Journal of Morphology 25 (1914)

From Embryology
Embryology - 29 Nov 2020    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" (textbooks, papers, people, recommendations) 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, interpretations and recommendations may not reflect our current scientific understanding.     (More? Embryology History | Historic Embryology Papers)

Journal of Morphology 25 (1914)

J Morphol. : 1 - 1887 | 2 - 1888-89 | 3 - 1889 | 4 - 1890 | 5 - 1891 | 6 - 1892 | 7 - 1892 | 8 - 1893 | 9 - 1894 | 10 - 1895 | 11 - 1895 | 12 - 1896 | 13 - 1897 | 14 - 1897-98 | 15 - 1898 | 16 - 1899-1900 | 17 - 1901 | 18 - 1903 | 19 - 1908 | 20 - 1909 | 21 - 1910 | 22 - 1911 | 23 - 1912 | 24 - 1913 | 25 - 1914 | 26 - 1915 | 27 - 1916 | 28 - 1916-17 | 29 - 1917 | 30 - 1917-18 | 31 - 1918 | 32 - 1918 | 33 - 1919-20 | 34 - 1920 | 35 - 1921 | 36 - 1921-22 | 40 - 1928 | 47 - 1929 | 51 - 1931 | 52 - 1931 |
Historic Journals: Amer. J Anat. | Am J Pathol. | Anat. Rec. | J Morphol. | J Anat. | J Comp. Neurol. | Johns Hopkins Med. J | Proc. Natl. Acad. Sci. U.S.A | J Physiol. | Ref. Handb. Med. Sci. | J Exp. Zool. | Yale J Biol. Med. | Anat. Anz. | Memoirs of the Wistar Institute of Anatomy and Biology | Quart. Rev. Biol.
Links: Historic Journals | Historic Embryology Papers

Founded by C. O. Whitman

EDITED BY

J. S. KINGSLEY

University of Illinois Urbana, 111.

with the collaboration of

Gary N. Calkins Edwin G. Conklin C. E. McClung

Columbia University Princeton University University of Pennsylvania


W. M. Wheeler

Buaaey Institution, Harvard University

William Patten

Dartmouth College


VOLUME 25

1914


THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY

PHILADELPHIA

THE WAVERLY PRESS

BALTIMORE, U. S. A.

Contents

1914

No. 1. MARCH

H. V. Neal. The morphology of the eye muscle nerves. Nine plates and four text figures 1


No 2. JUNE

Cora J. Beckwith. The genesis of the plasma-structure in the egg of Hydractinia echinata. Sixty-six figures (eight plates) 189

H. J. VanCleave. Studies on cell constancy in the genus Eorhynchus. Forty-six figures 253

James E. Ackert. The innervation of the integument of Chiroptera. Four plates 301

George A. Bates. The pronephric duct in elasmobranchs. Sixty-one figures (five plates) 345


No. 3. SEPTEMBER

Robert W. Hegner. Studies on germ cells. I. The history of the germ cells in insects with special reference to the Keimbahn-determinants in animals. Seventy-four figures 375

P. E. Smith. Some features in the development of the central nervous system of Desmognathus fusca. Fifty -eight figures 511


No. 4. DECEMBER

Fernandijs Payne. Chromosomal variations and the formation of the first spermatocyte chromosomos in the p]uropean earwig, Forfieuhi sp. Sixty-four figures -^^^

Nathan Fasten. Spermatogenesis of the American crayfish, Canibarus virilis and Cambarus immimis (?), with special reference to synapsis and the chromatoid bodies. One text figure and t(>n plates 587

C. E. McClung. a comparative study of the chromosomes in orthopteran spermatogenesis. Ten ])late8 651


THE MORPHOLOGY OF THE EYE MUSCLE NERVES

H. V. NEAL

Tufts College, Massachusetts

NINE PLATES AND FOUR TEXT FIGURES

CONTENTS

Introduction 2

1. The eye-muscle nerves as criteria of head segments 5

2. Technical methods 7

Histogenesis of spinal somatic motor nerves

1. General description 11

2. Protoplasmic connection between muscle and nerve center 16

3. Genetic relations of protoplasmic connections to cells 32

4. Genetic relations of protoplasmic connections to neurofibrils 35

5. The process theory versus the cell-chain theory 41

6. The growth of the neuraxon 43

a. The histogenesis of the Rohon-Beard cells 47

7. The origin of the cells of spinal somatic motor nerves 49

8. The fate of the cells of somatic motor nerve anlagen 51

9. The histogenesis of the neurilemma 53

10. The genesis of the sympathetic anlagen 54

11. Summary 58

Histogenesis of the eye muscle nerves

1. General description 60

2. Histogenesis of the oculomotor : {a to j) 62

3. Histogenesis of the trochlear : {a to j) 78

4. Histogenesis of the abducens : {a to j) 90

5. The bearing of the evidence upon the head problem 97

Relationships of the eye muscle nerves

1. The somitic nature of the eye muscles 98

2. The relations of the oculomotor 100

a. Relations to the profundus branch of the trigeminal 100

b. Relations to the ciliary ganglion 102

c. Relations to four eye muscles 102

d. The morphology of the oculomotor: Conclusions 103

3. The relations of the trochlear 106

a. Relations to ganglia 106

Relations to the superficial branch of the trigeminal 107

b. Relations to a somite innervated also by another nerve 108

c. The chiasma of the trochlear 109

d. The morphology of the trochlear: Conclusions 114

1

JOURNAL OF MORPHOLOGY, VOL. 25, NO. 1 .MARCH, 1914


2 H. V. NEAL

4. The relations of the abducens 117

a. The relations of the post-rectus muscle to post-otic muscles. . . 118

b. The bimeric distribution of the abducens JL22

c. The abducens of the cyclostomes 124

d. The morphology of the abducens: Conclusions 126

5. The metameric relations of the eye muscle nerves 126

6. The neuromeres as criteria of segmentation 132

7. The homology of head and trunk segments 138

8. Conclusions regarding pre-otic metamerism 140

Literature cited 144

INTRODUCTION

An advocate of the orthodox view of the vertebrate head as a structure derived from segmented pre-vertebrate ancestors is hkely to meet with an increduhty or an indifference quite unknown to morphologists of the last generation. The reasons for this are several. In the first place special creation has rested its case and evolution no longer feels the urgent necessity of strengthening the argument that the vertebrate head or any other organic structure has had a past history. In view of the general acceptance of the verdict in favor of evolution it appears to be a matter of minor importance just what this history has been, whether annelid, tunicate, arachnid or what-not.

Furthermore, the conflict with accumulating evidence and the disagreement in opinion among vertebrate morphologists; the divergence in results based upon the study of the ontogenesis of different vertebrates; the increasing conviction of the plasticity and mutability of ontogenesis; doubt regarding the actual specificity of the germ layers; all have tended to undermine the faith of our scientific forebears in the validity .of the fundamental law of biogenesis except in its most general features. Faith in phylogenetic inductions based largely on embryological data has been greatly weakened. A distinguished embryologist (McMurrich '12) recently expressed the opinion that more reliable results are to be obtained in the majority of cases from comparative anatomy." We must concede that the results of embryological investigation have disappointed those who anticipated general agreement regarding the phylogenesis of the vertebrate head.


