 2, 4, 8, 9). These are olfactory nerves. They were traced from the olfactory bulb obliquely across the septal mucosa into the vomeronasal organ. In the dog, the cat and man they branch many times just before their distribution in this organ. The vomeronasal organ in dog and cat is also innervated by several branches from the naso-palatins nerve; thus we see that this organ contains nerves from two distinct sources. In man, according to von Kolliker, these olfactory nerves are present only up to the third month of development and atrophy directly after that. Mihalkovics did not find them at all in a threa months human fetus. Long olfactory nerves resembling in every way those of the dog and the cat were seen on the septum of a child. These were traced to the vomeronasal organ. The naso-palatine branches were traced nearly to this organ, but the nerve was so fine that further dissection was not successful.


Olfactory Apparatus in Dog, Cat and Man 41

Histology of the Organon Yomeronasale.

Fig. 18 is a transection of the head of an embryo kitten in the region of Jacobson's organ. This shows the position of the organ, the cartilaginous capsule and the thickness of the epithelium. Figs. 54 and 55 will show the complete and partial capsule in the kitten. The fine structure is described by several investigators. I have been chiefly concerned with the sensory cells. In the kitten (Figs. 54a^ 55a) sense cells were found. These agreed in every way with the olfactory cells of the nasal mucosa. There are two processes : the thicker, peripheral one, and the fine, somewhat varicose, central fiber. The axone was followed for a considerable distance in the submucosa. No olfactory hairs were found, but in Fig. 54a indications of these are seen in the spike-like process.

I have no hesitation in calling these sense cells nerve cells, apparently identical with those of the olfactory mucosa. Free terminations mentioned by von Lenhossek, 1892, and Cajal, 1894, were not found, but we should consider those, from the gross dissection, to be the endings of the 5th nerve, as several branches of this nerve were traced into the organ. I believe, then, with others, that the vomeronasal organ is intimately connected with the olfactory sense.

Eesults. dog and cat.

1. The olfactory nerves are large and numerous in the dog and the cat, relatively less in the cat.

2. About one-half of the ethmo-turbinal folds are olfactory. This is a large distribution as compared with man.

3. All the folds of mucosa adjoining the cribriform plate are olfactory.

4. The mucosa is thick in the olfactory region; thin beyond this; the transition is sharply marked.

5. The mucosa of the septum is in two parts. The upper part is lined by the dorsal turbinated folds; the lower part is lined by a continuation of the mucosa of the cephalic part of the nose. About one-third to onehalf is olfactory.

6. The anterior ethmoidal nerve innervates the olfactory folds and septum ; its branches extend from the cribriform plate to the tip of the nose ; it also innervates the roof and upper lateral Avail of the nose. Small branches pass among the olfactory folds.


42 Effie A. Eead

7. The spheno-palatine nerve innervates the mucosa, cephalad of the ethmo-turbinal folds, the maxillary sinus, the lateral wall of the nose and the maxillo-turbinal folds, also the vomeronasal organ.

8. The vomeronasal organ is a tubular organ found on either side of the septum. It is innervated by olfactory and naso-palatine nerves.

9. The outer layer of the olfactory bulb is formed from the axones of the olfactory cells.

10. The glomeruli of the olfactory bulb are formed by the interlacing of the axones of the olfactory cells and the dendrites of the mitral cells. The number of mitral cells represented in a glomerulus varies in different animals.

MAN.

11. The olfactory nerves are relatively less in number in man than in the dog and cat.

12. They are distril^uted to the upper third of the septum and to nearly the whole of the superior concha (Figs. 30, 31).

13. The nose is innervated by two divisions of the 5th nerve, the anterior ethmoidal which innervates the anterior part of the septum and lateral wall, and a branch is also sent to the skin of the tip of the nose.

14. The spheno-palatine nerve innervates the lateral wall, the conchte and the ventral part of the septum.

15. The vomeronasal organ is much less developed in man than in the lower animals. A branch of the olfactory nerve passes to it, at least at the time of birth.

16. The axones of the olfactory cells form the outer layer of the olfactory bulb.

Conclusions, dog, cat and man.

1. The fusiform cells of the olfactory mucosa are true olfactory cells and true nerve cells. They lie in the deeper parts of the epithelium of the olfactory region.

2. The peripheral process is long and cylindrical and reaches the free surface of the epithelium, passing between the sustentacular cells. It bears the olfactory hairs.

3. The olfactory fiber is the axone of the olfactory cell; these collect to form olfactory nerve bundles and pass through the cribriform plate to end in the glomeruli of the olfactory bull). These nerve bundles do not anastomose to form a plexus.


Olfactory Apparatus in Dog, Cat and Man 43

4. The supporting cells are cylindrical and the inner process is often divided. Stellate cells are found at the base of the supporting cells.

5. The development of the sense of smell in the dog and the cat may be due to the large number of the olfactory nerves and to the extent of their distribution, and, according to Van Gehuchten, to the number of mitral cells with which each olfactory cell is associated.

6. The vomeronasal organ is intimately connected with the sense of smell. It contains, at least in the cat, sensory cells apparently identical with those of the olfactory mucosa.

7. There are free nerve terminations in the olfactory epithelium. These are the endings of the 5th nerve.

In the position of the olfactory nerve cell we apparently have a primitive condition. This is the only case in vertebrates where the nerve cells are within the epithelium, as are those for the tactile sense in many invertebrate forms. In the other organs of sense there is a gradual recession of the ganglion cell imtil, in the ganglion of the dorsal root of the spinal cord, the central nervous system is approximated. The branches for the tactile sense end freely either in special organs (tactile corpuscles) or in the free end-knots within the epithelium, but do not reach to the surface of it; while the branches of the nerves of other sense organs end freely among special cells, but do not reach the free surface of the epithelium. In the olfactory region the olfactory hairs are above the free surface of the epithelium and in direct contact with the air.


BIBLIOGRAPHY.

Papers starred (*) bave not been consulted, as they were not available. References have been made from articles where these papers were commented upon. Aenstein, 1887. Die Methylenblaufarbung als histologische Metbode. Anat.

Anz., Bd. 2. Barker, L. F., 1899. The Nervous System. 1904. Laboratory Manual of Anatomy. Bawden, H. Heath, 1901. A Bibliography of the Literature on the Organ

and Sense of Smell, Anatomy and Histology, 391 references. Journal

of Comparative Nenr^Jogy, Vol. 11. Blaue, Julius, 1884, Untersucbungen liber den Bau der Nasenscbleimbaut

bei Fischen iind Ampbibien, namentlicb iiber Endknospen als End apparate des Nervus olfactorius. Inaugural Dissertation, ArcMv f.

Anat. 11. Physiol., Anat. Abth.


44 Effie A. Eead

BouGERY ET JACOB, 1844. Anatoiiiie tie I'Homme. T. 3 — Atlas. T. 3 — Texte. V. Brunn, 1875. Untersuchungeu liber das Riechepithel. Archiv f. miJcr.

Anat., Bd. 11. 1880. Weltere Uutersuchungen iiber das Riechepithel iind sein Yerhalten

zum Nervus olfaetorius. Archiv. f. mikr. Anat., Bd. 17. 1892. Beitrage zur mikroskopischen Anatomie der menschlichen Nas?n hohle. Archiv f. mikr. Anat., Bd. 39. 1892. Die Endigung der Olfactoriusfasern im Jaeobson'schen Orgaue

des Schafes. Archiv f. mikr. Anat., Bd. 39.

Cajal, R., 18S9. Neuvas applicaciones del metode de colovaciou de Golgi. Barcelona. 1894. Les noiivelles idees sur la structure du Systeme Nerveux chez

rHomme et chez les Yertebres. Chaxjaeau a., 1873. The Comparative Anatomy of the Domesticated

Animals. CisoFF, 1874. Zur Kenntniss der Regio olfactoria. CentralTil. f. medic.

Wiss. 12. Jahrg. Cuk:xingham, D. J., 1903. Text Book of Anatomy. Dejerine, J.. 1895. Anatomie des Centres Nerveux, T. 1. Disse, .J.. 1896. Ueber Epithelknospen in der Regio olfactoria der Sauger.

Anat. Hefte, Al)th. I, Bd. VI. 1900. Riechshleimhaut und Riechnerv. bei den Wirbeltieren. Ergeb nisse der Anat. vnd EntKickehrngsgeschichte, Bd. X. This contains

a good bibliography. DoGiEL. A., 1887. Ueber den Bau des Geruchorganes bei Ganoiden, Knochen fischen und Amphibien. Archiv f. mikr. Anat., Bd. 29.

EcKEE, 1855. Ueber das Epithelium der Riechschleimhaut und die wahr scheinliche Endigung des Geruchsnerven beim INIeuschen und den

Situgetieren. Bericht iihcr die Verhandl. zur Beford. der Natiirwiss.

zti Freiburg, No. 12.

EcKHARD, 1855. Ueber die Endigungsweise des Geruchsnerven. Beitriigc zur Anat. und Phys., Bd. I, H. 1.

Ehelich, 1886. Ueber die Methylenblaureaktion der lebenden Nervensub stanz. Deutsche medic. WocJicnschr., No. 4. Ellenberger, 1891. Anatomie des Ilundes.

ExNEE, 1872. Weitere Studien iiber die Structur der Riechschleimhaut bei Wirbelthieren. Situngsbcr. d. Akad. d. Wiss. Wien, Bd. 65, 3 Abth.

GoLGi, 1875. Sulla fina Struttura dei bulbi olfattorii, Reggio Emilia. Grassi, B., t;nd Castronovo, A., 1890. Beitrag zur Kenntniss des Geruchs organs des Hundes. Archiv ficr mikr. Anatomie, Bd. 34. Hab\^y, 1882. Note on Jacobson's organ. Quart. Journ. Mic. 8c., Vol. XXII,

HiRSCHFELD, L., AND Leveill^, 1866. Traite et Inconographie du Systeme Nerveux. Atlas.


Olfactory Apparatus in Dog;, Cat and Man 45

Jayne, Hoeace, 1898. Mainiualian Anatomy, I'art 1.

Kallius, E., 1905. Sinuescrgaue, Abtli. I. Gerncbsorgau uiid Geschmacks organ. Von Bardeleben, Handbucli der Anatomie des Mensclieu,

Bd. V, Abth. I, T. 2. This contains a good bibliography. Klein, E., 1881. Contributions to the Minute Anatomy of the Xasal Mucous

Membrane. Quart. Journ. Mic. Sc, Vol. XXI. 1881. A further contribution to the Minute Anatomy of the Organ of

Jacobson in the Guinea Pig. Quart. Journ. Mic. Sc, Vol. XXI.

1881. The Organ of Jacobson in the Rabbit. Quart. Journ. Mic. Sc, Vol. XXI.

1882. The Organ of Jacobson in the Dog. Quart. Journ. Mic. Sc, Vol. XXII.

KoELLiKER, A., 1877. Ueber das Jacobson'sche Organ des Menschen. Gratulationsclirift d. Wiirzburger mediz. FakuUiit f. Rinecker. 1896 and 1902. Handbuch der Gewebelehre des Menschen. Bd. II, III.

V. Lenhossek, jNI., 1892. Die Xervenurspriinge und -Endigungen im Jacobson'schen Organ des Kaninchens. Anat. Anz., Bd. 7.

Meckel, J. F., 1832. Manual of General Descriptive and Patholgical Anatomy, Vol. 3.

Mebkel, Fr.. 1892. Jacobson sches Organ und Papilla palatiua beim Menschen. Anat. Hefte, Bd. I.

MiHALKOVics, V. v., 1898. Nasenhi3hle und Jacobson'sches Organ. Anat. Hefte, Bd. XI. This contains a good bibliography for Jacobsou's or the vomeronasal organ.

Milne-Edwards, H., 1874. Lecons sur la physiologie et I'anatomie de I'homme et des animaux, T. 11.

MiVABT, St. G., 1900. The Cat.

Owen, R., 1868. Anatomy of Vertebrates, Vol. III.

Piersol, G. a., 1907. Hvmian Anatomy.

PoiRER, P., AND Charpy^ A., 1899. Traite d'Anatomie Humaine, T. 3.

QuAiN, 1897. Elements of Anatomy, Vol. Ill, Pt. II. 1906. Elements of Anatomy, Vol. Ill, Pt. III.

Ranvier, 1889. Traite Technique d'Histologie.

Retzius, 1892. Die Endigungsweise des Riechnerven. Biol. Vntersuch., N. F. III. 1892. Zur Kenntniss der Nervenendigungen in der Riechschleimhaut. Biol. Vntersuch, N. F. IV.

Sappey, C, 1889. Traite d'Anatomie Descriptive, T. III.

ScRAPA, Antonio, 1785. Anatomicarum Annotationum, Liber Secundus.

ScHULTZE, Max, 1856. Ueber die Endigungsweise der Geruchsnerven und die Epithelialgebilde der Nasenschleimhaut. Monatster. der kdnigl. Akad. d. Wiss. su Berlin.


46 Effie A. Eead

ScHurxzE. Max, 18G2. Ueber den Ban der Nasenschleimbaut. AhhaiuU. d.

Xaturf. Gesellsch. s. HaUc. Spalteholz, W., 1903. Hand Atlas of Human Anatomy, Vols. I, III. Stowell. T. B., 1886. The Trigeminus Nerve In the Domestic Cat. Amer.

PhUos. Society, May 21, 1886, pp. 459-478. Van Gehuchten, A., 1890. Contributions h I'etude de la Muqueuse olfactive

Chez les Mammiferes. La Cellule, T. 6. 1891. Le bulbe olfactif chez quelques Mammiferes. La Cellule, T. 7. 1900. Anatomie du Systeme Nerveux de I'Homme. Wilder and Gage, 1886. Anatomical Technology.

TERMS AND THEIR ABBREVIATIONS IN THE EXPLANATION OF FIGURES IN PLATES I-XVII. Ax., axone. Bo., bulbus ocnli. B. olf., bulbus olfactorius. Crt. v-n.,cartilago vomerona sails. Cv. n., cavum nasi. Cc. eth., cellulse ethmoidales. Cbl., cerebellum. Cbrm., cerebrum. Ch. n. i., concha nasalis inferior. Ch. n. m., concha nasalis media. Ch. n. s., concha nasalis superior. Ch. n., conchfe nasales. Cranium, cranium. D. i., dens inferior. D. s., dens superior. Os., developing bone. Dct. n-1., ductus nasolacrimalis. F. 1. cr., foramen laminje cribroste. L. cr., lamina cribrosa ossis ethmoidalis. Lingua, lingua. Mnd., mandibula. Mt. n. i.. meatus nasi inferior. Mt. n. m.. meatus nasi medius. Mt. n. s., meatus nasi superior. Md. sp., medulla spinalis. M. n., mucosa nasi. M. s. n., mucosa septi nasi. M. sn. f., mucosa sinl frontalis. M. sn. mx., mucosa sini maxillaris. M. c, Meckel's cartilage. Nn. olf., nervii olfactorii. Nn. org. v-n., nervii organi vomeronasalis.


Olfactory Ajiparatus in Dog, Cat and Man 47

N. etli. a., nervus etbmoidalis anterior.

N. mnd., nervus niandibularis.

N. mx., nervus maxillaris.

N. nph., nervus nasopalatinus.

N. sph., nervus sphenopalatinus.

N. org. v-n., nervus organi vonieronasalis.

Orbita, orbita.

Olf. c, olfactory cell.

Olf. h., olfactory hairs.

Org. v-n., organon vomeronasale.

Zyg., OS zygomaticum. ' •

Palatinum, palatinum.

R. n. ext., ramus nasalis externus.

R. n. lat., ramus nasalis lateralis.

Rg. olf., regio olfactoria.

Sen. c, sensory cells of the vomeronasal organ.

S. n., septum nasi.

Sn. f., sinus frontalis.

Sn. mx., sinus maxillaris.

Sn. spb., sinus spbenoidalis.

St. c, stellate cell.

Sust. c, sustentacular cell.


PLATE lA (X 1-14).

Same as Plate I. To show the lining of the maxillary sinns and its innervation by the fiih nerve.


PLATE lA (X 1-14).

Same as Plate I. To show the lining of the maxillary sinus and its innervation by the Sih nerve.


OLFACTORY APPARATUS IN DOG, CAT AND MAN

EFFIE A. READ



AMERICAN JOURNAL OF ANATOMY--VOL. VIII


PLATE I. Fig. 1 (X 1.14).

Head of a dog, the bone has been removed from the lateral aspect of the nose and part of the orbit, exposing the deeper layers of the mucosa in which lie the olfactory nerve bundles and their branches.

Note the divisions of the 5th nerve and their distribution to the mucosa of the lateral wall of the nose and to the maxillary sinus.


OLFACTORY APPARATUS IN DOG. CAT AND MAN

EFFIE A. READ


srN.p



1


tAMERICAN JOURNAL OF ANATOMY— VOL.


PLATE II. Fig. 2 (X D Median section of the liead of a dog, witli the mandible and bonj' septum removed, to show the septal mucosa and its olfactory nerves ; the turbinated part of the septum and its olfactory nerves ; the vomeronasal organ and its innervation by the olfactory and naso-palatine nerves.

Note the anterior ethmoidal branch of the 5th nerve along the dorsal wall of the septum and its distribution to the septal mucosa.


OLFACTORY APPARATUS IN DOG, CAT AND MAN

EFFIE A. READ



^


Orc\Mrl\. '

Nn.orq.V-Ti.


AMERICAN JOURNAL OF ANATOMY— VOL. VIK


PLATE III. Fig. 3 (x 1-4).

Head of a cat, except the mandible.

The bone has been removed from the nose, orbit, and a part of the cerebruiu. exposing the attached surface of the nasal mucosa. This shows especially the olfactory folds with their nerves, the divisions of the 5th nerve, and their distribution to the lateral wall and mucosa of the maxillai'y sinus.

Note the small branches of the anterior ethmoidal nerve to the olfactory folds.

Fig. 4 (X 1-4).

Median section of the head of a cat, without the mandible ; the bony septum has been removed to show the deep or attached surface of the mucosa. This is to illustrete the mucosa of the septum and its olfactory nerves ; the turbinated part of the septum and its olfactory nerves ; the vomeronasal organ and its innervation by the olfactory and naso-palatine nerves.

The anterior ethmoidal branch of the 5th nerve passes along the dorsal part of the septum and is distri))uted to the septal mucosa.


OLFACTORY APPARATUS IN DOG, CAT AND MAN

EFFIE A. READ


PLATE III



N. mrvii


Fig. 3




Fig. 4


AMERICAN JOURNAL OF ANATOMY--VOL. VIII


PLATE IV. Fig. 5 (x 1-28).

The ethmo-turhinnls of n do,^, showing olf.-ictory nerves in the deeper hvj-ers of the nnicosa and the branching of the nerves whicli lie upon each other. The lining of the maxillary sinus is turned forward to show the thinner part of the folds ; these are thick in the region supplied by the olfactory nerves: the thin part is not olfactory. The anterior ethmoidal nerve is shown entering the skull through the "ethmoidal foramen.

Fig. 6 (X 1.8).

Same animal as shown in Fig. 5.. Dorsal view of the ethmo-turbinals to show the olfactory nerves. Note the anterior ethmoidal branch of the .5tQ nerve to the septal mucosa.

Fig. 7 (X 1-2).

Median section of the head of a dog with the bony septum removed to show the turbinal part of the septum and its olfactory nerves ; the septum and its olfactory nerves, the anterior ethmoidal and the naso-palatine nerves.


OLFACTORY APPARATUS IN DOG, GAT AND MAN

EFFIE A. READ


PLATE IV



AMERICAN JOURNAL OF ANATOMY— VOL.


PLATE V.

Fig. 8 (x 1-07).

Median section of the head of a cat with tlae bony septum removed, to show especially the olfactory nerves to the septal mucosa and the vomeronasal organ. Note the arterial plexus.

Fig. 9 (X 3.6).

The septal mucosa of a eat, with the blood vessels injected. To show especially the vomeronasal organ and its olfactory nerves, and the olfactory nerves of the septum. Note the numerous blood vessels. The thicker part of the mucosa is light, tlie thinner dark.


OLFACTORY APPARATUS IN DOG, CAT AND MAN

EFFIE A. READ


PLATE V



AMERICAN JOURNAL OF ANATOMY- VOL. VIII


PLATE VI.

Figs. 10-12. Photographs of the drawings by Antonio Scarpa of the nose of man.

Fig. 10.

Median section of the nose of man with the bony septnra removed to show the septal mucosa. Note the very large distribution of olfactory nerves. The naso-palatine nerve is also shown.

Fig. 11.

A median section through the skull, septum removed, to show the superior, median and inferior conehae of the nose. Note the perforated appearance of these bones. The olfactory nerve bundles pass through these canals and within them form an apparent anastomosis. To demonstrate the nerves the bony walls of these canals must be removed piece by piece.

Fig. 12.

Septal mucosa removed to show the superior, median and inferior conehae. Note the apparent anastomosis of the olfactory nerves and their extent ; also the distribution of the 5th nerve.


OLFACTORY APPARATUS IN DOG, CAT AND MAN

EFFIE A. READ


PLATE VI




AMERCAN JOURNAL OF ANATOMY— VOL. VIII


PLATE VII.

Fig. 13 (X •<54).

Head of a child, about one year. The kiteral wall of the nose and orbit have been removed to show the deeper layers of the mucosa in which lie the nerve bundles. Note the anterior ethmoidal nerve.

Fig. 14 (X 3.75).

Lateral wall of the nose of a child about one year. Same as 1.3, enlarged. To show especially the plexiform appearance of the olfactory region of the lateral wall of the nose, which has been interpreted by all workers to be a plexus of olfactory nerves. These cords, however, are formed not only of olfactory nerve bundles, but also of blood vessels and connective tissue, as shown by Figs. 24, 2G, 27. Note also the anterior ethmoidal nerve.


OLFACTORY APPARATUS IN DOG, GAT AND MAN

EFFIE A. READ


PLATE VII



AMERICAN JOURNAL OF ANATOMY— VOL. VIII


PLATE VIII.

Figs. 15-17 (X 5). Transections of the head of an embryo dog.

Fig. 15.

In the region of the vomeronasal organ. Note its relation to the vomeronasal cartilage. Maxillo-turbinals are shown.

Fig. 1G.

In the region of the maxillary sinus. Note the number of folds of the ethmo-turbinals.

Fig. 17.

Near the cribriform plate. Note the increased number of folds of the ethmo-turbinals and the maxillary sinus.

Figs. 18-20 (X 5). Transections of the head of an embryo cat.

Fig. 18.

In the region of the vomeronasal organ. Note its relation to the vomeronasal cartilage. The inferior turbinal is shown.

Fig. 19.

In the region of the maxillary sinus. Note the number of folds of the ethmo-turbinals.

Fig. 20.

Near the cribriform plate. Note the increased number of folds of the ethmo-turbinals.

Figs. 21-23 (X 5). Transections of the nose of a four months human fetus.

Fig. 21.

In the region of the vomeronasal organ. Note its relation to the septum and to the vomeronasal cartilage. Inferior concha is shown.

Fig. 22. Shows the relation of the inferior and median conchfe.

Fig. 23.

Shows the relation of the superior, median and inferior conchae. Note the olfactory bulb, the cribriform plate and a section of an olfactory bundle.


OLFACTORY APPARATUS IN DOG. CAT AND MAN

EFFIE A. READ



15





16



')t> fi


17



AMERICAN JOURNAL OF ANATOMY--VOL.


PLATE IX.

Fig. 24 (X 17-5).

The right lateral wall of the nose of a child about one year. This was stained with a differential stain to demonstrate that the olfactory nerves pass as dark bands within the cords of connective tissue and to show that the plexiform appearance is due, not to these olfactory nerve bundles, as has always been stated, but to the ramification of blood vessels and to the amount and arrangement of the connective tissue.


OLFACTORY APPARATUS IN DOG, CAT AND MAN

EFFIE A. READ


PLATE IX



AMERICAN JOURNAL OF ANATOMY— VOL. VIII


PLATE X.

Fig. 25 (X 9.2).

A part of the septal mucosa of the nose of a child about one year. (Same child as shown in Fig. 24.) This is stained with a differential stain to demonstrate that the large olfactory nerve bundles do not form an anastomotic plexus.

Fig. 2G (X 9-2).

A drawing from Fig. 25. The nerves were carefully followed out to show that any plexiform appearance is due to the crossing and recrossing of nerves. Joining was found in a few cases of the smaller nerve bundles.


4


olfactory:apparatus in dog, cat and man

EFFIE A. READ


PLATE X




y.-M







/;^\;l:M>


Z6


AMERICAN JOURNAL OF ANATOMY--VOL. VIII


OLFACTORY APPARATUS IN DOG, CAT AND MAN

EFFIE A. READ


'^rfy



Fig. 27 (X V2.S) .

Longitudinal section of tlie left lateral wall of the nose of a child about 1 year, same child as shown in Figs. 24 to 2('). This figure shows in section what is demonstrated in Fig. 24. that is, that tlie plexifnrni aiipearauce of this region is due to l>lood vessels, an<l not to nerves.


PLATE XI (i/L' natural size). Figs. 28, 29.

Von Brunn's 1892 diagram of nasal cavity of man to sliow the area of olfactory nerve distribution (blackened area). Fig. 28 = man 40 years; Fig. 29 = man 30 years.

Figs. 30, 31. Diagram of olfactory nerve distribution made from my dissection. Fig. 30=: child about 1 year. Fig. 31 =: man 30-40 years.


OLFACTORY APPARATUS IN DOG, CAT AND MAN

EFFIE A. READ


PLATE XI



30


3/


AMERICAN JOURNAL OF ANATOMY— VOL.


PLATE XII.

Fig. 32 (x 585).

Olfactory cells from the nasal mucosa of the dog (Golgi stain). Note the long axoues and the olfactory hair. (Enclosing lines indicate the thickness of the epithelium.)


OLFACTORY APPARATUS IN DOG, CAT AND MAN

EFFIE A. READ


PLATE XII



iP.


AMERICAN JOURNAL OF ANATOMY— VOL. VIII


PLATE XIII.

Figs. 33, 34.

Olfactorj' cells from the nasal luncosa of the cat. (Enclosing lines indicate the thickness of the epithelium.)

Fig. 33 (X S33.4). . Cells stained AA-ith methjiene blue.

Fig. 34 (X 625).

Cells stained by the Golgi method. Olfactory hairs are very distinct at the free end of several of the cells.


OLFACTORY APPARATUS IN DOG, CAT AND MAN

EFFIE A. READ


PLATE XIII



33



AMERICAN JOURNAL OF ANATOMY— VOL. VIII


PLATE XIV.

Fig. 35-38 (X TOO). Olfactory cells. (Enclosing lines indicate the thickness of the epithelium.)

Fig. 3.J. Olfactory cells from the cleg, stained with gold chloride.

Fig. 30. Olfactory cells from the dog. stained with gohl chlniide.

Fig. 37. Olfactor.v cells from the cat; isolated h.v f(irniald('hy<le dissociatcr.

Fig. 3S.

Olfactory cells from the nasal mucosa of man. stained with gold chloride and dissociated l)v formaldehyde. X(.te the short axones in two of the cells.


OLFACTORY APPARATUS IN DOG CAT AND MAN

EFFIE A. READ


PLATE XIV



3(o



.Olfcall


' Axorve.


3S


i ^


)'i



Olf.CC2.ll.



37


^3?:


AMERICAN JOURNAL OF ANATOMY— VOL. VIII


PLATE XV. Figs. 39-45. Fig. 39 (X 450). Olfactory epithelium of the cat; formaldehyde dissociation.

Fig. 40 (X 450). Isolated sustentacular cells of the cat.

Fig. 41 (x 300). Duct of Bowman's gland of the cat. Golgi stain.

Fig. 42 (x 450). Olfactory epithelium of an embryo dog ; gold chloride stain.

Fig. 43 (X 450). Ciliated cells from the respiratory epithelium of the nose of the cat.

Figs. 44, 45 (X 450). Free nerve terminations of the 5th nerve in the nasal mucosa of the cat.

• Fig. 44.

In the respiratory region.

Fig. 45. In olfactory region.


OLFACTORY APPARATUS IN DOG, CAT AND MAN

EFFIE A. READ


PLATE XV




'cV'Av'Si _.Ju.t:<i




1^




f

i

^0




t'i





/.w



^'«5


AMERICAN JOURNAL OF ANATOMY--VOL. Vlll


PLATE XVI.

Figs. 46, 47.

Glomeruli of the olfactory bnlb in tlie dog (Golgi stain), showing the end brushes of the olfactory axoues.

Figs. 48-50.

Glomeruli of the olfactory bulb in the dog (Golgi stain), showing the end brushes of the mitral cell dendrites.

Fig. 48 (X 310). Three glomeruli formed by the branching of one dendrite.

Fig. 49 (x 310). Glomerulus formed of dendrites from two different mitral cells.

Fig. 50. Glomerulus formed by the branching of one dendrite.

Figs. 51-53.

Glomeruli of the olfactory bulb of the cat. Golgi stain. Mitral cells and dendrites.

Figs. 51, 52.

Glomeruli formed of three dendrites from at least two mitral cells. Fig. 51 (X 310). Fig. 52 (X 4T5).

Fig. .53 (X 310). Glomerulus fcrn;ed by two dendrites from the same mitral cell.


OLFACTORY APPARATUS IN DOG, CAT AND MAN

EFFIE A. READ


PLATE XVI



AMERICAN JOURNAL OF ANATOMY-VOL. VIII


PLATE XVII.

Figs. 54-55a.

Fig. 54 (x 50).

Vomeronasal organ in an embryo kitten.

Section tbrougli the ceplialic region of the organ, the cartilaginous capsule entirely enclosing it. Two olfactory cells are shown in the upper part of the lining of the epithelium.

Figs. 54a, 55a (X 488).

Olfactory cells of the vomeronasal organ, with varicose axones. The cell with the longest axone was drawn from a different region of the same organ (enclosing lines indicate the thickness of the epithelium).

Fig. 55 (X 30).

Section through the middle of the vomeronasal organ ; cartilagenous capsule not entire. Note the difference in thickness of the median and lateral epithelium ; an olfactory cell is shown in the epithelium of the median wall.


OLFACTORY APPARATUS IN DOG, CAT AND MAN

EFFIE A. READ


PLATE XVII


^r^-:


-^ -SerA C-.


CH:\/•-r^.



,ho(i"


.P\xov\<z


Orq. V-r\.


-5 ///I



Orq. \/~n.


S^A


^S


AMERICAN JOURNAL OF ANATOMY-VOL. VIII


DISTEIBUTION OF THE SUBCUTANEOUS VESSELS IN THE TAIL EEGION OF LEPISOSTEUS.

BY

WILLIAM F. ALLEN.

From the Hersstein Marine Lahoratory, Pacific Gi'ove, Cal.

With 25 Figures.

Introduction.

The following is merely a continuation of a recent article^ on the distribution of the subcutaneous vessels in the head region of the Ganoids PoJyodon and Lepisosteiis; which was an attempt to see what light a study of these vessels in this group would throw on the general problem as to whether these vessels in fishes are veins or lymphatics; or a conunon system that may function for both; or if perchance they might not function as veins in the lower or more generalized forms and lymphatics in the higher or more specialized forms. In other words, to ascertain, if possible, what bearing, if any, they might have on the origin and phylogeny of the lymj)hatic system.

Material and Method of Procedure. — Two species of Lepisosteus were studied, namely, L. tristceclius and L. osseus. They were obtained from the Ohio and Mississippi Elvers -about Cairo, Illinois. The tails of these specimens were transversely severed immediately in front of the dorsal and anal fins; and the caudal vein and one of the lateral subcutaneous trunks were injected caudad from the cut ends with a Prussian blue gelatin mass. Ordinarily when injecting one of the lateral trunks, the opposite trunk and the dorsal and ventral subcutaneous trunks were plugged with cotton. With a few specimens the caudal artery was injected with Hoyer's lead chromate gelatin mass. In making the injections the cut end of the specimen was raised consid '"Distribntion of the Subcutaneous Vessels iu the Head Region of the Ganoids, Poltjodoii and Lepisosteus." Proc. Wasli. Acail. Sci.. Vol. IX, pp. 79158, 1907.

American .Tolexal of Anatomy. — Vol. VIII.


50 William F. Allen

erably higher than the tail end. After a specimen had been satisfactorily injected it was placed in a pail of cold water till the gelatin mass was hardened, and then preserved in a 4 per cent solution of formalin until needed for dissection. All of the dissections were made at the University of California Marine Laboratory, Pacific Grove, California. The observations made from these dissections were checked up from several transverse series of larva or small adult Lepisosteus, but mostly from a 90 mm. L. osseus tail, which had been killed and fixed in Teltyesniczky's potassium bichromate-acetic mixture, imbedded in paraffin, cut 10 microns, and stained in Heidenhain's iron hsematoxylin and counterstained in a concentrated alcoholic solution of orange G plus a little acid fuchsin.

Blood- AVASCULAR Supply for the Caudal Fin Eegion.

Before taking up this most interesting system of subcutaneous vessels it might be well, in order to avoid confusion, to first describe briefly the distribution of the blood vessels for this region.

The arterial trunk which supplies the caudal fin is the posterior continuation of the dorsal aorta or the caudal artery (Figs. 1-3, 9-19, C. A.). This trunk follows along in the haemal canal of the vertebral column immediately above the caudal vein. Under the ninth from the last vertebra it sends off a left caudal branch, which in company with the corresponding vein passes between the superficial and deep muscles of the fin to supply each. Then for a short distance the main artery travels along the right side of the ninth or tenth haemal spines of the tail (counting ventro-dorsad), between the vertebral column and the caudal subcutaneous or lymphatic trunk to send off a branch which pursues a dorso-caudal direction in company with the vertebral column; while the main stem continues caudad with the caudal trunk. Before the caudal fin is reached it crosses the right side of the subcutaneous trunk, arriving at a position between the trunk and the ninth or tenth hfemal spine of the tail, and finally when the distal end of the spine is reached it makes a sharp curve at right angles to travel ventrad through the basal canal of the fin. Its position in this canal is at first laterad of the caudal vein and in front of the caudal trunk, but further on in its ventral course it comes to lie in front of the vein. Throughout this canal it sends off two caudal fin ray hranclies (Figs. 1, 2, 18 and 19, C. E. A.), which run along the dorsal and ventral surfaces of each ray to give "off a network for the fin meuibrane.


Subcutaneous Vessels in Tail of Lepisosteus 51

Eunning parallel with the caudal artery, but below in the haemal canal, is the caudal vein (Figs. 1-7 and 9-19, C. V.). Although its walls are much thinner, yet, when expanded, its diameter is much greater than the arterial trunk, and below the tenth from the last vertebra it resembles a sinus. At this point it receives a large neural vein from above and the two caudal sinuses from above and the rear. The entrances of the latter are guarded by a pair of semi-lunar valves. Tracing the caudal vein posteriorly, it was found to leave the hgeraal canal at the seventh or eighth htemal spine of the tail (counting ventrodorsad), in company with, but ventral to, the left caudal sinus. Occasionally, however, the vein accompanies the right caudal sinus; in which case the left caudal vein is the main stem, and drains the tin. At first, the caudal vein is a more superficial vessel than either the caudal artery or the so-called caudal subcutaneous trunk, for it passes caudad between the superficial and the deep muscles of the fin. Upon reaching the basal canal of the tail, between the third or fourth and the fourth or fifth rays (counting dorso-ventrad), it immediately makes a curve at right angles to pass ventrad through this canal in front of the caudal subcutaneous trunk and either lateral to the caudal artery or between it and the caudal trunk. From each ray it receives two branches that traverse the upper and lower surfaces of the ray and collect a capillary network from the fin membrane. This vein usually lies betw^een the caudal ray artery and the caudal ray subcutaneous canal, but frequently, toward the base of the ray, it runs laterad with the artery, immediately proxamad of the caudal ray canal.

Fig. 23 is from a posterior view of a transverse microscopic section taken through the h^mal canal of a 90 mm. L. osseus. The caudal artery (C. A.) presents no special peculiarities. It is composed of a muscular layer, which is lined internally with an endothelial layer, and is filled with corpuscles, mostly red. The caudal vein (C. V.), on the contrary, is devoid of all muscular elements, and in specimens of this size is composed solely of a layer of endothelium, containing very few corpuscles. Both the artery and vein are supported in the hsemal canal by a spongy connective tissue, and are accompanied by two additional longitudinal trunks, which will be described later on.

In Fig. 11 Hopkins- figures the caudal vein of Amia (:=Amiatus) as ending under the eighth from the last vertebra and here receiving

^Hopkins, G. S. "The Lymphatics and Enteric Epithelium of Amia calva," The WiUlrj- Quarter-Ccntvnj Book. Ithaca, N. Y., pp. .307-38.5, 1893.


53 William F. Allen

the caudal sinuses. Likewise Hyrtl/ in Figs. 1, 2 and 4, V., represents the caudal vein in Esox (= Lucius) and Leuciscus as terminating under the last vertebra, and receiving a papilla from each of the caudal sinuses. With the bony fishes Sappey* (p. 46) states that the caudal vein takes its origin in two branches from the base of the caudal find and is so figured in the pike (PL XII, Fig. 2, 7). With the Selachians and Polyodon the structure of the caudal fin is so different that a comparison is hardly permissible. Mayer^ (p. 325) finds that the caudal vein extends much further caudad with the dogfish than with the roaches. With the latter it may become paired, then reuniting, finally terminates, and its place farther caudad is taken by branches that form a vasa vasorum with similar branches from the caudal artery.

The distribution of the caudal artery and vein doubtless conforms to the same general plan in all fishes, but in a great many excellent papers on the circulation of blood in fishes there has been a marked tendency to overlook the final ending of the caudal artery in the caudal fin and the origin of the caudal vein from the same. It may be that the caudal vein in some fishes is not continued into the caudal fin, but I mistrust in those instances that the injection mass has failed to reach the posterior end of this trunk. If such' were not the case, it would materially strengthen the hypothesis that the subcutaneous vessels which came from the tail and emptied into the caudal sinuses and then into the caudal vein were veins rather than lymphatics.

Intercostal Arteries and Veins. — These are among the most important branches of the caudal artery and vein. Ordinarily, as with the Teleosts, a neural and h^mal artery arise from under each alternate vertebra, and a neural and haemal vein empty into the caudal vein beneath the intermediate alternate vertebrge. The haemal vessels follow the hamal spines, and the neural vessels after crossing their respective vertebra run along the neural spines. From some of the neural arteries and veins, but not from all, a lateral artery or vein (Fig. 8, L. A. and L. V.) is given off or received. Tracing these vessels, laterad, they are found

^Hyrtl, Jos. "Ueber die Caudal uud Kopf-Siuuse der Fische, uud das damit znsammenhangende Seitengcfilss-System," Archiv fiir Anatomie und Physiologic, pp. 224-240, 1843.

Sappey, P. C. "Etudes sur I'appareil mucipare et sur le systeme lymphatique des poissons." pp. 1-G4, Paris, 1880. "Mayer, P. "Ueber Eigenthiimlicblveiten In den Kreislaufsorganen der Selacbier," MitthoiUmocn avs der zooJog'ischen Station zu Ncapel, VIII Bd., 1888.


Subcutaneous Vessels in Tail of Lepisosteus 53

to run along the septum between two myotomes to the great lateral subcutaneous trunk. Mesad of this canal they break up into numerous branches, which' for the most part follow, superficially, dorsad, or ventrad, along the septa between the myotomes, to form a network in the connective tissue that binds the skin to the body muscles.

One of the largest of the neural veins is the last one (Figs. 1, 3 and 14, Neu. V. (1)). It takes its origin in the posterior dorsal corner of the caudal peduncle. After passing obliquely cephalad for a short distance, it curves ventrad to cross the left surface of the eighth vertebra from the last. It then crosses the left sinus (x) and finally empties dorsally into the caudal vein a little anterior to the openings of the caudal sinuses. In Fig. 5 the abbreviation Neu V. (i) 0. marks the opening of this vein into the caudal. In the first dissections this vein was taken to be a part of the subcutaneous system opening into sinus (x).

Vascular Supply for the Dorsal Fin. — In the region of the posterior part of the dorsal fin two of the neural vessels are greatly enlarged to supply the dorsal fin. In the specimen from which Fig. 9 was drawn the dorsal fin artery (D. F. A.) took its origin from the caudal artery at about the level of the third vertebra from the posterior end of the dorsal fin. After crossing the right side of this vertebra it crossed the two preceding interneural spines obliquely to enter the basal canal of the dorsal near the middle part of the fin. Here it separated inta three branches; two of which supplied the anterior part of the fin and the third the posterior part. From these branches one or twa dorsal ray arteries (Fig. 9, D. E. A.) were given off to pass along either the anterior or posterior surfaces of the ray or both. In this specimen a second neural or a minor dorsal fin artery supplied the posterior part of the fin (see Fig. 9). With another specimen of L. tristcechus, where the dorsal fin artery was dissected out, it had its source about nine vertebrse cephalad of its position in Fig. 9, and approached the dorsal fin from in front. The dorsal fin vein (Fig. 9, D. F. V.) emptied into the caudal vein two vertebrge behind the origin of the dorsal fin artery. Tracing this vein peripherally, it was seen to cross the left side of the second vertebra, behind the dorsal fin artery. Continuing dorsally behind its neural and interneural spines it entered the basal canal of the dorsal spine from the rear, and traveling clear through this canal it was supplied from numerous dorsal fin ray veins (Fig. 9, D. R. v.), which may traverse either the anterior or posterior


54 • William F. Allen

or both surfaces of the rays, j^arallel with, but usually distal to, the corresponding arteries.

With the Selachians in addition to the subcutaneous vessels described by Sappey {op cit., p. 39) for the dorsal fin, Parker*^ (p. 720) and Mayer {op. cit., pp. 333-5) find a deeper vein, which is in connection with this system and helps drain this region. By Mayer it has been styled as the vena profunda (PI. XVI, Figs. 31, 23 and 24, v. prof.), and, in brief, Mayer sets forth the union of the so-called subcutaneous veins with the profundus as follows : The dorsal subcutaneous vein, at the insertion of the dorsal fin separates into two vence circulares, which encircle the fin and collect subcutaneous branches from it. Posteriorly they reunite at the base of the fin in a reservoir. This reservoir also receives one or two vena postica (V. P.), which travel along the distal edge of the fin cartilage and collect the blood from the inner parts of the fin. The vena profunda, in addition to collecting numerous branches from the fin muscles, which run parallel to the corresponding subcutaneous branches, communicates directly with the above mentioned subcutaneous reservoir at the posterior insertion of the fin, and eventually passes along the sheath that separates the great lateral muscles, to terminate in the caudal vein, or, acording to Parker, with Mustelus empties into the left renal portal vein. •

If in Lepisosteus we had the dorsal fin vein fusing with the dorsal subcutaneous trunk at the posterior end of the dorsal fin, or, better still, if the dorsal subcutaneous trunk (Figs. 1 and 2, D. T.), instead of passing through the basal canal of the fin, had formed two circular canals around the base of it, and had joined the dorsal fin vein in a sinus at the posterior base of the fin, we would have a condition of affairs almost identical with that found in the Selachians. Can it not therefore be possible that farther back in the phylogeny, or possibly in the embryo of Lepisosteus, the dorsal fin vein had such a comunication, and that as specialization advanced the subcutaneous vessels became more and more separated from the deep, and with this differentiation a change in function occurred? In such a case the vena profunda and the vena postica of the Selachians would be homologous to the dorsal fin vein and its branches of Lepisosteus.

Parker, T. J. "< )n the Blood- Vessels of Mustelus antarcticus : a Contribution to the Morpliology of the Vascular System in the Vertebrata," Phil. Trans., pp. G85-732, 1886.


Subcutaneous A'essels in Tail of Lepisosteus. 55

The distribution of the blood vessels to and from the anal fin was not studied. One, however, would naturally expect to find it supplied by an enlarged hsmal artery and vein, somewhat after the general plan of the dorsal fin.


SUBCUTAXEOUS YeSSELS OF THE CaUDAL FiX.

As with most fishes four longitudinal subcutaneous trunks, respectively, dorsal, lateral and ventral, are to be found immediately below the skin in Lepisosteus. In one way or another they discharge themselves posteriorly into two caudal sinuses, which empty into the caudal vein.

Ventral Subcutaneous Trunh (Figs. 1, 2, 3, 5 and 10-16, V. T.). — This important canal travels along the ventral median line, directly under the skin, in a mass of connective tissue that binds the two great lateral muscles. As stated in a previous paper {cp. cit., p. 114), this canal usually bifurcates anteriorly, each fork terminating in its respective pericardial sinus. Upon tracing this canal caudad it was found to enter the basal canal of the anal fin, and to receive an anal fin ray canal (Figs. 1 and 2, A. E. C.) from the anterior and posterior surfaces of each ray, which drains a network of canals from the membrane connecting each two rays. After leaving the anal fin, this trunk continues along the ventral surface of the caudal peduncle, to penetrate the basal canal of the caudal fin, and to become the caudal subcutaneous trunlc (Figs. 1-5, 7, 17-19, C. T.). It receives here, as within the anal,canal, two caudal fin ray canals (Figs. 1, 2, 18 and 19, C. E. C), which run along the dorsal and ventral surfaces of each ray, and receive a network of canals from the membrane joining two rays Upon reaching the third or fourth ra}' (counting dorso-ventrad) the caudal trunk makes a sharp curve at right angles to pass cephalad along the upper margin of the ninth or tenth of the last haamal spines (counting ventrodorsad), in company with, but below, the caudal artery, to empty into either the right or the left caudal sinus; oftener into the right.

Fig. 18 shows the caudal trunk of a 90 mm. Lepisosteus osseus, twice, in cross section. Toward the lower j^art of the figure it is seen traveling along through the basal canal of the caudal fin, and 1)elow the vertebral column it is seen again not far from its entrance into the right caudal sinus. In Fig. 23 we have a view of the posterior part of the caudal trunk higfhlv magnified. It consists solelv of a laver of endothelium


56 William F. Allen

imbedded in a mass of fibrous connective tissue. A few red corpuscles were found in this trunk throughout this series. In Figs. 18 and 19 several of the caudal rays, together with their blood vessels and subcutaneous canals, are seen in section. These subcutaneous canals or lymphatics are readily distinguishable from the blood vessels on account of their larger size, and by the additional fact that they frequently surround the blood vessels to a considerable extent.

With the Selachians, Mayer (op. cit., pp. 316-7) is the only writer to throw much light on the caudal termination of the vena ventralis, as he styles it. Like the dorsal vein, he finds it paired at the origin of the tail, each fork ending in its corresponding lateral vein, and these latter being in communication with the caudal vein. Parker {op. cit., p. 721) notes with Mustelus that the posterior ventral vein forms a loop around the anal fin, and after bifurcating at the cloaca, each fork anastomoses with its corresponding cloacal vein. iVs with the Selachians, there has been a notable silence concerning the caudal termination of the ventral subcutaneous trunk in the Teleosts. With Pleuronectcs, Sappey {op. cit., p. 49) finds that the ventral lymphatic tnmk (PI. XIT, Fig. YI, 4) is continuous with the caudal and dorsal trunks and forms an ellipse around the body, Avhich ends ventro-cephalad in the ductus of Cuvier and dorso-cephalad in the jugular vein. With Amiatvs, Hopkins {op. cit., pp. 372-3) represents the ventral trunk (PL II, Fig. 11, 0. and V.) as arising ,from the base of the caudal fin, and extending cephalad along the ventral side of the body to the heart, where it ends in the pericardial sinuses. In one instance an anastomosing trunk (Fig. 11, t) connects the tail part of the trunk with the lateral trunk, as the latter bends mesad to join the caudal sinus. This communication recalls a somewhat similar arrangement in Lepisostcus, where the caudal trunk or the posterior continuation of the ventral trunk was just described (p. 55) as terminating in the right caudal sinus after traveling through the basal canal of the tail.

Dorsal Subcutaneous Trunk (Figs. 1-3, 5, 10, 12-13, D. T.). — In many respects this trunk in Lepisosteus is similar to the ventral canal. It follows the dorsal median line in a layer of tough connective tissue that binds the great lateral muscles together. As stated in an earlier paper {op. cit., pp. 113-4), when not far from the posterior end of the skull it empties into the left branchial sinus, which is in direct connection with the cephalic sinus, and through the latter it reaches the jugular. When the dorsal fin is reached, instead of forming two


Subcutaneous Vessels in Tail of Lepisosteus 57

circular trunks about its base as in most fishes, it passes entirely through its basal canal, and receives two dorsal fin I'ay canals (Figs. 1 and 2, D. E. C), which traverse the anterior and posterior surfaces of each ray to receive a network of vessels from the fin membrane. In position the fin ray canals are more distal from the rays than the corresponding arteries and veins. Unlike the ventral trunk, the dorsal does not extend to the tail, but about midway between the posterior end of the dorsal fin and the beginning of the caudal it makes a sharp bend at right angles to continue ventrad across either the right or the left side of about the eleventh vertebra from the last, and here culminates in a longitudinal sinus, designated as sinus (x). With both L. tristcechus and L. osseus the dorsal trunk may empty into either the right or the left sinus (x), but in no specimen did the dorsal trunk bifurcate and each fork terminate in a sinus (x). In the specimens from which Figs. 2 and 5 were drawn the dorsal trunk crossed the right side of the vertebral column and ended in the right sinus (x) ; while in the specimen from which Fig. 3 was drawn it crossed the opposite or left side of the vertebral column and emptied into the left sinus (x). Fig. 12 shows the dorsal trunk of a 90 mm. L. osseus, in section, passing across the left side of the vertebral column, and Fig. 13 its termination in the left sinus (x).

Sinu^ (x) (Figs 1-6 and 11-16, x). — The two sinuses so designated are situated deeply along the lower and outer surface of the vertebral column, somewhere between the sixth and the thirteenth from the la?t vertebra. They are variable in length, position and in their mode of connections; both in different specimens and on the opposite sides of the same individual. Both of these sinuses receive a lateral subcutaneous trunk and a haemal longitudinal trunk, communicate ventrad with a caudal sinus that empties into the cardinal vein, and one or the other of them receives the dorsal subcutaneous trunk. In Figs. 1, 3, 4, 5 and 6 the anterior end of sinus (x) received the lateral trunk from the side and the haemal trunk from below and within; while in Fig. 2, which is the opposite side of the same individual as Fig. 1, the lateral trunk joined this sinus considerably caudad of the union of the dorsal trunk; in fact, the opening was about opposite the connection of sinus (x) with the caudal sinus. In the dissection of L. tristcechus from which Figs. 1 and 2, and the dissection of L. osseus from which Figs. 5 and 6 were drawn, the dorsal trunk terminated in the right sinus (x), about midway between the anterior and posterior extremities; while


58 William F. Allen

with the L. tristoeclius from which Figs 3 and 4 were drawn the dorsal trunk culminated in the left sinus (x), nearer its anterior extremity than its posterior. In Figs. 1 and 2, especially the former, sinus (x) extends some little distance caudad of its ventral comm^^nication with the caudal sinus; while in Figs. 3-6, sinus (x) might be described as discharging itself caudad and ventrad into the caudal sinus. As a matter of fact, sinus (x) in Figs. 3-6 appeared to be merely a deep continuation of the lateral trunk, which empties into the caudal sinus. No valves were observed at the entrance of any of the suljcutaneous trunks into sinus (x), or guarding the exit of the latter into the caudal sinus.

Some interesting observations should be recorded in connection with the dorsal trunk and sinus (x) from a series of transverse sections taken through the tail of a 90 mm. L. osseus: The right sinus (x) (Figs. 11-16, x) in this larva or small adult had a lengih of 1.42 mm. It began anteriorly as a blind sac, and upon following this series toward the tail the right sinus (x) was found to receive the hsemal longitudinal trunk .43 mm. and the lateral trunk .83 mm. caudad of its origin. The left sinus (x) has about the same length as the right, but is unlike it in that it is merely a continuation of the left hasmal trunk. It receives the left lateral trunk .6 mm. and the dorsal trunk .28 mm. cauded. The two sinuses (x) terminate in their respective caudal sinuses, and in section (see Fig. 16, x) their walls are continued some little distance into the caudal sinuses. Both sinus (x) and the dorsal trunk are surrounded by a mass of connective tissue. They are composed of a single layer of endothelium, and contain only a very few red corpuscles. No leucocytes were observed.

With the Selachians, Parker (op. cit.. p. 720), Sappey (op. cit., p. 38), and Mayer {op. cit., pp. 316-8, find a dorsal cutaneous vein or lymphatic trunk extending from the head to the tail. Sappey states with Squalus that it bifurcates anteriorly, each fork terniinating behind the eye in the jugular vein. Parker describes the lateral cutaneous veinin Mustelus (p. 721) as anastomosing caudad with both the caudal and the dorsal cutaneous vein; while Mayer portrays the vena dorsalis as anastomosing caudad with the vena lateralis. Of these three writers, Mayer gives us the most comprehensive description of the distribution of the subcutaneous vessels in the dorsal fin. He sets forth the vena dorsalis (pp. 333-4 and PL XVII, Fig. 17. v. d.) as separating into two vencB circidares (Fig. 17, v, circ). which encircle the fin, collect


Subcutaneous Vessels in Tail of Lepisosteus 59

subcutaneous vessels from the fin, and reunite in an unimpaired trunk at the posterior insertion of the fin. At the junction a reservoir of considerable size is formed, which is also in communication with the vena postica and the vena profunda, as has already been noted under the head of the dorsal vein (p. 54).

Concerning the caudal distribution of the dorsal trunk in Teleosts, Trois tells us with Lophius' (pp. 6-7) and Uranoscopus^ (p. 23) that the dorsal lymphatic trunk divides into three in the region of the dorsal fin; one branch passes through the median basal canal, and the other two encircle the base of the fin. With Lophius, Fig. 3, shows this trunk as having superficial connections with the lateral trunk, and Fig. 4 represents the dorsal, ventral, neural and haemal longitudinal trunks as anastomosing at the base of the tail; while the neural vessels connect the dorsal with the neural trunks, and the hsemal vessels the ventral and the haemal trunks. A somewhat similar arrangement was found by Sappey for the carp and the pike {op. cit., p. 47, and Pis. XI and XII, Figs. V, I and II, 4, 8 and 15), except that the dorsal trunk was not prolonged to the tail. In Pleuronectes, Sappey notes (p. 50) that the dorsal lymphatic trunk (PI. XII, Fig. IV, 1) is continuous with the caudal and the ventral, forming an elliptical trunk about the body, which ends dorso-cephalad in the jugular vein, and ventrocephalad in ductus of Cuvier. Throughout its course it was said to be a single trunk, passing through the basal canal of the fins. In a previous paper^ it was stated (pp. 54-5) that the distribution of the dorsal l}Tnphatic trunk of Scorpceniclitliys (Figs. 1 and 4, d. I. v.) was not dissimilar to Trois' description of it for Lophius. It was connected with the lateral trunk, the intermuscular or transverse vessels, and with the neural trunk through the neural vessels. Its caudal distribution has also been studied, but has been reserved for a separate paper.

As for the Ganoids, Hopkins contends {op. cit., p. 373) that the dorsal lymphatic trunk (Fig. 10, n.) of Amiatus is a single trunk, which extends along the dorso-meson from the caudal end of the body to the base of the cranium. At the cranium the dorsal trunk is said to

'Trois, E. F. "Ricercbe sul sistema linfatico del Lophius piscatorius," Atti del R. Institute Veneto di scienze, lettere ed arti, pp. 3-20, 1878.

'Trois, E. F. "Rieherche sul sistema linfatico dell' Uranoscopus scaber," Atti del R. Instituto Teneto di scienze, lettere ed arti^ pp. 19-36, 1880.

'Allen, W. F. "Distribution of the Lymphatics in the Head, and in the Dorsal, Pectoral, and Ventral Fins of Scorpsenichthys marmoratus," Proc. Wash. Acad, of Sci., Vol. VIII, pp. 41-90, 1906.


GO William F. Allen

bifurcate, each branch extending laterad to join its respective cephalic sinus. At the caudal end Hopkins states that "it anastomoses with the lateral lymph vessel, joining it just after the latter turns at right angles to enter the caudal sinus." Hopkins further believes that the dorsal trunk bifurcates, each fork terminating in a lateral trunk immediately after it bends to join the caudal sinus. This is so represented in Fig. 10, r.

From the above paragraphs it is evident that the distribution of the dorsal subcutaneous trunk of Lepisosteus most closely resembles Hopkins' description of a similar trunk for Amiatus. On the contrary, the dorsal trunk never forks caudad or terminates in the lateral trunk, but rather in one of the sinuses (x), each of which is a deep reservoir formed by the union of lateral and haemal trunks which culminates in its respective caudal sinus. As with Amiatus and Pleuronectes, it is a single median trunk in the dorsal fin region. Since there is no longitudinal neural trunk in Lepisosteus, there are no neural vessels to communicate with the dorsal trunk, as is the case with some Teleosts. Later on the lateral trunk will be described as collecting numerous intermuscular or transverse vessels, which doubtless communicate above with the dorsal trunk as in 8coq)a'nichthys, but this point was not determined for a certainty. In the basal canal of the dorsal fin the dorsal trunk received a canal from the anterior and posterior surfaces of each ray, which collected a network of canals from the fin membrane. These dorsal ray canals were accompanied more proximad by dorsal ray veins, which gathered a venous network from the fin membrane, and ultimately discharged into a dorsal fin vein that emptied into the caudal vein. As previously stated (pp. 7 and 8), this system of veins may be homologous to the deep dorsal fin veins of the Selachians, which were described by Mayer as anastomosing with the dorsal subcutaneous vein in a reservoir at the posterior end of the fin, but with Lepisosteus there was no anastomosis between the venous system and the subcutaneous system. The two systems were distinctly separate. "We have, therefore, in Lepisosteus a more specialized condition than is to be found in the Selachians. It might also be mentioned that these two systems are clearly unconnected in the dorsal fin region among the Teleosts. With ScorpcenicJitliys, however, there is no one vein that collects the entire venous blood from the dorsal fin, but several of the neural veins are extended to the fin and send off branches that traverse the fin rays and receive the capillary network from the fin.


Subcutaneous Vessels in Tail of Lepisosteus 61

Lateral Subcutaneous Ti-unks. — Unquestionably these subcutaneous trunks have received far more attention than any of the others.

With the Selachians, Parker {op. cit., p. 721) observed that the lateral cutaneous vein in Mustelus anastomoses posteriorly with both the dorsal cutaneous and the caudal veins. Sappey (op. cit., p. 38) found two lateral lymphatic trunks in Squalus. Le tronc lateral superieur (PI. X, Pig. Ill, 2) is represented as running along parallel with the mucous canal; caudad it expands into a fibrous caudal sinus, which opens into the caudal vein. Le tronc lateral inferieur (PI. X, Pig. Ill, 11) is portrayed as traveling along the median lateral line, parallel to the lateral line, but more superficially. Posteriorly this trunk is indicated as gradually rising higher and higher until it eventually anastomoses with the superior lateral trunk. Sometimes (evidently meaning in some species) Sappey notes that the inferior lateral trunk is absent. Certain cross vessels (Fig. Ill, 14, 12 and 3) connect the ventral trunk with the inferior lateral, the inferior lateral with tha superior lateral, and the superior lateral with the dorsal, and the cross vessels drain a subcutaneous network (13). Mayer {op. cit., pp. 316-7, and PI. 16, Figs. 2-4, v. I.) describes the vena lateralis as receiving the vena dorsalis and anastomosing with the vena ventralis, which has already been cited as emptying into the vena caudalis. In a footnote on p. 317, Mayer states that in a section through an injected specimen of 8. canicula he saw a minute connection between the lateral and the caudal veins. This communication is recorded as l3eing caudad of the anastomosis of the laterals with the ventral.

Hyrtl {op. cit., p. 233 and Fig. 7) represents the lateral lymphatic in various Teleosts as collecting numerous paired cross vessels, which receive a superficial network. On pp. 234-5, he states that each lateral trunk terminates in a caudal sinus (PL X, Figs. 1, 2 and 4, a) immediately behind the last vertebra. A longitudinal neural and a h^mal trunk culminate in one of these sinuses, and both of these sinuses empty into the caudal vein. Besides terminating in a caudal sinus, each lateral trunk in Salmo is described by Vogi^" (pp. 135-6 and PI. K, Figs. 3-5, 66) as continuing to the base of the tail, where it separates into a dorsal and a ventral sinus, each of which communicates on the opposite side with a similar sinus. Both of the caudal sinuses open into the caudal vein and the orifice is said to be guarded by a valve.

^"Vogt, C. "Anatomie des Salomes." Memoires de la societe des sciences naturellcs dc Xetichatel. 1846.


63 William F. Allen

Yogi found no transverse branches uniting with the lateral trunk, and he was of the opinion that such vessels were onty extravasations of the injecting mass. Trois' description of the caudal ending of the lateral trunk in Lophius (op. cit., p. 5) and that of Sappej' for the pike and carp (op. cit., pp. 46-7) are very similar to what Hyrtl found for the perch.

Concerning the Ganoids, Hopkins (op. cit., p. 373) notes that the lateral lymphatic trunk in Amiatus (Fig. 11, 1.) extends candad as far as the posterior end of the dorsal fin, where it suddenly bends at right angles toward the meson, to terminate in a caudal sinus, which lies beneath the vertebral column and empties into the caudal vein. Just as the lateral lymphatic trunk bends to empty into the caudal sinus it receives a connecting branch (Fig. 11, t) from the tail portion of the ventral trunk (o), and immediately before uniting with the caudal sinus it is joined by the dorsal trunk, or, as represented in Fig. 11, a fork of the dorsal (r).

In Lepisosteus there is but one lateral subcutaneous trunl; (Figs. 1-5, 8, 10-18 and 25, L. T.) on each side of the body. It occupies a like position to a homologous canal of other fishes, which is in a median lateral line, in a sheath of connective tissue, just within the skin. The anterior connections of the lateral trunk in Lepisosteus have been given in a previous paper (op. cit., p. 113). As with other fishes, it collects numerous paired intermuscular or transverse hranclies (Figs. 1 and 3, Intm. C), which travel along superficially in the connective tissue, joining two myotomes, and gather a network of vessels from the connective tissue that binds the skin to the trunk muscle. Doubtless these vessels are continued dorsad and ventrad to anastomose with the dorsal and ventral trimks as in Scorpaniclithys and other bony fishes, but this point remains undetermined. Wlien about half way between the posterior end of the dorsal fin and the base of the tail, the lateral trunk bends mesad, at right angles, to culminate in, and help form, what has been designated and described as sinus (x). Ordinarily, as shown in Figs. 1, 3, 4, 5 and 6, the lateral trunk joins the haemal in forming sinus (x), but in Fig. 3, which is the opposite side of the same specimen as Fig. 1, the lateral trunk did not join sinus (x) until after it collected the dorsal trunk. In fact, the point of union was considerably behind that of the dorsal trunk, being almost opposite the opening of sinus (x) into the caudal sinus. A description of the union of the htemal and the dorsal trunks with sinus (x) and the culmination of the latter in the caudal sinus has already been given under a separate paragraph.


Siibcntaneons Vessels in Tail of Lepisosteus 63

The following observations have been made from a series of transverse sections through the tail region of a 90 mm. L. osseus: As shown by Fig. 10, c, there is a conspicuous connecting vessel uniting the right lateral with the right haemal trunk. This communication, which is cephalad of the origin of sinus (x), was not noticed on the opposite side of this series or in any of the gross dissections. In Fig. 13, the left lateral trunk is seen in section uniting with the left sinus (x). Fig. 21 is drawn from a portion of the same section as Fig. 10, showing the right lateral trunk and its communication (c) with the right haemal trunk, greatly magnified. From this diagram it will be seen that the lateral trunk of a 90 mm. L. osseus lies imbedded in a mass of connective tissue directly mesad of the ramus lateralis vagi. Both the lateral trunk and its communicating vessel (c) are composed of a single layer of endothelium, and at the point of anastomosis is a mass of plasma and corpuscles; of these corpuscles the red greatly outnumber the white. The communicating vessel (c) is joined to the body myo'tomes by a very dense layer of fibrous connective tissue (F. Con. T.).

Fig. 25 is from part of a section taken through the lateral trunk of an adult L. osseus as seen with a high-power objective. When compared with Fig 21, it will be found to be almost identical in structure. Internally it consists of a layer of endothelium (End.), which is attached externally to a layer of fibrous connective tissue (F. Con. T.), and, as was the case with the 90 mm. specimen, the red corpuscles greatly predominate.

In Lepisosteus the lateral trunk has a most remarkable size. Its diameter exceeds that of the caudal vein plus the caudal artery, and it maintains this caliber throughout its entire lengih. Since it is in connection with veins at either end, the flow of lymph or blood, whichever it may be, can pass in either direction. The resistance should be about the same, unless the movement of the tail favored the forward movement.

Except for its caudal termination, the distribution of the lateral trunk in Lepisosteus is not notably different from other fishes, especially Amiatus, but instead of emptying into the caudal sinus after making its posterior-mesal bend as in Amiatus, it first empties into what has been designated as sinus (x), which has already been described as also receiving the haemal trunk, and sometimes the dorsal trunk, before culminating in the caudal sinus. In Lepisosteus there is never any connection between the lateral trunk and the posterior part of the


G4 William F. Allen

ventral or the caudal trunk as Hopkins describes for Amiatus; on the other hand, the caudal trunk empties directly into one of the caudal sinuses.

Hcemal Trunks (Figs. 2-4, 10-12 and 20, Hx. T. or E. and L. Hse. T.). — In Lepisosteiis two such longitudinal trunks traverse the hgemal canal. Their position is perhaps best portra3-ed in a transverse section, as, for example, in Fig. 10; here, on either side of the caudal artery, a little above the level of the caudal vein, are the right and left haemal trunks (E. and L. Hse. T.), A comparison of Figs. 20 with 22, shows us that in a 90 mm. L. osseus the size and structure of the caudal vein and the hgemal trunks are almost identical. Both consist of a single laj-er of endothelium (End.), and contain but few corpuscles, while the caudal artery (C. A.) had an additional muscular layer, and was filled with corpuscles. As previously stated, in the specimen from which Fig. 10 was taken, there was a connecting trunk (c) between the right lateral and the right ha3mal trunks. This communication was not observed on the opposite side or in any of the gross dissections. Its position in this specimen is some little distance cephalad of the point where the haemal trunk leaves the haemal canal to empty into sinus (x). Ordinarily, as shown in Figs. 2 and 3, the hasmal trunks leave the hsemal canal opposite the point where the lateral trunks bend mesad, and the two unite on the ventro-lateral surface of the vertebral column in what has been described as sinus (x) ; but in some instances, as in Fig. 1, which is the opposite side of the same specimen as Fig. 2. and on both sides of the 90 mm. L. osseus series, they leave the hasmal canal first, and form or empty into sinus (x), before the lateral trunk bends inward to join sinus (x). In Lepisosteus no hernial or intercostal vessels were found connecting the haemal with the ventral trunks, as is the case in the Teleosts. Since the lymphatics of the viscera have not been studied, the anterior termination of the hsemal trunks has not been traced out, but the natural supposition is that they would end in an abdominal sinus, situated either below the kidney or between the kidney and the vertebral column.

With the Teleosts, Trois (op. cH., p. 11 and Fig. 4, A.) finds in LopJiius, that what he terms as the ironclii linfatici spinall inferiori travels along in the haemal canal and collects the intercostal vessels. In Fig. 4 Trois portrays this trunk as anastomosing at the base of the tail with the superior spinal, the dorsal and the ventral lymphatic trunks, Le tronc sous-vertebral of the pike is briefly set forth by Sappey {op.


Subcutaneous Vessels in Tail of Lepisosteus (55

cit., p. 49) as following along in the same canal with the caudal artery and vein. A similar vessel is also noted by Sappey (p. 50) for Pleuronectes. In PI. XII, Sappey represents this trunk in the pike (Fig. II, 10) and in Pleuronectes (Fig. TV, 17) as receiving numerous hsemal or intercostal vessels, but its posterior ending is not given; from his figures, however, it appears to end blindly before the base of the tail is reached.

In addition to the caudal artery and vein, Mayer (op cit., p. 320) finds a longitudinal trunk in the haemal canal of Scyllium, Mustelus and Squatina, Avhich he takes to be a vein that collects a vasa vasorum, originating from little paired arteries that leave the caudal artery between the intercostal arteries (PI. 17, Fig. 11, avas.). Again, further on (pp. 325-6), Mayer states that the caudal vein is very irregular, changing from side to side, becoming paired, then unpaired, to finally disappear altogether, and its place is taken by a vasa vasorum (haemal trunks). After which there is nothing in the haemal canal but the caudal artery in the center, surrounded by connective tissue and small blood spaces. Mayer conjectures that possibly in the past these blood cavities had a different function, that they may have been derived from a degeneration of the body cavity. Hopkins does not record any such trunk for Amiatus.

Caudal Sinuses. — These sinuses have received considerable attention from all workers on the subcutaneous system of fishes. Naturally thev fall into two classes: pulsating hearts resembling the lymphatic hearts of the Batrachians, and non-pulsating sinuses. In some fishes or fishlike vertebrates that swim by a snake-like movement, we find a pulsating heart in the tail; while in other fishes which swim by a lateral movement of the tail, the caudal sinuses are simply reservoirs, and strange to say with the Selachians no such receptacle has been found; in this group the subcutaneous trunks empty directly into the caudal vein.

Greene^^ and Klinckowstrom^^ found two pulsating caudal hearts in the tails of Polistotrema {^= Bdellostoma) and Myxine. Each of these hearts, which are separated from one another by the median caudal

"Greene, C. W. "Contributions to tlie Physiology of the California Hagflsh, Polistotrema stouti : 1. The Anatomy and Physiology of the Caudal Heart," American Journal of Physiology, Vol. III. 1900.

'-Klinckowstrom. A. Title not known. TerJiandlungen dcs hiologischeii rcreins in StocJchohu, 1890 and 1891.


6G William F. Allen

cartilage, open dorso-eeplialad into the caudal vein, the orifice being guarded by a valve. In the ventro-cephalic corner there is a second opening, also guarded b}- a valve, through which the subcutaneous lymphatic or blood canals empty. According to Greene, the caudal hearts in the hagfish are not of themselves muscular or contractile, but are filled and emptied by a contraction of the musculi cordis caudalis, which lies laterad of the heart and presses it against the median cartilage. These muscles are said to be entirely separated from the heart, and are not to be regarded as part of it. In the hagfish, then, the caudal hearts or sacs receive the blood or lymph from the subcutaneous spaces and drive it into the caudal vein.

By far the best description of the caudal heart of the eel is given by Jones.^^ From pp. 676-9 and Figs. 1 and 3, the caudal heart is represented as separating, caudad, into a small dorsal branch (D) and a larger ventral branch (C). Between these two forks, or, to be more exact, between the ventral branch below and the caudal artery and dorsal branch above, there is a pulsating heart (E), which communicates anteriorly with the dorsal fork of the caudal vein, a short distance from its union with the ventral fork. A valve is said to guard the entrance of the heart into the vein. These observations of Jones', which were microscopic, were made upon the tail of a small eel that had been placed upon a thin plate of glass. Upon contracting, as is graphically shown in Fig. 2, a colorless stream of lymph was seen to enter the vein. Strange to say, Jones found no definite afferent lymphatic trunks emptying into the caudal heart, and contends that such lymphatic canals as were described by Miiller could not have escaped his attention. Furthermore, Jones points out that since the lymphatic vessels of Miiller are identical in position with the caudal blood vessels, that they must be such. On p. 679, Jones gives the credit of the discovery of the caudal heart in the eel to Hall in 1831, but it seems that Hall regarded it as a blood heart. Continuing, Jones insists that the relation of the blood vessels to the heart as set forth by Hall is incorrect, as Avill be seen by comparing Hall's figure (see p. 679) to Fig. 1. Likewise Jossifoo" overlooked the connection of the lymphatic system with the

"Jones, T. W. "The Caudal Heart of the Eel a Lymphatic Heart," Phil. Trans., 1868.

"Jossifoo, S. M. "Snr les voles prlncipales et les organes de propulsion de la lymphe chez certains Poissons," Archives d'anatomie microscopiqnc. 1906.


Subcutaneous Vessels in Tail of Lepisosteus C7

caudal heart in Anguilla. He failed to inject the caudal heart from the caudal vein, but states that when a pulsating caudal heart is seen through a microscope, it is colorless, except a little rose red, due to the adjacent blood capillaries, and that it is in great contrast to the connecting and surrounding blood vessels. Some year ago I injected some tails of the Mississippi Eiver Anguilla from the dorsal and the ventral subcutaneous trunks, and hope in a later study to be able to demonstrate their connections with the caudal heart.

In many Teleosts Hyrtl (op. cit., pp. 226-231 and 238, and Figs. 1, 2 and 4 a) finds two caudal sinuses situated immediately behind the last vertebra. These sinuses (see Fig. 4) are connected and both communicate anteriorly with the caudal vein, the orifice being guarded by a valve. The shapes of these sinuses are at variance in different species. According to Hyrtl, the caudal sinus is composed internally of a layer of endothelium, bounded by a layer of longitudinal fibers, outside of which there is an additional layer of transverse fibers. It is described as being contractile and comparable to the lymphatic hearts of the Batrachians. Furthermore, Hyrtl claims that the caudal sinus is no blood reservoir, for its serum is clear, and contains small, round, granular corpuscles. Vogt's description of the caudal sinus of Salmo (op. cit., pp. 135-6, and PI. K, Figs. 3-5, 54) is almost identical to HyrtFs for the perch. He finds a valve in the opening between the two sinuses, and each sinus is said to contain a few muscle fibers, is contractile, and contains a clear fluid in which there are a few granular corpuscle;]. Sappey (op. cit., p. 47, and PL XI, Fig. VI, 6, and PI. XII, Fig. II, 5) speaks of the caudal lymphatic sinuses of Teleosts as being non-contractile papillse, which are non-pulsating in the sense of the caudal hearts of the eel and Batrachians, and he was unable to find valves at tlie entrances of these sinuses into the caudal vein.

As for the Ganoids, Hopkins (op. cit., p. 372, and Fig. 11, s.) represents each caudal sinus in a 53 cm. Amiatus as being a reservoir about 1 cm. long by 3 to 5 mm. at its greatest width, situated ventrad of the last vertebrffi. It is said to empty cephalad into the caudal vein, the entrance being guarded by a valve, and to communicate mesad with its fellow sinus. As has already been quoted, the lateral lymphatic trunk, after receiving the caudal trunk and a fork of the dorsal, joins the caudal sinus from the side, near its anterior end.

The caudal sinuses of Lepisosteus (Figs. 1-6, 16 and 17, L. and K. Cau. S.) have almost exactly the same position as in Amiatus, and


G8 William F. Allen

about the same as in the Teleosts.'^^ Like Aniiatus, the caudal sinuses of Lepisosteus are situated ventrad and to the side of the last vertebrge. They do not lie in a longitudinal plane, but are tilted a little obliquely ventrad, following the general contour of the vertebral column in that region. Usually they cross the vertebral bases of the eighth and ninth or the ninth and tenth caudal hgemal spines (counting ventro-dorsad). These reservoirs are somewhat elongated, ordinarily deeper anteriorly than posteriorly, but in some instances, as shown in Figs. 1 and 4, they have little the appearance of sinuses, but rather maintain the same diameter throughout, which is slightly, if any, greater than the adjacent subcutaneous trunks. The contents of these sinuses was not examined, nor were they sectioned, save in a 90 mm. L. osseus and a few smaller specimens, where they were found to be composed of a single layer of endothelium, not unlike the longitudinal subcutaneous trunks or the caudal vein, and contained, especially, at their posterior ends and their junction with the caudal vein, a mucous or plasma-like substance, in which there were a few red and white corpuscles.

Considerable has already been said concerning the various openings of the caudal sinus. There is always a communication between the two sinuses (Figs. 1-7, o), wliich passes between the eighth and ninth or the ninth and tenth caudal haemal spines (counting ventro-dorsad). The aperture of this connection is through the mesal wall of the sinus, at about its center near the floor, but in the specimen from which Figs. 1 and 2 were drawn it was nearer the roof. N"o valves were found guarding this orifice, and in Fig. 17 (o) this communication is shown in cross section. One or the other of the caudal sinuses, more often the right, receives the posterior continuation of the ventral trunk or what has been described as the posterior trunk (Figs. 1-5 and 7, C. T.) from the rear. As has already been stated, each of the sinuses (x) (Figs. 1-6) joins its respective caudal sinus at the dorso-cephalic corner; while the opening through the ventro-cephalic corner leads into the caudal vein (see Figs. 1-7 and 15). The venous opening of the caudal sinus that receives the caudal trunk is always much larger than the other, as is shown in Fig. 7. Unlike the other orifices of the caudal sinus, the

"In both Lepisosteus aud Amiatus we have a masked heterocercal tail, which is more primitive than the tails of bony fish in that the caudal haemal and iuterhre-mal spines, although fused to each other, have not fused together in forming two hypural Imnes, as is the case in the Teleosts.


Subcutaneous Tessels in Tail of Lepisosteus 6',)

venous opening is guarded b}' a pair of semi-lunar rahrs (Figs. 4, 5 and 7, E. and L. Cau. S. V.), which open into the vein.

That part of the caudal vein which receives the caudal sinuses is enlarged almost into a reservoir, which in addition receives the last neural vein and the posterior part of the caudal vein from the tail.

With Lepisosteus the term caudal sinus is nothing more than an arbitrar}^ term applied to two non-contractile reservoirs situated ventrad of the last vertebrge, near the base of the tail, which collect four longitudinal subcutaneous trunks, together with two longitudinal haemal trunks, and empty into the caudal vein.


Summary and General Considerations.

In the tail region of Lepisosteus there are four longitudinal subcutaneous vessels and two profundus trunks in the haemal canal, which collect a subcutaneous network that so far as could be determined is entirely separated from the arteries or their capillaries, and which in one way or another terminate in two caudal sinuses that discharge themselves in the caudal vein.

The two caudal sinuses are elongated reservoirs situated ventrad of the posterior end of the vertebral column. They communicate mesad with each other, dorso-cephalad with what has been designated as sinus (x), and ventro-cephalad with the caudal vein, the latter orifice being guarded by a pair of semi-lunar valves.

Posteriorly one of the caudal sinuses, more often the right, receives what has been described as the caudal trunk, which is merely a prolongation of the ventral trunk passing through the basal canal of the caudal fin, where it runs parallel to a corresponding caudal artery and vein. From each ray. it receives two branches which traverse the dorsal and ventral surfaces of the ray and collect a rather coarse network of vessels from the fin membrane. This network is not continuous with arterials from the fin ray arteries. The caudal trunk in its course from the tail to the caudal sinus follows the caudal artery, and is in fact a deeper lying trunk than the caudal vein, which passes cephalad between the superficial and profundus muscles of the caudal fin.

The ventral trunk, of which the caudal is a posterior continuation, travels along the ventro-median line, just Avithin the skin. In passing through the basal canal of the anal fin it collects paired branches from each ray, which receive a network from the fin membrane similar to that found in the caudal fin.


70 William F. Allen

Each caudal sinus receives a most important communication through its dorso-cephalic wall, which has been described as sinus (x). These sinuses, which follow along the ventro-lateral surfaces of the preceding vertebrae, collect a lateral and a longitudinal haemal trunk, and one or the other of them, the dorsal trunk.

The dorsal subcutaneous trunk travels along the dorso-median line just below the skin, but when the dorsal fin is reached, instead of dividing into two circular canals at the insertion of the dorsal fin, as in the Selachians and most Teleosts, it passes completely through the basal canal of the fin, and, like the caudal trunk, receives a pair of canals from each ray. These canals drain a network of vessels which is decidedly lymphatic in the character of its meshes, and which, so far as I am aware, has no connection with the arteries. "When about equidistant from the posterior insertion of the dorsal and the base of the caudal fin it makes a ventral bend to cross the vertebral column, usually the right side, and culminates in the corresponding sinus (x), differing considerably in its termination from Hopkin's description for Amiatus, where it bifurcated, each fork uniting with the lateral trunk, immediately before the latter joined the caudal sinus.

As has been pointed out, the venous supply from the dorsal fin is collected by a pair of dorsal fin ray veins from each ray. In position these veins and the corresponding arteries are nearer the surfaces of the rays than the dorsal fin ray subcutaneous canal. In the basal canal of the fin they unite to form a dorsal fin vein, which passes through the basal canal in company with the dorsal subcutaneous trunk and the dorsal fin artery. Upon leaving this canal the dorsal fin vein in the two specimens in which it Avas traced out crossed the left side of the vertebral column to terminate in the caudal vein. The distribution of this vein recalls the vena postica of Mayer and Parker for the Selachians, and had it anastomosed with the dorsal subcutaneous trunk at the posterior border of the dorsal fin we would have had a condition of affairs almost identical with that found in the Selachians.

With Lepisosteus the lateral subcutaneous trunk offers few peculiarities not found in Amiatus or the Teleosts. In the Selachians and in Polyodon they are less sinus-like than in the bony Ganoids and in the Teleosts. In Lepisosteus, as in other fishes, they collect numerous paired transverse or intermuscular l)ranches, which receive a rather coarse network of branches from the connective tissue binding the skin to the m^'otomes. These transverse branches arc undoubtedly prolonged dorsad


fSubcutaneous Vessels iu Tail of Lepisosteus 71

and ventrad to anastomose with the dorsal and ventral trunks as in ScorpcBnichtliys and in other fishes, but this point was not determined for a certainty in Lepisosteus. Instead of emptying directly into the caudal sinus as in Amiatus the lateral trunk in Lepisosteus unites with a haemal trunk in forming sinus (x), which terminates in the caudal sinus. In a 90 mm. L. osseus series a connecting canal Avas found joining the right lateral trunk with the right haemal trunk. This communication was in advance of the union of the lateral with the haemal in forming sinus (x). It was not observed on the opposite side of this series or in any of the dissections.

Two longitudinal hsemal trunks were found above the caudal vein on either side of the caudal artery in the hgemal canal of Lepisosteus. They are undoulitedly homologous to what Mayer has described in the Selachians as vasa vasorum and to a similar trunk in the hgemal canal of the Teleosts. The cephalic distribution of this trunk was not traced out. No hgemal vessels were found connecting it with the ventral trunk as in the Teleosts, and, as stated above, it emptied into sinus (x).

In Lepisosteus no longitudinal neural trunk was found in the neural canal, as is the case with many Teleosts, and consequently there are no neural vessels to communicate above with the dorsal trunk.

Microscopic sections of a 90 mm. L. osseus showed the subcutaneous trunks, the caudal sinus and the caudal vein to be composed of a single layer of endothelium, surrounded by a mass of connective tissue. They contained but few corpuscles, the red predominating over the white.

No valves were found in the subcutaneous system of the tail region of Lepisosteus. save two semi-lunar valves guarding the entrance of each caudal sinus into the caudal vein.

Resume. — From the above summary the subcutaneous system of Lepisosteus should fall under the head of a lymphatic system, or a separate subcutaneous venous system that has no counterpart in the arterial system and which may function both for veins and lymphatics.

In favor of the hypothesis that this system in the tail region of Lepisosteus are veins, we find from microscopic sections of a 90 mm. L. osseus that the structure of the subcutaneous vessels and the caudal vein are almost identical, and while they contain but few corpuscles, yet the red predominate.

As opposed to this supposition and in favor of the hypothesis that they are lymphatics, this study has revealed the subcutaneous system of the tail region of Lepisosteus to be entirely separate from the blood


72 ^Yilliam F. Allen

vascular system, save at the points where the caudal sinuses empty into the caudal vein. This system of vessels collects a network which is decidedly lymphatic in the character of its meshes, is coarser than the blood capillaries, and so far as observed had no connection with the arteries. Furthermore, the peripheral regions are sufficiently supplied with veins. For most of the smaller subcutaneous vessels are accompanied by corresponding arterial and venous branches; the lateral arteries and veins supply the peripheral region of the trunk, the dorsal fin artery and vein nourish the dorsal fin, and the caudal artery and vein do the same for the caudal fin. If the subcutaneous vessels are classed as veins it would be necessary to consider them as a distinct venous system that had no counterpart in the arterial system.

We are compelled to admit that the evidence is insufficient to warrant anv sweeping statement as to the exact nature of this system of vessels in fishes. The little data we have, when considered in the light of certain recent studies on the embryology of the lymphatics in mammals, supports the supposition that, in the more primitive Selachians, certain subcutaneous vessels, probably veins, have become separated to some extent from the main venous system. , According to Sappey, in the skates the communications of this system with the veins are quite numerous. In Mustehis and Squahis, Parker and Mayer found these points of union less abundant, but in addition to the connections of the subcutaneous vessels with the caudal vein in the tail region of the Ganoids and Teleosts, they note that the so-called dorsal cutaneous vein anastomoses behind the dorsal fin with a deep dorsal fin vein. Such a vein was found in Lepisosteus, but no anatomosis Avith the dorsal subcutaneous trunk occurred, indicating, of course, that in Lepisosteus the separation of this subcutaneous system had become more complete.

In the head region we find that the same differentiation of the subcutaneous system has gone on as we pass from the Selachians to the Ganoids, and from the Ganoids to the Teleosts and Batrachians, but in this region it is more obscure. In an earlier paper it was pointed out that each of the branchial lymphatic trunks (nutrient branchial veins?) of Poh/odon anastomosed above with the subcutaneous system and below with the inferior jugular vein ; that they received a coarse network from the branchial arches and from their filaments, which, so far as could be ascertained, had no capillary connections with the arterial system. In Lepisosteus these branchial trunks were separated into dorsal and ventral l:)ranchial lymphatic trunks (nutrient branchial veins?), whicJi


Subcutaneous Vessels in Tail of Lepisosteus 73

were not connected, and which drained their respective halves of the arches. Those from the ventral portion of the arch emptied into the inferior jugular, while those from the dorsal portion terminated in dorsal branchial sinuses, Avhich were in communication with the subcutaneous system and with the jugular through the cephalic sinns. Similar dorsal branchial lymphatic sinuses were described in Scorpcenichthys as being in communication with the subcutaneous system and with the jugular through the cephalic sinus. No branchial lymphatic trunks were seen emptying into them, but both dorsal and ventral nutrient branchial veins were shown uniting directly with the jugular and the inferior jugular veins. It would seem that the only explanation of the above complicated condition of affairs is that a part of the subcutaneous system found in the region of the gills has become entirely separated and has reverted back to veins in the higher orders of fishes.

In the Teleosts there is a further differentiation or rather addition. Here we find that the rudimentary haemal trunk in the hsemal canal of the Selachians and Ganoids has developed into a conspicuous trunk with numerous hsmal branches that communicate ventrad' with the ventral trunk. There is also in Scorpceniclithys and in many Teleosts a large and important neural longitudinal trunk, which traverses the neural canal above the spinal cord and sends off numerous anastomosing branches to the dorsal trunk.

In conclusion it may be said that considerable anatomical data support the hypothesis that the subcutaneous vessels of the Teleosts and Batrachians, which are evidently lymphatics, have their homologue in the somewhat similar system of the Selachians, which has much the ap^warance of veins. The subcutaneous system of the Ganoids is apparently a sort of intermediary ; that of Poljjodon. one of the cartilaginous Ganoids, resembling the arrangement of the Selachians, and that of Lepisosteus, one of the bony Ganoids, approaching the system of the Teleosts and Batrachians.


LIST OF ABBREVIATIONS USED IN THE FIGURES. A. or P. prefixed to an abbreviation signifies anterior or posterior ; R. or L., right or left ; a series is numbered from cephalad to caudad. A. R., anal fin rays.

A. R. C, anal ray subcutaneous canals. A. R. M., anal ray levator and depressor muscles. C, in Fig. 10, communication between the lateral and haemal trunks.


74 William F. Allen

C. A., caudal artery. Cau. S., caudal sinus. C. F., caudal fin. Con. T., connective tissue. C. P. M., caudal fin profundus muscles. C. R., caudal rays. C. R. A., caudal ray arteries. C. R. C, caudal ray subcutaneous canals. C. R. v., caudal ray veins.

C. T., caudal subcutaneous trunk or posterior continuation of the ventral subcutaneous trunk.

C. v., caudal vein. Der., dermis.

D, F. A., dorsal fin artery. D. F. v., dorsal fin vein. D. R., dorsal fin rays.

D. R. A., dorsal ray arteries.

D. R. C, dorsal ray subcutaneous canals.

D. R. M., dorsal ray levator and depressor muscles.

D. R. v., dorsal ray veins.

D. T., dorsal subcutaneous trunk.

End., endothelium.

F. Con. T., fibrous connective tissue.

Hse. C, hsemal cauals.

Hge. S. (1) to (9), caudal hremal spines. Hypural bones of Teleosts.

Hse. T., hsemal longitudinal trunk.

I. Neu. S., interneural spines.

Intm. C, intei'muscular or transverse subcutaneous canals.

L. A., lateral arteries.

L. C. A., left caudal artery.

L. Cau. S., left caudal sinus.

L. Cau. S. O., left caudal sinus opening into the caudal vein.

L. Cau. S. v., semi-lunar valves guarding entrance of the left caudal sinus

into the caudal vein. L. Hae. T., left longitudinal h;Temal trunk. L. T., lateral subcutaneous trunks. L. v., lateral veins. My., myelon or spinal cord. Myo., myotomes of the great lateral muscles. Neu. S., neural spines. Neu. v., neural veins. Neu. V. (1), last neural vein.

Neu. V. (1) O., opening of the last neiu'al vein into the caudal vein. 0., connection or communication between the two caudal sinuses. Pig., pigment. PL, blood plasma.


Subcutaneous Vessels in Tail of Lepisosteus 75

K. C, red corpuscles.

R. C. A., right caudal artery.

R. C. S., right caudal sinus.

R. Cau. S. O., right caudal sinus opening into the caudal vein.

R. Cau. S. v., semi-lunar valves guarding the entrance of the right caudal

sinus into the caudal vein. R. Hge. T.. right longitudinal hfemal trunk. R. Lat. X.. ramus lateralis vagi. Uro., urostyle.

Ver., vertebral column or centrum. y. T., ventral subcutaneous trunlv. Wi. C, white corpuscles or leucocytes. X., a sinus in Lepisosteus formed by the union of a lateral, hfemal and

often the dorsal trunk, which terminates in one of the caudal sinuses.


DESCRIPTION OF THE FIGURES.

Figs. 1 to 9 were drawn to a scale from dissections of injected specimens ; 10-24 are from a transverse series tbrougJi the tail region of a 90 mm. Lepisosteus osseus; and Fig. 25 Is taken from a cross section of the lateral trunk from an adult L. osseus. In general, the subcutaneous or so-called lymphatic canals are colored yellow or drawn in outline, the arteries are stippled, and the veins cross-barred. A vessel drawn in dotted outline signifies that it passes within or behind a muscle, bone or other vessel. All outlines for the microscopic drawings were made with the aid of a camera lucida and the details were filled in afterward.

Fig. 1. Represents a lateral dissection of the tail region of a small Lepisosteus tristcechus as seen from the left side. The myotomes bordering the tail are removed in order to best portray the lateral trunk emptying into the left caudal sinus, the connection of the latter with the corresponding right sinus, and the termination of the left caudal sinus in the caudal vein, together with the origin and course of the vein. The outline for this drawing was traced from a photograph and the details were filled in afterward. X %•

Fig. 2. Is a similar dissection from the opposite or right side of the same specimen as Fig. 1. On this side the dorsal, hsemal and lateral trunks fuse to form sinus (x), which with the caudal trunk terminates in the right caudal sinus, and the latter empties into the caudal vein. The distribution of the caudal artery is also shown, and, like Fig. 1, the dissection was first photographed and the details were filled in afterward. X %•


SUBCUTANEOUS VESSELS IN TAIL OF LEPISOSTEUS

WILLIAM F. ALLEN


DT X • ^*^'


IntmC LJ. : -DRn


DRC DR



CF. 1


0. 'on



Fig. 7. Dorsal view of the caudal sinuses of a small L. osseiis; the dorsal walls of each being removed to show their median communication and the semi-lunar valves guarding their exit into the caudal vein. Natural size.

Fig. S. Shows distribution of a lateral artery and vein as seen from the left side of a small L. tristoechiis. In this dissection the great lateral trunk was removed, but its position is indicated by dotted lines, x %•

Fig. 9. The distribution of the dorsal fin vein and artery as seen from the right side of a small L. tristoecJius. All of the musculature in this region is completely removed. In another specimen where these vessels were traced out the artery took its origin seven or eight vertebrae cephalad of its position here, and approached the basal canal of the dorsal from in front, x ^.'■2 Fig. 10. From a transverse section through the caudal peduncle region of a 90 mm. L. osseus about midway between the dorsal and the caudal fins as seen from the rear (caudad) ; showing the caudal artery and vein, together with the dorsal, ventral, lateral and hsemal trunks in section. Note the connection between the right lateral and the right hfemal trunks. X ^■

Fig. 11. A portion of a transverse section .07 mm. caudad of Fig. 10. Here the caudal artery and vein, together with the right hremal and the two sinuses (x), are seen in section. In a section .04 mm. cephalad the right sinus (x) takes its origin blindly, and the left haemal trunk is leaving the haemal canal to become the left sinus (x). x 9


SUBCUTANEOUS VESSELS IN TAIL OF LEPISOSTEUS

WILLIAM F. ALLEN




Fig. 12. Transverse section taken .35 mm. caudad of Fig. 11 as seen from the rear (caudad). Left lateral trunk emptying into the left sinus (x), and the dorsal trunk crossing the left side of the vertebral column preparatory to emptying into the left sinus (x). The right haemal trunk terminates in the right sinus (x) .06 mm. caudad of this section, while the right lateral trunk does not join the right sinus (x) until a point .45 mm. caudad of this section is reached. X ^■

Fig. 13. Is from a section .22 mm. caudad of Fig. 12 viewed from the rear (caudad) ; dorsal trunk terminating in the left sinus (x) ; caudal artery and vein, ventral trunk and sinuses (x) seen in cross section, x 9.

Fig. 14. A transverse section .05 mm. caudad of Fig. 13 as seen from the rear (caudad) ; last neural vein crossing the left sinus (x) and emptying into the caudal vein. The caudal artery has bifurcated, and its two forks, together with the caudal vein, the ventral trunk and the two sinuses (x) are seen in section. X 9.

Fig. 15. Taken from a transverse section 23 mm. caudad of Fig. 14 examined from the rear (caudad). The two caudal sinuses «'ve seen In section emptying into the caudal vein. Both caudal arteries, both sinuses (x), the last neural vein and the ventral trunk are also shown in section, x '•^■

Fig. 16. From a transverse section .43 mm. caudad of Fig. 15 viewed from the rear (caudad). The left sinus (x) is seen in section terminating in the left caudal sinus, but the right sinus (x) does not empty into the right caudal sinus till a section .09 mm. caudad is reached. The caudal vein is bearing off to the left. Both caudal arteries, the right sinus (x) and the ventral trunk are seen in section, x 9 Fig. 17. A transverse section .67 mm. caudad of Fig. 16 as seen from the rear (caudad). The right caudal sinus communicates mesad with the left through the canal (o). The two arteries and caudal vein are seen in section, and the ventral trunk is traveling dorsad in the basal canal of the caudal fin to become the caudal trunk (C. T.). x 9


SUBCUTANEOUS VESSELS IN TAIL OF LEPISOSTEUS

WILLIAM F. ALLEN


DT



Ver


X

/VeuVw


C./|.


u^


RC.S.


l.Ciu.S


cv



LC./\ L Co U.S.



Fig. 18. Taken from a transverse section .47 mm. caudad of Fig. 16 as seen from the rear (caudad). Continuation of the right caudal sinus is observed in section a little below the vertebral column as the caudal trunlc (C. T.). The fin portion of the caudal trunk (C. T.), or the continuation of the ventral trunk, is becoming more and more dorsal in the basal canal of the caudal fin, and branches to it from the fin (C. R. C.) are seen in section, as are also the corresponding arteries and veins. Both caudal arteries and the caudal vein retain their former positions. X 9.

Fig. 19. Transverse section taken through the caudal fin 1.18 mm. caudad of Fig. 18 as viewed from the rear (caudad). Caudal vein and caudal trunk will be found passing through the basal canal of the fin; the caudal ray vessels are seen in section ; while the caudal artery has not yet left for the basal canal of the fin. x 9 Fig. 20. One of the hremal trunks from a cross section through the caudal peduncle of a 90 mm. L. osseus. Like the caudal vein, it is composed of a single layer of endothelium, and no corpuscles of any kind were found for a distance of several millimeters, x 225.

Fig. 21. From a transverse section through one of the lateral trunks of a 90 mm. L. osseus as it sends inward a connecting vessel (C) to join the haemal trunk. Note that the walls of both the lateral and the connecting vessels, which are composed solely of endothelium, are bounded by a fibrous connective tissue. Also that the red corpuscles greatly outnumber the white. X 225.

Fig. 22. A portion of a transverse section through a 90 mm. L. osseus, showing the caudal artery and upper part of the caudal vein imbedded in a mass of loose connective tissue within the hpemal canal. Corpuscles are very abundant in the artery, while in the vein they are very scarce. Note that the caudal veins consist of but a single layer of endothelium, which is almost Identical with the hirnial tnmks. X 225.


SUBCUTANEOUS VESSELS3IN TAIL OF LEPISOSTEUS

WILLIAM F. ALLEN


Ber.


\^ RUtX



Fig. 23. From a transverse section through the caudal trunk of a 90 mm. L. osseus. No corpuscles of any kind were found within this trunk throughout this series of sections. X 225.

Fig. 24. Transverse section through the caudal ray vessels of a 90 mm. L. ossetis. X -'-5.

Fig. 25. Is from a transverse section through the lateral trunk of an adult L. osseus. Note that the endothelial canal is bounded by a fibrous connective tissue and that the red corpuscles greatly predominate over the white. X 225.


Taussig FJ. The development of the hymen. (1908) Amer. J Anat. 8: 89-108.

The Development Of The Hymen

Fred. J. Taussig, M.D. From the Anatomical Laboratory of Washington University, St. Louis.

With 14 Figures.

Observations thus far collected concerning the origin and development of the hymen may broadly be divided into clinical and embryological. The clinical evidence is based on the study in adult life of the congenital anomalies such as hymen duplex with double vagina, hymen with absent vagina, etc. The interpretation of these anomalies is very difficult, and their value in an embryological study is really only of a confirmatory nature. Consideration of them alone can never result in a solution of the problem of hymeneal development. The embryological evidence on the development of the hymen is based on gross anatomical dissections, single microscopic sections, and on serial sections of the hymen and its surrounding structures. Accurate conclusions must be based on a correct valuation of these three sorts of evidence. For the proper study of so minute a structure as the fetal hymen, the last named method — serial sections — is of paraniount value.

An extensive review of the various opinions on hymeneal development has recently been made by Gellhorn ('04). From the standpoint of time, we may distinguish a convaginal theory, according to which the hymen is formed at the same time as the vagina, and a postvaginal theory, according to which the hymen is formed after the development of the vagina. It is, however, better to classify the various views from the standpoint of origin. They may be grouped under four heads :

(1) Vulvar Theorjr, Pozzi ('84).

(2) Bilamellate Vulvo-vaginal Theory, Schaeffer ('90).

(3) Uni-lamellate Vulvo-vaginal Theory, Nagel ('97). Budin ('79), Webster ('98), Klein ('94).

(4) Vaginal Theory, Dohrn ('75), Veit ('99), Gellhorn ('04).


(1) Vulvar Theory. Pozzi ('84) bases his theory wholly on the clinical findings in cases of malformation of the genital tract, above all on the presence of a hymen in the absence of the vagina, the occasional occurrence of a urethral hymen, and the presence of a single hymen in double vagina. Such anomalies he believes could only occur if the hymen developed from the vulva or sinus urogenitalis.

(2) Bilamellate Yulvo-vaginal Theory. In 53 out of 190 specimens of fetal hymens (28.8 per cent) Schaeffer ('90) was able to find a more or less distinct double hymen. The two folds were connected by bands of tissue, and according to his theory later coalesce to form the hymen. The one fold springs from the vulva, the other from the vagina. Thus we would have a four-layered hj^men, two layers from each fold.

(3) Uni-lamellate Vulvo-vaginal Theory. Budin ('79) explains this theory in the following way: The hymeneal ring is the outer end of the vagina. The latter opens into the sinus urogenitalis, at the same time pushing the walls of its canal outwards, just as the portio vaginalis uteri protrudes into the vagina. Thereupon an opening forms in the centre, but the peripheral ring-like protrusion remains, covered externally by the mucosa of the urogenital sinus, internally by the vaginal mucosa. Nagel's ('97) description (Fig. 9) differs only slightly from this. He says: "In embryos of the third month there is an increase and accumulation of the upper layers of the epithelium occurring at first just above the vaginal orifice, whereby the vagina becomes dilated at this point (in embryos of 7-10 cms. length). Through this dilatation arises the hymen. For since the edge of the original opening is not affected by the dilatation, the orifice, on the contrary, retaining its original narrowness, a ring must thereby be formed by which the vagina is shut off from the urogenital sinus. The opening of this ring, up to embryos of 20-22 cms. length, remains filled with epithelium."

According to this view, therefore, the hymen is made up of a single fold, one side of whieli is formed by the vagina, the other by the vulva.

(4) Vaginal Theory. Dohrns ('75) work supporting this theory is based on median sagittal sections through the pelvis of twenty-five fetuses from the ninth to the twenty-eighth week of development. N"o microscopic study Avas made. From the ninth to the fifteenth week he finds a stronger growth of the posterior wall of the vagina so that its canal becomes wider and bends more sharply forward. From the seventeenth to the nineteenth week there is a marked proliferation of the inner wail of the vagina, so that it seems made up of tooth-like projections. By the beginning of the nineteenth week the hymen is visible as a fold rising from the posterior wall of the vagina directly above the point of entrance of the vagina into the sinus urogenitalis. To meet this a shorter fold from the anterior wall grows downward. The two folds unite, leaving a semi-hmar opening. The growth of the hymen is very rapid. He continues :

"Der Umstand dass die Hymenalmembran in der Nahe der Stelle entsteht an welcher sich die Allantois imd Mueller'schen Gange in der Cloake begegnen, und der Sinus Urogenitalis abscheidet, hat wiederholt zu der Yermuthung geflihrt, dass der Hymen mit einem Entwickelungsgebilde der frliheren Zeit in Zusammenhang stiinde. Je genauer man aber die frliheren Entwickelungsstufen in ihrer Weiterbildung verfolgt, desto mehr wird man iiberzeugt, dass ein soleher Zusammenhang nicht vorliegt. Wir haben beim menschlichen Embryo einen langen Zeitraum, den Abschnitt von der 9-17ten Woche, in welchem wir den Mittelstufen zwischen Hymen und den an seiner Entwickelungsstelle friiher zusammengetrolfenen Gebilden nachspiiren konnen. Das Eesultat ist negativ. Der Hymen ist ledigiich eine spiitere Bildung, welche sich nicht in continuirlicher Fortenwickelung an friihere Formen anschliesst."

Gellhorn (•'04) also holds to the vaginal theor}^, although I believe ho does not sufficiently emphasize in his article the difference betAveen his conception and the above-mentioned vulvo-vaginal theory. Budin describes the vaginal bulbus as projecting into the urogenital sinus. Into this protruding conus, according to Gellhorn, the vaginal connective tissue grows, so that, with the possible exception of a thin layer of epithelium of the urogenital sinus on the outside, the entire hymen is of vaginal origin.

While it is thus seen that there is no lack of theories, their foundation is for the greater part the most meagre and inconclusive evidence. Only one man so far as I know studied serial sections of the hymen microscopically, and that one — Klein ('94)- — studied but a single case in this way.^ INTo one has thus far made use of serial sections in a number of embryos of various stages of development for an investigation of the

'I do not include in this consideration the two embryos tliat Klein cut in ti'ansverse serial section, hnt only the one sectioned sagittally. Transverse sections, imless some sort of reconstruction is made (and this was not done), are not favorable for a study of the hymen in its relations to neighboring structures, a point which Klein himself concedes.

hymen. It was this fact that induced me to study the five embryos at my disposal by this method. They were the following:

Embryo 1. 18 cms. long, well preserved. No abnoruialities of development as far as examined.

Embryo 2. 18 cms. long, slightly macerated, no abnormalities.

Embryo 3. 18 cms. long, well preserved, normal.

Embryo 4. 21 cms. long, removed by laparotomy, excellent preservation. No abnormalities in development.

Embryo 5. Six months gestation. Legs have been removed for another purpose, hence exact length could not be determined. Preserved in Mueller's fluid. No abnormalities of development.

The measurements of these embryos were taken from the vertex to the heel.

Fig. 1. Median sagittal section at entrance of vagina into urogenital sinus (Embryo 1). It is seen tliat the fold at tlie entrance is in connection partly with the vagina, partly with the sinns. F., fold ; U., urethra ; U. S.. urogenital sinus ; V., vagina. Magnified 20 X.

Paraffin was used in Embryo 1, celloidin in Embryos 2, 3, 4 and 5, as an imbedding medium. All sections were cut sagittally, in series, 25 microns in thickness and were stained, some with van Gieson, some with hematoxylin and eosin.

Before proceeding with the description of the microscopic findings, it will be well to point out the difference between the epithelium of the vagina and that of the urogenital sinus. Sinus epithelium as usually seen in section is a narrow, deeply staining band of small, round or spindle cells, with a nucleus almost filling the cell, whereas vaginal epithelium is a wide mass of large polygonal, faintly staining cells. Only the basal cells of the vagina are small and these take a deeper stain (see Fig. 8a). Klein gives the following measurements:

Sinus epithelium : thickness of entire layer, 0.073 mm. ; cells, 13 microns high, 6 microns broad; protoplasmic mantle about one-third the diameter of the nucleus. Vaginal epithelium : thickness of entire layer, 1.75 mm.; cells, 17-44 microns in diameter; nucleus one-fifth the size of the cell. The great variation in size of the vaginal cells is worthy of special attention.

Emhrijo 1 {Fig. 1). The vaginal canal in this 18 cm. fetus can be seen extending a distance of about one centimeter from the cervical indentation to the point where it breaks into the vulva. In its upper four-fifths there is no lumen, the central portion being occupied by an irregularly branching trunk of epithelium, four or five cells in diameter. Tn the lowest one-fifth the canal suddenly widens, the mass of epithelial cells becomes much thicker and a lumen is to be seen that is partly filled with desquamated epithelium. Beneath this epithelial layer lies an area of loose connective tissue cells possessing an embryonal character, with only here and there a connective tissue fibre stained pink by the fuchsin. This layer of embryonal cells is several times thicker than the rod of epithelium in the centre. The outer covering of the vaginal cylinder consists of a thin mantle of connective tissue fibres whose red color, when stained with van Gieson, serves to outline it sharply from the surrounding structures.

The point where the vaginal cylinder enters into the urogenital sinus can be folloAved in nineteen sections. The vaginal conus bends ventrally at its point of entrance into the sinus and thereby a fold is formed between its dorsal wall and the sinus. Microscopically this fold can be seen to consist of the following structures from within outwards: (1) vaginal epithelium, (2) vaginal embryonal connective tissue, (3) fully formed connective tissue fibres from the vagina, (4) fully formed connective tissue fibres from the vulva, (5) vulvar or sinus epithelium. This fold is, therefore, of vulvo-vaginal origin. It corresponds in shape and position to the hymen.

Emlryo 2 {Fig. 2). Fetus 18 cms. long. Owing to the poor state of preservation the epithelium is partly cast off and the hematoxylin stain rather diffuse. In general the state of development approximates that described in Embryo 1. A vaginal lumen can be seen, but it seems to be an artifact due to the desquamation of epithelium. There is no proliferative tendency in the vaginal connective tissue. The bulk of the vaginal conus is composed of embryonal connective tissue cells. The growth along the posterior vaginal wall is less marked, so that the fold left at the point of entrance into the sinus is less pronounced. However, such a fold can be distinctly followed through a considerable number of sections. From the direction of the connective tissue fibres it can be seen that both vulva and vagina enter into its composition.



Fig. 2. Median sagittal section through lower portion of genital tract (Embryo 2). F., fold; It., rectum; S., sphincter ani muscle; U., urethra; U. S., lu-ogenital sinus ; V., vagina. Magnified 20 X.

Embryo 3 {Fig. 3). Fetus 18 cms. in length, good state of preservation. The three layers of the vagina previously described are here also to be seen in about the same stage of development. No vaginal lumen can be distinguished. The convex bulb of the vaginal cylinder jjrojects but slightly into the cavity of the sinus urogenitalis. There is no special growtli of the posterior wall, and nowhere is there to be seen any fold^ such as was found in Embr3'os 1 and 2. There is no proliferation of the vaginal connective tissue.


-Since the completion of this article, I was able to obtain a sixth embryo and made serial sections of the genital tract. The embryo was 18 cms. in length and in an excellent state of preservation. The conditions were similar to those of Embryo 3. The vagina consisted mostly of embryonal connective tissue, with a central, somewhat branching core of epithelium, a few cells in thickness. At the point of its entrance into the sinus urogenitalis. there was no sign of a connective tissue fold such as Nagel describes (Fig. 9).


Embryo If (Figs. 11-13). Fetus 21 cms. in length. The vagina is seen to possess a lumen in its lower half, this lumen being filled almost entirel}^ with desquamated epithelial cells, whose nuclei and protoplasm, though greatly shrunken, can still be differentiated. The three layers of the vagina differ greatly in character from those described in Embryos 1-3. The inner epithelial layer is here the thickest of the three. The cells lie 12-15 rows deep and differ in size and staining character in a way similar to that of the adult vagina; i. e., the superficial cells are large, somewhat spindle-shaped, their protoplasm and



Fig. 3. Median sagittal section tbrougli eutrauee of vagina into urogenital sinus (Embryo 3). No fold is here visible. R., ventral wall of rectum; U., urethra : Y., vagina ; U. S., urogenital sinus. Magnified 20 X.


nucleus staining faintly; the deeper cells, especially the basal layer, are small, cubical, with deeply staining nucleus and scanty protoplasm. Into the layer of loose connective tissue cells outside this epithelial layer the latter sends finger-like processes, so that at times apparent islands of connective tissue cells are seen to lie in the midst of the epithelium. In serial sections these can be seen to be continuous with the connective tissue layer. Judged by the pictures in the previous cases, it would seem more rational to interpret this intertwining of connective tissue and epithelial process as due to an outward growth of the epithelium rather than an inward growth of the connective tissue.

The inner connective tissue layer is not as wide but more dense than in the embryos just described. The nuclei are smaller and stain more deeply, and here and there beginning connective tissue fibrillae are to be noted. The outer laver of connective tissue is not clearly differentiated from surrounding structures. We find the connective tissue fibres more developed and an admixture of unstriped muscular fibres.

Following the vagina down to its point of entrance into the sinus urogenitalis, we are struck by the difference in size and staining property of the vaginal and sinus epithelium. The sinus epithelium at this point consists of three or four layers of small cubical cells with deeply staining nucleus. This difference in the epithelium, already emphasized by Klein, together with the direction of the connective tissue fibres, makes it easy to determine how much is vagina, how much is sinus.

It is seen in studying the series that a crescentic fold of tissue attached to the dorsal and lateral aspect is left at the point of entrance of vagina into sinus. This fold is lined on the inner side by vaginal, on the outer side by sinus epithelium (Fig. 12). It is not by any means a well-formed membrane.

Just anterior to this fold the connective tissue of the vagina both ventrally and dorsally (but principally dorsally) sends a proliferating branch in through the epithelium. The two join to form a membrane that, with the exception of one small opening, completely closes the vaginal canal. It is clear that this membrane must be the hymen, and it is also indisputable that, in this case at any rate, it is of vaginal origin, since it lies internal to the point of junction between vagina and sinus urogenitalis, is lined on both sides by vaginal epithelium, and has its connective tissue directly continuous with the connective tissue of the vagina.

That this membrane is not one of the secondary folds occasionally to be seen internal to the hymen, Avliere there is a marked proliferative tendency on the part of the vaginal connective tissue, is evident by the fact that (with this one exception) there are to be seen no high papillary jDrojections. The sections in this case seem to represent about the same stage of development as do those of Klein, but his interpretation of the findings is different, as will be subsequently shown.

Embryo 5 {Figs. 4-8). Development that of about 5-6 months. The vagina is 3 cms. in length and from 2-6 mms. in diameter. The narrowest portion is the upper fifth, in which there is no lumen and the epithelium is only a few layers of cells in thickness. Here there is also little connective tissue proliferation. Further down the canal, and particularly near the vulvar end, this proliferation is very extensive, so that there appear bands, papillae and islands, depending on the way the sections happen to be cut through the projections. Here we should be at a loss to interpret the various structures were it not that we are able to follow them in series and thus to determine their relationship to their surroundings. Only two layers can be differentiated in this specimen, the middle layer of embryonal connective tissue being absent. The epithelial layer is similar to that in Embryo 4, except that larger



Fig. 4. Sagittal section a little to right of median line thvougn lower genital tract (Embryo 5). Section No. 179. F., fold; H., hymen; R., ventral wall of rectum ; S., sphincter aui muscle ; U., lu'ethra ; V., vagina. Magnified 5 X.


quantities of desquamated cells are found lying in the vaginal lumen. The connective tissue fibres take the fuchsin stain deeply and are more densely compacted than in Embryo 4.

We can again distinguish two folds or membranes. The outer fold is lined externally by vulvar epithelium, internally by vaginal epithelium (Fig. 8a). Its connective tissue fibres (Pig. 8) are partly continuous with the connective tissue of the vagina. In part they intermingle with the connective tissue of the vulva and perineum. Eeconstructed, this fold has somewhat the shape of a thin crescent, whose concave margin faces the urethra. It springs almost wholly from the dorsal wall. A short projection can be seen opposite on the ventral wall. Directl}- anterior to this fold is the true h}'ineii, a memhrane two to three millimeters in height and about one-half to one millimeter in thickness, likewise springing mainly from the dorsal wall (Figs. 4 and 5), with a smal oval ojjening high up near the urethra corresponding to the hymeneal orifice. Its epithelium and connective tissue are vaginal. Papillary proliferations are found on its inner and outer surfaces.

The evidence of these five embryos can be briefly summarized as follows :



Fig. 5. Median sagittal section (Embryo 5). Section No. 187. F.. fold; H., hj'men ; R., ventral wall of rec-tum ; S., sphincter ani nuiscle ; U., urethra ; v., vagina. Magnified 5 X.


In the fetus 18 cms. long (Embryos 1 and 2) there may be seen at the point of junction of vagina and sinus urogenitalis, rising mostly from the dorsal wall, a crescentic fold composed of elements coming both from the vagina and the vulva or sinus urogenitalis. Occasionally, as in Embryo 3, this fold is absent. In the fetus 21 cms. long this crescentic fold is again to be seen, Init not so well marked. Anterior" to it and lying wholly within the vagina is a thick membrane, the hymen, almost completely closing the vaginal canal, composed only of vaginal elements. No other similar folds or membranes are present.

In the fetus 25-30 cms. long the crescentic vulvo- vaginal fold is still recognizable, but the true hymeneal membrane is evidently anterior to it. It is here very well developed and is composed entirely of vaginal elements.

The explanation that suggests itself from the study of my sections is similar to Dohrn's ('75). It points to the hymen as of vaginal origin, independent of the place at which the vagina breaks into the urogenital sinus.^ This spot is already clearly to be seen in the 14 cm. fetus as shown in Nagel's illustration (Fig. 9). Within arises a fold of vaginal tissue, the true hymen, stretching almost completely across the vaginal canal. At the point where the vaginal bulbus breaks through,



Fig. 6. Sagittal section a little to left of median line (Embryo 5). Section No. 191. F., fold; H., hymen; R., ventral wall of rectum; S., sphincter ani muscle ; U., lu-ethra ; Y., vagina. Magnified 5 X.

the so-called j\Iuellerian eminence, a more or less well marked fold of tissue is left. As the fetus develops this fold becomes obliterated. In cases of arrested development we may have the fold persisting almost to birth, thus giving the picture of a bilamellate or double hymen.


'In this and subsequent arguments I have assumed that the vagina is entirely formed by the coalesced Muellerian ducts and not to any extent by the sinus urogenitalis. Practically the only testimony that v^^ould speak against this view is the occasional presence of epithelial areas that appear to come from the urogenital sinus. The interpretation of such epithelial areas is, however, a matter of great uncertainty, as has recently been pointed out by Meyer ('07) in a discussion on the remnants of the WolfHan ducts.


Let ns uow see how this explanation agrees with the findings, clinical and microscopic, that have been hronght forward by other observers.

Considering in the first place the clinical evidence, we would emphasize the variations in the shape of the hymeneal orifice. This has, I believe, not been given due importance. We have on the one hand authors as Kagel ('97) and Klein ('94), who hold that the formation of the hymen is passive, i. e., merely due to a bulging forward of the vaginal bulb, particularly of the dorsal wall, into the urogenital sinus, and a consequent thinning out of the intervening septum. On the other



Fig. 7. Sagittal section about 1 mm. to tbe left of plane of Fig. 6. Urethra and vulvo-vaginal fold are not to be seen in this section. No. 207. H., hymen ; R., ventral wall of rectum ; S., sphincter ani muscle ; V., vagina. Magnified 5 X.


hand, some investigators, as Dohrn ('75) and Schaefer ('95), consider its formation as active, i. e., a proliferation of connective tissue with the production of a membrane more or less completely shutting off the vagina from without. It seems to me the variations in position, shape and size of the hymeneal orifice point distinctly to a proliferative process. If we conceive the evolution as passive, we should expect a round or oval orifice near the upper portion of the hymen. Such a view cannot explain the cases of denticulate, cribriform and fimbriate hymens. Even Klein takes for granted, in the last named form, a papillary growth along the edge of the hymen. In other words, he claims the process is passive except when it is active. This seems irrational. Apparently the hymen does not represent a thinned out meml^rane, but a proliferation of connective tissue. That such a proliferating tendency of the vaginal connective tissue exists, all writers, including Klein ('94), agree.

The next clinical fact to be considered is that occasionally a hymen is to be found in the absence of a vagina. This point is emphasized by



Fig. S. Drawing of dorsal portion of hymen, vulva and vagina in median sagittal section (Embryo 1). Section No. 183. This shows clearly how the vulvo-vaginal fold is distinct from, and posterior to, the hymen. By their density and direction, the connective tissue fibres of the vagina are set off from the vulvar connective tissue. The difference in epithelium is also indicated in a general way. F.. fold ; H., hymen ; Va., vagina ; Vu., vulva. Magnified 40 X.


Pozzi ('84), as favoring his conception of the vulvar origin of the hymen. It is, however, here as elsewhere that the exception proves the rule. In the large majority of cases where the vagina is absent, a hymen is also not to be found. Thus the weight of the evidence favors the vaginal theory. Furthermore, as A^eit points out, the occurrence of a hymen in atresia vaginas can be readily explained. We know that, at some places, portions of the vagina may remain obliterated while at other points a lumen is formed. If the extreme lower end of the vagina be the only portion that so develops it might readily present the picture of a hymen in absence of the vagina.




Fig. 8a. Detail drawing ot the tip of the vulvo-vaginal fold seen in Fig. 8 to show the difference in character between vaginal epithelium and sinus epithelium. C. T., connective tissue ; Y. Ep., vaginal epithelium ; S. Ep., sinus epithelium. Magnified 100 X.


A few cases have been reported in which a single hymen was found with double vagina. This fact is brought forward by the upholders of the vulvar theory as proof of their contentions, in spite of the fact that here, too, the rule is that the hymen is double, one for each vagina. The burden of proof is here likewise against them. The unusual cases of single hymen can, moreover, be readily explained on the basis of an incomplete vaginal septum, that is, one in which the septum dividing the two vaginas does not fully reach to the hymeneal ring.

From the anatomical dissection of 190 specimens of fetal hymens Schaeffer ('90) concluded, as already stated, that this structure was composed of two folds. Gellhorn ('04) has raised the objection to these investigations that they Avere based upon patliological material. According to Schaeffer's own statement 42 per cent of his cases showed some maldevelopment of the genital tract. When we consider what we mean by maldevelopment, Schaeffer's cases acquire a distinctive value of their own. We mean not a different method of development, but an arrest of development.* If in 43 per cent of his specimens there was arrest of development in other portions of the genital tract, we have a right to



Fig. 9. Sagittal section through the posterior end of the vagina of a human embryo of 14 cm. body length (Nagel). 1, urethra; 2, vagina; 3, posterior surface of the hymen ; 4, widened portion of the vagina immediately back of the hymen.

expect a rather large percentage to show an arrest of hymeneal development. Now Schaeffer found in 28.8 per cent of all cases a bilamellate hymen, whereas other investigators — Klein ('94), Hart ('02), Gellhorn ('04) — found bilamellation but rarely or not at all. Is the inference not justifiable that the additional lamella represents a membrane left by some previous step of development that persisted instead of becoming obliterated? But what memljrane could that be? Only the membrane


^It is of interest in this connection that the hymen, according to Nagel, appears as a membrane only in the human race. In elephants, hyenas and other quadrupeds there is usually a constriction at the point of entrance of vagina into sinus urogenitalis, but no true hymeneal membrane. Corresponding to this constriction, we have at this point in man the vulvo-vaginal fold.


left at the point where the vaginal bulbus breaks into the sinus urogenitalis. It is this vulvo-vaginal fold, I believe, so clearly to be seen in my specimens, that in Schaeifer's cases persisted to a later date in a number of instances and gave the appearance of an additional fold of the hymen. This assumption is further supported by the following table in Schaeffer's work:

Number of Length of Fetus. Specimens.

16-25 cms 9

26-30 cms 8

31-35 cms 26

36-40 cms 41

41-45 cms 58

Over 45 cms 28

The steady decrease in the percentage of l)ilamellate hymen as the state of development increases is very striking and certainly would incline one to the belief that the bilamellate hymen represents a more primitive stage of development.

So much for the clinical and gross anatomical evidence on this subject. Coming now to the microscopic investigations, we must put aside as of secondary value all those ^jased on fetuses previous to the fifth month of development, at which time the hymen, according to the consensus of opinion, first makes its appearance.^ This would include ISTagel's ('97) sections of a 14 cm. fetus (Fig. 9) which gives pictures of a vulvovaginal septum similar to that found in my Embryos 1 and 2. In the series of sagittal sections of a 26 cm. fetus studied by Klein ('94) (Fig. 10) we have valuable evidence. A comparison of his illustrations with mine are very interesting. He finds a vulvo-vaginal fold that he interprets as the hymen at the point of junction of vagina and urogenital sinus. Just anterior to this fold and extending from both ventral and dorsal walls is a fold considerably thicker than the so-called hymen. Klein's explanation of this fold is that it is a column of vaginal tissue such as we occasionally find in hymen columnatus. Neither in his pictures nor in his description has Klein proven this, and a priori it is

"This excludes the work of Berry Hart ('02), who argues from the serial sections of two embryos of three months development that the hymen is formed fi-oni the Wolffian ducts. Webster ("98) has clearly pointed out the fallacy of his conclusions.


difficult to see how such an explanation is possible. In a sagittal median section through a hymen columnatns we should expect to iind the column as a broad surface continuous with the hymen. In Klein's sections it is comparatively narrow and there is a considerable hiatus between the column and hymen. A more plausible explanation would be that the so


Fig. 10. Sagittal section of the posterior end of the vagina aud the hymen (after Klein). Of the two folds Klein interprets the left hand one as a columnar branch of the right hand one, the hymen. Magnified GO X.


called column is one of the secondary folds or papilla? occasionally found in the vagina back of the hymen. These secondary proliferations are, however, found primarily in later stages of fetal development when there is a general proliferative tendency in the entire lower vagina. In Klein's embryo there is no such general reduplication of vaginal connective tissue. There are really only the two folds, just as in my sections, the one vulvo-vaginal, the other vaginal; the former a thin septum, the latter a thick membrane. In the absence of contradictory evidence, I feel justified in considering Klein's case as rather supporting than opposing my views.

The character of the epithelial covering also gives support to the vaginal theory. Klein lays stress on the marked differences between the vaginal and the vulvar epithelium. Gellhorn's ('04) microscopic sections of seven hymens at various stages of fetal development show that vaginal epithelium lines both sides of the hymen.

The direction of the hymeneal connective tissue fibres has been emphasized by Gellhorn ('0-i) as being of considerable importance. Even in fetuses at full term they could be seen running parallel and continuous with the vaginal connective tissue fibres. From the vulva no fibres enter into its composition.

I am well aware that the evidence of the serial sections in these five embryos is insufficient to firmly establish my contentions. Further investigations are necessary. Unfortunately, the material in a fresh state is not easily collected. Progress in this question can, however, only be made by the study of serial microscopic sections of the lower genital tract in fetuses of 18-30 cms. length.

Evidence of the sort that has heretofore been employed to support theories of hymeneal development, even if absolutely contradictory, cannot invalidate the views expressed in this paper. It would require evidence of the same character, serial sections of a number of embryos, to do this. Until such evidence is at hand, therefore, we must consider the hymen as a vaginal structure formed in the fifth month of fetal life by connective tissue proliferation directly anterior to the point where the vagina enters into the urogenital sinus.

In conclusion, I wish to thank Dr. E. J. Terry, Professor of Anatomy at Washington University, for assistance in getting material for this work and many helpful suggestions; also Dr. H. P. Wells for his excellent micro-photographs.

LITERATURE.

AcKEEEN, F. VAX. 1888. Beitriige zur Entwickelungsgescbichte der weiblichen Sexual-organe des ^Mensclien. Zeitschrift f. Wissen. Zool., Vol. XLVIII.

Bayer, H. 190.3. Entwickeluiigsgeschichte des weiblichen Genitalapparates. Vorlesungeu iiber Allg. Geburtsbilfe, Vol. I, Part 1, Strassburg.

BuDiN, P. 1879. Recherch. sur Tliymen et I'orifice vaginal. Prog. Med.,

p. 677. BuEHLER, A. 1905. Die Entwickelungsgeschichte der Keimdriiseu und ibrer

Ausfiihrungsgange. Hertwig's Handb. d. Entwiclv. d. Wirbeltbiere, Part

III, Cbap. 2.

Debierre, Ch. 1886. Sur I'aiiatomie de I'ovidnet. Ass. Fr. pour I'avanc.

des sci., Part 2, p. 540. Dickinson, R. L. 1904. Urethral Labia or Urethral Hymen. American

Medicine, Vol. VII, No. 9. DoHRN, F. A. R., voN. 1884. Die Bildungsfebler des Hymens. Zeitschrift

/. GeJ). u. Gyn., Vol. XI. DoHRN, F. A. R., VON. 1875. Ueber die Entwickelung des Hymens. Schrif ten d. Ges. z. Beford. d. ges. Naturwiss. z. Marburg, Vol. X, Suppl. Part 1. Fehling, H. 1S93. Lehrbuch der Frauenkrankbeiten, Stuttgart. Gellhorn, G. 1904. Anatomj', Pathology and Development of the Hymen.

Amer. Journ. Ohst., Vol. L, No. 2, pp. 145-178. Hart, Berry. 1902. The Development of the Urino-genital Tract. Brit.

Med. Journ., Sept. 13. Hoffmann, C. K. 1896. Artikel iiber "Hymen" in Eulenburg's Encyclopedie,

Vol. XI, p. 178. Job. 1898. De I'hymen en rapport avec Taccouchements. Th&se, Nancy,

No. 4. Kempe, H. a. E. 1904. Beitrage zu einer Entwickelungstheorie des

Hymens. Compt. rendu du Cong. Intern, de Zool., Bern, 1904, Bale, 190o,

pp. 315-318; rev. in Jahreslter. u. d. Fortsch. d. Anat. u. Entwick., Vol. XI,

Part III (1). Klein, G. 1894. Entstehung des Hymens. Festschrift zur Feier des fiinfzig jahrigen Jubilaums der Gesellscbaft fiir Geburtshilfe und Gyniikologie,

Wien. KoELLiKER, A. 1879. Manual of Human Microscopic Anatomy, p. 465. Meyer, R. 1907. Zur Kenntniss der kranialen und kaudalen Reste des

Wolff'schen Ganges u. s. w., Zentralbl. f. Gyn., Vol. XXXI, p. 293. MiKALKOvicz, G., VON. 1885. Untersuchungen iiber die Entwickelung des

Harn- und Geschlechtsapparates der Amnioten. Intern. Monatschr. f.

Anat. u. Histol., Vol. II, p. 41. MiNOT, C. S. 1892. Human Embryology, New York. Nagell, W. 1896. Die Weiblichen Geschlechtsorgane, in v. Bardeleben's

Handbuch der Anatomie des Menschen, Part 2, Jena. Nagel, W. 1897. Die Entwickelung und Entwickelungsfehler der Weiblichen Genitalien, in A^eit's Handbuch der Gyniikologie, Vol. I, pp. 521 561. Nagel, W. 1891. Ueber die Entwickelung des Uterus und der Vagina beim

Menschen. Arch. f. Microsc. Anat., Vol. XXXVII.

PoiRER. P., ET Charpy, A. 1901. Traite d'Anatomie Huinaiue, Vol. V, pp.

.".19-555. Pozzi, S. 1884. De la bride masculine du vestibule chez la femme et de

I'origine de I'liymeu. Annalcs dc Gijnec, Vol. XXI, p. 268. Retterer, E. 1891. Sur le developpemeut compare du vagin et du vestibule

des mammiferes. Compt. Rendu. Soc. Biol., Vol. Ill, p. 291. ScHAEFFER, O. 1890. Bildungsanoiiialien der weiblichen Geschleelitsorgane.

Arch. f. Gyndk., Vol. XXXVlI, p. 199. Strobel. 1893. Entwickelung und Anatomie der Vagina, Urethra und Vulva.

Inaug. Diss., Wiirzburg. TouRNEUx, F., ET Legaye, Cii. 1884. Memoire sur le developpemeut de

Tuterus et du vagiu. Joum. de VAnat. et de la Phijs., Vol. XX, p. 330. Veit, J. 1899. Handbucli der Gyullkologie, Kapitel liber "Die Erkrank uugen der Vulva," Vol. Ill, p. 189. Waldeyer, W. 1899. Das Becken, p. 6.54. Webster, J. C. 1898. Some Observations regarding the early Wolffian and

Muellerian Ducts, with Remarks concerning the Hymen. Trans. Anier.

Gyn. Soc., p. 44(>. Wertheimer, E. 1883. Recherches sur la structure et la developpemeut des

organes genitaux externes de la femme. Joum. de VAnat. et de la Phys., Vol. XIX, p. 551.


Figs. 11-13. Sagittal sections through the hymen (Embryo 4). The dotted line is drawn at the level of the entrance of the vagina into the urogenital sinus. It is to the left of this that the hymen is formed. Fig. 11 (Section No. Gl) strikes about the centre of the hymen, so that it appears as folds from above and below that do not meet (hymeneal orifice). To the right of it lies a high, narrow fold, rising from below, lined on the left ( anterior ) side by vaginal, on the right (posterior) side by sinus epithelium. In Fig. 12 (Section No. 63) the exact position of this fold and its relations can be better seen. The hymen can be clearly recognized to the left of it as a connective tissue membrane lined on both sides by vaginal epithelium. This is the most convincing pictiu-e in the series. Fig. 13 is taken lateral to this (Section No. 67). Only a teat-like remnant of the vulvo-vaginal fold can be seen below and to the right. The hymen is seen as a thick band. Papillte appear on the posterior side. Magnified 45 X.



On The Nature Of The Tectorial Membrane And Its Probable Role In The Anatomy Of Hearing

By

Irving Hardesty.

From the Anatomical Lahoratorp of the TJniversity of California. With Twelve Figures.

Tlds investigation has been undertaken with the hope of contributing something to the knowledge of the shape, character and intimate structure of the mammalian tectorial membrane. Of all the organs of special sense, the ear seems to be the most complex and its functional action least understood. As voluminous as is the literature upon it, investigators, as yet, by no means wholly agree even as to the detailed anatomy of the inner division of the auditory apparatus. In rimning over the papers published, one is struck by the fact that they may be quite definitely separated into two classes: anatomical investigations, and physico-physiological treatises, the latter being, in most cases. ]nirely theoretical. Quite often does it appear that the investigator in the latter class maintains that his conclusions are correct on the basis of his correct application of physics and mathematics to structures of unproven existence.

Of the four mechanisms deemed essential in the auditorv apparatus, all the recent papers practically agree as to the identity and action of the transforming and regulation mechanism and the conducting mechanism, and all agree that the hair cells of the organ of Corti comprise the stimulative mechanism; but they disagree as to what comprises the vibratory or resonance mechanism. Most of the older papers attribute the faculty of selective or sympathetic vibration, in accord with the waves imparted to the endolymph in the cochlea, to the basilar membrane. Of the four more important papers of recent date, two still assume that the basilar membrane is adapted for and serves as the vibratory structure, while the other two advance the idea, but occaAmericax Journal of Anatomy. — Vol. VIII, No. 2.


110 Irving Hardesty

siouallv held previously, that the tectorial membrane is the structure acted upon by the transferred sound waves. Either view admits that the hair cells are stimulated by impact or contact with the under surface of the tectorial membrane. The very evident disagreement of these papers in their descriptions and assumptions as to the character, form and intimate structure of the tectorial membrane, and the fact that none of the descriptions of the membrane seemed to apply correctly to some preparations of the cochlea in the possession of the author, suggested the undertakino; of the studv of the nature of that memln-ane.

Materials and Methods.

The material used for this study has been almost wholly obtained from the pig, and all the illustrations given are made from the preparations of cochleffi of this animal. Preparations from other mammals, including man, being obtainable with less ease and abundance, have only been used for occasional comparison with the preparations from the pig.

Pig fcetuses of al30ut term, that is, from 26 to 30 centimeters, and a litter of four pigs of about two weeks after Inrth were used. In all cases, the heads Avere severed, the vault of the cranium and the encephalon removed and the cochleoe then broken out from the inside and made use of as soon as possible after obtaining the material. The entire inner ear or bony labyrinth of the pig may be shelled out from its lodgment in the temporal bone from the fact that osseous fusion with this bone is of late occurrence. Kolmer, '07, mentions that even in the adult hog the cochlea is not so firmly fused in the petrous part of the temporal bone as it is in other mammals.

All the published illustrations and all the ordinary laboratory preparations of cochleae agree in indicating one fact, namely, that the tectorial membrane is invariably and badly shrunken and distorted by the action of the usual reagents used in fixing, dehydrating, etc. Since the chief deficiency of the papers describing the membrane appears to be failure to determine and study its normal characters, it was deemed especially necessary for the present study to obtain at least portions of it in the fresh and natural condition. Attempts toward this end were first made with frozen sections. The fresh cochlege were oriented and frozen on the Bardeen Freezing Microtome and sections clipped into petri-dishes, some containing normal salt solution alone, others 0.1 per cent methylene


The Nature of the Tectorial Membrane 111

blue iu normal salt. Though a few fairly satisfactory bits of the membrane were obtained from the methylene blue solution, the method had to be early discarded as being too coarse. The action of the section chisel in crashing through the bony labyrinth lacerated sorely the delicate structures within, the very action of the freezing appeared to result in distortion and dissociation, and positive orientation of the bits of membrane obtained was impossible.

Lightly crushing of the cochlea followed by gently teasing away the parts under fluid was next resorted to. A cochlea was held upon a solid surface and gently tapped around with a light hammer till the bony wall was sufficiently cracked without rupturing the wall of the membranous labyrinth and the whole placed in a petri-dish of amniotic liquor obtained from the sacs of younger foetuses. Under the dissecting microscope the bits of the crushed bone were then carefully removed with fine pointed forceps, and, with the same forceps and with teasing needles, the outer wall of the scala vestibuli gently torn away from as nearly as possible the entire length of the cochlea. Even with the greatest care, this tearing usually ruptures the vestibular (Eeissner's) membrane and thus seriously disturbs the structiires below it. "With practice, portions of the vestibular membrane could be removed separately and the tectorial membrane identified.

The tectorial membrane floats free from its attachment upon the labium vestibulare of the spiral limbus during the disturbance necessary for the removal of the structures and the consequent agitation of the fl.uid in which it is immersed, but, because of the fact of its being so very delicate and flexible and so extremely subject to surface tension, only short pieces could be obtained (3 to 5 millimeters in lengih) undistorted and free from adhering particles of bone and other debris. It was found that these pieces had to be isolated and handled .by producing currents in the fluid to waft them to desired positions, for upon touching them they would adhere to the point of the teasing needle so closely that freeing them meant distortion or destruction. These pieces were stranded upon the end of a clean slide, to which they were found to adhere less than to the ordinary section lifter, and some were mounted directly in glycerine and in glycerine jelly. Others were washed from the slide into a 0.1 per cent solution of methylene blue in normal salt and let it remain in this from 30 minutes to 2 hours. With the dish of stain over a white surface, these were again stranded upon the slide, rinsed as quickly as possible in salt solution to remove the surplus stain, and


112 Irving Hardesty

transferred for 5 to 10 minutes to a saturated aqueous solution of ammonium j^icrate. From this they were mounted in equal parts of pure glycerine and ammonium picrate solution. "Wliile the methylene blue brought out the structure of the membrane quite successfully at times, it was found, by comparing with those pieces mounted direct in glycerine, that the ammonium picrate produced in jDlaces a certain amount of shrinkage and distortion. Dehydration of these pieces, however gradually and carefully undertaken, was found impossible from the fact that in the higher grades of alcohol they invariably shrunk and crumpled almost beyond recognition. In this way, especially from the unstained pieces, a fair idea was obtained of the normal contour ancl appearance under the compound microscope^ and measurements of breadth could be made.

Owing to lack of sufficient amniotic liquor at times when material was obtained, it was found that teasing under normal salt solution gave apparently as good results, and most of the best bits of fresh tectorial membrane were obtained with its use.

In no case, however careful the procedure, could a whole or even* a half of the entire length of a tectorial membrane be obtained from a fresh cochlea. This was due to some extent to the violent disturbance necessary with the crushing and removal of the bone, but largely was because of the extreme flexibility and delicacy and the remarkable adhesiveness of the fresh membrane. Of the pieces obtained with portions undistorted, many would be covered with particles of debris by the time they could be mounted. Furthermore, extreme difficulty was met in orienting these pieces with certainty as to their upper and under surfaces. Others who have attempted examination of the membrane in the fresh condition seem to have encountered similar difficulties. The descriptions of Eetzius, '84, who worked in aqueous humor, of Kishi, '07, who used salt solution, and of others to be referred to, show that thev obtained but small pieces of the membrane, and that these were imfortunately distorted.

Therefore, the hope of obtaining an entire membrane in the fresh and normal condition had to be abandoned, and experiments were begun for a fixing fluid whose action would result in the least possible distortion of this evidently peculiar structure. This fluid should preferably be a decalcifying fluid as well. It was already apparent that the fluid could not contain alcohol unless accompanied by some ingredient which would counteract its shrinkage effect. Van Gehuchten's (Carnoy's)


The jSTature of the Tectorial Membrane 113

fluid was found unsuitable, though, in addition to its alcohol and chloroform, it contains acetic acid. PerenAd's fluid was found impossible, for, while it fixes with little or no shrinkage, and decalcifies well, it has a softening and macerating effect unless followed by washing in alcohol, which latter produces violent shrinkage. Depuis, '94, teased after fixing with Miiller's fluid and with chromic acid, both of which decalcify and neither of which causes shrinkage, but he does not seem to have obtained whole or even long pieces of the tectorial membrane. Both of these reagents were found to act so slowly that they were deemed undesirable. Gilson's fluid fixes and decalcifies more rapidly and was found to give much better results than any of the above.

The best results were obtained with Zenker's fluid, which contains the bichromate of potassium, the bichloride of mercury and acetic acid, all of which have a decalcifying action and none of which, combined, seemed to cause shrinkage or distortion of the tectorial membrane. Furthermore, as is well known, a prolonged application of this fluid is less injurious to tissues than are most of the fixing fluids, and cochlese immersed in it very seldom show the bubbles of gas within the membranous labyrinth, always produced by the more rapid decalcifying agents, and which disturb the arrangement of the structures within. It was found that 48 hours in this fluid resulted in good fixation and sufficient decalcification for the purpose in mind. The entire apex and the outer, thin bony wall of the entire coil were softened, while the thicker base and basal portion of the modiolus were still hard. This undecalcified base and core were found of considerable advantage in holding the specimen firmly during the careful process of teasing.

Eemoved from the fluid, the cochlea were washed an hour or more in water and teased under the dissecting microscope in a petri-dish containing enough water to well cover the specimen. Holding the specimen down firmly by the semi-circular canals and vestibular portion, it was found that, with practice and the exercise of care, the outer portion of the decalcified bony wall of the coil could be stripped off in pieces, leaving the membranous labyrinth intact and not even crumpled. The next task was to remove the outer membranous wall of the scala vestibuli without injury to the structures below it. The vestibular (Eeissner's) membrane was found to be fixed stiff in itself, and, being very loosely attached along its edges, it could be removed, board-like, and taken out of the way ; but in doing this the tectorial membrane, easily identified below it, was necessarily disturbed and sometimes injured. Only after the


114 Irving Hardesty

vestibular wall and membrane had been carefuly removed was it found advisable to touch the tectorial membrane. This, in the meantime, could be seen, parts waving upward and downward with the agitations in the water about it, parts floating entirely free, and, at times, parts more firmly held in place. These more stationary portions were afterward found to be small areas adhering to the organ of Corti due to the coagulative action of the fixing fluid, probably upon the albrmiens and mucoids in the intervening endolymph.

The action of the fixing fluid results in a decrease to some extent of that remarkable adhesiveness displayed by the fresh tectorial membrane and also renders it somewhat less flexible, but, .even after fixation it is so very sensitive to touch with the needles and to currents in the fluid about it that Reissner's and the basilar membrane appear as boards compared with a strip of thin silk. Fortunately, in the fixed condition, it comes away quite freely from its attachment upon the labium vestibulare ("Huschke's teeth") of the limbus spiralis. Wlien* ready to remove the membrane the water in the dish was drawn ofi: and replaced with clean to avoid the debris resulting from the teasing. Agitations of the water and occasionally lifting it lightly with the needle where it appeared most firmly attached were found sufficient to detach it. It was best removed from the cochlea to the floor of the dish by the gentle use of a small pipette, washing it from around the coil and over the torn edges of the cochlear wall, occasionally guiding it with a needle. This removal proved the most trying part of the whole process in attempts to obtain the membrane in its entirety. It must not he allowed to adhere to structures over which it passes, it must not be straightened too much from its natural coil, and, above all, it must not be allowed to double upon itself lest it cohere and the entire region so doing be spoiled.

Orientation of the entire meml)rane, or of large pieces, was not found difficult, from the fact that one becomes familiar with its differences of width while working with it, and, it being in the form of a coil, it does not turn over readily or may be easily seen when doing so.

AYhen washed out into the dish, the tectorial membrane remains for a time suspended in the water, and one has to watch it and quietly wait for it to settle to the bottom. In settling, it will return, to a certain extent, to its original coiled form.

Membranes thus obtained were stranded upon clean slides and some were mounted, unstained, in glycerine and in glycerine jelly, while


The Nature of the Tectorial Membrane 115

others were gently washed from the slide with and into the stain chosen. After trying several stains, including haematoxylin and methylene blue upon the fixed specimens, the Ijest results were obtained with 0.1 per cent aqueous fuchsin, applied from 30 to 60 minutes. While this stains the interfibrillar matrix of the membrane, it stains the fibers more deeply, and it is both permanent and desirably transparent. From the stain solution, the specimens were carefully rinsed in water and mounted in glycerine or gtycerine jelly, usually with the under surface upward. Some of the membranes removed entire were broken in handling during the staining; others were broken by the cover glass in mounting, hut at these stages the pieces could he oriented with ease, and studies and measurements of them could be made witK practical certainty as to locality. In all the author was able to obtain nine tectorial membranes deemed sufficiently intact and oriented for such measurements as are here given.

The study of the teased specimens was supplemented by the use of sections, some in celloidin and some in paraffin. Some cut vertical to the apex of the cochlea, some tangential and some horizontal. Certain of them were stained with the usual ha3matoxylin-Van Gieson stain; others, for microchemical studies, after being appropriately fixed and embedded, were stained, some with Meyer's muchsmatin, others with Mallory's stain for white fibrous tissue.

To study a point which arose concerning the development and later attachment of the tectorial meml^rane, sections were made of the cochleae of pigs varying gradually from 3 centimeters to 18 centimeters in length. These were fixed, some in Zenker's and some in Gilson's fluid, and all were stained l^y the lia^matoxylin-Tan Gieson method.

The Physical Characters of the Tectorial Membran^e.

The investigation of the mammalian cochlea may be said to have begun in 1847 with the researches of Todd and Bowman, but Corti's "Recherches sur I'organe de Void des mam mi fares appearing in 1851. really formed the starting point of the long series of observations which have appeared aliont equally distributed through the years since that time. It was in this monograph that Corti not only descriljed for the first time the organ which now l^ears his name, but also the structure which spans the spiral sulcus and extends above his organ. This superstructure was shortlv afterwards ag-ain described bv Claudius and


116 Irving Hardesty

Henle, the former of whom discovered it independently and gave it the name, membrana iectoria. Embryological investigations of the structures found in the cochlea began almost simultaneously with the purely anatomical. As early as 1854 Eeissner described in embryos the separate existence of the ductus coclilearis, discovering its limiting spiral lamella, afterwards called Eeissner's membrane by Kolliker and memhrana vestibularis by Henle. However, it was Kolliker who, in 1861, first described the formation of the ductus coclilearis from the ectodermal tube and suggested the process by which Corti's organ is elaborated and also the normal position and morphological signifieance of the tectorial membrane.

None of these earlier papers and but few of those following indicate satisfactory or complete observations as to either the consistency, bulk, extent, shape or actual structure of the tectorial membrane. Most of them saw in their preparations little more than an amorphous mass of very varying shape and very varying relation to the cells of the organ of Corti. Henle, in volume 2 of his Handbuch (1866), believed the membrane to be very delicate, yet firm and resistant, and of an especial elasticity; Bottcher, '69, also concluded that it possesses a high degree of elasticity; Nuel, '75, cited by Eetzius, denied elasticity, but claimed for it a soft gelatinous nature; Gottstein, '73, considered it strongly elastic, and Lavdowsky, '77, considered it a soft, elastic mass, fragile and distensible and in large part homogeneous. Eetzius, '84, examined it in aqueous humor and pronounced it transparent, soft, gelatinous, tolerably elastic, capable of some stretching, but that it would split under increased stress in the direction of evident fine fibers present in it, and found it more firm after the action of fixing fluids; Dupuis, '94, considered it not of a mucous consistency, but decidedly elastic; A^an Ebner, '02, describes it as a fragile, easily distorted, fibrous structure. Kishi, '07, who examined it in salt solution as well as in the fixed condition, found it to possess a marked elastic character and to shrink when subjected to the ordinary procedures; Avhile Shambaugh, '07, in a theoretical paper, considered it a lamellar structure of the same specific gravity as the endolymph in which it lies.

It will be noted that a majority of the authors cited above agree on one point, namely, that the tectorial membrane possesses elasticity. In mv trials to obtain intact specimens of the membrane in the fresh condition, several physical characters seemed uianifest. In the first place, even in proportion to its size, it is, in the fresh, most incon


The Nature of the Tectorial Membrane 117

ceivably delicate and flexible. It is by far more sensitively flexible in the transverse than in the longitudinal direction and the readiness with which it bends when touched or even agitated is beyond description. j\Iounted bits, carefully removed, in the fresh seldom showed longitudinal folds or crumples, but it Avas very difficult to avoid their being bent upon themselves or sometimes tied into impossible knots and gnarls. For the greater part, the tectorial membrane is much wider and thicker than a medullated nerve fiber, for example, but the manipulation of a fresh nerve fiber, especially under fluid, is quite an easy task when compared with attempts to manipulate this membrane. Again, its quality of adhesiveness is phenomenal. It is so subject to surface tension that, it matters not how clean the needle point may be, to touch the membrane is to have it stick, and to remove it from the needle usually means totally spoiling the specimen throughout the region of contact. Scraps of debris and shreds of tissue will adhere to it, and, when bent upon itself, it will often cohere. The membrane had to be moved chiefly by carefully inducing currents in the fluid in Avhieh the work was done.

Its specific gravity is manifestly slightly greater than that of the fluid in which it normally lies. The amniotic liquor of the pig is supposedly about isotonic with the blood serum and lymph of the animal. When freed in this fluid, the membrane, or parts of it, will remain suspended and be wafted about for several seconds, moved by the slightest agitations, but gradually it comes to rest upon the bottom of the dish. Bunge's Physiological and Pathological Chemistry gives 0.81 per cent sodium chloride solution as isotonic or physiologically normal for the pig. In this solution the membrane finally sinks just as in the amniotic fluid. After fixation in Zenker's fluid, it sank similarly in the tap water of this laboratory.

As delicate and flexible as the membrane obviously is, the behavior of my preparations indicated that it does possess elasticity. Judging from their descriptions, and especially from their illustrations, most of the observers cited above did no more than to assume an elasticity for the membrane from its apparent consistency and the various positions and shapes it presented in their hardened preparations. Eetzius and Kishi and the very few others who, in addition, observed the fresh membrane in indifferent fluids, pronounced it elastic from its behavior in these fluids, and the behavior of my preparations confirm their observations. Because of its extreme sensitiveness and flexibilitv, one


118 Irving Hardesty

has to observe its Ijehavior closely to perceive its elasticity. The elasticity ajjparent may be compared with that of a very thin strip of set gelatin immersed in cold water. If not injured by stretching or badly tangled, a piece of the tectorial membrane, as it floats around in a dish of unagitated salt solution, Avill gradually approach the coiled form similar to its natural shape when in the cochlea and thus come to rest upon the bottom of the dish. In case of whole membranes, or of longer pieces, the portion representing the upper or apical coils of the cochlea will assume the natural form more quickly than the other end or the portion from the basal coil, because of the fact that the membrane is thickest and broadest at the summit, gradually decreasing in size toward the base. Twists are far more common and more serious in the more slender region of the basal coil, and unless they disappear before the membrane reaches the bottom of the dish, the basal coil never assumes its natural form and thus pieces may come to rest with the apical coils well formed while the portion from the basal coils may be irregular or scarcely coiled at all. The more evident elasticity of the summit end of the membrane is explained as being due merely to the greater bulk of this end.

Whole membranes and pieces removed intact from the attachment upon the lal)ium vestibidare, unless containing twists, always come to rest flat upon tlie bottom of the dish, the under surface either upward or downward, notwithstanding the fact that the inner zone of the meml^rane removed from the labium vestibulare is very thin and tapers to an edge. This behavior, coupled with the fact that the direction of its greatest flexibility is transverse to the long axis of the membrane, together with its shape in transverse section, indicates that its greatest elasticity must ])e in the direction opposed to longitudinal rather than transverse stress. Elasticity in this direction can be largely instrumental in inducing the membrane to resume its natural shape when in fluid, and the membrane, being laterally attached along the inner zone, elasticity in this direction must tend to maintain the outer zone in its position above the organ of Corti.

Experiments, intentional and accidental, in stretching the tectorial membrane longitudinally indicate a small amount of elasticity opposed to such force. Pieces of the fresh membrane in salt solution, if stretched very slightly indeed will return to apparently their original thickness; if stretches more strongly, they will suffer attenuation, part of which becomes permanent; if stretched further, they may be drawn


The Nature of tlie Tectorial Membrane 111>

out into filaments of less than half the diameter of the original membrane and which seem to have lost all trace of elasticity. These filaments may be drawn asunder and the parts will remain with tapering" and pointed ends. Pieces of membrane from cochleae fixed in Zenker's fluid show a slight increase in rigidity and therefore greater elasticity when the slight strains are applied, but, if stretched further, they break with a transverse cleavage instead of suffering permanent attenuation. In the study of the structure of the membrane it was foimd that the lines of such cleavage follow the direction of the filu-ous components of the membrane.

The tectorial membrane suffers severely in the presence of all reagents which are hypertonic to it. Most all the recent papers touching upon its structure admit that it undergoes shrinkage with treatment for embedding and sectioning. I made several careful attempts to prepare pieces of the memlirane for mounting in balsam so as to get more permanent and probably more transparent preparations, but none were successful. However carefully the graded alcohols were applied, by the time dehydration was complete, the membrane would show shrinkage, constrictions and general distortion, and a bath in xylol would usually complete what the alcohols had begun. Eesults somewhat lietter were obtained by clearing with creosote after 95 per cent alcohol, but no preparations so made would bear comparison with the normal, mounted direct in glycerine, either as to shape, dimensions, evenness of contour or surface markings. The physical character of the membrane seems to be such that it is impenetrable to the alcohol molecule, or at least such that it cannot be dehydrated without shrinkage.

Under the dissecting microscope by reflected light it has a distinct glassy appearance and by transmitted light it is transparent. The compound microscope shows it to consist of fine colorless fibers embedded in a transparent matrix. This matrix has apparently tlie nature of a glutinous, collagenous semi-solid. It is hardened but slightly by fixing fluids containing no alcohol. It is about the first structure in the cochlea to be affected by maceration, for cochleae taken from pigs not freshly obtained show evidences of liquefaction of the tectorial membrane Avhen the organ of Corti may be practically intact. Sections of such cochleffi show above the organ merely a mass of tangled filaments resembling a coagulum. On the other hand, the membrane seems to be more readily fixed than the organ, for sections of cochlea pronounced


120 Irving Hardesty

imperfecth- fixed will often show the membrane with no sign of liquefaction when the cells of the organ are badly deranged.

It was thought that the matrix might probably have a mucous character from the fact that preparations are occasionally seen in which the membrane stains more or less deeply blue. To test this, specimens w^ere appropriately fixed, prepared into sections and stained with Meyer's muchasmatin and mucicarmine according to the procedure followed by Bensley in his study of the glands of Brunner, but in no case did the matrix take the stain in the way considered differential for mucus. It was further found that the membrane never stains deeply blue with hsematoxylin in preparations that have been carefully washed after fixing and decalcification. Orcein and Weigert's stain for elastic tissue likewise gave negative results. The matrix may be somewhat similar to that of certain cartil'ages only much softer, or, since studies of the development show the membrane to be a cuticular structure, and since it is derived from cells of octodermal origin, the matrix ma}^ be a variety of soft keratin.

The Shape and Dimensions of the Tectorial Membrane.

The cochlea of the pig is of the flat type. Gray, '07, has recently divided cochleae into two general types, the sharp pointed and the flat. The flat type is most common, being possessed by several varieties of mammals, including the primates and ungulata, to which latter the pig belongs, while the sharp pointed type is possessed by the carnivora and rodents, the Edentata having an intermediate type and the Marsupials both types.

The pig has one turn more in the coil of its cochlea than in that of the human ear. While the human cochlea is usually described as having 2y2 turns, Wiedersheim, '93, accredits it wdth nearly 3 turns, giving the pig 4 turns, the cat 3 turns, the rabbit 2l^, the ox 314, and the cetacea II/2 turns.

The tectorial membrane of the pig extends throughout the entire length of the ductus cochlearis (scala media), and thus it is about as long as the scala vestibuli, but not quite as long as the scala tympani. Therefore, it is not quite as long as the basal side of the membranous labyrinth. This relative extent is probably true for all the mammals. Dupuis, '94, who studied teased preparations after fixation with osmic acid, ]\Iuller's fluid, etc., and decalcification with hydrochloric acid.


The jSTatiire of the Tectorial Membrane 121

reports that for the cat, guinea pig, dog and rabbit the membrane extends throughout the cochlea. Measurements of the lengths of nine tectorial membranes, seven from the cochlea of pig foetuses at about term and two from pigs about two weeks old, gave an average length for the membrane of 25.5 millimeters.

In both breadth and thickness, the membrane gradually decreases from the apex of the cochlea toward the basal end (Fig. 1), and both the end at the apex and the basal end terminate bluntly, slightly tapering. The upper surface of the membrane is convex throughout its length (Figs. 2, 3, 4, 6 and 7), the height of the curvature being in the region which, in the natural condition, overlies the interclasped phalanges of the rods of Corti's organ. The under surface, or the surface next to the organ of Corti, is concave, but with a concavity of several curvatures. In the region of the apex of the cochlea, the outer edge of the membrane projects beyond the confines of the organ of Corti and here this edge bends downward slightly, resulting in an evident concavity of the under surface. The portion immediately imposed upon the surface of the organ of Corti is parallel with the surface of that organ and thus is usually plane; it may slant upward or downward from the horizontal plane of the cochlea or it may be slightly concave, depending upon the direction and the shape of the upper surface of the organ of Corti, which varies somewhat in the different regions. In the decrease in width of the tectorial membrane toward and in the basal coil, it gradually ceases to project beyond the organ of Corti, and thus finally there is no separate concavity of the outer edge to be considered (Figs. 4, 5 and 8). That zone of the under surface beginning with the inner border of the organ of Corti, spanning the spiral sulcus and including the inner, thin edge of the membrane, is, throughout, more deeply and regularly concave than any other portion. The greater concavity here is but an expression of the rapid decrease in thickness which is attained by the inner edge as it becomes attached upon the vestibular lip of the spiral limbus.

All the investigators mentioning it describe the inner edge as extending to the insertion of the vestibular (Eeissner's) membrane. In the main, this has been found to be true in my preparations of the pig. Occasionally, observation and comparative measurements from the free membrane and of the width of the labium vestibulare have indicated that the inner edge does not necessarily in all regions extend quite to the vestibular membrane. Hensen, '63 ; Eetzius, '84 : Barth, '89 ;


122 Irving Hardesty

Dupnis, '94:, and others have described the inner edge as sometimes notched. Xotches were frequently seen in my preparations, but they were always angular and could be explained as breaks suffered in freeing the membrane from its attachment.

The outer edge of the tectorial membrane in all my preparations is rounded and terminates bruskly. In discussing observations made by Kolliker in 1859 which denied an attachment of the outer edge of th'e membrane to the lamina reticularis and the cells of Hensen of the organ of Corti, Lowenburg, '64, described a delicate, almost invisible, seemingly frayed-out reticulum along the outer edge. Others have since mentioned this appearance, giving it the name, Loicenherg's border plexus. Dupuis, while admitting that it is very irregular, usually appearing in fragments, and not always present, referred to it as a third and outermost zone of the tectorial membrane, while A^on Ebner, '02, among others, refers to it as a filamentous structure, adhering to and collapsed upon the outer zone of the membrane and representing its original connection with- the lamina reticularis.

The existence of this border plexus as a normal appearance of the tectorial membrane is here denied.

Though the appearance observed may be seen at times in the preparations used here, there is evidence that it may be due to three causes, separately or in conjunction: First, it has always been observed in fixed preparations, and the fixing fluids and especially the alcohol produce coagulation of the albumens and globulins in the endolymph, Coagulum filaments, in forming, always appear most abundant on surfaces, and especially between surfaces which are close together. The places on the tectorial membrane where these filaments are most apt to be noticed are its outer edge and that portion of its under surface which is imposed upon the organ of Corti. Second, evidences of beginning liquefaction of the matrix of the membrane are, in the preparations used here, always most apparent along its outer edge and under surface. Portions of the fibers in the membrane, set free along the edge by the dissolution of the matrix, may produce a delicate, tangled, filamentous appearance which is no doubt augmented by coagulum filaments produced by the reagents. Third, in the study of the removed pieces of the tectorial membrane, I have now and then seen what appeared to be parts of Lowenberg's border plexus, but these appearances proved to be displaced portions of the accessory tectorial membrane described below (Figs. 3 and 5). In the region of the basal


The jSTature of the Tectorial Membrane 133

coil^ the outer edge of this accessor}^ membrane nearly coincides with the outer edge of the main body of the tectorial membrane;, and but slight displacements will result in its projecting far enough to be separately distinguished.

Corti, in first describing the tectorial membrane, expressed the belief that the outer edge was normally attached to the structures below it and located this attachment as far over as the epithelium upon the spiral ligament. Coyne and Cannieu, '85, agreed with Corti in so far as to place the attachment upon the cells of Claudius, and ever since Corti, investigators have looked for and claimed an attachment of the outer edge, though seldom agreeing as to the locality, and they have explained Lowenberg's Iwrder plexus as the outer edge of the membrane frayed by having been torn from its attachment. Quite recently, Kishi, '07, Avho seems to believe that the membrane is normally attached to the organ of Corti, states that* the border plexus, or third zone, is only an artifact which results from the tearing of the membrane from the organ of Corti, and which consists either of the frayed edge of the membrane alone, or of this edge together with a portion of the lamina reticularis removed with it. On the other hand, l^eginning with Hensen, '63, and Bottcher, '72, the outer attachment has been frequently denied. Kolliker and Yon Ebner admit, from the study of its development, that at one time the tectorial membrane is of necessity attached along its under surface, but that early, during the differentiation and elaboration of the organ of Corti, and the resultant displacement and adjustment, the attachment of the outer edge at least becomes obliterated. Ferre, '85, described the outer edge as bluntly rounded, and so it appears in the large majority of preparations, in both sections of fixed cochleae, always allowing for shrinkage, and also in the fresh condition in teased preparations.

Measurements of the ividth of the tectorial membrane of the pig were carefully taken at the different turns of the coil of the nine teased preparations obtained. These measurements were transversely across each tip or extremity and across the intervening portions at intervals of each half turn of the coil. IsTow and then partial breaks or slight distortions appeared in short portions of a specimen, and if these appeared in the particular portion of the turn to be measured and seemed sufficiently serious to invalidate the measurement, the measurement for that interval had to be omitted. However, these injured places were not very numerous, and, from the nine specimens, enough trustworthy measure


124


Irving Hardesty


ments at the different turns were obtained from which to compute very fair averages. Fortunately, one of the membranes from the pigs two weeks old was more successfully manipulated, and, after mounting, was so nearly intact and uninjured that a complete set of measurements was obtained from it alone. The measurements of the membrane from the foetuses at term averaged throughout so nearly identical with those from the pigs of two weeks that one could infer no essential dimensional differences between them. These transverse measurements were made with an ocular micrometer whose spaces were standardized in terms of microns. The measurements obtained from all the specimens gave the following averages for the total width of the tectorial membrane in the regions specified:


tip of basal end


58.1


near I 7th ! 6th basal j half- | halftip I turn j turn


107.9 I 124.5


5th halfturn


4th halfturn


3cl I 2d 1st

half- ! half- halfturn turn turn


149.4 170.2 190.9 207.5 257.3 ' 307.1


width tip of near tip end at at apex apex


215.8 166.0


The measurements of the tip at the apex were taken transversely through the point at which the inner edge terminates, and those of the basal tip, transversely through the point at which the outer edge terminates (see Fig. 1).

What is very apparent in the actual specimen may be perceived from these figures, namely, that the tectorial membrane of the pig varies widely in breadth, but varies gradually and somewhat progressively from the basal end toward the apex. Each end terminates bluntly, the widths of the tips taken at similar levels being much less than the widths possessed l)y the half-turns to which they belong. The width of the tip at the apex of the cochlea is nearly three times that of the basal tip. At the apex the membrane rapidly increases to its greatest width, which is attained in the first half-turn. From the basal end, it increases in width gradually till the width in the first half-turn is 2.8 times greater than that of the last.

Almost the entire variation of the tectorial membrane occurs in its outer, free, portion, or that strip spreading from the outer edge of the labium vestibulare (Huschke's auditory teeth) over the spiral sulcus and the organ of Corti. That portion, or inner zone, which lies upon and adherent to the labium vestibulare varies relativelv little in width.


The Nature of the Tectorial Memljrane 125

However, like the total width of the membrane, this inner, attached, strip is narrowest in the basal tnrn and increases gradually, thougli slightly, toward the apex to attain its greatest width in the first turn. Owing to the extent and shape of the labium vestibulare at the ends of the coil, this attached strip terminates in a point, and is the first zone to terminate at the apex, while at the base it terminates bluntly rounded and is the zone which persists farthest, wholly constituting the basal tip of the membrane.

Dupuis and others to be mentioned below, who studied the markings on the under surface of the membrane, distinguished a boundary line representing the outer extent of the labium vestibulare, or the boundary of the zone of attachment upon it. Barth and Bottcher considered this line an artifact. In all my teased preparations, but especially those fixed in Zenker's fluid, this line was distinguishable and, from further study, it is evident that Barth and Bottcher were correct in considering it an artifact in so far that it only represents the impress of the edge of the labium vestibulare, and has nothing further to do with the structure of the membrane. Measurements of the attached zone taken transversely to the inner edge of the membrane showed it to vary gradually from 62.3 microns in the last half-turn to 87.2 microns in the first half-turn.

These measurements subtracted from the total width of the membrane at the same regions, give the free, outspanning portion of the membrane a width of only 45.6 microns near the basal end, while the free portion of the first half-turn attains a width of 219.9 microns. Ivishi, '07, who studied the human tectorial membrane and that of a number of common mammals, states that this outer portion (second zones he calls it) is at least three times broader at the apex of the cochlea than it is in the basal coil. He does not give measurements. In Kolliker's Gewebelehre, Bd. 3, Second Half, Von Ebner states that the tectorial, like the basilar membrane, is widest toward the apex of the cochlea, and that, in man, the free portion has a width of 240 microns at the apex, and 120 microns in the basal region. These figures indicate that the free portion of the apical turn in man is only 2 times as wide as that of the basal region, while the above figures for the pig show the free portion in the first half turn to be 4.8 times as wide as it is in the last. However, Yon Eben does not state at what particular portion of the coil the measurements were taken nor whether they were made from fresh or fixed specimens or from sections. Kishi, though he only


126 Irving Hardest}'

publishes some very misleading photographs of vertical sections of the cochlea, made some of his observations from isolated bits of the membrane teased free in fluid, and probably these bits suggested to him the greater difference in the width of the outer portion in the two regions. Obviously, measurements of the membrane from sections, necessarily treated with alcohol, etc., and therefore shrunken, would average less and be less regular than if taken from teased and unshrunken specimens, and, as may be seen from Fig. 1, such results depend materially upon the particular localities at which measurements of the free zone are taken. Xone of the literature I have consulted indicates the successful isolation of appreciable lengths of the tectorial membrane of any animal.

Fig. 1 illustrates the shape of the pig's tectorial membrane as seen from the under surface and with the coil slightly opened, and it is an attempt to indicate its appearance under low magnification. In its natural position in the cochlea, it is more closely coiled, and, especially in the apical turns, the outer edge of one turn appreciably overlaps the inner edge of the turn below it. Thus it was necessary to draw it less closely coiled in order to show it entire. When mounted in glycerine, the cover-glass presses the apical coils of the membrane together and radially outward from the center. The basal coils could not be induced to retain their coresponding curvature with the application of the cover glass, though carefully arranged in the glycerine beforehand. Therefore, it should be stated that none of the preparations obtained showed the membrane as evenly coiled as it is shown in Fig. 1. In making this drawing, the character and extent of the coil was taken from the labium vestibulare laid bare under the dissecting microscope, after the removal of the membrane, and from the appearance of the membrane resting upon the bottom of the dish in fluid. A coiled line was drawn representing the inner edge of the membrane and imitating the character of the coil of the cochlea, but opened sufficiently to avoid overlapping of the edges of the membrane. Then radially from this line, the width of the membrane at each half turn, obtained by measurement of the actual specimen, was indicated by dots. The character and dimensions of the ends and the lines of demarkation of the under surface were likewise sketched in from measurements and from the study of mounted specimens. The scale decided upon gave each space of the ocular micrometer the value of one millimeter. The outline being made, the width and variations of the different zones were determined and other


The jSTature of the Tectorial Membrane 127

peculiarities of the difficult regions sketched in. The sketch was then transferred to a sheet of Eoss-board, upon which it was thought advisable to make the attempt to represent the glassy, delicately fibrous character of the membrane as it appears over a black surface. Figs. 2, 3, 4 and 5 were outlined in the same way, but under higher magnification, and are intended to show greater structural detail.

Measurements of the thiclcness of the membrane could be obtained only from vertical sections of cochlese made after the usual fixing, dehydrating and embedding. After getting some idea of the character of the membrane, a few specimens, carefvilly treated and embedded in celloidin, gave sections which showed considerably less shrinkage and distortion than is usually evident. Figs. 6, 7 and 8 were outlined from photographs of and drawn from a section of one of these specimens.

All the membranes showed some distortion and shrinkage. The latter especially is evident in a denser and more amorphous shell-like thickening about the entire periphery, which suggests a primary condensation due to the first attack of the shrinking agent, and which, once formed on the periphery may be less permeable and may result in the interior portion being subsequently more gradually and therefore less violently acted upon by the reagents. Fig. 8, from the seventh half -turn or basal coil, shows decided distortion and displacement due to shrinkage. This is from the more slender jjortion of the membrane, and, if the above suggestion is true, would necessarily be affected most by the reagents.

Measurements of the thickness taken from these specimens and corrected with the study of the isolated and supposedly unshrunken pieces indicate two relations :

(1) That, like the width, the thickness of the membrane begins with that of the outer zone in the rounded end at the apex (see Fig. 1), rapidly increases to its maximum in the apical region and thence gradually decreases toward the basal coil, in the end of which the minimum thickness occurs.

(2) That, throughout, the line of the greatest thickness, in the different transverse sections, runs approximately parallel to the edge of the labium vestibulare and approximately over or slightly to the inner side of the line of the enclasped phalanges of the rods of Corti.

The measurements of thickness were all taken from sections of the cochlefe of pigs at or very near term. They gave the following average tliir-kup-^sos for the regions specified:


128


Irving Hardesty


tip

at apex


1st halfturn


2d

halfturn


3d halfturn


4th halfturn


5th halfturn


6th half turn


49.8


74.7


70 5


62.3


53.9


42.4


31.1


7th halfturn


13.4


These averages are obtained from vertical, axial sections from each of four cochleae. The thickness of the tip at the apex is computed froin one measurement, for, unfortunately, the plane of only one of the sections was such as to pass transversely through the tip. Further, it is uncertain how far from the basal tip the measurements designated "seventh half-turn" were taken. Evidently they passed varying distances from it, for the measurements varied from 12.4 microns to 24.9 microns. The other regions did not vary so widely. All the measurements were made and recorded before the averages were computed and converted into the terms of microns.

It is seen from these' figures that the greatest thickness of the membrane occurs in the first half-turn, and thence the thickness decreases gradually and somewhat progressively toward the basal end. A comparison of Figs. 6 and 7 indicates that in this cochlea the greatest thickness occurred in the third half-turn instead of in the first. The shrinkage here seems to have resulted in a lateral crumpling and probable increase of thickness in the third half-turn. The drawings of this specimen were made before the average thickness was ascertained; otherwise a more average specimen might have been used. This specimen was among those measured. The first and second turns all showed much smaller differences than any other adjacent half-turns. That the differences between adjacent basal half-turns are greater than those between the half-turns of the apex is due to the fact that the basal turns are longer, the basal coil of the cochlea being several times longer than the apical, and thus longer strips of membrane intervened between the measurements in the basal half -turns. The lack of greater uniformity in progressiveness of decrease is probably due both to irregularities in the shrinkage effects of the reagents and to lack of identical orientation of the plane of section of the different cochleae, the plane passing nearer to the ends, especially the basal end, of some membranes than of others.

Yon Ebner quotes Eetzius for the statement that the thickest part of the human tectorial membrane measures from 24 to 25 microns. The particular turn from which this observation was made is not stated. The


The Nature of the Tectorial Membrane 129

figures seem rather low when compared with the 74 microns obtained here for the first half-turn of the membrane of the pig.

The Structure of the Tectorial Membrane.

Surface marJcings. Since the beginning of its study, the natural tendency has been to divide the membrane into zones. These zones vary in the different descriptions both as to their number and their boundaries, as is to be expected from the fact that some investigators have dealt wholly with appearances seen in vertical sections of the cochlea, others with the surfaces of isolated pieces of the membrane, and the majority with preparations more or less distorted by the action of reagents. Gottstein, "72; Coyne and Cannieu, '85; Barth, '89; Dupuis, '94, and others made three zones: (1) The inner zone, comprising the thin, attached strip spreading from the inner edge, at the insertion of the vestibular membrane, and terminating with Huschke's teeth or the edge of the labium vestibulare; (2) the middle or second zone, comprising the body of the membrane which spreads from Huschke's teeth to the outer, bruskly rounded border; and (3), believing Lowenberg's border plexus to be a part of the membrane, they called this the outer or third zone.

In 1863, Hensen, examining the surface of the membrane, observed a line running along the middle of the main body, approximately parallel with the edge of the labium vestibulare and apparently constant in occurrence. Later, this line was farther observed by Eetzius and Schwalbe who described it as a transparent, glittering strand on the under surface of the membrane, having the appearance of a hyaline thickening along the middle zone. By both of these investigators it was given the name "Hensen's stripe." This stripe or streak, while it has been reported absent in some animals (the rabbit, for example), has been used as a boundary line by those who have chosen to divide the membrane into three zones, exclusive of Lowenberg's border plexus whether admitted or denied. Divided in this wa}^, the first zone comprises the strip attached upon the labium vestibulare; the second or the middle is the strip between the labium vestibulare and Hensen's stripe, while the third or outer zone comprises the balance of the width from Hensen's stripe and including the outer edge of the membrane. Most of the more recent papers divide the membrane into zones in this way, though some, for example, Eickenbacher, '01, divide it into two zones: an inner, as


130 Irving Hardest} above; and an outer, comprising all of the outspanning portion and thus including Hensen's stripe.

Also, throughout the literature, occur descriptions of various lines and markings of the under surface in addition to Hansen's stripe and the line of impress of the labium vestibulare. Hensen's stripe has even been given various appearances, various widths and varied occurrence. Dupuis, '94, Avorking with the tectorial membranes from the dog, cat, rabbit and guinea pig, found Hensen's stripe apparently absent in some cases, while in others it seemed to be represented by three lines, which, however, he thought might have been due to optical defects produced by the variations in the thickness of the membrane compressed under the cover glass. Various extra lines have been described upon the under surface of the memtoane, some running longitudinally, others transversely or obliquely. Only recently, Kolmer, '07, working with various mammals, including the pig and man, describes a small piece of the membrane, giving it longitudinal lines upon its under surface which he interprets as bundles of longitudinally running fibers. In all of my preparations which proved to be shrunken, extra lines were always apparent; The prevailing direction of these was always longitudinal, and most of them appeared on the under surface. Especially, in all those pieces which I attempted to dehydrate and clear in oil, numerous lines were apparent, some of which ran obliquely and many of which anastomosed. Comparison with the fresh preparations and with those fixed with apparently little or no shrinkage made it very evident that these extra lines were due to the shrinkage effect of the reagents, and they were explained as crumplings and crimps of the shell-like peripheral condensation which is seen in all sections of embedded material and is probably produced in the first attack of the reagents upon the matrix of the membrane. Further shrinkage or extraction of water from the interior can but result in crimps of this outer, already condensed laA^er. As Kolmer says, such lines always seem to stain more deeply than the adjacent material, but this must be due to their greater condensation, and therefore greater capacity for holding stain. Though seeming more or less amorphous, as do all condensations of the membrane, a fibrous structure may be detected in them; but neither in these lines nor in any parts of the tectorial membrane of the pig have I been able to distinguish any longitudinally running fibers.