MORPHOLOGY OF EYE MUSCLE NERVES 6

Civil war in the camp of the morphologists between anatomists and embryologists has led some to exclaim, "a, plague on both your houses!" Morphology is the loser when embryologists and anatomists attack each other instead of morphological problems. The pot has called the kettle black and the luster of neither has increased. There seems to be no particular advantage in discussing the question whether comparative anatomy or embryology has been able to make the more valuable contribution to the solution of morphological problems. It appears not unreasonable to think that, notwithstanding frequent divergence of theoretical conclusions, the key to the solution of morphological problems lies in the hand of neither alone. The two sciences must develop together. The results by one method need to be 'controlled' by the other. *

The elaboration of other than the annelid theory of vertebrate ancestry upon which the orthodox conception has been largely based, has tended to throw doubt upon the view that the segmentation of the vertebrate head is an ancestral segmentation. A morphologist of the last generation might have thought it necessary to apologize for advocating the view that the vertebrate head is an unsegmented portion of an unsegmented ancestor. No apology would be required today. As a matter of fact, after a century of vertebrate morphology, we actually do not know whether vertebrates came from a segmented or an unsegmented ancestor; whether or not the present segmentation in the head region is a true metamerism; whether the pre-otic portion of the head is a coenogenetic addition to a primarily segmented body, or a palingenetic remnant of an unsegmented ancestor which has secondarily acquired a segmented trunk. Even among those who agree that the present segmentation is ancestral there is little accord as to the nature or number of these somites.

A grouping of morphological opinion, however, with reference to the extremes of expression shows a strong tendency on the part of morphologists to regard the head as metameric. The curve of variation in opinion is decidedly skew in the direction of morphological orthodoxy. Among the many papers dealing


4 H. V. NEAL

directly or indirectly with the problem of the morphology of the vertebrate head since the opening of the twentieth century, that of McMurrich ('12) is the only one adverse to the view that the head was primitively metameric. Vertebrate morphology, however, is deeply indebted to the skeptical but stimulating spirit which has constantly compelled orthodox opinion to reestablish its fundamental tenets.

To a considerable degree the skepticism regarding the comparability of head and trunk metameres may be traced back to the demonstration of the differences between cranial and spinal nerves, the serial homology of which seemed unquestionable to the earlier morphologists. The discoveries in the field of nerve components and of their various central and peripheral relationships seemed to many greatly to increase the difficulty of comparing the nerves of head and trunk. In consequence more stress came to be laid upon longitudinal columns in the central nervous system than upon a hypothetical primitive metamerism or neuromerism.

Nevertheless, the theory of the primitive metamerism of the head is "noch nicht aus dem Welt geschafft." The last decade has not seen a diminution in the numbers of those undertaking a reinvestigation of this perennial problem, notwithstanding the attractive fields in other departments of biology which have opened up during the period. With whatever incredulity or indifference schemes of head segmentation may meet as they appear, there are indications that students of head morphology are reaching an agreement with regard to the fundamentals of the metamerism of the head. The reinvestigation of many ontogenetic and histogenetic problems, concerning which earlier investigations had led to divergent results, have led to a remarkable agreement in conclusions and have tended to remove the prejudice against ontogenetic foundations for phylogenetic conclusions. It is now seen that much of the divergence in results of earlier embryological research was the result of inadequate methods or materials — of generalizations based upon insufficient data. There is good reason for thinking that further investigation of nerve histogenesis will show a similarity between


MORPHOLOGY OF EYE MUSCLE NERVES 5

cranial and spinal nerves which had been rendered doubtful by the incomplete researches of earlier students of nerve development. The recent elaborate studies of the histogenesis of the trochlear and oculomotor nerves by Dohrn ('07) and Gast ('09) have tended greatly to strengthen the conviction of morphologists that these nerves are comparable with spinal nerves. Belogolowy's ('10) careful monograph on the cranial nerves of the chick leads to the same conclusion. It is the purpose of the present study to make a contribution to the same end.

The writer takes pleasure in expressing his appreciation of the aid given him during the prosecution of this research by Professor John Sterling Kingsley, Director of the Harpswell Laboratory, during the summers of 1906-1913 arid for the unlimited supply of embryos of Squalus acanthias which residence at this laboratory has rendered available. He also acknowledges gratefully a grant of $50 by the trustees of the Elizabeth Thompson Fund, used in the attempt to get embryos of stages between 50 and 100 mm., not obtainable off the Maine coast during the summer months.

1. The importance of the eye-muscle nerves as criteria of the metamerism of the head

In a recent paper dealing with the head problem Ziegler ('08, p. 674) writes that, since he holds that the eye muscles arp relatively young muscles which have not arisen directly from segmental muscles, he is of the opinion that one cannot make use of their innervation for phylogenetic conclusions. Ziegler does not give his reasons for regarding the eye muscles as relatively young muscles. Moreover, he does not prove that they have not arisen directly from segmental muscles, and in drawing conclusions concerning the phylogenesis of the head he makes use of the relations of the oculomotor as an essential connecting link.

Cole ('98, p. 237) also refers to the eye-muscle nerves as the the least primitive of the cranial nerves" without giving any proof of the assumption.


6 H. V. NEAL

On the other hand, Gast ('09, p. 423) asserts that the eyemuscle nerves appear very conservative in their relations to the neuromeres, notwithstanding the considerable dislocation of their muscles. He therefore concludes that the eye-muscle nerves supply most important material for the study of the primitive metamerism. However profound the changes in their terminalorgans, the central relations remain unchanged, so that they serve as important criteria of the nature and number of pre-otic segments. "Die drei Augenmuskeln Nerven haben eine hohe phylogenetische Bedeutung," The writer would emphasize their importance even more than Gast has done, for it seems to him that the demonstration of the serial homology of head and trunk metameres depends largely upon the proof of the resemblance of eye-muscle 'and spinal somatic motor nerves.

The reason for such an opinion is plain when we consider that other evidences of cephalic segmentation, such as cranial ganglionic nerves and visceral arches, appear limited to the head region and to have no exact homologues in the trunk. The presumption that head and trunk were primarily undifferentiated must come not from such evidence but from structures which may be more readily compared. The eye-muscle nerves in connecting somites and neuromeres, are related to structures which extend through head and trunk. Moreover, in their central relations the eye-muscle nerves resemble spinal somatic motor nerves. Therefore, failure to convince morphologists of their meristic homology with spinal nerves would tend to undermine the foundation of the traditional conception of the head.

Consequently, the divergence of opinion regarding their true nature, as evinced by their ontogenesis in Selachians, has tended to obscure the more important issue of the history of the head. The repeated attempts to compare the trochlear and oculomotor with such nerves as the trigeminal and the facial, notwithstanding their obvious differences in the adult animal, suggests the necessity of renewed investigation of their histogenesis. Recent exhaustive papers by Dohrn ('07) and Gast ('09) have appeared in response to this need. In many matters of fundamental importance this paper confirms their results. The theoretical


MORPHOLOGY OF EYE MUSCLE NERVES 7

conclusions are not so fully in accord, but all agree that the ancestral vertebrate was a metameric acraniote, and there is no suggestion that the meristic elements of the head have had an exogenous source.

2. Technical methods

Considerable investigation upon the histogenesis of nervous structures has been made upon material prepared by methods unsuited to the special requirements of neurological investigation. To this may be attributed many of the fallacious inferences that have delayed our knowledge of the true method of nervous differentiation. Failure to use suitable neurological methods is no longer excusable, however, since the advances made by Cajal, Paton, Held and others. Tt is necessary for one who attempts today to base morphological conclusions upon the study of nerve histogenesis, to make sure by proper technique that he is dealing with real nervous structures and not with pseudo-nerves. Not every cellular strand or spindle-shaped cell, however closely associated with the central nervous system or its derivatives, is necessarily a nervous anlage. Some criterion by which a structure in the embryo may be identified as nervous is needed. Hence the importance of stains which will differentiate the neurofibrils from the time of their first appearance.