The normal markings readily detected on the under surface of the tectorial membrane of the pig when removed intact and unshrunken,


The Nature of the Tectorial jMembrane 131

and viewed under surface upward, are four only: (1) The fine fibers of which the membrane consists, the prevailing course of which is obliquely transverse and which are embedded in the transparent, collagenous matrix; (2) the line of impress, the boundar}^ of the zone of attachment, left by the edge of the labium vestibulare; (3) Hensen's stripe, always present, whose variations and significance will be mentioned below; (-1) a line on the outer zone which, in passing from the apical to the basal end, gradually approaches and comes to coincide with the outer edge of the membrane, and which constitutes the outer edge or back of the accessory tectorial merribrane to be described below, and from which may usually be detected short fibers extending obliquely towards Hensen's stripe. Some of the descriptions for Lowenberg's border plexus, some of the claims of attachment of the outer zone of the tectorial membrane to the cells of the organ of Corti, and some of the descriptions of extra lines on the surface were very probably dealing with this accessory membrane, for it seems to be easily displaced, and, in shrunken preparations, is diminished and distorted beyond recognition.

All the papers dealing with the appearance of detached pieces, seem to agree that the upper surface of the membrane is smooth and more or less evenly convex. Beyond the effect given by the component, obliquely transverse fibers of the membrane, the upper surfaces of my preparations showed no markings that could not 1)e interpreted as due either to injuries from manipulation, adherent debris or coagulum filaments. The highest point in the convexity of this surface remains throughout in about the same relation to the organ of Corti.

The -fibrous structure of the tectorial membrane is very evident in teased preparations, though the .exact internal arrangement of the fibers is somewhat difficult to determine. As mentioned above, the fibers are imbedded, or distributed throughout, in a transparent matrix, seemingly having the character of a glutinous or collagenous semi-solid. It is the nature of this matrix that allows of that most remarkable flexibility and sensitiveness possessed by the membrane and probably determines its low specific gravity. The matrix is stiffened Init slightly by fixing agents, shrinks severely in dehydration and stain tests indicate that it is not of a mucous nature. It is due to its shrinkage that the membrane so often appears in sections as an irregularly shaped, deeply staining, amorphous mass lying in various positions above the organ of Corti, and it is with this appearance the investigators have had frequently to deal.


132 Irving Hardesty

Lowenberg, "64, thought that the membrane consisted of layers, one above the other; Gottstein, '72. pictured it as structureless, and many others after these failed to comprehend its character. Most of the physiological and theoretical papers giving it any attention assume rather than study its anatomy. All the more recent anatomical studies admit the membrane to be fibrous. Shambaugh, '07, assumed the membrane to be constructed of an immense number of delicate lamellfe, and on the basis of this, propounds a very interesting theory of tone perception. Coyne and Cannieu, '85, were, I think, the first to make tangential sections of the cochlea, passing through the membrane transverse to its fibers, and to picture the transverse sections of the fibers as fine dots, as shown here also in Fig. 9. They also tried to analyze the course and arrangement of the fibers, and, with Barth, ^89 ; Dupuis, '94, and Kishi, '07, made, I think, a mistake in describing the fibers as coursing from the labium vestibulare in two layers, one on the upper surface and one on the under, with a less compactly assembled layer of fibers between the two. They probably mistook the peripheral condensation, explained above, as due to the action of the reagents, as special upper and loAver layers of fibers.

Seen from the upper surface, the fibers appear to course from the attached or inner zone outward, but, instead of coursing radially, they always slant from the labium vestibulare toward the apex of the cochlea. This slant is greater in the region of the inner zone, and deviates slightly toward the radius in the outer zone as shown in Fig. 4, tlie general course showing a slight tendency toward the shape of the letter >S^. This general course was also noted by Dupuis, '94, in his preparations. Kishi notes an oblique direction tending toward the apex in the animals he studied, and A^on Ebner states that the inclination is about 45° from the radius.

Kishi claims to have noted fibers of varying size and length, and Kolmer, '07, states that the fibers have a thicloiess of 14 micron. From casual observation the impression is very readily obtained that the fibers course individually across the entire width of the membrane, but closer study of the membrane of the pig under higher power convinces that they are far more numerous than first supposed, and tliat they are too fine to be followed individually or to be measured by the ordinary means.

Comparing the surface view with the transverse section (Figs. 2, 6 and 4) shows that, instead of all the fibers extending from the labium


The Nature of the Tectorial Membrane 133

vestibulare across to the outer edge of the membrane, the greater number of them at least course varying distances across, then curve downward and, still curving, reach the under surface, where they terminate, many running a short distance parallel to the under surface, prior to termination. The shape of the bluntly rounded, outer edge of the membrane is determined by the curvature downward of the outermost of the fibers. The outline of this edge is never perfectly even, but always appears finely and irregularly scalloped, as shown in Fig. 3, suggesting that the fibers may have a tendency to course and curve around the edge in bundles. In some of my preparations the under surface of the outer edge of the sections at the apex (similar to Fig. 6) appeared thickly studded with the down-hanging ends of fibers as though the under surface had been frayed either by the tearing away of the peripheral condensation or by partial liquefaction of the matrix. Along the inner edge, the sections show that all the fibers by no means are continuous from the inner, attached zone, but that the rapid Increase in the thickness of the free portion results from an enormous increase in the number of fibers concerned. The interior fibers of this region course and curve parallel in the main with those of the upper surface, till those nearer the under surface gradually come to sweep downward and then outward and parallel for short distances with the under surface which they form and in which they terminate.

Fig. 9 is made from a horizontal section of the cochlea, and represents a tangential section of the tectorial membrane passing about parallel with the surface of the labium vestibulare and splitting Huschke's teeth. It is taken from the neighborhood of the third halfturn, and a comparison of it with a line drawn in the same plane through Fig. 7 will make clear the relations and appearances shown by it. The plane passes through the dorsal aspect of the membrane, and, as is to be expected from the evident course of the fibers shown in Fig. 7, near Huschke's teeth (Ht) the fibers are cut transversely and obliquely in the zone indicated by a, while the fibers in the outer zone, indicated by &^ course more nearly in the plane of the section. In this section the fibers do not appear to run wholly parallel with each other. Apparently they even anastomose, and where they are cut transversely, they appear to be connected with each other by fine collateral filaments. How much this specimen was shrunken and how much this appearance is due to the reagents cannot be determined. The tangled appearance is no doubt due largelv to shrinkage and coagulation effects. The fine filaments


134 Irving Hardest} may represent the individxial filiers^, while the heavier lines and dots may consist of agglutinated bundles of them.

The under surface of the main bod}' of the membrane is thus given a transversely fibrous appearance when viewed on the flat, not by continuous fibers coursing across it individua'.ly, but by the general direction of the curved ends of many fibers. Beginning on the under surface, the fibers first lie parallel with this surface and form it, and then curving outward, upward and then inward, contribute to the upper surface. The summation of the course of these ends gives to the under surface the appearance of being striated in a direction oblique from the radial and, like the upper surface, the slant is from the labium vestibulare toward the apex of the cochlea. But the direction on the under surface is more nearly radial than that on the upper, as may be seen by comparing Fig. 4 with Figs. 2 and 3. On both surfaces, the direction of the fibers in the inner or attached zone of the membrane is inclined more toward the apex than that of the fibers in the main body and especially the outer edge. An examination of the under surface showed that the most nearly radial direction of the fibers occurs in the second and third turns, and an attempt is made to show this in Fig. 1.

Fig. 4 represents a drawing of the l^roken end of a piece of membrane from a cochlea fixed in Zenker's fluid. The preparation showed that the fixed specimen, at least, in breaking, has a tendency to break parallel with the direction of the fibers, and it suggests that the difference in the course of the fibers above and below may contrymte to the strength and elasticity of the membrane by a sort of interlocking arrangement. In a break like this, the portion of the fibers intervening between the under and upper surfaces must of necessity be broken across. In the break pictured, there were evident clumps of fibers which seemed to have been partly split off from the remainder before breaking, and which appeared something as slivers, each a mass of fibers held in a corresponding amount of the glassy matrix (Fig. 4, S).

From a careful comparison of the surface views with the sections one may conclude that the fibers contributing to the under surface of the tectorial membrane make their immediate approach (or, rather, considered from the standpoint of their origin, they leave this surface) from two general directions: (1) Most of those on the outer side of Hensen's stripe contribute first to the striation of the upper surface, then curve downward and then inward to contribute to the striation of the


The Xature of the Tectorial Membrane 135

Tflider surface; (3) most of those on the inner side of Hensen's stripe curve from tlie inner zone, first slightly outwards, then sharply downwards, and finally inwards to contribute to the striation of the under surface, thus following an S-shaped course. Apparently, these two general directions would result in a midregion either free from fibers, or, in which the general direction would be neither the one nor the other; but, in the sections, the transitions of curvature seem to be so gradual that this region is never definitely marked. Further, the two directions should result in a zone of the under surface in which the ends of the fibers from the two directions intercross. The sections show this intercrossing, and that, while it occurs to some extent in the body of the membrane, the greater part of it, as to be expected, occurs in the immediate under surface, and by no means all in one line (see Fig. 7).

The cut ends of the segments of the membrane represented in Figs. 2, 3, i and 5 were, of course, not possessed by those segments, but were outlined upon them and the arrangement of the fillers drawn in accordance M-ith the arrangement apparent from the study of the sections, in order to illustrate the relation between tlie fibers of the upper and under surfaces. The intercrossing of the ends of the fibers in the under surface is seldom apparent in the ordinary sections because such sections are usually more shrunlvcn than those represented in Figs. 6, 7 and 8, and the intercrossing, and indeed the whole parallel arrangement, is involved in the more deeply staining peripheral condensation of the substance produced by the reagents. In Fig. 7, Pc, for some reason or other during the manipulation, probably in the sectioning, this peripheral condensation seems to have partially peeled off from the under surface of the inner side of the membrane and the strip to have become stuck to the inner supporting cells of the organ of Corti.

Looking down upon and into the under surface of the freed membrane, with the compound microscope and with transmitted light, this surface, and all focal planes near it, appear studded with numerous fine dots which manipulation of the focus shows to be the ends of fibers. These ends are most numerous in and along either side of Hensen's stripe.

Hensen's stripe, or streak, in my preparations has a structure, and, I think, a significance. Hensen, who discovered it, and Eetzius and Schwalbe, who named it and described it as a hyaline, transparent strand on the under surface of the "middle zone," probably had to do with preparations shrunken at the periphery at least. Transparent


136 Irving Hardesty

preparations examined from the ventral surface, by transmitted light, show it to be fibrous and to be largely composed of a dense linear series of the intercrossing ends of the fibers approaching each other iu the ventral surface from the two directions. In fact, the stripe occupies approximately the position of the dividing line between the fibers coursing in this surface from the two directions. It is most probably distinguishable because it is a linear accumulation of these intercrossing ends which, as such, behave toward the light differently from the other regions. For the same reason, it was distinguishable in my preparations by reflected light, that is, when viewed over a black surface, as well as by transmitted light. It is strong enough to appear quite distinctly through the membrane when studied from the upper surface.

Von Ebner, '02, notes that Hensen's stripe may show from the upper surface, and thinks that it is produced by a thickening or bunching of the substance of the membrane, and he accepts the statements of others that it lies over the line of the inner hair cells of the organ of Corti. He, Coyne and Cannieu, '85, and Dupuis, '94, state that it is absent in some animals. The latter notes that it varies in different species, and that it may be modified by pressure of the cover glass, not leaving one to feel positive that the stripe is ever normally absent. Hensen himself thought the stripe a line of connection with the inner hair cells, and Shambaugh, '07, describes it as "a sort of fascet where, normally, the membrane is attached to the supporting cells just internal to the inner row of hair cells." The latter description is based upon the appearance of the membrane in sections, largely one section, manifestly distorted. In none of the sections of the pig's cochlea used here was it possible to locate Hensen's stripe by any means as positively as Dr. Shambaugh was able to do with his sections. It is of necessity always involved in the peripheral condensation. Denser, or more condensed, spots were often seen cut across in the under surface, but these were too numerous and too variously placed to be interpreted as other than sections of crimps in the peripheral condensation produced by the further shrinkage of the membrane during the preparation for sectioning.

In removing two of the preparations used here from cochleae fixed and decalcified in Zenker's fluid, the tectorial membrane was found to be stuck down to the organ of Corti in the regions of the first and second turns of the spiral. By careful use of the needle one of these was freed, and, though finally broken, the pieces were stained in fuchsin and mounted, under surface upward. Though disappointing and practi


The I^ature of the Tectorial Membrane 137

call}' useless for the purpose in mind, this preijaration was of interest with reference to the position of Hensen's stripe. Cells and portions of cells of the organ of Corti were adhering to the region involved, and, in one place, nearly the whole organ, including both rods, for a stretch of several cells. The ends of the hair cells could be distinguishel as such by means of the crescentic lines of the exit of their hairs, and it could be seen that the outer and inner rows of hair cells extended one on each side of Hensen's stripe. In other words, both by actually distinguishing it at times in their midst and by tracing it from the clean regions into the region involved, it was evident that, in preparations of the pig fixed in Zenker's fluid at least, Hensen's stripe lies between the inner and outer rows of hair cells, and therefore over the relatively smooth line occupied by the enclasped phalanges of the rods of Corti.

It is difficult to say whether or not Hensen's stripe exists as a ridge projecting beyond the general level of the under surface. A number of investigators have described it as such; others, less definite, mention it as a thickening or bunching of the material, and I remember none to have specifically denied its being ridge-like. Sections should be practically the only means by which this can be determined, but sections have proved very untrustworthy, because, even if a ridge is indicated in them, one is never sure it is not due to inequalities of surface shrinkage, and also because Hensen's stripe cannot be positively located among the variable thickenings and markings that different sections show. When viewed from the under surface, one gets a very decided impression of a ridge, but, under magnification, allowing the whole width of the membrane in the field, the greater part of the under surface appears roof-like with Hensen's stripe as the highest part. Careful focussing up and down under high power shows that the stripe is higher than the adjacent surface, but much less so than was to be expected, and that the roof-like appearance under low power was an optical effect, explained as due to the ends of the fibers converging, in the glassy matrix, from the two directions toward and into Hensen's stripe.

If the stripe runs along between the inner and outer rows of hair cells, one might infer that it is a ridge which represents a strip of the membrane never in contact with the projecting hairs, but which fits into the groove walled on either side by the hairs, and that it should have a width throughout corresponding to the width of the line of the exposed surfaces of the phalanges of the rods of Corti. The exposed


138 IrA'ing Hardesty

surface of the phalauges is widest at the apex of the cochlea and grows narrower towards the base of the coil, just as the rods and the angle they form decrease in size, and it was found that the stripe is somewhat wider at the apex and decreases in width toward the base, but at no level was it found to be as wide as the groove floored by the phalanges, nor does it vary in width as much as the groove. The only inference drawn from this study is that Hensen's stripe probably has the form of a low ridge the existence of which may be protected by the groove over which it lies.

In looking over papers dealing with the development of the tectorial membrane and the sections from pig foetuses prepared here for the purpose, the possibility was suggested that Hensen's stripe may have an embryonic significance: namely, that it represent? that area or line on the under surface which is last to become detached from the cells which give origin to the membrane.

On the Development of the Membrane.

It is not the object of this paper to describe in detail the development of the tectorial membrane. All the essential stages of the process have already been worked out quite thoroughly. Kolliker, '61, Middendorf, Eosenberg, Gottstein, Nuel, Eetzius, Schwalbe, Bottcher, Hensen, Prichard, Exner and Eickenbacher, '01, all agree that, in vertebrates, the membrane is a cuticular structure arising from the thickened epithelium of the inner side of the floor of the embryonic cochlear canal, and Kuhn and Hasse have described a similar origin for invertebrates. I find that in pigs of 2.5 cm. there has begun the marked thickening of the epithelium of that side of the cochlear canal which will become the portion to line the spiral sulcus and overlie the inner portion of the basilar membrane. At 3 cm. the thickening of the epithelium has increased and there appears above the entire thickening a cuticular film of quite appreciable thickness and decided fibrous character. The later stages, up to about 12 em., show increase in the width of the epithelial thickening corresponding to the increase in the diameter of the cochlear canal, consequent increase in the width of the developing tectorial membrane and marked increase in the thickness of the epithelium concerned. At 12 cm., or shortly before, there is differentiated a second and much smaller epithelial thickening along the immediate outer edge of the first one, thus giving the greater and the lesser epithelial thicken


The ^N'ature of the Tectorial j\Iembrane 139

ings, "'pads,"' noted by Kolliker, "61, and more fully described by Bottcher, '69, and Eickenbacher, '01. The lesser thickening is the first indication of the differentiation of the organ of Corti, while the cells of the greater give origin to the tectorial membrane and later gradually retract and sink back into the low, indifferent cells lining the spiral sulcus. Bottcher thought that the tectorial membrane takes origin from both the greater and the lesser thickenings. This can hardly be claimed, for the pig, for while the edges of the two thickenings are, of course, quite close together, the early membrane does not begin to extend over the lesser till there are evidences of the beginning of retraction on the part of the greater thickening. Evidences to the contrary of this are sometimes seen in the form of a thin, frayed reticulum sticking to the inner edge of the lesser thickening and continuous with the membrane, but serial study suggests that such appearances are due to distortion and to coagulum films produced by the reagents. Two years after its publication, Hensen showed that Bottcher's conclusion was based in part upon artifacts.

As Eickenbacher pointed out, the membrane, throughout its elaboration, undergoes no active growth whatever, but is purely a passive product of the continued activity of the epithelial cells below it. Not till pigs of about 14 cm. do my preparations show evidences of differentiation of the cells of the lesser thickening into what will become the different cells of the organ of Corti (Fig. 10), and at that stage none of the cells of the organ are in the least recognizable until traced backward from the more advanced stages. The contention of Ayres, '91 and '98, partially supported by Czinner and Hammerschlag, '97, that the tectorial membrane consists of a segregation of continuations of the hairs of the hair cells of the organ of Corti. is shown in my preparations to be fully met by the statement of Eickenbacher, namely, that the tectorial membrane is considerably developed some time before the hair cells are differentiated as such. Eickenbacher showed further that much of the findings of Czinner and Hammerschlag in this relation were artifacts, and cites Hensen in suggesting the same to have been true with Bottcher.

By a process of retrogression and displacement, first suggested by Kolliker and described later by Exner. '97. the tectorial membrane developed wholly from cells at the inner side of the organ of Corti, finally becomes so left that its outer part extends over the organ. At first, when the lesser epithelial thickening is not apparent in the prepa


140 Irving Hardesty

rations used here, a thin fibrous film may be discovered overlying the entire surface of the greater thickening and inward upon what will become the labium vestibulare. At the very beginning, the then very short fibers stand vertical to the epithelium and are even then agglutinated by a seemingly fluid-like stage of the transparent matrix. From pigs of 3 cm. up to 15 cm. the more important changes taking place in the floor of the cochlear canal are a marked increase in both the thickness and the width of the greater epithelial thickening and the appearance and differentiation of the lesser thickening.

At the stage when the labium vestibulare is evident, the greater thickening wholly occupies the area of the future spinal sulcus and consists of a high, pseudostratified epithelium whose surface is level with the upward pointing tip of the labium itself (Fig. 10). As to be inferred from this condition, the continuation of the epithelium upon the labium is very much lower than that of the thickening proper. In fact, this portion comprises the inner, tapering edge of the thickening at the stage before occurs the differentiation of the mesodermal tissue into the limbus spirale and its labium vestibulare and when the whole is covered by the delicate beginning of the tectorial membrane. The epithelium upon the labium is thus thin from the first, remains so throughout, and takes part in the production of only the thin, tapering, permanently attached, inner zone of the tectorial membrane. In realit}^, it ceases as an epithelium, and probably altogether, in the adult cochlea, for then the only superficial cells of the labium are those which are situated between and level with the horizontal connective tissue bundles comprising the auditory (Huschke's) teeth (see Fig. 9). Probably it ceases quite early to produce tectorial membrane, for, from pigs of 12 and 14 cm. upward, the inner, attached zone of the membrane gains apparently very little in thickness.

As the greater epithelial thickening increases in width during the earlier stages of the development, the fibers of the tectorial membrane, standing vertical to the epithelium at the very first, are gradually drawn outward from the more firmly attached inner zone, so that, by the time a membrane has gained a more appreciable thickness, 10 cm. pigs, the first formed portions of the fibers come to lie horizontal or nearly parallel to the surface of the epithelium while their continuations curve downward till the growing ends of the fibers only are vertical to the cells of the parent epithelium. In this way is produced the transverse striations apparent in the upper surface of the adult membrane and


The Nature of the Tectorial Membrane 141

the curved course of the fibers apparent in sections of it. All sections of foetuses show that the stages of development advance from the basal coil toward the apical, and so in the increase in thickness and width of the greater epithelial thickening, beginning at the base and advancing toward the apex of the cochlea, must lie the explanation of the fact that the outer direction of the fibers is inclined from the radial toward apex. Measurements from my preparations show that between 10 cm. and 15 cm., the width of the greater epithelial thickening increases from 124 microns to 174 microns. This increase consists of a very evident increase in the number of the epithelial cells concerned.

At about 14 cm., the lesser epithelial thickening, the future organ of Corti, has considerably advanced in form and appears as if it acts as a block to the further increase in width of the greater thickening, for, at that stage, the outer edge of the latter becomes bluntly rounded and bows upward (Fig. 10) with the result that in this edge, adjacent to the lesser thickening, there is formed the thickest part of the greater thickening. The outer edge of the developing tectorial membrane necessarily bends downward around this bluntly rounded edge of its parent epithelium. Up to this stage, the membrane never overlaps the lesser thickening and in confirmation of the statement of Eickenbacher, it must be said that at no stage is there good reason to assume that the cells giving rise to the organ of Corti ever have any thing to do with its development.

The order of the development of the membrane seems to be, first, the fibers and then the matrix, both being subadded gradually at the under surface. In all stages, from the very first, there always exists upon the immediate surface of the epithelium a thin, clear, parallel layer in which the matrix is either in a soluble stage of its development or wholly absent, and across which course the ends of the fibers from the epithelial cells into the body of the membrane above. In sections the thus unembedded ends of the fibers appear exceedingly tender, are often broken free and usually appear in agglutinated bundles (Fig. 10).

The stage of retrogression of the greater epithelial thickening, and the resultant displacement of the parts, begins with a cessation of growth of the epithelial cells, followed by their exhaustion and finally by a liquefaction of their remains, till, in the adult cochlea, the cells of the greater thickening are represented by the very low, indifferent epithelium which lines the spiral sulcus from the inner supporting


142 Irving Hardesty

cells of the organ of Corti to the edge of the labmm vestibulare. The exhaustion begins along the inner side of the greater epithelial thickening, at the edge of the labium vestibulare and gradually proceeds outward. Thus the first portion of the tectorial membrane to become free from its parent epithelium, that is, to cease growing, is the strip adjacent to the labium vestibulare. Sections from pigs of 15 and 17 cm. usually show a beginning recession of the epithelium in this region, and it is probably indicated in Fig. 10. (More strictly speaking, the portion of the membrane which first ceases to increase in thickness is, as mentioned above, the inner zone, attached upon the surface of the labium vestibulare.) The exhaustion of the epithelial thickening gradually passes outward till the last region of it that is active and attached to the tectorial membrane is its outer edge, or the region adjacent to the organ of Corti, which region, in the previous growth changes as noted above, became its thickest region. Finally this region recedes, the remains of the cells liquefy and the tectorial membrane is left entirely free from the cells which give rise to it. Transverse sections of the cochlear duct of pigs from 20 to 25 cm. often show a mass of disintegrating epithelial cells lying against the inner supporting cells of the organ of Corti, which mass represents the vestige of the greater epithelial thickening.

By the above processes of increase, recession and disintegration, the disposition of the fibers within and in the under surface of the tectorial membrane may be explained as well as the position which the membrane finally attains with reference to the organ of Corti (the lesser epithelial thickening).

The retrogi-ession of the epithelial cells undoubtedly results in a decrease in the width of the space occupied by the greater thickening (the space which becomes the floor of the spiral sulcus), and this decrease results in the inward displacement of the organ of Corti to its final position under the outer zone of the tectorial membrane. Counts of the cells comprising the greater epithelial thickening in the third half- turn of several cochleae from pigs of 14 to 15 cm. gave an average of 93 cells per transverse section. Counts of the cells lining the spiral sulcus, up to the inner supporting cells of the organ, in the same region of several fully developed cochlege gave an average of only 21 cells, the averages showing that the space in the fcetal cochlea contains four times as many cells as the developed. Measurements of the width of the floor of the spiral sulcus of the developed cochlea gave


The Nature of the Tectorial Membrane 143

an average of only 75 microns, while measurements of the width of the greater epithelial thickening in the 15 cm. pig gave an average of 174 microns, showing the space in the foetal to be more than twice as wide as in the developed cochlea. (The measurements indicate, further, what is shown in the specimens, that the epithelial cells of fhe fcetal cochlea, though much higher, are thinner than those of the developed.)

The direction of the fibers of the inner region of the body of the membrane is outward and almost parallel throughout. In this region multitudes of fibers are added to the body of the membrane, contributing to its sudden increase in thickness. The practically uniform outward direction here is due to the fact that these fibers were produced before the epithelium began to recede and the width of the thickening to decrease. Further outward, as far as Hensen's stripe, the fibers are directed outward, then downward, but still outward. At the region of Hensen's stripe, many fibers curve downward from their outward direction above and approach the under surface vertically. In fact, Hensen's stripe seems to be an expression of the period at which the retrogression of the epithelial thickening began. It also represents the line along which the thick, outer edge of the thickening was last attached and along Avhich growth was last contributed to the membrane.

With the exhaustion and retrogression of the cells of the inner portion of the thickening began the decrease in the width of the space occupied by it. Consequently, the portions of fibers produced by it after the decrease in width began were no longer directed outward or vertically, but inward, drawn inward, as it were, by the now receding outer edge of the thickening. As a result, there is in Hensen's stripe a crossing of the ends of fibers approaching it from the two directions, as noted above, in describing Hensen's stripe.

iVnd also as a result, the fibers coursing in the uiembrane on the outer side of Hensen's stripe are directed outward in the upper surface (the portions earlier formed), then curve downward (as do those on the inner side) and finally curve inward to approach the under surface at acute angles or to run short distances parallel with it. The tectorial membrane, firmly attached along its inner zone and sustained by its semi-solid matrix and its consequent elasticity, holds the position and shape in which it was molded; while the organ of Corti, during its elaboration, appears to follow the receding greater epithelial thickening and finallv comes to be situated under the outer zone of the tectorial


14:4 Irving Hardesty

membrane in such a position that the line of the phalanges of the rods of Corti about coincide with Hensen's stripe.

The greater width and thickness acquired by the tectorial membrane as its apical end is approached may be explained as due to the longer persistence and greater width attained by the apical portion of the greater epithelial thickening, noticeable in sections of foetal cochleae. All signs of the thickening may have disappeared in the basal coil when the coil of the apex may show considerable of the epithelium, and, in the adult, the floor of the spiral sulcus, as well as the entire spiral lamina, is left considerably wider in the apical than in the basal coils.

An Accessory Tectorial Membrane.

Though this structure as I shall describe it is undoubtedly present in my preparations, its description is entered into with some hesitancy because of the fact that in none of the papers I have been able to examine is such a structure noted for either the pig or other animals. The only explanation of why it has not been described before now is offered in the fact that it is so thin and fragile and so susceptible to the action of reagents that, with its existence not realized, it is impossible to distinguish in the ordinary preparations an}i;hing suggesting such a structure. Even after noting it in the unshrunken teased preparations, in only two of my transverse sections of the tectorial membrane, the best I have been able to get, could there be distinguished anything which might be interpreted as sections of this structure, however distorted. It was not evident on some of the broken teased preparations, because of the fact, soon learned, that it is apparently but very slightly attached to the main body of the tectorial membrane and is easily displaced or lost in the removal from the cochlea. In the teased preparations after reagents found to produce shrinkage, the "accessory membrane" was usually unrecognizable, appearing, from the" under surface, as an irregular belt of very delicate zig-zag lines which might be interpreted equally well as crumplings and shrinkage disarrangements of the surface and fibers of the main body.

The structure was first noted in a piece of tectorial membrane removed after Zenker's fluid and broken on the slide while mounting it in glycerine. Under low light, it appeared as an exceedingly thin transparent ribbon extending far beyond the broken end of the main bodv, in places slightly twisted, but for the most part flat and showing


The Mature of the Tectorial Membrane 145

parallel edges. Traced to the main body, it was found to be lifted off for some distance and to lie upon the main body much as a ribbon would appear if lifted by the end from slight adherence to a surface. Fig. 3 is an attempt to imitate this appearance of the structure, as seen in the preparation, the end of the main body, however, not being shown as broken, but being utilized to show the arrangement and relations of the fibers of the body. When viewed in its undisturbed position, it appears lying upon the under surface, as shown in Figs. 1 to 5, Ac, and so delicate it is, that one unaware of its presence, especially upon unstained or darkly stained pieces of the membrane, might consider the appearance produced by it as due to mere disturbances in the arrangement of the fibers of the under surface of the main body. That it lies practically free or, at least, very lightly adherent to the under surface of the main body is indicated by the apparent ease with which it comes away and by its occasional total absence from broken pieces. During its further study upon various preparations, short extents of it were at times seen crumpled away and free from the under surface, and at times deviating inward or outward from its general alignment. In several cases the deviation outward was sufficient for a small portion of its outer edge to project beyond the outer edge of the main body. Because it seemed to have a practically separate existence, it is referred to here as an Accessory Tectorial Membrane. Its structure, its delicacy, and its position explain the difficulty with which it is distinguishable and suggest that the probable reason it has not been noted before is because none of the investigators using teased preparations happened to obtain pieces of it isolated, extending free and undistorted, in the field of the microscope.

In shape and structure this accessory membrane somewhat resembles a long comb, the tapering teeth of which, however, are in the form of a net consisting of two sets of parallel fibers all continuous into the %ack of the comb" which is the outer edge of the membrane. The direction of one set of the fibers coincides with the direction of the fibers in that zone of the under surface of the tectorial membrane over which they lie. Thus, this set of the fibers cannot be distinguished when the accessory membrane is in position (Fig. 3). The other set of fibers crosses the first set obliqi^ely, arising at acute angles to outer edge of the membrane and coursing in a decided slant pointing toward the apex of the cochlea. The fibers are embedded in a thin film of transparent matrix of the same character as that supporting those of the


146 Irving Hardesty

tectorial membrane proper, and it is this matrix that maintains the shape of the accessory membrane and holds the two sets of fibers, and the fibers of each set, in their relative positions.

The inner edge of this membrane consists of nothing but the crossing inner ends of the fibers and is so exceedingly thin that it could not be distinguished at all when lying in position but for the optical effect produced by the obliquely coursing set of fibers. The outer edge (the back of the comb) is thicker than the inner, apparently resulting from a gradual increase in the amount of the supporting matrix. In the outer edge, the fibers of both sets seem to curl upward so as to produce a short abrupt curl in this edge of the membrane with the convex side toward the under surface of the tectorial membrane proper, as indicated in the drawing of the end in Fig. 3.

The question of the attachment of the accessory to the main body of the tectorial membrane coiild not be definitely settled. The accessory membrane was found so delicate that it would not stand manipulation; in more deeply stained preparations it could not be studied in its relations to the main body because of the light being obscured in passing through the whole structure; transverse sections, necessarily from dehydrated and embedded specimens, were entirely impossible, and, in the unstained and lightly stained mounts, if attaching fibers exist, they could not be distinguished in the maze of other fibers present on both sides of their probable locality. From the study made here, only this can be said : If the accessory membrane is attached other than by light surface cohesion to the main body, it is most probably attached along its outer, and thicker, edge. In several of my mounts, the inner edge appeared folded over for short distances as though it had been waving free in the fluids preliminary to the cover glass, while the outer edge retained its normal position as though it had behaved as a hinge. Further, the curled disposition of the fibers in the outer edge suggest that their ends may be continuous into the under surface of the tectorial membrane and that they may be broken, ripped asunder, in cases in which portions of the accessory membrane appear entirely free.

As to its position, the accessory membrane lies under the outer zone of the under surface of the tectorial membrane proper and to the outer side of Hensen's stripe. It exists throughout the length of the tectorial membrane.

Like the tectorial membrane proper, it is widest at the apical enrl and decreases in width toward the end of the basal coil. The decrease


The Nature of the Tectorial Membrane 147

in the width of the tectorial membrane proper occurs chiefly in the decrease in the width of the zone or strip on the outer side of Hensen's stripe. The decrease in the width of the accessory membrane is relatively much more gradual than that of this zone under which it lies. At its greatest width, in the apical turn (Figs. 1 and 2), its outer edge is considerably within the outer edge of the tectorial membrane jjroper, but its decrease is so much more gradual that by the time the seventh half-turn is reached, the outer edges of the two membranes coincide (Figs. 1 to 5). Also, at the apical end, its inner edge extends not quite to Hensen's stripe, but below, this edge comes to coincide with Hensen's stripe, and, at the basal end, the tips of its fibers may occasionally overlap the stripe. If Hensen's stripe coincides throughout with the line of the enclasped phalanges of the rods of Corti, then the accessory membrane can only come in contact with the hairs of the outer series of the hair cells.

Of the illustrations given of transverse sections. Fig. 7 alone shows a structure whose position warrants its being interpreted as a section of the accessory membrane. The structure is naturally attacked severely by the reagents, being very thin and fragile, and distortion renders it unrecognizable and often invisible in sections. One unaware of its existence would never discover it in even the best of sections.

Some of the observations recorded in the literature of the tectorial membrane may refer to this accessory membrane. It is somewhat probable that the irregular, transparent, reticular fringe described by Lowenberg, '64, as existing along the outer edge, and since referred to by some as "Lowenberg's border plexus" and by others as the third or outer zone of the tectorial membrane, may have been nothing more than the displaced outer edge of the accessory membrane projecting irregularly beyond the outer edge of the tectorial membrane proper. Coyne and Cannieu, '85, may have used this in their claim that the tectorial membrane is attached as far over as the cells of Claudius; Dupuis, '94, notes that Lowenberg's border plexus is very irregular and not alwa5rs present; Von Ebner, '02, in Kolliker's Gewebelehre, notes and illustrates it as filamentous; Kishi, '07, considers it an artifact. Hensen, '63; Bottcher, '72; Ferre, '85, and several others since them deny the existence of the border plexus. The latter may have been dealing with more nearly normal conditions. Further, several observers, to be mentioned below, claim filamentous attachments between the under surface of the tectorial membrane and the hairs of the hair cells and


148 Irving Hardesty

various other elements of the organ of Corti. It is probable that some of these attachments consist of crumpled bits of the distorted accessory membrane, augmented by coagulum filaments. Also some of the descriptions of various lines and markings on the under surface of the tectorial membrane, noted above, may have been dealing with appearances produced by the distorted accessory membrane as well as with crumplings of the peripheral condensation of the main body.

The processes by which the accessory tectorial membrane is developed from the greater epithelial thickening have yet to be worked out. Evidently it is one of the last structures formed, and probably its individual character may be due to its being formed during a period of rapid displacement of the parts.

The Attachment of the Tectorial Membrane,

All the observers invariably agree that the tectorial membrane is attached along its inner thin zone upon the surface of the labium vestibulare. This attachment is developed early in the development and is the line of fixation instrumental in the final course and arrangement of the fibers of the membrane resulting from, first the widening of the greater epithelial thickening and then followed by its retrogression and decrease in width.

Numerous investigators claim other attachments, but there is lack of agreement as to their location. There are to be found in the literature claims for attachment of the under surface of the membrane varying from attachment to the epithelium of the spiral ligament (Corti) and the cells of Claudius (Coyne and Cannieu), all the way across the lamina to attachment upon the inner supporting cells of the organ of Corti (Shambaugh). The majority of the claims are for attachment to the structural elements comprising the organ of Corti alone, especially to the hairs of either the inner or outer hair cells or both. Of the more recent papers, Kishi, '07, agreeing with Bottcher, '69, states that the membrane is connected with the lamina reticularis and with the hairs of both the inner and outer series. Siebenmann, 'OT), denies connection with the lamina reticularis in either the embryo or the adult. Von Ebner, '02, holds the membrane was attached to the lamina during development, but, in the adult, after the displacement of the parts, this attachment is necessarily lost and the membrane extends free over the spiral sulcus and organ of Corti. Eickenbacher, '01,


The Nature of the Tectorial Membrane 149

thinks, from his studies of the development of the membrane, that it is improbable, d priori, for the membrane to float freely in the cochlear duct. He did not carry his studies through the processes of the displacement. Shambaugh, '07, assumes that the membrane is attached to the inner supporting cells and that Hensen's stripe serves as a fascet for this attachment. Kolmer, '07, often found the membrane attached to various of the cells of the organ of Corti, but thinks these attachments were artifacts. As early as 1872 Gottstein described "transparent filaments" extending between the membrane and the inner supporting cells of the organ of Corti of foetal cochleae, but considered these a.s remnants of liquefying processes. Hensen, '07, in elaborating a theory for tone perception, assumes and pictures the tectorial membrane as extending free over the organ of Corti so that the hairs may brush against it during the agitations of the endolymph, resulting in auditory stimuli. The hairs have been described as sticking into the under surface of the tectorial membrane.

All the claims for the attachments of the membrane to any portion of the epithelium of the spiral lamina (including the organ of Corti) are based upon appearances presented in sections, and all describe the connections as filamentous. The studies made here tend strongly to the view that the membrane is attached only along its inner side and only upon the labium vestibulare.

(1) During the teasing away of the upper wall of the membranous labyrinth to expose the membrane preparatory to its removal, the outer portion could at times be easily seen under the dissecting microscope to wave upward and downward in response to agitations produced in the fluid with which it was covered during teasing. These movements seemed to occur along its entire course. They occurred within very small arcs and were suggestive of a certain amount of elasticity in the membrane, causing it to retain its horizontal position because of the attachment along its inner zone. If attached to the elements of the organ of Corti, or even to the hairs alone, while individually such attachments would be delicate and easily broken, it is hardly probable that the necessarily great multitude of such attachments would so readily disappear and allow the movements.

(2) Sections of fixed and dehydrated material are but little more suggestive of attachment of the membrane than they are of its being free from the organ of Corti in the adult condition. The membrane is alwavs more or less shrunken and distorted, and these effects of the


150 Irving Hardesty

manipulation may result either in its being drawn in contact with the organ below or in its being lifted well away from the organ. Sections often show it even pressed upon the organ, but the majority of sections show it directed in various angles, entirely free from the organ, and many of the older pictures of it are thus drawn. When caught more nearly in its normal position with its under surface parallel with and near the surface of the organ of Corti, delicate filamentous connections are often apparent, but such connections can be justly explained as brought about by crumplings or abrasions of the fibers in the under surface, or by bits of the very thin accessory membrane coming in contact with the auditory hairs or other elements, the appearance of continuity being accentuated by adhering coagulum filaments precipitated from the albumins and globulins of the endolymph.

(3) From the process of its development, it seem probable, contrary to Eickenbacher, but in accord with Von Ebner, that, a priori, the membrane is free from the underlying epithelial structures. During its development it is attached only to the epithelial cells giving origin to it, and since its outer zone acquires its adult position over the organ of Corti only by the recession of the parent epithelium and the consequent displacement of the organ, one must at least assume that, if attachment to the organ exists, it must be developed after the development of both the organ and the membrane. This is hardly probable and, moreover, it is hardly necessary to an explanation of the role of the organ in the phenomena of hearing.

On the Anatomy of Hearing.

It may be said that the tectorial membrane is comparable and analogous to the otolithic membranes of the cristas and maculae acusticae, except that it is neither developed by nor upon cells belonging in the group of special sensory cells with whose stimulation it is concerned, and except, further, that in the higher animals it is not beset with calcareous products.

Lavdowsky, '77, and others, including Ayers, '91, have called attention to the fact that, from the nature of the stimuli, the function of hearing must require three mechanisms: (1) A vibratory mechanism; (2) a regulation mechanism, and (3) a mechanism for the perception of sound stimuli.

Hensen, '63, was the first to advance the idea of the hairs of the hair cells being brushed against the tectorial membrane for the origin


The Nature of the Tectorial Membrane 151

of the stimuli in the perception mechanism. Lavdowsk}', Hensen, Nuel and others describe the basilar membrane, in the spiral lamina, as being composed of independent, radially disposed fibers. This and tlie further observations that the basilar membrane is considerably wider (and therefore its fibers longer) at the apex than in the basilar coils of the cochlea, led Helmholtz to advance the theory of hearing which bears his name. This theory involves the sympathetic vibration of the different "fibers"' of the basilar membrane in resonance with the atmospheric waves as transmitted to the endolymph by the tympanic membrane and auditory ossicles, and it assumes that the vibrations of those fibers involved by a given stimulus cause the hairs of the hair cells overlying the fibers to rub upward against the tectorial membrane. It thus assumes, further, that the cochlea has the power of analysis of sound or that the perception of tone is mediated by different parts of the cochlea. Ter Kuile, '00, in a purely theoretical paper dealing with the transmission to the hair cells of energy from the supposedly vibratory basilar membrane, shows that the hairs do not rub against the tectorial membrane, but strike its under surface vertically. Ewald, '99 and '06, accepts the idea of the basilar membrane and experiments with thin rubber membranes placed under fluid, finding that vibrations may be induced in them in response to atmospheric waves. In fact, with but two exceptions that I know of, all of the investigators of physiological acoustics since Helmholtz have accepted the assumption that the basilar membrane is the vibratory mechanism, and in their writings have merely put forth various modifications of the original conception. The two exceptions are Kishi, '07, and Shambaugh, '07, both of whom assume that the tectorial membrane is the vibratory mechanism, ]ioth claiming the power of resonance in it, as will be mentioned below.

As noted by Gray, '00, the explanations of hearing as now offered may be divided into two general theories: (1) The Eesonance Theory, suggested by the work of Hensen, ISTuel and others, propounded and elaborated by Helmholtz and variously modified ever since in its application to the cochlea ])y different investigators, and (3) The Telephone Theory.

The latter theory was suggested by Einne in 1865, supported by Voltolini in 1885, more fully elaborated by Eutherford in 1886, modified by Waller in 1891, and further modified by Max Meyer in 1898. It assumes that the transformed sound waves as imparted to the endolymph affect the structures of the cochlea as a whole.


152 Irving Hardesty

Eutherford first advocated that all the hairs of the hair cells vibrate equally to every note and that the nerve impulses thus aroused are mere vibrations similar in frequency, amplitude (intensity), and character (quality) to the soimd vibrations, and that, therefore, sound analysis is wholly cerebral.

Waller and Meyer assumed that the transformed sound vibrations in the endolymph act upon the basilar membrane as a whole, repeating the vibrations of the tympanic membrane. Meyer supposes that each wave, as it passes up the scala vestibuli, presses the basilar membrane downwards and does this just in the proportion in which amplitude of the wave is not decreased by the resistance it meets in passing upward toward the apex. Thus each wave will produce irritation of a certain number of hairs and the intensity of the irritation will diminish as the amplitude of vibration is diminished by resistance; and therefore a wave of greater amplitude will involve a greater extent of the basilar membrane, and of auditory hairs, than a wave of lesser amplitude. Certain waves, sooner than others, in passing up the scala will become too faint to sufficiently press down the basilar membrane. According to ^ this view, pitch depends upon vibration frequency or the number of stimuli per second, while intensity depends upon the total number of nerve endings irritated. It allows a certain amount of analysis in the cochlea.

The telephone theory differs from that of Helmholtz in that it assumes that the basilar membrane, instead of certain of its fibers vibrating in sympathy with given notes, vibrates as a whole to every note in so far as the original amplitude of the wave and resistance to propagation will allow, and that the auditory nerve fibers transmit to the brain stimuli of frequencies and intensities of the note or notes concerned. The resonance theory supposes analysis of sound by the cochlea; while the telephone theory supposes distinctions of sound, perception of tone, etc., to be accomplished by the brain, made possible of course by the varying quality of the stimuli in the peripheral organ, the cochlea. The resonance theory of necessity requires that each fiber of the basilar membrane and each arch, or pair of rods, of the organ of Corti be able to vibrate or move upward and downward independently.

Numerous objections to the resonance theory have been made, both physiological and anatomical. Of the physiological objections, two may be cited as examples of those the obviation of which has been recently undertaken bv adherents of the resonance theory.


The Nature of the Tectorial Membrane 153

(1) A mixture of notes of closely approximate beats, even having varying intensity, cannot be analyzed by the auditory apparatus; rather, only the maximum stimulus in the mixture is perceived. Gray, '00, attempts to obviate this with a modification of the resonance theory, in which he compares the sense of hearing with that of touch. He supposes that the basilar membrane does not vibrate as to its individual fibers, each in response to a given note, but rather in areas of its extent for each note, and thus he assumes that the mixture of approximate notes involves a given small area of the basilar membrane from which only the maximum stimulus is appreciated by the brain, all others of the mixture being neglected, just as when a small area of the skin is stimulated by a very bluntly pointed instrument and the "mind only pays attention" to the point of maximum stimulation; or, just as when two points on the skin, close enough together, cannot be distinguished from a single point when stimulated simultaneously. Evidently this modification also obviates the requirement of the resonance theory that each arch or pair of rods of the organ of Corti, as well as each fiber of the basilar membrane, must be able to move or vibrate independently.

(2) It is generally conceded that sensations of "noise" as distinguished from the perception and analysis of tones cannot be adequately explained by the resonance theory. Hensen, '07, has recently put forth an elaborate modification of this theory in which he tries, among other things, to account for this inadequacy. It is unnecessary here to discuss his ingenious paper further than merely to state that he retains the assumption that the basilar membrane is composed of fibers with the power to vibrate individually, allowing that probably several adjacent fibers, being of approximately the same length, may vibrate together with the respective note. To do this, he finds it necessary to add the assumption that the cochlea mediates only musical tones, and he calls it the "musical organ." This makes it needful for some other structure to mediate sensations of noise, etc., and for this he finds the cristas of the semi-circular canals and the maculae acusticse adequate.

Wliile, strictly speaking, the two objections or inadequacies cited are anatomical as well as physiological, there are some which may be considered more purely anatomical, and therefore physical, objections to the resonance theory.

From the nature of available information, the intangibility of the apparatus, it has been the rather common custom in discussing the physiology of hearing, to first assume the existence of anatomical


154 Irving Hardesty

characters and then to base explanations of the phenomena, many of which may be largely psjx-hic, upon the results of the application of excellent mathematics and physics to the structures assumed. The resonance theory must depend essentially upon the character it assumes for its vibratory mechanism. Therefore the anatomical objections to it must deal with these assumed characters. Among these objections may be enumerated the following:

(1) Supposing the basilar membrane to be composed of individual fibers capable of vibrating independently, its environment in the lamina spiralis is such that its vibration in response to many of the transferred sound waves, and certainly with any degree of sensitiveness, seems very improbable. The basilar membrane, the fibrous portion of the spiral lamina, is closely invested by two continuous layers of tissue on each of its sides. On its tympanic side, there is first the layer of endothelium lining the scala tympani (see Fig. 7), and second, the much thicker layer of epithelioidal cells which contains further, that anastomosing plexus of blood vessels known as the vas spirale, which is only approximately a single vessel in the end of the basal coil, but which elaborates with increasing complexity toward the apex of the cochlea. On its vestibular side, it is covered first by a "homogeneous" and well developed membrana propria of the neuro-epithelium (see A"on Ebner, loc. cit., page 927, Fig. 1450), and lasth', the whole is covered by, or supports, the neuroepithelium including that comprising the thick organ of Corti. Outside the animal body, structures of such relative thickness adhering throughout to a system of strings would effectually damp at least their resonant vibration. Further, if the basilar membrane were vibratory in the degree supposed by the theory, the circulatory pulsations in the vas spirale would very probably tend to at least confuse agitations produced by resonant vibrations of its directly overlying fibers, even with the habitual neglect by the auditory apparatus of the blood pulsations themselves.

(2) The resonance theory requires, and, truly, the earlier pictures of it assume (ISTeuel, "72, for example), that the basilar membrane be composed of a single set of radially disposed fibers. Ayers, '91, describes it in man as composed of four layers of fibers, three of which run radially from the lamina spiralis ossea into the ligamentum spirale to terminate at the base of the stria vascularis, and the fourth, a thin layer, running at right angles to the other three. If these four layers exist, there is nothing to be observed in either sections or teased prepa


The jSTature of the Tectorial Membrane 155

rations, indicating that they do not lie in direct contact and probably adhere to each other, and such an arrangement must tend to interfere with sensitive resonant vibration.

(3) Assuming that any or all parts of the basilar membrane proper consist of independent fibers, the question may be asked whether the membrane is capable of being thrown into vibration at all by sound waves. Ewald, '99, constructed a model in which he was able to so stretch a very thin rubber membrane that he succeeded in getting pieces as small as 0.5 mm. broad to vibrate when suspended in water by sound waves in the air. The recent measurements of the assumed vibratory width of the basilar membrane by Kolmer, '07, show this width to vary from only about 0.2 mm. (168 microns) in the basal turns to only 0.3 mm. (304 microns) in the apical turns. Shambaugh, '07, thinks that Ewald's model falls far short of proving that the basilar membrane, which he describes as being "much shorter, thicker and more rigid" than Ewald's rubber membrane, vibrates in response to sound waves transferred to the endolymph, and he mentions the fact that Helmholtz himself appreciated the doubt as to whether fibers so short as the width of the basilar membrane can be thrown into vibration by sound waves. Certainly its relative width and thickness, coupled with objections 1 and 2 above, make its vibration reasonably doubtful. Shambaugh describes the basilar membrane, in his preparations of the cochlea of the pig, as becoming so thick arid rigid in the basal coil, "a considerable distance" from its basal termination, as to preclude the idea of its being a vibratory structure. In one specimen he found complete absence of the basilar membrane in the basal end of the coil, but instead the perfectly formed organ of Corti, with its tectorial membrane, rested upon a direct junction of the crista of the spiral ligament with the labium tympanicum of the spiral limbus. In another specimen, he found the perfectly formed organ of Corti resting upon a solid bony plate bridging the width between the lamina spiralis ossea and the outer wall of the cochlea. Such absences of the basilar membrane under a completely formed organ of Corti, suggest that it is not essential to the sense of hearing either as a structure vibrating in accordance with the resonance theory or in accordance with the telephone theory.

(4) To the above anatomical objections to the resonance theory may be added evidence that the basilar membrane is not composed of individual and independent fibers at all. The idea of resonant vibration as implied in the theory demands the existence of such. Some of the


156 Irving Hardesty

older papers illustrate the fibers as separately distinct, and even counts of them are claimed, varying from 15,000 to 25,000. And measurements of their thickness have been given, varying from 1.5 to 2.3 microns. One of the most recent modifications of the theory, that of Hensen, is founded wholly upon the existence of separate fibers.

After becoming convinced by micro-chemical tests that the basilar membrane consists of white fibrous connective tissue, some horizontal sections of pig cochlea were prepared here and stained by Mallory's method for white fibrous tissue. The sections were necessarily cut quite thin in order to get areas containing the basilar membrane alone, exclusive of its coverings, so that its actual construction might be studied. Fig. 11 is an attempt to show the appearance of fiat sections of the membrane as brought out by Mallory's method. From these preparations, and afterwards even from the pieces obtained from the teased cochleae stained with fuchsin, it became evident that the basilar membrane is nothing more nor less than a thin flat tendon, and that its so-called "fibers" correspond exactly to the well known fiber-bundles in tendon fasciculi, being by no means independent of each other even as bundles. Each bimdle is, of course, composed of multitudes of fibers so fine that, even under oil immersion, one can never separate them with certainty as to the divisional units of the structure, much less measure them. The bundles (fibers of the basilar membrane), just as the bundles in a tendon fasciculus, are abundantly connected or continuous with each other by myriads of fine, silver-like collateral fibers inserted at all conceivable angles, the majority slanting in the direction of the bundles. In other words, the basilar membrane, instead of being formed of independent fibers, is a sort of a feltwork or modified reticulum of interwoven fibers, the general direction of most of which is radial to the axis of the cochlea, but withal, the resulting structure is such as to wholly preclude the idea of the independent vibration of adjacent fibers. True, in itself, it does not preclude the telephone theory of hearing. This radial or parallel tendency is no doubt due to the fact that, as the cochlea grows, the basilar membrane is developed from the mesenchyme under tension, just as tendons are, and thus it becomes stretched taut as the fioor of the ductus cochlearis and merely serves for the support of the neuro-epithelium in the position for its functioning. The membrane differs from an ordinary flat tendon in the distribution of its nuclei. These, instead of being arranged throughout in the varying columns of tendon cells, are absent under the region of the organ of Corti. As


The Nature of the Tectorial Membrane 157

shown in Fig. 11, they appear between the bundles, in a belt near the beginning of the spiral ligament in the outer edge, and, in the inner edge, the nuclei of the labium tympanicum encroach but little further than the foramina nervosa of the habenula perforata. The membrane no doubt grows at the edges, and the midregion being oldest, its cells are probably exhausted.

(5) The resonance theory, with its assumption of individual fibers in the basilar membrane, as mentioned above, carries with it the idea that the different elements of the organ of Corti, including the arches, or pairs of rods of Corti, must be capable of moving separately. Not only do anatomists agree that the neuro-epithelium forms a continuous membrane throughout the cochlea, but the component cells of the organ of Corti are easily seen to be intertwined and interwedged among themselves. Further, the rods of Corti's organ are probably more firmly adherent to each other in their series than are any other elements of the organ. This may certainly be said of the series of outer rods. In teasing the cochlea for the tectorial membrane, I found in several preparations when mounted, considerable extents of the outer rods adhering to each other, appearing on the flat as a grill-work, and wholly free from other elements of the organ. From one specimen, fixed and decalcified with Perenyi's fluid, which has a macerating effect upon tissues unless followed by washing in alcohol, I succeeded in obtaining a tier of the outer rods more than 5 mm, long and so firmly coherent that it floated about in the teasing fluid and withstood manipulation sufficient to mount it entire. Occasionally short tiers were found in the mounts of the tectorial membrane, sometimes adhering to its edge and lying out flat and otherwise free so as to be easily studied. Fig. 12 represents an end of one of the latter, and was chosen because it had been torn from the series in such a way that it afforded an excellent illustration of the shape, interrelations and character of the outer rods. This preparation was fixed in Zenker's fluid. The inner rods are manifestly not so resistent as the outer, for, while distorted scraps and fragments from their dissolution were found in the mounts, tiers of them alone were never found intact. The ends of the phalanges of the outer rods often showed frayed bits of the phalanges of the inner adhering to them. All appearances indicated, however, that the rods of Corti, both inner and outer, must be quite firmly coherent both the inner to the outer and to each other in their series.

(6) The resonance theory applied, as it has been from the beginning, to the basilar membrane, requires not only that the rods (and the other


158 Irving Hardesty

elements) of the organ of Corti resting upon the vibratory fibers should be able to move up and down independently, but it is usually inferred that both rods of the arch must rest upon the basilar membrane throughout the extent of the cochlea. .In all of my vertical sections of the cochlea of the pig, and Shambaugh notes the same in his preparations, usually throughout the basal coil the feet of the inner rods of Corti do not rest upon the supposedly vibrating basilar membrane, but reach over to stand upon the thicker edge of the labium tympanicum. Ter Kuile, '"00, recognized this, but made use of it to assume that contact of the hairs of the hair cells against the tectorial membrane during stimulation is not in a vertical direction, but occurs as a lateral, inward brushing motion, the outer rods alone being moved by the vibrations of the membrane while the inner rods merely act as fulcra in producing the brushing motion. This assumption gains nothing but an increase in the complexity of the resonance theory, since, above the basal coil, the feet of both rods do rest upon the supposedly vibrating portion of the basilar membrane. It is fairly well indicated from clinical evidence that the nerve fibers from the basal coil mediate sensations of notes having the higher pitch, and ter Kuile's assumption merely adds the anatomically improbable condition that the neuro-epithelium is agitated by these higher notes in a different way than it is by the lower notes supposedly affecting the coils toward the apex.

(7) To the anatomist, given structures in a given species of animal are strikingly identical in character. For example, all cochleae of normal individuals of a given species are strikingly alike in all details.

The power of tone perception and tone distinction by different individuals is not as constant and uniform as would be expected with the apparently constant anatomical mechanism for the mediation of sound. Individuals without special training (habits) differ markedly, and it is well knowTi that the "trained ear" in any special direction shows a far greater power of analysis of sound and distinction of smaller differences of tone than the untrained. Most of the investigators of the ph5'-siology of hearing, including Helmholtz, recognize this and usually try to obviate it by assuming modifications in the mechanism of the cochlea.

More probably, as in all the other senses, the whole question of analysis of stimuli and the power of finer perception, inherent or acquired by education, must be transferred from the peripheral or receptive mechanism to the central nervous organ. Just as in touch or smell or sight, so in hearing, all being physiologically similar, the variety of


The Xature of the Tectorial Membrane 159

the sensation experienced depends only upon the quality and intensity of stimulus applied to the peripheral nerve endings. The specialization of sensation is doubtless with reference to general groups only, and this specialization dependent upon the locality of the cerebrum to which the impulses aroused by external stimuli are borne. The optic nerve, stimulated by touch (pressure) conveys impulses which are interpreted as sensations of light; the acoustic nerve, stimulated by touch (tumors, etc.) gives rise to sensations of sound. And it is very probable that within tlie general groups, most manifest in touch, identical nerve terminations acted upon by external stimuli varying in quality, give rise to sensations varying according to the quality of the stimuli.

It is considered neither possible nor necessary in this paper to consider fully the numerous modifications and additions to the resonance theory. Many of the papers are merely attempts to explain away the physiological (and psychological) difiiculties and anatomical objections to the theory as applied to the basilar membrane and the action of the organ of Corti. It seems to the author that these objections, taken collective^, are such that neither the resonance theory nor the telephone theory as applied to the basilar membrane are any longer tenable.

The fact has been frequently noted, and especially by ter Kuile, that the tectorial membrane, in all species, is always entirely coextensive with the organ of Corti, and is always sufficiently wide to cover the hair cells, whether these are in linear series on either side of the rods, as in mammals, or whether they are scattered throughout the entire epithelium of the organ of Corti, as in birds and reptiles. Since the same cannot be said of the basilar membrane, it seems very probable that the tectorial membrane is more essential to the sense of hearing. Further, because of its shape, its more logical position, above the auditory hairs, and its far more delicate sensitiveness, the tectorial membrane seems much better qualified to serve as the vibrating mechanism than the basilar membrane.

Siebenmann, '00, seems to have been the first to suggest that the tectorial membrane is the structure agitated by sound waves, though the inference drawn by Ayers, '91, amounted to something similar. Quite recently Kishi and Shambaugh, realizing the anatomical disqualifications of the basilar membrane, have attributed the power of resonance to the tectorial membrane.

Kishi, '07, after citing four anatomical objections, concludes that the basilar membrane is not qualified for the vibrating mechanism of either


160 Irving Hardesty

the resonance theory or of any other theory of hearing. He substitutes the tectorial membrane as being more suited to the purpose, and then, from some rather distorted sections of the cochlea (illustrated with photographs), he assumes that the tectorial membrane is attached along its outer edge to the laminea reticularis of the outer supporting (Hensen's) cells of the organ of Corti, and thus claims that it is held down firmly upon the organ and always at right angles to the auditory hairs. Being also firmly attached along its inner edge, as all admit, Kishi describes and diagrams it as a membrane stretched tautly over the organ from the labium vestibulare of the spiral limbus to the lamina reticularis of Hensen's cells. He then assumes that the length of the fibers composing the membrane ocrresponds throughout to the breadth of the membrane in the respective turns of the coil. Admitting that his assumed absolute length of the fibers (the breadth of the membrane) cannot be determined because of shrinkage and distortion produced 'by the technique, he finds, from measurements of the width of the membrane in his sections, that the relative length of the fibers in the apical region is at least three times greater than those of the basal coil. Therefore he assumes that the fibers have different spans, vibrating lengths, in the different regions of the cochlea with all conceivable variations between the longest in the apex and the shortest in the basal end, and thus it is to be inferred that the tectorial membrane is a resonant structure composed of fibers capable of sympathetic vibration to all sound waves of lengths between and including those which may affect its longest and shortest fibers. He realizes that the fibers course obliquely across the membrane, slanting toward the apex, and that, therefore, their actual lengths are greater, especially in the apical region than the width of the membrane as seen in radial sections.

Against Kishi's conception of the tectorial membrane, three objections may be urged:

(1) From the processes of development of the membranes, it is very improbable, if not impossible, that any of its fibers in any region extend across the entire breadth of the membrane.

(2) From the processes of its development, it is probable if not certain that the great majority of the fibers are attached at neither of their ends, but merely lie embedded in the transparent matrix.

(3) From the studies made here and by others mentioned of preparations as nearly normal as possible, it is concluded that the tectorial membrane has but one attachment in life, namely, its inner edge upon the


The Nature of the Tectorial Membrane 161

labium vestibulare of the spiral limbus. In the apical turn, at least, its outer edge cannot be attached to the lamina reticularis, for this edge in the first half-turn (Fig. 6) may extend into the space of the ductus cochlearis a distance of nearly half its width beyond the organ of Corti.

Shambaugh, '07, urges the objections, accredited to him above, against the possibility of the basilar membrane being a resonance mechanism, and, from appearances in some of his sections, concludes (1) that the tectorial membrane does not lie free in the endolymph over the organ of Corti, but is attached along Hensen's stripe to the inner supporting cells of the organ; (2) that "the hairs of the hair cells project normally into the under surface of the tectorial membrane"; (3) that "the size of the membrana tectoria near the apex of the cochlea is many hundred times its size near the beginning of the basal coil"; and (4) he concludes further that, in structure, the tectorial membrane consists of "au immense number of delicate lamellae taking their origin from the portion of the membrane which rests upon the labium vestibulare." From these conclusions and the conclusions generally accepted as to the relation of the tectorial membrane and the hair cells to stimulation of the acoustic nerve terminations, he advances a theory that the tectorial membrane consists of a series of resonators (the lamellae) which are capable of responding to the most delicate impulses passing through the endolymph, claiming that the great variation in size of the membrane from one end to the other, suggests the possibility that, by acting the part of a resonator, it is capable of responding in different parts to impulses (tones) of different pitch. Shambaugh, in applying his theory, claims with reason that all the observations by Helmholtz and his followers supporting their resonance theory apply more readily to the tectorial membrane as a resonator than to the basilar membrane, and proceeds to explain the pathological phenomena of "tone islands" "diplacousus hinauralis dysharmonica" and "tinnitus aurium" on the basis of a resonating tectorial membrane.

Some of the observations both made and cited in this paper are not in accord with the premises from which Shambaugh's theory follows. In the first place, it is considered probable that the tectorial membrane does lie free over the organ of Corti and that the auditory hairs do not project into it. In the second place, it is not a lamellated structure. Ever since 1869, when Bottcher teased portions of it and found them to contain fibers, the fibrous structure of the membrane has been conceded by all who have studied it with reference to its structure. Sections in


162 Irving Hardest}'

different planes, as made by Coyne and Cannieu, '85, and here (Fig. 9), indicate clearly its fibrous nature.

It is not the intention of the paper to elaborate another theory of hearing and to enter into a necessarily prolonged application and defense of it. Leave is asked to merely suggest a modification of a theory already advanced, namely an application to the tectorial membrane of the telephone theory, heretofore applied exclusively to the basilar membrane. Some of the considerations upon which this suggestion is based are the following :

(1) The prevailing acceptation is that the cochlea is the peripheral organ of the auditory apparatus, and that the construction of this organ is such as to be especially capable of serving, in conjunction with the central nervous system, in the analysis of sound.

(2) From their anatomical relationship, it is conceded that auditory impulses are aroused in the acoustic nerve through the mediation of the hair cells of the organ of Corti about which they terminate.

(3) It is usually conceded and here accepted that the hair cells are stimulated through the agitation of their hairs, but that the hairs are neither suitably constructed, long enough, nor vary sufficiently in length to be themselves acted upon selectively by the sound waves as transferred to the endolymph.

(4) All the recent conceptions of the process of hearing accept the idea that the hairs of the hair cells are agitated by contact with the under surface of the tectorial membrane, either by brushing against it or by perpendicular impingement, produced either by vibrations of the basilar membrane below or movements of the tectorial membrane above.

(5) The numerous physiological and anatomical objections to it are considered sufficient to render untenable the idea that the basilar membrane is the vibrating mechanism either of the kind demanded by the resonance theory or in accord with the telephone theory.

(6) While experimental investigation has not yet been able to ascertain the exact form of the wave motion in the endolymph of the cochlea, it has been determined by observation of the action of the tympanic membrane and auditory ossicles that the force of the motion produced in the tympanic membrane by the atmospheric sound waves may He increased thirty times in the transformation and transference of the motion to the basis of the stapes and the membrane of the fenestra vestibuli, and that the amplitude of the atmospheric waves may be reduced as much as seventy-six times (or more). In this transference


The Nature of the Tectorial Membrane 163

of the energy to the cndolymph, the reduction of amplitude, as well as the increase of the force, depends not only upon the lever arrangement of the ossicles, but upon the tensity of the tympanic muscles and the pressure of the air in the tympanic cavity. Therefore, the wave motions imparted to the endolymph by the basis of the stapes correspond to the atmospheric sound waves, of which they are transformations, but resemble them only in frequency of vibration. The quality of the motion imparted to the endolymph depends of course upon the quality or form of the atmospheric waves acting upon the tympanic membrane, but the quality of the two may not be identical in detail. While the canals of the cochlea are probably entirely filled with endolymph normally, and under a pressure which depends largely upon the blood pressure of the different conditions of the body, the membrane over the fenestra cochleie (rotunda) and the direct continuation of the endolymph of the cochlea with that of the vestibule are considered sufficient to allow compensation for the incompressibility of the fluid and to allow propagation of the motion in the form of true compression waves. Pressure applied by the stapes to the endolymph in the cochlea may not only be compensated or relieved by the membrane below and by the ductus reuniens connecting it with the sacculus, utriculus and semi-circular canals, but, by way of the sacculus, such periodic pressure may be compensated through the ductus endolymphaticus into the cranial cavity. Furthermore, the layer of softer tissue between the epithelium lining the scalge and the solid walls of the bony labyrinth may aid in maintaining the form of the wave motion. Therefore there is reason to conclude that the motions imparted by the stapes at the basal end of the cochlear coil are propagated in the endolymph toward the apex of the cochlea in the form of compression waves, however short, with longitudinal direction and transverse vibration. The waves are very probably similar in many respects to pulse waves.

(7) That the agitation of the hairs of the hair cells is brought about through the activity or actual movements of the tectorial membrane induced by the wave motion transferred to the endolymph is suggested by the peculiarities of that membrane.

. (a) It is in the logical position for such action. It lies above the hair cells adjacent to the portion of the endoh-mph to which the waves are first imparted by the basis of the stapes. The basilar membrane is situated below the tectorial and is both covered and obstructed.

(6) In the mammalia and in all animals possessing it, the tectorial


164 Irving Hardesty

membrane is entirely coextensive with the organ of Corti. The basilar membrane as such is not entirely coextensive.

(c) From its specific gravity, its evident lateral elasticity and from its most remarkable transverse flexibility and sensitiveness to agitations in a fluid surrounding it, the tectorial membrane appears to be far more admirably qualified to serve as the vibrating mechanism than does the basilar membrane.

(d) From the evidence that the tectorial membrane is attached along but one of its sides; from the fact that, even if it were attached along both sides, none of its fibers are coextensive with its width; from the fact that it is so constructed that relatively very few of the fibers, even if vibrating, can in any way come in contact with the hairs of the hair cells, and from the fact that the great majority of its fibers are evidently attached at neither of their ends, it is very improbable that the membrane is capable of acting as a sympathetic resonator,

(e) , In addition to its suggestive consistency and structure, the tectorial membrane is so shaped and proportioned as to suggest that it may act with reference to vibrations in the endolymph in a way by which the peripheral components of the phenomena of hearing may be explained; namely, the membrane may be acted upon as a whole, the extent, region, amplitude and quality of its vibrations depending upon the force, frequency, amplitude and quality of the vibrations acting upon it.

From the various pathological phenomena, observation of which have been followed by post-mortem examinations of the cochlea, it is quite generally accepted that sensations of the higher pitches are mediated by the structures of the basal coil of the cochlea, and from this it has been assumed, naturally, that the sensations of lower pitch are mediated by the apical coil.

The tectorial membrane is narrowest and thinnest in the basal coil and gradually increases in both width and thickness to acquire its greatest proportions in its apical end. Being attached along its inner edge throughout, it must be affected by sound waves in the endolymph very much as a very flexible ribbon so attached would be by agitations running parallel with its length in a fluid in which it floats.

The extent to which the tectorial membrane can be agitated must depend (1) upon the energy of the wave motion, or upon the extent and point at which the wave motion in the endolymph is neutralized by overcoming the inertia of the fluid and by the resistance offered by the


The Nature of the Tectorial Membrane 165

walls of the cochlea, and (2) upon the breadth and thickness of the membrane and thus upon the extent to which the wave energy is exhausted in overcoming the inertia of the membrane itself.

That notes of higher pitch or greater vibration frequency are more apt to produce vibrations in the thin, narrow, basal coil of sufficient amplitude to stimulate the auditory hairs than they are to produce such vibrations in the apical coil is considered probable for the following reasons :

(1) The thin, basal end of the membrane lies nearer the fenestra vestibuli (ovalis) or the point at which the waves are imparted to the endolymph. Of sounds of equal intensity, or amplitude of vibration, but of different pitch, those of higher pitch or greater vibration frequency are sooner overcome by the resistance of the medium (do not travel so far) as those of lower vibration frequency. In the atmosphere, sound waves of high frequency are damped out before those of low frequency and their speed of transmission continuously decreases as they become fainter. This damping out must occur much more quickly in a medium like the much more viscous endolymph. Therefore it is possible that sound waves of the highest perceivable pitch may affect only the end of the basal coil of the tectorial membrane and be damped out wholly before reaching the upper coils, or at least to such an extent as not to agitate the upper portions of the membrane sufficiently for stimulation of the hair cells.

(2) The natural vibration period of the thin, narrow strip is of greater frequency than that of the thicker wider strip. It is possible to subject a strip of material to vibrations of such frequency, either above or below its natural vibration period, that it will not vibrate at all or vibrate weakly or irregularly. It is probable that no portions of the tectorial membrane, when subjected to sound waves transferred to the endolymph, will undergo vibrations of sufficient excursion to impinge upon the auditory hairs except those portions whose natural periods correspond to the vibration frequencies of the waves affecting the endolymph. Portions adjacent to these, having approximately the same natural periods (thickness and width), would of course be also affected, but to a degree decreasing as the distance from the portion most affected increases. Or, again, the effect of loading a vibrating body is to lower its vibration frequency or pitch. If the body be of uniform proportions and the load be distributed uniformly, the vibration frequency of all its components will be lowered; if the load be placed at one part of the


166 Irving Hardest}^

body, one of the resulting complications will be a lowering of the vibration frequency of that part. Should the tectorial membrane be considered a vibrating body carrying a load so distributed as to gradually increase toward its apical end, then its vibration period or frequency must decrease as this end is approached. Therefore, it is possible that the thin, narrow basal coil of the tectorial membrane may be sufficiently affected by waves of high frequency to stimulate the auditory hairs when other portions of the membrane are not. Several waves, of course, may pass in a medium simultaneously in the same direction.

This idea suggests a sort of resonance quality in the tectorial membrane when the latter is considered as a whole, and thus it is not fully in accord with the idea originally suggested in applying the telephone theory to the basilar membrane. It is slightly analogous to some of the features of Waller's and Meyer's modification of the telephone theory.

(3) The natural consistency of a strip attached along one edge determines in considerable measure the extent and form of movements induced in it by wave motion in the medium surrounding it. The basal end of the tectorial membrane, being narrower and thinner, though of the same material, is more flexible than the end at the apex. Therefore, it must offer less resistance to waves of high frequency (is more easily crumpled into abrupt and frequent folds) than the thick apical end. The viscosity of a vibrating body, while it may affect the vibration frequency (pitch) but little, aids materially in causing the amplitude of vibration, the excursion or intensity of movement of the body, to gradually decrease and dwindle away as the waves pass along the body. It is therefore possible that waves of high frequency are capable of throwing the thin basal coil of the tectorial membrane into waves of considerable excursion or amplitude, while, as they pass along toward its apical end, they may be gradually absorbed in overcoming the inertia of the membrane ; first, becoming too faint to throw it into vibrations, or folds, of their frequency and of sufficient amplitude to agitate the auditory hairs below, and finally, as the broad, thick end is approached, a region is reached in which the waves of high frequency wholly dwindle away.

In the same way, for the above reasons, it may be suggested that waves of lower frequency than those which sufficiently agitate the basal end of the membrane only, can, according to their frequency or length, affect respectively the remaining regions sufficiently to stimulate the auditory hairs, because such longer waves travel farther in the given medium, are less rapidly overcome by friction of the walls of the scala


The Nature of the Tectorial Membrane 167

of the cochlea, and because the broader and thicker portions of the membrane probably can be thrown into vibrations by the waves of lower frequenc3\ In other words, the vibrations of higher frequency are more rapidly absorbed and dwindle away in passing along the membrane, while waves of greater length or lower frequency may throw the thicker portions of the membrane into undulations of sufficient amplitude for impingement upon the auditory hairs. Thus the first half-turn of the cochlea would be a portion of the peripheral organ of the auditory apparatus which is thrown into activity only by sound waves of the lowest frequency to which the mechanism is capable of reacting.

Aside from the possibility that the different widths of the tectorial ■ membrane possess different natural vibration periods and therefore may exercise a sort of synchronous selection or resonance with reference to the waves passing in the endolymph, there is nothing in the above suggestions to claim it improbable that the thin, basal coil is not made to undulate by waves of low frequency also. It very probably is, and under the following conditions :

There is little information available as to the properties of sound waves traveling in a substance whose elasticity is not the same in all directions. It is known that in strips of wood the waves travel more rapidly, are less impeded, in the direction of the fibers than across them. The tectorial membrane is most elastic in the direction of its fibers, that is, it is more resistant to stress applied transversely to them. It is exceedingly flexible and sensitive to motion applied transversely to its long axis. Such motions are, in the main, parallel to the direction of its fibers and thus can meet less resistance from them. Since, as shown- above, the fibers are not so arranged nor so attached as to act as a system of resonators, and since the membrane is attached only along one side instead of at each end or along both sides, it is very probable that it is largely a passive structure. Attached along its inner edge and held in its position over the organ of Corti by its elasticity, it extends, suspended in the endolymph and subject to be acted upon by whatever motion may be imparted to the latter. Sound waves transferred to the endolymph by the stapes at the fenestra vestibuli must travel along the membrane from basal end to apex with excursions of amplitiide transverse to its long axis and thus in the direction which may bring into play its remarkable flexibility. Waves of a given frequency of vibration will affect corresponding vibrations in the membrane to just that extent and amplitude to which they are not damped


168 Irving Hardesty

out by the endolymph and resistance of the walls, and not absorbed in overcoming the viscosity and bulk of the membrane itself. The waves of lowest frequency, or greatest length, may produce corresponding undulations in the entire membrane; waves of the highest frequency may produce in the thinner, and therefore most flexible, end of the membrane alone corresponding undulations of sufficient excursions for the necessary impingement upon the auditory hairs. Undulations of high frequency would result in a greater number of impingements per unit of length of the organ of Corti than would waves of low frequency, and thus both the nature and the locality of such stimulation would be different. Waves of higher amplitude, or intensity, would produce impingements of greater intensity and thus give sensations interpreted in degrees of intensity.

Functioning in this way, the peripheral organ of the apparatus is functionally as well as morphologically comparable to the so-called otolithic organs, the eristse and maculae acusticge. And, further, it is comparable with the other organs of special sense. The varieties of optic sensations, for example, depend upon the intensity and quality of the stimuli, that is, upon the number and variety of waves of light impinging upon a unit area of the retina. The various sensations obtained by touch depend upon the number, intensity and quality of the stimuli applied to the skin. Differences in quality of stimuli at the periphery are perceived and interpreted as differences by the central nervous organ, A number of stimuli applied to a unit area of the skin gives rise, within the possibilities of skin-innervation, to a different sensation or interpretation than does a single stimulus applied to that area. The same stimuli applied with different intensity give rise to interpretations of different intensity. So, in general, for taste and smell, all being dependent upon contact, whether the energy applied be simply mechanical or more strictly chemical.

In the auditory organ the stimuli arousing the sensations are thought to be mechanical. The greater numbers of stimuli applied to unit areas of the organ of Corti result from waves of high frequency and give rise to a sensation interpreted and named by the central nervous organ as high pitch. It happens that the region of the cochlea anatomically adapted for mediating sensations of highest pitch is its basal coil. Conversely, smaller numbers of stimuli applied to the unit area are interpreted as low pitch, the stimuli from waves of the lowest frequency being distributed most sparingly, but throughout the extent of the cochlea,


The Nature of the Tectorial Membrane 169

including the basal coil. Greater amplitude would mean intensity of impingement and interpretations of intensity of sound. Physical reinforcement of wave motions would mean reinforcement of tones; quality of impingement upon the auditory hairs would mean quality in the sensations as interpreted; the number of a given variety of stimuli applied to the unit area of hair cells would determine the predominant variety of the sensations resulting.

If the accessory tectorial membrane described in previous pages of this paper should prove to be a true and constant structure, it would be considered capable of separate vibration. Varying slightly in width, being narrowest at the basal end, this membrane suggests a considerable increase in the possibilities of function of the cochlea. It would undulate in accord with waves in the endolymph too faint to agitate the main body of the membrane at all, while its vibrations would be subject to the same conditions and would vary much as those of the main body. With waves of greatest amplitude it would act in conjunction with the main body. It is interesting to note that from its position it would act alone, only upon the outer and wider series of the auditory hairs.

In all the preparations showing the coherent series of the outer rods of Corti's organ, an interesting interrelationship of the rods was suggested. As seen on the flat, the shafts or slender mid-portions of the rods invariably showed a tendency to be grouped (Fig. 13). From those preparations in which the locality of the coil could be determined, this grouping seemed to be different in different regions. In the first turn of the organ of Corti, the shafts were grouped in twos and threes; in the second turn (Fig. 12) they were grouped in threes and fours; in a longer strip of them, fully five millimeters, which, during the teasing process, was seen to float out from the region of the fifth turn, the shafts were arranged in groups of from five to eight. None were obtained as coming from the basal coil.

We have no information as to whether or not the rods stand in this relation during life. They possibly stand straight and equidistant and cohere into groups during the manipulation before and during the mounting on the slide. In forming the arch bounding the tunnel of Corti, the shafts of the rods stand approximately straight in profile view, while the portions braced by the foot-cells always curve as they approach their rest upon the basiliar membrane (Fig. 12). It was thought possible, then, that the grouped appearance of the shafts on the slide might result from the straightening of the curved foot-portion


170 Irving Hardesty

when the series comes to lie flat on the slide. However, if this were true, one would expect all of the shafts to be curved, whereas those in the middle of each group are usually straight. Again, if the shafts stand straight and equidistant during life, noticeable depressions in the surface of the phalanges opposite the spaces between the groups would probably result from the bending of the shafts in forming the grouping. In all the preparations, the outline formed by the cohering phalanges appeared straight. Further, in the grouped conditions shown in the preparations, the granular protoplasm dispersed between the rods appeared continuous throughout the smaller spaces between the rods of each group, but in the spaces between groups it appeared above and below with a thin film along the sides of the bounding rods, leaving a definitely bounded vacant space in the center (p.. Fig. 12). In the split shown in Fig. 13, the upper area of this protoplasm appeared split also-. Since this preparation came from a cochlea fixed in Zenker's fluid, the protoplasm must have been hardened while in position in the cochlea and therefore was firm enough to be broken, as shown when the rod-series was split by the teasing process.

Whether equidistant or grouped in life, the preparations suggest an anatomical arrangement which may increase the functional possibilities of the organ. As shown by Joseph, '00, and by Held, '02, and as shown in Fig. 12, the rods proper consist of bundles of longitudinally placed fibers held together by a seemingly hyaline matrix, and therefore may be distinguished from the granular protoplasm distributed between the rods and forming the foot cells. From all appearances, the rods are relatively rigid, the outer more so than the inner series, and their presence is generally conceded as serving for the support of the neuroepithelium at each side. If they are grouped during life, then the two phalanges which come together opposite the larger spaces between the groups (a., Fig. 12), might be less resistant to vertically applied force than would the phalanges of the rods of the middle of the groups, b, and thus, also, would the hair cells immediately supported by these phalanges be held in varying degrees of rigidity against the impact of the tectorial membrane. On the other hand, a separation of the rods, as shown in the groups, into straight and equidistant relations would produce slight bulges in the outline of the coherent phalanges, and thus, if the rods are straight and equidistant during life, this condition probably results in hair cells at intervals standing higher than others and more apt to be stimulated by slight undulations of the tec


The Nature of the Tectorial Membrane 171

torial membrane, and to be stimulated more strongly by undulations of greater amplitude. Either relation (that is, if the hair cells do not lie in the same plane), would be a factor in determining the number (and also the intensity) of the stimuli applied to a unit area of the organ of Corti. Such conditions are merely suggested by the slides. Though constant in my preparations, the grouping may be artifact and may mean nothing at all.

As evident in the above pages, I have entered very lightly into an application of the laws of physics to the anatomical role suggested for the tectorial membrane. Many features must be subjected to further anatomical investigation and must receive further substantiation before a wise application of the known can be made. For the same reason, the physiological phase of the question, which, based upon the many auditory phenomena, could be prolonged indefinitely, is touched upon but little. Just how far the cochlea, in itself, takes part in the analysis of sound is the difficult question: In the physiological literature the cochlea is spoken of as analyzing^ noticing, appreciating phase, distinguishing and perceiving tones, etc. It is well known that several waves may pass, in the same direction, simultaneously through a medium. In the cochlea, acting as above suggested, their resulting stimulation of the auditory hairs would be somewhat segregated as to pitch. A mixture of sounds of different qualities but of approximate pitch would act upon approximately the same area of the organ of Corti, complicate the process and probably be interpreted as "noise/' The matter of the finer analysis of sound is no doubt almost wholly cerebral. It certainly is largely a matter of education. It is of course probable that cochlese, as other organs of the body, may differ as to their anatomical excellence in different individuals. I wish to suggest that most of the phenomena of hearing which are actually peripheral may be explained on the basis of the relationships suggested.

Impulses borne by the cochlear division of the acoustic nerves are conveyed to given localities of the cerebrum. Probably any stimulus aroused in these fibers gives rise to sensations of sound. Gray, '00, reports a case of deafness and "singing noise in the ear in which the post-morten examination showed all the divisions of the ear to be in perfect condition, but, on further search, a tumor was found in the medulla oblongata encroaching upon the entering trunk of the acoustic nerve. Tinnitus aurium could be produced not only in this way, but, if produced temporarily, or permanently as sometimes happens, by a


172 Irving Hardesty

violent auditory stimulation (explosions, etc.), it might result from the tectorial membrane being thrown by the excessive amplitude of the vibrations so violently upon the auditory hairs that it becomes stuck to them for a while by means of its glutinous matrix and thus would give rise to continuous stimulation. Allien teased out in the fresh condition, the membrane adheres to the needle point with a readiness very embarrassing to the operator, and sections often show it adhering to the hair cells upon which it has evidently been crumpled by the manipulation.

Calcareous deposits in the matrix of the membrane would not be surprising, considering the developmental analogy it bears to the otolithic membranes. If uniformly distributed, these would produce a stiffening of the membrane and a general impairment of the hearing; if localized they would result in "tone islands." Calcification may be the cause of some of the peculiarities of the "failing ears" of old age.

■Hensen, '07, in his recent paper upholding the resonance theory as applied to the basilar membrane, claims that the tectorial membrane, considered free by him, cannot be made to impinge upon the auditory hairs by wave motion in the scala vestibuli, assuming that waves in this canal would produce simultaneous depressions of the whole spiral lamina below and thus of the hair cells, so that the usual space separating them from the tectorial membrane when the cochlea is at rest would be maintained. And he thinks, therefore, that sensations of sound do not arise till the wave motions have passed through the scala vestibuli and, therefore, only on their return in the scala tympani below does the resonance action of the fibers of the basilar membrane force the hair cells upward to impinge upon the under surface of the membrane. This |dea seems to me subject to the following suggestions :

(1) It seems probable that confusion would result from the anatomical necessity, involved in his idea, that the two sides of his resonant membrane would be subjected simultaneously to different sets of waves, those passing above and those below a given point.

(2) Or, if the basilar membrane is capable of undulating so freely to the waves in the scala vestibuli, these waves would be imparted through it to the endolymph in the scala tympani below, and the same waves passing in the same direction in both canals would meet at the helicotrema and would either be reflected there by the wall of the labyrinth and return towards the middle ear, meeting other waves traveling in the opposite direction, or they would be damped out in the confusion at the helicotrema.


The Nature of the Tectorial Membrane 1T3

(3) The scala tympani increases in diameter in passing from the helicotrema to the fenestra cochlese (rotunda) . Therefore, supposing that the sound waves passing in the scala vestibuli return without interference in the scala tympani, they would not only be fainter because of the resistance of the endolymph and walls of canal through which they have passed, but their amplitude would be further decreased because of their being dispersed over the increasing space of the scali tympani. So, Hensen's idea must assume that the stimulating resonant action of the basilar membrane is accomplished with waves of less amplitude than when entering the cochlea.

(4) Anatomical studies of the basilar membrane suggest that it is wholly incapable of moving as sensitively as the tectorial membrane to waves in the scala vestibuli. Not only is the basilar membrane firmly continuous along its either edge with the walls of the labyrinth and invested above and below by layers of other tissue, but even when torn away as a free strip, it is far more stiff than the tectorial membrane. When teased free in fresh preparations it is rigid enough to more than retain its shape, and its flexibility, compared with that of the tectorial membrane, is as a wooden board compared with a strip of tissue paper. It must at least resist wave motion in the endolymph far more than the tectorial membrane Avhich lies suspended in the endolymph, one edge free, and by position is the first to receive the impact of the waves.

To these objections to Hensen's idea may be added the ol3Jections enumerated on another page against the idea that the basilar membrane is a resonance mechanism at all.

From the anatomical standpoint, the basilar membrane may be considered as nothing more than a thin, fiat tendon, thicker along its edges, whose purpose is to so strengthen the fioor of the cochlear duct that it may firmly support the organ of Corti and at the same time be thin enough to allow the presence and necessary caliber of the scala tympani below it.

The vestibular (Eeissner's) membrane is considered as having little to do with the function of the organ of Corti. By development, it is the remains of the outer wall of the embryonic cochlear canal, after the liquefactions of the mesenchymal tissue resulting in the scala vestibuli. It may serve as a protection to the organ below, damping violent agitations in the scala vestibuli and preventing upward displacement of the tectorial membrane. It is weakest along its edges, which condition gives, in teased cochlea, somewhat the impression of its being hinged to


174 Irving Hardesty

the walls of the lab3a'inth. Grasped in the forceps, long strips of it can be removed intact, so easily does it come aAvay. Therefore, sound waves passing in the scala vestibuli may be readily imparted to the endolymph in the cochlear dnct.

Summary.

(1) The tectorial membrane of the pig, occupying a cochlea of about four turns, has an average length of 25.5 millimeters. It is about three times as broad and five times as thick in the apical half-turn as it is in the last half-turn, its dimensions decrease gradually from its apical toward its basal end, and its ends terminate bluntly.

(2) Its specific gravity is but little greater than that of the fluid in .Avhich it lies.

(3) It possesses a small amount of elasticity, barely sufficient to cause the thicker, apical region to resume its normal coils while the membrane is suspended in fluid after being freed from its attachment. Its greater elasticity is in the direction opposed to stress applied longitudinally. It is remarkably flexible to stress applied transverse to its long axis.

(4) Its structure consists of multitudes of delicate fibers of unequal length embedded in a transparent matrix of a soft, collagenous, semisolid character with marked adhesiveness. The general transverse direction of the fibers inclines from the radius of the cochlea toward the apex, which inclination is greater in the upper than in the under surface. Most of the fibers, by coursing and curving through the membrane, take part in producing the fibrous appearance of both the upper and under surfaces.

(5) From the study of both its adult structure and the processes by which it is developed, it is concluded that none of its fibers extend the entire width of the membrane, none are attached at both ends, and the greater number of them are attached at neither of their ends.

(6) Hensen's stripe is explained as a line of the intercrossing ends of the fibers of the under surface resulting from the processes by which the growth of the membrane terminates.

(7) From its study in both the fresh condition and in preparations considered least shrunken and distorted, it is concluded that the tectorial membrane projects free over the organ of Corti and is attached only along its inner zone upon the labium vestibulare of the spiral limbus.


The iSTature of the Tectorial Membrane 175

(8) A thin, exceedingly delicate, accessory tectorial membrane is described lying along the under surface of the onter zone of the main body to which its outer edge is lightly attached. It varies in width somewhat as does the main body and its fibers extend inward toward Hensen's stripe, but only extend over the outer series of hair cells.

(9) To the several objections advanced by others to the assumption that the basilar membrane performs resonant vibration, there is added evidence that the basilar membrane is nothing more than a flat tendon composed of a lamina of interconnected bundles of white fibrous connective tissue whose purpose is merely to strengthen the floor of the ductus cochlearis and the position of the organ of Corti, and which are too rigid and firmly associated to allow of resonant vibration. And, further, even if resonance were anatomically possible in the membrane, the two layers of tissue on each of its sides would be sufficient to damp such action.

(10) To the evidence adduced that the different elements of the organ of Corti are incapable of being moved separately by vibrations in the basilar membrane, there is added evidence that the outer series of the rods of Corti are especially so associated as to be incapable of separate motion. The outer rods are more resistant to maceration than the inner rods.

(11) The theories in which the basilar membrane is considered the vibrating mechanism in the cochlea are considered untenable, and an application of the telephone theory to the tectorial membrane as the vibrating mechanism is suggested on the basis of its logical position, its extent, shape, proportions, consistency and structure, and the probable character of the transformed and transferred sound waves in the endolymph of the cochlea.

BIBLIOGRAPHY.

A VERS, Howard. Die Membrana tectoria — was sie ist, unci die Membrana basilaris^ — was sie verrichtet. Anat. Anz., Bd. 6, 1891.

Avers, Howard. On tTie Membrana basilaris, the Membrana tectoria, and the Nerve Endings in the Human Ear. Zool. Bulletin, Vol. I, No. 6, 1898.

Barth. Beitrag zur Anatomie der Schnecke. Anat. Anz., Bd. 4, No. 20, 1889.

BoTTCHER, A. Ueber Entwicklungsgeschiehte und Bau des Gehorlabyrinthes. Yerhandlungen der Kaiserl. Leop. Karol-deutschen Akad. der Xaturforscher., Bd. 35, 1869.


176 Irving Hardesty

CoRTi. Reclievches sur Torgane de rouie des Mauimiferes. Zeitsclirift fiir wiss. Zoologie, Bd. 3, 1851.

Coyne, P., et Cannieu, A. Contribution a I'etude de la Membran de Corti. Jour, de I'Anat. et de la Physiol., Ann6e 31, 1895.

CziNNER und Hammerschlag. Beitrag zur Entwicklungsgescbichte der Corti'schen Membran. Arcbiv fiir Obrenbeilk., Bd. 44, 1807.

Dupuis, A. Die Corti'scbe Membran. Anat. Hefte, Bd. 3, H. 10. 1804.

YoN Ebner, V. In Kolliker's Handbucb der Gewebelebre des Menscben. Bd. 3, 2d Halfte, 1902.

EwALD. J. R. Zur Pbysiologie des Labyrintbes. — VI. Mitteilung eine ueue

Hortbeorie. Arcbiv fiir die gesammte Pbysiologie (Pfliiger's Archiv), Bd. 76, 1899.

EwALD, J. R., und Jaderholm, G. A. Auch alle Gerauscbe geben, wenn sie intermittirt werden. Intermittenztone. Pfliiger's Arcbiv, Bd. 115, 10<>»;.

ExNEB, S. Mittheilung iiber Beitrag zur Entwiclvlungsgescbicbte der Corti'scben JMembran, von Ignaz Czinner und Viktor Hammerscblag. Sitz. Bericbt. der Akad., Wien., Abtbl. 2a, Heft 3 and 4. 1897.

Ferr^. Contribution j\ I'etude du Nerf Auditif. Soc. Zool. de France, T. 10, 1885.

GoTTSTEiN, J. Ueber den feineren Bau und die Entwicklung der Gehorscbnecke der Menscben und der Sauger. Arcbiv fiir mik. Anat, Bd. 8, 1872.

Gray, A. A. On a Modification of tbe Ilelmboltz Tbeory of Hearing Jour, of Anat. and Physiol., Vol. 34, 1900.

Gray, A. A. The Labyrinth of Mammals. Vol. 1, J. and A. Churchill, London, 1907.

Held, H. Zur Kenntniss des Corti'schen Organs und der iibrigen Sinnesapparate des Labyrintbes der Sauger. Abhandl. der Math.-pbys.-Klasse der stichs. Gesell. der Wiss. Leipzig, Bd. 28, 1902.

Vox Helmholtz, H. L. T. Die Lehre von den Tonempfindungen als physiologische Grundlage fiir die Tbeorie der Music. Ausgabe 5, Braunschweig, 1896.

Henle, J. Handbucb der Eingeweidelebre des Menscben. Handbucb der systematischen Anatomie des Menscben, Bd. 1, Braunschweig, 1866.

Hensen, V. Zur Morpbologie der Schnecke des Menscben und der Saugetiere. Zeitschr. fiir wiss. Zool., Bd. 12, 1863.

Hensen, V. Die Empfindungsarten des Scballs. Pfliiger's Arcbiv. Bd. 119,

1907. Joseph, H. Zur Kenntniss vom feineren Bau der Gehorscbnecke. Anat.

Hefte, Bd. 14, H. 46, 1900.


The Xatiire of the Tectorial ]Meinhrane 177

KiSHi, K. Covti'scbe Membraii unci ToDempflndungstheorie. Pfliiger's Arebiv, Bd. 116, 19C»T.

Tek Kuile, E. Die ricbtige Bewegnngsform der Membrana basilaris. Pfluger's Avcbiv, Bd. 79, 1900.

Ter Kuile, E. Die Uebertragimg der Energie vou der Grnndmeuibran anf die Ilaarzellen. Pfliiger's Aroliiv, Bd. G9, 1900.

KoLLiKEK, A. Entwiclvlungsgesebicbte des Menscben und der bobeven Tiere, Leipzig, ISGl.

KoLJiER, W. Beitriige znr Keuntuiss des feinereu Baues des Gebororgaiis mit besonderer Beriicksicbtigung der Haussaiigetiere. Archiv fiir mil^. Anat., Bd. 70, H. 4, 1907.

Lavdowsky, B. Uutersucbnugen iiber den alcnstisc-ben Eudapparat der Siiugetiere. Arebiv fiir mik. Anat., Bd. 13, 1877.

LowENBERG. BeitrLige znr Anatoniie der Sclnieoke. Arebiv fiir Obrenbeilk., Bd. 1, 1864.

Meyer, M. Several papers in Zeitscbr. fiir Psycliol. und Pbysiol. der Sinnesorgane. Bd. 16 and 17, 1898.

NuEL. Beitrag znr Keuntuiss der Silngetbierscbnecke. Arebiv fiir mik. Anat., Bd. 8, 1872.

NuEL. Recbercbes mieroseopiqiie sur I'anatomie du lima(;ou des Mammiferes, 1875. Cited by Retzius, 1884.

Retzius, G. Das Geburorgau der Wirbeltiere. Biologiscbe Uutersucbungeu, Bd. 1, Stockholm, 1884.

RicKENBACHER, O. Untersucbuugen iiber die embryonale INlenibraua tectoria der Meerscbweincben. Anat. Hefte, II, 16, 1901.

Shambaugh, G. E. a Restudy of tbe Minute Anatomy of the Structures in tbe Cocblea, with Conclusions bearing on tbe Solution of tbe Problem of Tone Perception. Amer. Jour, of Anat., Vol. 7, No. 2, 1907.

Siebenmann, F. Mittelobr und Labyrintb. Bardelebeu's Handbucb der Anat. des Menscben, Bd. 5, 1900.

Spee. Mitteilungen zur Histologic des Corti'scben Organs und der Geborscbnecke des erwacbsenen Menscben. Verbandl. der Anat. Gesell., Bonn. 1901.

Steinbrugge. H. Ueber das Verbalteu des menscblicben Ductus cocblearis im Yorbofblindsack. Anat. Hefte, Bd. 3, 1894.

WiEDERSHEiM, R. Gruudriss der vergleicbendeu Anatomie der Wirbeltbiere, Jena, 1893.


178 Irving Hardesty

EXPLANATION OF FIGURES.

Abbreviations in common.

Ac, accessory tectorial membrane.

Hs., Hensen's stripe.

Iz., inner zone of tectorial membrane.

Lv., Labium vestibulare.

Mb., Basilar membrane.

Mv., Vestibular (Reissner's) membrane.

St., Scala tympani.

Fig. 1. — Representing unsbruuken appearance of an entire tectorial membrane teased from the cochlea of a pig two weeks old. Under surface upward. Shape constructed from measurements from several specimens and from study of specimen represented. Drawn with coil more open than normally to avoid overlapping of edges in apical turns. Arrangement of fibers drawn in from specimen.

Fig. 2. — Block from first half-turn of whole specimen. Viewed from under surface. Cut end showing course of fibers is added from study of sections of other membranes. Li., line of impression of edge of labium vestibulare.

Fig. 3. — Block from third half-turn of whole specimen, imder surface upward, showing course of fibers and appearance of accessory membrane as torn loose from main body. Li., line of impression of edge of labium vestibulare ; Lr., portion of lamina reticularis removed from labium with inner zone of tectorial membrane.

Fig. 4. — Block from fifth ha If -turn of whole specimen, showing the general course of the fibers as viewed from the upper surface, and indicating the character of cleavage in a broken end of the membrane. S., slivers of fibers and matrix formed in the breaking.

Fig. 5. — The end of the basal coil of the tectorial membrane, showing the outer edge of the accessory membrane coinciding with the outer edge of the main body, and showing how the inner zone (Iz.) or the attached portion of the membrane curves so as to constitute the tip of the termination.

Fig. 6. — Vertical section of the tectorial membrane in position and througli the first half-turn of its coil, showing the extent to which its outer edge projects beyond the organ of Corti and the course and arrangement of its fibers. Pc, peripheral condensation interpreted as produced by the shrinkage action of the reagents ; Vs., vas spirale, here a f)lexus instead of a single vessel ; Ss., spiral sulcus.

Fig. 7. — Vertical section from third half-turn of same cochlea as Fig. 6. with the organ of Corti represented entire and showing the character of the epithelium of the spiral sulcus iSs.) and an appearance on the under


The Xature of the Tectorial Membrane 179

surface of the outer zone of the tectorial membrane, Ac, which is interpreted as representing the much shrunken accessory membrane in section. P^., below, shows what seems in the section to be a strip of the peripheral condensation torn off in the manipulation and adhering to the inner supporting cells of the organ of Corti. E., endothelium lining scala tympani, St.; Vs., vas (plexus) spirale. The curved shape of the section of the tectorial membrane is due to shrinkage.

Fig. 8. — Vertical section from seventh half-turn of same cochlea as Fig. 7. The curve in the upper surface of the outer zone of the tectorial membrane is explained as due to shrinkage produced by the techique as well as the fact that the membrane does not appear to extend entirely over the organ of Corti. From the size of the tectorial membrane, as compared with Figs. G and 7, it is assumed that the section passes some distance from the basal end. Lettering same as in Figs. 6 and 7.

Fig. 9. — Horizontal section of tectorial membrane at region of second turn of coil, showing appearance and arrangement of fibers as cut transversely, obliquely and longitudinally. Irregularities of course of fibers explained as due largely to collapsed condition produced by reagents, a., region of transversely cut fibers; &., longitudinally cut fibers; Ep., "Epithelial cells" inserted between the white fibrous tissue bundles; lit. (Huschke's teeth) of the labium vestibulare.

Fig. 10. — Vertical section from the third half-turn of a cochlea from a foetal pig of 14 centimeters, showing a stage in the development of the tectorial membrane just before the beginning, along the inner edge, of the retrogressive changes of the greater epithelial thickening, Get. Let., lesser epithelial thickening, enlarging to form organ of Corti ; Mt., tectorial membrane ; 8s., cells in region of what will become spiral sulcus ; other lettering same as in Figs. 6 and 7.

Fig. 11. — T'rom thin horizontal section of basilar membrane, stained by Mallory's method for white fibrous tissue, showing the structure to be composed of bundles of white fibers, FJ)., connected with each other by numerous fibers less compactly arranged. Ls., side toward spiral ligament ; Lt, side toward labium tympanicum of spiral limbtis : Fn., foramina nervosa in babenula perforata.

Fig. 12. — Portion of a line of the outer rods of Corti, adhering together and showing grouping as appeared in teased specimen from second turn of cochlea, a., phalanges assumed to be less resistant to downward pressure than those designated by 6.; f., fibrous portion of curved foot of rod which rests over basilar membrane ; Ir., fragments of inner rods left in their separation from the outer ; P., granular protoplasm between rods and probably continuous with that forming the foot cells.


THE NATURE OF THE TECTORIAL MEMBRANE

IRVING HARDESTY



AMERICAN JOURNAL OF ANATOMY--V0L. VIII, NO. 2


THE NATURE OF THE TECTORIAL MEMBRANE

IRVING HARDESTY



Fig. 2


AMERICAN JOURNAL OF ANATOMY— VOL. VIM, NO. 2


THE NATURE OF THE TECTORIAL MEMBRANE

IRVING HARDESTY



Fig. 3


AMERICAN JOURNAL OF ANATOMY~VOL. VIII, NO. 2


THE NATURE OF THE TECTORIAL MEMBRANE

IRVING HARDESTY



Fig. 4


AMERICAN JOURNAL OF ANATOMY— VOL. VIII, NO. 2


THE NATURE OF THE TECTORIAL MEMBRANE

IRVING HARDESTY



Fig. 5


AMERICAN JOURNAL OF ANATOMY--VOL. VIM, NO. 2


the:nature of the tectorial membrane

IRVING HARDESTY



Fig. 6


AMERICAN JOURNAL OF ANATOMY--VOL. VIII, NO. 2


THE NATURE OF THE TECTORIAL MEMBRANE

IRVING HARDESTY



Fig. 7


--Mv



Fig. 8


AMERICAN JOURNAL OF ANATOMY--VOL. VIII, NO. 2


THE NATURE OF THE TECTORIAL MEMBRANE

IRVING HARDESTY


Ep


Ht







Fig. 9



Fig. 10


Fb



Fig. 11


AMERICAN JOURNAL OF ANATOMY--VOL. VIII, NO. 2


THE NATURE OF THE TECTORIAL MEMBRANE

IRVING HARDESTY



Fig. 12


AMERICAN JOURNAL OF ANATOMY--VOL. VIII, NO. 2


EARLY DEVELOPMENT OF THE CERVICAL VERTEBRAE AND THE BASE OF THE OCCIPITAL BONE IN MAN.^

BY

ckARLES RUSSELL BARDEEN,

University of Wisconsin.

With 3 Figures.

During the earlier stages of deA^elopment the cervical vertebrae resemble those of the thoracic region. The two regions soon become differentiated from one another by the much greater development of the costal processes of the thoracic region. The seventh cervical vertebra alone, as a rule, has a large costal process, and this does not extend far beyond the transverse process of the neural arch (Fig. 1). In the cervical, as in the thoracic vertebrae, the development of a region of loose tissue in the base of the primitive ventral process serves to separate the costal element from the transverse process. In this loose tissue an anastomosing artery extends from the intersegmental artery on the posterior to that on the anterior side. The anastomosing artery between the costal element and transverse process of the seventh cervical vertebra remains small, but the more anterior anastomosing arteries give rise to a large continuous vessel, the vertebral artery, which extends anteriorly between the costal processes and transverse processes of the root of the vertebral artery.

In the costal processes of the seventh cervical vertebra centers of chondrification are found at the period when similar centers appear in the ribs. Centers of chondrification in the costal processes of the rest of the cervical vertebra appear much later, usually not until the embryo has reached the length of from 16-18 mm.

As in the thoracic vertebra?, there are two bilaterally placed centers

^The early development of the thoracic, sacral and coccygeal vertebrae has previously been described in The American .Tournal of Anatomy, Vol. IV, 1905. American Jouenal of Anatomy. — Vol. VIII. Ko. 2.


182 Charles Russell Bardeen

of chondrification for each of the vertebral bodies. These soon fuse with one another ventral and dorsal to the chorda dorsalis. In the first two vertebrae the ventral fusion takes place before the dorsal fusion.

There are separate centers of chondrification for the neural arches. In the more distal cervical vertebrge these centers are similar to those of the thoracic vertebra. In the more proximal cervical vertebrae the centers of chondrification appear as basal plates lateral to the anterior end of the bodies of the vertebrae. With these they soon fuse. From the plate-like base chondrification extends rapidly into the main part of the arch. From the neural arches, laminar, articular and transverse processes are developed. The costal elements have separate centers of chondrification which soon fuse proximally with the bodies of the vertebrae and distally with the tips of the transverse processes of the vertebrae. The dorsal growth of the laminar processes and the formation of the spinous processes of the cervical vertebrae take place in the main like that of the thoracic. WTien fully formed, however, the cartilaginous cervical vertebrae have essentially the shape of the adult osseous cervical vertebrae. Even before the end of the second month of development distinct cervical characters may be distinguished (Figs. 1 and 2).

Some investigators hold that the neural arches of the mammalian vertebrae contain elements of both the ventral and dorsal arches found in the lower vertebrates (see Schauinsland, Hertwig's Handbuch, 1902). There are theoretical grounds for believing that the ribs primitively belong to the ventral arches. In the higher mammals and man, however, the presence of the ventral arch elements is manifest merely in the caudal region, where temporary haemal processes are developed, and in the upper cervical region, where, in the membranous stage, there are differentiated from the ventral margins of the primitive discs bands of tissue which connect the bases of the neural processes. These bands of tissue have been called by Froriep the hypochordal braces (Spangen). In reptiles and birds the hypochordal brace becomes converted into cartilage and connects the cartilaginous arches of each side with one another. It finally becomes fused with "the ventral portion of the proximal end of the vertebral body. In the primitive type of development the chondrification takes place from two bilaterally placed centers, each of which, according to Schauinsland, represents a ventral hemiarch. In other instances chondrification takes place from a single center situated in the median line. According to Froriep, in the cow a


Early Development of the Cervical Vertebras


18^


median center of chondrification appears in all the hypochordal braces, but, except in the first two vertebrae, the brace disappears before cartilage is actually formed. In the hypochordal brace of the second



nalis hypoglossi 'Atlas arcus ventralis


Costa I


Fig. 1. — Lateral view of a model of the occipital cartilage, the cervical vertebras the first thoracic vertebra and the proximal end of the first rib in an embryo 20 mm. long. The costal elements of the cervical vertebrae are cartilaginous and are connected by ligamentous tissue with the transverse processes of the neural arches and with the vertebral bodies. The intervertebral discs are shown except between the fourth and fifth, and the fifth and sixth vertebrae. Dense interarticular tissue is shown between the articular processes from the second cervical to the first thoracic vertebra and, posterior to this, interlaminar connective tissue membranes and a ligamentous band running from the neural process of the atlas to the thoracic region.


18J:


Charles Eussell Barcleen


vertebra the cartilaginous anlage is very transitory, in the first vertebree it forms the anterior portion of the arch of the atlas. Weiss describes two bilaterally placed centers of chondrification in the hypochordal brace of the atlas in the white rat. No cartilage is fonnd in the more distal hvpochordal Ijraces. In man a hypochordal brace becomes well



Corpus sphenoidale (posterior portion)

Chorda dorsalis Os occipitale pars basilaris Atlas arcus ventralis



Costa I


Vertebra thoracalis

Pedicle' Processus articularis anterior/


Lamina


Fig. 2. — Postero-iuedial view of the model represented in Fig. 1. The anterior extremity of the base of the occipital and the posterior part of the body of the sphenoid are also shown. Between the posterior part of the base of the occipital and the epistropheus the intervening dense tissue has in part been removed so as to reveal the ventral arch of the atlas. The intervertebral discs are omitted between the fourth and fifth, fifth and sixth cervical, and between the seventh cervical and first thoracic vertebra?.


Earh' Development of the Cervical ^'erteljrte


185


developed merely in connection with the atlas. It Ijecomes cartilaginous later than the neural arch. There are indications of two hilaterally placed centers of chondrification, but fusion with one another and with the neural arches takes place as soon as chondrification is well under way.

Specific mention must be made of the mode of development of the epistropheus, of the atlas, and, in connection with the latter, of the base of the occipital.

Epistropheus. — The general mode of development of the epistropheus is like that of the other cervical vertebrae. Its marked distinction comes from its union with the body of the first cervical vertebra. This union


Basioccipitai



Hypochordal brace


Fig. 3. — Sagittal section tlu'oiigb the lateral part of the cervical region of the spinal column of an embryo 14 mm. long.


takes place through the transformation of the intervertebral disc into cartilage, first lateral to the mid-sagittal plane and later in this plane.

Atlas. — The base of each neural hemi-arch of the atlas becomes temporarily fused with the body (14 mm. embryo), but this fusion is incomplete and soon is followed by the development of dense fibrous tissue between the arch and the body (Fig. 3). At the same time the hypochordal brace becomes cartilaginous and unites the arches of the atlas in front of the Body. Each costal process becomes fused medially to the basal process, laterally to the transverse process of the corresponding hemi-arch.


186 Charles Eiissell Bardeen

The articulation between the lateral mass of the atlas and the superior articular surface of the epistropheus seems to be formed rather in the interventral than, as in the other intervertebral diathroses, in the interdorsal membranes. This is also true of the atlanto-occipital diathrosis. For a brief period (14 mm. embr3'o) the bases of the neural arches of the atlas and epistropheus, together with the tissue intervening between the bases of arches of the atlas and the occipital, become fused into a nearly continuous mass of precartilage (Fig. 3).

Hagen, His' Archiv, 1900, gives a somewhat different account of the development of the atlas and epistropheus in man. He concludes (1) that the dens epistrophei arises from the region of the body of the epistropheus and a portion of the body of the atlas; (2) that the massse lat. of the definite atlas arise from the rest of the primary anlage of the body of the atlas, and (3) that the short piece which unites them in front arises from the fusion of both neighboring septa.

Basi-occipital. — Opposite the last occipital myotome the axial mesenchyme is differentiated, like that of the spinal sclerotomes, into a light anterior half and a dense posterior half or scleromere. In the spinal region each scleromere joins with the light half of the sclerotome next posterior in giving rise to the body and arch processes of a spinal vertebra. In man the occipital scleromere is not thus associated with the light half of the first spinal sclerotome. On the contrary, it becomes associated with the lighter tissue of its own segment and with the tissue into which this is continued anteriorly. Laterally the tissue differentiated at the side of the anterior half of the first spinal sclerotome, the interventral membrane, becomes temporarily converted into a precartilaginous membrane (Fig. 3).

Chondrification of the base of the occipital begins in two bilaterally situated centers in the posterior portion of the occipital anlage. The union of these centers takes place posteriorly ventral to the notochord and anteriorly dorsal to the notochord (Fig. 2). The neural processes of the posterior part of the occipital anlage seems to have separate centers of chondrification, but these centers fuse almost immediately with the centers of chondrification of the body.

From the tissue derived from the first and second sclerotomes and not utilized in the formation of the atlas and the epistropheus are derived the various ligaments which unite these bones. The details of the formation of these ligaments are too complex for description here.


A .


THE PHARYNGEAL POUCHES AXD THEIE DERIVATIVES IN THE MAMMALIA.

BY

HENRY FOX, Ph.D.

With 73 Figures.

The present paper is an outgrowth of an earlier unpublished article submitted to the Faculty of the University of Pennsylvania in partial fulfilment of the requirements for the degree of Ph.D. The original article gave the results of a study of six or seven different stages in the development of the pig. Subsequently, through the kindness of Dr. C. S. Minot, I was enabled to study the extensive series of mammalian embryos in the collection of the Harvard Medical School. Of these I studied most thoroughly the series of pigs and cats, but also gave some attention to the later stages in the rabbit. The results of this additional study, along with those included in my former article, I noAV offer in the present paper.

My aim is to give a complete history of the pharnygeal pouches and their derivatives as typically exemplified in the mammalia. The main facts of this history had been largely determined previous to my starting the investigation, but the interpretations attached to these facts by various authors differed considerably, and, moreover, there remained a number of details about which there was much confusion. These unsettled matters seemed to me to warrant a full investigation of the subject.

Soon after I had begun my observations an important article by Hammar appeared treating of the development of the fore-gut in man. Hammar had in his possession a large series of embryos, and from these he made out a full and consistent history of the middle ear and Eustachian tube. He also compared with his own results the statements made by earlier authors, and, through the more abundant material at his command, was enabled to show how their conclusions were in most cases the result of mistaken interpretation based on an insufficient body of facts.

The American- .Journal of Anatomy.— Vol. YIII, No. 3.


188 Henry Fox

While jDrimaril}' concerned with the development of the pharynx in man, Hammar also examined a series of rabbit embryos, and while he does not in his article treat of them particularly, he yet mentions that he finds an essential agreement in the formation of homologous parts in both forms. Accordingly, he is inclined to assume that the essential features of the development in man will hold good in the case of other mammals.

Hammar's first article was followed by a second on the fate of the second pharyngeal pouch. This is the last article of his I have seen, and, so far as I know, he has not published any articles on the fate of the last two pouches.

The appearance of Hammar's paper seemed to me at first to do away with the necessity of further study of the first two pairs of pouches, but as Gaupp had already expressed the idea — ^based upon the conflicting statements of earlier investigators — that the formation of these parts probably differed considerably in different species of mammals, I concluded that a further contribution on the subject in the three species examined by me would not be without value. Moreover, as Kastschenkd, the chief authority on the process in the pig, had declared that the middle ear tube did not arise in any way from the first pharyngeal pouch, I considered this an additional reason for continuing my investigation.

The results of this investigation, so far as the first and second pharyngeal pouches are concerned, are largely confirmatory of the conclusions reached by Hammar in man and the rabbit. The probability therefore is that the development of these parts is essentially similar in the majority of placental mammalia.

In the case of the third and fourth pharyngeal pouches I have obtained results which clear up certain details about which there has been much conflict of opinion. Among these may be mentioned the determination of the origin and structure of the carotid gland, a reconciliation of the conflicting statements regarding this structure made by Kastschenko and Prenant, a confirmation of the ectodermal origin of the so-called thymus superficialis of Kastschenko, and finally the origin of a second structure — beyond doubt the glandule thyroidienne of Prenant — from the fourth pouch.

My earlier studies were made by the aid of the wax reconstruction method. Later, owing to the lack of facilities for continuing the use of this method, I adopted the method of graphic projection, making dorsal, ventral and lateral views of each stage studied.


The Pharyngeal Pouches in the Mammalia


189


During the progress of this investigation I received much assistance from a number of investigators. To Dr, E. G. Conklin I am greatly indebted for his kind encouragement and helpful suggestions, and for these I desire to express my hearty thanks. To Dr. C. S. Minot I am under special obligations for his kindness in allowing me to examine the fine series of embryos in his charge. I also desire to thank Dr. C. B. Davenport for permission to continue part of this work in the laboratory at Cold Spring Harbor, Long Island.


I shall present the results of my studies under the following headings :

I. The Formation and Structure of the Pharnygeal Pouches.

II. The Later Modifications and Fate of the Pharnygeal Pouches.

In this study I was enabled to examine the following stages, which I here present in the order of progressive development:


Pig.

(3) 6.5 mm., (M^) (1)

(6) 9.0 mm., (M^) (2)

(7) 10.0 mm.. No. 401 (4)

(8) 12.0 mm., " 518 (5)

(9) 13.5 mm., (M')

(10) 14.0 mm., No. 65

(11) 17.0 mm., " 51

(12) 18.0 mm., (M^) (15) 20.0 mm.. No. 542

(17) 24.0 mm., " 64

(18) 25.0 mm., (M^)

(22) 32.0 mm.. No. 74

(23) 35.0 mm., (M)


Cat. 4.6 mm.,


6.2 mm., 9.7 mm., (14) 10.7 mm., (16) 15.0 mm., (21) 23.0 mm., (24) 31.0 mm.,


No. 398 " 413

" 380 " 446 " 474 " 436 466 500


Rabbit.

(13) 14 days, 10.0 mm.. No. 157 • " ~ - 576


(19) 16*

(20) 18

(25) 20

(26) 21


17.8 mm., 29.0 mm.


172


I. The Formation axd Structure of the Pharyngeal Pouches,

Under this heading I shall describe all stages leading up to the complete formation of the four pairs of pharyngeal pouches characteristic of the mammalian embryo. The stages here considered include Nos. 1 to 4, inclusive.

The earliest stage of development of the pharynx and its appendages was shown in a cat embryo of 4.6 mm., No. 398 of the Harvard collection (Figs. 55 and 56.) The embryo is approximately straight, the headfold is distinctly differentiated, but the posterior two-thirds of the enteric cavity opens widely into the yolk vesicle (at x in the. figures). The neural tube is closed except anteriorly, where a narrow cleft still persists. The optic vesicles are present, but there is no sign of the optic cups.


190 Henry Fox

The pharynx anteriorly is in contact with the stomatodeal plate (St.). As a whole it is a relatively wide, dorso-ventrally flattened sac. Only two pairs of i^haryngeal pouches are present as wide lateral diverticula of the pharjTix. Of these the first (Ph. P. I) alone reaches the ectoderm and joins with it for a short distance. The second pair (Ph. P. 2) are only barely indicated as faint outbulgings, the one on the left being the more distinct of the two.

The pericardium is of small size, in striking contrast to the enormous bulk it attains in later stages. It contains the inner portions of the great vitelline veins (v. v.), which are joined together only at their extreme anterior ends. Only one pair of fully developed aortic arches is present — the first or mandibular (ao. i). These extend dorsally in front of the first pouch and join the paired dorsal aortse. Two prominent out-bulgings from the sides of the vitelline veins are probably destined to form the common trunk from which the remaining aortic arches subsequently arise.

A cat eml^ryo, No. 413 of the Harvard collection, shows the next stage in advance (Fig. 57). The posterior part of the body is still approximately straight, but the head portion is strongly flexed upon it and is of relatively much greater extent than before.

Anteriorly the stomatodeal plate has disappeared. The phar}Tix is, as before, a wide flattened sac, but its width in its anterior portion is somewhat greater than in its posterior part. Its floor is somewhat deeper than before and close to the mouth is produced into a deep median groove — the median oral groove (M. GE.).

The hypophysis (HYP) appears as a blunt protuberance from the dorsal side of the pharynx close to its anterior extremity. Three pairs of phar}Tigeal pouches are now present. The first two pairs reach tlie ectoderm and join with it for a considerable extent (see light areas of Ph. P. 1 and 2, Fig. 57). Of the third pair, the pouch on the right side reaches the octoderm, while that on the left is still removed by a slight interval from it.

The first pharyngeal pouch forms a relatively large transverse fold. The greater part of its lateral margin is in contact with the ectoderm. The area of contact is widest dorsally and diminishes progressively in width toward the ventral side. The extreme ventral part of this margin extends a slight distance below the region of contact as a free edge, which then turns suddenly inwards as the ventral margin of the pouch. This part of the pouch projects slightly below the floor of the pharynx and thus forms a ventral diverticulum of the pouch.


The Pharvnoeal Pouches in the Mammalia 191


■ J "fc


The second pharyngeal pouch, although considerably smaller than the first, is essentially similar to it. It has a ventral diverticulum, which is somewhat less prominent than the same part in the first pouch.

The third pharyngeal pouch is considerably smaller than its two predecessors. It forms a finger-like outgrowth, which extends outward^ and downwards and, in the case of that on the right side, joins the ectoderm. The left pouch does not quite reach the latter.

The fourth pharyngeal pouch is only barely indicated by a slight bulging of the walls of the pharynx behind the base of the third pouch.

The first three pairs of aortic arches are now fully developed and a fourth is beginning to develop.

In a pig embryo of 6.5 mm. (M- of my collection) and a cat of 6.2 mm. (No. 380, Harvard collection) all the pharyngeal pouches and their associated parts are typically developed. The two embryos show almost the same relative stage of development, but that of the pig shows a slightly more primitive condition. It will, accordingly, be considered first.

The pharynx (Figs. 1, 2 and 3) shows four complete pairs of pharyngeal pouches, all of which have a more or less extensive contact with the ectoderm of the corresponding grooves. Between the first two pairs of pouches the pharynx is considerably wider than in the region between the last two pairs. Anteriorly the hypophysis (HYP.) projects forward as a blunt protuberance, and immediately back of it arises a minute conical process, the representative of Seessel's pocket.

The pharyngeal pouches in general have the form of vertical winglike expansions projecting outwards and slightly backwards from the side walls of the pharynx. Typically, they are joined to the pharynx by a relatively narrow base and only laterally dilate into the wing-like expansions mentioned. Each pouch is attached laterally to the ectoderm. The extent of this attachment is shown by the clear areas in Fig. 1. As these show, it varies greatly, being most extensive in the second, where it includes almost the entire lateral margin. A similar relation is noted by Hammar in the corresponding stage in man, and, as the figures of the next stage show (see Fig. 58), it holds in the cat.

A conspicuous feature — shown best in the second and third pouches — ■ is the presence of deep ventral projections to the pouches (V.D. 1-4). They reach to a greater or less extent below the floor of the pharynx. Hammar calls them the ventral diverticula. They appear, from all published figures examined, to be constant at the corresponding stage in all mammals so far investigated.


192 Henry Fox

A ventral view (Fig. 3) shows a number of important features. Projecting from each side of the pharynx are the four pharyngeal pouches, each with its ventral diverticulum. That of the first pouch (Y. D. 1) fornls a low narrow ridge extending from the infero-lateral angle of the pouch inwards and slightly backwards quite to the median line, where it joins the corresponding ridge of the opposite side. There is thus formed a complete transverse V-shaped fold, the apex of 'the fold being the meeting point of the two opposite limbs. The ridge corresponding to the median oral groove begins immediately in front of this apex. The shallow impression between corresponds to the tuberculmn impar of His. Just behind the apex is the median thyroid. The latter consists of two lobules joined to each other and to the pharynx by a slender epithelioid cord.

The ventral diverticula of the next two pouches are much deeper than that of the first, but are largely confined to the lateral half of the pharynx. A faint ridge, however, extends from the base of the second diverticulum to the median line, where it is joined by a similar ridge from the third pouch (see Fig. 60 of the next series for this condition). The two sets of ridges thus converge to form a rather low protuberance immediately above the thyroid and in front of the tracheal ridge.

The presence of these inner low ridges connecting the opposite ventral diverticula of the second and third pouches shows their essential agreement in this respect with the first pouch. Only, in the case of the two former, the lateral half of each ridge is produced far below the level of the inner portion, while in the first pouch the depth (its height) of the ridge is throughout approximately uniform.

Owing to the form of the ventral diverticula of the second and third pouches there is left between their opposite lateral halves a considerable space, in which is lodged the apical portion of the heart along with the large arteries radiating from it (Fig. 3). The prominent aortic arches at this time are the third (carotid), fourth (aorta typica) and the fifth (pulmonary). The latter has a small posteriorly directed branch — the later pulmonary artery (Fig. 3, Pul.), The first aortic arch is much reduced in size and has lost all connection with the dorsal aorta. The second is also extremely reduced and is only connected with the dorsal aorta by an extremely narrow (apparently functionless) vessel.

Immediately back of the common origin of the aortic arches begins a sharp median ridge, which deepens posteriorly. It represents the future lar}Tix and trachea.


The Pharyngeal Pouches in the Mammalia 193

The first pharyngeal pouch has a greater lateral extent than any of the succeeding. As the figures show, the lateral extension of the pouches decreases regularly from before backwards. The first pouch blends with the lateral portion of the pharynx by a broad base, so that it is impossible to draw any definite line between the two. The limits assigned by Hammar, whose usage in this matter I adopt, will be given presently. Laterally the outer extremity of the pouch is produced upwards as a blunt prominence — the dorsal diverticulum of Hammar — which projects considerably above the roof of the pharynx. The dorsal diverticulum terminates in a narrow apex — ^the dorsal apex (d. a. 1) (Eecessus tympani anterior, Hammar).

Hammar includes in the first pouch the following parts: (1) The sulcus tubo-tympanicum. This is Moldenhauer's term for the prominent ridge (Fig. 2, S. T. T.) representing the antero-lateral border of the pouch. It begins externally at the dorsal apex and extends downwards, inwards and forwards, terminating close to the base of the hypophysis. (3) The latero-ventral ridge and its continuation, the ventral diverticulum. This is a narrow ridge which in its dorso-lateral portion is joined to the ectoderm. The connection includes the dorsal two-thirds of its lateral extent. The lower third forms a free edge, and this, at its lower outer angle, turns sharply inwards and backwards to form the ventral diverticulum. (3) The sulcus tensoris tympani (S. T. Ty.)- This is a term applied by Hammar to the border extending from the dorsal apex backwards and inwards to join the next part along the inner border of the hyoid arch. (4) The sulcus tympanicus posterior (S. T. P.). This term, also given by Hammar, includes the longitudinal ridge forming the inner boundary of the hyoid arch and connecting the sulcus tensoris tympani with the base of the second pouch. (5) The impressio cochlearis. This Hammar defines as a conspicuous depression on the dorsal wall of the pouch close to its origin from the pharynx. The auditory sac lies immediately above this area.

The second pharyngeal pouch is characterized, as already mentioned, by the extensive contact of its lateral margin with the ectoderm. Only at its extreme lower end does this border have a free margin. So far as the present specimen is concerned, there is no communication between the lumen of the pouch and the exterior. The closing membrane is exceedingly thin, but examination shows no break in its continuity.

The ventral diverticulum of this pouch forms a prominent quadrangular fold. The mesial half forms only a faint ridge, but the lateral


194 Henry Fox

portion is very deep. The deepest part is represented by the blunt angle immediately below the lower end of the lateral border.

Posterior to the region of the second pouch the pharynx diminishes considerably in width. Its lateral margin forms a low ridge connecting the second pouch with the third. Between this ridge and the median dorsal ridge of the pharynx is a shallow longitudinal furrow, in which is lodged the dorsal aorta.

The third pharyngeal pouch is slightly smaller than the second. It is joined by a relatively narrow base with the pharynx, but distally expands into a broad wing-like fold with a prominent ventral diverticulum. A slight dorsal diverticulum is also present. The lateral margin is in contact with the ectoderm for almost its entire length.

As in the case of the second pouch, the deep portion of the ventral diverticulum is limited to the lateral half of the pharynx. Its mesial portion is represented by a low ridge, which extends from the root of the lateral half forwards and inwards close to the median line, where it joins with the same part of the second pouch. The extreme ventral tip of the ventral diverticulum is turned toward the mesial side.

The fourth pharyngeal pouch is the smallest of the series. It is divided by a shallow constriction into two parts, a dorso-posterior portion (Ph. P. IV), which projects laterally and at one point comes into contact with the ectoderm, and a ventro-anterior bulge, which terminates blindly and corresponds to a ventral diverticulum.

From the base of the ventral diverticulum a low ridge extends forwards to the base of the third pouch. It corresponds to the mesial extension of the ventral diverticulum.

In the cat embryo of 6.22 mm. (No. 480, Harvard collection) the condition of the pharynx is essentially similar to that just described in the pig. As in the latter, four pairs of pharyngeal pouches are present.

The characteristic features of this stage are shown in Figs. 58, 59 and 60. Fig. 58 shows the lateral aspect. The clear areas on the lateral margins of the pouches show the extent to which these are attached to the ectoderm. It will be noticed that they are essentially similar to the same parts in the preceding specimen. The dorsal diverticulum of the first pouch is somewhat more elevated. The ventral diverticulum of the third pouch extends to a slightly lower, level. The fourth pouch shows more clearly its division into two portions. The dorso-posterior portion is somewhat bulbous. Its more dorsal part is


The Pharyngeal Pouches in the Mammalia 195

flattened and is produced outwards as a thin process which reaches the ectoderm. The ventral diverticulum projects almost directly forwards.

The median thyroid is of a relatively large size. It has lost all connection with the pharynx and lies at a lower level than in the preceding specimen.

Owing to the more rapid growth of the mandibular and hyoid arches as compared with that of the arches posterior to tliem, the originally almost transverse plane of the second and third pouches becomes postero-lateral. Their originally anterior and posterior surfaces thus become antero-lateral and postero-internal, respectively. Their lateral margins thus come to project backwards.

The increased antero-postcrior growth of the mandibular arch leads to a change in the direction of the tubo-tympanal border of the first pouch. The latter is at first almost transverse, but later assumes a more anterior direction. The more antero-mesial direction of this border in the present stage as compared with that in the preceding shows the beginning of the change. As growth continues the border progressively lengthens, thus giving an increased width to the basal portion of the pouch. These relations are clearly indicated in the dorsal view (Fig. 59).

Figs. 59 and 60 show the relations of the more posterior pouches to the now fully formed sinus prsecervicalis — relations which are of considerable importance in view of later developments. Owing to the great increase in bulk of the hyoid arch the posterior border of the latter projects backwards. The third and fourth arches increase but slightly in bulk and thus remain at a considerably lower level than the arches in front. The sinus is thereby formed as a deep recess, the bottom being formed by the arches mentioned. Just within the anterior margin of the sinus opens the second pharyngeal groove. The third groove occupies the middle of the inner walls. Dorsally it meets the upper extremity of the fourth groove. From this point the latter turns strongly downwards, backwards and inwards to where it meets the fourth pouch. As the latter lies at a considerably lower level than the other pouches, this part of the sinus projects inw^ards as a prominent, pointed process.

The ventral view (Fig. 60) shows some additional features. On the right side the section is taken at a slightly lower level than on the opposite side, and accordingly it shows the entire exterior of the first two arches, together with the ventral extension of the first pharyngeal


196 Henry Fox

groove. It also shows how the antero-internal angle of the sinus praecervicalis is continued ventrally into the ventral extension of the second groove. The latter has a decided anterior course, and at its mesial end meets the first groove.

On the left side the ventral wall of the sinus prsecervicalis is represented as having been removed, so that its interior is clearly shown. The internal process of the sinus is less deep than the same part on the right.

The continuous transverse fold formed by the ventral diverticula of the first pouch is clearly shown in this view. As in the case of the pig, there is no contact between this fold and the corresponding ventral extension of the groove, the two being separated by a considerable thickness of mesenchyme.

II. The Later Modifications and Fate of the Pharyngeal

Pouches.

Owing to the more or less independent course which the different pouches take in their later history, I think it will conduce to greater clearness if I consider them separately, and accordingly I subdivide the above topic as follows:

A. The Modifications of the First Pharyngeal Pouch, (a') The Formation of the Primary T}rmpanic Pouch.

(a") The Differentiation of the Tympanic Cavity and Eustachian Tube.

B. The Modifications and Fate of the Second Pharyngeal Pouch, (b') The Retrogressive Modifications of the Pouch.

(b") The Formation of the Tonsillar Fold.

C. The Metamorphoses of the Third Pharyngeal Pouch and its Derivatives.

(c') The Elongation of the Ventral Diverticulum and the Formation of the Thymus, (c") The Origin and Structure of the Carotid Gland, (c'") The Sinus Praecervicalis and its Eelation to the Thymus.

D. The Fourth Phar}Tigeal Pouch and its Transformation into the Tiateral Thyroid and Glandule Thyroidienne.

A. THE modifications OF THE FIRST PHARYNGEAL POUCH.

(a') The Formation of the Primary Tympanic Pouch. The pharynx is essentially alike in a pig of 9 mm. (Series M^, my collection) and in a cat of 9.7 mm. (ISTo. 446, Harvard series). Fig. 61


The Pharyngeal Pouches in the Mammalia 197

gives a ventral view in the latter. The first pharyngeal pouch is wider in the antero-posterior direction than before — a change connected withr the anterior growth and elongation of the oral cavity and the consequent prolongation in the same direction of the attached tubo-tympanal rim. The ventral diverticula are slightly less prominent. Together they form a low V-shaped elevation on the floor of the pharynx. Just external to the median apex formed by the convergence of the arms of the V each is joined by one of the pair of folds forming the outer line of the tuberculum impar (Tub.). Close to the lateral margin each arm is crossed by the broad alveolo-lingual fold (AL.F.). A slight distance in front the latter meets the vestibular fold (V.F.). Immediately back of the point of convergence a lateral fold (SM.F.). is given off, which extends obliquely outwards and backwards over the lateral ridge of the latter. The formation of this fold marks the initial step in the development of the later important submeckelian fold.

The tuberculum impar arises as a result of the bipartition of the median oral ridge. The crest of the latter widens and its middle part then becomes depressed to form a shallow concavity — the ventral counterpart of the tuberculum.

In the pig of 10 mm. (No. 401, Harvard series. Figs. 4-6) the pouch is joined to the ectoderm by only the dorsal third of its lateral ridge. The remainder of this border is now free and forms a low fold separating the antero-lateral and postero-lateral surfaces of the pouch. Ventrally it is continuous with the ventral diverticulum. Where the transition takes place the alveolo-lingual swelling cuts across it at right angles, forming here the line of demarcation between the pouch and the pharjTix.

The ventral diverticula present no new points of interest. The swellings which marked the lateral boundaries of the tuberculum impar are now relatively inconspicuous, having been absorbed along with the adjacent parts of the pharyngeal floor in the broad depression (representing the anlage of the tongue) lying between the alveola-lingual ridges.

Dorsally the pouch projects relatively higher than hitherto and terminates in a more acute apex. This condition is not due to the growth dorsalwards of the pouch, but is a result of a ventral displacement of the phar}Tix. As a comparison of the figures shows, the formation of the neck of the animal is attended with a ventral (caudal) flexure of the posterior half of the pharynx. The flexure also affects to a minor


198 Henry Fox

degree the remainder of the pharynx, tending to displace it to a lower level. This tendency, however, is checked by the fact that the first pair of pouches is still attached to the ectoderm l)y their lateral extremities. These points are accordingly relatively fixed in position, and, as the basal portion of the pouch sinks in response to the general lowering of the pharynx, the structure attains the pronounced ascending course characteristic of it at this stage.

In consequence of this change the basal portion of the pouch has assumed an almost horizontal plane, while its peripheral part ascends almost vertically. Where the two parts meet there is on the lateral surface a slight ridge extending from the lateral ridge to the base of the vestibular fold. It corresponds with the fold mentioned in the preceding stage as forming the beginning of the submeckelian fold (Fig. 6, S.M.F.).

This fold subdivides the antero-lateral wall into two surfaces, an external, dorso-lateral and a mesial ventro-lateral surface. The former forms an elongated triangular area limited dorsally by the tubotympanal crest and posteriorly by the lateral ridge. The latter forms a smaller triangle bounded internally by the alveolo-lingual fold and posteriorly by the lateral ridge.

The anterior prolongation of the tubo-tympanal ridge is more pronounced than in the preceding stage. The difference is due to a continuance of the process, already mentioned, of anterior elongation of the oral cavity.

A pig of 12 mm. (No. 518, Harvard series. Figs. 9-11) shows the pharynx only slightly larger than in the stage just described. The continued anterior elongation of the oral cavity has given the tubotympanal crest a decided antero-posterior course. The pouch retains its connection with the ectoderm only at its dorsal apex. The lateral ridge forms only a low prominence extending from the dorsal apex to the ventral diverticulum.

The dorsal apex appears 1)roader and more depressed than in the preceding stage. This condition, I think, results from the lateral flexure of the apex in consequence of the general growth in width of the head.

At this stage the pouch has the essential features of the primary tympanic pouch of Kastschenko. This investigator considered the pouch as merely a widened diverticulum of the lateral wall of the pharynx and regarded the lateral ridge as alone representing the first phar}Tigeal


The Pharyngeal Pouches in the Mammalia 199

pouch. As Hanimar shows and my observations confirm, Kastschenko's conception of the pouch was entirely too limited and was doubtless due to his not examining earlier stages in which it is more typically developed.

The primary tympanic pouch at this stage is a dorso-ventrally flattened triangular fold which arises by a broad base from the pharynx and terminates peripherally in the dorsal apex. The pouch as a whole lies almost horizontally, but towards the lateral edge it turns sharply upwards. Its walls are medio-dorsal and lateral. The former is limited laterally by the tubo-tympanal crest, dorsal apex and posterior tympanal borders. All below these limits is embraced in the lateral wall. This is divided by the lateral ridge into two surfaces, antero-lateral and latero-posterior. The antero-lateral surface is further subdivided into two areas — dorso-lateral and ventro-lateral — by the submeckelian fold.

The ventral diverticula now form a pair of low swellings. Mesially they are interrupted by a shallow longitudinal groove connecting the tongue concavity with the deep hollow in front of the larynx.

In a pig of 13.5 mm. (Series M^ of my collection) the condition of the pouch is intermediate between that last described and the next. The only feature that calls for remark is the presence at the dorsal apex of a short narrow process by which the pouch retains its last connection with the ectoderm (Fig. 37).

In the pig of 14 mm. (No. 65, Harvard series. Figs. 14-16) the primary tympanic pouch has separated entirely from the ectoderm and now lies some distance below it — a condition due to the greater lateral growth of the head compared with that of the pouch.

The pharynx at this time begins to show modifications due to its own differential growth. The increase in width of the anterior half is considerably greater than in the posterior portion. Thus, while the distance between the apices of the first pair of pouches has increased appreciably since the last stage, that between the same parts in the second pair remains approximately the same. In consequence of this the pouch now shows a more pronounced lateral projection. The tensortympani crest turns sharply inwards and joins the posterior tympanal border at an obtuse angle. The latter border also shows a tendency to assume a more transverse trend. The second pouch appears as a rounded prominence at the postero-internal angle of the tympanic pouch.

The ventral half of the lateral ridge and its continuation, the ventral diverticulum, have disappeared. Their former position is only indis


200 Henry Fox

tinctly indicated by low swellings on the under side of both pouch and pharynx. The dorsal half of the lateral ridge, however, is continued into the submeckelian fold, and these are now slightly more prominent. Together they now form a continuous crescent-shaped fold extending from the dorsal apex to the base of the vestibular fold. It underlies, for the greater part of its length, Meckel's cartilage. For this reason I have called it the submeckelian fold. The shallow depression in the lateral wall which it subtends I call the Meckelian fossa.

The paired ridges which formerly limited the tubercalum impar laterally have now become blended with the epithelium covering the tongue anlage. The formation of the latter has been accompanied by the progressive downgrowth of the surrounding alveolo-lingual crests, particularly in their anterior portion. The deep space thus enclosed is filled with the tissues of the organ. Posteriorly this space is now connected by a deep groove with the space in front of the larynx.

An early stage in the formation of the external auditory meatus is shown by the conical indentation projecting under the pouch. Its inner angle terminates a short distance below the latero-posterior surface. The two structures are nowhere in contact, a moderately thick layer of mesenchyme intervening between them.

In the pig of 17 mm. (No. 51, Harvard series. Figs. 19-21) the primary tympanic pouch is slightly more expanded and depressed. The dorsal apex has become flattened out to a low rounded prominence and has sunken to a lower level, so that it scarcely projects above the level of the pharyngeal roof. The tubo-tympanal crest in consequence is almost horizontal. Anteriorly it turns sharply inwards to form the relatively short tubal portion, the remainder forming the tympanic part (see Fig. 20).

On the lateral wall the submeckelian fold forms a prominent, projecting ridge. It extends from the dorsal apex downwards and forwards to the latero-inferior edge, when it projects as a convex ventral pocket. In front of this region it suddenly dies out, forming only a low fold (Fig. 19, y.), continued to the base of the vestibular fold. The interval outside of this part is occupied by Meckel's cartilage. The latter ascends from the mandibular arch in the angle between the submeckelian and the vestibular folds and thereby comes to lie in front of and above the former. The presence of the cartilage in the angle mentioned has probably some close connection with the separation of the two folds. Its presence would inhibit continued lateral extension


The Pharyngeal Pouches in the Mammalia 201

of the connecting portion, while it would not interfere with such growth in the remainder of the fold. The latter would then continue to expand laterally and would thus give rise to the prominent projection which it forms at this stage.

The form of the pharynx is essentially the same in a pig of 18 mm. (Series M% my collection) and a rabbit of 14 days (No. 157, Harvard series. Fig. 70). It differs but slightly from that last described. The greater part of the tympanic pouch lies in an almost horizontal plane, only its extreme lateral portion being slightly upturned (Figs. 41-46). The dorsal apex forms only a low eminence, the meeting point of tubotympanal border, submeckelian fold and tensor tympani border.

The most important feature of this stage consists in the definite segregation of the neighboring skeletal structures, particulary Meckel's cartilage and the auditory capsule. Their formation is so intimately associated with certain later modifications of the pouch that a short description of their essential characteristics is necessary. Meckel's cartilage (Mck.) is a stout rod, which, as already mentioned, rises from the mandibular arch in the angle between the submeckelian and vestibular folds and then turns obliquely backwards above the former fold (Figs. 41-43). Close to the posterior margin of the fold it sends down the stout manubrium which curves around the back of the fold and terminates in a slight depression — the manubrial fossa — immediately beneath (Figs. 43-44). The submeckelian fold is thus wedged in the angle between Meckel's cartilage and the manubrium and is thus relatively fixed in position (Fig. 43). This relative fixity of the fold is an important factor in the final transformation of the pouch into the definitive tympanic pouch and Eustachian tube.

The auditory capsule occupies the depression (Impressio cochlearis) between the dorso-internal surface of the pouch and the roof of the pharynx.

(a") The Differentiation of the Tympanic Pouch and Eustachian Tube.

In a pig of 20 mm. (No. 542, Harvard series, Figs. 23-27) we observe the beginning of the changes leading to the final transformation of the primary tympanic pouch into the definitive pouch and Eustachian tube. The transformation appears to be closely connected with a continuance of the processes already indicated. Of these we may recall (1) the ventral (caudal) flexure and elongation of the posterior half of the pharynx in connection with the formation of the neck, (2) the


202 Henry Fox

anterior extension and flexure of the mouth, and (3) the relative fixation of the primary tympanic pouch b}' the differentiation of the surrounding cartilages.

As a result of the flexures of the pharynx and mouth the common structure now has the form of an arch (Fig. 23), the apex of the arch heing that part lying between the primary tympanic pouches. From . /. side of this part each pouch projects as a broad, flattened fold, which towards its periphery turns strongly upwards so that the apex again extends some distance above the roof of the pharynx. Together the tubo-tyrapanal and tensor tympani borders form an arched curve, the apex being formed by the dorsal apex (Fig. 23). The submeckelian fold (S-M.F.) has much the same appearance as before. It is completely separated from the vestibular fold. It, however, no longer projects below the ventral line of the pouch, but lies a slight distance above it on the lateral surface. This change has been effected by a process which only becomes noticeable in the region of the pouch at this time. This is the downward growth and posterior extension of the alveolo-lingual folds (A.L.F.). As these gTow down they carry with them the adjacent ventro-lateral wall of the pouch, and thus the latter loses its original horizontality and assumes an inclined position. Its surface thus comes to be more nearly continuous with the plane of the dorso-lateral portion. Since the submeckelian fold forms the dividing line between the two portions, it comes, in consequence of this change, to occupy its present relatively higher level on the side of the pouch.

An important, but at this stage inconspicuous, feature is a shallow indentation on the posterior tympanal rim between it and the second pouch (see Figs. 26-27, z.). The latter pouch is now so small that the exact line of demarcation is not easily recognizable. A slight ridge (Fig. 26, p-m.f.), however, which extends from the indentation to the submeckelian fold, enables one to fix upon this point as being between the two structures. The same ridge, showing the same relations, is present in the immediately preceding stage when the second pouch was still clearly recognizable. This ridge later becomes the prominent elevation limiting the manubrial fossa posteriorly.

In the cat of 10.7 mm. (No. 474, Harvard series) the tympanic pouch shows a slight advance. The indentation between the pouch and the second pouch is slightly deeper and consequently the posterior tympanal crest now forms a rounded lobe projecting dorsally. In all other respects this stage is so similar to the preceding that further description is unnecessary.


The Pharyngeal Pouches in the ]\Iammalia 203

The cat of 15 mm. (Xo. 436, Harvard series, Fig. 63, 64) presents the next stage in the modification of the tympanic pouch. The incision, which in the preceding stages had just begun to form Ijetween the tympanic poucli and the dorsum of the second pouch, is now much deeper. The postero-Iateral lobe of the pouch in c<.«-Lsequence protrudes more strongly in the dorso-posterior direction '^ 'iks a result of the incision a new ventro-mesial border (Y.M.E.) has begun to form between the base of the pouch and the pharynx. Posteriorly this border connects by a rounded angle with the posterior tympanal border (Fig. 64, s. t. p.).

The connection of the pouch with the pharynx is both relatively and actually of lesser extent than in the preceding stages. This condition represents the commencement of the gradual constriction of the connecting part as a result of the anterior extension of the incision.

Owing to the increased depth of the intervening incision the tympanic pouch is now completely separated from the second pouch. This condition is apparently produced in the following manner : It will be recalled that the pouch has now become relatively fixed in position by being included between Meckel's cartilage with its manubrial process and the auditory capsule. The neighboring lateral walls of the pharynx, on the other hand, are continuously being displaced to a lower level by the downgrowth of the alveolo-lingual margins. Among the parts thus carried down is the concavo-convex fold representing the dorsal remnant of the second pouch (Ton.F.). In consequence of this displacement of the second pouch and the relative fixity of the tympanic pouch, the incision (z) between the two spreads dorsally over the second pouch and reaches the longitudinal ridge (P-S.F.) lying immediately internal to the base of the pouch (cf. Figs. 26, 70 and 64). Thus the base of the tympanic pouch is placed in connection with this ridge, which is the extel-nal expression of the groove extending backward from the Eustachian opening between the levator cushion and the salpingo-pharyngeal fold. For convenience in description I shall speak of it as the post-salpingeal groove.

As a result of the process just described the dorsum of the second pouch comes to lie on the lateral surface of the pharynx below the posterior margin of the tympanic pouch. With the subsequent downgrowth of the alveolo-lingual folds it is carried farther ventralwards, and, as will be described later, is finally transformed into the tonsillar recess.


204 Henry Fox

On the expanded lateral wall of the tympanic ponch two prominent outstanding folds are now present. The anterior is the snbmeckelian fold; it shows the now deep Meckelian fossa on its dorsal side. The posterior fold is less prominent; it corresponds to the ridge formerly mentioned as forming the posterior limit of the manubrial fossa. The latter now forms a depression of considerable depth.

A rabbit of IGi/s days, 17.8 mm. (Ko. 576, Harvard series. Fig. 71) shows the constriction of the tympanic pouch still further progressed. The post-salpingeal fold (p. s. f.) is more convex. The dorsum of the second pouch (ton. f.) is separated by a short interval from the base of the tympanic pouch. The remaining features of the pouch are essentially like those in the following stage.

This stage is represented by a pig of 24 mm. (No. 64, Harvard series, Figs. 29, 30). The constriction of the tympanic pouch has now reached a stage where its connection with the pharynx embraces about twothirds of its former extent. The ventro-mesial margin is accordingly of considerable length. The posterior half of the poiich projects strongly backwards as a wide, cup-shaped fold.

The submeckelian fold forms a wide, almost horizontal shelf (Tigs. 47-50, s. m. f.). Laterally it reaches considerably beyond the dorsal apex, so that it is clearly visible from above (Fig. 30). On lateral view it appears at a considerably higher level than before. This position it has obtained partly as a result of its lateral extension and the consequent flattening of its dorsal surface and partly from the ventral downgrowth of that portion of the pouch lying immediately below it (cf. Figs. 48-50, with Fig. 43).

The manubrial fossa forms a shallow impression between the submeckelian (s. m. f.) and post-manubrial folds (p. m. f.). The latter is much less prominent in the pig than in the equivalent stage of the cat.

The pig of 25 mm. (Series M^, my collection), while slightly more advanced than the preceding, is essentially similar so far as the tympanic pouch is concerned. Figs. 47-51 give views of several transverse sections of the structure.

The next step in advance is shown by a cat of 23.1 mm. (No. 466, Harvard collection. Figs. 66-67). In this case the tympanic pouch and Eustachian tube are first clearly differentiated from each other. The former is a wide, cup-shaped expanse, concave dorsally. As a whole, it has a decided ascending direction. The postero-lateral border (P-L.B.) forms a highly elevated ridge. Immediately back of the


The Pharyngeal Pouches in the Mammalia 205

dorsal apex it is interrupted by a deep incision — the incissura tensoris (I.Tn.). Anterior to the apex is the submeckelian fold (SM.P.) facing at this stage in the antero-dorsal direction. Laterally its margin is so far upturned as to hide from view the adjacent part of the tubotympanal border.

The nianiibrial fossa (]\In.r.) is very dee]). It lies immediately l:»elow the tensor incision^, bounded anteriorly by the submeckelian fold and posteriorly by the post-manubrial fold (P-M.F.). The external auditory tube lies a short space below the fossa, but is still separated from it by a considerable thickness of connective tissue.

The most noteworthy feature of this stage is the initial division of the pharynx into its oral and nasal portions by the backward extension of the palatine incisions. The oral cavity has been entirely separated from the nasal tube, but the constriction of the pharynx has only begun in its more anterior part. The constriction, as the figures show, takes place in the part lying below the post-salpingeal fold, between it and the tonsillar fold.

The rabbit of 18 days (Harvard series), while showing a slight difference from the preceding, is yet so closely similar that a full description is unnecessary. Its general features can be seen by consulting Fig. 73. The most marked feature is the greater posterior extension of the palatal constriction. The broad grooves continued back from the latter over the sides of the pharynx represent the posterior palatine grooves (Arcus pharyngo-palatinus), p. pi.

In the pig of 32 mm. (ISTo. 74, Harvard series, Pigs. 32, 33, 52) the Eustachian tube is more constricted than in the preceding stage. The tympanic pouch projects sharply outwards and backwards. Close to where it joins the tube it gives off the still prominent submeckelian fold. Where the latter and the tubo-tympanal borders meet is the dorsal apex (Eecessus anterior, D.A.I). Below the latter on the lateral wall is the now crescent-shaped post-manubrial fold. The Tatter arches around underneath the manubrial fossa (Mn.F.) and becomes continuous with the portion extending to the dorso-lateral margin of the pouch. Immediately below the fossa the ventro-lateral surface of the pouch is flattened and is adpressed against the inner part of the external auditory tube. Only an exceedingly thin layer of connective tissue intervenes between the two structures (Fig. 52).

The cat of 31 mm. (No. 500, Harvard series. Figs. 68-69) gives the final stage in its species. The Eustachian tulie is very narrow, while


206 Henry Fox

the tympanic ponch is widel}^ expanded, particularly in its posterior part. It still retains its cup-like form, the concave surface fitting closely against the ventro-lateral wall of the auditory capsule. The submeckelian fold (S-M.F.) is relatively not so prominent as earlier. The dorsal apex or anterior recess (D.A.I) projects strongly outwards. The manuljrial fossa (Mn.F.) forms a deep hollow on the more dorsal portion of the lateral surface. It is largely surrounded by the now high and conspicuous post-manubrial fold (P.M.F.). Below the fossa is the surface already mentioned as being in close relation with the external auditory tube. The remaining posterior extension of the pouch is simply applied to the neighboring part of the auditory capsule.

The latest stage studied is shown by a rabbit of twenty-one days. Fig. 73 shows the more important points. The manubrial fossa (]\In.F.) is still rather deep in its dorsal half, but vertrally becomes very shallow and there tapers out to a point. The area entering into the constitution of the tympanic membrane is more extensive than in the preceding stage. It now includes a considerable part of the surface in which tbe manubrial fossa is located.

The ]Meckelian fossa (Mk.F.) is almost ol^literated ; it persists as a very shallow impression on the antero-dorsal margin close to the union of the pouch with the Eustachian tube.

Review and Comparisons.

The foregoing results render it highly probable that the developmental history of the first pharyngeal pouch is essentially the same in the three species of mammals studied.

This history I have subdivided into three periods, as follows :

(1) The period of formation of the typical pouch.

(2) The period of transformation of the pouch into the primary tympanic jjouch.

(3) The period of differentiation of the tympanic pouch and Eustachian tube and their subsequent modifications.

Period I. The formation of the pharnygeal pouches takes place in the usual manner — beginning with the most anterior and ending with the most posterior.

When typically developed the first pharyngeal pouch has the form of an approximately transverse vertical fold. At its dorsal lateral


The Pharyngeal Pouches in the Mammalia 207

angle it projects dorsalwards as a narrow prominence, the dorsal apex (recessus anterior). From this apex three prominent ridges diverge, i. e., antero-lateral, lateral and postero-lateral. The first extends diagonally inwards and slightly forwards. It forms the sulcus tubotympanicus of Moldenhauer. The lateral ridge is that by which attachment to the ectoderm is effected. The area of attachment includes nearly its entire extent. Yentrally this ridge is continued into the ventral diverticulum. The postero-lateral ridge extends obliquely inwards and backwards from the dorsal apex to the dorsum of the second pouch.

The ventral diverticula of the first pair of pouches are at first more prominent than those of the succeeding, but they are soon outstripped by the latter. Typically they form a pair of low, but sharp folds, which at first are continuous across the median line of the pharynx.

Period II. The most important changes leading to the transformation of the first pouch into the primary tympanic pouch are the following :

The gradual separation of the pouch from the ectoderm. This process begins on the ventral side and progresses dorsalwards until complete separation has been effected. In consequence of this separation the lateral ridge becomes greatly reduced and partly absorbed int"S the neighboring walls of the pouch.

The tubo-tympanal border becomes prolonged in the anterior direction. This change is produced as a result of the elongation in the same direction of the adjacent part of the oral cavity.

The ventral diverticula first become interrupted in the mid-line, and later gradually disappear as a result of absorption into the fioor of the pharynx.

The basal or mesial portion of the pouch is displaced ventralwards in consequence of a corresponding displacement in the adjacent part of tlte pharynx itself. This portion of the pouch thus assumes an almost horizontal position. At first the peripheral part, owing to its continued attachment to the ectoderm, retains its ascending course, joining the mesial portion at a sharp angle. Later, after complete separation from the ectoderm, the peripheral portion also sinks, and thereby assumes a plane more nearly like that of the mesial portion.

The submeckelian fold is formed by the union of the dorsal remnant of the lateral ridge with the diagonal fold separating the basal and peripheral portions of the pouch. At first the fold is continuous an


208 Henry Fox

teriorh' with the lateral margin (vestibular fold) of the oral cavity. Subsequently this connection is interrupted and the submeckelian fold then grovrs out as a prominent shelf-like protuberance underlying Meckel's cartilage.

Period III. The transformation of the primary tympanic pouch into the definitive tympanic pouch and Eustachian tube is marked by the following features :

The peripheral jjortion of the j)0uch becomes relatively fixed in position by the segregation of Meckel's cartilage with its manubrial process and the auditory capsule.

The basal j)ortion, on the other hand, continues to be carried down by the downgrowth of the alveolo-lingual fold.

The combined effect of these two processes is to give the pouch a peripherally ascending course.

An incision forms at the postero-internal angle of the pouch between it and the dorsum of the second pouch. This incision rapidly extends forwards as an ever-widening cleft between the base of the pouch and the wall of the pharj^nx.

In consequence of this process the connecting part of the pouch is progressively constricted until it forms a narrow tube, the Eustachian tube.

The remainder of the pouch retains its original wide extent and forms the tympanic pouch.

The later changes relate mainly to modifications in the detailed structure of the tympanic pouch. Among them are the formation of the manubrial fossa, the reduction of the submeckelian fold and the formation of the tympanic membrane.

The manubrial fossa lodges the ventral extremity of the manubrium. At first it is a shallow impression on the lateral surface immediately underlying the posterior part of the submeckelian fold. With the formation of the definitive t3^mpanic pouch it rapidly deepens to form a cup-like depression. Subsequently this elongates at its ventral extremity to form the fissure-like groove characteristic of its final ^fage.

The submeckelian fold is at first very prominent and partly encloses a Meckelian fossa. The latter later assumes a more flattened form and the fold at the same time liroadens until it is al)sorbed into the wall of the pouch. In the latest stage the submeckelian fold forms only 'an inconspicuous swelling on the outside of the tubo-tympanal border.

The tympanic membrane is formed by the progressive approximation of


The Pharyngeal Pouches in the Mammalia 209

the venlro-lateral portion of the tympanic pouch and the neighboring dorso-internal surface of the external auditory tube. At first the two surfaces are separated by a considerable interval filled with connective tissue. This interval later becomes narrower until it is reduced to an exceedingl}^ thin layer — the membrana propria of the definitive membrane. The formation of the tympanic membrane begins on the ventrolateral surface of the pouch, but subsequently it extends dorsalwards so as to include the portion containing the manubrial fossa.

After its differentiation the j)ouch as a whole increases in width both laterally and longitudinally. Its posterior portion extends backwards as a prominent projection (posterior recess). The margins become upraised and thus the pouch as a whole assumes a cup-like form. The concavity on the dorsal side, corresponding to the promontory, lies close to the latero-ventral surface of the auditory capsule.

As already mentioned, my results make it highly probable that the developmental history of the first pharyngeal pouch is in all important respects similar in the three types studied. This probability is still further heightened when the results are compared with those obtained by other investigators. Thus Piersol has described and figured tlie earlier stages in the rabbit. They agree in every important particular with the corresponding stages in the cat and pig.

The most complete comparison can, thanks to the work of Hammar, be made with the human species. Hammar figures nearly every stage from the typical first pharyngeal pouch to the end of foetal life. I have carefully compared Hammar's descriptions and figures with mine and find that in every important particular they are applicable to the types examined by me. It is in fact difficult to recognize any importaiJt differences, at least as late as the stage when the tympanic pouch and Eustachian tube have been fully differentiated. In the ease of the human sjDecies the tympanic pouch during the later fcetal life gives rise to several outgrowths from its dorso-lateral margin. Prom one of these the mastoidal cells arise as a complex series of buds. In the rabbit these outgrowths had not formed at as late a stage as that of an animal of 21 days. Wliether they are present at the same stage of development in the other two forms I am, at present, unable to say. The latest stages of each, which I was enabled to examine, showed no trace of them.

Hammar does not lay as much stress as I on the submeckelian fold. He describes its formation correctly, but apparently fails to note its separation from the vestibular folds and its later lateral expansion. His figures, however, leave no doubt that in these particulars the human


210 Henry Fox

species agrees with the other types. In some of his figures I am noi certain whether Hammar means to include the submeckelian fold as a part of the recessus anterior or to limit the latter to the dorsal apex. His descriptions seem to me to favor the latter alternative. He applies, at any rate, no distinctive term to the fold, and accordingly I have felt free to call it the submeckelian fold.

The foregoing remarks make it apparent that the same essential type of development of the tympanic pouch and Eustachian tube holds in species belonging to four different orders of mammals, i. e., Eodentia (Lepus), Ungulata (Sus), Carnivora (Felis), and Primates (Homo). So far as known, other species, which have been much less thoroughly investigated, agree with this type. Accordingly, it seems reasonable to suppose that the same type prevails in the majority of ordinary placental mammals and that it represents the typical development of the structures in the class. In forms which are adapted to a special environment (Cetacea, for example) or which are farther removed from the main phylogenetic series (Edentata) it may show important modifications. So far as I am aware, these forms have not yet been investigated in regard to this point. The Marsupials and Monotremes have also not been sufficiently investigated to allow of any assertions being made concerning them. It may be added that a figure by Maurer, showing an early stage of the phar^mx in Echidna, bears a striking likeness to that of my 6.5 mm. pig and 6.2 mm. cat.

B. THE MODIFICATIOXS AND FATE OF THE SECOND PHARYXGEAL POUCH.

(b') The Eetrogressive Modifications of the Pouch.

We left the second pharyngeal pouch fully and typically developed in a cat of 6.2 mm. (Figs. 58-60). Its form at that stage is that of a postero-laterally projecting, vertical fold, which is connected by its entire peripheral margin with the ectoderm of the corresponding groove. At its dorso-lateral angle it is produced into a slight elevation forming a dorsal apex (D.A.2) similar to that of the preceding pouch, but considerably less prominent. On its ventral side the pouch is continued as a prominently projecting ventral diverticulum (V.D.2). The deep portion of the latter is limited to the lateral half of the i^harynx, its internal border forming a free edge (Fig. 60). At the base of this edge the diverticulum is continued mesially as a low fold similar to the same part in the first pouch. Like the first pouch, the second has four borders and three surfaces. The borders are antero-lateral, posterointernal, lateral and Central. The surfaces are antero-lateral, medio


The Pharyngeal Pouches in the Mammalia 211

posterior and dorsal. The antero-Iateral border (Ton.F.) extends from the dorsal apex diagonally forwards and inwards to the postero-internal angle of the first pharyngeal pouch. It forms the dividing line between the dorsal and antero-Iateral surfaces. The postero-internal border is approximately crescentiform. Laterally, owing to the posterior flexure of the pouch, its course is almost longitudinal, but basally it bends first inwards and then posteriorly and connects with the lateral ridge extending to the third pouch. It separates the dorsal and postero-internal surfaces. The lateral margin forms the part connected with the ectoderm. It separates the antero-Iateral and postero-internal surfaces. Ventrally it is continued into the ventral margin, which forms the free edge of the ventral diverticulum.

In a cat of 9.7 mm. (ISTo. 446, Harvard series, Fig. 61) the second pouch, beyond a slight increase in size, shows but few new features. For a short distance below the dorsal apex it has separated from the ectoderm — the initial step in the process which in this case begins at the dorsal end and progresses- towards the ventral. The separation is accompanied by the ingrowth of mesenchyme into the intervening space.

The latero-ventral angle of the ventral diverticulum is produced into a slight projection. As a result the inner border of the diverticulum ascends more diagonally to the floor of the pharynx.

In the pig of 10 mm. (No. 401, Harvard series) a departure from the preceding stage is shown by the slightly lower level of the second pouch. In the preceding examples the dorsal line of the pouch lay a short distance above the same line of the pharynx, while in the present stage it lies below it. This condition is probably produced by the changes taking place in the neighboring parts. The hyoid region increases in thickness more rapidly in its dorsal portion than in its relatively passive ventral part. The dorsal portion thus projects outwards over the lower and consequently the second ectodermal groove assumes a more inclined direction than before. With the latero-ventral rotation which the dorsal half of the groove undergoes it naturally results that the attached dorsal portion of the internal pouch accompanies it, at least in part, in the same direction, and thus assumes a more lateral, as well as lower, position.

In the 12 mm. pig (N"o. 518, Harvard series. Figs. 9-12) the second pouch comes to a standstill so far as any lateral growth is concerned. Thus the distance between the lateral margins of the two opposite pouches remains the same as in the preceding stage. The first pharyngeal pouch, on the contrary, continues to extend rapidly in that direc


212 Henry Fox "

tion, and at the same time carries with it the attached adjacent parts. As already mentioned, the antero-lateral margin of the second poucTi is continuous anteriorly with the latero-posterior border (S.T.T.) of the first pouch. At first the two join at a considerable angle, but as this border of the first pouch is carried outwards by the growth of the pouch, the attached antero-lateral border of the second pouch (Ton.F.) follows it and thus the angle tends to become drawn out and the borders to form a continuum. Consequently at this stage the antero-lateral border extends diagonally outwards instead of inwards, as in preceding stages.

The lateral flexure of the antero-lateral margin causes it to arch outwards above the underlying antero-lateral wall. The latter thus forms a well-marked concavity, which is limited internally by the low fold (later part of the alveolo-lingual sinus) connecting the ventral diverticula of the first and second pouches. Owing to its inclined position, this surface will henceforth be called ventro-lateral (Fig. 12, c. v.).

Corresponding to the depressed condition of the ventro-lateral surface, the dorsal surface, which is everywhere closely adpressed against the underlying wall, is raised into a low dome-shaped convexity. The latter I shall call the dorsal prominence (Fig. 10, D.Pr.).

The ventral diverticulum (v. d. 2) is reduced to about three-fourths of its former vertical extent. • This change I am inclined to attribute, in part at least, to the outward extension of the antero-lateral portion of the pouch. The latter would set up a tension in the remainder of the pouch which would lead to a partial absorption of the diverticulum into the adjacent portion of the pouch. A fact favoring the existence of such a tension is the presence upon the ventro-lateral wall of a narrow fold extending obliquely upwards from the base of the diverticulum (Fig. 12).

The lateral margin of the second pouch is now largely free from the ectoderm, the connection with the latter persisting only in its more ventral portion, where the corresponding ectodermal groove forms a deep, vertical pit (Fig. 11).

The pig of 14 mm. (ISTo. 65, Harvard series, Figs. 14-17) shows the second pouch slightly reduced in vertical extent, but produced at its ventro-lateral angle into a long, fine process (Fl.P.), the distal end of which is attached to the ectoderm. Elsewhere the pouch is free and is removed by a wide interval from the ectoderm. This condition is the result of the rapid growth in thickness of the hyoid region. As the


The Pharyngeal Pouches in the Mammalia 213

figiires (12 of last stage and 17 of this) show, this growth has not been accompanied by a corresponding increase in the pouch, which at this stage remains of the same width as in the preceding stage. Consequently, as the second pharyngeal groove is displaced more and more lateralwards, the attached ventro-lateral angle becomes drawn out into the process here shown. Owing to its form, I designate the latter the filiform process (fl. p.).

The lateral margin of the pouch is considerably less prominent than hitherto. This change appears to be produced by an actual regression of the margin. This is indicated by the fact that the distance between the lateral margins of the two opposite pouches is slightly less than in the preceding stage. The regression is probably attributable to the tension exerted upon this margin by the continued lateral extension of the adjoining antero-lateral margin with which it now joins at a wid'S angle.

As just mentioned, the antero-lateral margin (ton. f.) has continued to extend in the lateral direction. It thus has a decided antero-lateral course. Por this reason it is inappropriate to call it by the term hitherto used, and accordingly I shall hereafter speak of it as the dorsolateral margin.

The dorsal apex (D.A.2) of the pouch now forms only a slight protuberance at the posterior extremity of this margin (Fig. 15).

In consequence of the extension laterally of the dorso-lateral margin the underlying ventro-lateral surface has acquired the form of a deep concavity (Fig. 17, c. v.). The overlying dorsal wall is correspondingly raised as a broad dome-shaped prominence (Fig. 15, D.Pr.).

The ventral diverticulum (v. d. 2) has almost ceased to exist as a distinct feature. Only in its more peripheral part does it project to a fair degree below the ventro-lateral line of the pharynx. Its middle part has largely disappeared owing to the downgrowth of the alveololingual ridge (Fig. 17, al. f.) and the union of the latter with the sinus piriformis (Fig. 17, s. pi.). The place of the original diverticulum is indicated by a widening of the continuous ventro-lateral fold thus formed.

The more internal part of the diverticulum persists as a slight ridge on the inner side of the ventro-lateral fold (Fig. 17).

In the 17 mm. pig (Figs. 19-22) the second pouch has entirely severed its connection with the ectoderm, leaving the filiform process terminating l^lindly in the mesenchyme.


214 Henry Fox

Owing to the continued lateral extension of the dorso-lateral margin (Ton.F.), the original lateral border now forms a continuum with it. This leaves the dorsal apex as a minute protuberance at its posterior extremity.

The dorso-lateral margin shows no increase in length, but anteriorly it has been carried farther outwards in consequence of the extension of the tympanic pouch in that direction. It thus acquires a coiirse almost in line with the postero-lateral margin (s. t. p.) of the latter. Only a slight notch remains to indicate the dividing line between them.

The ventral diverticulum (v. d. 3) has now nearly disappeared, having been absorbed by the downgrowth of the ventro-lateral margins of the pharynx.

The features of the second pouch in the pig of 17 mm. are essentially duplicated in cats of 10.7 mm. and 12 mm.

A rabbit of 14 days (No. 157, Harvard series. Fig. 70) shows a stage somewhat intermediate between that just described and the next.

The same remark is also applicable to an 18 mm. pig (Series ]\P, of my collection).

In a pig of 20 mm. (No. 542, Harvard series. Figs. 23-27) the second pharyngeal pouch is chiefly modified as regards length and direction. As indicated by the fig-ures, these modifications are related to the ventral (caudal) flexure of the posterior half of the pharynx. As already mentioned (see account of tympanic pouch), the tympanic pouch has at this time become relatively fixed in position l)y the formation about it of the related cartilages. Consec[uently, as the pharynx continues to bend toward the ventral side, the attached second pouch tends to be drawn out and flattened (Fig. 23). This process is also accelerated by the continued deepening of the ventro-lateral margin of the pharynx.

The hitherto jDrominent dorsal prominence is now reduced to a low swelling located at the postero-internal angle of the tympanic pouch. No lateral extension of this part of the pouch has taken place since the 17 mm. stage. Its middle and posterior portions, on the contrary, have shrunken slightly, due probably to the tension produced by its elongation in the ventral direction and the continued downgrowth of the adjacent ventro-lateral margin (A-L.F.) of the pharynx.

Posteriorly the second pouch thus forms a low ridge situated on the outer side of the ventro-lateral fold (= conjoined alveolo-lingual and piriform sinuses).

The shrunken filiform process is still recognizable.


The Pharyngeal Pouches in the Mammalia 215

The dorsal apex has been absorbed into the neighboring surface of the dorsal prominence.

The ventral diverticulum forms only an inconspicuous fold in the same situation as hitherto.

(b") The Formation of the Tonsillar Fold.

The cat of 15 mm. (ISTo. 436, Harvard series, Figs. 63-64) gives us the initial step in the transformation of the remnant of the second pouch into the tonsillar fold.

In this stage the dorso-lateral fold, representing the second pouch, no longer forms a continuum with the adjacent border of the tympanic pouch, but is separated from the latter by an indentation which extends quite across its dorsal side to the longitudinal ridge (P.S.F.), forming its mesial boundary.

The second pouch thus forms an arched lateral fold. The lateral margin (Ton.F.) of the fold lies on a level with the ventro-lateral line of the pharynx. The ventral side is concave; the dorsal correspondingly convex. Anteriorly the fold is continued immediately under the ventrointernal angle of the tympanic pouch and extends to the base of the vestibular fold. Internally the fold is limited on the ventral side by the alveolo-lingual fold, on the dorsal by the adjacent surface of the pharynx. The structure thus defined is the tonsillar fold (tonsillar sinus). The concavity on its ventral side corresponds to the tonsillar prominence (tonsillenhocker).

There is no trace at this stage of the filiform process.

The rabbit of I6I/2 days (No. 576, Harvard series. Fig. 71) shows an essentiall}' similar condition. The fold (ton. f.) is more strongly arched and its lateral edge lies some distance above the lower edge of the alveolo-lingual groove. The fold is widest immediately imder the post-salpingeal ridge (p. s. f.). Its anterior continuation forms a low ridge, which probably represents an extension of the fold over the adjacent surface of the pharynx.

In the 24 mm. pig (No. 64, Harvard series. Figs. 29-30) the tonsillar fold is removed by a considerable interval from the base of the tympanic pouch (Fig. 30). Between them the surface of the pharynx is depressed, forming the palatal constriction. The present position of the fold is due to the formation of this constriction and to the continued downgrowth of the alveolo-lingual margin (A-L.F.) with which it is closely associated.


216 Henry Fox

In form the tonsillar fold (Ton.F.) of the pig of this stage hears a greater resemblance to that of the cat than to the same structure in the rabbit. The tonsillar fold of the latter has a more decided ascending plane than the others.

In the cat of 23.1 mm. (N"o. 466, Harvard series, Figs. 66-67) the tonsillar fold (Ton.F.) has attained its definite position. Tlie palatal constriction has now separated the nasal cavity from the mouth and has begun to encroach upon the pharynx. The tonsillar fold forms a wide, diagonally ascending arched fold on the side of the oral portion. Its ventro-lateral surface is, as usual, deeply concave.

The pig of 32 mm. (No. 74, Harvard series, Fig. 32) shows the tonsillar fold (Ton.F.) more nearly erect than in the cat. In outline it is approximately quadrangular and its outer (= ventral) surface is less concave than in the cat. Ventrally it is limited by the alveololingual ridge (A-L.F.), which at this stage no longer forms the lower line of the pharynx, but lies on the outer side of the giosso-epiglottic fold (vallecula glosso-epiglottica).

In the 31 mm. cat (jSTo. 500, Harvard series. Figs. 68-69) the tonsillar fold (Ton.F.) has essentially the same form as in the 23 mm. cat. As in the pig last described, its lower boundary- — the alveolo-lingual ridge — now lies on the outer side of the giosso-epiglottic fold (G.Ep.).

The palatal constriction has now completed the division of the pharynx into nasal and oral portions.

The narrow cord shown in the figure parallel with the tonsillar fold is an epithelial structure which lies free in the connective tissue to the outer side of the fold. Its significance I have not been able to solve.

In none of the stages so far studied did I observe any clear indications of the formation of lymphoidal tissue in connection with the tonsillar fold.