Equally important in addition, is the use of a method which will demonstrate cell boundaries and relations, especially in those early stages when nervous connections between nerve center and end-organ are estabhshed. Since no single method is known which will effect both of these results with equal precision, different methods for comparison and control of results appear necessary. Trust in a single neurological method, however highly perfected, may mislead as much as a less specific stain has done. The importance of methods of preservation and staining is evinced by the great divergence in the results, both technical and theoretical, of Paton ('07) and Held ('09), each of whom worked with a highly perfected technique.

Therefore, while it may be granted that the older embryological methods are no longer adequate for the purposes of neu


8 H. V. NEAL

rological investigation, the results by no single neurological stain, however specific, may be wholly trusted. The methods suited to demonstrate the neurofibrils in vertebrate embryos give very unsatisfactory pictures of cell boundaries and relationships. But, as a basis for theoretical conclusions, the cellular relationships are quite as important as phases of differentiation of neurofibrils. Moreover, it must not be forgotten in the enthusiasm created by the newer discoveries that, provided embryological material be abundant, the identification of an embryonic structure as nervous may be determined by tracing its fate through successive stages of development. The demonstration of neurofibrillae therefore is not absolutely indispensable in the determination of its nervous character.

In the course of the present study different methods have been used. Among those which have given the best results are Cajal's nitrate of silver, Paton's modification of Bielchowsky's method, Held's molybdic acid — hematoxylin stain and vom Rath's picro-acetic-osmic-platinic chloride — pyrogallic acid treatment. All the drawings in plates 1 to 8 of this paper were made from preparations by the latter method. They have been confirmed by observations upon material in which the neurofibrillae were specifically stained.

The Vom Rath ('95) method is as follows:

1. Fix in the dark for one to three days in the following mixture (use plenty and change each day) :

Saturated and filtered solution picric acid 200 cc.

Glacial acetic acid 2 cc.

Platinic chloride (dissolve in 10 cc. water) 1 gram

Osmic acid 2 per cent 25 cc.

OxN-ing to the great brittleness of embryos fixed in this fluid all changes of liquid should be made A\ith pipette in the same dish, avoiding as far as possible any movement of the embrj-os.

2. Stain in 0.5 per cent pyrogallic acid in the dark for twenty-four to forty-eight hours ^^dth several changes.

3. Grades of alcohol from 35 per cent slowly by the siphon capillary drop method to avoid shrinkage. Xylol, to which paraffin is added as it dissolves.

4. Imbed in rather hard paraffin of best quality.

5. Thin sections, not over 8 micra, preferably 4 to 6 micra.


MORPHOLOGY OF EYE MUSCLE NERVES 9

Vom Rath's method is not specific for the neurofibrils, which are nevertheless deeply stained. Cell boundaries are shown with special distinctness and shrinkage is slight. The process is advantageous in demonstrating cell relations in the stages, when nervous connections of tube and somite are effected.

Flemming's stronger formula gives excellent fixation of selachian embryos but does not allow the use of pyrogallic acid for subsequent staining. Fixation seems quite as faithful as in Vom Rath preparations, but cell boundaries are not so distinct as in the latter. Iron hematoxylin gives the best stain, subsequent to the use of Flemming's fluid, but it is necessary to paint the sections with 0.5 per cent celloidin in order to prevent their loss in staining on the slide.

For the specific purpose of demonstrating the neurofibrils Cajal's method has given uniformly satisfactory results, which appear somewhat less refined than those obtained by the Bielchowsky-Paton process. The Cajal method is as follows:

1. Fix in absolute alcohol and 1 per cent ammonia for forty-eight hours.

2. Wash for one-half to three minutes in distilled water.

3. Pyridine for twenty-four hours.

4. Distilled water — many changes — for twenty-four hours.

5. Two per cent aqueous solution of silver nitrate for three days at 35°C. in the dark.

6. Rinse in distilled water.

7. Four per cent pyrogallic acid in 5 per cent formalin for one to two days.

8. Paraffin sections.

The Simarro-Cajal silver reduction method, following fixation in 70 per cent pyridin, which has given such splendid results when apphed to mammal and other amniote embryos has proved a complete failure in the case of Squalus embryos.

Excellent results in the differentiation of the neurofibrils have followed the use of the molybdic-acid hematoxyhn process as developed by Held ('09). Tissues may be fixed by various methods including Zenker's fluid and Rabl's picro-sublimate. The stain is effected by a solution of molybdic acid in a 1 per cent solution of hematoxylin in 70 per cent alcohol. The stain


10 H. V, NEAL

is better after months or years. Immediately before use, several drops of this tincture — depending on the strength wanted — are dissolved in distilled water and the sections are stained warm on the slide at 50° C, or for a longer time cold. The sections may be stained directly or they may be mordanted in iron alum. The neurofibrils are differentiated by the Bielchowsky-Paton process but, like the Cajal method, this does not demonstrate the fibrils within the neuroblast cell in the earliest stages of histogenesis. By this method the neurofibrils are stained a dark brown or black, while other tissues are light brown or yellowish brown. In the process only tested distilled water and absolutely clean glass ware and glass or bone spatulas^-no metal — should be used:

1. Fix and keep embryos in 10 per cent formalin neutralized or made slightly alkaline "v\dth magnesium carbonate.

2. Wash in running tap-water for twelve hours.

3. Wash in three or four changes of distilled water for a half-hour.

4. Place in pure 1 per cent nitrate of silver for six days in the dark at a temperature of about 70°C. Tissues must become reddish-brown in color. If they become yellowish-brown, throw away.

5. Place in a solution of silver nitrate freshly prepared as follows : 20 cc. of 1 per cent silver nitrate. Add 2 drops of 40 per cent caustic soda to form a gray precipitate. Then add 20 to 30 drops of strong ammonia, enough to dissolve the precipitate while stirring. . Allow to remain at least forty-five minutes.

6. Wash quickly in two baths of distilled water and quickly place in distilled water, to every 20 cc. of which five drops of glacial acetic acid has been added. Leave in this five to fifteen minutes or until the reddish-brown becomes yellowish bro"«ai.

7. Wash quickly and place over night in 1 per cent hydroquinone containing 5 per cent neutral formol.

• 8. Wash quickly in distilled water, run up through alcohols rapidly and imbed in paraffin through benzole or chloroform.

9. Cut sections and fix on slide. Dry well, then paint slides with 0.5 per cent celloidin. This is followed by absolute alcohol-xylol, and absolute alcohol-xylol again. Then absolute alcohol to 95 per cent alcohol down to water (distilled) .

10. Then one to two hours in 0.1 per cent solution of gold chloride neutralized with lithium carbonate. Grubler and Hollborn's gold chloride should be used (flavum, not fuscum),

11. Rinse in distilled water and fix in 5 per cent hyposulphite of soda for fifteen minutes. Wash in running tap water for two hours. Then alcohols up to absolute. Then absolute and eosin for a minute. Absolute alcohol, xylol, and mounlT in neutral balsam.


MORPHOLOGY OF EYE MUSCLE NERVES 11

HISTOGENESIS OF SPINAL SOMATIC MOTOR NERVES

1. General description

Somatic motor nerves make their first appearance in embryos of Squalus acanthias of about 4.5 mm. and approximately 30 somites. With their advent begins the rhythmical bending of the embryo. The first pair of somatic motor nerves to appear are those of VanWijhe's seventh somites, the first somites differentiated, and those which form the first permanent myotomes.

Previous to the appearance of this first pair of somatic motor nerve anlagen, no protoplasmic connections or plasmodesms between tube and somite are discernible. In the intercellular space between neural tube and somite there is present a plasmoid substance or a liquid with a minimal amount of coagulable material, which, when treated by the usual reagents, assumes a vacuolated appearance. By the use of such fixing fluids as neutral 10 per cent formalin, followed by intense stains such as acid fuchsin it is possible to demonstrate this coagulable substance with considerable clearness. By the- use of other fixing fluids — the plasmolyzing action of which is less intense than that of formalin — and the use of less powerful stains, the plasmoid material is almost invisible and has usually been ignored in embryological studies. Whatever may be the chemical nature of the granules of this plasmoid substance, it does not have the staining properties of protoplasm and its presence may not therefore be taken as evidence of a primary protoplasmic connection between tube and somite, although it is indisputably the medium through which the growing nerve anlagen find their way to the adjacent myotomes.

While the possibility that this plasmoid material may contribute to the growth of the nerve paths may not be denied, yet even so the contribution must be very slight or quite negligible, since the amount of coagulable material in it is very small. To represent it at all, as in figure 1, greatly exaggerates its distinctness and amount as seen in sections. There is nothing in its vacuolated structure that would suggest 'paths' or 'plasmodesms' suggested by the Hensen hypothesis. Moreover, the


12 H. V. NEAL

proximity of tube and somite is so close that the absence of predetermined paths for the growing nerve is not surprising. Supporters of the Hensen hypothesis have greatly exaggerated the mechanical difficulty of the acquisition of a connection between tube and myotome in the absence of primary paths. But even the most ardent supporter of the Hensen hypothesis will admit the capacity of cells to throw out amoeboid processes into the surrounding medium, and for distances quite as great as that which separates somite and neural tube.

As a matter of fact, protoplasmic connection between tube and somite to form the anlagen of somatic motor nerves in Squalus is effected in precisely that way, as was described by Dohrn ('88) many years ago. The process is one of amoeboid outflow, as stated by that accurate observer, arid as shown in figures 4 to 7 of this paper. Beginning with the seventh somite of the selachian embryo protoplasmic connection with each of the successive somites of the body is effected in the same way, that is, by amoeboid processes of medullary cells of the ventrolateral wall of the neural tube opposite the middle of the somites. Since the successive somatic motor nerve anlagen are successively formed, beginning with the most anterior, all stages in the establishment of these connections of nerve and muscle may be seen in serial cross-sections of Squalus embryos of all stages beginning with embryos of 30 somites. The process goes on in the caudal region until quite late stages are reached.

The protoplasmic connections which form the anlagen of somatic motor nerves are not formed by the differentiation of plasmodesms which have existed from the beginning as the result of incomplete cell division, but are formed secondarily by the free outgrowth of amoeboid processes of medullary neuroblasts. The process is one of protoplasmic movement, analogous to that seen in the remarkable culture preparations of Harrison and to the outgrowth of the neuraxons of the Rohon-Beard cells illustrated in plates 3 and 4 of this paper. Stages in the establishment of these connections between tube and myotome and in the growth of the nerve anlagen are shown in plates 1 and 2 of this paper.


MORPHOLOGY OF EYE MUSCLE NERVES 13

In correlation with this movement of medullary protoplasm there occurs a similar movement of the protoplasm of the cells of the sclerotomic portion of the somite, which is inaugurated by an extension of amoeboid processes toward the neural tube (figs. 2 and 3) followed in later stages by a movement of entire cell bodies. In many cases, but not invariably, the movement of the sclerotome cells precedes that of the medullary cells. The protoplasmic material derived from the two sources in exceptional cases unites to form a protoplasmic strand or plasmodesm in which it is difficult to distinguish ectodermal and mesodermal constituents, as is shown in figures 8 and 9. This difficulty obtains, however, only in the earliest stages, if it occurs at all. Even in such cases the sclerotome cells soon separate to form a loose mesenchyma of lightly staining and much vacuolated cells, while the medullary portion persists as a compact strand of deeply staining protoplasm which qtiickly takes on a fibrillar appearance and also becomes cellular through a process of cellular migration from the neural tube (figs. 11 and 12).

Evidence that the cells of the somatic motor nerve anlagen of Squalus are chiefly, if not exclusively, of medullary origin has been given in an earlier paper (Neal '03) and need not be reviewed in detail at this time. Their medullary derivation seems sufficiently attested by evidence of continued migration in successive stages; by their close apposition to the fibrillar bundle or their inclusion within the fibrillar portion of the nerve anlage (fig. 12); by the change in the contour of the neural tube in successive stages as seen in cross-sections; and by the relations of the outer limiting membrane of the tube to the nerve anlage. If mesenchymatous cells are added to the nerve anlage at all, it appears to be in later stages of histogenesis only.

While spinal somatic motor nerve anlagen are primarily protoplasmic and non-fibrillar this stage is quickly passed and, as the anlage extends ventrad along the median surface of the myotome between the myotome and sclerotome of the somite, it assumes a fibrillar appearance. That these fibrils are neurofibrillar in the Apathy sense is proven by the fact that they stain intensely in either Cajal or Paton preparations. Evidence


14 H. V. NEAL

has been given in an earlier paper that the cells of the nerve anlagen have no genetic relations to the fibers. The fibers may be traced to deeply-staining bipolar neuroblasts in the ventrolateral wall of the neural tube.

WTien sections in metameres immediately posterior to those in which protoplasmic connections between tube and somite have just been established by the outflow of medullary protoplasm are studied; that is, in metameres in which it may be assumed that protoplasmic connections are about to be established by a similar outflow, before the medullary cells have thrown out amoeboid processes, deeply staining cells may be found which may be inferred to be the neuroblasts of that metamere. Their more deeply staining properties may be ascribed to the presence of a deeply stained neuroreticulum Vithin them. Analogy with the histogenesis of the Rohon-Beard cells favors this inference, since within the latter, from the time of their differentiation from the surrounding epithelial cells of the tube, a reticulum with a strong affinity for stains makes its appearance and seems to be genetically related to the neurofibrillae which later make their appearance within the long processes of these remarkable cells (plates 3 and 4). In Cajal preparations similar cells in similar relations show a deeply-stained reticulum. There is therefore reason for thinking that the neurofibrillae of somatic motor nerve anlagen have their origin in medullary neuroblasts and are not exogenous in their derivation as suggested by Paton ('07). In all stages the fibers of the nerve anlagen may be traced to deeplystained bipolar cells in the somatic motor column of the neural tube. While it is difficult to obtain positive proof of the fact, the evidence, so far as it goes, favors the opinion th^t the cells which form the first protoplasmic connections between tube and somite are true neuroblasts and not indifferent cells which form protoplasmic paths for the nerves, as suggested by the cell-chain hypothesis or by the similar hypothesis of the Hertwigs. In short, connections between tube and somite are effected primarily by neuroblasts and not by indifferent cells. These primary connections appear protoplasmic and not fibrillar, only because


MORPHOLOGY OF EYE MUSCLE NERVES 15

they are formed by the amoeboid terminations of the neuraxon processes. The protoplasmic and fibrillar relations are precisely analogous with those seen in a Rohon-Beard cell such as is represented in figure 21. The advancing end of the neuroblast cell is protoplasmic and amoeboid, while the more proximal portion of the cell has a neurofibrillar structure. There appears no good ground for doubting the analogy between the phenomena of growth of the Rohon-Beard cells and those of the medullary neuroblasts which form the fibrillar constituents of somatic motor nerves. The evidence of the exogenous derivation of the neurofibrils advanced by Paton seems unquestionably based on incompletely stained preparations. The assumption of the participation of indifferent cells in the formation of 'paths' for the growing nerve fibers has an equally insecure foundation.