In the rabbit of 21 days (Fig. 73) the tonsillar fold (Ton.F.) is approximately vertical. Its lateral surface is deeply concave and lies between a dorsal and a ventral fold. The former corresponds to the supra-tonsillar recess and evidently represents the derivative of the second pouch. The ventral fold I am inclined to homologize with the infra-tonsillar recess (Y.T.) which is a derivative of the pharynx. Hammar, however, who describes a similar stage in the rabbit as well as in man and several other mammals, fails to mention this fold as the part in question. I regret that with the relatively few later stages at my disposal I have not been able to solve this problem satisfactorily.


The Pbarvngeal Pouches in the Mammalia 217

Review and Comparisons.

My investigations make it probable that the history of the second pouch, so far, at least, as its earlier stages are concerned, is similar in the forms studied. Unfortunately, ni}' rabbit and cat material was not sufficiently abundant to enable me to make this statement without qualification. However, the specimens I did examine agreed very closely with corresponding stages in the pig series. The later stages were not sufficiently numerous to enable me to make comparisons. In its general features the development of the tonsillar fold seems to agree in all forms; in details, there are undoubtedly considerable differences in tKb different species.

The history of the pouch, as mainly determined in the pig, divides itself in two j)eriods — ^the first characterized by a series of retrogressive changes in the pouch, the second by a series of progressive changes converting the remains of the pouch into the tonsillar fold.

When typically developed the second pouch has the form of a posterolaterally projecting vertical fold. Dorsally the dorso-lateral angle is produced as a dorsal apex. Yentrally it shows a prominent ventral diverticulum. Connection with the ectoderm is more extensive than in any other pouch, the entire lateral margin taking part in the formation of the versclilussmemhran.

The earlier modifications of the pouch are connected ■with the rapid lateral growth of the hyoid region. The pouch, on the other hand, remains stationary. Parts of it are, however, connected with adjoining structures, and, as these undergo displacement connected with subsequent growth, the pouch becomes profoundly modified.

Separation of the pouch from the ectoderm begins on the dorsal side and extends progressively toward the ventral. The last point to remain attached is the ventro-lateral angle of the ventral diverticulum, which becomes drawn out into a thin cord, the filiform process. The latter subsequently separates and then shrinks in length and disappears.

Largely as a result of the lateral extension of the adjoining tympanic pouch the dorso-anterior portion of the second pouch is drawn farther outwards. Its margin, which originally extended forwards and inwards, acquires an antero-lateral course and thus comes to form a continuum with the posterior border of the tympanic pouch. The underlying antero-lateral surface becomes ventro-lateral and its wall becomes depressed to form a deep concavity, which corresponds to the later ton


218 Henry Fox

sillar projection. The closely adpressed dorsal wall is correspondingly raised into a dome-shaped swelling, the dorsal prominence.

After its separation from the ectoderm the original lateral margin of the pouch recedes towards the median line. At first it forms a slight projection at the postero-internal angle of the pouch, but later this is absorbed and then forms a continuum with the dorso-lateral fold.

The ventral diverticulum early diminishes in size and later is absorbed by the downgrowth of the alveolo-lingual fold.

At the termination of the first period the remains of the second pouch form a laterally ascending arched fold, which lies at the postero-internal angle of the tympanic pouch. On its ventral side it is concave and on the dorsal correspondingly convex. Its inner boundary is formed by the alveolo-lingual ridge.

The second or progressive stage is marked (1) by the separation of the second pouch from the tympanic pouch, (3) its ventral displacement to its definitive position, and (3) its progressive modification to form the tonsillar fold.

The separation from the tympanic pouch takes place by the extension of the indentation between the two structures over the dorso-lateral surface. The second pouch thus comes to lie at a slightly lower level than the base of the tympanic pouch.

The ventral displacement takes place in connection with the continued downgrowth of the alveolo-lingual ridge and the accompanying formation of the palatal constriction. The latter forms between the base of the tympanic pouch and the dorsum of the tonsillar fold and as it enlarges the fold is pushed farther ventralwards, where it attains its final definitive position.

In its principal features the tonsillar fold in the later embryonic stages is similar in the species studied. It then forms a prominent arched fold on the lateral surface of the oral portion of the pharynx parallel to the alveolo-lingual sinus. Its ventro-lateral surface is concave, its dorsal convex.

The later modifications are concerned with the assumption of its definitive form. Owing to lack of materials, these modifications were not traced. The rabbit of 21 days shows that they may be considerable. I shall consider them further in the comparative part.

Contributions to the developmental history of the second pouch have been made by a number of investigators, chief among whom are Born, His, Rabl, Piersol, Kastschenko and Hammar. The results of these


The Pharyngeal Pouches in the Mammalia 219

authors, while agreeing in certain respects, are hopelessly discordant in others. The most satisfactory account is that given by Hammar of the development in man, Hammar also describes a few stages in the formation of the tonsil in several other species of mammals. In the main my results are in harmony with his. The only statement of his to which I cannot subscribe is that the "hiemengang" (== my filiform process?) is an ectodermal derivative. As his figures show, this structure occupies the same relative position as my filiform process. In Hammar's view this is formed by the passive deepening of the ectodermal groove produced by the growth of the hyoid region. He also pictures it as protruding above the margin of the pouch as a dorsal organ. In the pig, on the other hand, this structure is perfectly continuous with the ventro-lateral angle of the pouch and appears as a prolongation of the latter. In none of the specimens examined by me did I notice any communication between the lumen of the filiform process and the exterior. As for a dorsal organ projecting above the ventro-inferior edge of the pouch, I find no evidence of it. Accordingly, I am disposed to think that the filiform process and the "kiemengang" are independent structures. The latter would then be absent in the pig, while the former would be lacking in the human species. This view is supported by the observations of Piersol on the rabbit. He speaks of the ventro-lateral angle as continued in a blind tube. This evidently corresponds to the filiform process. He later speaks of the latter as cutting off from the pouch and undergoing changes reminiscent of the thymus. I find nothing of this in the pig. In the latter the process simply disappears — at least I have not seen any trace of it later than a 17 mm. animal. On the other hand, Piersol describes a long epithelial tube, which according to him arises from an insinking of the ectodermal groove. This corresponds to the "hiemengang" as described by Hammar. Piersol states that it originates while the filiform process is still present (twelfth day), but later (fourteenth day) it disappears without leaving a trace. If the facts as described are correct, the rabbit shows the filiform process and "Tciemengang" as independent formations. The latter is evidently an extremely temporary structure. As already mentioned, I saw no trace of it in the pig, and in this I am in agreement with Eabl and Kastschenko, both of whom studied the same animal. The short duration of the "kiemengang" may perhaps have led to its being overlooked in this animal.


220 Henry Fox

I may add here that the term kiemengang was first applied by Eabl to the endodermal structure here called filiform process. Hammar considers Eabl's account as contradicted by his results as determined in the human species, and accordingly applies the same term to the ectodermal structure. In case two independent structures, one endodermal and the other ectodermal, are found to exist in mammals, it will be necessary to return to Eabl's original use of the term.

My rabbit and cat series throw no light on this puzzling matter. In the former I did not examine sufficiently early stages, while in the latter the stages which would show the structures under consideration were lacking in the collection.

The remaining results agree fairly closely with those of Hammar. The form of the early tonsil in the forms studied by me differs somewhat from that in man, though the difference is a minor one. In regard to the rabbit of 21 days, Hammar does not speak of it as having an infratonsillar sinus. My specimen, on the other hand, shows a fold which, in my opinion, corresponds to this sinus. As, however, I was unable to examine forms in which the latter is undoubtedly present, I will not urge this homology.

The above comparisons would indicate that, although the chief features in the history of the second pouch agree in all species of mammals studied, there are considerable differences in detail. The matter of the filiform process and kiemengang would illustrate this. In later stages, as Hammar shows and as my specimens indicate, the tonsillar fold differs considerably in its structural details in various species of mammab.

C. THE METAMORPHOSES OF THE THIRD PHARYNGEAL POUCH AND ITS

DERIVATIVES.

(c') The Elongation of the Ventral Diverticulum and the Formation of the Thymus.

When typically developed, as in the cat of 6.2 mm. (Figs. 58-60), the third pharyngeal pouch bears a considerable resemblance to the second. Like the latter, it bears a deep ventral diverticulum (V.D.3), which likewise is limited to the lateral half of the pharynx, its inner half forming only an inconspicuous ridge. The deep part of the diverticulum is only slightly prolonged in a ventro-mesial direction.

The lateral margin is joined to the ectoderm for almost its entire extent, in this respect also resembling the preceding pouch (see clear area in Fig. 7).


The Pharyngeal Pouches in the Mammalia 221

In the 9.7 mm. cat (No. 466, Harvard seiies, Figs, 61, 62) the ventral diverticulum (Y.D.3) of the third pharyngeal pouch is slightly more elongated at its ventro-internal angle. In other respects it is essentially similar to the stage last described.

The pig of 10 mm. (No. 401, Harvard series, Figs. 4-8) shows clearly the initial steps in the formation of the thymus duct. The ventrointernal angle of the pouch is now clearly elongated in a ventro-mesial direction and ends in an acute angle which is wedged in the angle between the roots of the carotid and aortic arches (Fig, 7). The downgrowths of the two sides are not quite symmetrical, the right being slightly larger and ending in a more acute angle than that on the opposite side. Laterally the lower part of the pouch has separated from the ectoderm, leaving only its dorsal half in contact with the latter. Below the point of contact the lateral border turns obliquely inwards and downwards and is continued into the ventro-lateral edge of the thymus downgrowth.

On the dorsal side the peripheral portion of the pouch projects slightly above the upper end of the verschlussmemhran as a dorsal diverticulum (Figs. 4-8, D.A.3).

The carotid gland (Figs. 7-8, C.Gl.) lies closely adpressed against the anterior wall of the pouch and on its dorsal side projects slightly above the upper margin of the latter (Fig. 8). We shall defer further consideration of the gland until later.

In the 12 mm. pig (No. 518, Harvard series. Figs. 9-11, 13) the ventral diverticulum (V.D.3) is further elongated. In addition to the ventro-internal direction which it took in the earlier stage, it now shows a pronounced anterior trend, an effect of the ventral rotation of the adjacent part of the pharynx. At this time it has acquired a distinct tubular form.

The union of the pouch to the ectoderm is limited to a short stretch immediately below its dorsal apex. The rest of its lateral border is free and is continued ventrally into the outer edge of the tubular downgrowth.

The part connecting the pouch with the pharynx is considerably more constricted than in the preceding stage.

The tubular downgrowth is still further elongated in a 14 mm. pig (No. 65, Harvard collection. Figs. 14-16, Thy.). Its blind ventral extremity lies a slight distance below the level of the pericardio-cervical


222 Henry Fox

groove and thus occupies the upper part of the pericardial region. It retains its tubular form, but shows a differentiation into two portions — a terminal, swollen portion and an intermediate, relatively narrow canal, which connects the former with the remaining dorsal body of the pouch. The latter now forms a compressed epithelial plate, bearing upon its anterior surface the voluminous carotid gland. It is still connected with the pharynx by a narrow connective.

The dorsal apex of the pouch has disappeared. The peripheral portion of the pouch is attached for a short distance to the anterior wall of the fundus prsecervicalis (F.Pc).

A lumen is present in the tubular downgrowth and in the pharyngeal connective, but has disappeared into the dorso-peripheral part by approximation of its anterior and posterior walls. The latter thus assumes the form of a vertical plate, connected with the fundus prsecervicalis at its dorso-lateral angle and continuous with the intermediate cervical canal of the thymus at its ventro-internal angle.

In the 17 mm. pig (No. 51, Harvard collection. Figs. 19-21) the thymus downgrowth (Thy.) has still further elongated and now shows clearly its segmentation into three portions. These include (1) the considerably swollen thoracic vesicle, (3) the intermediate cervical connective (CV.C), and (3) the dorsal plate (D.Pl.), with which is closely associated the carotid gland (C.Gl.) and fundus prscervicalis (F.Pc).

Only a few fragments of the original lumen now remain, mostly confined to the thoracic vesicle.

Connection with the pharynx is still maintained by an extremely thin, solid connective (Fig. 45, of an 18 mm. pig.).

The dorsal body of the pouch, as already noticed, has become reduced to a flattened plate and now constitutes a part of thymus. It has separated from the ectoderm of the surface of the body, but remains attached to that of the fundus prascervicalis, which constricts from the superficial ectoderm and accompanies the thymus as the latter passively recedes from the exterior. The detailed account of this process will be reserved for later treatment. It suffices at this time to state that this ectodermal structure remains in close connection with the dorsal extremity of the thymus for a considerable period.

In an 18 mm. pig the thymus shows no essential deviations from that in the preceding stage. It is slightly longer and has a more vertical course than in the latter.


The Pharyngeal Pouches in the Mammalia 223

I shall at this stage treat of the factors which appear to be operative in producing the present condition of the thj'mus. So far as I have been able to determine, the modifications which the third pouch has undergone are referable, in large part at least, to the operation of purely mechanical factors. The pouch itself shows but little power of active growth. Compared with the surrounding parts, it is relatively passive. The first important factor is found in the unequal rate of growth of the neck and of the ventral tubular process. The former carries the roots of the large arteries from their original position immediately under the pharyngeal pouch region to their definitive position in the upper part of the thorax. The ventral extremity of the thymus, it will be recalled, is from the start in close relation with the bases of the carotid and aortic arteries, and as the latter become displaced to successively lower levels, this portion of the thymus is carried down with them. The elongation of the neck takes place, however, at a more rapid rate than that of the thymus, and thus produces on the latter a tension tending to carry it downwards. This, however, is prevented by the fact that the dorsal extremity of the thymus is, during the same period, relatively fixed in position by its attachment to the skin and pharynx. The result of such conditions would be to produce a strong tension in the intermediate connecting portion, which in consequence would become drawn out in the form of a thin cord, similar to that shown in the present stage (Fig. 21, Cv.C). This view is supported by the fact that the diameter of the cord is actually, and not merely relatively, much less than in preceding stages.

As already mentioned, the dorsal extremity of the thymus is relatively fixed in position. This condition is readily explained when it is recalled that this portion is attached to both the ectoderm and the pharynx and also bears the voluminous carotid gland, the mere bulk of which alone would hinder any ready displacement, even through such a plastic medium as the surrounding mesenchyme.

An additional factor, first pointed out by Kastschenko, is found in the behavior of the hypoglossal nerve. The latter underlies the fundus prsecervicalis. As this structure in harmony with the attached thymus becomes displaced in the ventral direction, it comes in contact with the nerve, which is also carried downwards with it until it forms a strong angular flexure, over which the outer part of the fundus curves like a hook over a cord. The elastic reaction of the nerve would naturally form a strong hindrance to any ready ventral movement of the head of


224 Henry Fox

the thymus. That a decided tension of the kind indicated is actually produced is evidenced by the later behavior of the nerve — a subject which I shall consider more fully when considering the modifications of the fundus prgecervicalis.

In a cat of 10.7 mm. the thymus has approximately the same characteristics as in the pig just considered. It shows, however, no clear lumen in any part.

In a pig of 20 mm. (No. 543, Harvard collection, Figs. 23-25) the thymus (Thy.) is considerably longer than hitherto and its ventral extremity is slightly lobed. Its lumen has entirely disappeared, owing to the thickening of its walls. The whole organ is thus composed of small-celled epithelioid tissue, which in every respect l)ears a close resemblance to ordinary lymphoid tissue. The dorsal plate is closely wedged in between the carotid gland and the fundus prgecervicalis (Fig. 23). The th3'mus on the left side has completely separated from the pharynx, while that on the right is still connected with it by an extremely thin cord (Fig. 28).

The separation from the pharynx is probably connected with the inward flexure of the sinus piriformis and the consequent decrease in the lateral diameter of the pharynx. The dorsal extremity of the thymus and the attached carotid gland, however, retain their original position, with the result that the connecting cord is drawn out to an exceedingly thin strand, which subsequently constricts.

In the cat of 15 mm. (Fig. 63) the thymus shows no special features. It closely resembles the same structure in the pig just considered, but is located at a relatively lower level. In the pig its dorsal extremity is opposite the sinus piriformis, while in the cat it lies some distance below it.

In the 24 mm. pig (No. 64, Harvard collection. Figs. 29-31) the thymus on each side is completely separated from the pharynx. Its terminal thoracic portion has grown considerably l^elow the level of the thjrroid. It is more swollen than hitherto and its distal portion is subdivided into a considerable number of convolutions (Fig. 31, Thy.).

The intermediate cervical connective persists as an exceedingly thin solid cord (Cv.C).

The dorsal plate is so intimately fused with the carotid gland and fundus prgecervicalis (F.Pc.V.) as to be distinguishable from them only with difficulty.


The Pliaiyngeal Pouches in the Mammalia 225

The cat of 23.1 mm. (jSTo. 466, Harvard collection) shows the thymus entirely free from the carotid gland, that on the left side being separated from the latter b}^ a considerable interval. At its thoracic portion the organ is subdivided into numerous lobules.

In the 33 mm.. pig (No. 74, Harvard collection) only the dorsal end of the thymus was specially examined. It still forms a flattened plate associated with the carotid gland and the lobules of the fundus prsecervicalis. It is connected by the cervical connective Avith the now large jind much lobed thoracic thymus.

In the 33 mm. cat (JSTo. 500, Harvard collection. Fig. 68) the thymus is entirely unconnected with the carotid gland. The dorsal extremity of the cervical cord (Cv.C.) is placed immediately outside of the ventral edge of the lateral wing of the thyroid. The cord is somewhat convoluted. It is continuous at its ventral extremity with the thoracic thymus, which is much enlarged and sul^divided into numerous lobules. The latter are solid and are composed of small-celled epithelial tissue. I could observe no indications of the formation of true lymphoid tissue in it.

In a rabbit of 20 days the thoracic thymus is very large and on each side it is subdivided into numerous lobules similar to those seen in the cat last described. It is connected by the cervical cord with the carotid gland, which is located close to the dorsal edge of the lateral wing of the thyroid.

(c") The Origin and Structure of the Carotid Gland.

The carotid gland first appears in a 9 mm. pig (M^ of my series, Fig. 36) as a mass of intertwined solid vesicles (C.GL), representing a series of folds of the anterior wall of the third pharyngeal pouch. The gland is therefore a purely epithelial structure of endodermal origin. The carotid artery (Car.) lies immediately in front of it and shows a small branch extending back toward the gland.

In the 10 mm. pig (Figs. 7-8, C.Gl.) the carotid gland is somewhat larger. As in the preceding stage, it is continuous with the peripheral half of the anterior wall of the pouch and even projects a slight distance beyond its lateral margin (Fig. 8), It here meets the ectoderm of the anterior wall of the third pharyngeal groove. Dorsally it projects a short distance above the upper margin of the pouch and partly curves backwards over it. The carotid artery (Car.) gives off a small branch, which, on reaching the gland, divides into numerous capillaries which form a rich network interpenetrating it in all directions. Yen


226 Henry Fox

trally they unite into a common vessel which opens into the inferior jugular vein.

In the 12 mm. and 13 mm. pigs the carotid gland forms a mare regularly circumscribed, voluminous mass (Figs. 9-10, C.GL). It retains the same topographic relation to the third pouch as before, but has increased considerably in size (Fig. 13). Its follicular structure is clearly shown, and reminds one of that of the liver. Like the latter, it consists of a reticulum of numerous, closely intertwined follicles interpenetrated by a rich system of capillaries, the latter derived from the carotid artery (Fig. 53).

The gland and the associated solid wall of the pouch (= dorsal plate of the th5'mus) are intimately connected with each other (Figs. 38-40). In many specimens, owing to unsuitable staining, the two parts cannot be clearly distinguished from each other. In those appropriately double stained with haematoxylin and Bordeaux red or with alum-cochineal and orange G the definitive reticular structure of the gland is clearly shown.

In the 14 mm. pig (Figs. 14-15) the carotid gland shows no specially noteworthy features. On the dorsal side it projects considerably above the upper edge of the thymus and there comes in contact with the inner part of the fimdus praecervicalis (F.Pc).

In the 17 mm. pig (Figs. 19-20) the carotid gland appears to difffer only in size from that just described. The same remark applies also to an 18 mm. animal (Figs. 45-46).

In the 20 mm. pig (Figs. 23-24) the carotid gland forms a moderately large ovoid organ lying to the outside of the sinus piriformis (S.Pi.). It is distinctly follicular in structure, but its individual follicles show no lumen.

In cats of 10.7 and 15 mm. the carotid gland is similar in essential respects to that in the pig last described. It appears to be less compact than the latter, the interspaces between the follicles being relatively larger (Fig. 63, CGI.).

In the pig of 24 mm. (Figs. 29-31) the carotid gland (C'.Gl.) shows no peculiar characteristics.

In the cat of 23 mm. (Fig. 66) the carotid gland shows no special features. It resembles essentially that in a 15 mm. example.

In the pig of 32 mm. (Figs. 34-35, C.Gl.) the carotid gland has the same characteristics as hitherto, but is closely invested by the lobules of the proliferated fundus praecervicalis.


The Pharyngeal Pouches in the Mammalia 227

The cat of 31 mm. (Fig. 68, CGI.) shows the carotid gland as an ovoid body, located near the antero-dorsal angle of the lateral wing of the thyroid. It shows its follicular structure very clearly. It has lost its connection with the cervical cord of the thymus, the latter terminating opposite the ventral border of the thyroid.

In a 20-day rabbit (No. 172, Harvard series) the carotid gland shows its typical form and structure. It lies outside of the antero-dorsal angle of the lateral wing of the thyroid and is connected by the cervical cord with the thoracic thymus.

In the 21-day rabbit the gland occupies a depression in the outer surface of the lateral wing of the thyroid. It is now entirely unconnected with the cervical cord, the latter lying at a much lower level.

(c"') The Sinus Prascervicalis and its Eelation to the Thymus.

It will be recalled that in a 6.2 mm. cat (Figs. 59-60) the sinus prfficervicalis (S.Pc.) forms an approximately funnel-shaped depression. The inner posterior portion corresponding to the stem of the funnel forms a deep pit, which extends diagonally inwards and backwards and terminates in a sharp edge, which at its ventral extremity is in contact with the lateral process of the fourth pouch. The outer relatively wide vestibule forms an approximately triangular depression, surrounded on all sides by the overhanging prominences of the adjacent parts — anteriorly by the hyoid arch, posteriorly by the anterior cervical region and ventrally by the pericardial prominence. Dorsally it becomes shallow and there blends with the side of the head without any perceptible break. At its ventro-anterior angle it is continued into the pericardio-cervical groove (Fig. 60). The inner wall of the sinus is formed by two low prominences representing the third and fourth pharyngeal arches.

In the 9.7 mJii. cat the sinus prsecervicalis (Fig. 61, S.Pc.) is deeper and narrower than in the stage just described — a difference due to the continued outgrowth and approximation of the adjacent hyoid and anterior cervical regions. That part of the third arch which immediately adjoins the hyoid region is rotated outwards with the latter and thus comes to face obliquely baclcwards. In this way the outer opening of the deeper part comes to lie on a level with the third phar5rngeal groove (Ph.G.3). We shall henceforth designate this deeper part of the sinus the fundus prgecervicalis (F.Pc), a term applied to it by Kastschenko.

The opening of the fundus is triangular, narrow above, wider below (see right side of figure). At its dorsal apex the third pharyngeal arch


228 Henry Fox

and anterior cervical prominence converge. The fourth pharyngeal arch^ which is located at the inner extremity of the fundus, is thus j)artly hidden in lateral view.

The fundus has much the same features as hitherto. Its inner extremity, while still connected at one point with the fourth pouch (Fig. 62, Ph. P. 4), is prolonged above the latter and forms a slight dilatation situated a short distance back of the dorsal edge of the third pouch (Fig. 62, V.Pc). This free, dilated portion is evidently to be. compared with the vesicula prsecervicalis (= vesicula thymicus of Kastschenko), which will be more fully considered in the descriptions of the

pig The 10 mm. pig (Figs. 5-8) shows the sinus in a condition somewhat intermediate between the two last described. On the right side the fundus and the fourth pharyngeal pouch are still connected with each other, but on the opposite side they are entirely separate and are removed from each other by a considerable interval, which is largely occupied by the aortic arch proper (Ao.). The large size of the latter suggests that it may have been an active agent in effecting the separation of the pouch from the fundus. Thus, on the side where the two structures are still connected, the ventral side of the artery is closely pressed against the connecting part and is evidently exerting a pressure upon it which would tend to effect its separation. As we shall see, this soon takes place.

The inner extremity of the fundus is formed by a narrow, obliquely vertical groove — the fourth pharyngeal groove (Fig. 8, Ph. G.4). Where the fourth pouch retains its connection, it is confined to the ventrointernal angle of the fundus. From this point the remainder of the groove ascends diagonally forwards and at its dorsal extremity meets the corresponding part of the third groove (Fig. 7, Ph.G.3). There is thus included between the two a triangular convexity which represents the fourth pharyngeal arch. The third groove in its entire extent is connected with the underlying pouch. The latter is not joined to the bottom of the groove, but to its anterior wall, the lateral margin of the pouch reaching some distance beyond the deep part of the groove (Figs. 7-8).

On the side where it is no longer joined to the fourth pouch the blind, inner end of the fundus is turned obliquely backwards and comes into close relation with the inferior ganglion of the vagus. This part we shall hereafter designate the vesicula prascervicalis. In my estima


The Pharyngeal Pouches in the Mammalia 229

tion, this is a more appropriate name for it than the term vesicula thymica, applied by Kastschenko.

In a 13 mm. pig, in consequence of the passive behavior of the sinus and the continued outgrowth of the surrounding parts, the sinus is still deeper and its margins are so near each other that, with the exception of its most external part, the entire sinus may be considered as included in the fundus (Pig. 13). The external opening of the latter is now relatively small. On its dorsal side the adjacent borders of the third pharyngeal arch and anterior cervical region have fused, leaving only a faint groove to mark the earlier extension of the sinus in that direction. On its ventral side the part of the head underlying the sinus has grown out and has united with the ventral extremity of the third arch, at the same time obliterating the groove earlier connecting the sinus with the pericardio-cervical fissure.

The outer limit of the fundus is approximately formed by the middle of the third pharyngeal arch. Prom this point it extends inwards and slightly backwards as a deep pocket, the blind inner extremity of which — the vesicula prsecervicalis — terminates close to the inferior ganglion of the vagus. This part is formed by the third and fourth pharyngeal grooves and the intermediate fourth arch. In consequence of the diminished depth of the fourth groove, the fourth arch does not form as prominent a convexity as in the preceding stage.

In the 14 mm. pig (Pigs. 14-16 and 18) all that remains of the sinus prfficervicalis has become, by virture of its passive deepening and constriction, included in the fundus prsecervicalis (F.Pc), which now forms a deep, blind pocket opening to the exterior by a much reduced opening placed immediately under the posterior rim of the hyoid arch. The outer half of the fundus forms a relatively narrow duct — the ductus prsecervicalis of Kastschenko — leading to the external opening (D.Pc). The inner half is relatively wider and at its mesial end is continued into the vesicula (V.Pc). In this part the earlier prominence of the fourth arch has flattened out and thus the third and fourth grooves cease to be longer distinguishable. In this way the inner end of the fundus assumes the form of a bulb.

In the next stage, i.e., a 17 mm. pig, the narrow ductus forms a solid cord, which has just severed its connection with the external ectoderm (Figs. 20-21). The inner portion of the fundus now forms a relatively broad, flattened band, which is slightly concave on its posterior side (Pig. 21, P.Pc). By its anterior wall it is closely connected with the


230 Henry Fox

carotid gland and the epithelial plate now forming the dorsal end of the thymus, but representing originally the peripheral body of the third pouch (D.Pl.). With the exception of its vesicula, the fundus is without a distinct lumen. The vesicula originates at its vent ro -internal angle, then bends backwards and terminates, as before, in the anterior part of the ganglion of the vagus.

In the 20 mm. pig the ductus has shrunken to a mere remnant (Fig. 28, f. pc). The remainder of the fundus prgecervicalis, excepting the vesicula (v. pc), forms a flattened band, which is closely wedged in between the lateral surface of the carotid gland (c. gl.) and the hypoglossal nerve (xii), over which its free end curves after the manner of a hook. At the lower posterior side of the carotid gland it expands to form the vesicula, which retains the same relation to the ganglion of the vagus as hitherto. The dorsal extremity of the thymus is wedged in between this part of the fundus and the carotid gland (Fig. 45),

It will be noticed that the fundus now has an ascending course. Beginning at its vesicular extremit}^, it extends diagonally upwards and outwards over the lateral surface of the carotid gland to the upper side of the hypoglossal nerve, over which it curves like a hook. This condition, as I shall attempt to show presently, is to be correlated with the changes in position of the nerve mentioned.

The condition of the fundus prsecervicalis is essentially similar in a cat of 10.7 mm. The left ductus, however, is considerably longer than that on the opposite side.

In a cat of 15 mm. the ductus has disappeared. The rest of the fundus (Fig. 63, F.Pc.) is coiled over the top and back of the carotid gland. It shows a distinct vesicle, located on the posterior surface of the gland.

The pig of 24 mm. shows the fundus as a much coiled, mostly solid, band. A slight lumen persists in the vesicula prsecervicalis, which Is now located on the postero-inferior surface of the carotid gland, some distance in front of the ganglion of the vagus. From this part the fundus curves upwards over the outer surface of the gland as an exceedingly thin, flattened ribbon, which at its dorsal extremity expands slightly into the hook-shaped process, which, as before, is curved outwards over the hypoglossal nerve (Fig. 31, F.Pc. and F.Pc.V.).

In a pig of 25 mm. the condition of the fundus is very similar to that just described. The part of it adjoining the vesicula is somewhat broader and is sub-divided into a number of small lobules.


The Pharyngeal Pouches in the Mammalia 231

In the 33 mm. pig (Pigs. 34-35) the fundus shows a remarkable increase in size and now forms a prominent, irregularly lobed mass appendaged to the carotid gland (F. Pc). That portion (P. Pc. V.) adjoining the vesicula has subdivided into several lobules, from which the original vesicula itself is not clearly distinguishable. From this part the thin band curves dorsally around the outer side of the carotid gland, and at its upper extremity is continued into the hook-shaped process, which has undergone a remarkable proliferation into a relatively immense, much convoluted mass (F. Pc). The hypoglossal nerve (XII) partly divides it on the left side (Fig. 34) into two unequal lobes — a large outer and a smaller internal. On the right side (Fig. 35) the division is complete, the outer being separated from the inner by a thin plate of connective tissue. The outer, free portion represents the so-called thymus superficialis of Kastchenko. (Fig. 35; Th.S.).

This division of fundus is evidently — as indeed Kastchenko first pointed out — a result of the pressure produced by the hjrpoglossal nerve. On the left side (Fig. 34), where the two divisions of the prgecervical mass are united, this nerve lies in the constriction between them immediately under the connecting cord. Where, as on the right side (Fig. 35), the division is complete, the nerve lies entirely above the prgecervical mass. These relations naturally suggest that the displacement of the nerve itself has been the active cause in producing the present condition of the structure. This view is further supported by the relations between the two structures as observed in earlier stages. In pigs of 10, 13 and 14 mm, the nerve, after descending behind the sinus prgecervicalis, curves forward some distance below its ventral margin (Fig. 40). In the 17 mm. pig it lies immediately under the fundus close to where the latter meets the carotid gland (Figs. 45-46), It has thus assumed, relative to the fundus, a more dorsal position — a change associated with the ventral displacement of the thymus owing to the elongation of the neck. Later, as these alterations in position continue, the nerve produces an upward pressure on the fundus, and thus causes it to assume an ascending course with its peripheral free extremity hanging loosely, like a hook, over the nerve. This is the condition observed in a 35 mm. pig (Fig. 31). This portion of the fundus then undergoes a rapid proliferation, perhaps an indirect result of the mechanical irritation produced by the nerve, and thus attains the


232 Henry Fox

form characteristic of the present stage, when the prjEcervical mass is being finally constricted into two separate parts.

As this is the latest stage in the series of pig embryos which I have examined, I can state nothing as to the future history or fate of the praecervical body in this animal. It .seems probable from its large size that it would be present at birth. Kastschenko observed it in an 80 mm. pig. Prenant claims that it is present at birth in the sheep.

In cats of 23.1 mm. and 31 mm. I was imable to find any certain traces of a fundus prscervicalis. This fact indicates that in this animal the later behavior of the organ must be less complicated than in the pig. It is probable that it rapidly disappears.^

In two late stages of the rabbit I could find only extremely uncertain traces of the fundus. In a twenty days' foetus a slight process is present at the dorsal apex of the left carotid gland. In a twenty-one days' example this has apparently disappeared, but some distance above the gland and entirely disconnected from it is a minute lymphoid body. This may possibly represent the transformed process seen in the twenty days' individual. However, this point cannot be settled until additional material is examined.

Review and Comparisons.

The third pharyngeal pouch when typically developed closely resembles the second. Like the latter, it has a prominent ventral diverticulum.

The pouch becomes transformed into the thymus. The greater part of the latter, i. e., its thoracic portion and cervical cord, are formed by the downgrowth of the ventral diverticulum.

The carotid gland is a derivative of the dorsal portion of the pouch. It arises as a series of follicular outgrowths from the anterior wall of the latter.

The pouch does not separate entirely from the ectoderm, but remains attached to that of the sinus prsecervicalis. Separation from the superficial ectoderm takes place by the deepening and subsequent constriction of the sinus, which accompanies the pouch in its passive withdrawal from the surface of the body.

The connection of the pouch with the pharynx is at first formed by a wide opening. Later this is reduced to a solid cord, which subsequently constricts, thus leaving the pouch as an entirely independent body,

^Verdun asserts that it had entirely disappeared in an embryo of IG mm.


The Pharyngeal Pouches in the Mammalia 233

The dorsal body of the pouch after the separation of the attached fundus prfficervicalis from the skin forms the relatively inconspicuous dorsal extremity of the thymus. It loses all trace of a lumen and thus forms a solid epithelial plate wedged in between the carotid gland and fundus prsecervicalis.

The fully formed thymus is differentiated into three parts — a ventral, thoracic thymus, an intermediate cervical cord and a dorsal plate to which the carotid gland is attached. In the cat the connection of the thymus with the carotid gland is interrupted in the later stages of development. In the rabbit the two structures become disconnected by the twenty-first day of development.

The carotid gland is typically an ovoidal body located in the neck close to the outer side of the lateral wing of the thyroid. Structurally it is a reticulum of solid follicles, interpenetrated by a system of capillaries derived from the carotid artery.

The sinus prsecervicalis, as a result of its passive deepening by the outgrowth of surrounding parts, is transformed into a deep recess, the fundus prsecervicalis. The latter is finally cut off from the ectoderm, and, by retaining its connection with the dorsal plate of the thymus, comes to lie at a considerable distance below the surface. Its inner extremity forms for some time a vesicle, which enters into close relation with the inferior ganglion of the vagus.

In the cat the fundus apparently is early atrophied. In the pig, however, it undergoes a strong proliferation, giving rise to a prominent, irregularly convoluted mass closely associated with the carotid gland. The peripheral lobe is separated from the remainder by the constriction effected by the hypoglossal nerve. This portion represents the so-called thymus superficialis of Kastschenko.

In late stages of the rabbit all remains of the fundus prsecervicalis have largely, if not entirely, disappeared. Only very doubtful traces of it remain.

The formation of the thymus as described in this paper is in harmony with all the more recent observations. These prove that the organ is of purely endodermal origin and that all but an insignificant portion arises from the ventral diverticulum of the third pouch.

With regard to the final lymphatic transformation of the thymus, I can say little, owing to the fact that I did not have at my disposal a. sufficient number of older stages to enable me to form any decided opinion as to the process by which the change took place. In the rabbit


234 Henry Fox

in which I examined a large series of relatively late foetal stages the follicles as late as the twenty-first day maintained the same histological character, that is, they were composed of small-celled epithelial tissue. The latter, however, bears a strong resemblance to ordinary lymphoid tissue, and its persistence at this late period in this animal suggests that the fundamental tissue of the definitive thymus may be really epithelial and not lymphoid tissue. Such a view has recently been supported by Stohr. On page 9 of his article "Ueber die Thymus" (Sitzungs-Berichte der physikalisch-medicinischen Gesellschaft zu Wiirzburg) he writes, "Die Thymus ist ein epitheliales Organ von Anfang his zu Ende, so gut wie etwa sine Speicheldriise." My own observations do not cover sufficiently late stages to enable me to give any strong support to this view.

Regarding the origin and structure of the carotid gland there has been considerable diversity of opinion. Steida first described the organ and postulated its origin from the endoderm. This view was later supported by Fischelis, Kastschenko, on the other hand, makes no distinction between the carotid gland and the associated dorsal extremity of the thymus, both of which are included in his nodulus thymicus. The latter he regards as simply the much swollen dorsal end of the thymus. He apparently overlooks the vesicular structure of the gland — an oversight which is not surprising when one bears in mind that this structure is only shown in sections suitably double-stained.

Steida, while correct in his derivation of the gland from the endoderm, errs when he states that it later separates from the thymus anlage and comes into contact with the carotid artery. Kastschenko maintains that the nodulus thymicus, with which he considers the carotid gland of Steida to correspond, never separates from the thymus. He shows that the body which Steida took for the gland in later stages is of an entirely different character. It forms ein verldngerter ellipsoider Knoien, which surrounds the internal carotid at the bifurcation of the common arfery. This he maintains is merely a local thickening of the adventitia of the artery.

My observations show that the contention of Kastschenko is correct. The carotid gland does not separate from the thymus, at least not in any stages examined by the two investigators mentioned. I find the same peculiar thickening of the adventitia of the internal carotid artery as described by Kastschenko in both the pig and cat. It is particularly large and prominent in pigs of 17-24 mm. and cats of 23-31 mm. In


The Pharyngeal Pouches in the Mammalia 235

both it is located at a considerably higher level than the true carotid gland.

Piersol in his study of the rabbit does not, like Kastschenko, distinguish between the carotid gland and the dorsal extremity of the thymus. It, however, is probably present, as it is clearly distinguishable in the later stages of the same animal. Its presence in the earlier stages has been shown by Verdun.

According to Prenant, the merit of having determined that the carotid gland is a proliferation from the epithelium of the third pouch belongs to de Meuron. Prenant describes and figures correctly the histological structure of the organ. His work is based upon the sheep, but his results are in all respects in harmony with what I have observed in the pig.

In Verdun's work, "Derives branchiaux chez les vertebres superieurs," the term "la glandule hranchiale III" is applied to this organ. The term "carotid gland" he applies to the conjunctival proliferation surrounding the carotid artery at its bifurcation. There has been much confusion in the use of this term. As already mentioned, Steida applied it to both structures, though in the first part of his description, i. e., of the earlier stages, he applies it to the endodermal derivative. I have, therefore, retained it for the latter. The conjunctival swelling itself is no gland and consequently does not deserve to be called one.

The exact share taken by the ectoderm in the formation of the thymus has been a puzzling problem. His early advanced the view that the entire thymus was an ectodermal structure, but soon abandoned it. Fischelis would apparently consider it as half endodermal, half ectodermal. Kastschenko derives the bulk of the organ from the endoderm, but considers that its dorso-peripheral portion, which he designates by the term, thymus superfieialis, is of ectodermal origin.

Since the last 'investigator the part taken by the ectoderm in the formation of the thymus has been largely ignored. The prevailing opinion regards the thymus as of purely endodermal origin. M} own observations, however, corroborate the statements of Kastschenko so far as his facts are concerned, but, unlike the latter, I do not consider the so-called thymus superfieialis as of sufficient constancy or importance to warrant the application to it of this rather pretentious term or to be considered as an actual constituent of the thymus.

Kastschenko describes correctly the origin of his thymus superfieialis by the constriction of the fundus prgecervicalis.- His observations were

rhis is "Le fond du troisiime sillon ectodermique" of Verdun.


236 Henry Fox

made upon pigs, and the large size which the outer free lobe of the fundus attains in these creatures probably led him to consider it as an important part of the dorsal extremity of the thymus. Its similarity in histological structure to the thoracic thymus was another fact upon which he based his view as to its importance.

To me these facts do not warrant the ascription of the ectodermal structure under consideration to the thymus. First, as regards the prominence of the outer lobe of the fundus, it is evident from my observations on the cat that this condition is not general among mammals. Even in the pig it does not seem to be absolutely constant, since Kastschenko himself speaks of an animal of 30 mm., in which he could find no trace of a thymus superficialis. In the sheep, according to Prenant's account, it is evidently similar to that in the pig. It might be inferred from these facts that a prominent fundus prsecervicalis is limited in late foetal life to the ungulates,^ but is probably more or less early atrophied in other forms.

With regard to the histological structure of the so-called thymus superficialis, it will suffice to state that its similarity in this regard to the thymus is no greater than that which any branching epithelial mass shows. In specimens which I examined the lobuli of the fimdus prsecervicalis bore as strong a resemblance to those of the salivary glands as to the same parts in the thymus. The resemblance is therefore unimportant.

I would therefore conclude that the fundus praecervicalis is to be looked upon as an associate of the thymus, but not as an integral part of it.

Prenant evidently considers the vesicula thymica of Kastschenko as including all the derivatives of the fundus praecervicalis. He regards it apparently as of endodermal origin and hence as a part of the thymus. He says, "La tete du thymus se developpe aux depens de la 3d poche entodermique et d'un diverticule de cette poche; celui-ci, qui est sans doute identique a la vesicule thymique Kastschenko, s'enforce dans le ganglion du vague." This statement is erroneous. The vesicula th}anica of Kastschenko is not a diverticulum of the third pouch, but represents the inner blind recess of the fundus prascervicalis, which I prefer to call the vesicula prsecervicalis.

Verdum correctly considers his "le fond du troisieme sillon ectodermique" as corresponding to the "vesicule thymique" of Kastschenko.

^Verdun, however, states that it disappears entirely in an 18 mm. calf.


The Pharyngeal Pouches in the Mammalia 237

He, however, does not trace its transformations in the pig, merely stating that he had found it and its outer connective in a 19 mm. embryo.

D. THE FOURTH PHARYNGEAL AND ITS TRANSFORMATION INTO THE LATERAL THYROID AND GLANDULE THYROIDIENNE.

In a 6.2 mm. cat we noticed the division of the fourth pouch into two segments by a lateral constriction (Figs. 58-59). The more dorsal of these constitutes the body of the pouch; it projects strongly backwards and on its outer side gives off a slender process, which connects with the ectoderm of the inner extremity of the sinus preecervicalis (Fig. 60). The remaining segment forms a ventral diverticulum •(V.D.4), which at this period extends for a short distance downwards, forwards and inwards.

The 9.7 mm. cat shows nearly similar conditions. The ventral diverticulum (Fig. 61, V.D.4) has lengthened slightly and begins to assume a more tubular aspect. The dorsal extremity forms a more distinct posterior process. The lateral process still persists, and, as before, connects with the ectoderm of the fundus prsecervicalis.

In a 10 mm. pig (Student collection. Harvard) the fourth pharyngeal pouch of one side has entirely separated from the ectoderm, while on the opposite side the slender connecting process still persists (Figs. 7-8). The dorsal process (Gl.T.) projects strongly backwards. The ventral diverticulum forms a relatively short, rounded protuberance. Its ventral extremity lies immediately under the aortic arch. It bears the same relation to the latter that the corresponding part of the preceding pouch does to the carotid arch.

In a second 10 mm. pig (No. 401, Harvard series) the fourth pouch of each side has severed its connection with the ectoderm (Figs. 4-6). The ventral diverticulum (V.D.4) forms a compressed, antero-ventrally projecting sac. The lateral process has disappeared. The dorsal process (Gl.T.) retains the same characteristics as hitherto.

A 12 mm. pig is apparently exceptional in that the fourth pouch of one side retains its connection with the fundus prsecervicalis (Fig. 10). On the opposite it has essentially the same features as in the preceding stage.

In two pigs of 13 and 14 mm., respectively, the fourth pouch has been transferred to a considerably lower level by the downward flexure of the sinus piriformis. The ventral diverticulum (Figs. 14 and 20, La.T.) shows increased length and forms a tubular process, whose axis


238 Henry Fox

is more nearh' vertical than in the preceding stages. Its distal extremity is somewhat bulb-like in form (Fig. 39, la. t.), but its basal portion has become constricted to a narrow duct, opening into the sinus piriformis (Fig. 40). The dorsal portion is much reduced; it is represented only by the dorsal process, which forms a spheroidal body attached by a narrow stalk to the duct (Figs. 14, 16, Gl.T.). It contains only a slight lumen. Otherwise it forms a solid mass, whose walls are apparently thrown into series of tight folds. These produce an appearance simulating that of the carotid gland — an organ produced from the homologous part of the preceding j)ouch.

At this period the distal extremity of the ventral diverticulum is without any connection with the median tliyroid. The latter at this time is rather small and lies in the mesial plane above and between the ventral extremities of the thymus downgrowths.

In both the 17 and 18 mm. pigs the ventral diverticulum has still further elongated, largely as a result of the lengihening of its duct, which now appears as a slender, solid cord (Figs. 19, 21). A lumen is present only in the terminal vesicular part. The dorsal process (Gl.T.) presents the same appearance as before; its lumen, however, has disappeared and its structure more clearly resembles that of the carotid gland (Fig. 45, c.gl.). We shall henceforth designate it the "glandule thyroidienne" — a term applied to it by Prenant.

The median thyroid now occupies a position considerably posterior to that occupied by it ^in the preceding ststge. It has grown considerably and has assumed a horseshoe shape, owing to the outgrowth of its lateral wings, which at their outer extremities almost touch the vesicles of the ventral diverticulum.

In a cat of 10.7 mm. the ventral diverticulum has separated from the pharynx and forms a pear-shaped vesicle — ^the lateral thyroid vesicle — lying free in the mesenchyme by the side of the trachea. Its lumen is reduced to a mere slit — a result of the internal proliferation of its walls. It bears the same relation to the th}Toid as in the pig of the stage last described.

I was unable to distinguish in this individual any clear evidence of the presence of a "glandule thyroidienne."

In a 20 mm. pig the lateral thyroid vesicle shows only fragments of a lumen. Otherwise it is a solid structure and is composed of several layers of small cells of epithelial nature, but closely resembling lymphoid tissue. They have essentially the same character as the elements forming the lobules of the thvmus.


The Pharyngeal Pouches in the Mammalia 239

The median thyroid has moved backwards to a position immediately in front of the trachea (Figs. 23, 25), Its lateral wings have grown back over the outer side of the lateral thyroids, the latter thus being partly embedded on their inner sides.

An exceedingly fine duct still connects the lateral thyroid with the pharynx.

The dorsal process (Gl.T.) forms a small spheroid attached to the remains of the duct. It is clearly composed — like the carotid gland — of a close network of solid follicles, the interstices of which are traversed by a system of capillaries (Fig. 54).

A cat of 15 mm. (Pig. 65) shows each lateral thyroid embedded in a depression on the inner side of the corresponding lateral wing of the median thyroid. Its minute structure is easily distinguishable from that of the latter organ by its solid lymphoid character.

A "glandule thyroidienne" does not appear to be present.

In a pig of 24 mm. the lateral thyroid is also largely surrounded by and embedded in the lateral wing of the thyroid. Only its more dorsal portion projects above the latter (Pig. 29, La.T.).

The "glandule thyroidienne" is present, at least on the right side. I found no clear trace of it on the opposite side.

In a 25 mm. pig conditions are similar to those just described, but both glands are present. That on the left, however, is much reduced, being only about a fifth the size of that on the right. I may add that the same difference in size between the glandules of the two sides was noticed in an 18 mm. pig.

In a 23.1 mm. cat the lateral thyroid bears the same relation to the lateral wings of the thyroid as in the pig last described. I could, however, find no readily distinguishable traces of the "glandule thyroidienne."

In a 31 mm. cat such lateral thyroid is deeply embedded in a concavity on the inner surface of the corresponding wing of the thyroid. No part of it projects beyond the periphery of the latter at any point.

Two minute bodies (Pig. 68) were observed, one on each side, close to the lateral border of the oesophagus. They apparently correspond in position with the glandules, but, as I could not readily determine their minute structure in the specimen examined, I think it improbable that they represent these. They are probably independent structures, lymphatic in origin.

In a 21-day rabbit the lateral thyroid forms a solid, ovoidal mass deeply embedded in a flask-shaped depression on the inner side of the


240 Henry Fox

lateral wing of the th3^roicl. As in earlier stages, it is readily distinguishable from the latter by its different histological structure.

Review and Comparisons.

The fourth pharyngeal pouch resembles the third in producing two distinct structures, the lateral thyroid and the "glandule thyroidienne."

The lateral thyroid is formed by the elongation of the ventral diverticulum. This at first is perfectly continuous with the side of the pharynx, but the connecting part early becomes constricted, assumes the form of a solid cord and subsequently separates from the pharynx. The remaining ventral portion forms a piriform vesicle, which soon becomes solid and later by baclrward growth of the median thyroid becomes embedded in the lateral wings of the latter.

The dorsal body of the pouch early loses its connection with the ectoderm and undergoes partial atrophy. A considerable portion, however, is transformed into the "glandule thyroidienne of Prenant. In the cat I have not been able to trace the history of this structure, but Verdun, who examined a large series in this type, asserts that along with the lateral thyroid it becomes embedded in the lateral lobe of the median thyroid. In the pig it persists for a considerable period, but never forms any connection with the median thyroid.

The results of my study of the fourth pouch and its derivatives are, as a whole, corroborative of previous investigations. The main facts in its development had been ably presented by Kastschenko, although he overlooked the "glandule thyroidienne" which was described by Prenant. The latter gives a full account of the microscopic structure of the organ and distinctly asserts its homology with the carotid gland of the preceding pouch.

A^erdun derives from the fourth pouch another structure which he terms "th3^mus IV" on account of its supposed homology with the thymus of the preceding pouch. This, he states, arises as a "diverticule externe et ventral." In all the examples examined by me I have noticed nothing to suggest this structure. Verdun describes it most fully in the case of the cat, but states that it is only exceptionally, and then only slightly, developed in the rabbit. In the camel and ox he finds it doubtfully represented by certain lobules associated with the "glandule thyroidienne." In the other forms — man, mole, opussum, dog, pig, sheep — ^he gives no very convincing evidence of its presence. In view of these facts, i. e., its exceptional presence in one form and its doubtful


The Pharyngeal Pouches in the IMammalia 241

presence in certain others, it seems to be that this thymus IV cannot in the mammals have the significance Verdun attributes to it. At least, to my mind, the actual facts do not bear it out, whatever its theoretical support may be.

The lateral thyroid Verdun regards as an autonomous structure, representing the post-branchial bodies of lower vertebrates. To me it appears that this is done for the most part on purely theoretical grounds. So far as actual facts are concerned, the lateral thyroid in the mammals closely follows the thymus in its behavior, and I can see no clear and convincing reason for regarding it as other than the ventral diverticulum of the fourth pouch and therefore homodynamous with the thymus of the preceding pouch. I Ivnow of no decisive evidence against the view advocated by Verdun, but the facts adduced by him are not sufficient to establish his point. I therefore follow the usage of most writers in regarding the lateral thyroid as a part of the fourth pouch.

REFERENCES.

1. Bell, E. T. The Development of the Thymus. Amer. Jour of Anat.,

Vol. V, 1905.

2. BoEN. Ueber die Derivate der embryonalen Schlundbogen und Schliind spalten bei Saugethieren. Archlv f. mlkr. Anat, Bd. 22.

3. FiscHELis, P. Beitrage zur Kenntniss der Entwicklungsgeschichte der

Gl. Thyroidea und GI. Thymus. Archlv f. mikr. Anat., Bd. 2.5, 1885.

4. Froriep. Ueber Anlagen von Sinnesorganen am Facialis, Glossopharyn geus und Vagus u. s. w., Archiv f. Anat. u. Physiol., Anat. Hefte, 1885.

5. Flint, J. M. On the Framework of the Glandula parathyroidea. Amer.

Jour, of Anat, Vol. IV, 1904.

G. Gradenigo, G. Die embryonale Anlage der Gehorknochelchen und des tubo-tympanalen Raumes. Centralbl. f. d. med. Wiss., 1886.

7. Ham MAR. Studien iiber die Entwicklung des Vorderdarms und einiger angrenzenden Organe : — Die Entwicklung des Mittelohres und des ausseren Gehorganges. Archiv f. mikr. Anat., Bd. 59, 1902.

Das Schicksal der zweiten Schlundspalte. Zur vergleichehden Embi'yologie und Morphologie der Tonsille. Ihid., Bd. 61, 1903.

S. His, Wilhelm. Ueber die Sinus prrecervicalis und iiber die Thymusanlage. Archiv f. Anat. u. Physiol., Anat. Abth., 1886.

9. Kastschenko, N. Das Schicksal der embryonalen Schlundspalten bei Saugethieren. Archiv f. mikr. Anat., Bd. 80, 1887.


242 Henry Fox

JO. Mall, F. P. The Braucliial Clefts of the Dog. Studies from the Biological Laboratory of Johns Hopkins University, Baltimore. Vol. IT, 1888.

11. MiNOT, C. S. A Laboratory Text-book of Embryolog5\ Phila., 1903.

12. MoLDENHAUEK, W. Die Entwicklung des mittleren und des ausseren

Ohres. Morph. Jahrbuch, Bd. 3, 1877.

13. PiEKSON, G. Ueber die Entwicklung der embryonalen Schlundspalten

und Ihre Derivate bei Siiugethiereu. Zeitschr. f. wiss. Zoologie. Bd. 47, 1888.

14. Prenant. Developpement organique et histologique du thymus, de la

glande thyroide et de la glande carotidienne. La Cellule, T. 10, 1894.

15. Stohe, p. Ueber die Thymus. Sitzungsberichte der phys. med. Gesell schaft zu Wiirzburg, 1905.

16. SuDLER, M. T. The Development of the Nose and of the Pharynx in

Man. Amer. Jour, of Anat., Vol. I, 1902.

17 Urbantschirtsch. Ueber die erste Anlage des Mittelohres und des Trommelfelles. Mittheil. aus dem embryologischen Institut der k. k. Univ. in Wien, 1877.

18. YERDCisr, M. P. Sur les derives de la quatrieme poche branchiale chez

le chat. Compt. rend. Soc. Biol., 1897.

19. Verdi'a^ et Tourneux. Sur les premiers developpements de la thyroide

du thymus et des glandules parathyroidiennes chez I'homme. Jour, de I'Anat. et de la Physiol., 1897.

20. Verdun et Soule. Sur les premiers developpements de la glande

thyroide, du thymus et des glandules satellites de la thyroide chez le lapin et chez la taupe. Journ. de. I'Anat. et de la Physiol., 1897.

21. Verdun. D6riv§s branchiaux chez les vertebres superieurs. Toulouse,

1898.


EXPLANATION OF FIGURES.

Fig. 1. — Lateral view of the pharynx in a G.5 mm. pig, M^ for parts consult Fig. 58. X 60, reduced %.

Fig. 2. — Dorsal view of the pharynx in the same embryo. D. A. 1, dorsal apex of first pharyngeal pouch ; HYP., hypophysis ; Ph. P. 1-4, pharyngeal pouches ; S. P., Seessel's pocket ; S. T. P., sulcus tympanicus posterior ; S. T. T. sulcus tubo-tympanicus ; S. T. Ty., sulcus tensoris tynipani ; V. D. 1-3, ventral diverticula of the pharyngeal pouches ; P.P. 4, posterior process of the fourth pouch. X 50, reduced Vs.

Fig. 3.^ — Ventral view of pharynx and larger blood-vessels in the same embryo. HYP., hypophysis ; Ph. P. 1-4, pharyngeal pouches : V. D. 1-4, ventral


The Pharyngeal Pouches in the Mammalia 243

diverticula ; M., moutli ; Tr., trachea ; Ao. 2-5, aortic arclaes ; D. Ao., dorsal aorta ; Pul., pulmonary artery ; T. Ao., truncus arteriosus ; Tyr., thyroid ; Ch. Ty., chorda tympani. x 50, reduced Vs.

Fig. 4. — Lateral view of the pharynx in a 10 mm. pig. No. 401, Harvard medical collection. D. A. 1-3, dorsal apices of first three pouches ; Gl. T., dorsal process of the fourth pouch, the primordium of the glandule thyroidtenne ; S. M. F., submeckelian fold ; S. T. T., sulcus tubo-tympanicus ; S. T. Ty., sulcus tensoris tympani ; Ton. F., dorso-lateral region of the second pouch, later transformed into the tonsillar recess ; V. D. 1-4, ventral diverticula ; V. F., vestibular fold. X 60, reduced V^.

Fig. 5. — Dorsal view of the pharyngeal region in the same specimen. Symbols as before. X 00, reduced Vo.

Fig. 6. — Ventral view of the same region in the same embryo. Symbols as before. X 00, reduced Vo.

Fig. 7. — Ventral view of the region of the third and fourth pouches in a 10 mm. pig, students' collection. Harvard medical collection. F. Pc, fundus prfecervicalis ; Ph. G. 3, third pharyngeal groove ; 10, vagus ; other symbols as in Fig. 8. X 80, reduced %. Figure inverted.

Fig. S.^Dorsal view of the same region shown in Fig. 7. Ao., aorta ; Car., carotid ; C. Gl., carotid gland ; D. A. 3, dorsal apex of third pouch ; Gl. T., dorsal process of fourth pouch, later the glandnle thijroidicniie ; Ph. G. 4, fourth pharyngeal groove, forming the inner edge of the fundus proecervicalis ; S. Pc, sinus pi-recervicalis ; V. D. 2-3, ventral diverticula of second and fhird pouches. X 80, reducea %.

Fig. 9. — Lateral view of the pharynx in a 12 mm. pig. No. 518, Harvard medical collection. C. GL, carotid gland ; remaining symbols as in Figs. 4-6. X 60, reduced Vb Fig. 10. — Dorsal view of the pharyngeal region in the same specimen. D. Pr., dorsal prominence of the tonsillar fold ; F. Pc, fundus praecervicalis ; V. Pc, vesicula praecervicalis ;. other symbols as in preceding figures. X 60, reduced Vs Fig. 11.- — Ventral view of same parts as in Fig. 10. Symbols as in preceding figures. X 00, reduced Vs Fig. 12. — Region of the second pharyngeal pouch in a 12 mm. pig. No. 518, Harvard medical collection, viewed from below, cv., concavity on the lower surface of tonsillar fold (ton. f.) ; o., ridge connecting second and third pouches ; v. d. 1-2, ventral diverticula of first and second pouches, x 30.

Fig. 13. — Region of sinus prsecervicalis in a 12 mm. pig, students' collection, Harvard medical collection, ventral view. G. Nod., ganglion nodosum. Remaining symbols as before, x 80, reduced %.

Fig. 14. — Lateral view of the pharynx in a 14 mm. pig, No. 65, Harvard medical collection. A-L. F., alveolo-lingual ridge ; C. Gl., carotid gland ; C. V., concavity on ventro-lateral wall of tonsillar fold; D. A. 1-2, dorsal apex of


244 Henry Fox

first and second pouches ; F. Pc, fundus prsecervicalis ; Fl. P., filiform process of second pouch ; Gl. T., glandule thyroidienue ; La. T., vesicle of lateral thyroid ; P. T. R., posterior tympanal margin ; S. M. F., submeckelian fold ; S. T. T., sulcus tubo-tympanicus ; S. T. Ty., sulcus tensoris tympaui ; Thy., thymus ; Ton. F., tonsillar fold (^ sinus tonsillaris) ; Tyr., median thyroid; V. D. 2, ventral diverticulum of second pouch ; V. F., vestibular fold ; y., interval between vestibular and submeckelian fold, x ^0, reduced Vb Fig. 15. — Dorsal view of pharyngeal region in the same embryo, ectoderm shown onlj- on one side. D. Pr., dorsal prominence of tonsillar fold ; F. Pc, fundus preecervicalis ; V. Pc, vesicula prpecervicalis ; other symbols as before. X 60, reduced Vs.

Fig. 16. — Ventral view of the parts shown in Fig. 15. X 60, reduced 'A.

Fig. 17. — The region of the second pharyngeal pouch, ventral view, in a 14 mm. pig, No. 65, Harvard collection, a-1. f ., alveolo-lingual ridge ; fl. p., filiform process of second pouch ; s. pi., sinus piriformis ; s t p., sulcus tympanicus posterior. Other symbols as in Fig. 12. x 30.

Fig. 18. — The region of the sinus prsecervicalis in a 14 mm. pig, same' specimen, ventral view. D. Pc, ductus prsecervicalis ; Fl. P., ventral end of filiform process of second pouch ; F. Pc, fundus prtecervicalis ; G. Nod., ganglion nodosum ; Ph. P. 3, third pharyngeal pouch ; V. Pc, vesicula prajcervicalis. X 80, reduced %.

Fig. 19. — Lateral view of pharynx in a 17 mm. pig, No. 51, Harvard medical collection. A-L. F., alveolo4ingual fold ; C. V., concavity on ventro4ateral wall of tonsillar fold ; Cv. C, cervical cord of the thymus ; D. A. 1-2, dorsal apex of first and second pouches; D. PL, dorsal plate of thymus (the carotid gland is closely associated with this part, but in this specimen it was not sufficiently differentiated by the stain to enable its outline to be accurately traced) ; Fl. P., filiform appendix of second pouch ; F. Pc, fundus pmecervicalis ; La. T., lateral thyroid vesicle, the glandule thyroidierme projects backwards from its dorsal end ; S. M. F., submeckelian fold ; S. Pi., sinus piriformis ; S. T. T., sulcus tubo-tympanicus; Ton. F., tonsillar fold; Thy., thymus; Tyr., thyroid; V. F., vestibular fold; y., space between vestibular and submeckelian folds. X 60, reduced Vb Fig. 20. — Dorsal view of same parts shown in Fig. 19. S. T. Ty., sulcus tensoris tympaui ; S. T. P., sulcus tympanicus posterior ; D. Pr., dorsal prominence of tonsillar fold; C. Gl., carotid gland. X 60, reduced Vc Fig. 21. — Ventral view of same parts shown In Figs. 18-19. E. Au., external auditory tube; Gl. T., glandule thyroidienne. Other symbols as in Fig. 19. X 60, reduced Vc Fig. 22. — Region of the second pharyngeal pouch in a 17 mm. pig, No. 51, seen from below. Symbols as in Figs. 12 and 17. X 30.

Fig. 23. — Lateral view of pharynx in a 20 mm. pig. No. 542, Harvard medical collection. D. A. 1, dorsal apex (recessus anterior) of tympanic


The Pharj'ngeal Pouches in the Mammalia 245

poucli ; Mn. F., manubrial fossa ; S. Pi., sinus piriformis ; z., indentation between tympanic pouch and tonsillar fold. X '^j reduced Vs.

Fig. 24. — Dorsal view of pharynx in same specimen, x 60, reduced Vb Fig. 25. — Ventral view in same specimen. X ^' reduced V5.

Fig, 26. — A portion of the pharynx, including tympanic pouch and tonsillar fold in a 20 mm. pig, viewed from the posterior side. pm. f., post-manubrial ridge; ps. f., post-salpingeal fold. Other symbols as before, x 30, reduced %.

Fig. 27. — The same structure as in Fig. 26, viewed from anterior side. Symbols as in preceding figure. X 30, reduced Vs Fig. 28. — Thymus and associated parts in the same animal, front view, 12, hypoglossal nerve ; other symbols as before, x 30.

Fig. 29. — Lateral view of pharynx in a 24 mm. pig. Harvard medical collection. X 60, reduced Ve Fig. 30. — Dorsal view of the same pharynx as in Fig. 29. X 60, reduced Vb Fig. 31. — Left thymus of same animal with carotid gland and prsecervical appendix. Cv. C, cervical cord ; F. Pc, hook-shaped portion of fundus priecervicalis; F. Pc. V., ventral segment of the fundus — this is .ioined to F. Pc. by a narrow connective extending over the outer surface of the carotid gland, but hidden from view in figure by the oblique position of the latter, x ^'^' reduced %.

Fig. 32. — Lateral view of pharynx in a 32 mm. pig. Harvard medical collection. G. Ep., glosso-epiglottic fold ; Mk. F., Meckelian fossa ; Mn. F., manubrial fossa ; other symbols as before. X 60, reduced Vb Fig. 33. — Dorsal view of same structure as in Fig. 32. The naso-pharynx is represented as cut across, thereby showing its Interior. X 30, reduced %.

Fig. 34. — Left carotid gland and associated prsecervical mass and hypoglossal nerve in 32 mm. pig, anterior view. For symbols, see i'ig. 31. X 80, reduced V^.

Fig. 35. — Right carotid gland and associated parts in the same animal, posterior view. The carotid gland is represented as sectioned in the plane of the paper. The prgecervical mass has divided into two parts, internal, F. Pc, and external, Th. S. The latter is the thymus superficialis of Kastschenko. X 80, reduced %.

Fig. 36. — Part of a sagittal section of the head of a 9 mm. pig, showing origin of carotid gland. C. Gl., carotid gland ; Per., pericardial cavity. Other symbols as in preceding figures, x 5^' reduced Vs.

Fig. 37. — Part of a transverse section of a 13 mm. pig (M') showing the final separation of the first pouch from the skin. Coch., cochlea; Ch. Ty., chorda tympani ; D. A. 1, dorsal apex with the part by which the pouch is last connected with the skin; Impr., impressio cochlearis ; Ph. G. 1, the first pharyngeal groove, its undulating course causes it to be sectioned twice in the section ; V. D. 1, ventral diverticulum of first pouch. X 50, reduced %.


246 Henry Fox

Fig. 38. — Part of a transverse section of the region of the third pouch in a 13 mm. pig (ivr). d. pi., dorsal lamina of the thymus. Other symbols as in preceding figures, x 50, reduced i^.

Fig. 39. — Similar section slightly posterior to last. ph. g. 2, second pharyngeal groove. X 50, reduced I/2.

Fig. 40. — Section posterior to last, passing through fundus prtecervicalis, f. pc. ; or 3, ectodermal organ of third pouch ; pul., pulmonary artery, x 50, reduced 1'2.

Fig. 41.- — Part of a transverse section through the head of an 18 mm. pig, M^ The section passes close to the oral extremity of the tympanic pouch, al. f., alveolo-lingual fold ; G. Gas., gasserian ganglion ; s. p. c, superior petrosal nerve ; 5^ inferior maxillary division of trigeminal. X 50, reduced i/^.

Fig. 42. — Part of a similar section, a few sections behind the preceding. X 50, reduced 1/2.

Fig. 43. — Similar section, posterior to preceding, d. 1. s., dorso-lateral wall of tympanic pouch ; G. Gn., geniculate ganglion ; p. g. 1, first pharyngeal groove. X 50, reduced %.

Fig. 44. — Similar section in region of the dorsal apex of the tympanic pouch. G. Au., auditory ganglion ; hy., hyoid ; thr., thyroid cartilage ; 7, main trunk of facial, showing its two divisions, the outer being the basal portion of the chorda tympani. x 50, reduced %.

Fig. 45. — Section near hind margin of tympanic pouch. The latter lies in the upper right-hand corner, dc. t., duct of lateral thyroid ; f . pc, fundus praecervicalis — the dorsal plate of the thymus lies between this and the carotid gland, c. gl. x 50, reduced %.

Fig. 46. — Section slightly posterior to preceding, gl. t., glandule thyroidienne ; ps. f., post-salpingeal fold ; v. pc, vesicula prrecervicalis. x 50, reduced %.

Fig. 47.- — Part of a transverse section of the head of a 25 mm. pig, M^ The section passes near the front end of the tympanic pouch, v. f., vestibular groove. X 50, reduced %.

Fig. 48. — Similar section near anterior edge of submeckelian fold in same animal. Mk. f ., Meckelian fossa ; s-m. f., submeckelian fold ; vis., ventrolateral wall of tympanic pouch. X 50, reduced %

Fig. 49. — Section through middle part of the submeckelian fold in same animal, e. an., external auditory meatus. X 50, reduced %.

Fig. 50. — Section slightly posterior to the point where tympanic pouch and pharynx are connected, pig, 25 mm. The post-salpingeal fold projects from the sides of the pharynx immediately internal to the tympanic pouch, mn., manubrium of the malleus. X 50, reduced %.

Fig. 51. — Section through the posterior segment of the tympanic pouch, p. m. f., post-manubrial ridge; s. t. ty., sulcus tensoris tympani; t. s., tensor muscle ; v. m. r., ventro-mesial border. X 50. reduced ^2


The Pharyngeal Pouches in the Mammalia 247

Fig. 52. — Part of a transverse sectiou of the bead of a 36 mm. pig, showing relation of the external auditory meatus. E. Au., to the tympanic pouch, prom., promontory of the tympanic pouch; the projection below the manubrium is the post-manubrial ridge. X 50, reduced %.

Fig. 53. — Transverse section through carotid gland and adjacent part of third pouch. Pig, 13 mm., M\ x 300 circa, reduced %.

Fig. 54. — Cross section of a glandule thyroidlenne in a 20 mm. pig. Harvard medical collection, x 300 circa, reduced i/^.

Fig. 55. — Ventral view of the pharyngeal region of cat of 4.6 mm., No. 398, Harvard medical collection, showing associated blood-vessels. St., stomatodeum ; v. v., vitelline veins ; x., opening of enteron into yolk sac. X 50, reduced %.

Fig. 56. — Lateral view of the same region, x 50, reduced %.

Fig. 57. — Lateral view of the pharyngeal region in a cat. No. 413, Harvard medical collection, showing pharynx and associated aortic arches. M. Gr., median oral groove ; Ph. P. 1-3, first to third pharyngeal pouches. X 80, reduced i/^.

Fig. 58. — Lateral view of the pharynx in a 6.2 mm. cat, No. 380, Harvard medical collection. Ton. F., antero-lateral border of the second pouch, in later stages the tonsillar sinus, x 60, reduced %.

Fig. 59.^ — Dorsal view of the pharynx in the same cat, showing also the related ectodermal parts. S. Pc, sinus prnecervicalis ; F. Pc, fundus priBcervicalis. X 60, reduced 14.

Fig. 60. — Ventral view of the pharyngeal region in the same cat. X 60, reduced V-2.

Fig. 61. — Ventral view of the pharyngeal region in a 9.7 mm. cat, No. 446, Harvard medical collection. On the right the plane of section is higbe) than on the left side. SM. F.. submeckelian fold, barelj^ visible at this stage ; on the right the interior of the sinus prtecervicalis is shown ; Tub., tubercuhua impar ; V. F., vestibular fold. X 60, reduced %.

Fig. 62. — Region of the sinus prjiecervicalis in the same cat. F. Pc, fundus pi'fecervicalis ; Ph. P. 4, fourth pharyngeal pouch, only its outer tip showing on the left side ; V. D. 2-3, ventral diverticula of second and third pouches, their inferior parts cut off ; V. Pc, vesicula prtecervicalis. X 80, reduced i/^.

Fig. 63. — Lateral view of the pharynx in a 15 mm. cat, No. 436, Harvard medical collection. Mn. f., manubrial fossa ; P. S. F., post-salpingeal ridge ; S. Pi., sinus piriformis ; V. M. R., ventro-mesial border. X 60, reduced Vs Fig. 64. — Dorso-posterior view of pharynx in the same animal. The figure shows the initial separation of the tympanic pouch and tonsillar fold, p s. f., postsalpingeal ridge; s. t. p., sulcus tympanicus posterior; ton. f., tonsillar fold ; v. m. r., ventro-mesial border ; z., incision between tympanic pouch and tonsillar fold. X 30, reduced %.


248 Henry Fox

Fig. C5. — Median thyroid, lateral thyroid, la. t., carotid gland, c. gl., and prsecervical body, f. pe., In a 15 mm. cat. X 30.

Fig. 66. — Lateral view of pharynx in a 23 mm. cat, No. 4(50, Harvard medical collection. I. Tn., incissura tensoris ; Mn. F., manubrial fossa ; PI. B., postero-lateral border of the tympanic pouch ; P. M. F., post-manubrial ridge ; S. M. F., submeckelian fold ; Ton. F., tonsillar fold. X 60, reduced Vb.

Fig. 67. — Dorsal view of the pharynx in a 23 mm. cat. P. R., recessus posterior ; S. R., recessus superior, x 60, reduced V^ Fig. 68. — Lateral view of pharynx and mouth in a 32 mm. cat, Harvard medical collection ; a doubtful body of lymphoid nature, x 60, reduced Vb.

Fig. 69. — Dorsal view of pharynx in the same cat. X 60, reduced Vs.

Fig. 70. — A part of the pharynx, including tympanic pouch and tonsillar fold in a 14 days rabbit, viewed from the posterior side. Symbols as in Fig. 26. X 30.

Fig. 71. — Part of the pharynx in the region of the tympanic pouch and tonsillar fold, rabbit, 161/2 days, x 30.

Fig. 72. — Part of the pharynx in an 18 days rabbit. Harvard medical collection, i. tn., incissura tensoris; p. pi., palatine constriction; p. r., recessu* posterior ; ton. f., tonsillar fold. X 30.

Fig. 73. — The posterior portion of the pharynx of a rabbit of 21 days, Harvard medical collection, viewed from the front. The nasopharynx is represented as sectioned close to the base of the Eustachian tubes and the oral pharynx immediately in front of the tonsils. E. Au., external auditory meatus; Epi., epiglottis; G. Ep., glossi-epiglottic fold; I. Tn., incissura tensoris; Mk. F., Meckelian fossa; Mn. F., manubrial fossa; S. Pi., sinus piriformis; Ton., F., tonsillar fold; V. Ton., inferior tonsillar recess. X 30, reduced %.

ABBREVIATIONS.

A-L. F., alveolar-lingual fold. Ao. 2-5, 2d to 5th aortic arches. Car., carotid artery.

C. Gl., carotid gland. Ch. Ty., chorda tympani. Coch., cochlea.

CV., concavity on ventro-lateral surface of tonsillar fold. Cv. C, cervical cord of the thymus.

D. A. 1-3, dorsal apex of first to third pouches. D. Ao., dorsal aorta.

dct., duct of the lateral thyroid.

d. 1. s., dorso-lateral wall of the tympanic pouch.

D. Pc, ductus prsecervicalis.

D. PI., dorsal plate of the thymus.


The Pliaryngeal Pouches in the Mammalia 249

D. Pr., dorsal prominence of tonsillar fold.

E. Au., external auditory tube.

F. Pc, fundus prsecervicalis.

Fl. P., filiform appendix of 2d pouch.

F. Pc. v., ventral portion of fundus prjecervicalis.

G. Au., auditory ganglion. G. Ep., glosso-epiglottic fold. G. Gas., Gasseriau ganglion. G. Gn., geniculate ganglion.

Gl. T., glandule tliyroidienne, dorsal process of 4th pouch.

G. Nod., ganglion nodosum.

Hy., hyoid.

Hyp., hypophysis.

Impr., impressio cochlear is.

I. Tu., incissura tensoris.

La. T., lateral thyroid.

M., mouth.

M. Gr., median oral groove.

Mck., Meckel's cartilage.

Mk. F., Meckelian fossa.

Mn. F., manubrial fossa.

Ph. G. 1-4, pharyngeal grooves.

Ph. P. 1-4, pharyngeal pouches.

PI. B., postero-lateral border of the tympanic pouch.

P-M. F., post-manubrial fold.

P. PL, palatine constriction.

P. R., recessus posterior of the tympanic pouch.

Prom., promontory of the tympanic pouch.

Ps. F., post-salpingeal fold.

P. T. R., posterior tympanic margin.

Pul., pulmonary artery.

S-M. F., stibmeckelian fold.

S. P., Seessel's pocket.

S. Pc, sinus prjEcervicalis.

S. Pi., sinus piriformis.

S. R., recessus superior of the tympanic pouch.

St., stomatodeum.

S. T. P., sulcus tympanicus posterior.

S. T. T., sulcus tubo-tympanicus.

S. T. Ty., sulcus tensoris tympani.

T. Ao., truncus arteriosus.

Thr., thyroid cartilage.

Th. S., thymus superficialis.

Thy., thymus.

Ton. F., tonsillar fold of the 2d pouch.

Tr., trachea.

Tub., tuberculum impar.


250 Henry Fox

Tyr., median thyroid.

V. D. 1-4, ventral diverticula of the pharyngeal pouches.

V. F., vestibular fold of mouth.

V. L. S., ventro-lateral wall of the tympanic pouch.

V. M. R., ventro-mesial border of the tympanic pouch.

V. Pc, vesicula prrecervicalis.

V. v., vitelline veins.

X., opening of enteron into yolk-sac.

y., interval between vestibular and submeckelian folds.

z., indentation separating tympanic pouch and tonsillar fold.


PHARYNGEAL POUCHES IN THE MAMMALIA

HENRY FOX



r'-i.


Oi.:-..


-••-5..


DAJ.





Fro. 2

THE AMERICAN JOURNAL OF ANATOMY-VOL. VIM, NO. 3



PHARYNGEAL POUCHES IN THE MAMMALIA

HENRY FOX


■%W/


JiCs.__^_V


n



x\


lidS--


tout '■



~—)—vdl


-SI

^ — t-i


FIG. 12


'h


/./



Ht AMERICAN JOURNAL OF ANATOMY- -VOl . VIII, NO. 3


PHARYNGEAL POUCHES IN THE MAMMALIA

HENRY FOX


'■ -'- -^

m



.J^'




PHARYNGEAL POUCHES IN THE MAMMALIA

HENRY FOX




THE AMERICAN JOURNAL OF ANATOMY— VOL. VIII, NO. 3


PHARYNGEAL POUCHES IN THE MAMMALIA

HENRY FOX


r^



Fic.


22


THE AMERICAN JOURNAL OF ANATOMY— VOL. VIM, NO. 3


PHARYNGEAL POUCHES IN THE MAMMALIA

HENRY FOX


1/


/ /



THE AMERICAN JOURNAL OF ANATOMY-VOL. VIII, NO. 3


PHARYNGEAL POUCHES IN THE MAMMALIA

HENRY FOX



THE AMERICAN JOURNAL OF ANATOMY--VOL. VIII, NO. 3


PHARYNGEAL POUCHES IN THE MAMMALIA

HENRY FOX


—dal



Fig. 26



d.a.i


tont—


K/


flJ!


FIG. 37



FIG. 28


THE AMERICAN JOURNAL OF ANATOMY -VOL. VIII, NO. 3


PHARYNGEAL POUCHES IN THE MAMMALIA

HENRY FOX


Z'




rx.


ALF. v.i;


N

Hi


G/.r



THE AMERICAN JOURNAL OF ANATOMY— VOL. VIII, NO. 3


PHARYNGEAL POUCHES IN THE MAMMALIA HENRY FOX



FIG. 36 THE AMERICAN JOURNAL OF ANATOMY— VOL. VIM, NO. 3


PHARYNGEAL POUCHES IN THE MAMMALIA

HENRY FOX

1,1




PSi



..-41 .;■- '. \


l^»



>0 ^



efes


FIG.V


THE AMERICAN JOURNAL OF ANATOMY—VOL. VII!, NO. 3


PHARYNGEAL POUCHES IN THE MAMMALIA HENRY FOX



THE AMERICAN JOURNAL OF ANATOMY— VOL. VIM, NO. 3


PHARYNGEAL POUCHES IN THE MAMMALIA

HENRY FOX



-VD


FIG CJ


f.MF



THE AMERICAN JOURNAL OF ANATOMY— VOL. VIII, NO. 3


PHARYNGEAL POUCHES IN THE MAMMALIA

HENRY FOX



Fia 65




.9"


u



-4.Pc.


■^ -^


\(/


."' J



THE AMERICAN JOURNAL OF ANATOMY— VOL. VIM. NO. 3


PHARYNGEAL POUCHES IN THE MAMMALIA

HENRY FOX


m


tf


\ 1 r ',


Uii


Wi


Ton F ^T-T

/


s^m




DA^


' \--P-MF


/


THE AMERICAN JOURNAL OF ANATOMY--VOL. VIM, NO. 3


PHARYNGEAL POUCHES IN THE MAMMALIA HENRY FOX


to»i s.Pi.


FIG. 70



-P-^-t


smif — -J dnl-—^

hiut —


P-s.f-"



-ton.f.


■.Pi.


FIG. 71


^


ton.f.

flr/.f


FIG. 72


THE AMERICAN JGUKNAL (


-VOL. VIII, NO. 3


THE LATER DEVELOPMENT OF THE NOTOCHOED IN

MAMMALS.

LEONARD W. WILLIAMS.

Fro7n the Laboratory of Comparative Anatomy, Harvard Medical School.

The fate of the notochord in mammals has received, in recent years, very scant attention. This is well exemplified both by the briefness of the discussions of the subject in recent text-books and also in reference works, and by the contradictory statements found in different, or even in a single book.

The following shades of opinion are found in volumes which have appeared very recently:

"The notochord here" (in mammals) '^persists longer intervertebrally than intravertebrally, but it disappears entirely by the time the adult condition is reached." (Wiedersheim's Comparative Anatomy, fith German edition, 190G, p. 62, and 3d English edition, 1907, p. GO.)

"They (the intervertebral discs) are developed, like the bodies around the notochord, persisting parts of this structure forming a central core to each disc." (Piersol's Anatomy, 1907, p. 132.)

"The notochordal remains lying between each pair of vertebrae with the perichordal tissue grow and remain throughout life as the nuclei pulposi of the intervertebral discs." (Bonnet, Ent^vicklungsgesehichte. 1907, p. 381.)

"The notochord is essentially an embryonic structure in mammals, although it does not completely disappear, for traces of it are to be found throughout life in the middle of the intervertebral discs. Wlien fully developed it is a cylindrical rod composed of clear epithelium-like cells, enclosed within a special sheath of homogeneous substance. These cells, although they may become considerably enlarged and vacuolated, undergo no marked histogenetic change and take no part in the formation of any tissue of the adult. . . . Within the (intervertebral) disc the notochord is enlarged and afterwards converted in each, along with

The American Jouexal of Anatomy. — -Vol. VIII, No. 3.


253 Leonard W. Williams

the siirrountling tissue, into the nucleus pulposus." (Bryce, in Quain's Anatomy, Vol. I, Embryology, 1908, p. 49 and 252.)

'"I'lie notochord not only remains intervertebrally, but gi'ows conlinuously at that point, showing therewith the tendency (Neigung) after the loss of its sheath, to fuse with the surrounding connective tissue. (Leboucq, 1880.) . . . The nucleus pulposus or gelatinosus of the intervertebral ligament (intervertebral disc) consists in every case in the adult manmial, of such common growth of the notochord and of the tissue hdng next to it. G. Jager is indeed right when he compares, as mentioned above, the intervertebral longitudinal ligament (Liingsband) of birds with the pnlpy nucleus of the disc of mammals." (Schauinsland, in Hertwig's Handbuch d. vergl. u. exper. Entwickelungslehre der Wirbeltiere, Bd. 3, Teil 2, 190G, p. 517.) The statement referred to is as follows: "Only inconspicuous remnants of it remain finally in the interior of the intervertebral ligament; they lie there enclosed in a longitudinal band which, as the 'ligamentum suspensorium' binds together the successive vertebrae (G. Jager, 1858)."

"Finally the notochord disappears from the vertebral regions, although a canal, representing its former position, traverses each body for a considerable time, and in the intervertebral regions it persists as relatively large flat discs, forming the pulpy nuclei of the fibro-cartilages." (McMurrich, The Development of the Human Body, 1907, p. 170.)

"La corde dor sale, par exemple, constitue a elle seule tout le squelette axial chex les Chordes primitifs, tandis qu'elle disparait entierement dans les formes superieures." (L. Vialleton, Un Probleme de I'Evolution, 1908, p. 87.)

The cause of this unflattering state of our .knowledge is that the theory concerning the fate of the notochord of- mammals, which was widely accepted years ago, was not well founded and does not explain the known facts. In the three decades between 1850 and 1880 the subject aroused considerable interest. Two theories were in the field: one proposed in 1852 by Luschka, and advocated by Kolliker, H: Miiller and Lowe; the other originated by Virchow, and defended by Luschka after 1856, Eobin, Dursy, and Leboucq.

The nucleus pulposus in man was carefully described, in 1852, in the first edition of "Die Halbgelenke," by Luschka, who found that it arises from the intervertebral expansion of the notochord.

Virchow later contended that the nucleus pulposus of the new-born


Development of the Xotochord 253

child is formed by the liquefaction of the central portion of the connective tissue of the intervertebral disc. He also described a tumor found upon the inner surface of the base of the skull, and, believing it to be a growth of cartilage, named it Ecchondrosis physalifora.

Luschka, in 1856, adopting in part Virchow's theory of the origin of the nucleus pulposus, found that its characteristic tissue,- which consists of whitish clusters of vacuolated cells in a transparent gelatinous matrix, arises primarily by the liquefaction of the intervertebral notochord expansion, but that it is augmented by the liquefaction of the surrounding fibro-cartilage. He also accepted Virchow's theory of the nature of Eccliondrosis physalifora.

Two years later, 1858, Luschka published the second edition of his '"Plalbgelenke," in which he described the nucleus pulposus at various ages. He found further evidence of the close relationship of notochor- . dal tissue and cartilage. The sharp boundary line between the notochord and the intervertebral disc which occurs in the new-born child, owing to the formation of papillae of fibro-cartilage which project into the notochordal tissue, disappears during childhood. liiquef action finally leads to the obliteration of the boundary between the two tissues. In old age the nucleus pulposus becomes of dirty green or gray color. It loses its gelatinous and elastic character while cheesy masses, the products of fatty degeneration, as well as calcareous, masses, appear in it. A large number of cartilage cells with thick, strongly laminated walls are found in it. Some of these are iirunense mother cells with innumerable daughter cells. Fat vacuoles are visible in many cells.

In the same year Heinrich Miiller noted that portions of the notochord persist for a long time after birth in the base of the skull, in the odontoid process of the axis and in the coccyx. He tried to demonstrate that Eccliondrosis physalifora is produced by excessive growth of notochordal tissue.

Kolliker says in the fifth edition of his "Histology:" — "In 1858 I pointed out that the intervertebral ligament of a child of one year contains ordinarily a pear-shaped cavity which is filled with the continuously growing mass of the notochord ; and that this mass, which consists of a soft matrix and many cells with characteristic vacuoles arranged in clusters or in a network of strands, forms in the adult a large part of the nucleus pulposus in which, in certain cases, the characteristic fcetal notochordal cells can be recognized. . . . This soft mass" (of fibro-cartilage, which Kolliker regards as the peripheral


254 Leonard W. Williams

part of the nucleus pulposus) "which often bears the irregular processes, first described by Luschka, surrounds the notochordal remnants (Reste) so that the two structures interlock with one another and a sharply marked cavity, such as occurs in the child, does not exist." The original article is in the Verh. phys. med. Gesellsch., Wiirzburg, 1859, IX, pp. 193-242.

This clear and precise statement by so acknowledged a leader as Ivoliiker did not suffice to settle the question, for in the following year Eobin, describing carefully a large series of stages in the development of the notochord of mammals, maintained that the notochord undergoes mucoid degeneration and that its intervertebral expansion is gradually invaded by papillae of fibro-cartilage which, meeting at its center and expanding gradually, finally replace the degenerated notochordal tissue.