After the fiber bundle of the dorsal (somatic sensory) nerve unites with that of the somatic motor nerve a cluster of cells makes its appearance median to the mixed bundle at about the level of the dorsal aorta, forming the anlage of the sympathetic ganglion. The derivation of these cells from the neural tube has sometimes been assumed. The majority of opinion has incHned to the view that they are derived from the dorsal ganglia. Positive proof of either inference has never been given on the basis of direct observation. Experimental data appear not absolutely trustworthy. The present paper makes no contribution toward the solution of this important question. Analogy with the formation of the sympathetic in the head appears to the writer to favor the view that the most of the sympathetic cells have their source in the dorsal ganglia. The problem, however, needs renewed investigation.

Turning now to a more detailed discussion of some of the mooted points in the problem of nerve histogenesis, we may take up first the important question of the primary connection between nerve and muscle.


16 H. V. NEAL

2. Are muscle and nerve connected with each other ah initio? Are

protoplasmic connections between myotome and

tube primary or secondary?

With the statement of this question we plunge at once into the most controverted point in neurogenesis — the problem of how nerves become connected with their terminal-organs. The earlier students of head morphology avoided the details of nerve histogenesis. Yet it is clear that the morphological resemblance of nerves is determined quite as much by their cellular and central connections as by the character of their terminal-organs. In the adult elasmobranch and in the higher vertebrates these relationships are frequently so greatly modified and complicated that to unravel their intricacies requires the facts which the ontogenesis of the lower vertebrates alone can give. Avoidance of the details of histogenesis and reliance upon embryological procedure unsuited to neurological investigation has led to the inclusion of cell clusters or strands among cranial nerves without proof of their nervous character and to erroneous inferences regarding the phylogenesis of the head. Minot ('96) suggested this weakness of earlier researches and the path of later investigation when he wrote that 'Hhe attempt has been made to solve the most difficult questions of the morphology of cranial nerves without answering the inconvenient question of nerve fibers and their sheaths." To-day the student of phylogenesis is no longer able to shut his eyes to the necessity of a thorough investigation of the genesis of nerves before drawing conclusions regarding the past history of the vertebrate head.

In the introduction to his masterly discussion of nerve histogenesis, Held ('09) states that notwithstanding the numerous observations of the past fifty years and the multiplicity of hypotheses, the question as to how nerves are formed in the embryos of vertebrates is one of the most burning questions of embryology :

Many new methods have been applied to its solution and a surfeit of new views have appeared, following one another in a short period of time, but as yet no agreement has been reached regarding the true principle of nerve development. This is due to a great extent to the


MORPHOLOGY OF EYE MUSCLE NERVES 17

lack of a fundamental histological method which effects a satisfactory stain for embryonic as well as for adult nervous tissue. But the disagreement is also partly due to the fact that it has been the concept rather than the reality that has been observed. Finally the continued conflict of opinion has been due to the fact that no comprehensive observations have been brought together which make it. possible to demonstrate in one and the same animal all of the essential stages in the development of a definite nerve path, and to extend in a comparative research over the chief groups of vertebrates, in order to find, in the diversity of phenomena seen in different embryos, the same fundamental principle of nerve development.

Held undertook this herculean task and has made a most valuable contribution to our knowledge of the histogenesis of the neurofibrillae, which, since the emphasis laid upon them by Apathy, have come to be regarded as the essential elements of the nerve fiber. In large measure Held's monograph is an attempt to reconcile the divergent hypotheses of nerve histogenesis.

Three chief theories of nerve histogenesis have been advanced. Chronologically arranged these are:

1. The theory advanced by von Baer in 1829 and later elaborated by Hensen ('64, '76, '08) of the primary connection of the nervous center with its later innervated terminal-organ. The fundamental idea underlying this von Baer-Hensen theory is that of the necessity of a primitive connection between the nerve and its terminal organ. In other words that it is impossible to conceive of a nerve fiber finding its area of distribution or its terminal-organ as the result of the free outgrowth of the nerve cell. Hensen brought forward the idea of primary nerve paths, which was founded on the actual observation that in the embryo adjacent organs are connected with each other by means of protoplasmic bridges. These were interpreted as remnants of incomplete cell divisions and they were supposed to form the primary nerve paths which later become transformed into the definitive nerves or are used for nerve formation. How the transformation occurs was unknown to Hensen, but he was convinced that they did not develop from cell outgrowths. The theory appeared to afford a satisfactory explanation of how nerves find their way to their terminal-organ. Chiefly on the ground of general principles, Gegenbaur ('98), Fiirbringer ('97), Kerr

JOURNAL OF MORPHOLOGY, VOL. 25, XO. 1


18 H. V. NEAL

('04) and Braus ('05) have added to the Hensen hypothesis the assumption of an original and unchangeable connection of motor nerve and muscle.

Held ('09) suggests that the assumption of primary — ab initio — connectJon between muscle and nerve is not an essential part of the Hensen hypothesis. With this view, however, Hensen ('03) does not seem to agree. Hensen states his hypothesis in his Vorrede as follows: Nervous connections are not established through the free outgrowth of the nerves in the embryo, but central and peripheral organs remain in connection with each other from the time of their formation (Sonderung) until the complete differentiation of the nervous trunks."

2. The second theory is that advanced by Schwann ('39) and later revived by Balfour ('77) as the cell-chain hypothesis, according to which the nerve fiber arises from the union of a series or chain of primary cells, which later accompany the neuraxon as the so-called Schwann's or neurilemma cells. According to this hypothesis nerve fibers are formed by the fusion of primary cells, whose nuclei become the nuclei of the differentiated fiber, so that the nerve fibers are in consequence multicellular in origin. This theory has been supported by Marshall ('78), VanWijhe ('82, '86, '89), Beard ('85, '88, '92), Miss Piatt ('94, '96), Sedgwick (94, with some modification), Hoffmann ('96), Kupffer ('90, '91, '94), Rafaelle ('00), Bethe ('00-'07, with modifications), Brachet ('05, '07), Cohn ('05, '06, '07), Oscar Schultze ('04-'07).

According to the cell-chain hypothesis, the peripheral nerve fiber is the common product of a chain of cells, on the end of which, in the case of a motor nerve, or in an intermediate position, in the case of a sensory nerve — the larger cell forming the ganglion cell is situated. The cell chain not only forms the ganglion cell but also the cells of the neurilemma sheath, which are not merely sheath cells but also the cells that produce the fiber — as nerve-fiber cells or 'neurocytes' (Kupffer), 'nerve-cells' (Bethe and Apathy) or 'neuroblasts' (O. Schultze). The nerve fiber is thus an elongated mosaic consisting of a series of cells, each cell of which has added to what another cell before it has formed, and has continued into the following cell. Bethe ('00


MORPHOLOGY OF EYE MUSCLE NERVES 19

'07) has modified this view by assuming that mitotic cells form the first anlage of the nerve and the development of the nerve proceeds by cell multiplication. Not every fiber, however, arises from a series of cells, but a series of cells produces a large number of fibers.

3. The third theory of nerve histogenesis, first formulated by Kupffer and Bidder ('57), but finally developed as the neuroblast theory by His ('79), holds that the nerve fiber develops as the process of an embryonic ganglion cell. Even before this theory had been advanced it had been noted by Remak ('55) that nuclei are absent in embryonic nerves. His ('79, '88, '90, '93, '04) developed this process theory of Kupffer and Bidder as the 'neuroblast theory' of nerve histogenesis, according to which nerves are formed as processes of special cells or 'neuroblasts.' The first outgrow^th of these cells forms the axis cylinder and the cell becomes a ganglion cell. A single process is formed by the neuroblasts of the neural tube, while two are produced by the neuroblasts of the peripheral ganglia. Other branched processes formed later become the dendrites of the ganglion cells. Thus every single nerve fiber represents a single and definite cell of origin which constitutes its genetic, nutritive and functional center. These nervous units do not unite with each other in protoplasmic continuity, but a transfer of impulses occurs without continuity. This view has been supported by von Koelliker ('86, '91, '92, '00, '05), Sagemehl ('82), Ram6n y Cajal ('90-'07), von Lenhossek ('90-'03), Retzius ('93), Neal ('98, '03, '12), Gurwitsch ('00), Harrison ('01-'12), Bardeen ('03), Lewis ('06, '07), Dohrn ('07), Belogolowy ('10) and Burrows ('11).

Waldeyer ('91) has still further developed the neuroblast theory as the neuron theory, according to which the nervous system, both central and peripheral, consists of independent functional units — cells or neurons — which never become fused or united but transmit impulses by transfer or induction from one neuron to another.

Later investigators have advocated various modifications of these three theories, but not hypotheses that may be sharply distinguished from the three theories mentioned above. They


20 H. V. NEAL

may be regarded as variations of a theme rather than as original themes. The brothers Hertwig, for example, on the basis of evidence derived from the study of the development of nerves in medusae, suggested ('78) the hypothesis that the terminal organ, primarily separated from the nerve center, becomes secondarily connected with it by means of a simple protoplasmic path which is later differentiated into the actual nerve. The connection between nerve center and terminal-organ is formed secondarily, is at first protoplasmic and not nervous, and is the product of indifferent cells and not those which form the nerve fibers. Such protoplasmic connections become the paths along which the nerves develop and are established when nerve center and terminal-organ are in close proximity. This hypothesis, in common with that of Hensen, lays stress upon the protoplasmic paths within which the definite nerve fibers are differentiated, but differs from the Hensen hypothesis in assuming the secondary connection between center and terminal-organ. The hypothesis, however, as Braus ('05, p. 472) suggests, may not be sharply distinguished from the cell-chain hypothesis which has been rendered untenable or unnecessary by Braus' experiments upon amphibian larvae. Braus admits, however, the possibility of the formation of protoplasmic bridges before the appearance of the neurilemma cells. But, since it has not yet been proved that the protoplasmic bridges are the products of indifferent cells — in other words, that they are not the product of the same neuroblasts which produce the neurofibrillae — the hypothesis remains little more than a speculation, and the phenomena upon which it seems to rely are equally well interpreted by the process theory.

Other investigators, such as Apathy ('97), Joris ('041, Pighini ('04, '05), Besta ('04), London and Pesker ('06), Brock and Gieriich ('06, '07), and Held ('09) have presented some divergent opinions, but no wholly distinct point of view. The peculiar views of Apathy possibly deserve special recognition since they have aroused so great an interest. According to Apathy ('97) the embryonic ganglion cells acquire their fibrillar structure, not by a process of cell differentiation, but secondarily


MORPHOLOGY OF EYE MUSCLE NERVES 21

from special cells, which, as actual 'nerve cells' — that is, as neurofibril-forming cells — are intercalated in the path of the nerve fibers and which push their differentiated product into the interior of the muscle cell or gland cell as well as into the ganglion cell. This doctrine, which conceives of the ganghon cell and the nerve cell as entirely distinct kinds of cells has been also supported by Bethe ('00-'07). According to Bethe the neurilemma cells are the 'nerve cells' in the Apathy sense of the word. That is to say, they are the cells which have formed the neurofibrils and persist as sheath cells.

Held attempts ('09) to reconcile the divergent views of nerve histogenesis, finding something of truth in them all. In_ common with Hensen he emphasizes the importance of protoplasmic paths which furnish material for the growing nerve; agrees with the Hertwig brothers that these may be secondarily formed by indifferent cells, and supports the Kupffer-Bidder theory in holding that the neurofibrils are the product of the neuroblast through a process of centrifugal growth. While his observations seem to support the process theory in most details, he denies the doctrine of free outgrowth of nerve fibers.

None of these hypotheses of nerve histogenesis has been abandoned entirely by partisan supporters. With the introduction of special nerve methods the confusion has seemed to increase rather than to diminish, so that there is not a single result of the most recent investigation which is not directly opposed by another. The chief problem raised by Apathy as to the genesis of the neurofibrils, has not yet been solved; and the more important question of how muscle and nerve become connected seems as far from solution as ever. Relief from conflict of opinion has not fgllowed recourse to experimental morphological methods as is shown by the divergence in the views regarding neurogenesis held by Harrison ('03, '04) and Braus ('05). While the former supports the process theory, the latter thinks that his results confirm the Hensen or Hertwig hypothesis of predetermined paths for the growing nerves.

• Upon the answer to the question, whether or not nerve and muscle are primarily connected, depends, in large measure, not


22 H. V. NEAL

only the answer to tKe problem of the morphology of the eye muscles and their nerves but also general conclusions regarding the phylogenesis of the vertebrate head. For it may readily be seen that if nerves are ab initio connected with their terminalorgans, all the modifications of the latter would naturallj' be accompanied by associated changes of the former. If, for example, a post-otic nerve innervates a pre-otic muscle, upon the assumption of an immutable connection of nerve and muscle, the inference that the muscle had migrated from behind the ear into its present position and carried its nerve with it, might appear the most reasonable interpretation of such a peculiar relationship. But if, on the other hand, a nerve is not invariably associated with a given muscle throughout its phylogenetic history, but may grow into muscular territory primitively foreign to it, such an assumption would greatly modify our views of nerve and muscle phylogenesis and possibly our views of the history of the vertebrate head.

The h.istogenetic problem of neuro-muscular relations therefore appears fundamental to any problem, such as that of the morphology of the vertebrate head, involving nerve and muscle in complex and obviously modified relations. Upon the answer to this problem depends, for example, the answer to the problem of the present singular muscular relations, of the trochlear and abducens nerves. While it is an unquestioned fact that nerve and muscle tenaciously retain connections once acquired, and while this conservatism of relation may be taken as a basic assumption in morphology, the possibility still remains that under changed conditions new neuro-muscular relations may be acquired. If Gegenbaur and Fiirbringer be correct, such changes are as incredible as, for example, the inheritance of mutilations are believed to be. But, on the other hand, to all of those who have come to accept the process theory of nerve development and who therefore assume the secondary connection of nerve and muscle, the invasion of new territory and its piratical seizure by exotic nerves appears a possibility. A discussion of the present problem, therefore, does not lead the morphologist far afield. The history of the head must be written primarily in terms of nerve histogenesis.