Kolliker's theory of the origin of the nucleus pulposus was again attacked in 1869. Dursy found that the nucleus pulposus is formed primarily by the liquefaction of the connective tissue of the intervertebral dis6 and that the notochord takes no part in the formation of the definitive nucleus pulposus. This view is like Luschka's second theory, except that he found that the notochord is liquefied first and later fuses with the liquefied connective tissue.

Ten years later Lowe entered the field in defense of Kolliker's theory. He agreed essentially w^ith Kolliker, but found that the entire nucleus pulposus of the adult rat is formed by the notochord. This difEerence is merely one of definition, for, as was noted above, Kolliker considered that the looser fibro-cartilage of the disc should be regarded as the peripheral portion of the nucleus pulposus, which, by this definition, is composed of the notochordal tissue and the loose fibro-cartilage of the disc.

Heiberg, 1878, says: "dann muss man zu dem Schluss kommen dass die Chorda dorsalis beim Menschen keinen Antheil an der Bildung der Pulpa des Intervertebralligamentes nimmt."

In 1880 Leboucq published the last considerable article upon the question of the fate of the mammalian notochord. He found that in the human and in other mammalian embryos, the notochordal tissue is practically destroyed long before birth. In the vertebra the notochord degenerates and is absorbed, but the intervertebral notochordal expansion is first invaded by connective tissue and is ultimately replaced by it.

This piece of work, coming after a long controversy, was widely ac cepted in spite of the fact that it contradicted practically all previous


Development of the Notochord 255

writers, for it indicated that the notochord disappears much earlier than had been believed by others.

Fric, however, in his "Handbuch der Gelenke" (1904), accepts Kolliker's work and, figuring the nucleus pulposus, describes its peculiar tissue as the remnant (Eest) of the notochord.

Turning now to Miiller's contention that Virchow's tumor, Eccliondrosis physalifora, actually arises from the notochord, not from cartilage, we find Virchow's view universally accepted until 1894, when H. Steiner, working under Eibbert, published a careful study of a case of the tumor and found that Miiller was right in believing it to be an abnormal growth of the notochord.

Eibbert, a year later, proved that the tumor can be produced experimentally in the rabbit, by puncturing the intervertebral ligament so as to allow a portion of the nucleus pulposus to escape. This tissue, lying in the connective tissue and muscle near the ligament, grows for some time and forms a characteristic Ecchondrosis pliysalifora. This knowledge led Eibbert in his "Geschwiilstlehre," 1904, to propose and use the name chordoma, instead of Virchow^s name for the tumor. Fischer and Steiner, working in Eibbert's laboratory, described a case of malign chordoma in 1906.

The generally accepted interpretation of the formation of the vertebrae in the Amniota is the theory of Eemak of the resegmentation of the vertebrae. Among recent exponents of Eemak's attractive generalization are Schultze, Schauinsland, Weiss, and Bardeen. According to this theory the vertebrae are formed by the division of the sclerotomes and by the subsequent fusion of the sclerotome halves adjacent to each intersegmental plane to form an intersegmental vertebra. The fissure of von Ebner, or the intervertebral fissure, a mid-segmental transverse diverticulum of the myocoele which in mammals, however, arises independently of the myocoele and after its disappearance, divides the sclerotome into half segments. The anterior half of each sclerotome is formed of looser tissue than the posterior half. The latter is apparently a mesenchymal condensation and has been called "the primitive vertebra" by Eemak, the "scleromere" by Bardeen. Its central portion is the "hypochordal rod" of Froriep, the "horizontal plate" of Weiss, and the "primitive disc" of Bardeen. Its lateral portions are the "vertebral arches" of Froriep, and the "costal and neural processes" of Bardeen. Weiss divides the lateral portion of the primitive vertebra into a "vertical plate" which extends backward from the outer


256 Leonard W. Williams

end of the "horizontal plate" to the intersegmental artery, and the "vertebral arch" which extends outward from that artery between the myotomes. Weiss finds that the transverse portion of the primitive vertebra forms the annulus fihrosus of the intervertebral disc; Schnltze and Bardeen, that it forms part of that structure and also the anterior portion of the vertebra. Schultze asserts that the rib belongs to the posterior half of the segment, and Bardeen that from this point of origin it moves into an intersegmental position. Weiss maintains that the vertical plate and the arch fuse with the vertebra, Bardeen that only the arch does so.

In this paper I have undertaken to trace the development of the notochord, in the pig, from the time of the appearance of its segmental waves. It was also found necessary to review the formation of the vertebrae in order to determine the exact relation between the notochordal waves and other metameric structures.

The development of the notochord in man, the rabbit, cat, dog and opossum has also been studied, but in less detail. The shape of the intervetebral notochordal expansions has been found to be sufficiently characteristic in each species to justify a brief description of the form of the notochordal enlargements.

The study has been made possible by the extensive series of marmnalian embryos in the Harvard Embryological Collection, and the series referred to by number belong to this collection.

I am greatly indebted to Prof. Charles S. Minot for many helpful suggestions for the prosecution of the work.

The Formation of the Notochordal Sheaths and of the Precartilaginous vertebrae of the pig.

The notochord of a very A'oung pig embryo (5.5 mm., H. E. C. !N"os. 915, 916, 917) is a dorso-ventrally fiattened rod with major and minor axes approximately 25 and 50 micra long. Each cross section contains about eight wedge-shaped cells whose exposed walls form a thin notochordal sheath. As was shown by Dr. Minot in 1907, the notochord and the fioor of the spinal cord of young mammalian embryos are thro^ATi into a series of segmental undulations. In the pig the crests or dorsal curves of the notochord are nearly intersegmental, for they occur very slightly in front of the transverse plane of the intersegmental vessels, the troughs being nearly mid-segmental.


Development of the Notochord 257

In the head the notochord is nearly midway between the medulla and the pharynx, but posteriorly it gradually approaches the cord until, in about the 25th segment, the two are in contact. In front of this point, the mesenchyma of the sclerotomes extends from the myotomes to the notochord as a dense mass which is interrupted at the level of the notochord by scarcely perceptible light transverse zones which connect each pair of intersegmental vessels, and by a narrow longitudinal median zone of similarly light tissue.

In order to determine accurately the density of the nuclei I have counted the nuclei in an area 24 microns square in sections 10 microns thick, or in 5,760 cubic microns. In one or two cases the series are cut at a different thickness, and it was necessary to calculate the number of nuclei in this volume from the data for another volume. The percentage of error in this calculation is less than that of the counting. In every case the number given is the mean between at least two counts in different places. At the level of the notochord there are 36 nuclei in 5,760 cubic microns in the transverse light zones, 48 in the longitudinal zone and 63 in the denser regions. Below the notochord the mesenchyma is of nearly uniform density and has 54 nuclei in 5,760 culnc microns.

Higher up, beside the spinal cord, the structures of the anterior half of the segment, the spinal ganglion and the roots of the spinal nerves, are surrounded by loose mesenchyma; and the dense tissue, like that below the spinal cord, is confined to the posterior half of the segment. The cavities of the myotomes have closed. The fissure of von Ebner, which in the Sauropsida divides the sclerotome into anterior and posterior halves, has not appeared and consequently the sclerotome is not divided into anterior and posterior portions.

Behind the point where the spinal cord and notochord first come in' contact, there is a region in which the notochord is not in contact with the sclerotomes, but lies in a small cavity. Farther back the notochord is in contact with the spinal cord, the post-anal gut, and the somites.

Slight condensations of mesenchyma become visible in embryos of 6 mm. They occur in the middle of the segments and lie just under and perhaps a trifle behind the troughs of the notochordal waves. They are produced by the extension of the intersegmental zones of loose tissue, forward and backward toward the centers of the sclerotomes, and are the intervertebral discs of Froriep or the primitive discs of Bardeen. Their position is mid-segmental in the pig as in the cow


258 Leonard W. Williams

(Froriep), not in the posterior half of the sclerotome as in man (Barcleen).

Tliese structures are more clearly defined in still later embryos, 7.S mm. (Fig. 1), and it is evident that the differentiation of the axial structures is brought about as much by the spreading apart of the mesenchymal cells as by their aggregation or condensation. There are in 5^,760 cubic microns about 100 nuclei in the disc and 29 in the light transverse zones. The enumeration of nuclei in the dense regions is too difficult to yield reliable results,^ and the anT)roximate number only can be given. Below the level of the notochord the tissue of the disc is much less dense than at the level of the notochord. The dense tissue of the discs does not extend much above the notochord, but the discs are now united by a median cord of dense tissue, the perichordal septum, which surrounds the notochord and forms a dense band below it. The rounded intersegmental zones of looser tissue are deeply constricted between the notochord above and perichordal septum below, but the two lateral portions of each seem always to be connected at least by a slender cord of loose tissue which passes under the crest of the corresponding notochordal undulation.

The notochordal sheath appears first in embryos of 7 mm. and is apparently fully formed in embryos of 7.8 mm. It is an anhistic membrane about 1 to 1.5 microns thick and it is faintly striated concentrically. The formation of this sheath, or of the inner sheath which appears later, does not affect the proper walls of the notochordal cells which can still be seen inside the sheath.

The loosening up of the axial mesenchyma reaches its maximum in embryos of 9 mm., for there are 20 nuclei in 5,760 cubic microns in the light zones and 63 in the discs (Fig. 2). In embryos of 10 mm. the mesenchyma is denser, and the loose tissue of the vertebrae has 50 nuclei in the same volume. The mesenchymal tissue above the notochord has increased largely, for in the 5.5 mm. embryo the notochord was separated from the spinal cord by a small plate of mesenchyma scarcely thicker than the notochord; it is now separated by more than thrice its diameter. The perichordal septum surrounds the notochord and the vertebral anlages. The septum, nevertheless, is incomplete, for, as stated above, a small isthmus imder the notochordal crest connects the lateral masses of the vertebral anlage.

The multiplication of the cells of the discs and of the perichordal septum has been sufficiently rapid, up to this age, to maintain approxi


Development of the Notoeliord 259

mately the same density of tissue in them despite the rapid increase in volume of this part of the embryo, but the cells of the vertebrae do not keep pace with the general growth and are consequently drawn apart.

The notochord is larger and about 15 cells occur at the periphery and 3 or 4 at the center of each transverse plane.

The fissure of von Ebner is present in the trunk. It scarcely reaches downward to the level of the notochord and it does not reach inward as far as the intersegmental arteries. Its position is such that, if it were extended downward and inward, it would divide the intervertebral disc. Bardeen finds that in man this fissure is mid-segmental and that the "primitive disc" lies in the posterior half of the sclerotome in early stages, but that the intervertebral disc is later formed upon the site of the fissure of von Ebner. In the pig, however, the fissure of von Ebner does not divide the sclerotome into anterior and posterior portions ; on the contrary, the sclerotomes fuse with one another in the median line and longitudinally, as we saw in the 5.5 mm. embryo, and, in the axial rod thus formed, appear the loose transverse zones which will form later the bodies of the vertebrte.

The dense mesenchyma of the intervertebral disc extends outward to the spinal nerves and then divides into an anterior and a posterior process. The former, the interdiscal membrane of Bardeen, extends forward on the inner side of the spinal nerve to the preceding disc. The latter, representing the "costal and neural processes" of Bardeen, extends outward and backward, and downward and upward. The upper, or "neural" process, extends upward behind the spinal ganglion and upon the inner side of - the posterior half of the myotome. The lower, or "costal" process, extends downward and outward between the divergent lower ends of the myotomes.

As Bardeen found in man, all the axial mesenchyma is as yet blastemal, and I believe that, although this tissue has greatly increased in volume, almost all visible differentiation has been effected by the separation of its cells from one another.

The apparent condensations give rise "to cartilage, perichondrium and ligaments" (Bardeen) and consequently the blastemal "scleromere" of Bardeen, which is composed of the intervertebral discs with the costal and neural processes, cannot justly be regarded as a morphological skeletal unit. In short, definite skeletal differentiation, the formation of cartilage or precartilage, has not as yet begun in the "primitive vertebrae," but is foreshadowed in the definitive vertebrge by the


260 Leonard W. Williams

loosening up of the blastemal tissue which, at least in the spine of the pig, always precedes the condensation that forms precartilage.

An extraordinary multiplication of the cells of the vertebra, disc, and neural and costal processes has begun in pig embryos of 11 mm. A further differentiation of the discs and the vertebra accompanies this new phase of growth. The tissue of the vertebrae has become precartilage, for the nuclei stain less intensely and, although the proto])lasmic network remains, it becomes attenuated and stains less readily tlian elsewhere. The cells and cytoplasm of the discs, on the other hand, continue to take stains as before.

The edges of the discs are continuous Avith a similar but less dense tissue which completely surrounds the vertebras and, in the median line, fuses with the upper and lower edges of the perichordal septum. The neural processes are possibly separated from the vertebral bodies for a time by this sheet of tissue, but long before chondrification l)egins the neural processes or arches are united to the vertebrae.

'J'he notochordal cells have lost all definite arrangement and are move or less vacuolated. They are flattened antcro-posteriorly and are closely packed together.

The number of cell divisions in the vertebra^ apparently reaches a maximum in embryos of 12 mm. (Fig. 3), and there are 54 nuclei in 5,760 cubic microns. In addition to the exceptionally large number of mitoses, one sees many elongated and dumb-bell-shaped nuclei, as well as a few pairs of top-shaped nuclei united by their points at a large acute or an obtuse angle. The three nuclei in Fig. 13, as well as a number of similar nuclei, were found in a single section of one vertebra (Section 894, H. E. C. No. 5). This embrj^o is well preserved, and similar nuclei occur in two other 12 mm. embryos (ISTos. 6 and 518), but seem to be absent from a fourth embryo (No. 7). I am inclined to believe that the rapid cell division which accompanies the transformation of blastemal tissue into precartilage is partly mitotic and partly amitotic. Eod- and dumb-bell-shaped nuclei occur in embryos of 10 and 14 mm., but they are rare and do not furnish acceptable evidence of amitosis.

The precartilage of embryos of 12 mm. has reached its maximum density. The nuclei are surrounded by small quantities of cytoplasm which forms a delicate network.

A considerable quantity of loose mesenchyma separates the vertebrae and intervertebral discs from the spinal cord. Anteriorly the base of the spinal cord has lost its segmental undulations, but posteriorly


Development of the N'otochord 261

they persist even in enibr3'os of 14 mm. Only a small part of the neural and costal processes has been transformed into precartilage. At the level of the base of the spinal cord, the neural process of the scleromere is approximately triangular in frontal section. Two acute and equal angles are directed inward and outward and its obtuse angle is directed forward. The last lies close behind the spinal nerve and is continuous with the "inter dorsal" and the "interdiscal ligaments" (of Bardeen) which lie respectively upon the outer and inner sides of the nerve. The outer angle projects between the myotomic muscles and is separated from the blastemal tissue of the anterior end of the interdorsal and interdiscal ligaments of the next segment by the ramus dorsalis of the spinal nerve. Above the ramus dorsal is there is a rounded blastemal mass which cannot be assigned to one or the other segment. The outer angle of the neural process remains blastemal for some time and seems finally to form the myoseptum. The inner angle is continuous below and anteriorly with the costal process and with the periphery of the intervertebral disc. The precartilage of the neural arch appears upon the posterior side of the inner angle of the blastemal neural process. It reaches only to the upper edge of the myotomes. The costal process is largely blastemal, but it contains the small elongated precartilage of the rib.

In embryos of 14 mm. (Fig. 4) the vertebrae are larger and are more definitely outlined. A sheet of elongated, closely-placed nuclei, formed by the extension of the interdiscal ligament and by its fusion with the perichordal septum, surrounds the vertebra and binds together the successive intervertebral discs. It represents the fibrous tissue of the perichondrium and of the dorsal and ventral common ligaments. A mass of dense blastemal tissue, which is perforated by the ramus dorsalis, extends from the neural arch to the rib.

The vacuolization of the notochord has continued and an inner sheath, which is much thicker than the outer sheath, has been formed. The inner sheath and the vacuoles of the notochord are composed of mucin or a mucin-like substance, for they are stained by mucicarmine. For convenience this substance is referred to hereafter in this paper as mucin, but I do not intend to convey the impression that its composition is even approximately known. The notochord is surrounded within its inner sheath by an apparently continuous wall which is formed by the exposed walls of its superficial cells.

The intervertebral disc begins a little later, in a pig of 14.8 mm., to differentiate into a looser central portion, with nuclei irregular both


262 Leonard W. Williams

in shape and arrangement, and an outer and larger region with nuclei which are elongated longitudinally and are united by strands of protoplasm into layers concentric with the center of the disc. The inner portion later (in 24 mm. embryos) forms the cartilage which serves, as has been shown by Schultze, Minot, and Weiss, to bind together the successive vertebrae in a continuous rod of cartilage, the chondrostyle. It should be noted in passing that the development of the cartilage of the intervertebral disc at this late period, as compared with the vertebral cartilage, is another indication that the apparent condensation, the scleromere, is not the first but the last portion of the vertebral column to be differentiated. It is undifferentiated rather than precociously differentiated tissue.

The vertebral precartilage is still further differentiated in embryos of 17 mm. Each nucleus is now surrounded by a small cell body in which are enclosed one, or more commonly two, vacuoles, each of which is nearly or quite as large as the nucleus. The cytoplasmic network has disappeared and the cells lie in a homogeneous matrix.

The perichordal septum has ceased to be recognizable as such, but its upper portion now remains as the fundament of the dorsal common vertebral ligament. It lies in a deep groove in the precartilage of the body of the vertebra. The ribs and vertebras are sharply marked off by the perichondrium from surrounding tissues.

Chondrification of the vertebrae begins before embryos are 20 mm. long. (Fig. 5.) At this time the vertebral cells have much the same character as before, but the vacuoles are less conspicuous and the cytoplasm is more granular and stains more heavily. Each cell, however, ir now separated from the matrix by a heavily stained capsule. A considerable space often separates the cell from its capsule; this, however, may be due to shrinkage.

The cartilage is now surrounded on all sides by a layer of small, closely-packed, rounded nuclei. This layer, together with the fibrous tissue surrounding the vertebra, forms the embryonic perichondrium. The central portion of the intervertebral disc, wdiich includes from one-third to one-quarter of its diameter, is now precartilaginous. The outer portion of the disc is gradually becoming more fibrous. The interdorsal ligament is now well differentiated from the blastemal portion of the ligament and from the neural arch. Its upper edge and outer portion still remain blastemal. The same is true of the upper end and outer portion of the neural process. The upper blaste


Development of the Notochord 263

mal tissue will later differentiate into the upper portion of the interdorsal ligaments and of the neural arch which now extends but to the middle of the side of the spinal cord. The outer blastemal tissue of the neural process, as noted above, seems to form the myoseptum. A column of blastemal tissue from which will be formed the transverse process of the vertebra, the tubercle of the rib, the costo-transverse ligaments, etc., extends from the rib to the neural arch.

The notochord has also undergone fundamental alteration. The cell walls, which up to this time have remained intact, are now breaking down (or are being absorbed) and the mucin from the cell vacuoles escapes. The cells are united by strands of cytoplasm and the notochordal tissue now resembles mesenchyma. A part of the mucin remains in the cytoplasmic mesh, some of it, escaping, helps to thicken the inner sheath of the notochord, and a large quantity collects within the intervertebral portion of the notochord whose sheaths are compressed slightly by the intervertebral disc. The vertebral portion of the notochord, owing to the escape of the mucin from its vacuoles into the inner sheath, or into the intervertebral portion of the notochord, is much reduced in size and is much denser than before, but the corresponding portion of its sheaths is dilated. The notochord is thus dilated intervertebrally and contracted vertebrally, but the reverse is true of its sheaths, the outer sheath being of greater diameter, and the inner sheath both of greater diameter and also of greater thickness in the vertebra. The notochordal undulations are obliterated by these changes in the notochord and vertebrae.

A brief discussion and summary of the relation of the notochordal undulations and the ribs and vertebra to the segments is desirable at this point.

Dr. Minot has shown that the segmental waves of the notochord of the pig are somewhat different from those of other mammals. The crests of the undulations are a very short distance in front of the intersegmental arteries, and are in the posterior fourth of the segment; the troughs are in the second fourth; the ascending or anterior slope in the third; and the posterior slope in the first fourth of the segment. The intervertebral disc, which is formed from the transverse portion of the primitive vertebra, is mid-segmental, and lies just behind the trough of each notochordal undulation. The edge of the blastemal intervertebral disc abuts upon the posterior edge of the spinal nerve and from this point the dense tissue of the primitive arch extends backward


264 Leonard W. Williams

and outward into the neural and costal processes which belong to the posterior part of the segment. These relations persist nntil the formation of precartilage begins, when the blastemal primitive vertebra breaks up into the intervertebral disc, the neural arches, ribs, myosepta and such diverse structures as cartilage, fibrocartilage, fibrous connective tissue, and perichondrium.

Finally, I am convinced that the delimitation of the "^primitive vertebra" is not due to its becoming differentiated before the surrounding structures, but to the more rapid differentiation of the definitive vertebrae which leaves the more slowly developing blastemal tissue between the successive vertebrae as the "primitive vertebra." These considerations suggest that the scleromere is not a morphological unit or anlage; it is rather a residual mass of undifferentiated sclerotomic tissue which later forms such diverse morphological units as the annulus fibrosus and the fibro-cartilage of the intervertebral discs, the rib, the neural arch and the myoseptum. In short, the "primitive vertebra" is a lager rather than an anlage, a store of rudiments, not a rudiment. If this is true, the conception of the resegmentation of the primitive vertebrae is without foundation, for the "primitive vertebra" is not a vertebra at all. Moreover, the evidence presented by Bardeen in support of his belief that the intervertebral disc is formed by the union of the tissue from the anterior surface of the primitive disc and from the posterior surface of the anterior half of the sclerotome, is not conclusive. He says (p. 165) : "During the period of differentiation of the scleromeres the myotomes undergo a rapid development. The median surface of each myotome gi-adually protrudes opposite the fissure of von Ebner. The dorsal and ventral processes of each scleromere are then slowly forced into the interval between the belly of the myotome to which it belongs and the one next posterior, and thus finally they come to occupy an intersegmental position." Again (p. 166), "During the development of the interdiscal membranes, the primitive discs become hollowed out on the posterior surface. A comparison of Fig. 2 with Fig. 3 demonstrates this." On page 167, "Each primitive disc has become further hollowed out at its posterior surface, owing in all probability to the conversion of its tissue into that of the area between the discs. The tissue of each segment immediately anterior to the primitive disc has become greatly thickened and the line between it and the disc indistinct." These facts are summarized on page 167, "The primitive discs become hollowed out posteriorly by a loosening up of their tissue


Development of the Xotochord 265

and strengthened anteriorly by a condensation of the tissue immediately bounding the fissure of von Ebner." The evidence for the most important point, whether or not the primitive disc is partitioned between the vertebra and the intervertebral disc, is very slight; and if the process described in the first and second quotations were to continue for some time, it would perfectly account for the fact stated in the third and fourth quotations. However, Weiss found in the white rat, and I find in the pig, that the primitive disc becomes the intervertebral disc. The fundamental mistake, I believe, in all work upon the primitive vertebrae, is the assumption that the primary sclerotomic condensations either are precartilage or are skeletal anlages. The fact is that they are neither the one nor the other. Precartilage, of the mammalian vertebrae at least, arises from such primary condensations only after a preliminary loosening up and subsequent condensation.

The Segmentation of the Notochord.

The notochord shows a most marked change in pig embryos of 24 mm. (Fig. 6). The advancing chondrification of the vertebrae is the apparent cause of a considerable expansion which, on the one hand, presses the notochord, together with the greater part of its semifluid inner sheath, from the vertebra toward the intervertebral discs; and on the other hand draws the disc away from the notochord so that a cavity is formed within the disc for the reception of the notochord. The outer sheath seems not to be broken and the inner sheath adheres to it so that, at no point, does the notochordal tissue come in contact with the cartilage of the intervertebral disc or even with the outer sheath. The intervertebral cavity is fusiform and the notochordal enlargement is irregularly diamond-shaped. The dense tissue from the vertebral portion of the notochord usually forms two slender cones whose broadened bases, opposing the bases of similar cones from the adjacent vertebrae, compress the intervertebral part of the notochordal tissue and flatten antero-posteriorly the masses of mucin within it. The two notochordal sheaths are very much compressed at the center of the vertebra, but less and less so the farther from its center. The greater part of the mucin which forms the inner sheath is forced into the intervertebral cavity; but a small part of it, and occasionally a few notochordal cells, are retained within the vertebrae. The notochordal tissue retains its syncytial character, and there begins at this time a more rapid increase of its mass, affecting alike nuclei, cytoplasm and mucin.


266 Leonard W. "Williams

The chondrostyle is now complete anteriorly, for the tissue of the inner portion of the intervertebral disc is passing in this stage into true cartilage. The outer part of the disc is more and more fibrillar and is clearly destined to form the annulus fibrosus. The edge of the disc is still attached to the head of the rib, but the formation of the articulation is indicated by a small cavity which has now appeared between the two structures.

The cartilage of the vertebrte is more advanced. A large quantity of matrix intervenes between the cell-capsules, and many cells, owing to division, have two nuclei. The notochord forms approximately .2, the cartilage .3, and the 'fibrous tissue .5 of the diameter of the intervertebral disc.

Comparatively slight changes are seen in the notochord of an embryo of 39 mm. The sjTicytial network has been enlarged both by growth and by the formation of a relatively greater number of vacuoles of mucin, and, consequently, the nuclei are farther apart. The cytoplasm of the syncytium forms in places a regular continuous boundary, but at other points the vacuoles seem al)out to escape to the exterior. The small fragments of the notochord which have been enclosed within the vertebrae are degenerating. The first step in this process is apparently indicated by the separation of the cells from one another and by a change in the C}i;oplasm and nucleus which causes the former to take Orange G. more intensely and the latter to stain more deeply with hsematoxylin. Somewhat later the nucleus becomes deeply and irregularly constricted and it is finally broken up into small pieces. The cytoplasm later is fragmented in the same manner.

Calcification of the centers of the vertebrse and of the notochordal sheaths within them begins in embryos of 32 mm. (Series 136) and has affected the greater part of the vertebral bodies in an embryo of 39 mm. Within the intervertebral disc, the outer notochordal sheath is no longer recognizable in the larger embryos, having been, in all probability, stretched to excessive thinness. The inner sheath has the same character and the same relative size as before.

Mitotic figures are very frequent in the cartilage just outside the region of calcification, and many characteristic cell-nests have been formed.

The cartilaginous portion of the intervertebral disc has increased in volume with advancing chondrification much more rapidly than its fibrous portion, and now forms about .41 of the disc, the notochord form


Development of the Xotochord 267

ing about .25 and the fibrous tissue .33. The cartilage of the disc now has the structure which was earlier (17 mm. embryos) characteristic of the cartilage of the vertebrae; that is, each cell contains one or two vacuoles as large as or larger than the nucleus. A few cell capsules contain two nuclei. The fibrous tissue has begun to assume its characteristic arrangement in alternating layers whose fibers almost form right angles with one another and are inclined at an angle of 15° or 20° to the longitudinal axis of the spine.

A great change in the shape of the notochordal enlargement has occurred in an embryo of 75 mm. (Fig. 7). It is flattened antero-posteriorly and in vertical or horizontal section it is elliptical except for slight projections forward and backward into the open ends of its sheaths. The notochord now forms .41 of the diameter of the disc, the fibrous tissue .36, and the cartilage .23. The notochordal sjiicytium is larger but retains the same genera] character. It is boimded by a clearly defined cytoplasmic layer which has fewer vacuoles and more nuclei than the central . portion of the notochordal tissue. The cells of the tibrocartilage of the disc are flattened as though by the radial pressure produced by the gro^vtll of the notochord, and the cells of the numerous cell-clusters are arranged in rows parallel to the adjacent portion of the notochordal sheath. Periosteal buds have filled the centers of the vertebrae and bone formation has begun. In a few places the calcified iiotochordal sheaths have been destroyed by the periosteal buds. This process finally destro3'S completely the vertebral portion of the notochord and hereafter the notochord is confined to the intervertebral discs.

The Formation of the ISTucLErs Pui.rosus of the xAdltlt Pig.

The flattening of the notochordal enlargement continues until, in the cervical region of an embryo of 150 nmi., it is thrice as broad as thick and forms one-half the diameter of the disc, the cartilage having shrunken to .15 and the fibrous tissue having remained of the same relative size (.36). The notochord in an embryo of 250 mm. forms .58, the cartilage only .08 and the fibrous tissue .3-4 of the disc's diameter. In short, the notochord is expanding at the expense of the fibro-cartilage, which, being attached to the cartilaginous faces of the vertebrae near their common axis, is stretched over their faces by the expanding notochordal tissue. Consequently the cartilage finally forms a thin capsule which surrounds the notochordal disc. The diameter of the mass


2GS Leonard W. Williams

of notochordal tissue is about six times as great as its thickness. The peripheral portion of the notochordal syncytium is more continuous and regular, and is also denser. The notochordal tissue (Fig. 15) contains a few very large vacuoles at its center, but elsewhere is filled with a multitude of small vacuoles.

The vast increase in nuclei which accompanies the growth of the notochord is apparently due entirely to mitotic division, for in wellpreserved material mitotic figures are abundant and there is no suggestion of amitotic division. In poorly preserved tissue mitosis cannot be recognized, and the irregularity of certain nuclei suggests amitosis, but this condition is probably due entirely to improper fixation.

In a larger embryo (250 mm.) the large central vacuoles of the notochord are apparently moving toward the periphery, and in a few places they have broken through the dense peripheral layer into the inner sheath. As they reach the surface these vacuoles often tear off portions of the dense peripheral layer which form rounded isolated masses. I find upon the lower edge of a single disc in this embryo, an interdigitation of processes of the cartilage and notochord such as Kolliker and Luschka describe in the adult man. The cartilage has constricted off a few small nodules of notochordal tissue which arc assuming the structure characteristic of adult notochordal tissue. Large vacuoles which do not consist of mucin are forming in the cells which have become visible in the syncytium.

In the half-grown pig the notochord has encroached yet farther upon the remainder of the disc and forms about .74, and the fibrous tissue and fibro-cartilage .2G of the diameter of the disc. In shape the notochordal expansion is a very thin, lenticular disc. It is still surrounded by its inner sheath of mucin, which has become more dense, and after fixation in Zenker's solution it appears very finely and somewhat irregularly fibrillar. It now stains with hsematoxylin more strongly than the fibro-cartilage. The formerly continuous peripheral sheet of dense syncytial tissue is now broken in many places by large masses of mucin, and in other regions it contains large vacuoles which seem about to escape into the inner sheath. The formation of mucin has continued until the center of the notochordal mass consists of a large quantity of mucin in which the slender syncytial network seems suspended. Between the very loose central mass of the notochord and the much broken dense peripheral layer, there is a zone which contains moderately small vacuoles in a sjmcytial mass. The mucin is gradually replacing a large


Development of the Notochord 269

portion of the syncytial tissue and, beginning at the center, is gradually coming to surround the notochordal tissue instead of being surrounded by it. The mucin vacuoles at first are imbedded in the syncytium; the vacuoles, enlarging, touch and finally unite with one another, leaving a coarse network of syncytium; the strands gradually become attenuated and finally break, so that masses of notochordal tissue, which are very small near the center and large at the periphery of the nucleus pulposus, are isolated.

The cartilaginous portion of the intervertebral disc is represented by a very thin sheet of fibro-cartilage which lines the cavity in which the notochord lies. The portion of this sheet which lies upon the calcified cartilage of the epiphysis is more fibrillar than the portion which stretches around the notochord from vertebra to vertebra. The fibrous portion of the disc is very dense, and its inner portion, like the cartilaginous portion of the disc at an earlier stage, is being pressed radially outward by the notochord and forms a capsule around the notochord.

In the adult (Fig. 8) the notochordal enlargement is relatively thicker than in the half-gi'own pig. It is also relatively smaller, for it now forms but .4 of the diameter of the disc, the fibrous tissue forming the remaining .6. The expansion of the fibrous tissue is produced by the thickening of its various layers, not by the addition of new layers. The fibro-cartilage forms, as before, a thin lining of the notochordal cavity.

The mucin of the notochord retains the same character, but the tissue has undergone a most astonishing modification. The mucin now divides the notochordal tissue into a great number of relatively large masses which in turn are divided by smaller quantities of mucin into subsidiary masses. Each of the latter consists of a number of very peculiar cells or cell-like structures (Fig. 16) which are bound together by small quantities of syncytial cytoplasm. These cells each contain two, or more rarely one or three, large vacuoles which are surrounded by thin cytoplasmic walls and are separated by a small amount of cytoplasm in which lie, in the great majority of, if not in all cells, two small nuclei. All efforts to determine the nature of these vacuoles have failed. No stains affect them. They do not shrink in absolute alcohol but they do swell in water. Each cell of material which has been fixed in formalin and immersed in water for some time, owing to the swelling of the vacuoles, becomes elongated and constricted in the middle. Cells with but one vacuole resemble fat cells, but the vacuoles are not fat.


270 Leonard W. Williams

for they are not stained by osmic acid, Sharlach E, or Sudan III. The cytoplasm of these cells contains granules of glycogen.

It is a cause of regret that I have not been able to obtain notochordal tissue from immature pigs of various ages' and from very old animals in order to follow carefully the process of formation of these cells and the ultimate modifications of the nucleus pulposus. It should be noted that although there are several points of similarity between notochordal tissue and cartilage in the pig and, as we shall find later, greater similarity in other animals, nevertheless, they are distinct tissues. The nucleus pulposus is formed entirely by the notochord.

Kolliker describes and figures clusters of cells from a child of one year which are essentially like the cell clusters of the adult pig; and Fric describes and pictures the same tissue from the nucleus pulposus of the human adult. Both men, however, consider that this notochordal tissue forms only the central part of the nucleus pulposus and that the weak fibro-cartilage of the disc is the peripheral portion of the nucleus pulposus. No transitions occur between the two tissues, and the inclusion of the fibro-cartilage in the nucleus pulposus is merely a matter of definition, not a question of fact; but unfortunately the description of the nucleus pulposus as formed of notochord and cartilage has led many to believe that it is produced by a fusion of the two tissues. The two tissues remain as distinct in the adult, despite their interlocking papillse, as in the new-born child, in which they are separated by a sharp boundary. Both Kolliker and Fric call the notochordal tissue of the nucleus pulposus a "remnant" (Rest) of the notochord. This terminology is not allowable because it makes the remnant of a part greater than the whole; for the notochordal tissue of a single disc is much greater than the entire notochordal rod of the embryo.

The Notochord in Other Mammals.

It has not been possible to follow out the development of the notochord in other mammals as carefully as in the pig, nor is it necessary, for unless there appears in the literature of the subject or in the tissues of the adult some confusion or doubtful evidence, it is permissible to assume that similar structures in mammals have a similar developmental history.

The notochord of an opossum embryo of 7.5 mm. is a slender rod without segmental undulations. Anteriorly it is surrounded by an


Development of the jSTotochord 271

anhistic sheath, and its tissue is s}Ticytial and vacuolated. At the base of the tail, the notochord is cellular, vacuolated, and the somewhat thickened exposed cell-walls form its only sheath. In the tail, vacuoles are absent, and I am unable to find either cell-walls or sheath. A mesenchymal sheath surrounds the notochord in the tail and in the posterior part of the trunk, but there is no indication of a perichordal septum.

The dense sclerotomic tissue is interrupted, at the level of the notochord, by broad intersegmental zones of looser tissue. The myocoele has closed, and I do not find the fissure of von Ebner. The lighter zones are broad anteriorly, and in the neck and thorax they have become precartilage. The intervertebral discs lie in the "third and fourth fifths of the segments, and are consequently a trifle farther back than in the pig. In an embryo of 8 mm., the centers of the discs of the anterior part of the spine have become precartilage.

The chondrostyle is well formed in an embryo of 11 mm. (Series 925), and is a cylindrical rod which encloses the notochord and bears the ribs and neural arches. Near the tip of the tail, the vertebrae are represented by broad zones of loose blastemal tissue which are separated by the denser tissue of the intervertebral discs. In the middle of the tail ..the notochordal undulations have appeared, and the crest of each undulation lies in an intervertebral disc, the trough in the precartilage of the vertebra. The center of each intervertebral disc in the sacral region is now precartilage, and the peripheral portion of each disc forms a slight thickening of the continuous perichondrium of the chondrostyle. The enlargement of the vertebrae, which is caused by their chondrification, is forcing the notochord in the lumbar region from the vertebrae into the intervertebral discs and is also both compressing the cells of the disc and carrying the lower and lateral parts of the disc away froui the notochord. This process makes the vertebrae, which before chondrification are smaller than the intervertebral discs, larger than the intervening discs. In the trunk and neck, however, the chondrification of the cartilage of the discs has caused them to become of nearly as great diameter as the vertebra. In this region, therefore, the discs are recognizable only by the notochordal enlargement, by a slight compression of the cartilage cells, and by the slight thickening of the perichondrium which represents the fibrous portion of the disc. The result of all these processes is that the spine of this embryo is represented anteriorly by the nearly cylindrical chondrostyle; in the


272 Leonard ^Y. Williams

posterior part of the trunk by the cartilaginous vertebrae and the constricted precartilaginous or blastemal intervertebral discs; and posteriorly by the constricted blastemal vertebrge and the dense blastemal discs.

The notochordal crests correspond, at first, quite closely with the intervertebral discs; but later (assuming that the posterior portion of the notochord represents an earlier, and the anterior part a later phase of identical processes) the crests appear to be a little in front of the centers of the discs. As the notochord is driven from the vertebrge, it forces its sheath downward and outward so that its lower limit in the intervertebral disc is brought down as far as the troughs of the notochordal undulations (Fig. 10). The enlargement consequently becomes irregularly fusiform, its lower surface being flat. As more tissue is forced into the disc, the enlargement bulges downward sharply at a point near the middle of the disc and somewhat behind the crest of the notochordal wave. This process continues until, in an embryo of 12 mm., the notochordal enlargements are roughly diamond-shaped (Fig. 17). The lower angle is always less acute and prominent than the upper or primary angle, which represents the crest of the notochordal wave. The notochordal sheath is not broken by the expansion of the notochord. The inner notochordal sheath appears late and is relatively thin. The accumulation of mucin within the intervertebral enlargement of the notochord is quite large. The anterior end of the notochord terminates, at a point midway between the hypophysial fossa and the foramen magnum, in a rounded knob. The cranial portion of the notochord, with the exception of the knob just mentioned, lies upon the upper surface of the cartilage, under the perichondrium, and forms a distinct ridge. The head of each rib is continuous with the intervertebral disc, and the tubercle is continuous with the neural process of the vertebra behind the disc. The transverse process (or cervical rib) of each cervical vertebra is continuous with the neural arch above and with the body of the verteljra at the base of the arch. The head of the rib seems to be displaced forward in the trunk of the opossum.

The process of differentiation of the vertebral column of the opossum seems to be identical with that in the pig, but the scantiness of the material at hand does not permit a precise determination of the conditions in the opossum. I am quite sure, however, that in both animals there should be recognized four distinct processes of vertebral differ


Develojiment of the ISTotocliord 273

entiation: the blastemal stage, in which the mesenchymal tissue is loosening up; the precartilaginous, in which a rapid multiplication of cells occurs ; the cartilaginous and the osseous stage.

In the adult opossum the nucleus pulposus forms a large part of the intervertebral disc. In the tail, it forms 63 per cent of the disc; the remainder being formed, as would be inferred from the great size of its cartilaginous portion in the embryo, almost entirely of fibro-cartilage. The notochordal tissue has much the same character as in the pig, but the cells have ordinarily but one nucleus and the vacuoles are smaller and of very variable number. The cell or tissue clusters are much larger than in the pig.

The notochord of an adult mouse forms about 62 per cent of the cervical intervertebral discs. Of the remainder, the portion below the nucleus pulposus is of about twice the diameter of the part above it. Fibro-cartilage forms about two-thirds of the former, and but onethird of the latter. The notochordal tissue forms practically a single mass which contains very little mucin, but is surrounded by a thick layer of it, the inner sheath. The cells have usually one nucleus and relatively small vacuoles.

The structure of the nucleus pulposus of the guinea-pig is quite different from that of the mouse. Its relatively small and nearly spherical cells form small strands or clusters that are suspended in a large mass of mucin. They commonly contain small vacuoles. The cartilaginous portion of the disc, as in the pig, is very small. In a few jjlaces the notochord interdigitates with the fibro-cartilage.

In the adult dog (of advanced but unknown age) the annulus fihrosus encloses a mass of soft friable tissue of yellowish color which appears, at first sight, to be homogeneous. A small rounded but irregular mass, however, forms the center of the disc. This is quite distinct, and can be lifted out whole, leaving a cavity of sharp and smooth contour. Sections of the soft center of the disc at first seem to be composed of a single tissue, but it is seen that the cell clusters at the center of the disc are of much more variable size than those of the peripheral part and also that they are of different composition. The peripheral portion of the center of the disc is clearly cartilage containing immense cell nests. Not having been able to make out fibrils in the matrix, I am inclined to believe that this is hyaline cartilage with a very soft matrix. Luschka figures and describes, in the human adult, papillge of fibro-cartilage which project into the notochordal tissue and


274 Leonard W. AVilliams

which contain similar but much smaller nests of cells. The central mass consists of a firm and apparently fibrillar matrix in which are embedded, without any regularity, clusters of cells of various size and appearance. Certain clusters contain small non-vacuolated cells which are from 10 to 14 microns in diameter and stain intensely with Orange G. The nuclei have a diameter about one-third as large as that of the cells and each contains small masses of chromatin and a nucleolus. Many cells contain nuclei, and the cells are often arranged in pairs or fours like cartilage cells. These cells are probably non-vacuolated notochordal cells. Other cell clusters are larger and are enclosed in a definite rounded cavity. Many of the cells in these clusters are of the type just described, the others contain vacuoles of various sizes. The cells with large vacuoles are similar to the notochordal cells described by Luschka, Kolliker and Fric in man, and are somewhat like those of the pig's notochord. There can be little doubt that these are also notochordal cells. The boundary of the central mass or notochord can be recognized under the microscope by a slight difference in staining property and texture between the matrix of the cartilage and of the notochord. Not having studied the formation of the intervertebral disc of the dog, I am unable to assert more than the probability that its soft center is formed of notochordal and cartilaginous tissue, and that as age advances the two tissues become more and more alike.

The morphological and physiological meaning of the segmentation of the notochord is cjuite clear. The notochord and its membranous, cartilaginous or bony sheath have been assumed to be developed in inverse ratio to one another : the former being the predominant structure in less, and the latter in more specialized forms. This is, in a general way, true, but the notochord does not degenerate in mammals. On the contrary, while it loses its primary continuity and surrenders to the vertebrae a part of its primary function, nevertheless the notochord continues to perform a part of its primary function, but in a somewhat different way and in connection with the segmented spine. The notochord is primarily surrounded by a continuous sheath of connective tissue, in which later appear isolated metameric cartilaginous elements. In mammals the cartilaginous elements unite, as we have seen, at such an early stage in development that they may scarcely be said to exist as separate units. The chondrostyle is deeply constricted near the center of each segment by the fibrous tissue of the intervertebral disc. Without these constrictions the chondrostyle would be too rigid


Development of the Notochord 275

to admit ready flexion, and with them the chondrostyle would not offer a sufficient resistance to axial stresses; hence the need of the nucleus pulposus which, being incompressible and also being closely invested by the fibro-cartilage and fibrous tissue of the disc, serves as a pad upon which the vertebrse turn. When the spine is unbent, the nucleus pulposus forms a rounded mass which is bound in on all sides by the fibrous tissue and fibro-cartilage of the disc, which, being attached to the heads of the vertebrse, forms a capsule whose layers are concentric with the nucleus pulposus. When the spine is bent forward, for example, the posterior portion of the annulus fibrosus is stretched straight, forcing the nucleus pulposus forward as a wedge-shaped mass between the inclined faces of the vertebrae, while the anterior part of the disc is pushed forward, its surfaces being drawn together, in a sharp curve or in one or more folds. Corresponding changes occur as the spine is flexed in other directions or is circumducted. The nucleus pulposus contributes largely to the strength of the spine and to its flexibility. The chondrostyle is partially replaced by bone, but its intervertebral portions persist.

The tissue of the notochord is at first cellular and epithelial. Later it becomes syncytial and resembles closely mucoid connective tissue. It finally becomes cellular a second time and then is very similar to cartilage. Kotochordal tissue is perfectly distinct from all other tissues of mammals, and passes through a very cliaracteristic cytomorphosis.

The Shape of the jSTotochordal Enlargements in Mammals.

A description of the shape of the notochordal enlargements is very unsatisfactory alike because of the difficulty of accurate description; because the shape of any particular dilation changes with growth; and because, although the shape of the enlargement in each species is remarkably characteristic, the great amount of variation which occurs renders it difficult to determine the normal type of expansion. In the opossum, as has been noted above, the crest of the notochordal undulation makes the upper contour of the expansion. From this boundary the enlargement gTows downward to the level of the troughs (Figs. 10 and 17). Later a small ventral process appears somewhat behind the crest of the enlargement and, as this process enlarges, the expansion becomes somewhat diamond-shaped, but the upper angle always remains in ad


276 Leonard W. Williams

vance of the lower. Thus^ although the shape of the notochordal enlargements is constantly changing, it is characteristic of the opossum at each stage.

In the pig, the trough of the notochordal wave lies slightly in front of the center of the intervertebral discs, and, as in the opossum, the convex wall of the notochord gives the cue to subsequent changes of form. The upper or concave wall of the notochord moves upward and, at the same time, the chondrification of the vertebra forces, or at least seems to force, the notochordal crest or summit forward until it comes to lie at the posterior edge of the disc. The notochord now makes a sharp descent from the posterior to the anterior edge of the intervertebral disc, and as it gradually expands the upper point moves forward to the center of the disc and the lower moves backward until a symmetrical diamond-shaped expansion is formed.

The notochord of the rabbit is first enlarged vertebrally, as in the pig and opossum, but to a greater extent. These dilations appear in the anterior vertebrae of an embryo of 10.5 mm. which are just passing over into precartilage. The convex side (the lower) of the notochord is usually more expanded than the upper, and there is usually formed a sharp ventral expansion of its sheath. The vertebral enlargements are larger and more symmetrical in an embryo of 12 mm., in whicli the vertebrse are precaTtilaginous. They occur throughout the trunk and they have apparently obliterated the notochordal undulations. Chondrification forces the notochord from the vertebrse into the intervertebral expansions, and in an embr3'o of 14.5 mm. (Fig. 18) the vertebral expansions have disappeared. The centers of the intervertebral discs are now loosening up in preparation for the formation of precartilage. The notochord first expands upward in the disc and later downward also. In an embryo of 18.8 mm., the enlargements are quite rounded and somewhat later they become of greater diameter than length. The upper moiety is usually somewhat larger than the lower, and somewhat in front of it (Fig. 9). The notochord is vastly more vacuolated in the embr3'o of the rabbit, and the vacuoles are more evenly distributed than in the other mammals studied.

In the cat, as in the pig, opossum and rabbit, the first notochordal expansions are vertebral. These appear in an embryo of 10.7 mm. in the anterior vertebrae which consist of very loose mesenchymal tissue; and in an embryo of 12 mm. they are larger and more numerous, the vertebrae being precartilaginous. The vertebrae are cartilaginous in an


Development of the Notochord 277

embryo of 15 mm., and the discs are precartilaginons. The intervertebral expansions (Fig. 19) are forming, but, unlike those of other mammals, the vertebral expansions persist at least until the cartilaginous vertebrae are calcified (in embryos of 39 mm.) (Fig. 11). The first indication of the intervertebral expansion is a slight angular point which appears upon the convex or upper side of the crest of the notochord at the middle or near the anterior edge of the disc. As this increases in size, a similar but smaller and usually more rounded ventral point appears somewhat behind the first. The two angles usually become nearly equal, but they retain for a long time their asymmetrical position, and the enlargements are usually more flattened antero-posteriorly than in other mammals. There is also greater variation of the shape of the enlargement than in other mammals.

It should be noted in passing that a large part of the upper edge of the annulus fibrosus of the eighth to the seventeenth discs of the cat, and of the eighth to the sixteenth in the rabbit is converted into a transverse intercostal ligament which binds together the heads of each pair of ribs. In older rabbit embryos these ligaments are less distinct than in younger embryos.

The notochord of a sheep embryo of 14.6 mm. is of uniform diameter, and the chondrification of the vertebras has just begun. In an embryo of 16.1 mm. enlargements have appeared in the first few vertebrae, but at 17 mm. the enlargements are being compressed. The chondrification of the vertebrae is far advanced in the next older embryo in the collection, 26.1 mm. (Fig. 14). The centers of the intervertebral discs are precartilaginons and a somewhat top-shaped enlargement of the notochord has formed in each at the summit of the notochordal wave. The vertebral enlargement has been divided and the parts have been driven forward and backward toward, but not into the adjacent discs. The notochordal enlargements are thus intervertebral, but each is a deeply constricted cord, consisting of the top-shaped central or intervertebral lobe and a pair of somewhat irregular and larger lobes which lie at a considerably lower level in the ends of the adjacent vertebrae.

The human notochordal expansions are of yet another type. The notochord is situated considerably below the center of the vertebral column and vertebral expansions do not occur. In an embryo of 22 mm. chondrification of the vertebrse has advanced considerably and the notochord is sharply compressed in the center of each vertebra. It is correspondingly and symmetrically dilated intervertebrally. The centers


278 Leonard W. Williams

of the discs are composed of precartilage. In an embryo of 33 mm. (Figs. 12 and 20) the chondrostvle is practically complete. The otherwise fusiform intervertebral expansion of the notochord is compressed above as by the sharp edges of the vertebras, and consequently bears dorsally a small angular process which projects into the broad but thin cartilage of the disc.

The shape of the notochordal enlargements, in the mammals which have been studied, is perfectly characteristic at each stage of their development until they are transformed into the nuclei pulposi of the intervertebral discs.


The Eelation of the Notochord to Chordoma.

The course of the notochord in the skull of the human embryo, taken in connection with the results that have been reached in the preceding part of this paper, offers some suggestions as to the origin and nature of chordomia. It will be seen in Fig. 20 that the notochord makes a single large sigmoid curve in the base of the skull and that it lies near the surface of the cartilage at four points. It is near the upper surface, in the hypopliysial fossa, a short distance behind the fossa, and near the foramen magnum and near the lower surface at a point midway between the hypophysial fossa and the foramen magnum. This is the normal course of the notochord in the skull of human embryos. Its curve is due to the fact that after the formation of the notochord, the mesenchyma, growing inward between the base of the brain and the pharynx, surrounds the anterior and posterior ends of the cranial portion of the notochord ; but, since the notochord is attached to the epithelium of the vault of the pharynx longer than elsewhere (until embryos are 9 or 10 mm. long), it collects above the central portion of the notochord. As the parachordal cartilages unite, they surround the notochord and hold it in this position. These facts were discovered by Froriep (1882), who described the later history of the notochord in the skull, and Gaupp cites them in his article upon the skull in Hertwig's "Handbuch der Entwickelungslehre."' The middle section bears a large but variable number of kinks, short branches, thickenings, and other irregularities. These I find often involve the pharyngeal epithelium, which is here thickened and often invaginated. This section of the notochord has been found by Froriep to degenerate first, but in one of the embryos in the H. E. C, Xo. 851, 22 mm., it forms small


Development of the Notochord 279

masses of tissue imbedded in the retropharyngeal connective tissue which are very similar both to the adult notochordal tissue pictured by Kolliker and Fric and to the tissue of chordoma. The posterior part of the notochord is forced upward and backward and forms a large mass upon the upper surface of the skull-base a short distance in front of the tip of the tooth of the axis. The anterior part of the notochord is forced forward and, forming a large mass between the cartilage and the perichondrium of the hypophysial fossa, persists longer than elsewhere in the skull. "We have seen that notochordal tissue, which is enclosed in a large mass of cartilage, is either forced from the cartilage or is enclosed in it and degenerates. If the tissue escapes from the cartilage it undergoes a typical cytomorphosis and forms adult notochordal tissue which is in all essentials like chordomal tissue. If this same process takes place in the skull we should expect to find notochordal tissue forced by the first chondrification of this region, that of the dorsum seller., either forward into the hypophj^sial fossa, or backward and upward upon the dorsum sellce. The chondrification of the posterior end of the parachordal plate would, under the same conditions, force the notochord backward toward the apex of the odontoid process of the axis, or forward. In the latter case the notochordal tissue would be forced either out under the skull or forward to the junction of the sphenoidal and occipital cartilages or bones. Chordoma occurs at all these points, except between the pharyngeal epithelium and the skull, and only at these points. It occurs most frequently upon the dorsum sellce, less frequently in the hypophysial fossa and at the spheno-occipital junction, and in the maligiiant case reported by Fischer and Steiner it was found upon the upper surface of the basi-occipital bone. It should be noted also that the tumors which occur at the spheno-occipital junction lie in the marrow spaces, as though the tissue of the tumor had been forced into the bone under great pressure, as would be the case if notochordal tissue were compressed by the growth of both bones.

It seems to me probable that at least the majority of chordomas are comparable to cranial nuclei pulfosi, and that chordoma should not be regarded as an abnormal growth of notochordal tissue, but merely a normal growth in an abnormal position. I am confident that chordoma also occurs beneath as well as above the spheno-occipital junction, but no such cases have been reported.


280 Leonard W. Williams

Summary.

1. The primitive vertebra of Eemak or the scleromere of Bardeen is not a morphological unit and, in the pig, it is not resegmented to form the posterior part of the intervertebral disc and the anterior part of the following vertebra. Its central part forms the annulus fibrosus and the intervertebral portion of the chondrostyle from which 'arises the fibro-cartilage of the intervertebral disc. Its lateral portions form a large variety of structures, among which may be mentioned the ribs, the neural arches (or parts of them), the costo-transverse articulations, ligaments, myosepta, and perichondrium. In short, the primitive vertebra is a mass of undifferentiated mesenchyma which is never segmented longitudinally.

2. The cartilage of the vertebra does not arise from a primary condensation of mesenchyma but from a secondary condensation which follows a loosening up of the relatively dense tissue of the scleromere. In the pig, mesenchymal tissue of nearly uniform density collects around the notochord of the trunk before the embryo is 7 mm. long. From this time until the embryo is 9 mm. long, the intersegmental or vertebral portions of this tissue become constantly looser while the midsegmental or intervertebral portions probably become slightly denser. A secondary condensation of the vertebral tissue takes place as the embryo grows from 9 to 12 mm., and at the same time there occurs a loosening up of the central part of the intervertebral disc preparatory to its secondary condensation. The secondary condensation of the tissue of the vertebrae and of the intervertebral discs produces precartilage. Chondrifieation begins when embryos are 14 to 17 mm. long,

3. The notochord expands slightly in each vertebra at the time of the formation of precartilage in all mammals studied except possibly man. This vertebral expansion is usually obliterated as the vertebra chondrifies, and the vertebral portion of the notochordal sheaths and small pieces of notochordal tissue occasionally or regularly retained in the vertebra are destroyed before the ossification of the vertebra. Most of the notochordal tissue is forced into the intervertebral disc and, growing, forms the nucleus pulposus.

4. Notochordal tissue undergoes a characteristic cytomorphosis. It is primarily cellular and epithelial; later it becomes a syncytial network with a niucin-like substance in its vacuoles; and finally it becomes cellular and closely resembles cartilage.


Development of the Notochord 281

5. The form of the notochordal enlargements in each species studied, is characteristic of that species at each stage of its development.

BIBLIOGRAPHY.

Bardeen, C. R., 1905. The development of the thoracic vertebrae in man.

Amer. Jour. Anat., IV, 163-174. Carius, F., 18S8. Ueber die Entwickelung der Chorda und der primitiven

Rachenhaut bei Meerschweinchen und Kaniuchen. Diss. Marburg.

33 pp. DuRSY, E., 1SG9. Zur Entwiclceluugsgeschichte des Kopfes, pp. 232, Atlas.

Tiibingen. Fischer, B., und Steiner, 1906. Ueber ein maliugues Chordom der Schiidel Riickgratshohle. Beitr. z. Path. Anat, Jena, XL, 109-119. Fric, R., 1904. Handbuch der Gelenke, Jena. 512 pp., 161 figs. Froriep, a., 1882. Kopfteil der Chorda dorsalis bei menschlichen Embry onen. Beitr. z. Anat. u. Embryol, als Festschrift f. J. Henle. 1883. Zur Entwickelungsgeschichte der Wirbelsiiule, insbesondere des

Atlas und Epistropheus und der Occipitalregion. I. Beobachtung an

Hiihnerembryonen. Arch. f. Anat. u. Phys., Anat. Abt. 1886. II. Beobachtung an Saugetierembryonen. Ibid. Gaupp, E., 1905. Die Entwickelung des Kopfskelettes, Hertwig's Handbuch

der vergl. u. experimentalen Entwickelungslehre der Wii'beltiere.

Jena. Bd. 3, Teil 2, 573-874. Gbahl, Otto, 1903. Eine Ecchondrosis physalifora spheno-occipitalis unge wohnlichen Umfangs mit interessanten klinischen Folgen. I. D.

Gottingen. Heiberg, J., 1878. Ueber die Zwischenwirbelgeleuke und Knochenkerne der

Wirbelsiiule. Mitth. a. d. Embryol. Inst. d. K. K. Univ. Wien. I,

1880. 119-129. Jageb, G., 1858. Das WirbelkiJrpergelenk der Vogel. Sitz. Wiener Akad.

XXXIII. Keibel, Fr., 1889. Zur Entwickelungsgeschichte der Chorda bei Sfiugern.

Arch. f. Anat. u. Phys., Anat. Abt, 329-388. Klebs, 1864. Ein Fall von Ecchondrosis spheno-occipitalis amylacea. Vir chow's Arch., XXXI, p. 396-399. KoLLiKER, A., 1859. Ueber die Beziehungen der Chorda dorsalis zur Bildung

der Wirbel, etc. Verb, phys.-med. Ges. Wiir/Aturg, IX. 193-242, T.

I-III. 1867. Handbuch d. Gewebelehre des Menschen. Leipzig, 5th ed. 749

pp., 524 figs. 1883. Chordahohle und Bildung der Chorda beim Kaninchen. Sitz.

phys.-med. Ges. Wiirzburg, pp. 2-9.


282 Leonard W. Williams

Leboucq, H., 1880. Recherches sur la mode de disparition de la corde dorsale

Chez les vertebres superieurs. Arch, de Biol., I, 718-73G, pi. 29, figs.

1-19. LiEBEBKUHN, 1882. Uebor die Chorda bei Siiugethieren. Arch. f. Anat. u.

Phys., Anat. Abth., 396-436. 1884. Ibid. 435-452. Lowe, L., 1879. Zur Kenntniss der Siiugethierchorda. Arch. f. mikr. Anat.,

XVI, 597-612, pi. 29. LuscHKA, H., 1852. Die Halbgelenke.

1856. Die Altersveriind. der Zwischenwirbelknorpel. Virchow's Arch., IX., 309-327.

1857. Ueber gallertartige Auswiichse am Clivns Blumenbachii. Virchow's Arch., XI., S-12, 4 figs.

1858. Die Halbgelenke, pp. 144, T. VI. Berlin.

MI^-OT, C. S., 1907. Segmental flexnres of the Notochord. Anat. Rec, No. 3,

1907, 42-56, 6 figs. MtJLLER, H., 1858. Ueber das Vorkommen von Resten der Chorda dorsalis

beim Menschen. Zeitschr. f. rat. Med., II., 202, Taf. III. RiBDERT, H., 1895. Ueber die experimentelle Erzeugung einer Ecchondrosis

physalifora. Verb. XIII, Congr. f. innere Med., 455-464, pi. V. 1904. Geschwiilstlehre. Bonn. 062 pp., 596 figs. Robin, C, 1868. Memoire siir revolution de la notocorde. Paris, 212 pp.,

12 pis., 63 figs. Rosenberg, E., 1875. Ueber die Entwickelung der Wirbelsiiule und des

centrale carpi des Menschen. Morphol. Jahrb., I., 83-197, PI. III-V. ScHAUiNSLAND, H., 1905. Die Entwickelung der Wirbelsaule. Hertwig's

Handbuch der vergl. und experimentalen Entwickelungslehre der

Wirbelthiere. Jena. Bd. 3, Teil 2, 339-672. ScHULTZE, O., 1896. Ueber embryonale und bleibende Segmentierung. Verb.

anat. Gesellsch. Berlin. 87-93. 1 fig. 1897. Grundriss der Entwickelungsgeschichte des Menschen und der

Silugethiere. Leipzig. Steinek, H., 1894. Ueber die Ecchondrosis physalifora spheno-occipitalis.

Centralbl. f. path. Anat., V., 4.57-461. ViRCHOw, R., 1857. Ueber die Entwicklung des Schadelgi-undes. Berlin. • Weiss, A:, 1901. Entwick. der Wirbelsaule der weissen Ratte, Zeit. f. wiss.

Zool., 69, p. 492-582, pi. 38-39.


Development of the Notochord 283

EXPLANATION OF FIGURES.

Figs. 1 to 3 are photographs of the same magnification of similar frontal sections of pig embryos of 7.8, 9, and 12 mm. The notochord, the spinal nerves, the intersegmental vessels, the loose tissue of the vertebrae and the dense tissue of the "scleromere" should be noted.

Fig. 1. 7.8 mm. pig embryo, section 381, Harvard Embryological Collection, No. 430, X 306.

Fig. 2. 9 mm. pig embryo, section 595, H. E. C, No. 54, X 306.

Fig. 3. 12 mm. pig embryo, section 562, H. E. C, No. 6, x 306.

Figs. 4 to 7 are photographs of median sagittal sections of vertebrae of pig embryos of 14, 20, 24, and 75 mm.

Fig. 4. Fourth dorsal vertebra, embryo of 14 mm. H. E. C, No. 66, section 184, X 133.

Fig. 5. First dorsal vertebra, embryo of 20 mm. H. B. C, No. 60, section 24S, X 110.

Fig. 6. Fifth vertebra, embryo of 24 mm. H. E. C, No. 63, section 28, X 76.

Fig. 7. Eleventh dorsal vertebra, embryo of 75 mm. x 30.

Fig. S. The nucleus pulposus and portions of the anmihts fihrosus and of the epiphyses of an adult pig. X 1^ Figs. 9 to 12 are photographs of median sagittal sections of vertebrae of embryos of the rabbit, opossum, cat and man.

Fig. 9. Twelfth dorsal vertebra of a rabbit embryo of 29 mm. H. E. C, No. 171, section 182, x 60.

Fig. 10. Fourteenth dorsal vertebra of an opossum embryo of 12 mm. H. E. C, No. 616, section 196, X 112.

Fig. 11. First dorsal vertebra of a cat embryo of 39 mm. H. E. C, No. 394, section 195, X 56.

. Fig. 12. Eighth dorsal vertebra, human embryo of 32 mm. H. E. C, No. 292, section 176, X 56.

Fig. 13. Nuclei, which are possibly undergoing amitotic division, from a vertebra of a pig embryo of 12 mm. H, B. C, No. 5, section 894, x 1000.

Fig. 14. Third dorsal vertebra of a sheep embryo of 26.1 mm. H. E. C, No. 1112, section 235, X 36. The trilobate character of the intervertebral notochordal enlargement is shown.

Fig. 15. Notochordal syncytium with mucin-filled spaces from a pig embryo of 150 mm. x 800.


284 Leonard W. Williams

Fig. 16. Three vacuolated cells from the nucleus pulposus of an adult pig. X 452.

Figs. 17 to 20 are reconstructions of a median section of the notochord and chondrostyle of embryos of the opossum, rabbit, cat and man.

Fig. 17. From an opossum embryo of 12 mm. H. E. C, No. 616, x 1^ Fig. is. From a rabbit embryo of 14.5 mm. H. E. C, No. 162, X 13.

Fig. 19. From a cat embryo of 15 mm. H. E. C, No. 437, X 12.

Fig. 20. From a human embryo of 32 mm. H. E. C, No. 292, x "•


DEVELOPMENT OF THE NOTOCHORD

LEONARD W. WILLIAMS




Fig. 1


Fig 2



Fig. 3


Fig. 4







Fig. 5


Fig. 6


THE AMERICAN JOURNAL OF ANATOMY— VOL. VIII, NO. 3


OEVELOPMENT OF THE NOTOCHORD

LEONARD W. WILLIAMS



Fig. 7



'^^:k^M$


Fig. 9



-'*^



Fig. 11


Fig. 12


THE AMERICAN JOURNAL OF ANATOMY— VOL. VIII, NO. 3


DEVELOPMENT OF THE NOTOCHORD

LEONARD W. WILLIAMS


^^^mm



^^fe^^^!^:g:;?"ff%; ;^


Fig. 13


Fig. 14


"^.Jig:y:'y"^


1 'if' m


m^^'"^


^ ; f -'-^



Fig. 15


Fig. 16


THE AMERICAN JOURNAL OF ANATOMY~VOL. VIM, NO. 3


DEVELOPMENT OF THE NOTOCHORD

LEONARD W. WILLIAMS



Fig. 17


THE AMERICAN JOURNAL OF ANATOMY— VOL. VIII, NO. 3


DEVELOPMENT OF THE NOTOCHORD

LEONARD W. WILLIAMS



Fig. 11


THE AMERICAN JOURNAL OF ANATOMY--VOL. VIII, NO. 3


DEVELOPMENT OF THE NOTOCHORD

LEONARD W. WILLIAMS


I



Fig. 19


THE AMERICAN JOURNAL OF ANATOMY— VOL. VIM, NO. 3


DEVELOPMENT OF THE NOTOCHORD

LEONARD W. WILLIAMS



THE AMERICAN JOURNAL OF ANATOMY--VOL. VIII, NO. 3


T^iE PERIPHEEAL NERVOUS SYSTEM IN THE HUMAN EMBRYO AT THE END OF THE FIRST MONTH (10 mm.).


GEORGE L. STREETER, M.D.,

Professoi- of Anatomy, University of Michigan, Ann Arbor.

(With 3 Plates and 1 Text Figure.)

The period of development under consideration represents the completion of what we may call the primary stage in the growth of the nervous system. The primary neurones forming the peripheral nerves are by the end of the first month well laid down; all their chief peripheral branches and plexuses are indicated, and centrally the nerve roots can be traced to their distribution in the brain and spinal cord, where the nuclei of the motor roots can be outlined and the sensory roots can be recognized as definite bundles extending up and down in the wall of the neural tube. The higher neurone systems, however, are still in a most rudimentary state, and in sections through the brain and cord at this time we see only the primary apparatus differentiated. Such co-ordinating centers as the pons, olive and cerebellum are still undeveloped, and the forebrain is not much more than an undifferentiated thin-walled tube. It can thus be seen that the period with which we are dealing represents a definite stage in the growth of the nervous system, the stage of the primary brain, and a stage which is of particular importance for a proper conception of the embryology of this system.

A general view of the nervous system as it exists in embryos at the end of the first month is represented on Plate I. It can bo seen that the reconstruction shown there corresponds in age almost exactly with the well known His reconstruction of his embryo KO of 10.3 mm. Nl. That investigator early recognized the significance of the stage of the primary brain. However, since His, '88, published his monograph and description of the embryo KO there have been introduced many improvements in the methods of work, and before all others should be mentioned the Born wax plate procedure. It is with such The American Jodenal of Anatomy.— Vol. VIII, No. 3.


286 George L. Streeter

aid and by means of the increased supply of available embryos that it was possible by Lewis, '02, to elicit further detail and obtain greater completeness in our conception of the formation of the brachial plexus and the development of the nerves of the arm, and Bardeen, '07, the nerves of the leg, and myself the nerves of the occipital region (Streeter, '04). With this in mind it seemed desirable to go over the same ground covered by His in his reconstruction of the embryo KO with the purpose of bringing out the same structures with more accurate detail and in more plastic form, and at the same time to incorporate the results of the work of the investigators just mentioned.

Some of the features brought out by this study with regard to the cranial nerves were reported at the Chicago meeting of the Association of American Anatomists (Streeter, '08). It is the purpose of the present paper to include the whole peripheral nervous system of the same embr}'0.

Material and Methods. The embryo on which this study is based constitutes No. 3 of Professor Ruber's Collection, and was kindly loaned by him for this purpose. After fixation in alcohol the specimen measured 10 mm. greatest length. Its estimated age is 31 days. Unfortunately there was no clinical history obtainable. The embryo is cut in a series of 5 micra sagittal sections, and the tissues are in an excellent state of preservation. The description and drawings are based in part on wax plate models and in part on profile reconstructions made in the usual way by superimposing transparent papers.

Brain" and Spinal Cord.

The general form of the brain and spinal cord and their relation to the body outlines is shown in Plate I. The spinal cord is largest in the cervical region, and from there slowly tapers down, being somewhat larger again in the lumbar region, and finally abruptly tapers off at the coccyx. In transverse section the cord possesses a trapezoid outline with rounded corners, the width of the ventral half being a little greater than that of the dorsal half. A large ventricle or central canal extends throughout its whole length. By the lanceolate border of this canal the cord is divided on each side of the median line into an alar and basal plate.

Toward the head the spinal cord enlarges and gradually merges into the rhombencephalon. In this transition the basal plates become thicker,


Peripheral JSTervous System in Human Embryo


287


and the alar plates become wider, and at the same time flare apart. As the alar plates spread apart the narrow dorsal seam, that exists between the two in the spinal region, widens out into the broad roof of the fourth ventricle. The most striking feature of the rhombencephalon



Fig. 1. Composite sagittal view of a 10 mm. human embryo (Huber collection No. 3), showing the rhombencephalon and the so-called rhombic grooves, which exist temporarily as transverse furrows in the floor of the ventricle. They are connected by the fifth, seventh, ninth and tenth nerves with the branchial arches. The fourth groove (d.) is an exception and has no corresponding visceral nerve. Arising from it can be seen the sixth nerve ^ extending forward to reach the anlage of the external rectus muscle.


in this embryo is its large size as compared with the rest of the brain. Its general form is shown in tbe drawing. It will be noticed that that part of the alar plate cephalad to the trigeminal nerve, which is to take part in the formation of the cerebellum, shows as yet little signs of differentiation.


288 George L. Streeter

The rhombic grooves present in the floor of the fourth ventricle were described in the paper previoiisly mentioned (Streeter, '08), and their relation to the cranial nerves was also referred to. It will only be necessary here to refer to the accompanying Fig. 1., in which is shown a composite sagittal view of the embryo. The six rhombic grooves are indicated by the letters a. to f. Extending from the rhombic grooves to the branchial arches may be seen the trigeminal, facial, glossopharyngeal and vagus nerves. The correspondence of the arches, the nerves and the grooves strongly suggests the branchiomeric character of the whole group. Arising from groove d. is the n. abducens extending forward to the region median to the trigeminal. This is included in the drawing, though it is recognized that it may have no determinative connection with the fourth groove with which it stands in relation.

The outlines between the mesencephalon, thalamencephalon and prosencephalon are easily made out, and correspond closely with the His KO embryo (Tab. II, Fig. 3). It is to be noted, however, that in this case the olfactory bulb is not as well developed as indicated by the His drawing, the beginning olfactory pouch being represented by only a slight depression in the brain wall just lateral to the lamina terminals.

The general structure of the neural tube at tliis stage is shown in a series of cross sections given in Figs. 20-26 in the monograph of His, '88, mentioned above. They show the division of the wall into three layers, the so-called ependymal, mantle and marginal zones; the first is made up of deeply staining primitive cylindrical cells, the second of clumps of budding neuroblasts, and the third is almost wholly fibrous. In the mantle zone are developed the nuclei of origin of the motor nerve roots, which can be outlined as shown in the reconstructions. The marginal zone consists of a reticulated framework through which the various fibre tracts make their way. At this time the only tracts that stand out prominently are the dorsal funiculi of the spinal cord, consisting of processes from the dorsal roots of the spinal nerves, and in the brain the corresponding tractus solitarius and spinal tract of the trigeminal nerve. In addition to these a portion of the marginal zone can be outlined which contains developing longitudinal fibres, and represents the so-called ventral ground bundle. In the cord it serves to connect the different levels of the mantle layer. It is continued through the hindbrain, and longitudinal fibres grow through it in that region, eventually resulting in the three main longitudinal tracts : the fasciculus longitudinalis medialis, the lemniscus and the tractus pyramidalis. This


Peripheral Nervous System in Human Embryo 289

ground bundle has been modelled out and forms the support to the reconstruction shown in Plate II.

Cekebral Nerves.

For purposes of description the nerves of the head will be grouped in accordance with their relation to the four functional systems of Gaskell, the significance of which has been emphasized in the recent writings of Johnston, C. J. Herrick, and others. According to Gaskell the functional systems are determined by the two chief activities of the organism; first, actions in relation to the external world (somatic), and second, internal activities having to do with the processes of nutrition, etc. (visceral). In each case there is the double activity on the part of the nervous system, sensory and motor, making in all four primary functional divisions. Now some of the cranial nerves consist of elements belonging exclusively to one functional division, for example, the n. abducens which consists entirely of somatic motor fibres, while others are complex nerves containing elements of more than one system, such as the n. glossopharyngeus which contains elements from three functional divisions, somatic sensory, visceral sensory and visceral motor. The nerves will therefore be grouped according to the predominance of their functional elements as follows:

Somatic Sensory. Somatic Motor Visceral

Olfactory Oculomotor Trigeminal

Optic Trochlear Facial

Acoustic Abducens Glossopharyngeal

Hypoglossal Vagus and Accessory

In general the basal plate of the neural tube is motor and the lateral plate is sensory; thus the somatic motor group arises entirely from the basal plate, while the visceral group arises in large part from the lateral plate and in lesser part from the basal plate. The somatic sensory group is specialized and these nerves have individual processes or lobes of the nervous system devoted to them, i. e., the olfactory bulb, the eye bulb and stalk, and the tuberculum acusticum.

Nerves of Special Sense Organs.

These nerves belong to that large group of afferent nerves that connect the integument with the central nervous system, the somatic


290 George L. Streeter

sensory nerves. The general cutaneous nerves supplying the whole surface of the body compose the bulk of this group, and they are represented in the head by fibres belonging to the trigeminal, ninth and tenth nerves. In addition there are in the head region special somatic sensory nerves. The union of nerve and integument has resulted in these cases in the formation of special sense organs, that is the olfactory organ, the eye and the ear, composed partly of nerve elements and partly of integument. In this sense the olfactory nerves, the retinal ganglion cells and the acoustic nerve, though differing so widely in their adult morphology, may be considered as analogous in development and function.

Nn. olfadorii. At the end of the first month the ectodermal olfactory pit is definitely formed and is already in the process of differentiation, evidenced by the increase of body-protoplasm of its epithelial cells. A corresponding pouch is just beginning to form at the olfactory area at the base of the cerebral hemisphere, the anlage of the olfactory bulb. The fibres which later connect the two, nn. olfactorii, cannot yet be made out.

N. opticus. In case of the optic apparatus the chief contribution on the part of the integument is the lens; and the nerve contribution is the retina. The optic nerve corresponds to the olfactory stalk, and it is formed by the conversion of the optic stalk into a fibre tract connecting the retina with the brain. In the embryo studied the optic stalk consists of a thick-walled hollow tube whose cavity still freely communicates with the general brain cavity. The lower border is indented by the choroidal fissure which gives it a crescentic outline in cross section. The structure of this stalk is like that of the wall of the hemisphere, consisting of a thick ependymal layer covered in by a thin fibrous marginal zone. This marginal zone later thickens at the expense of the ependymal layer and is converted into a framework through which the nerve fibres from the retina make their way to the brain. At the end of the first month we cannot yet refer to this stalk properly as the optic nerve.

N. acusticus. This nerve is in a more advanced stage of development than the optic and olfactory nerves. The epithelial part of this apparatus, the ear vesicle, is already pinched off from the skin, as is shown in Plate I, and consists of a closed sac consisting of a double pouch having an upper vestibular portion and a lower cochlear portion; at the junction of the two on its median side is attached the ductus endolymphaticus.

The acoustic ganglion lies closely against the cephalic and median border of the ear vesicle. Two groups of fibres from the ganglion can


Peripheral Nervous Sj'stem in Human Embryo 291

be made out entering into connection with the cells of the vesicle; the upper group corresponds to the nerves to the superior and lateral ampullae and the utricle, and the lower group the nerves to the saccule and posterior ampulla. The ganglion is somewhat elongated and consists of an upper and lower portion. The upper portion is wholly vestibular, and the lower is partly vestibular and partly cochlear. The cochlear part, or ganglion spiral, is the thickened border of the lower division' of the ganglion of which it is a derivative. The acoustic nerve emerges from the proximal end of the ganglion and enters the brain just lateral to the n. intermedins. The nerve at this time consists almost wholly of vestibular fibres. It is shortly after this that a well defined separate cochlear trunk can be made out connecting the ganglion spirale with the brain. What has been mistaken in younger embryos for a cochlear trunk are the fibres from the lower division of the vestibular ganglion. The acoustic fibres enter the brain wall near its border opposite the third and fourth rhombic grooves, and spread out in the marginal zone just dorsal to the spinal tract of the trigeminal nerve to form the anlage of the tuberculum acusticum.


Somatic Motor Group.

The hypoglossal and the three nerves to the extrinsic eye muscles (nn. oculomotorius, trochlearis and abducens) that compose this group are all shown in Plates I and II. As can be seen in Plate II, their nuclei of origin may be considered as a cephalic continuation of the ventral motor column of the spinal cord. This is particularly evident in case of the hypoglossal nerve, whose nucleus and emerging fibres form a continuous line with the ventral roots of the cervical nerves, and their close relation is shown by the tendency of the two to unite in the formation of a plexus. The eye nerves are in the same series with the hypoglossal, and always maintain a similar position near the median line and directly beneath the floor of the ventricle, though they are separated longitudinally at varying intervals. The character of the muclei of origin is the same in all four nerves.

There are no ganglia on these nerves such as are found in the spinal region, though occasionally a ganglion and also at times a dorsal root is associated with the more caudal roots of the hypoglossal. In such cases the ganglion is to be regarded as a precervical one, the exact counterpart of the spinal ganglia. This ganglion when present is the so-called Froriep ganglion.


292 George L. Streeter

The n. oculomolorius arises from a group of neuroblasts forming the ventral part of the mantle layer in the mesencephalon. These neuroblasts converge to form small rootlets which pass through the ground bundle and emerge on the ventral surface of the neural tube in the concavity of the cephalic bend. Here they unite into a common trunk, which at first passes directly ventralward and then making a slight angle bends lateralward and finally breaks up median and between the first and second divisions of the trigeminal nerve, in the cellular mass which is to form the eye muscles.

The n. troclilearis arises from a group of neuroblasts similar and lying just caudal to those of the oculomotor. The rootlets derived from them, instead of emerging directly ventralward, curve dorsalward to reach the roof of the isthmus, where they decussate and emerge as a solid trunk which passes down lateral to the neural tube and breaks up among the cells which are to form the superior oblique muscle.

The n. abducens arises from a group of neuroblasts forming the median part of the mantle layer directly beneath the fourth rhombic groove. See Fig. 1. The rootlets arising from these neuroblasts pass directly ventralward and after emerging they unite to form the main trunk, which bends forward immediately at an angle of nearly 90°, and passes forward into the region of the terminal rootlets of the oculomotor, as shown in Plate I. The relative positions of the abducens and the facial nerve will be referred to under the description of the latter nerve.

Visceral Group.

The facial, glossopharyngeal and vagus form a series of similar nerves which consist almost wholly of visceral fibres. As can be seen on Plates II and III, the visceral motor fibres arise from the nucleus ambiguus, which consists of a column of neuroblasts continuous with the lateral horn cells of the cord. The visceral sensory fibres arise from the peripheral ganglia and enter the alar plate of the neural tube and form a longitudinal strand which in the adult we know as the tractus solitarius. In addition to these visceral fibres there is a small number of somatic sensory fibres, supplying the integument of the adjoining region, which arise and have a course similar to the visceral sensory fibres. In aquatic vertebrates there are also the special somatic sensory fibres of the lateral line system, whose fibres join the rootlets of the facial, glossopharyngeal and vagus to reach the brain, and the ganglia from which these fibres are derived become incorporated in the geniculate, petrosal and nodosal


Peripheral Nervous System in Human Embryo 293

ganglia. We still have in the human embryo a trace of these organs in the form of a temporary thickening of the ectoderm directly over the ganglia of these three nerves.

The 11. facialis is characterized by a large predominance of visceral motor fibres. They make up the bulk of the adult nerve, and in the higher vertebrates they have been made use of to supply the muscles of expression. These motor fibres arise from a group of neuroblasts situated beneath the third rhombic groove. Fibre bundles are assembled and pass directly lateral under the floor of this groove gradually converging to form a solid trunk which emerges from the neural tube just median to the acoustic ganglion. The trunk curves backward, as can be seen in Plate I, and breaks up among the cells of the hyoid arch, from which the facial musculature is to be derived. The sensory fibres of this nerve are derived from the geniculate ganglion. From the proximal end of the ganglion the fibres are assembled to form the so-called n. intermedins which enters the alar plate of the neural tube and forms the beginning of the tractus solitarius as shown in Plate III. From the peripheral end of the ganglion the fibres pass down to form the chorda tympani, and finally leave the main trunk of the nerve to enter the mandibular arch, eventually joining the third division of the trigeminal nerve. The great superficial petrosal is another peripheral derivative of this ganglion which makes its appearance a little later. Though the facial and acoustic nerves are closely united in position, and connecting fibres are usually present between the two in the adult, it must still be remembered that this is only due to the fact they lie close together. Further than that they have nothing in common, being nerves which belong to entirely different embryological and functional classes. The term facial-acoustic complex should only be used in the sense of position.

In a paper already referred to (Streeter, '08,) it was pointed out that the facial (motor root) and abducens nerves occupy positions in the embryo which are relatively reversed in the adult. The facial at first lies directly under the third rhombic groove, while the abducens is more caudal and is under the fourth rhombic groove. As development progresses the nuclei of these two nerves shift their relative positions, the abducens migrating forward. This migration results in the bending of the motor root of the facial out of its original course and produces the genu facialis which is characteristic of the adult.

The n. glossopharyngeus forms a more typical visceral nerve than either the facial or vagus. It possesses a ganglion of the root and


294 George L. Streeter

ganglion of the trunk, the latter being temporarily connected with the placode over the third arch. As can be seen from the relative size of the ganglia the nerve consists mostly of sensory fibres, connected peripherally with the structures developing from the second (r. tympanicus) and third (r. lingualis) arches. The tympanic branch is not well defined until we come to embryos between twelve and fourteen mm. long. Centrally the rootlets enter the brain wall and, joining with the fibres from the facial, extend caudally (Plate III) as the tractus solitarius. The motor rootlets of this nerve arise from a group of neuroblasts in the nucleus ambiguus series, situated beneath the floor of the fifth rhombic groove. The motor bundles extend directly lateral beneath this groove and pass under the spinal tract of the trigeminal and then emerge from the brain wall and join the main trunk of the nerve.

The n. vagus like the facial represents a branchial nerve the motor fibres of which have in man undergone special development, for the purpose of supplying a group of muscles derived from its branchial arch or arches. The large motor trunk of the facial nerve, as we have seen, is distributed to the muscle cells of the hyoid arch and, as these cells group themselves into the muscles of expression and spread forward over the face, the facial branches are drawn along with them. In a similar way the more caudal rootlets of the vagus become predominantly motor and form a distinct bundle which we know as the spinal accessory nerve, and this bundle is distributed to a group of muscle cells originally belonging to the more caudal branchial arches, and in man are destined to form muscles for the arm girdle, the mm. sternocleidomastoideus and trapezius. As these muscles spread out into their eventual position the nerve is drawn down across the neck with them. Coincident with the increased importance of this musculature as we ascend the vertebrate scale we meet with increased development of the spinal accessory, and it obtains additional rootlets of origin by spreading down into the region of the spinal cord. As can be seen in Plate I, the spinal accessory may reach as far down as the fourth cervical segment. The nucleus of origin of the spinal accessory and other motor rootlets of the vagus constitutes the nucleus ambiguus of the medulla oblongata and the lateral horn of the spinal cord, the two being directly continuous. This is best shown in Plate II.

The neuroblasts of the basal plate of the neural tube form two columns, a larger median one and a smaller lateral one ; the median column furnishes somatic motor fibres to the ventral spinal roots and hypoglossal,^


Peripheral Nervous System in Human Embryo 295

and at intervals further forward the motor nerves of the eye ; the lateral column furnishes visceral motor fibres to the dorsal spinal roots and to the vagus (spinal accessory), glossopharyngeal, facial and trigeminal nerves. In the 10 mm. embryo these neuroblast columns are longitudinally continuous from the spinal cord into the vagus region, and there is present a continuous series of median (ventral) and lateral rootlets. As other structures develop the columns become interrupted and particularly the lateral column; thus in the nucleus ambiguus of the adult we have a broken series of discrete nuclei.

The motor neuroblasts of the vagus point dorso-lateralward and form rootlets which emerge just ventral to the entrance of the sensory roots. After emergence they turn forward and form a common trunk which in the spinal region lies between the dorsal roots and the side of the spinal cord. The more caudal rootlets are devoid of sensory fibres; but as we go forward we meet with ganglionated rootlets. In the vagus as in the glossopharyngeal there are the ganglia of the roots and the ganglion of the trunk (ganglion nodosum). The ganglia of the roots represent a series diminishing in the caudal direction, as shown in Plate I. In the adult the more caudal ganglia usually disappear except for traces of scattered ganglion cells found occasionally on the rootlets of the spinal accessory division. These more caudal vagus ganglia are not to be mistaken for the Froriep ganglion, which represents a persistent precervical ganglion. In the one case we have a series diminishing from the head toward the tail, and in the other it is in the opposite direction. Owing to the tendency to regression on the part of more caudal of the vagus root ganglia the vagus complex, as was shown in a former paper (Streeter, '04), becomes differentiated into a fore part or vagus division which is predominantly sensory, and a back part or accessory division which is almost wholly motor. In the 10 mm. embryo there is no division between the two parts.

On entering the wall of the neural tube the sensory fibres immediately unite in a longitudinal tract continuous with similar fibres from the facial and glossopharyngeal thus completing the formation of the tractus solitarius. In Plate III is shown a cross section of the neural tube in the vagus region, and on it is indicated the position of the tractus solitarius. The marginal or reticular zone of the alar plate in this region is mostly made up of the longitudinal fibres of this fasciculus, and directly ventral lies the similar group of fibres belonging to the trigeminal nerve. The relation of the two is best shown in Plate II.


296 George L. Streeter

Peripherally the fibres from the ganglia of the roots together with the remaining motor fibres that are not included in the accessory bundle are collected into a common trunk and pass down caudally to be lost in the ganglion nodosum. After they emerge at the distal end of this ganglion they bend medialward and lose themselves on the wall of the oesophagus.

The n. trigeminus possesses the largest ganglion of the whole embryo. From this ganglion the three large peripheral divisions extend ventralward. The ophthalmic division passes forward and subdivides into the frontal and nasociliary nerves, the latter forming a long slender well defined branch passing just dorsal to the eye stalk. The maxillary and mandibular divisions pass downward and break up in their terminal branches among the cells of the maxillary process and mandibular arch respectively. Centrally the ganglion is connected with the brain by a thick short trunk which enters the wall at the pontal bend and opposite the first and second rhombic grooves. Within the wall the fibres immediately form a flattened longitudinal tract, as shown in Plate II, part of which extends caudally as the spinal tract, and part extends forward and upward to enter the cerebellar ridge. These fibres must be considered as containing both somatic and visceral elements, between which no difference could be made out embryologically.

In its motor elements the trigeminal nerve departs somewhat from the type represented in the other three nerves of this visceral group. In the others the nucleus of origin is in the basal plate, and the nerve rootlets exhibit a characteristic curved course to reach the point of emergence; while in the trigeminal the nucleus is more lateral and lies directly against the entering sensory fibres, so that the fibres of the motor root pass directly ventralward to fuse with the mandibular division.

It is possible that the motor nucleus of the trigeminal is to be considered as an hypertrophied example of one of the dorsal motor nuclei found in the adult ninth and tenth nerves, both median and lateral to the tractus solitarius^ and which have not yet been recognized in the embryo. In that case we must conclude that either the nucleus ambiguus and ventral motor root are not present in this nerve, or that they are represented by the mesencephalic motor root. In analyzing these nerves

Tor a description of the motor nuclei connected with the tractus solitarius see E. L. Melius, '02.


Peripheral ISTervous System in Human Embryo 297

it is to be remembered that in man the typical visceral cranial nerve has three central terminations:

1. Sensory root (tractus solitarius).

2. Curved ventral motor root (nucleus ambiguus).

3. Straight dorsal motor root (nucleus vagi dorsalis).

These three elements may be represented in the different nerves in different proportions. The ninth nerve approaches the mean and all elements are fairly represented. In the vagus the curved ventral motor roots are increased in proportion in the caudal portions and form the spinal accessory. In the facial the sensory root (n. intermedins) is diminutive, while the curved ventral root becomes the main trunk of the nerve. In the trigeminal it is the straight dorsal motor root that forms the principal motor supply, while the curved ventral motor root is either not present or is represented by its mesencephalic root.

Spinal Nerves.

At the end of the first month each segmental nerve of the trunk possesses a sharply outlined spinal ganglion, whose constituent cells are in the early stages of differentiation. On examination it can be seen that many of these cells consist of a prominent nucleus surrounded by a thin rim of ill defined body protoplasm. In other cells the body protoplasm has increased in the form of a process at one or both ends. Cells of this kind are clustered so that their processes unite to form fibrous strands. These strands in turn fuse into larger bundles and lead toward the two poles of the ganglion. At the proximal pole they become grouped into the dorsal roots, which enter the spinal cord in an uninterrupted longitudinal series. In the cord they unite and extend up and down in the marginal zone in the form of a flattened band of fibres which later constitutes one of the dorsal funiculi of the cord.

The fibres from the distal pole of the ganglion imite in a common bundle which is almost immediately joined by the fibres of the ventral root, the two together forming the main trunk of the nerve. The ventral roots consist of a continuous series of rootlets emerging from the ventrolateral border of the neural tube, and derived from the neuroblasts of the mantel layer of its basal plate. These neuroblasts form a longitudinal column which, as shown in Plate II, is continuous with and of the same character as the nucleus of the hypoglossal nerve.


298 Geora-e L. Streeter


b"


At the same time that the dorsal and ventral roots imite to form the main trunk they give off lateral fibres to form the dorsal branch, the so-called posterior primary division, which turns back dorsalward and covers in the distal part of the ganglion, and breaks up among the cells which are to form the long muscles of the back.

The remainder of the nerve trunk is continued forward as the ventral branch or anterior primary division. From its median side there is given off the ramus communicans, which extends medianward to the region of the aorta and ends in the sympathetic ganglion cord. The rami communicantes and the sympathetic cord are not shown in Plate I. The main trunk terminates in two branches, the anterior and lateral terminal branches, which correspond to the anterior and lateral cutaneous branches of the adult, and which in the thoracic and abdominal regions end among the cells giving rise to the musculature of the front and lateral body wall.

Throughout the spinal region there is a tendency for the nerve trunks to unite at the level of the lateral terminal branches and form intersegmental loops. This loop — or plexus — formation may involve either the lateral or the anterior terminal branches, or both. "We thus have produced the cervical, brachial and lumbosacral plexuses.

In the cervical region the anterior and lateral terminal branches form two separate plexuses; the former produces the deep cervical plexus and the latter the superficial cervical plexus. The superficial cervical plexus consists of the union of the lateral terminal branches into loops from which are given off the cutaneous branches to the auricular, cervical and occipital regions. The deep plexus results in the formation of the ansa hypoglossi and the phrenic nerve. The former is produced by the fusion of the second and third cervical nerves into the descendens cervicis, which unites in a loop with the hypoglossal, together with which the first cervical has been incorporated above. From this loop are given off the short branches which end among the cells that are to form the hyoid musculature. The main trunk of the hypoglossal bends sharply medianward to end in the tongue anlage. The deep cervical plexus was studied in an older embryo (14 mm. — Mall embryo No. 144) and described in a former paper (Streeter, '04). The ansa hypoglossi is shown in Fig. 11 and is essentially the same as in the present case, differing only in that the communication between the first cervical and the hypoglossal could not be traced. In Plate II of that paper the r. descendens is labelled wrong; the leader should extend to a point above


Peripheral ISTervons System in Human Embryo 299

where it is joined by the descendens cervicis, which can be seen as a slender nerve made up of branches from the second and third cervical nerves.

The phrenic nerve is formed by anterior terminal branches from the fourth and fifth cervical nerves. A contribution on the part of the sixth could not be made out. A view of the nerve can be seen through the arm in Plate I. Owing to the position of the diaphragm at this time the course of the nerve is almost directly ventral, passing over the lung anlage which is not represented in the drawing. Later, as pointed out by His and Mall (Mall, '01), the points of origin and insertion of the nerve draw gradually apart, due, on the one hand to the descent of the diaphragm and the lengthening of the thoracic cavity, and on the other hand to the subsequent elevation of the cervical nerves which accompanies the development of the structures of the neck. It is thus that there results the long caudal course of this nerve that is characteristic of the adult.

The brachial plexus is shown in Plate I by representing the arm as transparent. In the region of the fifth cervical to the first thoracic nerves there is an exuberant growth of both the anterior and lateral terminal branches, resulting in a solid flattened mass of fibres, which in turn is split by the skeletal anlage into two laminae, from which the various nerves arise. Arising from the anterior or ventral lamina one can recognize the nn. musculocutaneous, medianus and ulnaris, and from the posterior or dorsal lamina the nn. axillaris and radialis. These nerves pass down into the arm and break up in the muscle masses, which they are to supply.

The lumbosacral plexus as compared with the brachial plexus is somewhat retarded in its development. It is formed by the fusion of the trunks of the five lumbar and upper three or four sacral nerves. These nerves and their ganglia, as with the cervical nerves, are enlarged as though stimulated to extra growth by the presence of the limb bud. The nerves unite into a flattened mass of fibres which enters lateralward into the base of the leg bud, the division into anterior and lateral terminal branches being lost in the formation of the plexus. The further course of the fibres is determined by the framework of the leg. Owing to the cell masses of the bony pelvis and the femur the fibres become grouped into four bundles arranged in two pairs, each consisting of a median and lateral trunk. Of the upper pair the median trunk corresponds to the obturator nerve, and the lateral the femoral nerve.


300 George L, Streeter

The lower pair corresponds to the sciatic nerve, and its median bundle constitutes the future tibial nerve, and the lateral the common peroneal. From these larger nerves smaller branches split off and become isolated as discrete nerves, and enter the muscle masses between which the main trunks lie.

The Sympathetic System,

The ventral migration of the cells derived from the neural crest, that are destined to become sympathetic ganglion cells, is completed by the end of the first month. The elements derived from the successive segments have by that time fused together, on each side of the body, to form a longitudinal column of proliferating cells, situated lateral to the aorta and directly in front of the developing bodies of the vertebrae.

This column extends as a rather sharply outlined continuous cellular strand from the occipital region to the level of the lower sacral vertebrae. The differentiation of the individual cells is already under way and nerve fibre formation can be recognized throughout its length, being most pronounced in the thoracic and lower cervical region. This process however has not advanced far enough to produce a breaking apart of the column into segmental nodes, as is characteristic of the adult ganglionic chain.

Dorsally the column remains in connection with the nerve roots and spinal ganglia by means of the rami communicantes , which consist of sharply outlined fibrous bundles, varying from one to three bundles to each segment. The rami communicantes are largest and best developed in the thoracic and lumbar regions. In the cervical region they are least well developed, and owing to the slanting course of the upper cervical nerves they could not be satisfactorily traced in sagittal sections for the upper three segments.

Ventrally in the region of the sixth to the eleventh thoracic segment a definite fibro-cellular plexus extends forward to form the splanchnic nerves and cceliac plexus. The splanchnic nerves extend almost directly forward. As is the case with the phrenic nerve their caudal course is brought about later by the subsequent caudal displacement of the viscera relative to the bodies of the vertebrae. At this time the diaphragm and caudal surface of the heart lie opposite the first and second thoracic vertebrae, and the stomach is opposite the fifth to tenth thoracic vertebrae so that the coeliac region is directly ventral to the origin of the splanchnic nerves from the ganglionated cord.


Peripheral Nervous System in Human Embryo 301

The cardiac plexus and the cranial sympathetic ganglia owing to the more complicated architecture of those regions could not be outlined with any accuracy. The cephalic end of the ganglionated cord can be traced median to the hypoglossal nerve to the region situated between the ganglion nodosum and the wall of the pharynx. Its extension along the internal carotid artery and commuication with the ganglia of the head could not be made out.

LITERATURE.

Bardeen, C. R., 1907. Development and variation of ttie nerves and the

musculature of the inferior extremity and of the neighboring regions of

the trunk in man. Amer. Jour. Anat., Vol. VI. His, W., 1888. Zur Gesehichte des Gehirns sowie der centralen und pevi pherischen Nervenbahnen beim menschlichen Embryo. Abhand. math. phys. CI. Kgl. Sachs. Gesell. d. Wiss., Bd. XIV. Lewis, W. H., 1902. The development of the arm in man. Amer. Jour. Anat.,

Vol. I. Mall, F. P., 1901. On the development of the human diaphragm. Johns

Hopkins Hospital Bulletin, Vol. XII. Mellus, E. L., 1902. On a hitherto undescribed nucleus lateral to the

fasciculus solitarius. Amer. Jour. Anat, Vol. II.

Streeter, G. L., 1904. The development of the cranial and spinal nerves in the occipital region of the human embryo. Amer. Jour. Anat., Vol. IV.

Stbeeter, G. L., 1908. The nuclei of origin of the cranial nerves in the 10 mm. human embryo. Proceed. Assoc. Am. Anat., Anat. Record, Vol. II.


Plate I. Lateral view of a wax plate reconstruction of a 10 mm. human embryo (No. 3, Huber collection), showing the origin and distribution of the peripheral nerves. The ganglionic masses are represented by darker and the fibre bundles by lighter shading. For purposes of orientation the diaphragm and some of the viscera are shown. The arm and leg are represented as transparent masses into the substance of which the nerve branches may be followed. The original model is enlarged 40 diameters and the drawing is about 12" diameters.


PERIPHERAL NERVOUS SYSTEM IN HUMAN EMBRYO

GEORGE L. STREETER


PLATE I


Veslcula auditiva Gang, acustlcum


Gang, radicis n.I.\ Gang, petrosuni

Gang, radicis n.X


N. frontalis



I Co. N. tibialis

N. peroneus


Tubus digest.^


Gang. Proriep


N. hypoglossus I.C

Gang, nodos.

-N. desc. cerv. -Rami byoid.

(Ansa hypoglossi) N. Musculocutan. N. axillaris >J. phrenicus N. medianus N. radialis -N. uluaris

I Th.


N. femoral

N. obturator


R. posterior

R. terminalis lateralis

R. terminalis anterior Mesonephros


Xn. ilioing. et hypogastr.


THE AMERICAN JOURNAL OF ANATOMY— VOL. VIII, NO. 3


Platk II. Median view of a model of the cTanial nerves in the same embryo shown in Plate I. A portion of the spinal cord is represented and above that everything is cut away, excepting the sensory bundles and motor nuclei of the different nerves, together with that portion of the marginal zone which is to form the funiculus anterolateralis. The somatic motor nuclei are colored red, and it can be seen that they form a column that is practically continuous with the cells of the ventral horn of the spinal cord. Enlarged about 30 diameters.


PERIPHERAL NERVOUS SYSTEM IN HUMAN EMBRYO

GEORGE L. STREETER


PLATE II



THE AMERICAN JOURNAL OF ANATOMY— VOL. VIII, NO. 3


Plate III. Reconstruction of the facial, glossopharyngeal and vagus nerve group in the same embryo shown in Plate I, showing the series of motor nuclei of origin and the tractus solitarius. It can be seen how the latter is formed by the sensory roots on entering the marginal zone of the neural tube, analogous to the dorsal funiculi formed by the dorsal roots of the spinal cord. A section of the neural tube is included in the reconstruction to show its relation to these different structures. Ganglionic masses are represented by lighter shading than the fibre bundles, and the point at which the rootlets enter the neural tube is indicated by dark rings. Enlargement about 35 diameters.


PERIPHERAL NERVOUS SYSTEM IN HUMAN EMBRYO

GEORGE L. STREETER


PLATE Mi


Tractus isolitarius


Nucl. mot. n.lX III', sens. IX et X


Xuel. mot. n.X (ambiguus)


Pars intermedins n.VII



Gang, genie


N. chorda tympan

X. facialis


-Lam. alaris


Lam. basalis


N. accessorins


(iaug. uodos. n. vagi Gang, pcti-os. n.IX


THE AMERICAN JOURNAL OF ANATOMY--VOL. VIII, NO. 3


ON THE ORIGIN" OF THE MESENTEKIC SAC AND THORACIC DUCT IN THE EMBRYO PIG.

BY

WALTER A. BAETJER.

From the Anatomical Lahoratort/ of the Johns Hopkins University.

The modern literature on the development of the lymphatic system may be said to begin with the publications of Ranvier^ in 1895. Prior to this time, the most widely accepted theory was that the earliest lymphatics existed as spaces in the mesenchyme, which gradually became confluent and formed lymphatic ducts. The first break from this idea came as early at 1868, when Danger^ published his observations on the lymphatics in amphibians. He noted that the lymphatics in tadpoles did not resemble tissue spaces, but had rounded endothelial-lined ends, which often had long strands of endothelium coming from them. This gave them the appearance of developing blood-vessels, and Langer interpreted his observations to mean that lymphatics grow like blood-vessels by the sprouting of endothelium.

Ranvier noted the same appearances in the developing lymphatics in tadpoles and in pig embryos and also interpreted them to mean that the lymphatic vessels grow by budding. This made the original theory seem even more doubtful and suggested the close association between the lymphatics and the veins. The proof of the venous origin of the lymphatics — suggested by Ranvier — was not definitely established, however, until the work of Dr. Sabin^ on pig embryos in 1902. In this study it was first shown that the lymphatic system begins by the formation of sacs or hearts, which arise from the veins. Four of these sacs were described — two in the neck and two in the posterior portion of the body, from which lymphatic vessels grew out and invaded the rest of the body.

^Ranvier. Comptes-Rendus de rAcademie des Sciences. Tome 121, 1894, 1895 and 1896, and Arch. d'Anatomie, 1897.

"Langer. Sitz. d. k. Akad. d. Wissensch., Bd. LVII, 1868.

Sabin. Am. Jouk. Anat., Vol. I. The American' Journal of Anatomy. — Vol. VIII. Xo. 3.


304 Walter A. Baetjer

This view Avas later confirmed by Dr. Lewis* in a paper on the "Development of the Lymphatic System in Eabbits," published in 1905. He worked out in more detail the origin of these sacs, showing them to come from transformed veins. He also stated that they arose in several places in the body and described two more sacs, viz., one in the root of the mesentery and one along the external mammary vein.

In the present work, which has been guided by Dr. Sabin, further evidence will be offered in favor of the direct venous origin of the lymphatics from a study more especially of the origin of one part of this system — the mesenteric sac. For this work serial sections were made of embryos from 16 mm. to 30 mm. long, in which the blood vascular system had been injected, and of embryos from 33 mm. up with lymphatic injections. From this time the primitive system is complete and can be injected through the thoracic duct without difficulty.

In the study of the origin of this system it has been found that there are certain primitive sacs existing in different regions of the embryo which represent the earliest lymphatics. As has been previously shown (Sabin, Lewis, McClure and Huntingdon) the first of these appears in the cervical region, near the internal jugular vein, in the pig, between 14 and 16 mm (Sabin). In a later stage — 21 mm. — according to the reconstruction, made by Lewis, definite lymphatic vessels have arisen in three regions (Fig. 8).^

(a) In the cervical region in association with the jugular veins.

(&) Along the vertebral column, dorsal to the aorta, in the exact location of the developing thoracic duct.

(c) In the root of the mesentery, just ventral to the renal anastomosis of the sub-cardinal veins. At this stage the lymphatics exist independently of each other and of the veins, as definite, well formed spaces in the embryonic s}Ticytium and lined by endothelium similar to that of the veins.®

Between these stages, in embryos 15 mm., in which lymphatics are found only in the cervical region, and in those 23 mm., where they exist

Lewis. Development of Lymphatic System in Rabbits, Am. Jour. Anat., Vol. V, No. 1, 1905.

^Lewis, Ameb. Jour. Anat., Vol. V, No. 1, 1905.

"In general, the study of my own specimens agrees with the reconstruction, though I think this condition occurs at a later stage — 23 mm. — since at 21 mm. the lymphatics in the region of the thoracic duct have not yet been differentiated. However, this difference may be due to variations in the methods of making measurements.


Mesenteric Sac and Thoracic Duct in Embryo Pig 305

in the three areas mentioned above — an embryo can be found (between 21 and 23 mm. in length) with lymphatics in the cervical region and in the mesentery before there is any evidence of similar vessels in the location of the future thoracic duct. At this stage and in all the succeeding ones up to 23 mm., all of the vessels dorsal to the aorta can be completely injected through the umbilical artery or vein, though the branches of the azygos veins are in the exact location of the thoracic duct which replaces them in the later stages from 23 mm. upwards. Below 21 mm., true lymphatics exist as such only in the cervical region, the mesenteric sac being represented by a plexus of capillaries in free communication with the veins. This is well shown in Figs. 1 to 4, which represent sections through the mesenteric sac. The capillaries at these stages are all engorged with blood and injection mass, and in all the sections large openings into the veins can be easily made out.

Before proceeding, however, to the consideration of the origin of this sac, which is indeed the principal object of this paper, several points may be set forth concerning the origin of the thoracic duct proper. This work, however, which is still incomplete, is to be the subject of a future article; therefore, only those points will be stated here which have a direct bearing on the main subject of this paper.

It has been shown that the lymphatics in the cervical region are derived directly from the veins, and this conclusion leads us naturally to expect that the primitive thoracic duct and mesenteric sac would in all probability be found to have a similar origin. This seems to be definitely established for the mesenteric sac; but for the thoracic duct, though all the evidence now at hand seems to favor a like origin for this branch of the lymphatic system, still it does not as yet seem as conclusive as could be desired.

The thoracic duct proper, or the part of the lymphatic system dorsal to the aorta, is not seen until the mesenteric sac is almost completely differentiated from the veins, this area being occupied by numerous branches of the azygos veins, all of which can be completely injected. Lewis, in his work on the development of the lymphatics, concludes that "lymphatic vessels develop along the course of the azygos veins apparently from independent venous outgrowths, all of which unite to form a continuous system, later acquiring new and permanent openings into the veins." The evidence which could be gathered from my own series, I think, in general, supports this conclusion, for sections taken from the same levels show all the vessels dorsal to the aorta injected in an embryo


306 Walter A, Baetjer

of 20 mm., while in the stage between 22 and 23 mm. the uninjected duct occupies the same relative positions. Although this evidence does not seem definite enough to warrant an empirical statement as to the origin, it is certainly suggestive, for such a rapid development could hardly be ascribed to new growth and is therefore most logically explained by assuming the presence of pre-formed channels which then become differentiated, such as will later be shown to be the case in the mesenteric sac. The evidence, then, in connection with the thoracic duct, is that it is preceded by a series of veins from which it is suddenly rather than gradually transformed into lymphatics. The question which seems to me of interest is, Does the receptaculum first form as one of the primitive sacs and the thoracic duct grow from it to meet the lymphatics growing down from the cervical region, or does the thoracic duct form from a number of segmental anlagen all homologous to the primitive cervical or mesenteric ? Dr. Lewis' figures would seem to suggest the second; but, although the primitive duct is not uniform in calibre and contains numerous dilated areas connected by much smaller trunks, I have been unable to find in any series a stage in which these dilated portions exist as independent sacs. However, it should be added that the connecting trunks are, in some cases, so narrow as to suggest this as a possibility. It will thus appear that there is no evidence whatever that the thoracic duct in the pig forms as Sala'^ has described for the chick,

Sala described the duct in the chick as forming out of solid cords of mesenchyme cells, in which a lumen subsequently developed. Thus, although all the evidence now available seems to indicate the direct venous origin of the thoracic duct, it is not, I think, nearly so conclusive as can be shown in the case of the mesenteric sac. .

This sac, located in the roof of the mesentery, between the Wolffian bodies, and just ventral to the renal anastomosis of the sub-cardinal veins, was first noticed by Dr. Lewis in following the transformations of the vena cava inferior in rabbits.^ In this paper, in his plate illustrating the vessels of this region, the lower portions of the sub-cardinal veins are detached from the rest, and "though colored blue, like the veins, they are described as spaces in the mesentery, suggesting the lymph-hearts of the chick." It is also stated that they may be subcardinal derivatives. In a later work by the same author on the "De 'Sala. Recerche a. lab. di Anat. norm. d. r. Univ. di Roma. A'ol. 7, 1900. 'Lewis. Ameb. Jour. Anat., Vol. I.


Mesenteric Sac and Thoracic Dnct in Embryo Pig 307

velopment of the Lymphatic System in Eabbits," he states that ^a portion of the sub-cardinal veins seems to become detached from the rest to form lymphatics," and "that some lymphatics in the mesentery accompanying the superior mesenteric and gastric veins may have arisen as branches of these veins." In his conclusion it is stated that "similar though smaller sacs than the jugular sac arise from the sub-cardinal and mesenteric veins at a slightly later date."

Thus, though the probable early venous connection of this sac is mentioned, no definite account is given of the time and method of origin, nor of its differentiation from the venous system and subsequent connection with the l}Tnphatics dorsal to the aorta — the thoracic duct proper. To trace the origin and development of this sac, serial sections were made of pig embryos in which the blood-vessels had been injected, the embryos ranging in size from 16 mm. to 30 mm.^ In an embryo of 16 mm. there is as yet no evidence of the blood capillaries which later form this sac, whereas at 30 mm. the sac has been completely differentiated from the venous system and is abundantly connected with the thoracic duct by large channels on each side of the aorta. This is well shown in Fig. 9, which is taken from an embryo of 30 mm. ; the aorta is seen suspended, as it were, in lymph.

Cross sections through this sac in an embryo 22 mm. long just after its differentiation, show shreds of tissue extending into the lumen and very irregular margins, presenting a picture highly suggestive of tlie fusion of many small vessels (see Fig. 7). The early stages were then cut to ascertain whether it was possibly formed by the coalescence of numerous capillaries and thus of direct venous origin, as had already been shown for the cervical lymphatics. This was found to be the case.

At 16 mm. no vessels exist in this region and the mesenteric attachment is a homogeneous network of embryonic connective tissue, while at 17 mm., the next stage examined, a very few small capillaries have appeared (Fig. 1). These are shown in Fig. 1 as small, injected vessels, lying in the root of the mesentery, just ventral to the renal anastomosis of the sub-cardinal veins. Their course is extremely short, running only the length of a few sections, and parallel to the long axis of the

"It has been found by H. Mc.L. Evans that all injections should be made through the umbilical artery, whether the embryos are dead or alive, since venous injections are never so complete and are much more likely to cause extravasations which are particularly prone to occur in this region. If the embryo has only recently died, the heart will usually commence beating again as soon as the injection mass reaches it.


308


Walter A. Baetjer


embryo. Toward the lower end they turn dorsalward, emptying finally into the renal anastomosis by distinct openings, as shown also in Fig. 1. This is quite in contrast with their mode of venous connection in the later stages in which there is definite fusion into much larger channels before their final termination (Fig. 3). Through the stage of 18 and up to 19 mm. there is a gradual increase in these small mesenteric veins;



Fig. 1. — Transverse section through the renal anastomosis of the sub-cardinal veins of an embryo pig, 17 mm. long, showing the small veins in the root of the mesentery, which are the anlage of the mesenteric sac A., aorta ; R. A., renal anastomosis ; W. B., Wolffian body ; M. C, mesenteric capillaries ; Mes., mesentery; O. A., genital anlage.

they become more numerous and of definitely larger calibre, while at the same time there is an increase in their length (Fig. 2). Fig. 2, taken from an embryo of 18 mm., shows well the initial steps in this process. The region occupied by the minute capillaries in the embryo of 17 mm. (Fig. 1) is seen to contain many more veins, all of which are definitely larger than those seen in the preceding section. A later and


Mesenteric Sac and Thoracic Duct in Embryo Pig


309


more pronounced stage in this development is shown in Fig. 3 from an embryo of 19 mm. During this time, also, there is a beginning of the process of fusion — a process which eventually leads to the formation of the sac in the later stages — by which many of the smaller vessels first unite to form larger channels before acquiring their venous outlet. Fig. 3 illustrates this process in its inception. In this the root of the mesentery contains many veins of much larger calibre formed by the coalescence of the capillaries seen in the earlier stages. Even now, how


Fig. 2. — Transverse section through the renal anastomosis of the sub-cardinal veins of an embryo pig, 18 mm. long. In this the small capillaries seen in Fig. 1 have become more numerous and of definitely larger calibre. A., aorta ; R. A., renal anastomosis ; W. B., Wolffian body ; M. C, mesenteric capillaries ; Mes., mesentery ; G. A., genital anlage.


ever, the fusion is not at all marked. The openings into the sub-cardinal veins are much larger than at 17 mm. (Fig. 1), and they are still numerous and quite distinct.

In all these stages, as well as in the succeeding ones, both before and after the complete separation of the sac, conclusive evidence for its venous origin is derived from a special study of the sections through the region of the meso-nephric arteries. Fig. 6 is taken from a section through this region in an embryo of 32 mm. In this the early mesenteric sac is seen occupying the area between these arteries; it is com


310


Walter A. Baetjer


pletely differentiated from the neighboring veins which are filled either with blood corpuscles or the injection mass. In the earlier stages, however, this area is studded with numerous small veins. It is thus evident that this structure, which later becomes an integral part of the lymphatic system, is represented in the earlier stages by a plexus of small veins.



Fig. 3. — Transverse section through the renal anastomosis of the sub-cardinal veins of an embryo pig, 19 mm. long. The veins in the root of the mesentery have fused to form the larger channels seen in this section. This is the beginning of the processes of fusion which leads to the formation of the mesenteric sac in the later stages. At this stage, however, these vessels can all be traced definitely into the renal anastomosis. A., aorta; R. A., renal anastomosis ; W. B., Wolffian body ; M. C, mesenteric capillaries ; Mes., mesentery ; G. A., genital anlage.


Up to this time, i. e., 19 mm., the increase both in the number of capillaries and the amount of fusion, has been very gradual. From now on, however, the development in both of these phases goes on with great rapidity; in fact, the active process in the formation of the sac may be said to begin here. Between 19 and 20 mm. the number, calibre and lensrth of these vessels become markedly increased : there is much


Mesenteric Sac and Thoracic Duct in Embrvo Pig


311


greater fusion with tlie formation of several definite channels bearing a close resemblance to the sac in the later stages; but still having large openings into the veins. This is shown in Figs. 4 and 5; in Fig. 4, representing a section through the renal anastomosis of the sub-cardinal veins, in an embryo of 20 mm., the root of the mesentery is occupied by



Fig. 4. — Transverse section through the renal anastomosis of the sub-cardinal veins in an embryo pig, 20 mm. long. In this section the venous channels in the root of the mesentery are beginning to show definite evidences of fusion and sac formation, though they are still connected with the veins, as is shown in the figure. A., aorta ; R. A., renal anastomosis ; W. B., Wolffian body ; M. C, mesenteric capillaries ; Mes., mesentery ; O. A., genital anlage.


large venous channels between which there are extensive communications. At this stage there are still numerous openings into the veins from which the vessels can all be definitely injected, but not as easily as in the earlier embryos, since the venous connections are now being gradually obliterated. One of these openings is shown in the figure, with the ink entering the sac. In Fig. 5, taken from a section a little lower down in


312


Walter A. Baetjer


the same embryo, the sac is even more definitely formed. It also shows that the sac is still injected from the vein. The drawing shows the beginning of the mesonephric arteries arising from the aorta, the area between them being occupied by the large channels still definitely connected with the veins, whereas, in Pig. 6, from a section through the


4 fS >?SWyA^is^^i/m^i%^:w^iif' ■n,A'^si^Ajt^/^hH



Fig. 5. — Transverse section through the beginnings of the mesonephric arteries in an embryo pig, 20 mm. long. The section shows the beginnings of a definitely formed sac which, however, can still be traced into the veins. The venous connection at this stage, however, is gradually becoming obliterated. A., aorta ; P. C F., post-cardinal veins ; M. C, mesenteric capillaries ; Mes., mesentery ; W. B., Wolffian body ; li. A., beginnings of mesonephric arteries.


same level in a 22 mm. embryo, the sac has lost almost all venous connection and has become a definite part of the early lymphatic system.

This active process of development continues between 20 and 21 mm. until, at the latter stage, there is formed a definite sac with irregular margins, due to the coalescence of numerous capillaries; and similar in


Mesenteric Sac and Thoracic Duct in Embryo Piff


313


appearance to that seen in Fig. 7, though this is taken from an older embryo. At the 21 mm. stage, however, there are still definite venous openings, although they are gradually becoming lessened by a gradual process of differentiation which becomes complete between 22 and 23 mm., before the appearance of any uninjected vessels in the region dorsal to the aorta, thus leaving the sac, for a short time, independent of either venous or lymphatic connections (Figs. 7 and 8). Fig. 7 shows this sac, with its irregular margins, located in the root of the mesentery just




m




^21,





-^jy


Fig. 6. — Transverse section through the mesonephric arteries in an embryo pig, 22 mm. long. The mesenteric channels at this stage were uninjected, as shown in the figure, while the adjoining veins were filled with blood or injection mass. The venous connections of the sac have been entirely obliterated. A., aorta ; W. B., Wolffian body ; S. C. V., sub-cardinal veins ; M. 8., mesenteric sac ; Mes., mesentery ; M. A., mesonephric arteries ; P. G. V., postcardinal veins ; G. A., genital anlage.


ventral to the sub-cardinal veins, that is, in the exact location of the venous plexus in all the earlier stages. It has lost all venous connections and has not yet acquired any communication with the rest of the lymphatic system, but exists for a short time, as an independent sac in the embryonic mesenchyma. Fig. 8 is a section from a lower level in the same embryo and shows the beginnings of the mesonephric arteries with the area between occupied by an independent sac which has replaced the large venous channels of the preceding stages.


314


Walter A. Baetjer


Between 23 and 25 mm. the thoracic duct appears, extending at first from about the level of the renal anastomosis up to a point approximately opposite the arch of the aorta. It is bilateral in the thoracic region, but fuses below the diaphragm to form a much dilated channel, just dorsal to the aorta — probably the primitive receptaculum. From 35 mm. on the sac gradually develops its connection with the rest of the



Fig. 7. — Transverse section through the renal anastomosis, in an embryo pig, 23 mm. long. This is the first appearance of a definite sac in the exact location of the venous plexus in the earlier stages. It will be noticed that the irregular margins suggest the fusion of many small vessels. At this stage no connection can be traced between the sac and either the lymphatic system or the veins. A., aorta ; R. A., renal anastomosis ; M. f>., mesenteric sac ; W. B., Wolffian body.


lymphatic system in the following manner: small capillaries grow from the sac and primitive receptaculum; these are shown in the earliest stage in Fig. 8 as small vessels extending up from the sac, on each side of the aorta, where they eventually meet and anastomose, finally forming definite channels along the lateral margins of the aorta, complete by 30 mm. (Fig. 9) from which stage they can be definitely injected by way of the thoracic duct.

Prior to this time, however, when the embryo is between 25 and 26 mm. in length, communication is established between the early thoracic


Mesenteric Sac and Thoracic Duet in Enibrvo Piff


315


dnct and the cervical lymphatics, so that when the embr3'o reaches 30 mm. the three integral parts of the lymphatic system have all become united to form a completed duct. The connection between the duct and the cervical lymphatics is established as follows : In an embryo between 25 and 26 mm. the duct, which is at first bilateral, extends up to a point about opposite the arch of the aorta, where the two main trunks become fused. The cervical lymphatics, by this time, liave orown down




Fig. 8. — Transverse section through the beginnings of the mesonephric arteries in an embryo pig, 23 mm. long. The figure shows the definite mesenteric sac in the region between the mesonephric arteries which was occupied by the venous plexus in the earlier stages. A., aorta ; M. 8., mesenteric sac, showing the upgrowth of capillaries around the aorta ; S. C. V., sub-cardinal veins ; Wl B., Wolffian body ; G. A., genital anlage ; M. A., mesonephric artery.


to this same level and connection between the two is established by small anastomosing channels, scarcely larger than capillaries. Gradually, as development takes place, these become distended, until a duct of uniform calibre is formed when the embryo reaches a stage, above 33 mm., from which time complete injections of the lymphatic system can be made without difficulty. Such an injection is shown in Fig. 10, made from a cleared specimen 5.5 cm. long.

This specimen gives a good idea of the primitive lymphatic system. The cervical h^mph sac is shown now turning into lymph nodes. The


316


Walter A. Baetjer


thoracic duct appears as a plexus of vessels along the aorta in which one can see two definite channels, one on either side. Or one may say that the duct is double with numerous anastomoses. Between the kidneys and reproductive organs is the large mesenteric sac from which vessels



Fig. 9. — Transverse section through the early cisterna chyli and mesenteric sac in an embryo pig, 3 cm. long. The section shows the connection of the receptaculum chyli and the mesenteric sac, by large channels along the lateral margins of the aorta. A., aorta; C. C, cisterna chyli; E., kidney; M. 8., mesenteric sac; /., Intestine; P. C. V., post-cardinal vein.


spread to the capsule of the kidneys. The cisterna chyli is directly behind the mesenteric sac, separated from it by the aorta and hence cannot show in the figure. The posterior lymph sacs are opposite the lower end of the Wolffian bodies, and from these sacs vessels are seen passing to the legs.

To sum up briefly, then, the following conclusions may be noted :


Mesenteric Sac and Thoracic Duct in Embryo Pig 317

1. The first lymphatics appear in the cervical region near the internal jugular vein, in the pig, between 14 and 16 mm. (Dr. Sabin).

2. The thoracic duct appears first at 23 mm., and is located in the exact position of the branches of the azygos veins in the earlier stages. In all probability, it is derived from branches of these veins.

3. The mesenteric sac begins as a mass of blood capillaries, lying ventral to the renal anastomosis of the sub-cardinal veins. They become fused into a definite sac at first having extensive venous openings, which later become gradually obliterated, thus making the sac independent at one time of either venous or lymphatic connections.

4. By a growth and fusion of capillaries from both the sac and the lower portion of the thoracic duct proper, the former becomes a definite part of the lymphatic system.


Fig. 10. — From a cleared specimen of a lymphatic injection in an embryo pig 5.5 cm. long. The two main ducts in the thoracic region are seen to fuse into single channels both below the diaphragm and above the arch of the aorta. Numerous channels are seen crossing the aorta. For the lymphatics below the diaphragm, the drawing is slightly misleading, for the specimen was exceedingly transparent, and includes lymphatics of two levels. This complicated area will be made more clear in a paper of Dr. Heuer's soon to be published in this Journal. Between the two kidneys and Wolffian bodies is the large mesenteric sac, lying in front of the aorta ; the leader marked C.C. runs to this mesenteric sac. The cisterna chyli dorsal to the aorta, is entirely concealed by the mesenteric sac. The ducts running anteriorly from the posterior lymph sacs empty into the cisterna chyli dorsal to the mesenteric sac ; they do not empty into it as they appear to do in the figure. C, cervical lymph sac ; C. C, mesenteric sac dorsal to which is the cisterna chyli ; D., primitive thoracic ducts ; W. B., Wolffian bodies ; T., testes ; K., kidneys ; P., posterior lymph sac.


MESENTERIC SAC AND THORACIC DUCT IN EMBRYO PIG

WALTER A. BAETJER



Fig. 10


THE AMERICAN JOURNAL OF ANATOMY— VOL. VIII, NO. 3


THE GROWTH OF THE BRAIN AND VISCERA IN THE SMOOTH DOGFISH (MUSTELUS CANIS, MITCHILL)

BY

WILLIAM E. KELLICOTT

From the Biological Laboratory, The Woman's College of Baltimore, Md.

Current ideas regarding the growth of animals are based to a considerable extent upon observations on the weight or length of the entire organism. And yet it is well known that the proportions of many external parts regularly change during growth and that at least one internal organ, the brain, does not increase in weight at the same rate as the whole animal. The question comes quite naturally then whether other parts than the brain may not also have their own rates or cycles of growth which may differ more or less from the type of growth given by the entire organism. Physiologists are showing- that in many instances certain organs or tissues may show regular, sometimes recurrent, growth cycles quite independent of the growth of the body as a whole and we are led to inquire whether the normal growth of an animal may not be actually a complex of growth cycles of component parts. It is quite possible to examine this question from the morphological as well as from the physiological side, and the present paper represents an attempt to discover whether the brain and viscera of the dogfish grow similarly or in diverse ways, as somewhat independent units of growth.

We are remarkably deficient in our knowledge regarding the normal growth of the viscera or of parts of the body other than the brain. Doubtless much valuable information of this kind regarding man lies concealed in hospital and clinical records. But, as far as I have been able to discover, the meagre data collected by Welcker and Brandt, '02, and by Vierordt, '06, represent the extent of our knowledge regarding the growth of the viscera in man and other vertebrates.

It seems that this lack of information concerning the growth of parts has led to a partial misconception of what is involved in the growth of

The American Journal of Anatomy. — Vol. VIII, No. 4.


320 William E. Kellicott

the animal body, and that the growth of one or two bulky tissues has been mistaken often for the gro^Yth of the organism. Further, some of the problems of growth have been overlooked because of the failure to bear in mind the ph3'siological distinction between determinate and indeterminate growth among animals. The birds and mammals are unusual among vertebrates in that they continue to live for a long time after maturity without continuing to grow meanwhile. This is certainly true of man and seems to be true of the other mammals ; and yet the mammals are the forms whose growth has been studied most extensively and from which many fundamental conceptions of growth have been derived. The problem of growth is somewhat different among all of those lower forms wdiose growth is indeterminate and continues, though slowly to be sure, throughout life, and which represent a more primitive condition in this respect. The present paper presents a fairly complete series of data regarding the increase in weight of the brain and certain of the viscera in a form which grows indeterminately, and suggests a possible interpretation of their significance for the general problems of growth.

SuM:\rARY

A series of 315 dogfish (Mustelus canis) including specimens from birth (length 32.8 cm., weight 7G.2 grams) up to a female 135.1 cm. in length, weighing 8434 grams, has been examined with a view toward getting precise information regarding the relative growth of the brain and viscera. Accurate w^eighings were made of the brain, heart, rectal gland, pancreas, spleen, liver and gonads. It is found that these viscera as well as the brain are relatively largest in the smaller individuals and that they decrease in size relatively throughout life although they do not cease to increase in actual weight. The period of maximum relative size of these parts is about that at which the growth of the whole organism is most rapid.

While the ages of these fish can not be determined exactly it seems that the dogfish of one meter in length, weighing approximately 2750 grams, is probably about five years old. Females are not ordinarily mature before this size is reached, although the males may mature considerably before this. Absence of the time factor in these observations is not important because their factors, such as food and temperature, are known to be of more importance than age in determining the size of fish.


Growth of Brain and Viscera in Dogfish 321

The weights of the different organs are plotted for each individual separately and smooth curves derived. Each organ seems to have its own rate of growth more or less independent of the others, only the general features of its growth being adapted to that of the entire organism.

While the weight of the whole animal is increasing by the addition of definite equal amounts the heart also increases by the addition of equal increments, but the other organs, except the gonads, increase by the addition of gradually decreasing increments. Peculiarity in the growth of the liver is explained by the accumulation of fat in this organ. Peculiarities in the growth of the gonads apparently are due to the fact that at first these organs do not have a reproductive function; later a second cycle of growth appears which coincides with the approach of sexual maturity during which these organs are actually reproductive.

The sexes can not be distinguished with respect to either the absolute or relative weight of these parts, excepting the gonads.

It is not the organism then, but the organ or tissue that is the growing unit, the growth of the organism being a composite resultant of the growth of its parts. The muscles and supporting tissues form about 75 per cent of the total weight, and it is chiefly the increase of these parts which is measured when the weight of the total organism is taken as the basis for describing growth. The growth of these predominating tissues masks the differing rates of growth of parts of co-ordinate importance though of lesser bulk and an erroneous conception of the growth of organisms results.

Comparison is made with other data and certain general conclusions suggested. The muscles and supporting tissues seem to outgrow the brain and viscera, a relation leading ultimately to a loss of physiological balance within the organism. We should regard the condition of determinate growth found in the birds and mammals as secondary and as arising from the primary condition of indeterminate growth as an adaptation such that the muscles and connective tissues cease their growth while the more slowly growing brain and metabolizing organs are still competent to carry on the work of the whole mass of the organism.

The hypothesis is suggested that the regulation of the normal growth of individual tissues or organs may be carried on through specific internal secretions as is known to be the case with the growth of certain occasionally developing organs or in certain pathological growth phenomena. The evidence here given is purely morphological and therefore indirect.


322 William E. Kellieott

Material and Methods

The material from which these data were drawn was secured at the Laboratory of the Woods Hole Station of the Bureau of Fisheries.^ The data were collected with quite another purpose in view, but the present questions have arisen and may be considered at this time.

The data are given in full in Table III, pp. 343-350.

The determinations were made with the accuracy usual in statistical work; the weighings were made to 0.01 gram. The organs determined were the brain, heart, rectal gland, pancreas, spleen, liver and gonad. These parts were removed within one-half hour after the fish ceased to react to touch, and in practically every case the heart was still beating when it and the other organs were removed for weighing. The attachments of the viscera were cut through along the surfaces of the organs and the organs rolled gently in a towel until the blood was expressed. The brain was sectioned from the cord in situ, in a transverse plane extending just along the posterior tip of the cerebellum. Anteriorly the olfactory tracts were cut off along the contours of the hemispheres. The infundibulum and pituitary body were included, but the cranial nerve roots were removed along the surface of the brain and the outer membranes removed. This of course does not give a complete brain, but the limits chosen seemed to give the best compromise between completeness of l)rain and rapidity of removal. This last factor might be one of considerable importance on account of the rapidity with which the weight of the brain changes after death. In the heart no satisfactory landmark could be found along which to section the sinus venosus from the veins nor from the auricle: consequently the entire venous end of the heart was removed along the auriculo-ventricular groove which is very definite. Anteriorly the heart was sectioned at the junction of the conus and truncus. What is referred to as "heart" includes therefore only the ventricle and conus arteriosus. These parts make up by far the greatest part of the weight of the heart and at the same time are the chief functional elements in the fish. The rectal gland was removed as nearly as possible along the line where its duct commenced.

In preparing the plates the weight of each organ in each individual was recorded separately. The curves were derived from these records

'I take pleasure in acknowledging my indebtedness to Commissioner George M. Bowers, of the Bureau, and to Director Francis B. Sumner, of the Laboratory, for the privileges of a research room and a large share of the dogfish material collected during the seasons of 1906 and 1907.


Growth of Brain and Viscera in Dogfish 323

by calculating a series of average weights of each organ in successive groups of individuals, and the line formed by connecting these averages was then smoothed to a curve so as to reduce to a minimum the plus and minus deviations of the averages. The groups from which the averages were derived varied in extent from 100 to 1,000 grams in different regions of the entire group, according to the rate at which the character of the curve was changing.

The fish themselves were taken, with very few exceptions, from a single trap in Buzzards Bay, and in nearly every case they were examined within twenty-four hours after their removal from the trap : during this interval they were kept in tanks abundantly supplied with running water. A few specimens had been kept for a few days before weighing in a large floating car in very favorable conditions. In determining total weights the fish were dried off with a towel and the weight of stomach contents subtracted: the ovaries and oviducts, even when containing embryos, were included.

The total number of fish examined was 315 (176 females, 139 males). Of these, the brains of 65 and the gonads of 15 were not weighed. These fish did not form a "random sample," but were of sizes selected so as to give as complete and uniform a series as possible. Practically all of the larger individuals taken were examined, but only a small proportion of the smaller and medium sizes. Consequently many of the usual statistical constants ernployed would be of no comparative value as descriptive of this group and they have not been determined. As may be seen from the plates the fish ranged in size from those just born, having an average weight of 76.2 grams and average length of 32.8 cm., up to a female weighing 8434 grams, length 135.1 cm. The specimens examined at birth, of which there were thirteen, were born in the laboratory from a single female (I believe this is the maximum number of young recorded for this fish) and it should be borne in mind that in speaking subsequently of the condition of certain organs "at birth," these specimens alone are referred to. Since these were all obtained from a single female and that the largest taken during two seasons' collecting, the weights found may not represent accurately the precise average condition at birth.

The series of fish examined shows, then, an increase in weight of over 110 times and in length of over four times. From a small number of fish collected wholly at random the average weights and lengths were found to be about, males, 1050 grams, 73 cm. ; females, 1800 grams, 87


32-t William E. Kellicott

cm., or, including both sexes, about 1425 grams, 81 cm. : probably these figures are fairh' close to the actual averages. It is to be regretted that the exact ages of these fish can not be told. I have been unable to find any statement regarding the time rate of growth of any of the Elasmobranchs. While we may nevertheless continue to use the expression "rate of growth" in speaking of the brain and viscera, it must be remembered that not the time rate but the comparative rate of growth is meant, the weight of the entire organism serving as the basis of comparison.

The Data

The Brain. — At birth the average weight of the brain is found to be 0.855 gram. After birth, as shown in Plate 1, its weight increases rapidl}', but at a slightly diminishing rate, so that in fish of about average (not middle) size the brain weighs approximately 3.5 grams. Among the laro-er individuals the diminution is much slower but is continued, though the growth of the brain does not cease during life. The heaviest brain weighed 7.2 grams. While the total weight had increased over 110 times the brain had increased but 8.4 times. The curve shows that this heaviest brain was somewhat larger than the expected weight, which would l)e only about 6.4 grams, giving an increase of 7.5 times.

The curve showing increase in absolute weight of the brain is, however, less significant than that of its relative weight: this also is shown in Plate 1. At birth the ratio of brain weight to body weight is very high — 1.116 per cent. But this falls very rapidly while the animals are increasing up to 300 or 400 grams. Then gradually the ratio decreases until in fish of average size it is only 0.25 per cent. This decrease continues at a diminished rate throughout life, so that it is still falling in the largest specimens examined: in the largest it was but 0.085 per cent. In other words the brain of this largest fish was, compared to the total weight, only one-thirteenth as large as in those just born.

It is not possible to distinguish between the sexes with respect to brain weight. The heaviest male examined weighed 3010.5 gi-ams, or considerably less than one-half the heaviest female, but the weights of the brain, both absolute and relative, in males and females of the same total weight were not sensibly different, as can be seen from inspection of the plates where the sexes are charted distinctively. The curves given by the brain weights are remarkably smooth and the individuals grouped closely about the curve.


Growth of Brain and A'iscera in Dogli^h 325

The Heart. — At birth the average weight of the heart is 0.078 gram. Almost from this time the increase in weight is uniform. Plate 3 shows this as the only instance among the organs measured of a perfectly regular increase in weight, the line of absolute weights being a straight line after passing the region representing a total weight of about 170 grams. This means of course that equal increments are added successively as the total weight increases by the addition of equal increments. The heaviest heart weighed 6.75 grams, showing a total increase in the series of 84 times.

The curve of relative weights is not so simple. At birth the ratio is 0.11 per cent. Just after birth the heart grows more rapidly than the entire organism, and this relation continues until the total weight is 160-170 grams — ^double the weight at birth. This maximum ratio is about 0.12 per cent, a comparative increase of about 9 per cent. (One individual shows a ratio of 0.133 and two others 0.130.) From this maximum the ratio decreases, at first rapidly then more slowly, until in fish of average size it is only about 0.087 — considerably less than at birth and about two-thirds the maximum. M with the brain the ratio continues to decline steadily though slowly throughout the series, and in the largest fish falls to O.OS per cent. That is, compared to the total weight, the heart of the largest fish is only two-thirds as large as at its maximum size or less than three-fourths as large as at birth.

There is no distinction between the sexes in either absolute or relative weight of the heart.

The Rectal Gland. — The average weight of tlie rectal gland at birth is 0.0304 gram, the heaviest weighed 1.62 grams — an increase of over 53 times. Between these two extremes we find, as shown in Plate 3, the weights distributed along a nearly straight line much as with the heart except that there seems to be a tendency among the largest specimens for this curve to rise somewhat more rapidly.

The curve given by the relative weights of this gland is of much the same character as that of the brain, and of the heart after its maximum point. At birth the ratio averages 0.0398 per cent. At first this falls rapidl}^, then more slowly to a ratio of about 0.0225 in fish of average size, and after this much more slowly to 0.0192 per cent in the largest individual. In this specimen the rectal gland is relatively a little less than one-half as large as at birth. After a weight of 2500 grams is reached this curve drops scarcely at all, the line being practically parallel with the base.


336 William E. Kellicott

Here again there seems no certain distinction between the sexes, although it is possible that in the largest males the rectal gland is slightly smaller than in females of corresponding sizes. The plate shows that in males of 2000-3000 grams the records for the males are mostly below those of the females, although to be certain of a difference here a larger number should be examined.

The Pancreas. — The average weight of the pancreas at birth is 0.061 gram: the maximum weight observed is 7.78 grams. The distribution of weights between these extremes is quite uniform, but Plate 4 shows that this is along a slightly curved line. The total increase in the pancreas weight was more than 127 times.

The curve of relative weights is quite similar to that of the heart. At birth the ratio is 0.08 per cent; it increases very rapidly until the total weight reaches about 200 grams when it reaches a maximum of 0.137 per cent. Then falling more slowly than the heart it is about 0.105 in fish of average size, higher still than at birth, and finally drops to probably about 0.075 per cent. The largest specimen measured showed an actual ratio of 0.092, but inspection of the curve shows that this is considerably in excess of the expected average as indicated by fish of 5000 to 7000 grams, among which it averages 0.0787 per cent. We may therefore assume about 0.075 as the probable final ratio for comparative purposes. This is not far below the ratio at birth, but it is only about one-half the maximmn.

There is no distinction betAveen the sexes in either absolute or relative weight of the pancreas.

The Spleen. — At birth the average weight of the spleen is 0.097 gram. Even at this time there is indicated the condition of great variability which continues throughout the series. Thus at birth the extremes in weight are 0.06 and 0.15 gram among individuals whose total weights vary only between 69.5 and 84.0 grams. And we find that while the curve given by the serial weights is regular — see Plate 5 — the individual weights are distributed about it much less closely than in any of the organs yet considered. We find the spleen of the heaviest fish is not the largest absolutely. The largest spleen weighed 12.36 grams, and was found in a female weighing 3638 grams, less than half the weight of the heaviest fish. This was an exceptionally large spleen and for purposes of comparison we should consider rather the condition shown by the curve to be more typical. Thus at 8400 grams total weight the spleen averages about 9.75 grams, giving an increase in the entire series of more than 100 times.


Growth of Brain and Viscera in Dogfish 327

At birth the ratio of spleen weight to total weight is 0.126 per cent. This ratio rises very rapidly just after birth and in fish of about 200 grams it reaches an average of about 0.475, that is it has nearly quadrupled. This is the most rapid initial rise observed in any of these organs. The extreme variability is again shown by the range of the ratios among individuals of approximately 200 grams, the limits being 0.257 and 0.895; this latter ratio is about six times the maximum ratio found at birth. From this point the ratio diminishes gradually throughout the series much as in the pancreas, falling in fish of average size to 0.3, and finally among the largest specimens to 0.106 per cent, or slightly less than at birth and only about one-fourth the maximum average ratio.

There is no distinction between the sexes in either the absolute or relative weight of the spleen.

The Liver. — The weight of the liver at birth averages 2.4 grams. As with the spleen the variability of this organ is so great that the condition in a single specimen may differ very considerably from the condition shown by the examination of a large number to be typical. Plate 6 shows that at 8000 grams the average weight of the liver is about 365 grams. This would show an increase of 152 times. The liver of the largest fish weighed 339.7 grams, while the heaviest liver weighed 557 grams, and was from a fish weighing 6811 grams. Between the extremes of the series the rate of increase was seen not to be quite regular, but gives a rather complex curve unlike that given by any other organ examined.

At birth the average ratio of liver weight to total weight is 3.12 per cent. This rises rapidly to a primary maximum of 5.5 per cent at a total weight of about 200 grams, the usual point of these maxima. It then declines as usual, but ceases when about 4.2 in fish of 700-800 grams and then commences to rise a second time, more slowly now, to a second maximum of 7.4 per cent in fish of about 3700 grams. From this point the ratio falls again until, among the six largest individuals it reaches about 5.5 per cent — about the same as at the first maximum. As will be pointed out later it seems possible to explain the peculiar form of this curve in such a way as to show that the liver really does not differ essentially from the other organs described with respect to its comparative rate of growth.

There is no difference between the sexes in either the absolute or relative weight of the liver.

The Gonads. — At birth the gonads of the two sexes are not unlike in weight either absolutely or relatively. The average weight is 0.274


]28 William E. Kellicott gram, and the average ratio to total weight is 0.358 per cent. But from birth onward the two sexes become entirelj^ distinct in both respects and therefore must be described separately.

Male. — In the male the weight of the gonads increases rapidly as shown in Plate 7 to a maximum average of about 29 grams : the heaviest testes weighed 31.8 grams. The total increase is nearly lOG times.

The relative weight of the testes increases rapidly after birth from 0.358 to a first maximum of 0.775 per cent among males weighing about 400 grams. Then after falling to a ratio of 0.60 at about 900 grams it recovers and rises rapidly to a ratio considerably higher than the first maximum and then more gradually to a final maximum ratio of 1.15 (1.03 as averaged from the eight heaviest males). This ratio of 1.03 per cent is nearly three times that at birth and one-third higher than at the first maximum. This is the only organ measured whose weight continues throughout the life to increase relatively to the total weight.

Female. — From the time of birth the gonads of the females increase at a somewhat slower rate than do those of the males. The rate is only approximately uniform, so that the line given by plotting the weights (Plate 7) is not a simple curve. The gonad weights of the larger females are not strictly comparable with each other because, at the season when they were weighed (July and August), some of the ovaries still contained one or more large yolk-filled ova, while from others the mature ova had been completely discharged so that the weight of the ovary was considerably altered. In Plate 7 the ovaries that contained large ova are enclosed in small circles. Only one female of less than 3600 grams (the exception weighed 2266 grams) was found with large eggs in the ovary. Similarly only one of less than 3250 grams (the exception weighed 1998 grams) contained developing embryos in the oviducts, while all of the 26 specimens observed above this weight contained embryos. The females of this dogfish therefore do not become mature until they have reached a weight of at least 2000 and usually 3000 grams; the length of such fish is roughly 90 and 100 cm., respectively, and their age probably four or five years.

The relative weight of the ovary rises like that of the testes to a primary maximum in fish of about 400 grams. This maximum is not so high as in the male, being only 0.675 per cent. This is followed by a fall to a ratio of 0.43 in fish of about 1700 grams. After remaining at about this point for some time the ratio tends to rise a second time as in the male, though not to the same extent. Probablv this


Growtli of Brain and Viscera in DoiiTi^li 329

tendency to rise a second time here is periodic and seasonal. For if we consider the distribution of only the ovaries containing large ova we see a considerable rise in relative weight, but the weights of the ovaries from which all mature ova have been discharged do not give an increasing ratio but one that is practically uniform. Probably therefore this general curve of relative weights is a composite and should be resolved into two components as in Plate 7, Curves A and B.

Discussion of the Data

Before attempting to interpret these data or to compare them with other more or less similar data from other forms we may compare the data concerning the various organs among themselves.

Of course these curves do not contain a known time factor and this must be borne in mind during their consideration and in comparing them with other data. The measurement of a large number of dogfish taken at random might give us some information regarding the average time rate of growth for the first few years of life, but so far as I know we have no such data. Of the specimens which I examined the first eighty were collected at random, since they included practically all of the fish taken in almost daily hauls of the trap for six weeks. Upon plotting these fish according to weight and length we get curves w^ith several maxima in their early parts, indicating the total weight and length with age as follows :

Age

Birth

1 Yr

2 Yr

3 Yr


Weight IN Grams

75


Length IN Cm.

32.5


300


45.0


750-800


63.0


1400


78.5


200-2250


90.0


2750


99.0


There is considerable overlapping, and after the fifth year the individuals are so scattered that they form a more or less continuous and irregular line. The distribution commences to become uneven after what is apparently the second year. It can not be told with much probability whether or not the sexes grow at the same rate, though it seems likely that the males at first grow faster than the females, but after the third


330 William E. Kellicott

year this relation is reversed. Probably the actual average weights and lengths at different ages would be found to differ somewhat from these figures, but this is the best approximation I can make with the available data. This is precisely the type of distribution which has been found repeatedl}' in the measurement of teleosts. (See, for example, Moenkhaus, '95, and Fulton, '01, '06).

It is undoubtedly true that here, as in the teleosts observed, age is not the chief factor in determining size, but that, as in so many invertebrates, the factors of food and temperature are the most important. Therefore to relate total size with age in these and similar forms would not add much to the interpretation of these size relations. As affecting the size of the brain and viscera probably the age is of still less importance, so that for these organs we are upon a more instructive basis in considering their weights as compared with that of the entire organism without reference to age. Consequently the absence of a known time factor is not the serious matter it might seem at first thought or probably would be were we considering a higher vertebrate such as a bird or mammal.

As compared with these curves in which total weight irrespective of age is alone the basis of comparison, the curves given by the time rate of increase would differ in their precise form, since different lengths of time would be required to cover what are here indicated as equal distances along the base line. If the time element could be introduced the curves would probably be shortened chiefly in the region between the total weights of 1000 and 2000 grams and somewhat less below and above these weights, and then lengthened considerably from the middle part onward.

Considering in the first place, then, the absolute weights of the organs measured, we find, first, that while the curves are all of essentially the same character, yet, second, each is distinguished by certain details of form, and, third, excepting only the liver and gonads, they are all of simple character with no indication of being formed by two or more dissimilar elements or cycles. No attempt has been made to describe any of these curves mathematically. With the exception of the curve of gonad weights which is obviously of special character, it is possible to arrange the curves in a series between the members of which the differences are minimal.

Commencing with the spleen (Plate 5) we have a curve which is distinctly concave to the base line throughout its entire course, al


Growth of Brain and A^iscera in Dogfish 331

though the concavity decreases considerably toward the upper end. The curve most nearly similar to this is that of the brain (Plate 1) in which the concavity is well marked, but does not extend over the entire curve, the upper end being straight though at an angle with the base. Following this we have the curve of the pancreas (Plate 4), which is considerably less concave and in which the entire upper part has become straight. This tendency to straighten is continued in the curve of the rectal gland (Plate 3) where only the very beginning shows any concavity and that very slight. Finally we reach the condition found in the heart (Plate 2) where the line becomes straight almost from its very beginning. The curve given by the liver weights (Plate 6) is more like this last than any other, but is not simple in character for reasons to be mentioned presently. This tendency for the concave curves to straighten is carried to an extreme in the gonad curves (Plate 7) which actually become convex to the base; this is most pronounced in the curve of the testes.

Of course the alteration here of a concavely curved line into a straight line indicates that an organ passes from a condition in which it grows by the addition of constantly diminishing increments, to one of growth by the addition of successively equal increments, the total weight meanwhile increasing in either case by the addition of equal amounts. The spleen, for example, throughout its growth increases by the addition of constantly decreasing increments, while the heart almost from the first grows by the addition of equal increments. The other organs show the various intermediate conditions described. The obliquity of this line to the base is determined by the size of the increment added, just as the amount of the curvature is determined by the rate at which the increments are decreasing. The rectal gland is the only one of the viscera except the gonads which gives a curve becoming convex to the base, and this occurs only at its extremity, this means that this organ increases, among the largest females, by the addition of increments of increasing size.

We may conclude then that these curves though all of the same general type, are individually distinct so that the rates of increase in weight of these organs are not determined by precisely the same factors or, at any rate, by factors operating with equal intensity at corresponding periods throughout this series of organs. This is shown also by the curves showing the relative weights of these organs which seem more instructive and which aid in the possible interpretation of these curves just described.


332 William E. Kellicott

These curves of relative weight are again all of the same general type though individually distinct. The heart, pancreas and spleen (Plates 2, 4, 5) are essentially alike in showing first a rapid increase in relative weight followed by a somewhat less rapid fall and then a long and slow but steady decline through to the end of the series. Individually these differ in the extent of the initial rise, in the rate and duration of the first and second phases of decline and in the total amount of decline. In the brain and rectal gland (Plates 1, 3) the initial phase of rapid rise is omitted, the entire curve resembling in general the declining parts of the preceding curves. Of all the curves, that of the rectal gland shows the most nearly horizontal line in its latter part. The first parts of the curves given by the liver and gonads (Plates 6, 7) are similar to those first mentioned in showing the initial rapid rise and similarly this is followed by a considerable decline. This decline, however, is checked and the curves soon become somewhat irregular, the testes showing a very pronounced rise, the ovary and liver considerably less.

It is hardly necessary to point out that the relation between these curves and the base line of body weights expresses the relation between the relative rates of growth of the particular organ and the entire body. Should these lines be parallel the two are growing at relatively equal rates; when the curve rises or falls the particular organ is growing respectively faster or slower than the body as a whole. The growth of an organ by the addition of constantly equal increments, as described above, would give a curve or line, depending upon the actual size of the increment, tending to approach the base line, which simply expresses the fact that equal increments added become smaller relatively as the organ and the body increase in size. With certain initial size relations among organ, increment and total weight, addition of constantly equal increments would give a curve showing first a rapid rise and then a slower and long continued fall, much as we find in some of these curves, for example, in the heart, which increases regularly at the rate of 0.85 gram per 1000 grams total weight. Should the added increments constantly decrease in amount the curve would approach the base line more rapidh% or should the increments increase the curve might become parallel with the base as in the rectal gland, or it might even rise as in the testes.

In discussing the possible significance of these curves it is convenient to consider the liver and gonads separately later. It is evident that


Growth of Brain and \'iseL'ra in Dogfish 333

during a certain early period, the exact time and extent of wliicli may vary, the hrain, heart, rectal gland, pancreas and spleen grow at a faster rate than do the other parts of the body: and that after this period these organs grow at a slower rate than the other parts. As other parts we must include chiefly the muscular and skeletal elements, gonads and probably fat. The conditions of the excretory system and of the peripheral parts of the circulatory and nervous systems are unknown in this form, but these constitute relatively small parts of the total weight with which these other organs are being compared, and in forms where something is known about them they too grow more slowly. The alimentary canal seems to share this relation and in the dogfish its length increases at a constantly diminishing rate.

It appears therefore that while the various tissues or organs grow at different rates, those forming the greater part of the bulk of the body, namely, the locomotor and supporting tissues, continue throughout life to increase in mass at a more rapid rate than do the brain, heart and chief viscera. Ultimately, of course, were this relation continued, a condition would be reached where these parts mentioned would become physiologically incompetent to carry on the work of the constantly increasing mass of the body. That is, the muscles and connective tissues actually outgrow their controlling and metabolizing sources, the physiological balance of the organism becomes unstable and death may result naturally or the animal ma}^ become an easy victim to adverse circumstances.

Many features of the curves are partly explainable by assuming that the mass of an organ is related in a general, though not a precise, way to the extent of its functional activity. Thus at the time of birth the relative size of the heart and digestive glands is increasing rapidly and this increase continues for a time after birth, while the young fish are undergoing the most critical period of their existence and are making the most numerous and rapid adjustments of their lives. The young contained in the oviducts of the mother are almost inactive and food is supplied abundantly and in a condition which permits of its easy assimilation. Immediately after birth the small fish must capture and prepare for assimilation their own food and the activity of the neuromuscular and digestive tissues is enormously increased. Probably the relatively immense size of the brain at birth is a provision for this period which comes on so abruptly, and the initial phase of rapid increase in relative size of this organ which we have noted is absent from our


334 William E. Kellicott

curves, is actually thrown forward into the latter part of "foetal" life and is only apparently omitted. It is not now possible to consider the rectal gland from this point of view because we have no positive knowledge regarding its function, but the similarity between its curve of growth and that of the brain is worth noting, though probably merely accidental.

Following this period of rapid and extensive adjustment comes that of the most rapid growth which occurs during that time the animal is dependent upon its own organs of direction and metabolism. During the year of embryonic or "foetal" life it attains a weight of approximately 75 grams. During the first year of postnatal life, if our determination of time rate of growth is correct, it grows to a weight of about 300 grams — an increase of four times. During its second year, however, the fish only little more than doubles its weight. It is, therefore, when of from 75 to 300 grams weight that, as an independent organism, it is growing the most rapidly (as measured by Minot's method) and that all its independent nutritive processes must be relatively at their maximum, and it is just at this time that the metabolizing organs are relatively the largest. The spleen gives an excellent single example of this relation; in the fish this is the chief hsmatoplastic organ and during this period of rapid growth the circulating tissues must be of prime importance, and correspondingly we find the spleen relatively very large. After this early period of rapid growth increase in bulk is slower and the relative size of the metabolizing organs decreases. But as already mentioned, late in life the brain and viscera cease their growth more rapidly than the muscles and supporting tissues, which make up about 75 per cent of the total weight. This relation is obviously non-adaptive, but may, probably will, be found quite frequent if not typical among indeterminately growing forms.

Donaldson, '95, has noted from the fragmentary observations of Bischoff a similar relation in man and pointed out that there the muscular system has grown more extensively than the brain and nutritive system, suggesting that this was because the constructive processes become less active in older individuals. His suggestion is well supported by these more complete data from the dogfish, although here the problem is somewhat different because the muscles and connective tissues do not stop growing at a definite time as in man and other mammals.

We may now attempt to interpret briefly the peculiar growth relations of the liver and gonads. The liver seems to differ from the other


Growth of Brain and Viscera in Dogfish 335

digestive organs and to agree more nearly with the muscular and skeletal tissues with respect to its comparative rate of growth in that it does not show on the whole a pronounced relative decrease in weight. The first part of the liver curve (Plate 6) is almost precisely like those of the other organs, but the usual steady decline following the maximal relative size is replaced by only a brief period of decline and then the liver weight rises again to a fairly high level, tending to fall again only among the very largest specimens. It seems probable that this peculiarity is the result, at least in large part, of the storing of fat in this organ. Most higher vertebrates and some lower forms, under good nutritive conditions, lay up fat not only within many tissues but quite usually within or around some of the organs contained in the body cavity, often in large masses. In the dogfish one notices at once the complete absence of fat tissue in this region. The liver, however, particularly in the larger specimens, contains largo quantities of fat or oil, as the manufacturers of cod-liver oil could testify, and everyone who has used these animals in the laboratory can not have failed to observe this annoying fact.

In order to fix this point Dr. C.' L. Alsberg, of the Harvard Medical School, has very kindly made, at the Bureau of Fisheries Laboratory, a series of determinations of the amount of fat in livers taken from dogfish of different sizes. His results show clearly, first, that the amount of fat does increase rapidly with the size of the fish or liver to an extremely high percentage in tlie larger fish; and, second, that livers above or below the average weight in fish of a given size have relatively higher or lower percentages of fat respectively.

In the dogfish, therefore, as in the teleost (Fulton, '06) the liver must be regarded not as a simple glandular organ, but, in the older individuals, largely as a fat reservoir. This fact offers an interpretation of the form of its growth curves. Only in the younger fish is the liver chiefly a glandular structure, and in such fish we find the growth curve typical for such parts. In animals of 600 to 800 grams we find the accumulation of fat masking the actual condition of growth of the liver tissue proper, so that apparently but not really this organ seems to afford an exception to the rate of growth typical of the viscera. If we could subtract the amount of stored fat from the total weight we should have left a series of weights representing actual liver substance, which would give a curve probably not unlike those already described for the other viscera. The extreme variability of the liver may also result largely from this same fact, since the amount of fat stored up would depend


336 William E. Kellicott

quite largely upon the constantly varying nutritional conditions of the fish.

The somewhat similar form of the growth curves of the gonad weights must of course have a totally different significance. The gonads (Plate 7), while at first like the other viscera, are unique in showing later both an absolute and a relative increase in weight throughout life ; this occurs earlier and is more pronounced in the testes. The gonads are the only organs whose curves of growth give evidence of being composed of two overlapping cycles : in the males a second growth C3'Cle commences when a total weight of 800 to 1000 grams has been reached, and in the females when the total weight reaches 2000 to 3000 grams. The periods at which this second increase in size of the gonads occur, namely, of sexual maturity in each case, give a clue to its cause and real nature.

The first cycle which is in progress at birth, is that during which these organs reach their primary maximum soon after birth ; this probably is equivalent to the only cycle present in the growth of the other organs and occurs at about the same time. During this cycle the gonads are not reproductive in function; they are growing as more or less undifferentiated mesodermal organs, not wholly undifferentiated, for they must be producing internal secretions which affect the rate of growth of certain other parts in such a way as to form at any rate the secondary sexual characters. As this phase of growth commences to wane just as it does in the other viscera, the gonads become truly reproductive in structure and in function and as such organs they enter upon a second cycle of growth which carries the size of these parts to a much higher point, and which is never completed in the sense that it reaches a maximum and then declines during life. This phase of growth is more pronounced in the testes than in the ovaries because the latter are not so largely composed of strictly reproductive tissue, but contain relatively a much larger amount of connective tissue.

We have already mentioned the apparently composite character of the latter part of the growth curve of the ovary. These weights were being determined during the season of ovulation and in Plate 7 the curves A and B show respectively the average weights of the ovaries before and after ovulation. There must be consequently a regular rhythm in the weight of the ovary which extends over a period of one or two years — it is not certain whether this dogfish can produce young annually or only biennially, although probably the latter is the case. The curve of ovaries which have recently discharged their ova drops quite to the horizontal.


Growth of Brain and Viscera in Dogfish 337

indicating that they just maintain their relative size. It is only the ovaries containing large maturing ova, filled with 3'olk, that show clearly the second growth cycle; this agrees with the interpretation offered of their peculiar growth curve.

Comparison with otiiek Data

Upon attempting to consider these data comparatively w^e meet serious limitations at once. The growth of the brain has been studied cpiite extensively in mammals, chiefly in man (see Donaldson, '95, for references) and in the guinea pig and the rat, which are typical for the group, and in frogs of various species (Donaldson, '98, '03, '08, Donaldson and Shoemaker, '00). Concerning the growth of the viscera we seem to have no sufficiently extended data for comparative use except for man, and these are not homogeneous and are far from satisfactory. The data collected by Welcker and Brandt, '02, are the best, but these observations are based upon such a small number of individuals that their usefulness is very limited. In dealing with such variable structures, records of single individuals of a given size may be very far from representing the actual average condition.

The work of Donaldson ('95, '98, '03, '08) has made it well known that in mammals after birth and in frogs after metamorphosis the brain does not grow as rapidly as the remainder of the body, so that it becomes relatively smaller as the animal becomes larger or older. Dubois. '98, (this paper contains a good bibliography up to 1898), Dhere and Lapicque, '98, Donaldson, '03, Beck, '07, Hatai, '08, and Eobertson, '08, have proposed formulae according to which the weight of the brain can be calculated quite accurately from the weight or weight and length of the body or some other external characters. All of these formulae are of such character as to show that in a given species the brain becomes relatively smaller as the individual increases in size. To illustrate, Vierordt, '06, publishes data (quoted by Donaldson, '95, from a previous edition of Vierordt) showing that in human males the brain at birth composes 12.29 per cent of the total weight, while at 25 years of age only 2.16 per cent. A similar scries of data for females gave 12.81 and 2.23 per cent respectively. During an increase of more than 21 times in total weight the brain had increased only about 3.5 times. Welcker and Brandt, '02, give a few data from several sources which show practically • the same decrease. They give measurements also upon the human


02

<! Pi

O ►J


H

o


E-«  K O 1— I


PQ

H

P> i-i Eh

<J

O

<

U <1










1




■3 r 2 5 & > fc " ■< OO


: ^

^


^


^


^


fe^


^


EiS



5aS§:^

J N <j m <;


• • o ■ • O



CO



O

CO



CI


CO


CO




lO




-*



I>



H^




i








H


CQ


to


en


CO


CO


CO



CO to 1


Z X


a>


<v


o


OJ


0)


2



s


a>


•^ o


r


r






c



^









c











a a












. . -^^


-t^


+-<


••-^


^

■<^



-*-a


-IJ



o







iC


CM


P3 g 1^


■ • r^


t

IC


"—1


CO


CO



t^


'^


CO


CO


O


-*


00


rt



CI


00


"^W




""*








f-l


tc


a:


Ul


CO


(O


CO



m


CO




(D


— ,


,~^^


, — • —


^■^~


' ' — ^


,— ^— .


'~'^~"


/-^-<


2° ^










.




-Ht^



00


00 00


CO



co'io


a f* >J Ed


C5 •* O -* (M


Oi«D


O -V


05 CO


00 CO


00 05


t^ OS


rt 00


(M t^ 1-H rl -*


(N ^


CO o


r-l CI


00 o


^CM


CO.-H


a>-*


--^ o


, O o a

O 00 O Tj4 CI


!M (M


CO --1


CO o


i-H O


oo


CM O


I— I 1—1


rH O


(M.-i,-i


^









a&:











g





















E a;


00



-H



iC lO


CO IC


'. lO


00 CM


iC


'J


CO



lO


<M


CM —1


COC2


. ^


OCM


■^ ,


lO O C' o c


C: O


CD.-H


0> 00


O C^l


oo


o


oo


00 CM


z < 5


<M (M O CD O


1— 1 )-H


i>o


O CO


do


o o


o

oo


o t

OiOOOr-,


00 fO


00








(S


CO^'*


cc^









PQ


T-H T— 1


I— 1









L4











! ^









«0 N


3 ^


T—




in


CM




o-*



in o in o o


q q


(£> (O


<o t^


CO q


t- to


<o


■* «*


CM O


^1


oi -r^ o S o


d d


d 00


oi in


T^ CO


<N ^


CO


d T^


<o >*


T- a> <o ^ o


o o


CO 00


CM 0>



T- ■*


CM


I^ CO


tJ ^^


"^^ CO o> CJ


T- CM


CM r^


in


CO





^


2 z


CM CO 00


CO to


"^







00


b "


CO in


to









H.












op ^




73 "^D







a o



oi:


CO

11


-Is


.-3


^3


CO ^

-73 T3



go

H

1^



g-<


CMM


■<


\<


co-<




H-2


1-H






1-5





-^ <<








(M


o


o^


(M


00


00


CI


. c^J




p


p


o


CO



^p




^


'^


-.


-.


bc" ^


■^ t/J -^


TS

Cj

oi r

11



i

1



r-' -i^ J


o-o


2s 32


E, be (I) 73



cd

bO


1


c3


>



c3


o Q




o

Q


Growth of Brain and Viscera in Dogfish 339

fcetiis showing that at three and six months the percentage weight of the brain is 20.29 and 18.74 respectively, at birth 16.10 and adult 2.42 per cent.

For the bull-frog (Kana catesbiana) Donaldson, '98, gives data showing that while the total weight is increasing from 1.32 up to 313.0 grams the relative weight of the brain falls from 1.89 to 0.068 per cent. And in three other species of frogs (E. pipiens, R. esculenta, R. temporaria) the same author, '08, gives data showing that the brain falls from 0.57, 0.4G5, 0.43 per cent to 0.24, 0.24, 0.23 per cent respectively, while the total weights were increasing from 11.6, 12.4, 14.1 grams to 47.0, 45.0, 32.8 grams respectively. We should note that in determining the total Aveight ol the frogs the ovaries were included only when not pigmented, if pigmented their weight Avas subtracted. If these had been included throughout, the ratios given for the large individuals would be still lower.

And we haA'e seen that in the dogfish the brain decreased from 1,116 to 0.085 per cent of the total weight during growth from birth to maximum size.

Obviously these figures can not be used comparatively because the relative ages and conditions are so unlike, but just to emphasize the fact of this universal and very considerable falling off in relative brain weight these data and some others are condensed in Table I.

The idea that the viscera, like the brain, show a relative decrease during growth is perhaps less familiar. Most of the available data have been collected and tabulated by Welcker and Brandt, '02, and some additional data for man by Vierordt, '06. These are partially given in a condensed form in Table II. In considering this table it should be noted that only in the frog and dogfish are the figures the result of the examination of a considerable series of specimens so that the truly representative character of those included in the table is certain. In the other forms often only a single individual has to serve, and when we consider the great normal variability of the weight of viscera we see that these figures may not give very precise information. Nor except in man and the fowl have we information as to the relative size of these parts at different stages of growth. In these two instances we have a few scattering data given by Welcker and Brandt, '02, from their own and various other sources. These data suggest that in these forms more complete observations would show that many of the viscera w^ould give curves showing a rise and fall as in the dogfish, but facts are so scanty that nothing definite can be said on the point. Incomplete as




o

I— I




I— I <^




w


u

^





hJ


r^


-^



H


k<


O



H


K




\^



o


^


H


^


"A


3


« 


(^


w



f^



r/5



52;



-!l



C



e;





fe







2 'S S §

^




00






00

c


00


d




d


CI


0.092 (prob. = 0.075)


1 -1


^- \ :


cocoo

O't CO

.-00







«5 r-4


11^

|d


II 10

(N t^

So


00 ■305

^co

-go


lip

o|o


II

CO |>« 



00

in

CO

d


i


i-i

HO





05


CO 10


05



d




IN


IN


05 IN



00

CO

CO





CO

in


00 CO


05


in

IN






IN



. los '.


K

II

li

< .

CO CO



00

in

N


rt c


05

d


in


<N

d




_


00


in

05

d





'



M (P JO

g 3-S

<; S.CO j •


d

SI"


00


00 10



_




to

IN

CO


<5

in



1

'



(R. esculent a)


IK « 

bC tn eS ^ <u


in



10

CO

d


CO


OS 05


00




d



c


•*


d



IN


05 j

in

CO


SI


in in


e<5


d



00


CO

d



d



d



00 IN

d


00

IN


to


COcK

d


00

(N



i

1




in

a m


05 CO

i


00.2 00'


d


to d


05




00

d


« 


lO

d


IN

m d


0^


■a S-3--S


in


a


« 21



CO


d



d



CO



CO




■5"

s

§1



in


00

00


X5

OS

C5


05


Si to-;



10 i2 d£


00



d


CO (N





00


d




"5


CO


^1


-H

CO CO


be

05^

°13


CO (N

d



CO IN



05

in


to

CO



in

00

in


S5 <




00


8

in


1^1


(N


(N 00

to (N

C


01


d


d


in

IN


(N

in

d


d


05

co'


d ;S 2 5

in



CO


06



CO

CO IN


s§3

^1



d


(N ■*

d d


in

CO

d


in

d


§

10


CO

d




. d ©^ >.



CO"" OOP


05

IN

d


CO '


d


in



in



CO 00


CO


d


d


to in

CO




2

0"


d c

a; ro


0)

13

J


I


d 1




c "a

02


"3 a

u

-3


i %


3


■d

s


to

■i c



s


Growtli of Brain and Viscera in Dogfish 341

tlie data are they represent the state of our knowledge and are given in the table for what they are worth.

The net result is to show that among the vertebrates generally the parts which increase relatively the most during growth are the skeleton and muscles and perhaps the skin and subcutaneous tissue. Together these make up roughly three-fourths of the total weight. The viscera as well as the brain show a falling off at various rates.

Conclusion

We see then that it is a general law among vertebrates that the organism does not grow entirely as a unit but as the resultant of the growth of its parts, and that these component parts of the body do not all grow at the same or corresponding rates. What is ordinarily measured and described as growth of tbe organism is really not growth of the whole organism, but the growth chiefly of the locomotor, supporting and protective tissues, and probably frequently of fat also. The curve of growth of an organism is a composite affair made up of many dissimilar components, with the growth of those mentioned as the predominating elements. Such curves therefore may be misleading unless this relation is borne in mind, since disproportionate importance is thus allowed to elements of the body which should be considered of only co-ordinate importance. Indeed it is not clear why, in the growing organism, the condition of the brain and viscera should not be the more significant, and 5^et the growth of these parts is almost completely masked by the increase of the bulkier muscles, skeleton and skin.

It seems that some confusion has arisen in the study of growth through the failure to remember that in all of those forms whose growth has been studied most extensively, growth is determinate: the organism reaches, about the time of sexual maturity, a fairly definite average size, then stops growing and may continue to live for a considerable time thereafter. This is true for mammals and birds at least and perhaps also for some of the lower vertebrates. In some forms the organism or parts of it may actually diminish in weight. But among fishes quite another condition prevails: growth is indeterminate and the fish may continue to increase in weight, slowly it is true, as long as it remains alive, with an average food supply. This seems to be the condition among most invertebrates, except insects, and has been demonstrated in the teleost fishes (Fulton, '01, '06), and from these data it must be true


342 William E. Kellicott

also of the dogfish. Observation of many of the lower vertebrates in nature (Fulton, '01, '06) and in captivity, such as the giant salamander and some reptiles, shows that these grow indeterminately; Agassiz's, '57, observations i;pon Chrysemys are typical. As a recent example of the failure to make this distinction we might mention the work of Eobertson '08, who has devised certain formulae for the description of growth and has brought out the very suggestive fact that the growth curve of an organism or organ or tissue is similar to that given by an autocatalytic reaction. These formulae hold good upon the assumption that the organism or organ has a definite period of growth at the end of which increase in size ceases. This is true for the higher vertebrates, but for all the indeterminately growing forms we can not determine any such "final weight" of the body or organ upon which to base a formula. We could not assume the maximum discovered size as the "final weight" because this is subject to such extreme variation; in the dogfish, including both sexes, we might find the "final weight" anywhere from 2000 to 8000 grams and even higher.

We have already suggested, since the muscles, supporting tissues and skin increase relatively much more rapidly than the other parts and since this relation continues through life, that a time must come at which the brain and metabolizing organs become incompetent as physiological elements in the organism and death must result. As showing that this relation is not one of mere bulk alone, Ave might recall some of the observations of Hardesty, '05, on the frog. Here the number of spinal ganglion cells as well as the number of dorsal and ventral root fibers increases throughout growth, but inspection of his data shows that this increase is at a constantly diminishing rate so that the relative number of ganglion cells is constantly decreasing. Thus in a frog of 7.0 grams there are in the ganglia of the V, VII, IX nerves respectively 103, 77.8, 549 cells per gram of body weight, while in a frog of 63.4 grams the corresponding numbers are 15.9, 18.1, 72.7, although the actual numbers of cells contained have increased 41, 111, 20 per cent respectively. In general the same is true for the number of fibers in the dorsal and ventral roots. Hardesty points out that this increase in number of ganglion cells is opposed to the tradition regarding the nerve cells of vertebrates which is supported by Hatai, '02, who found in the rat no increase in the number of ganglion cells during growth, but only increase in the number of mature fibers.

Beddard, '03, found in a giant salamander which died in the Zoological Society's Gardens that the only visible cause of death was the small


Growth of Brain and Viscera in Dosfi&h


343


TABLE III. The Data. Weights, in Grams, of Total Fish, Brain and Viscera






MALES.