MORPHOLOGY OF EYE MUSCLE NERVES 23

The attempt has been made recently by Held ('06, '09) and Paton ('07) to combine the two views of nerve histogenesis, claiming a. primary connection of neural tube and myotomes by means of protoplasmic bridges or plasmodesmata, along which, as paths, later the true nerves — neurofibrillae — by centrifugal (Held) or centripetal (Paton) growth — effect the definitive nervous connection with the muscle. True nervous connection with muscle is thus secondary, but the connection takes place along predestined paths — the plasmodesmata.

WTiether the plasmodesmatous connections are secondary or primary remains undetermined by Paton. Held (p. 277) holds with the Hertwig brothers ('78) that they are secondary and he states that Hensen's hypothesis that all nerves have arisen through the incomplete separation of the origin and end cells cannot be correct. Held, however, does not appear to have established the truth of this assertion on the basis of a detailed study of their genesis in the vertebrate embryo. The theoretical importance of the manner of establishment of these 'primary' protoplasmic connections of neural tube and their genetic relations to the cells which form the neuro-fibrillae is so great that it appears worth while to consider this portion of the histogenetic problem with thoroughness and in great detail. The functional importance of the primary protoplasmic connections has been established by the valuable experiments of Paton ('07), who demonstrated that in selachian embryos, in stages before the appearance of neuro-fibrillae, the embryo responds by muscular movements to external stimulus. If such reactions be effected through the medium of the central nervous system, then it is certain that the efferent paths of the impulses must be the "undifferentiated protoplasmic bridges" described by Paton and identified by him as the 'plasmodesms' of Held ('06), To some this evidence might seem sufficient to prove their 'nervous' character, but not so to a disciple of von Apathy, to whom there is no nerve without neuro-fibi411ae." With regard to these protoplasmic bridges, Paton has the following to say (p. 555) :

In the region of the ventral roots in a 5 mm. embryo of Pristiurus there is a marked bulging forward of the protoplasm and in many sections a 'bridge' al)out 4 to 5 /i in width similar to those shown by


24 H. V. NEAL

Kerr ('04) to exist in Lepidosiren may be seen spanning the distance between the edge of the medullary substance and the inner border of the myotome. The structure of the protoplasm of which these strands are composed seems similar in all respects to that forming the general matrix and does not at any point, except at the places to be described later on, present evidence of a fibrillar structure. Less frequently strands of protoplasm are seen in other localities, as for example at the point where later in the development of the embryo processes of giant ganglion cells (Beard) emerge from the cord, or where the matrix surrounding the cells of the ganglionic masses is in contact with the periphery. Any attempt to determine the moment when these bridges appear and the manner in which they are formed necessitates the consideration of questions of fundamental importance.

Are these structures the product of a single cell or do several elements contribute protoplasm to span the interval between two given points as widely separated as the periphery and cord? Hensen's idea that these bridges are originally thrown across from one cell to another and then as the embryo grows these threads are pulled out to many times their original length, is an exceedingly ingenious and suggestive hypothesis but has not yet been proved. There cannot be any ground however for doubting the existence of these structures. The clebatable point is merely in regard to the manner in which they are formed.

Fui^her on he says:

Neal ('03) in an interesting paper on the development of the ventral nerves in Selachians, says that he has been unable in any of his sections to show the existence of a protoplasmic connection 'even of the most attenuated kind between the somite and the neural tube before the first neuraxon makes its exit from the neural tube.' At first it was difficult for me to reconcile this statement with the results of my own observations as well as those of other investigators who have repeatedly observed these bridges in Selachans. At the time when Neal's investigations were conducted there was no method of stainmg which was capable of differentiating the component parts of the neuraxon, and it is not at all improbable that the structure which he had reason to believe was the growing end of a neuroblast was only the undifferentiated protoplasmic band or bridge. In the second place the method of fixation undoubtedly has something to do with the failure to detect the existence of these structures which are much more easily demonstrable in sections fixed in corrosive — acetic or neutral formol than they are in solutions containing picric acid. For the reasons mentioned the structures represented by Neal as neuraxons cannot be accepted as such without further proof.

Paton goes on further to say (p. 556) that:

Probably the long processes depicted by this investigator as being projected from medullary cells are in reality made up of two com


MORPHOLOGY OF EYE MUSCLE NERVES


25


ponents: a short process and the long undifferentiated protoplasmic strand or bridge with which it is apparently fused so as to give, in specimens stained by certain methods the appearance of a single long process. In plate 23, figure 1, one of the bridges is represented, the proximal side of which has fused with the matrix of the cord while the distal is united Avith that of the myotome. (See text-figure A.)

A similar condition is also depicted in figure 2. At later stages one may find connections present between the cord and group of cells, which eventually form the spinal ganglia, and between the latter and



Text fig. A Taken from Paton's figure 2, plate 2.3; my., myotome; neur., neuraxon (neurofibril, Paton) ; rx.v., anlageof somatic motor nerve (plasmodesma, Paton); scl., sclerotome (myotome, Paton); th.n., neural tube, Paton's figure 1, plate 23, is in all essentials like this, but does not show the neuraxon process (neurofibril) stained.

Text fig. B Taken from Dohrn, 1907, figure 7, plate 21; my.,- VanWijhe's second myotome (m.obl.sup.); ir'ch.a., peripheral chiasma of the trochlearis; tr'ch., trochlearis anlage; vs.opt., optic vesicle; V. opt.su., ramus ophthalmicus superficialis trigemini; VII.opt.su., ramus ophthalmicus superficialis facialis.


the periphery. Further there is abundant opportunity to study these plasmodesmata (Held) in the region of the cranial nerves, where undifferentiated links of protoplasm frequently unite the existing ganglionic masses either with the central nervous system or with the periphery. In the case of the oclomotorius and trochlearis the existence of these bridges is very problematic.

The present writer ('03) made the attempt to discover whether the protoplasmic bridges of Paton were primary or secondary


26 H. V. NEAL

connections of neural tube and myotome and he reached the same conclusion as did Dohrn ('88), namely, that they are secondary in origin and that they arise by the extrusion of processes of medullary cells which Neal regards as neuroblasts. Froriep ('04) stated with equal positiveness that the protoplasmic bridges are connected with intramedullary cells. Against the assertions of these three observers we have the statement of Paton that probably the long processes depicted by this investigator as being projected from medullary cells are in reality made up of two components: a short process and the long undifferentiated protoplasmic strand or bridge with which it is apparently fused so as to give, in specimens stained by certain methods, the appearance of a single long process."

This assertion taken in connection with Paton's criticism (quoted above) of my results as based upon inadequate methods of preservation and staining suggests the possibility that Paton himself has been unable to reach positive conclusions regarding the histogenesis of the 'protoplasmic bridges' and the relations of theii* cellular components because of too implicit reliance upon a method primarily suited to demonstrate the neuro-fibrillae but quite unsatisfactory for general cytological purposes. However adequate — or inadequate — the Bielchowsky-Paton method for demonstrating the histogenesis of the neuro-fibrillae, it appears quite unsuited for demonstrating the cellular boundaries and relations. For this purpose, the vom Rath method as used by Neal ('03) is superior. The difference between the cytological results obtained by the Paton method and those obtained by the vom Rath method are strikingly shown by .a comparison of Paton's figure 2, plate 23, and figure 11 of the present paper, which represent almost exactly identical stages in the histogenesis of somatic motor nerves. The most striking difference consists in the absence of cell boundaries in the former and in their clear definition in the latter. (See fig. A, p. 25.)