Total


Brain


i Rectal Gland


He.\rt


Pancreas


Spleen


Gonad


Liveh


C


695


0.77


0.03


0.08


0.06


0.09


0.22


1.77


L


71-3


0.885


0.03


0.08


0.055


0.085


0.24


1.78


J


71.8


. 805


0.025


0.075


0.055


0.085


0.24


1.96


E


76.7


0.81


0.02


0.07


0.06


0.095


0.27


2.45


I


78.2


0.84


0.035


0.085


0.06


0.09


0.26


2.76


H


78.8


0.85


0.04


0.095


0.065


0.15


0.275


2.79


D


79 I


0.851


0.04


0.086


0.065


0.12


0.31


2.325


A


80.3


0.86


0.03


0.09


0.06


0.08


0.32


2.73


247


124.6


1.04


0.04 '


0.155


0.13


0.52


0.57


6.68


168


128.8


0.99


0.045


0.13


0.18


0.43


0.51


5.52


206


157-3


1.11


0.05


0.185


0.205


. 555


0.915


9.13


197


160.9


1*18


0.055


0.175


0.225


0.585


0.84


7.51


214


172.5


1.22


0.05


0.21


0.21


0.77


0.99


7.45


185


186.0


1.14


0.055


0.205


0.25


0.53


0.925


8.57


231


198.5


1.17


0.085


0.25


0.29


1.31


1.10


12.54


205


214.8


1.36


0.065


0.25


0.265


0.66


1.36


11.78


198


223.0


1.35


0.07


0.25


0.32


1.07


1.46


, 11.16


229


232.5


1.35


0.08


0.22


0.30


0.91


1.62


11.21


238


242.5


1.415


0.08


0.26


0.335


2.17


1.89


12:67


216


248.2


1.33


0.10 1


0.28


0.405


0.92


1.48


10.98


252


252.0


1.35


0.085


0.26


0.30


1.16


1.93


12.68


224


279.0


1.23


0.11


0.315


0.325


1.55


2.05


16.17


64


281.0



0.09 :


0.29


0.47


1.70


2.48


16.00


225


282.0


1.45


0.10


0.28


0.34


1.12


1.73


14.23


59


311



0.12


0.32


0.45


2.10


2.47


18.71


62


324.0



0.12


0.32


0.37


1.25


2.00


17.40


294


348.0


1.51


0.11


0.39


0.41


1.39


2.76


24.40


38


362.0



0.07


0.31


0.25


0.95


0.85


7.89


122


368.6


2.08


0.09


0.355


0.52


1.25


2.11


12.75


125


425 -7


2.065


0.14


0.49


0.72


1.51


2.30


17.92


112


435


2.21


0.105


0.44


0.60


1.33


3.23


14.00


164


437


2.21


0.13


0.42


0.62


1.38


2.60


19.42


227


458.0


2.085


0.105


0.485


0.495


1.78


4.035


19.92


124


528.0


2.19


0.175


0.57


0.73


2.23


4.69


24.25


26


537 -o



0.15


0.56


0.76


3.22


4.82


15.74


20


558.0



0.12


0.53


0.75


2.07


3.50


19.26


175


560.5


2.23


0.205


0.57


0.65


2.39


7.60


31.98


228


594


2.21


0.16


0.555


0.76


3.05


4.40


23.57


27


610.0


i


0.15


0.55


0.70

1


2.21


3.80


21.54


344


William E. Kellieott


TABLE III. (Continued.) The Data. Weights, in Grams, of Total Fish, Beain and Viscera


MALES


< « 


Tot.^l


Brain


Rectal Gland


Heart


Pancreas


Spleen


Gonad


Liver


157


616.3


2.625


0.175


0.605


0.83


3.11


3.71


33.53


43


628.5



0.12


0.57


0.63


2.37


4.31


27.34


180


632.7


2.78


0.195


0.615


0.785


1.90


3.81


18.13


28


663.5



0.16


0.50


0.80


3.17


4.50


23.06


128


678.3


2.33


0.205


0.65


0.85


2.42


3.85


22.41


41


680.0



0.22


0.58


0.69


3.24


6.12


32.74


22


693 5



0.16


0.56


0.73


2.135


3.92


26.64


106


707.0


2.44


0.16


0.70


0.91


2.39


3.95


22.91


39


715



0.18


0.60


0.86


4.21


5.31


39.69


170


735


2.555


0.22


0.655


0.95


2.39


4.03


25.79


19


743



0.20


0.67


1.025


5.73


6.86


34.36


45


772.0



0.15


0.60


0.66


2.39


3.84


40.44


6


779



0.18


0.865


0.88


2.46



30 . 65


220


793 5


2.92


0.245


0.79


0.955


3.205


4.05


24.92


258


816.7


2.48


0.22


0.77


0.985


4.275


6.11


34.05


303


818.0


2.25


0.22


0.71


1.18


2.38


6.80


56.2


223


839


2.76


0.22


0.72


1.03


3.16


4.48


53.0


259


842.0


2.68


0.22


0.865


1.08


3.25


5.02


36.36


262


845 5


2.66


0.22


0.71


0.80


2.87


3.95


41.03


89


863.7


2.95


0.23


0.86


1.14


2.71


5.13


37.76


63


878.0



0.21


0.72


0.81


2.35


4.10


52.0


61


893



0.19


0.81


0.87


4.045


4.69


49.79


79


900.4


2.76


0.21


0.78


0.81


2.76


4.79


40.4


10


908.0



0.27


0.74


0.86


2.87



32.31


99


912.5


2.81


0.29


0.96


1.48


3.34


5.18


43.0


1


917 5



0.20


1.00


1.12


2.61



46.87


277


938.0


2 . 365


0.26


0.91


1.39


5.625


5.41


58.8


270


946.0


2.92


0.23


0.91


0.89


3.96


6.55


57.7


215


958.0


2.98


0.23


0.935


0.97


2.50


6.73


39.1


177


970.8


2.99


0.295


0.98


1.105


3.78


5.37


42.1


246


999 5


2.58


0.24


0.85


0.78


3.80


5.03


72.5


131


1024.5


3.035


0.28


1.03


1.14


3.765


6.70


38.3


80


1087.5


2.95


0.24


1.05


1.17


3.17


8.08


32.4


249


1088


2.87


0.25


1.10


1.11


3.34


6.08


35.6


162


1 1 06


3.02


0.29


0.99


1.29


3.52


7.36


40.4


21


1112.5



0.23


0.97


1.01


4.48


9.47


74.39


23


1123.0



0.235


0.85


1.30


3.80


5.70


47.15


2


1124.5



0.26


0.95


1.405


4.70



52.75


3


1150



0.27


1.06


1.16


2.93



44.77


Growth of Brain and Viscera in Doafisli


345


TABLE III. (Continued.) The Data. Weights, in Grams, of Total Fish, Brain and Viscera


MALES


2 n

IS S

02 2;


Total


Brain


Rectal Gland


Heart


Pancreas


Spleen


Gonad


Liver


245


I162.O


2.97


0.21


1.07


1.035


3.29


7.57


67.0


11


1 194 5



0.29


1.09


1.32


3.54



59.08


85


1208.4


3.28


0.31


1.05


1.47


3.67


7.08


36.75


178


1221.3


3.18


0.28


1.28


1.11


3.47


6.33


39.7


29


1246.5



0.26


1.07


1.15


2.82


8.13


57.15


149


1259 5


3.09


0.29


0.98


0.96


4.21


7.13


58.1


137


1277


2.99


0.285


1 .22


1.425


3.24


8.01


89.8


296


1291 .0


3.325


0.29


1.095


1.81


3.48


7.02


75.7


209


1300.0


3.55


0.36


1.23


1.535


3.275


10.045


64.8


171


1315 5


3.47


0.30


1.13


1.27


4.53


7.96


79.6


14


1324.0



0.27


1.07


1.26


3.95



68.11


46


1327.0



0.33


1.11


1.62


2.90


8.69


73.56


5


1337



0.24


1.24


1.525


3.56



59.55


78


1338.5


3.84


0.29


1.19


1.64


4.87


11.81


77.4


181


1370.0


3.21


0.30


1.10


1.36


3.07


7.53


44.5


286


1400.0


3.50


0.30


1.28


1.63


3.49


8.95


59.3


56


1416.0



0.23


1.05


1.175


4.45


9.52


86.87


17


1452.5



0.30


1.27


1.56


3.55


12.51


80.01


47


1453



0.30


1.14


1.40


3.90


9.57


80.0


34


1454



0.27


1.20


1.40


5.18


9.55


79.99


172


1462.0


3.53


0.32


1.54


1.62


4.00


11.07


58.5


136


1463


3.655


0.33


1.32


1.72


5.13


13.61


50.1


154


1468 . 5


3.92


0.32


1.42


1.79


4.90


12.71


56.1


121


1471.0


3.42


0.27


1.31


1.28


3.50


14.20


77.7


109


1481 .0


3.64


0.28


1.21


1.47


4.55


10.62


90.9


293


1495


3.47


0.37


1.42


1.65


3.91


13.47


94.3


179


1512.3


3.55


0.39


1.32


1.88


5.08


14.18


92.2


156


1521.0


3.87


0.305


1.24


1.67


3.30


11.00


59.0


261


1529 5


3.52


0.30


1.29


1.26


4.56


12.38


76.8


13


1563



0.31


1.30


1.83


4.65



80.29


160


1S67 7


4.17


0.285


1.36


1.72


3.34


13.46


82.2


92


1625.0


3.71


0.33


1.55


1.765


2.61


13.42


64.0


297


1651.0


3.32


0.36


1.425


1.61


4.45


13.56


114.2


218


1655


4.235


0.37


1.45


1.30


4.715


15.13


79.4


7


1681.0



0.41


1.36


1.81


4.55



92.34


213


1749


3.62


0.345


1.60


1.80


5.29


19.55


103.7


113


1756.3


3.82


0.42


1.60


1.85


5.69


17.12


66.4


248


I 760 .


3.51


0.34


1.24


1.29


3.39


12.78


79.7


90


1760.6


4.13


, 0.45


1.48


2.13


5.79


16.09


54.7


346


William E. Kellicott


The Data.


TABLE III. (Continued.) Weights, in Grams, of Total Fish, Brain and Viscera


MALES


si

K S


Total




219


1821 .0


291


1860.0


95


1914-5


93


1975 9


83


1986 .


256


2002 .


266


2046 .


166


2046 .


115


2070 . 5


165


2130.0


114


2185.5


273


2197.0


52


2302.0


140


2398 5


111


2464.5


240


2482 5


268


2563


69


2590.0


44


2649.0


108


2651.0


67


2791.0


94


3010.5


Br.un


4.095

3.64

4.01

3.89

4.88

3.98

4.385

4.24

4.06

4.46

4.28

4.15

4.81 4.74 4.37 4.02 4.80

4.99 4.77 5.38


Rectal Gland


0.35

0.33

0.39

0.44

0.40

0.335

0.41

0.45

0.47

0.41

0.48

0.405

0.40

0.48

0.43

0.455

0.46

0.50

0.47

0.50

0.45

0.61


Heart


1.965


1.50 1.65 1.11 2.17 2.005 1.88 1.815 2.06 1.50 1.81 1.89 1.67 2.23 70 015 96 33


2.25 2.39 2.72 2.94


PaNCRE-'^.S


Spleen


1.77


5.09


1.96


5.66


2.01


5.59


2.07


4.52


2.00


6.31


1.87


5.45


2.21


4.69


2.32


5.24


2.24


5.41


2.345


5.85


2.24


6.05


1.31


7.41


1.96


5.64


3.23


5.62


3.07


6.63


2.55


7.27


2.225


5.00


2.28


4.75


2.24


7.56


2.29


8.44


2.35


7.08


3.46


4.95


Gonad


17.16 20.40 20.59 14.59 23.28 19.97 23.89 18.99 16.00 19.63 21.63 20.06 20.78 24.65 28.36 31.80 24.22 22.70 26.83 29.30 29.15 23.97


Liver


93.8

96.3

108.1

66.4

85*.

79.5

136.3

87.7

88.5

147.0

70.1

136.0

148.16

86.4 104.2 140.0 97.8 141.0 175.59 136.4 204.8 171.9


FEMALES


K


71.7


0.83


F


71.7



B


75


1.02


G


83.1


0.84


M


84.0


0.905


232


152. 1


1.07


196


159 5


1.10


167


160.0


1.07


186


189.0


1.27


187


189.0


1.39


57


208.0


.... !


195


209.0


1.42


191


217.5


1.155


169


220.8


1.315


0.025

0.02

0.03

0.035

0.035

0.05

0.065

0.06

0.06

0.07

0.08

0.065

0.065

0.085


0.075

0.085

0.085

0.08

0.100

0.17

0.20

0.18

0.215

0.245

0.27

0.25

0.22

0.26


0.06

0.06

0.07

0.06

0.065

0.20

0.20

0.27

0.23

0.245

0.27

0.305

0.32

0.275


0.06


0.25


0.095


0.23


0.10


0.27


0.09


0.35


0.120


0.325


0.57


0.60


0.50


0.88


0.36


0.855


0.50


0.93


0.54


0.94


0.80


1.06


1.245


1.02


1.02


1.075


1.50


1.39


1.95

1.96

2.50

2.51

3.720

7.48

7.92

8.19

9.17

9.08

13.33

11.87

11.02

13.45


Growth of Brain and Viscera in Dogfish


347


TABLE III. (Continued.) The Data. Weights, in Grams, of Total Fish, Brain and Viscera






FEMALES





m p


1 Total i


Brain


Rectal Gland


Heart


Pancreas


Spleen


Gonad


Liver


253


221 .O


1.285


0.07


0.28


0.33


1.60


1.52


13.63


53


226.0



0.09


0.23


0.34


1.56


1.50 1


11.82


37


238.0



0.09


0.25


0.32


1.07


1.45


10.79


50


240.0



0.10


0.24


0.38


1.18


1.64


11.87


194


251-4


1.35


0.08


0.27


0.30


0.725


1.27


14.39


204


259 7


1.40


0.07


0.27


0.35


1.01


1.41


12.09


237


269.5


1.37


0.08


0.28


0.33


0.71


2.10


12.00


221


308.5


1.385


0.10


0.30


0.41


0.95


1.54


17.01


58


333



0.12


0.35


0.46


2.11


2.72


18.74


300


345 5


1.54


0.12


0.395


0.45


0.85


1.95


19.44


288


352.5


1.45


0.12


0.37


0.49


1.51


2.25


23.13


132


390.7


1.96


0.10


0.43


0.56


1.41


4.71


16.63


144


441.4


2.44


0.135


0.465


0.68


1.29


2.88


18.29


25


482.5



0.13


0.49


0.81


1.48


1.97


16.14


289


485.0


1.81


0.145


0.55


0.62


1.405


2.63


35.45


202


513-5


2.36


0.16


0.57


0.58


2.15


3.65


15.88


199


513-7


2.33


0.12


0.545


0.66


1.59


3.93


19.95


176


525


2.28


0.16


0.53


0.81


1.91


2.56


23.68


188


527 5


2.56


0.15


0.56


0.635


1.97


3.065


20.53


51


547



0.15


0.61


0.65


2.25


3.37


23.46


150


550-3


2.315


0.16


0.555


0.70


2.06


2.795


23.19


123


552-7


2.46


0.15


0.485


0.70


1.81


3.03


18.45


287


555


2.00


0.13


0.55


0.675


1.90


3.00


21.53


126


561. 5


2.325


0.16


0.54


0.87


1.79


2.75


19.27


159


574


2.35


0.14


0.63


1.01


2.745


4.61


25.08


207


630.5


2.23


0.19


0.65


0.615


2.29


3.95


33.00


16


699.0



0.17


0.61


0.94


3.24


3.02


23.88


31


699.0



0.18


0.58


0.69


2.02


5.21


23.15


60


710.0



0.21


0.61


0.81


1.45


2.86


25.48


161


747.0


2.53


0.23


0.75


1.10


2.98


4.80


48.8


35


761.0



0.22


0.70


1.00


3.24


4.72


45.19


183


771.0


2.42


0.195


0.80


0.86


3.465


4.57


34.6


182


809.4


2.38


0.18


0.70


0.86


1.86


3.77


36.1


84


824.4


2.87


0.18


0.67


0.93


2.15


5.32


23.64


36


829.0



0.23


0.70


1.07


2.42


6.75


34.78


54


837



0.22


0.63


0.97


2.81


5.20


50.64


15


868.0



0.27


0.825


1.10


5.74



37.10


65


874.0



0.18


0.70


0.85


2.25


4.32


51.03


158


890.3


2.93


0.24


0.86


1.185


3.04


4.86


39.3


348


William E. Kellicott


TABLE III. (Continued.) The Data. Weights, in Grams, of Total Fish, Brain and Viscera


FEMALES


2 n as a p


Total


Brain


Rectal Gland


Heart


Pancreas


Spleen


Gonad


Liver


203


919 5


2.815


0.21


0.865


0.95


2.97


4.355


53.2


163


939 5


3.24


0.23


0.80


1.20


1.69


2.86


30.14


174


988.0


2.97


0.25


0.91


0.95


3.10


3.30


45.3


299


995


2.90


0.25


0.85


1.01


2.91


5.35


43.0


155


995 5


3.0


0.23


0.93


1.08


3.27


4.91


41.05


302


1000.


2.56


0.27


0.82


1.26


2.87


6.60


71.8


276


1014.0


2.88


0.255


0.82


0.945


2.55


5.76


54.9


230


1026.0


3.23


0.27


1.03


1.63


3.96


4.42


46.1


102


1042.4


3.13


0.275


0.95


1.32


4.65


3.82


40.3


235


1064.0


3.19


0.27


0.895


1.10


3.07


6.05


49.3


88


1073.2


3.11


0.33


0.94


1.32


3.72


5.00


74.4


135


1085.0


3.08


0.27


0.96


1.125


3.15


4.76


35.24


153


1086.0


3.17


0.22


0.99


1.18


2.85


4.37


29.33


98


1130


2.93


0.20


0.81


1.07


2.70


4.57


34.25


234


I 143.0


3.18


0.27


1.08


1.02


5.275


5.16


66.0


201


"45


3.29


0.27


1.02


1.18


3.00


7.00


53.6


127


I 163. I


3.14


0.265


1.05


1.78


4.935


6.79


52.52


104


I 176.0


3.06


0.28


0.95


1.28


5.07


6.19


102.0


236


1187.0


3.575


0.275


1.11


1.36


4.40


5.08


53.8


226


1196.5


3.07


0.255


1.08


1.30


4.54


4.65


76.5


55


1230



0.32


1.02


1.23


3.82


7.45


60.90


210


I 249 .


3.455


0.25


1.23


1.19


5.35


4.57


90.4


118


1267.0


3.16


0.28


1.13


1.30


3.45


5.42


67.8


97


1295 5


3.36


0.34


1.06


1.87


4.47


6.24


66.2


33


1296



0.27


1.09


0.95


3.77


7.17


74.49


9


1297 5



0.35


1.11


1.23


4.43



95.49


151


1303 5


3.76


0.31


1.45


1.72


4.70


7.89


64.75


72


1310 6


3.41


0.28


1.17


1.32


2.79


5.81


43.5


75


1312 5


3.66


0.29


1.02


1.23


2.81


4.91


41.6


298


1343


3.11


0.25


1.26


1.56


3.25


6.54


81.3


260


1352.0


3 . 735


0.31


1.255


1.70


4.02


8.92


59.2


200


1371


3.66


0.29


1.26


1.48


5.24


6.88


53.6


278


1392


3.11


0.315


1.17


1.46


4.09


7.47


102 . 5


272


1412 5


3.21


0.36


1.215


1.24


4.67


7.35


65 . 3


134


1439


3.90


0.35


1.32


1.34


3.35


8.0


44.4


285


1440


3.49


0.32


1 .22


1.50


2.61


5.19


79.5


24


1488



0.32


1.18


1.55


4.35


7.29


76.34


12


1500



0.30


1.21


2.04


3.80



111.59


251


1500


3.27


0.375


1.23


1.45


4.85


6.51


103.5


Growth of Brain and Viscera in Dosffisli


349


TABLE III. (Continued.) The Data, Weights, in Grams, of Total Fish, Brain and Viscera


FEMALES


K

< « 

M P


Total


i Brain


Rectal Gland


Heart


Pancreas


Spleen


Gonad


Liver


8


iSiQ 5



0.35


1.27


1.69


4.72



83.21


279


1537 o


3.48


0.38


1.34


1.61


5.97


6.03


77.1


129


1538.3


3.95


0.36


1.50


2.04


4.19


7.50


57.9


76


1550.0


3.37


0.285


1.29


1.53


3.77


5.38


" 69.2


71


1575.2


3.43


0.37


1.265


1.52


3.25


6.83


62.9


265


1603.0


3.24


0.285


1.32


1.58


4.47


5.84


95.0


292


1617.0


2.98


0.37


1.36


1.59


3.98


7.35


155.5


77


1642.0


3.72


0.38


1.30


1.645


4.98


7.01


166.7


86


1660.0


3.61


0.31


1.50


1.76


4.75


7.0


80.7


250


1688.0


3.96


0.34


1.40


1.74


6.00


9.24


94.0


103


1736.5


3.83


0.40


1.56


1.52


4.59


9.83


84.0


301


1737


3.205


0.385


1.325


1.60


5.21


8.53


141.0


119


1761 .0


4.05


0.36


1.50


1.83


2.69


7.23


38.8


73


1789 5


4.14


0.39


1.56


1.77


4.27


6.02


59.7


152


1791 .0


4.06


0.345


1.82


1.91


3.82


7.77


86.2


269


1795


3.34


0.405


1.38


2.13


3.53


5.44


129.5


130


1795


3.98


0.38


1.41


1.90


4.94


9.15


105 1


211


1803.5


4.19


0.42


1.88


1.89


4.68


10.22


106.6


100


1847.0


3.98


0.33


1.44


1.55


4.26


6.59


81.7


146


1890.0


3.73


0.375


1.72


2.35


4.595


9.12


81.4


244


1913.0


4.21


0.34


1.49


1.89


4.16


5 . 25


120.5


295


1948 .


3.51


. 365


1.69


1.55


5.48


10.94


148.5


184


1952.8


4.17


0.46


1.60


1.49


3.50


9.14


57.0


145


1998.0


4.24


0.45


1.915


2.40


6.30


13.47


74,3


40


2090 .



0.44


1.80


2.01


3.97


6.37


103.0


105


2107.0


3.88


0.445


2.06


2.44


4.78


9.51


121.9


120


2115.0


3.945


0.41


1.60


2.34


5.30


11.58


93.0


48


2130.0



0.43


1.50


1.64


5.21


8.55


109.34


49


2146.0



0.62


1.79


1.72


5.26


8.85


113.94


173


2175.0


4.05


0.49


1.82


1.87


4.57


13.74


75.2


212


2219.0


4.24


0.40


1.98


2.23


4.955


9.48


115.7


32


2230.0



0.35


1.75


1.75


6.13


11.20


143.76


101


2266.0


4.46


0.42


2. 21


1.61


5.50


30.28


75.6


290


2290.0


3.83


0.46


1.585


2.57


6.28


10.78


180.0


42


2302.0



0.45


1.86


2.42


5.51


7.62


171.51


133


2330 5


4.36


0.42


2.00


2.49


5.075


11.70


105.3


74


2332.0


4.29


0.405


1.97


2.55


4.79


13.06


98.1


4


2350.0



0.48


1.96


2.77


5.22



145.76


208


2380 .


4.35


0.44


2.055


2.28


5.775


13.15


157.4


110


2394 3


4.39


0.50


1.93


2.51 1


5.61


9.86


152.1


143


2427 5


4.39


0.51


2.38


2.61


5.40


11.35


149.8


275


2432


3.57


0.43


1.87


1.95


5.70


8.61


163.5


350


William E. Kellicott


TABLE III. (Concluded.) The Data. Weights, in Grams, of Total Fish, Brain and Viscera


FEMALES


B


Totai.


Brain


Rectal Gland


Heart


Pancreas


Spleen


[ Gonad


Liver


kZ










264


2534


4.24


0.54


2.05


1.82


4.08


9.78


139.4


91


2610.7


4.63


0.43


2.175


2.49


4.71


9.38


104.8


66 2642.0


4.3


0.45


2.5


2.1


4.2


9.5 j


129.0


142 2652.0


4.815


0.65


2.73


3.30


5.13


12.17


96.7


30 2860.0



0.51


2.05


2.32


6.70


13.25


175.2


274 2865.0


4.30


0.65


1.975


2.25


4.88


9.13


166.8


87 2900.0


4.52


0.66


2.27


2.28


7.00


11.8


230.5


141 2911.0


4.94


0.49


2.265


3.00


5.39


10.20


159.6


257 2960.0


4.57


0.67


2.66


2.19


6.50


15.09


232.2


239 3001.0


4.62


0.50


2.30


2.95


6.90


14.075


169.1


222 3039


5.33


0.58


2.96


3.48


5.17


11.63


228.3


242


3114


4.73


0.86


2.98


3.00


7.14


13.77


206.8


148


3250.0


4.61


0.66


2.59


2.48


6.60


18.0


154.6


192


3304 5


4.745


0.56


2.47


2.91


5.535


19.16


171.3


241


3312.0


4.80


0.60


2.62


2.925


6.46


14.59


326.5


189


3490.0


4.80


0.60


2.84


2.92


6.73


26.91


196.8


255


3513


4.87


0.80


2.98


3.08


8.14


18.90


325.5


271


3522.0


4.63


0.63


3.08


2.66


7.67


15.81


366.8


96


3581.0


4.71


0.77


2.92


3.20


8.28


15.50


253.0


254


3609


4.82


0.53


3.55


2.62


6.04


16.59


117.0


243


3638.0


5.08


0.74


3.64


4.98


12.36


27.91


368.3


193


3760.0


5.01


0.705


3.10


4.345


6.23


15.90


419.0


117


3810.0


5.34


0.65


3.31


3.59


8.96


27.27


206.2


147


3995


4.91


0.81


3.30


4.17


7.44


32.57


401.5


107


4152.5


5.28


0.655


3.30


3.56


8.96


26.04


.357.6


233


4172.0


5.08


0.72


3.72


3.50


10.25


17.82


283.3


116


4264.5


5.35


0.75


4.15


6.52


8.90


23.65


329.5


82


4275


5.63


0.94


4.05


4.03


6.08


47.5


188.0


217


4420 .


5.105


0.845


4.255


3.845


7.95


26.84


270.2


284


4510.0


5.16


0.78


3.53


3.91


8.47


17.30


247.0


263


4540.0


5.39


1.32


5.35


3.645


8.87


19.53


403.0


281


4550.0


5.48


0.905


3.98


3.47


8.27


26.77


259.5


68


4600 .


5.32


0.81


3.39


3.90


5.66


46.5


175.0


18


4754



0.93


3.36


3.46


8.33


23.60


355.95


139


4862 .


5.68


0.96


3.90


4.56


9.95


26.53


274.0


267


4863 .


4.98


0.92


4.735


4.16


6.34


20.39


274.7


282 5634 -6


5.51


0.95


4.36


4.43


7.03


16.26


296.5


280


5850.0


5.95


1.04


4.19


4.02


10.96


29.75


273.3


190


6123.6


5.55


1.17


5.72


5.43


8.57


28.37


394.5


138 6633


5.98


1.36


5.93


4.76


10.23


40.17


287.2


283 6790.0


5.915


1.36


5.19


5.84


8.12


24.97


557.0


70 8434.0


7.20


1 1 . 62


6.75


7.78


9.00


118.00


! 339.7


Growth of Brain and Viscera in Dogfish 351

size of the valves of the conus from which the walls of the conus had apparently grown away, thus rendering them incompetent. He found these valves relatively larger in a smaler specimen and in a brief letter to "Nature" suggested that such "normally unequal growth" might be here and elsewhere a cause of natural death. From the facts we have given here it seems that we must expand this idea of "normally unequal growth" to include the general and normal relation, among the lower forms at least, between the locomotor, supporting and protecting tissues on the one hand and on the other the controlling, circulatory and metabolizing organs, either singly or collectively.

From this point of view the condition of determinate growth which we find in higher vertebrates is secondary and is derived from that of indeterminate growth as an adaptation upon the part of the organism, such that the muscles and supporting tissues cease their growth at such a point that the brain and viscera remain competent to maintain a physiological balance. We have seen that in the dogfish the brain and viscera decrease rapidly in relative size after an early maximum until about the time of sexual maturity and after that the decrease is much slower. If only the muscles and connective tissues could be made to stop their growth at that time (i. e., become determinate), the animal apparently might continue to live for a comparatively long time after maturity. Apparently this is just what happens in the higher forms.

The recent work of the physiologists upon growth of certain tissues or organs suggests a mechanism through which this normal growth of the tissue or organ might be controlled and which might account for this comparatively more rapid cessation of growth in the determinately growing forms. They have demonstrated that the growth of occasional organs or tissues often results from the presence of internal secretions or of hormones. For example we might mention the effect of foetus extract upon the growth of the mammary glands, of ovarian secretion upon the growth of the placenta, and so forth. And also in pathological growths internal secretions are known to be involved, for example, the effect of the thyroid secretion upon growth of the brain or connective tissues and of pituitary secretion upon the growth of bones. Probably the best example of this effect of internal secretion upon the growth of pans is given by the whole group of secondary sexual characters which result fi'om secretions of the gonads. These secretions seem to affect growth in either a positive or a negative way — either by their presence or by their withdrawal after having been present for some time.


352 " William E. Kellicott

It is quite possible, therefore, that the normal growth of the individual tissue or organ may be similarly regulated by the presence or absence of specific internal secretions formed in various parts of the body, \ATiether these secretions cause growth to continue by their presence or by their absence could be told only through experiment, the relation might differ with different parts, but the general hypothesis of the control of normal growth through internal secretions seems legitimate and may afford a simple explanation of the growth of the organism as a composite collection of more or less independently growing units. The data presented here seem to give morphological evidence of a control of some such nature.

Woods Hole, Mass., August, 1908.

REFERENCES

Agassiz, L., '57. — Contributions to the Natural History of the United States

of America, Vol. I, Pt. 2. North American Testudinata, Boston. 1857. Beck, F. R., '07. — Eine Methode zur Bestimmung des Schadelinhaltes und

Hirngewichtes an Lebenden und ihre Beziehungen zum Kopfumfang.

Zeitschr. f. Morph. u. Anthrop., Bd. X, Heft 1, 1906-07. Beddard, F. E., '03.— Normally Unequal Growth as a Possible Cause of Death.

Nature, Vol. LXVIII, 1903. Dh£r6 et Lapicque, '98. — Sur le rapport entre la grandeur du corps et la

developpement de I'encephale. Arch, de Physiol., T. X, 1898. Donaldson, H. H., '95. — The Growth of the Brain. New York, 1895. Donaldson, H. H., '98. — Observations on the Weight and Length of the

Central Nervous System and of the Legs, in Bull-Frogs of different Sizes.

Jour. Comp. Neur., Vol. VIII, No. 4, 1898. Donaldson, H. H., '03. — On a Formula for determining the Weight of the

Centi-al Nervous System of the Frog from the Weight and Length of its

entire Body. University of Chicago, Decennial Publications, Ser. I, Vol.

X, 1903. Donaldson, H. H.. '08. — The Nervous System of the American Leopard Frog,

Rana pipiens, compared with that of the European Frogs, Rana esculenta

and Rana temporaria (fusca). Jour. Comp. Neur. and Psychol.. Vol.

XVIII, 1908. Donaldson, H. H., and Shoemaker, D. M., '00. — Observations on the Weight

and Length of the Central Nervous System and of the Legs in Frogs of

different Sizes (Rana virescens brachycephala, Cope; pipiens, Schreber).

Jour. Comp. Neur., Vol. X. 1900. Dubois, E., '98. — Ueber die Abhangigkeit des Hirngewichtes von der Korper grosse beim Menschen. Arch. f. Anthrop., Bd. XXV, 1898.


Growth of Brain and Viscera in Dogfish 353

Fulton, T. W., "01.— Rate of Growth of Sea Fishes. Fishery Board for Scotland. Ami. Kept. XX, Pt. 3, 1901.

Fulton. T. W.. "OG.— On the Rate of Growth of Fishes. Fishery Board for Scotland. Ann. Rept. XXIV, Pt. 3, 1900.

Hardesty, I., '05. — On the Number and Relations of the Ganglion Cells and Mediillated Nerve-Fibers in the Spinal Nerves of Frogs of Different Ages. Jour. Comp. Neur. and Psychol., Vol. XV, 1905.

Hatai. S., '02. — Number and Size of the Spinal Ganglion Cells and Dorsal Root Fibers in the White Rat at Different Ages. Jour. Comp. Neur., Vol. XII, 1902.

MoENKHAUs. W. J., '95. — A'ariation of North American Fishes. II. The Variation of Etheostoma caprodes, Rafinesque, in Turkey Lake and Tippecanoe Lake. Proc. Indiana Acad. Sci., No. V, 1895.

Robertson, T. B., '08. — On the Normal Rate of Growth of an Individual, and its Biochemical Significance. Arch. f. Entw.-Mech., Bd. XXV, Heft 4, 1908.

ViERORDT, H., '06.- — Anatomische, Physiologische und Physikalische Daten und Tabellen, 3 Aufl., Jena, 1906.

Welcker, H., und Brandt, A., '02.^Gewichtswerthe der Korperorgane bei dem Menschen und den Thiereu. Ein Beitrag zur vergleichenden Anatomie und Entwicklungsgeschichte. Arch. f. Anthrop., Bd. XXVIII. 1902.


PLATE 1. Curves showing the actual and relative increase in the weight of the brain in a series of dogfish (Mustelus canis). In this and in the following plates the sexes and individuals are plotted separately and the smoothed curves drawn from averages.


GROWTH OF BRAIN AND VISCERA IN DOGFISH

WILLIAM E. KELLICOTT



Wt. % d* = O X

? = . -^









1


« 



>.


~^r^--r^


- •





w


■J0-'









/'■■

--^


— .^



- ' ^ ...^ ^










10 Total Weight


DO 20 - Grams


no 30


DO 4000 50


00 6000


7000 8000


THE AMERICAN JOURNAL OF ANATOMY— VOL.


PLATE 2.

Curves showing the actual and relative increase in the weight of the heart.


GROWTH OF BRAIN AND VISCERA IN DOGFISH

KELUCOTT


6.0



wt. %

C^ = O X

9 = • +








^



♦"' ',"C '* A*\ »


'i, ♦. ,



+ * * "1


. ^


^


^




5.0




4.0



X






» — • ^^■'^



+




5.0

a




a 2.0



« 









t l.o



^


1000 2000

Total Weight - Grams


OF ANATOMY— VOL. VIII, NO.


PLATE 3.

Curves sbowiug tbe actual and relative increase in the weight of the rectal rlaud.


GROWTH OF BRAIN AND VISCERA IN DOGFISH

WILLIAM E. KELLICOTT



Vlt. %

c? = ° ^






/^


/


1.82








^







.


-^


^







..•


^^


^






1/;


^ •


^^i^










-I.J i • "


' ,.


y








^V^/x^f _^v








^"'










0.045 o


1000 Total Weight - Grams


, 5000


HE AMERICAN JOURNAL OF ANATOMY--VOL


PLATE 4. Curves sbowiug the actual and relative increase in the weight of the pan


GROWTH OF BRAIN AND VJSCERA IN DOGFISH

WILLIAM E. KELLICOTT



1000 Total Weight - Grams


THE AMERICAN JOUhNAL OF ANATOMY--VOL. VIM, NO. 4


PLATE 5. Curves showing the actual and relative increase in the weight of the spleen.


GROWTH OF BRAIN AND VISCERA IN DOGFISH

I E. KEULICOTT





12.36

!







wt. % S ^ o +

$ = • X




j^-


' —


"





,^-<^







fWv'



' *







♦*« "' ."' )^'




'








■(>


^ +. X



"



■ r


'


1000 Total Weight - Grams


HE AMERICAN JOURNAL OF ANATOMY-


PLATE 6. Curves showing the actual and relative increase in the weight of the liver.


GROWTH OF BRAIN AND VISCERA IN DOGFISH

WILLIAM E. KELLICOTT


^ 100








557





»t. "S

2 = . .



. .



^^^_


-^






••


■./"


-^^







k '"^


,'■ *•♦ « ^ ;


+ • .^^ * +


<^






' '"'""i^








■ =•



1000 Total Weight ■• Gra


AMERICAN JOURNAL OF ANATOMY-VOL.


PLATE 7. Curves showing the actual and relative increase in the weight of the gonads. The small circles enclose the records of individuals with ovaries containing large yolk-filled ova. Curves A and B show the average maximal and minimal sizes of ovaries before and after the discharge of the ova.


GROWTH OF BRAIN AND VISCERA IN DOGFISH

WILLIAM E. KELLICOTT



1000 Total Weight - Urams


IICAN JUUHNAL OF ANATOMY— VOL. VIII, NO, 4


(

COl


THE MOEPHOLOGY OF COSMOBIA; SPECULATIONS CONCERNING THE SIGNIFICANCE OF CERTAIN TYPES OF MONSTERS.

Bt

HARRIS HAWTHORNE WILDER. With 4 Plates and 32 Text Figures.

Introduction.

Should one wish to learn the methods of a conjurer, he might vainly watch the latter's customary repertoire, and, so long as everything went smoothly, might never obtain a clue to the mysterious performance, baffled by the precision of the manipulations and the complexity of the apparatus ; if, however, a single error were made in any part or if a single deviation from the customary method should force the manipulator along an unaccustomed path, it would give the investigator an opportunity to obtain a part or the whole of the secret. Thus, although the simile must not be pushed too far, it seems likely that through the study of the abnormal or unusual some insight may be obtained into that mystery of mysteries, the development of an organism, an insight denied to those who study only the usual and normal; and this is especially likely to be the case where the abnormalities studied are not deformities, such as are caused by failure of nutrition, mechanical injury, or other external cause, but where they are due to some modification in the germ itself, leading the organisms to develop in accordance with laws as definite and natural, though not as usual, as those governing normal development.

That any of the cases usually classed as "monstrosities" can be as natural and symmetrical in their development as are normal individuals, and be thus as legimate a subject for biological investigation, seems not to be generally believed, an attitude which has been fostered by the customary practice of denying them a place in the text-books of general anatomy and embryology, and banishing them to a sort of extra-mural

The American Journal of Anatomy. — Vol. VIII, No. 4.


356 Harris Hawthorne \\ ilder

ghetto among all sorts of malformations and deformities under the general head of "teratology", an omnium gatherum in which a few have sought and found valuable material but which for the most part have been left to the curiosity-seekers. Indeed, with the exception of the brilliant suggestion of Fisher^ who in 1866 separated the double monsters from the rest, there has been little or no attempt to distinguish between monsters that develop in accordance with the laws of growth inherent in the organism and the various deformities due to external causes. Classifications of monsters are not wanting, indeed the earlier teratologists did little else but classify, each by his own method, but a careful distinction between the two sorts does not seem to have been made.

That this distinction, that between an unusual form of development and a deformity, is a real one, is easily shown. It takes but a moment's consideration to see that such a case as that of the Siamese or other conjoined twins does not belong in the same class with an acephalus, or with a monster showing distortion or truncation of limbs, since the former shows a perfectly normal development in respect to bodily symmetry and the normal condition of the organs and tissues. It is an unusual type of being, but is not a deformity or malformation ; and from this standpoint it is but a step to include also all double monsters formed of equal components, the "diplopagi" of my previous paper (1904). At that time I made a sharp distinction between those double monsters in which the components are equal and those in which one component is more or less reduced (i. e.. '"parasitic" monsters), but since then I have changed my views on this point and include them with equal diplopagi, a point concerning which I shall have something to say later on. Passing this over for the present, however, I may say that what I then recognized of the definiteness and order characterizing the structure and development of diplopagi has been corroborated and emphasized by the opportunities I have since had of investigating many more cases, and their symmetry and regularity in anatomical details have led me to insist upon a sharp distinction between them and other forms of anomalies and to look upon the former as beings as orderly and perfect in their development as are the usual and normal types of being. AhnormM they certainly are in the sense of not being the usual form in which a given species manifests itself, but they are not deformed.

Furthermore, it seems also necessary to extend this distinction lietween orderly and deformed beings so as to include, not only diplopagi with both equal and unequal components, and normal individuals, but also


The Morphology of Cosmobia 357

the numerous cases of primarily symmetrical beings that are less than a normal being, such as cyclocephali and symmeli (eyclops and siren monsters). These, like the diplopagi, are represented by complete series of forms which connect the extreme cases with the normal by imperceptible gTadations. The similarity of the two sets of symmetrical abnormalities that lie upon either side of the normal is apparent from the exact correspondence in detail between similarly incomplete members of both sets, such as a median double leg or a median double eye, ichether they represent the two normal components as in the one case or the two supernumerary ones as in the other.

This recognition of the kinship between these forms of defective monsters and diplopagi seems an important logical step, since it enables us to construct an almost unbroken series (or rather several related series, differing in their geometrical relations) which begin at the most defective evclocephalus or symmelus; progress step by step until the normal condition is attained, and then again, passing this point, run through the various grades of diplopagi to the stage represented by separate duplicate twins. There are even suggestions of possible extensions of the series beyond this point, for we have not only identical triplets and still higher numbers of separate "duplicate" individuals, but there are also certain cases of monsters which suggest intermediate stages between these and duplicate twins, such as cases of twin births in which one is a normal individual and the other a double monster^, or cases in which a multiplication involving a part of the organism has progressed beyond the degree of two components and thus represents a stage between two and three components. This series may be stated as follows, it being borne in mind that the

inter-relation of the components may involve more than one geometrical

possibility and that the phenomenon may not include the entire bodv : —

1. Cases in which the entire individual, or the part involved, is

less than a normal individual.

Va) The "Monstrum Anglicum," born at Fishertou Anger, near Salisbury, England, in 1664 (to Mrs. John Waterman). This was an imperfect female isfhiopagus, with legs upon one side only. At the same birth there was a normal daughter who lived to grow up. Licetus, ed. 1665, p. 316.

(b) Fisher's case 43, born near Berlin, Germany, 1773 (to Frau Anna Maria Woblack). This was a male of the "Toeci" type (A VI of my Table, '04, PI. A). At the same birth there was a normal male child.

Both of these cases are rather old. but seem to have been accepted by later teratologists.

Forster, Taf. IV, Fig. 12, Tricephalus.


358



Harris Hawthorne Wilder

n III




VIII



Fig. 1. Diagram showing a related cosmobiotic series.

Stages I-V, Various degrees of Cyclopia.

Stages VII-IX, Normal beings.

Stages X-XIV, Various stages of Diprosopy.


The Morphology of Cosmobia 359

2. A normal individual.

3. Cases in which the entire individual, or the part involved, is

more than one individual and less than two.

4. Separate duplicate twins, both normal.

5. (a). Separate duplicate twins, one of them a diplopage.

(6), Cases involving a single part, in value between two and three components.

6. Separate duplicate triplets.

Beyond this lie the possibilities of identical quadruplets or even higher numbers, with a tendency to a farther partial duplication on the part of one or more of them, thus continuing the series almost indefinitely.

This idea may be best explained by taking the details of a given organ in an actual series, such a one, for example, as is furnished by the eyecomponents in a series that includes the various degrees of cyclopy, diprosopy, and normal beings (Fig. 1). In this we may begin with the most reduced type of Cyclops, and find, externally, a mere palpebral slit, symmetrical in outline and with, perhaps, a double set of lacrimal punctae. This may be followed by cases in which the palpebral opening becomes gradually larger and the median doiable eye-ball more and more visible. Beyond this the progressive stages are shown by changes upon the eye-ball itself, which becomes continually more visible through an increased palpebral opening; first, a double pupil in an oval iris, two distinct pupils on a figure-8 iris, a double iris on a single sclera, two distinct irides, and finally an eye-ball which is plainly double. (Stages I-V.) In all of these cases the nose, which is prevented from coming down in the usual manner through a downward growth of the fronto-nasal process, remains above the double eye and presents a shape something like a proboscis, decidedly abnormal, but characteristic of all monsters in which there is no space between the eye-components. [Cf. the imperfect face of an unsymmetrical Janus.]

In the next stage, however (stage VI), which closely approaches a normal type, the eye-balls are distinct and a small and narrow nose rudiment, usually with a single median nostril, succeeds in pushing its way down the narrow interval between them, and thus appears in the normal position.

To continue the series further it is necessary to select several types of faces which may be found at any time in an average street crowd. We may begin with an individual with the eyes unusually close together and with the thin file-like nose that always accompanies such a condition


300 Harris Hawthorne Wilder

(stage VIII), and end with one that possesses a very broad face, has the eyes far apart, and a broad nose between them (stage IX).

Between these extremes we may find all perceptible gradations, but for our purpose here but one will suffice, an average human face, with eyes and nose of average proportions (stage VIII).

But the series does not end here even, for just beyond the face with the eyes far apart occurs a stage in which there is a slight doubling of the nasal septum or of some other parts of the nose (stage X) ; then one with two distinct nasal components and a minute palpebral opening between (stage XI). This median eye is in all points similar to that of the type with which we began the series, save that here, if we should analyze the case more completely, we would find that the components of this median double eye would have their potential outer sides together, while in the former case the relation of the eye components is that of a normal pair of eyes, with the inner aspects together. These relations become apparent if we compare members of the series a little more doubled, with the two components of the double eyes more completely developed.

After this the stages that follow, not all of which are shown in the diagram, repeat, in respect to this median double eye, the external appearances shown in the first part of the series; there comes first the doubling of the pupil, then that of the iris, and lastly that of the entire eye-ball, although internally there is always the reverse relationship of the muscles and nerves, as will be shown later on. After the eye-ball is completely doubled there follow several stages in the gradual separation of the ear components, and finally two complete heads upon one neck, a typical Dicephalus. The relations of this form of monster to other forms of diplopagi are too well known to necessitate repetition here, but the reader may be referred to my former paper on the subject, in which is shown a farther continuation of the Dicephalus series, and the relation of this line to the various forms of Ischiopagi.

That such a complete series may be made by the use of both normal and abnormal types, while not in itself constituting a proof of any real relationship between them, is still highly suggestive of a similar cause at the basis of all, and that one which is fundamental, most probably existing in the germ itself. The exact similarity of form, even to minute anatomical details, between parts of the same degree of development on either side of the normal, that is, in both "defective" and "excessive" mon>ters (monstra in defectu et monsira in excessu of the older teratol


The Morphology of Cosmobia 3G1

ogists), leads to the conclusion that both sorts of monsters are clue to the same cause or kind of cause, and that they should be considered together in any general treatment of the subject, especially in all discussions concerning the cause of these monsters. It seems, furthermore, that all members of this series, both normal and almormal, are equally subject to definite and orderly laws of development, the impulse to the formation of which lies, in the one case as in the other, within the organism, and leads in all cases to the formation of beings which are primarily symmetrical and free from all pathological tissue or anything which is out of harmony with the organism as a whole. It must be remembered, however, that the word "primarily" is always to be understood in connection with the above statement, for even in the case of perfectly normal germs, later causes, mainly external, may lead to very great deformities, through which the resulting organism is led to deviate from the original goal. Since this is often so in individuals primarily normal, where all the conditions of development have become long adapted to an embryo of a certain definite form and size, how much more likely to become secondarily deformed must be an organism unusua] and unwonted in these particulars? If we consider the perfect adaptation of the uterus, the placenta, the yolk-sac, the egg-shell, and the other adjuncts to development that appear in the different classes of vertebrates, we wonder that any embryo of abnormal shape can ever attain an advanced stage without secondary deformations rather than that some of them should become thus. It seems also very probable that monsters vary in their susceptibility to secondary deformation, and that, while certain types usually come to maturity and are even viable, others may inevitably encounter some adverse mechanical principle at an early embryonic stage. Thus the deformed proboscis that represents a nose in several of the types of the series in question, probably develops without deformation up to the time at which the descent of the fronto-nasal process should begin, and the deformity that then begins to make its appearance, although an inevitable one, is due to no deficiency in the germ, but to an unfortunate mechanical relationship which appears at this time Should an embryo of such a monster ever be obtained at a time before the fronto-nasal process begins its downward growth it is safe to predict that there will be nothing in this region which may be considered a deformity, but that the parts will be svmmetrically and orderly arranged as in the case of those parts which are not hindered during development.


362 Harris HaAvthome Wilder

In carrying on the study of these forms from this standpoint it becomes thus a matter of no moment if or at what time a secondary deformation occurs. If we have either a true double monster or one of a symmetrical series less than a normal individual we must assume that any lack of symmetry or other deformation is secondary in nature and that the embryo was not deformed at first. We must try to look through all such deformations, which by the very nature of the case are bound to occur frequently, and endeavor to find what was the essential condition of the original type which Nature attempted to produce, to learn the intention of the germ, if the expression be allowed. To do this it will be well to consider the nature of the causes which may produce secondary deformation of an abnormal embryo.

Naturally the chief of these in cases where the components are together greater than a single individual, is lack of room, and the commonest result of this disadvantage would naturally be the reduction in size of the less favored component, producing a monster which would be classed under the head of "autosite and parasite," the "parasitic monster" of most authors. In my previous paper I followed the usual custom and sharply distinguished this sort from "true diplopagi," i. e., symmetrical ones. This view I at present reject, and, while not quite prepared to accept all cases of parasitic monsters as deformed instances of primarily symmetrical ones, I feel sure that the most of them are, and that they difi'er from symmetrical monsters merely in the accidental conditions to which they have individually been subjected during development. The criterion of a true diplopage which should in all cases be insisted on is that of homologous union, that is, that the parts of each component by which they are united to each other should be anatomically the same, a criterion which is indeed difficult of application in cases in which the lesser component is very much reduced, but which is evident in by far the greater number of cases. This would still leave open the possibility of the occurrence of other forms of association, such as that of the secondary fusion of two blastomeres on a common yolk, in which the points of union would be the chance points at which the two embryos first came in contact with one another, and would not be homologous. This latter form of monster is probably common among the Sauropsida, since the occurrence of such fusion in the early stages has been frequently observed, although it is not likely that these cases would be able to develop far. A similar secondary fusion of two geometrically imrelated embryos might account for cases of included fetus (fetus in


The Morphology of Cosinobia 363

fetu), and I would wish to remain noncommittal for the present in regard to the majority of dermoid cysts and other embryomata where the parasite is too amorphous to apply the test of homologous union. HoAvever, excepting all these doubtful cases there still remains a large class in which the lesser component is properly related to the greater to constitute with it a primarily symmetrical diplopage, though secondarily deformed.

A second cause of deformity, at least as regards bilateral symmetry or equality of components, and one which is especially operative in assisting in the secondary deformation of a diplopage, is found in the striving among the parts during growth for the best physiological efficiency. In a vertebrate embryo certain of the organs, especially those of circulation, and, to a lesser extent, digestion, are physiologically active from a very early period of development. The former, for example, is early called upon to solve certain mechanical problems connected with the transportation of the blood, and although much of the general arrangement of these parts is probably inherent in the germ, the details are mainly left to the exigencies of the particular eases, as is shown by the great amount of individual variation in the adult of a given species, especially in the smaller, later appearing vessels. Now in those diplopagi in which the heart is represented by two separate components, which yet form parts of one system, the problem is presented in a more definite way than in the case of two rival blood-vessels that supply the same part, since there are here two pulsating organs to direct one circulation. It is inevitable in such a case that one of the hearts should early become a little stronger than the other and gain either a control of vessels beyond the median line separating the two components or else secure for its vessels a little more of the blood; in either case the result would soon show in a lack of symmetry between the two components or between corresponding parts, although at first the two components were exactly equal. This point will be brought out farther on by a comparison of the circulation in several cases of the gi'oup known as "Janus" monsters, where the asynmietry in the circulatory system is plainly of a secondary nature and due to the causes above outlined. As an instance of secondary asymmetry due to the rivalry among digestive systems I may cite the case of a two-headed lamb (Teras XV of my collection), which had come to full term and lived and fed for perhaps four weeks. Aside from the doubling of the head it appeared absolutely normal, i. e., single, and tlie two heads were exact duplicates, of each other and per


364 Harris Hawthorne Wilder

feet in development.^ It had, however, probably by chance, used one mouth exclusively in suckling and that exercise of functional activity, even during the brief life of the animal, had given a slight twist to the neck so that the feeding head came to be carried as if it were the continuation of the median axis of the body while the head that took no nourishment lay a little to one side. It cannot be doubted that the persistence of this treatment through several years would have increased this tendency so that in the adult state a person who knew nothing of the early conditions would have classified the case as that of a lamb with a parasitic head.

Since, now, in the case of Sauropsida and Mammalia, so much of the development is gone through with before the incident of hatching or birth, it is to be expected that in most cases of primarily symmetrical monsters there will be found secondary differences at least in the internal organs, at the time of birth, and that these differences will be greatest in those systems which are actively functional from an early period, such as the circulatory and digestive systems ; while the least annount of modification is to be expected in such systems as the skeleton and muscles. That such is actually the case will be seen in the descriptive part of this paper where the numerous modifications in the circulatory system of otherwise perfectly symmetrical monsters will be contrasted with the symmetry of median eyes, formed from two components, where each muscle, nerve, or other part is repeated on the two sides with at least as much faithfulness as in the two sides of a normal bilateral being. In the case of the circulatory system the modifications would naturally affect mainly the heart and the main vessels, while the arteries distributed to symmetrical components would be as regular and symmetrical as the nerves or muscles. [Cf. the cephalic arteries in double-headed monsters, as shown by Miss Bishop.]

A third cause, or rather a large class of causes, producing secondary deformation of a primarily symmetrical monster, and one likely to induce all sorts of pathological conditions, is the mechanical hindrance to the carrying out of a given plan of development through the interposition of some organ which cither encroaches upon the space required for sometliing else or actually blocks the way and renders farther develop ^This specimen, Teras XY, was one of those used by my pupil, Miss Bishop, in lier paper upon the arteries of dioephalous monsters. The exact equivalence of the two head components is well shown in the equivalence of the arterial supply.


The Morphology of Cosniobiu 3G5

ment in a certain direction impossible. Such is the case of the proboscis-like nose given above. Still more serious M'ould be the difficulty if the encroaching organ should hinder the full efficacy of some system of functional importance to the embryo, as for example the circulatory or lymphatic organs. Thus there might arise those evidently pathological beings that fill the pages of general works on teratology, yet in which may still be traced a definite original plan capable of finding a place in some series of symmetrical and non-pathological beings. It is probable, indeed, that certain types of primarily symmetrical monsters have such a mechanical configuration as to render development impossible without becoming secondarily pathological through this cause, and in such cases one must either learn to eliminate the secondary modifications as remarked above or, perhaps, by some fortunate chance, as rare as the discovery of some long-sought link among the strata of the earth's crust, obtain and study an embryo of the type in question representing a stage previous to the appearance of the causes of the pathological condition.

Aside from the above and probably other causes affecting the development of abnormal embryos in general, certain groups of vertebrates are undoubtedly subject to special hindrances due to some peculiar and specific mode of development. The result of this may be that only certain t3pes are possible within a certain group of vertebrates; hence, in completing a given series of symmetrical monsters it may easily happen that a certain stage sought cannot be found in mammals but may be of frequent occurrence among amphibians, or that a stage in the series that always appears in a secondarily distorted condition in birds may be found without such modifications in reptiles. This suggests an explanation of the fact that certain types of monsters are of far more frequent occurrence in some animals than in others, a field of inquiry that would undoubtedly yield much if it were carefully investigated. For example, birds have at least two developmental peculiarities that would tend to modify the development of monstrous embryos, namely, the enormous yolk and the early twisting of the embryo, and as a matter of fact, avian monsters, when hatched, are more limited in variety than those of mammals and are usually unilateral and somewhat distorted. All of these suggested causes of malformation have this in common, that they exert their modifying influence at some time during embryonic life upon what must be at first an undeformed, though abnormal, embrvo. It follows, therefore, that we would eliminate these secondary modifi


366 Harris Hawthorne Wilder

cations if we could only study these types of monsters as early embryos. Such material is^ however, all but impossible to obtain, but with the gradually increasing use of the embryos of various vertebrates in general laboratory work in colleges and universities, an increasingly large number is yearly sorted over, and thus the chance of such fortunate discoveries is always on the increase. Embryo avian monsters are freqiiently met with, but these, for the reasons above cited, are rather unsatisfactory material for work, although Kaestner has turned these very disadvantages into points of much significance and has obtained therefrom highly important results. Early mammalian embryo monsters, of the types referred to in this paper, are almost unknown, but since in my collection I already have two of them, Terata III and XXX, (see p. 369, foot-note; also Fig. 27) such research appears to be quite within our reach.

Aside from the obvious advantages of getting rid of secondary modifications, the study of "teratembryology" offers another, which is very great. Just as certain of the types which belong naturally in a teratological series are bound during later development to meet with certain mechanical difficulties which prevent them from surviving birth, the so-called "non-viable" monsters, it is also very probable that others which exist so far as we know only in a theoretical series actually begin embryonic existence but meet with difficulties which set a term to their life while still embryos. Such forms would then be expelled in a disintegrated condition, or absorbed, and would thus never come within the ken of the teratologist.

If now we may assume that there is a class of monsters which are primarily as symmetrical in structure and as normal in their tissues as are the beings we usually consider normal, and if we may hold that they, as well as normal beings, owe their structure to some germinal variation, there is great need of a distinct term, which is broad enough to include both these forms of monsters which are merely deformities and pathological cases. In defining this term a set of duplicate twins, whether separate or conjoined, should have the value of a single unit.

For such an organism (or set of organisms), whether normal or abnormal, whether less or more than or equivalent to a normal being, and lastly whether perfect or deformed, provided the deformity is due to a secondary cau^e as explained above, I propose the name Cos:\roBiON' (plural CosmobiaJ^ an orderly living being. In this term, the meaning of which exactly expresses my idea, the only violence to classic


The Morphology of Cosmobia 367

Greek seems to lie in the formation of a neuter noun, /3tov, a living thing, from the abstract masculine, /8ios, life; a formation which is abundantly supported by analogy, and fully in accordance with the genius of the language.* Another possible criticism of the word lies in the fact that the word Kdo-/Aos, order, has been commonly used, even by the Greeks themselves, in its derived meaning of the Universe. The literal meaning is, however, exactly what I desire, and in addition to being euphonious and simple, it readily admits of the adjective form CosMOBioTic and is applicable to plants as well as to animals if desired. Whether such forms exist among plants I cannot say, but, in order to be analogous, they must belong to the sexual generation and hence be looked for among the lower Cryptogams.^

The theory which I have here set forth, and which I may call the theory of CosmoMa, is not wholly a new one, since the recognition of certain series of monsters with imperceptible gradations between them has been pointed out by teratologists for a long time. What may be considered new is, first, the recognition of the relationship between the symmetrical anomalies on either side of a normal being; the inclusion among Cosmobia of certain types of secondarily deformed and misshapen monsters resulting from abnormal conditions during development; and, thirdly, the possibility of considering in a single series both these forms with less and those with more than the normal number of parts, including also normal beings. Whether either of these points may be in accord with the actual facts can only be shown through much investigation of the structure of all sorts of Cosmobia, together with experimental work on the artificial production of these forms, a field which, though often tried, has until recently yielded but little.® At the best my theory can be

We have already the exactly analogous word Rhizobia, the organisms living in the root nodules of leguminous plants.

"The various sorts of double or twinned flowers, fruits, and other parts of plants are phenomena of quite a different kind from cosmobiotic organisms, since the appearance is localized and affects certain parts only. These may perhaps be compared to the doubling of fingers, limbs, or other lateral parts among animals, which, although they may be germinal in origin, do not modify the entire organism.

'As far as can be learned from figures ana descriptions of artificiall pyroduced monsters, they all seem to be merely cases of deformation, with no genuine Cosmobion among them. In the case of the reported artificial production of "double monsters," lilje those claimed by Panum or Gerlach, it is probable that these were simply natural cases of monsters, quite frequent among birds, and not due to the experiments. Experiments thus far seem to


368 Harris Hawthorne W ilder

notliiug more than a working hypothesis, which may point out the direction in which needed work may be done, and its establishment, modification or refutation are a matter of indifference so long as our knowledge of the subject is advanced. Concerning the probability of each of the points suggested the reader must judge after the consideration of the following report, which presents the results of my investigation of the subject.

As material I have had the opportunity of studying the following monsters ; the numbers given in connection with them being those under which I have filed them in my notes. These will be used for convenient reference throughout the paper.

I. I'ig embryo; Diprosopus tetrophthalmus. [Lambert.] II. Two-headed snake (Storeria). [Davison.] III. Human Synote (imperfect Janus), advanced fetus. [Baldwin.] lY. Human Thoracopagi, one parasitic ; advanced fetuses. [Wistar Inst. Coll. No. 2884.] This specimen is figured by Hirst and Piersol, Part IV, Pis. XXXVIII and XXXIX. ' V. Chick embryo, two bodies, head apparently single, but imperfectly formed. [Smith College Laboratory, incubated.] VI. Human Cyclops, child at term. [Wistar Inst. Coll. Xo. 6956.] VII. Pig Cyclops, large fetus. [Wistar Inst. Coll. Xo. 2913.] VIII. Human Paracephalus. [Wistar Inst. Coll. Xo. 4926.] This specimen is figured by Hirst and Piersol, Part III, PI. XXIA^. IX. Pig embryo, perhaps a Cyclops, but very imperfect. [Lambert.] X. Human Omphalopagi, advanced fetuses. [Wistar Inst. Coll. Xo. 4996.] This specimen is figured by Hirst and Piersol, Part IV, PI. XXXV.

emphasize the position talcen here, that true oosniobia of all sorts, normal, excessive, and defective, are due to a cause existing in the germ, or applied during the very early stages of development, and it is there that our efforts should be directed if we may hope to produce such an organism artificially, a result that we can hardly expect to reach by purely mechanical means. As the tendency to produce duplicate twins and other sorts of abnormal cosmobia seems inherent in certain organisms, and to be transmitted by heredity, it Is quite possible that we may be able to breed certain of the viable forms. In this connection the remarkable experiments of Stockard, which are now being carried on, are of the greatest moment, and will yield important conclusions.


The Morphology of Cosmobia 369

XI. Chicken, newl^y hatched, Diprosopus triophthahiuis, also with four legs, the inner ones smaller. [B. G. Wilder.] XII. Pig, new-born, Diprosopus triophthalmus. | B. G. AVildei!.] XIII. Duck, adult, with a supernumerary leg. [B. G. Wilder.] XIV. Pig, new born (?), two bodies in the form of Omphalopagi, but with a single head with double tongue and lower jaw. [Lambert.] From the Barnum Museum of Tufts College. XV. Lamb, three or four weeks old, with two equal heads. [B. G. Wilder.] Born at Ludlowville, IST. Y. XVI. Chick embiyo, perhaps a "Janus." [Smith College Laboratory, incubated.] XVII. Kitten, new-born; parasitic thoracopagus. [Mead.] XVIII. Chick embryo, perhaps a "Janus." [Gorham.] XIX. Chick embryo; double primitive streak. [Gorham.] XX. Turtle, small Chrysemys; dicephalus. [Mead.] XXI. Chick, a few days after hatching; double median leg in addition to the normal pair. [Lambert.] XXII. Chick, a few days after hatching; a supernumerary leg upon

one side. [B. G. Wilder.]^ Of these monsters, which are listed in the order in which they have been received, Xos. VIII, XIII, and XXII are probably not Cosmobia, at least not typical ones, and are hence not considered in any way in the present paper. The others may be placed in groups as follows : —

A. Janus-Omphalopagus group : Xos. Ill, IV, Y, X, XIV, XVI,

XVII, XVIII.

B. Diprosopus gi-oup : Xos. I, II, XI, XII, XV, XIX, XX.

C. Cyclops group : X"os. VI, VII, IX.

From this material it is my purpose to select for the use of the present paper certain topics, the consideration of which may be of direct bearing upon the theory above enunciated rather than to attempt an anatomical description of the several forms, a kind of work of which

'Since the coiupletion of this paper I have received froui Professor B. (.}. Wilder a large and valuable collection of monsters, which arrived too late for incorporation in this list. One of them, however, Teras XXX. is of such value in this connection that it deserves special mention. This is a kitten embryo of 34 mm., with a partially doubled head, and practically the counterpart of Dr. Lambert's diprosopic pig, Teras T. This I have already sectioned and hope to use it later in connection with further work on Teras I. I am also inserting into this paper a drawing of it in connection with the description of this latter monster. (Fig. 271).)


370 Harris Hawthorne Wilder

there is already an abundance among teratological literature. Two such topics will be treated in this paper; the first of which is a general description of Teras III, with especial treatment of the auditory ossicles and the circulatory system. This monster belongs to the type called "Synote" and I shall refer to it as the "Baldwin Synote" from the donor of the specimen. The second topic is a comparison between various types of double eyes, as they occur in all three of the groups given above. In a work of this nature, more than in perhaps any other, the investigator is dependent for his material upon the thoughtfulness and generosity of others, and it is thus with especial earnestness that I wish here to personally thank those who have so kindly aided me in this respect. Dr. F. D. Lambert, who inspects many hundreds of pig embryos yearly, has sent me everything abnormal which has come under his inspection, and I wish especially to mention the beautiful specimen, No. I, which came to me faultlessly preserved, so that I was able to section the entire specimen and make wax-plate models of several of the parts, an opportunity that has seldom if ever come to the teratologist. The human Synote, ISTo. Ill, also beautifully preserved, was given me by Dr. James F. Baldwin, the Chief of Staff of Grant Hospital, Columbus, Ohio, who preferred to have the specimen used for scientific investigation rather than to keep it "simply as a curiosity." In using such a valuable specimen I feel the responsibility for good results and only hope that they will justify the sacrifice. An especially valuable aid in my investigations has come from the Wistar Institute of Anatomy, in Philadelphia, largely through the kindness of its director. Dr. Milton J. Greenman, who has loaned me numerous specimens for study and even dissection. Prof. B. G. Wilder, of Cornell University, has been untiring in obtaining for me Cosmobia of various sorts, his most valued contribution being perhaps the diprosopic pig, No. XII, which, delivered at term, belongs in the same series as Dr. Lambert's beautiful specimen. No. I, and thus enables me to trace the later development of this form of monster; the fact that the two differ slightly in the degree of separation of the two components enhances the value of both specimens.® The two-headed snake, Storeria occipitomaculata (Teras II) was sent me by Dr. Alvin Davidson, of Lafayette, and Drs. Mead and Gorham, of Brown University, have furnished me with several valuable specimens.

•One of Professor Wilder's latest specimens, Teras XXX, mentioned in a previous footnote and figured here as Fig. 27&, will perhaps vie with Teras XII in point of value. This came too late for proper treatment here, but will serve for later research, especially in connection with Teras I.


The Morphology of Cosniobia 371

Part 1. The "Baldwin Synote/' a human monster of the imPERFET Janus type.

A. General consideration of the Janus-Omphalopagus Series. This is an extensive natural series of Cosmobia, certain of the stages of which are among the most frequent of mammalian symmetrical monsters. In this series the two components are placed vis-a-vis, that is, with the ventral sides applied to each other, and with always a common umbilicus. Individuals of this series may vary in two ways, not dependent upon each other; (1) in the extent of longitudinal union, and (2) in the amount of lateral torsion of the two components in respect to each other. These will be considered in order.

1. Variations in the extent of longitudinal union. The posterior limit of this union is always fixed by the umbilicus, to which it extends in all cases and beyond which it never goes, leaving, posterior to this point, in undeformed Cosmobia, two perfect hinder parts, facing each other. Even in cases with extreme involution of one side (as described under (2)) this peculiarity becomes less and less posteriorly, and by the time the umbilical level is reached the components exactly face each other.

The stages in such a series are shown in the accompanying diagram. (Fig. 2.) Each stage figured rests upon actual cases, and typical examples are indicated by the name of the authority and the date of the descriptive paper, so that the cases can be easily found by refernce to a teratological bibliography. As the diagram represents merely the extent of union longitudinally, each case is represented with the components exactly opposite, that is, with no lateral torsion. The series thus begins with a "Janus symmetros," in which the line of union begins at the vertex and extends, as in all the cases, to the innbilicus. In the second stage the brows, eyes, and noses become free, although the latter may be hindered from full development through lack of room. In the third stage the face is free as far as the chin, and in the fourth the heads and necks are free and the union begins at the manubrium sterni. In the stage following the union begins at the mesosternum, but in the next it involves the xiphisternum alone. The extreme type of this series is one in which the entire sternum is free and the soft parts alone are involved. Such were the famous "Siamese twins," Chang and Eng.

Beyond this there is introduced a stage in which the united parts include only extra-embrj^onal structures, as, for instance, a portion of the umbilical cord or even the placenta alone. Such a case would be a true Cosmobion, but as the united structures are cast away at birth


372


Harris Hawthorne Wilder






YIII



Fig. 2. A related series of Diplopagi, the "Janus-Thoracopagus" series (monophaliens, G. St. Hilaire) : drawn in diagrammatic form from recent cases, as follows :

I. PROSOPOPAGUS. The "Baldwin Syuote," described in this paper. II. GNATHOPAGUS. Williamson. 1895. III. TRACHELOPAGUS. Fkaseb, 1890. ly. STERNOPAGUS. Sebart, 1896-97. V. THORACOPAGUS. Scott, 1889. VI. XIPHOPAGUS. BoETTCHEB, 1871. VII. OMPHALOPAGUS. The "Siamese Twins," b. 1819; d. 1874. VIII. URACHOPAGUS. Any genuine case of "Duplicate Twins." In the last case (VIII) the imited part does not usually involve the free portion of the umbilical cord, but is confined to the placental portion. If the portion of cord shown be considered to represent the entire extra-embryonal part of the embryo, the diagram correctly expresses the condition. Such twins are thus properly double monsters in which the parts in connnon are extra-embryonal and are cast away at birth.


The Morpliology of Cosmobia 373

there results a pair of free components, in short, duplicate twins ! During early embryonal life, when there is no clear distinction between embryonal and extra-embryonal parts, it would be without question a Diplopage, but the subsequent abandonment of those parts in which the union occurs leaves the two comporients free from each other, a ditference of no moment to the morphologist but a most vital one from the standpoint of the individuals concerned.

As the terms that have been applied to the various stages of this series are numerous and varied, I have hoped to simplify matters a little by suggesting a series of terms, the most of which are already in use, which designate the condition in the simplest way; namely, using in each case the Greek term for the point at which the union begins anteriorly, joining to this the usual term for a monster with two components, — pagiis, (-n-qytvixL) . The term omphalo- is often added to the compound, as "Thoraco-omphalopagus," but this is hardly necessary as the umbilicus must of necessity always form the posterior limit of such monsters. We have thus the terms, Cephalopagiis, Gnatlwpagus, etc., as given in the diagram, a set of terms which may be freely added to if we wish to express grades between those given.

2. Variations in the amount of lateral torsion of the two components; with involution of certain of the parts upon the less favored side. It rarely happens that the components so exactly face each other in the head region that the two resulting compound lateral faces are perfectly equal. Such a case is called a "Janus symmetros" and is one of the greatest teratological rarities. Usually in this region there is what may be termed a lateral torsion, with the result that the compound face upon one side is complete while that upon the other suffers a greater or less involution, that affects the median parts at least, but, with a greater degree of involution gradually affects parts that are more laterally placed. The series formed by this principle is shown in diagrammatic form in Fig. 3, which begins with a Janus symmetros and ends with so complete a degree of involution that the external features of the face are entirely suppressed upon the imperfect side, giving the compound head the external appearance of a single one, facing in a direction that is really lateral in respect to the actual position of the components. This extreme condition is that realized by my Teras XIY, a pig with eight legs, two bodies, and apparently a single head placed sideways with respect to the bodies. In all cases the torsion is most pronounced anteriorly and becomes gradually less towards the umbilicus, at which point it disappears and the two bodies are opposite each other.


[■i


Harris Hawthorne Wilder


B. The Baldwin Synote and its place in the series. This monster [Plate I] is an "imperfect Janus" (Cephalopagus) of the degi-ee of torsion represented by Stage V of the diagram [Fig. 3]. The face



Fig. 3. Diagram showing different degrees of lateral torsion (involution) of monsters of the Janus group.

I indicates the relation of the two components ; the others represent a^ual cases, as seen in cross-section at the level of the eyes. II is a "Janos symmetros," in which the two components are exactly opposite, and the two faces are consequently equal and perfect. Ill to V are the common types, in which one face is more or less suppressed, thus suggesting the employment of such words as "anterior" and "posterior," "front' and "baclv," etc., to express the two aspects. Type IV is designated as an "Iniops," Type V as a "Synote." In Type VI one face is wholly suppressed, at least externally, and appears like a normal head to which two bodies are attached laterally.


of one side is complete, while the features that appear upon the other are (1) a large nasal proboscis, consisting of a stalk and a larger terminal bulb; (2) a minute palpebral opening, but with externally no


The Morphologj^ of Cosmobia 375

trace of an eye-ball; and (3) the two external ears, meeting each other in the median line, whence the name Synote.

For convenience in description and in order to emphasize the compound nature of many of the structures, it will be convenient to designate the two components as A and B, A being the one on the left when the specimen is held with the more perfect side towards the observer. This rule, which can be readily understood by the help of the illustrations to this article, which are designated in accordance with this system, is applicable in the case of almost every specimen belonging to this series, since cases in which the components are placed exactly opposite each other are extremely rare, and since the rule can be easily applied whenever there is the slightest lateral torsion. If now the right and left lateral aspects of each component, viewed by itself, be designated by the letters r and I, the composition of a given compound organ may be indicated as Ar-j-Bl or Al-|-Br, as the case may be. Thus the "perfect" face, in reality composed of half faces contributed by the two components, has the composition Ar+Bl, while that of the imperfect side, of which a few features alone appear, is composed of the two half faces Al-]-Br. This would indicate, for example, that the two external ears, which are situated together upon the imperfect side, consist of the left one of A and the right one of B, as a moment's inspection will prove.

As with all monsters of this type, the two components become more nearly opposite each other posteriorly, and at the level of the common umbilicus they are practically quite opposite, that is, their median sagittal planes coincide. This may well be shown by a comparison of the parts at different levels, beginning with that of the mouths. The mouth of the perfect side leads into a pharynx of normal appearance, furnished with a good tongue, while the imperfect side has a much narrowed pharynx, without external opening, and fitted with a narrow, tapering tongue. Tongues and pharynges thus correspond to the two apparent faces, the apparently normal ones for the perfect face and the narrow ones for the imperfect face. Each is thus not an organ belonging to a single component, but compound in the same way that the faces are. The same relationship obtains in the case of all organs that face the same way as these, namely, towards the sides of the components.

In the center of the transverse pharyngeal plane there lies a single common oesophagus, into which both pharynges open. Upon the apparent ventral side of each neck, lateral in respect to the components, there is placed a larAmx, one upon either side of the oesophagus. Each


376 • Harris Hawthorne ^^'ilcle^

is apparently a single organ, Ijut that of the imperfect side is narrower than normal, yet not incomplete, while that of the perfect side is apparently a normal one. From each larynx a trachea leads down into a pair of kings, the pair npon the perfect side being in reality Ar-f-Bl and that of the imperfect side, Br-|-Al, reading in each case from left to right, as is natural.

In the chest region the imperfect side is being gradually rolled out and is thus becoming more nearly the equal of the other. This may be readily shown by ascertaining the degree of inrolling at two levels, for example, (1) at the acromia and (2) at the nipples, and comparing th^ two. The distance l^etween the acromion processes of the perfect side is 7.1 mm., and between those of the imperfect 48 mm. In the case of the nipples, which are situated at a lower level, the distance between them on the two sides is 36.5 mm. and 29.5 mm., respectively. If in each case we consider the measurement for the perfect side as 100, that for the imperfect side at the level of the acromia is but 67.6 per cent of it, while at the nipple level the measurement of the imperfect side is 80.8 per cent of that of the perfect one. The sternum of the imperfect side shows this change throughout its length, as it is somewhat narrowed anteriorly and a little rolled in, as if by the close approximation of the shoulders, but posteriorly it is practically like that of the other side.

The alimentary canal well illustrates the principle of the secondary modification of an organ through embryonic functional activity, since for some weeks previous to birth the intestine has been made the receptacle for the meconium and has had more or less to do with the function of nutrition. The common oesophagus leads into a common stomach, though evidently one formed of two components, since it presents two cardiac enlargements, one on either side of the cesophagus. Tl o outline of the stomach is thus heart-shaped, but is not quite symmetrical, since the cardiac lobe of component A is a little larger than that of B. The stomach leads through a single pylorus into a common intestine, whicli continues single throughout about three-quarters of its length (80.9 cm.) and then divides into two, which extend to the two colons located in the separate components, the connecting piece of A being 29.3 cm., and that of B 25 cm. in length. Beyond these extend the cseca, appendices, and colons of each individual, that of A being 21.6 cm. in length, and that of B 27 cm. As the intestinal canal is a common one from the united phar^^nges to the bifurcation at the lower fourth of the small intestine, the decision as to physiological function comes at this latter


The Morphology of Cosmobia 377

point, aud, owing undoubtedly to tlie chance of gi'avitation or of some temporary mechanical condition, such as a fold of the gut, the first of tlie fecal matter selected the branch belonging to component B as its functional outlet. B's intestine and colon therefore became filled and distended from this point on, while those of A, although perfectly normal, remained empty. In one feature there is a secondary deformation, and



21.6 cm.


27.0 cm


29.3 cm.


25.0 om.

Fig. 4. Alimentary canal of tlie Baldwin Syuote, Teras III. The measurements given were taken between tlie points designated with an X.


that is, the entire intestine of A, beyond the bifurcation, has pinclied itself off from the rest, leaving a short stub at the place where the intestines have become individual. It seems probable that this cutting off occurred subsequent to the adoption of B's canal for functional activity, although there was probably some direct connection between the two events. Perhaps the pinching off occurred as the result of inactivity, or perhaps, which is more likely, it was the final result of a restriction


378 Harris Hawthorne Wilder

or some other slight cause responsible for the original decision. It would be of much interest to examine the intestines of other Jani and see whether one side is preferred for the meconium, whether, if this he the case, the unused side becomes separated from the other, and whether the same side is alwa^ys preferred. This latter point could of course be determinable only in cases with some lateral torsion, and in view of the varying relation of the aortic arches in different specimens, as described below, it is most unlikely. The cutting off of A's intestine, in view of its perfect development otherwise, appears to have been a late event, due to some mechanical cause, and in no way the result of a primary inequality in the two components.

It has been claimed for certain cases of duplicate twins that, in some particulars at least, they are the symmetrical equivalents of each other, but in this case, at all events, the two components are not thus related, but each is perfectly normal in this respect. Thus the appendix of A lies superficially upon the perfect side, and that of B upon the imperfect, the right side of each component, and the same thing is shown by the spleens, which lie diagonally with reference to the double stomach, but each upon the left side of the component to which it plainly belongs, and therefore noriual in its relations. The double liver is an enormous organ, and upon each aspect covers the entire width of the visceral cavity. The face of this mass seen upon the perfect side consists of the right lobe of A, with the gall bladder of that component, and the left lobe of B, without a bladder; the opposite face, that of the imperfect side, which at this level is nearly as wide as the other, is composed of the right lobe of B, with its gall bladder, and the left lobe of A, without one. In these organs, then, except for the continuity of the liver masses belonging to the separate components, the relations are the normal ones which would be found in two separate individuals that stand facing each other, and show no trace of looking-glass symmetry. At this level, also, the other organs are mainly individual, and not shared by the two components.

As a detailed study of all the parts of this specimen is not possible here, we may select two, the details of which seem to present especially good material for discussion: (1) the middle ear of the synotic side, with its included ossicles, and (2) the heart and the larger blood-vessels. The first of these is a part that has not yet become functional, and would thus naturally be expected to show the primary and unmodified symmetry characteristic of a typical Cosmobion; the second is one that shows the


The Morphology of Cosmobia 3T!)

modifications due to functional activity from almost the beginning of development, and thus exhibits the solution of what may be considered a new physiological problem, a study in Experimental Zoology made by Nature herself.

C. The Compound Middle ear of the SynoUc side. This region may be taken as an especially good example of one which, in part through lack of function during embryonic life, has retained its primary symmetry. The parts are at the same time sufficient in number and complexity to express this symmetry in a high degree, and the complete correspondence, even of the smallest details, is highly suggestive in the present argument.

a . ■ b



Fig. 5. Baldwin Synote, Teras III.

(a) Normal os tyrupanicum of the right ear of corupoiient A.

(b) The two reduced ossa tynipanica siirrounding the synotie ossicles. These consist of the right one of B and the left one of A.

All are drawn to the same scale.

Within the single external meatus lay the middle ear chamber, its entrance framed in by a pair of tympanic bones, symmetrical in respect to each other but each somewhat smaller than the normal ones of the perfect side [Fig. 5]. Normally each forms a nearly complete circle, the frame for the outer tympanic membrane, but on the synotie side the two together framed in an oval space, its greatest diameter lying transversely; of this each tympanic bone formed a little less than half the circumference. Undoubtedly there was also here a tympanic membrane, but as the parts had been subjected to maceration in caustic potash before examination this point cannot be determined.

Lying within the oval tympanic frame there appeared the auditory ossicles, a doubled malleus upon a doubled incus, and two distinct stapedes


580


Harris Hawthorne Wilder


diverging right and left from the longer crura of the two latter. [Fig. 6.] The details of the separate ossicles are best learned from Figs. 7-9, which represent the synotic ossicles, in most cases accompanied by a



Fig. 0. Baldwin Synote, Teras III. General view of the double ossicles of the synotic side ; external aspect. This complex was situated within the oval frame formed by the two tympanic bones of this side, Fig. 3 (b).

Right of A. Right of B.

Capitulum [



Capitulum Right of B. /^'^ Right of A.


Proc. ant. 1*1 '/ Collum

Proc. n -^3 M. tens. tymp. lat, "^ ^^



Fig. 7. Baldwin Synote, Teras III.

(a> Double malleus; external aspect. The normal mallmis. taken from the right ear of the perfect face (= A, r) is placed beside It for comparison.

(b) The same ossicles; internal aspect. Upon the double malleus note the slender median processus anterior [Folii], directed downwards and suggesting the former location in the embryo of a double median Meckel's cartilage.

Both bones are drawn to the same scale.


corresponding one from the perfect side, which appears in the case of each bone perfectly normal. In the macerated specimen it was of course impossible to demonstrate the associated soft structures, but by tlie


The Morphology of Cosmobia


381


existence of tciulons^ processes, and other features, each occurring so far as possible in the normal position, and showing always a complete bilateral symmetry, we may feel sure that tlie soft parts were as regular and symmetrical as are the osseous features.

In the synotic malleus [Fig. 7] there is present upon either side the tendon of M. tensor tympani, also a median processus anterior [Folii]. This latter is well developed and suggests the presence in the earlier embryo of a median rudiment of Meckel's cartilage, the equivalent of both halves of the mandible. The double capitulum is small and narrow in proportion to that of the normal one of a single side, but it is per


Right of A.


Right of B.


Left of A.



Crus longum ( (^ Proc.

lentiformis


Left of A Right of B.


Right of A.


b



Crus longum


Proc. lentiformis


Fig. s. Baluwiis^ Synote, Teras III.

(a) Double incus; external aspect. The normal incus, taken from the right ear of the perfect face (=A, r.), is placed beside it for comparison.

(b) The same ossicles; internal aspect. Both bones are drawn to the same scale.


fectly symmetrical and obviously similar to more familiar reduced median parts that occur in Cyclops and Siren monsters. The conspicuous foramen in the median line appears to correspond to nothing in the normal malleus, but is suggestive of the double nature of the piece, and reminds one of the old explanation of a "fusion" of two originally separate components, a theory which has in its favor only such appearances and which is entirely inadequate to explain the phenomena.

Like the malleus, the double incus [Fig. 8] exhibits upon either side the features characteristic of the normal bone, save that the processus lentiformis appears to be present upon one side only. This part is. how


38:3 Harris Hawthorne Wilder

ever, extremely small and easily detachable, and it may easily have been lost in the preparation.

The two stapedes [Fig. 9] differ from the normal in one noticeable feature, and that is the lack of the characteristic foramen. As its presence is due to the formation of a stapedial artery, which is transient and embryonic in j\Ian, but persistent in certain other mammals, its lack here suggests the failure of this artery in the earlier embryo, just the sort of disturbance in the course of the developing carotid that might be expected.

Perhaps nowhere in my study of Cosmobia have I found the problem expressed so simply as here; reduced to its lowest terms, as it were. Instead of a complicated organism, with, perhaps, some degree of secondary deformation, we have here, taking either one of the compound auditory ossicles, a simple bone, median in position, perfectly bilateral, fur


FiG. 9. Baldwin Synote, Teras III.

(a) Normal stapes from the "perfect" side.

(b) Stapes from the synotic complex; lacking the perforation. Both bones are drawn to the same scale.

nished with accessory structures, and without a trace of pathological tissue. If we can explain this, we shall probably hold the key to the solution of all other abnormal Cosmobia, both less and more than a normal being ; perhaps, also, that of the latter as well. Modern biolog}^, in attacking the problem of development, has wisely directed its attention mainly upon the germ cells, and the first cell generations that proceed from them. To the earlier investigations in this field, which, observational in their nature and conducted under the microscope, were applied directly to the germ cells, has been recently added a line of investigation by means of experiment upon early embryos, with careful observations of the results. In Cosmobia we have such results, and we have reason to believe that the initial causes, which are at the present time unknown, lie in the early germ, perhaps even in the unfertilized ovum. It is even permissible to suggest that the cause may lie in a doubling or a deficiency of the primi


The Morphology of Cosmobia 383

tive granules, by whatever name they may be known, and if so it would seem impossible to produce true Cosmobia as the result of experiment, except by the application of means that would modify the developing germ. On the other hand, if this can be done, the cause would not seem to be so fundamental. In the former case the only way in which the matter can be investigated is by studying the results as they are produced by nature, and it may thus be that these unusual beings, which are at the same time orderly and definite in development and structure, may prove a most important factor in the solution of the great biological problems.

D. The hearts and the main Hood-vessels. As characteristic of this entire series there are two hearts, each composed of a half from each component, and placed so that they correspond to the two apparent faces and not to the two components. With each of these hearts is associated a pair of lungs, which bear apparently normal relations to the two apparently normal hearts, but which are also made up of a lung from each individual. There thus lies back of each double sternum a set of thoracic viscera which, except for some slight and evidently secondary deformation, look like those of a single individual, yet, by their relationship to the two bodies lower down, are seen to be composites, each part the property of two individuals.

To analyze this in greater detail, the thoracic viscera facing the perfect side consist of a heart, made up of A's right and B's left half, to which are attached A's right and B's left lung. The trachea and larynx, which are in their usual place with reference to this composite set of organs considered as a unit, are single in appearance but formed of halves contributed by each of the two component bodies. In the same way, upon the imperfect side, there lies a set of viscera that consists of A's left and B's right lung, the two associated with a heart composed of A's left and B's right half. Between these two sets of viscera, each of which is, as it were, backed up against it, lies the oesophagus, a single tube common to both individuals. The vertebral columns with the spinal cords, as will be remembered, are never compound in this type of Cos •'It is hardly necessary to state that the above was written before the results of the recent experiments of Stockaed became known. Since, however, the entire paragraph was intended to awaken speculation, and not to assert a dogmatic position. I have thought it best to allow it to stand as originally written. The discussion of Stockard's work is given below, mainly under the review of the recent literature.


584


Harris Hawthorne Wilder


mobion, but belong to the separate individuals and lie upon the sides of the two sets of thoracic viscera, their ventral aspects facing the median oesophagus. An aorta and a posterior vena cava, each individual, run down the ventral aspect of each.

The two sets of thoracic viscera are separated from one another by a partition of pleura, forming two thoracic cavities, the larger one associated with the perfect, the smaller with the synotic side. The partition is, however, placed obliquely with reference to the compound thorax, so that the large chamber occupies not only the space immediately l^ehind the sternum of the perfect side, but also that framed in by the vertebral column and ribs of component B : the smaller thoracic cavity is simihirly related to the synotic side and component A. The results of the dis


Ant. Vena Cava.


0*^



Fig. 10. attached.


Baldwin Syiiote, Teras III. Heart of perfect side, with lungs


sections of these chambers and their organs may be treated separately, l)earing in mind the composition of each set of viscera.

The heart and lungs of the larger chamber, that of the perfect side, are exposed by laying back the sternum and costal cartilages of this side [Fig. 10]. They appear nearly normal, so much so in fact that it seems pfobable that the slight deformation they display has been due to development in too cramped surroundings, and that they must have been entirely normal at an earlier stage of development. The heart is large, with well defined auricles and ventricles in normal relation. Upon the sides of this organ, and partly adherent to it, lie the lungs, rather incomplete in their development, 1)ut normal in respect to the number of lobes, the right (A's) with three, the left (B's) with two. The large


The Morphology of Cosmobia


385


blood-vessels are somewhat abnormal in their relations, and may be made out by the comparison of Figs. 10-13. Each ventricle possesses an aortic arch, both of which pass to the ventral aspect of the vertebral coliinm of component B, where they meet an arch from the right ventricle


B


Art. Pulm. to Rt. Lung of A. Ant. Vena Cava


To Aorta of B. ■from Lft. Ventr

To Aorta of B.from Rt. Ve



A


.Pulm. Vein from Right Lung of A.

Pulm. Vein from Left Lung of B.


Post. Vena Cava


— /. Right Ventricle A.


Fig. 11. Baldwin Syiiote, Teras III. Heart of perfect side; inner aspect. Tlie dotted line across the auricular region marks the position of the incomplete partition separating the two auricles. The lungs and trachea have been removed.



Left Ventricle; B.


Right Ventricle A.


Fig. 12. Baldwin Syiiote, Teras III. Heart of perfect side; inner aspect, c-ut open to show the interior chambers. The auric-ulo-ventricular openings are marked by arrows.


of the heart of the synotic side and together form B's dorsal aorta. Since, in Cosmobia belonging to this series, the dorsal aorta of each component is usually made up of the union of two arches, one from each of the ventricles that belonged originally to the side in ques^tion, the only departure from the type lies in the fact that the arch that arises


o8G Harris Hawthorne Wilder

from the right ventricle of the heart on the perfect side (A's) crosses over to the aorta of B and does not turn the other way and join the dorsal aorta that belongs to A's vertebral column. The fact that the arch from the right ventricle of the synotic heart comes over to B's aorta is the usual thing in such Cosmobia, since this heart