But Paton's failure to determine the histogenesis of the 'protoplasmic bridges' is not wholly due to his failure to use a method of fixing and staining which demonstrates cell boundaries and relations, l)ut there is evidence also that he did not make a


MORPHOLOGY OF EYE MUSCLE NERVES 27

careful study of the earlier stages in the differentiation of the somite and of the 'protoplasmic bridges.' A study of those earlier stages when protoplasmic connections are formed might have convinced him that Dohrn, Froriep and Neal have stated the facts correctly. That Paton had neglected the study of the genesis of these bridges is evident from his failure to recognize the sclerotome in his sections. For had he been thoroughly familiar with the earlier stages, when the 'protoplasmic bridges' first appear, he would have seen the formation of the sclerotome associated with the development of the bridges and consequently would not have been led to the inference that the primitive fibrils of somatic motor nerves make their first appearance within the myotome. Held ('09, p. 253) has already called attention to this.

The origin of the neurofibrillae and their relation to the protoplasmic processes which form the protoplasmic bridges will be discussed in a later section of this paper in connection with the problem of the genesis of these plasmodesms. This is the essential point of difference between the results of Paton and the writer. I have regarded the medullary processes which form the 'protoplasmic paths' as neuraxon processes, since the neurofibrils are differentiated within them. These processes were therefore spoken of as neuraxons and the statement was made that "before the first neuraxon makes its exit from the neural tube, sections show no protoplasmic connection, even of the most attenuated kind, between the neural tube and the somite." Paton, holding the Apathy view of the independent and exogenous origin of the neurofibrils recognizes no neuraxon — not even an undifferentiated neuraxon — in the absence of neurofibrils. And yet, as Paton has demonstrated in the case of the Rohon-Beard cells, the neurofibrils become secondarily differentiated within the neuroblast and its process, which consisted primarily of undifferentiated protoplasm. The proof that the protoplasmic bridges are formed by neuraxon processes will be stated later in this paper.

If Paton ('07, p. 556) be correct in thinking that the trochlear and the oculomotorius acquire their nervous connections with their muscles without the participation of plasmodesmatous


28 H. V. NEAL

bridges, a conclusion reached by Neal ('98), Dohrn ('07) and Gast ('09), there appears no good reason why he should not accept the conclusion that spinal somatic motor nerves become differentiated and connected with their myotomes in the same manner. If a nerve like the trochlear, which traverses a great distance in reaching its peripheral organ, can do so without the aid of a plasmodesmatous bridge, there would seem very little ground for assuming a different mode of histogenesis for spinal somatic motor nerves. That, as a matter of fact, their genesis is essentially alike is supported by the evidence presented in this paper.

Held ('06, '09) Hke Paton ('07) emphasizes the importance of the plasmodesmatous connections of neural tube and myotome, and concludes, in agreement with 0. and R. Hertwig ('78), that these plasmodesmatous bridges are secondarily formed as cell outgrowths. Not only cells of the myotome and spinal cord but also those of the chorda are supposed by Held to participate in their formation. Held ('09, pp. 91-92) refers to the network between chorda^ tube and myotome as 'Scilly's Fasernetz' and characterizes it as usually a cell- or nuclear-free tissue, since it consists exclusively of the basal cell processes of the germ layers united together. Secondarily, mesenchymatous cells wander into this network.

He further states (p. 281) that at the time of the appearance of the first motor nerves, which have grown from the neural tube, there is everywhere present an abundance of fine connections. The process of transformation of this — a pre-nervous plasma path — into nerves depends upon the sj^ecific power of His' neuroblasts.

With regard to their primary origin, he says (p. 277) that Hensen's hypothesis that all nerves have arisen by the incomplete separation of the origin- and end-cells cannot be correct, since it cannot be thoroughly applied to the details of the development of definite organs in their reciprocal relations. The motor spinal nerve of a frog-larva, for example, can be developed from no mitotically-formed primitive nerve. On the contrary, Held, in agreement with the Hertwigs, regards the plas


MORPHOLOGY OF EYE MUSCLE NERVES 29

matic connections which form the nerve paths as secondary in origin. He has, however, failed to make a detailed study of the genesis of those plasmodesmata which, according to his view, are later transformed into the somatic motor nerve aniagen. Furthermore, one searches in vain through Held's admirable monograph for a convincing proof of the assertion (p. 274) that the growth of the nerve substance is correlated with a resorption of the plasmatic path so that the latter passes over into the former and is utilized in its formation." Similarly unproven is the dogmatic assertion (p. 280) that the 'plasmatische Ausfliisse' of Dohrn ('88) are really not such. Dohrn's characterization of the first protoplasmic connections of tube and myotome as 'plasmatische Ausfliisse' seems admirable. That they quickly lose this undifferentiated protoplasmic character may readily be granted, so that Held's assertion is correct except for the very first stages in the appearance of the protoplasmic bridges.

A number of important questions are involved in the issue raised by Held and Paton: (1) Are protoplasmic connections between myotome and tube primary or secondary? (2) What cells participate in their formation? (3) Have these protoplasmic paths a genetic relation to the neurofibrils? That is, do the same cells which form the plasmodesms also differentiate the neurofibrillae? The writer ('03). concluded that the protoplasmic connections are secondary; that they are formed exclusively by processes of medullary neuroblasts; that within them the neurofibrils* are differentiated. Paton ('07) was not able to determine whether they are primary or secondary; suggests (p. 560) that mesenchymatous cells may participate in their production; and concludes that the neurofibrils have no genetic relation with the protoplasm, of the bridges, but arise independently within the myotome and grow centripetally into the tube along the bridges. On the other hand Held ('06, '09) regards the plasmodesms as secondary in origin; but holds that they are formed by the fusion of cells of chorda, tube and myotome; maintains that they have no genetic relations to the neurofibrils which grow into them secondarily from medullary neuroblasts, for which the paths provide a protoplasmic covering. The great theoretical importance


30 H. V. NEAL

of these problems warrants a reinvestigation of those stages in vertebrate embryos which may contribute to their elucidation. Taking up these problems in order we may first ask whether or not protoplasmic connections between myotome and tube are primary or secondary.

1. Are muscle and nerve connected with each other ab initio? The Hensen view of the primary connection of nerve and muscle is not supported by the evidence presented by Squalus embryos as shown in the drawings represented in figures 1, 2 and 3, which demonstrate the relations of tube and somite in stages previous to the appearance of the anlagen of somatic motor nerves. Figure 1 may seem the most significant of the three since it is taken from a series in which no somatic motor nerve has made its appearance anywhere in the embryo. The section is typical of the trunk region of embryos of this stage and the finer histological features are in all essential respects like those presented by standard methods of preservation and staining. That the space between tube and somite is normal and not an artifact is evinced by the fact that it appears in the living embryo. In sections this region appears to be filled by a vacuolated non-staining or slightly staining plasma. The most significant fact presented in the section is the absence of protoplasmic bridges between tube and somite. Before the first anlagen of the motor roots make their appearance in later stages, it is impossible to find protoplasmic connections between the two organs. The kind of fixation or the stain used makes no diff^erence in the phenomena, provided the space is not obliterated through shrinkage. Reagents, which like formalin cause considerable plasmolysis and vacuolization, and the use of which is likely to be followed by considerable shrinkage, may exaggerate the vacuoles and cause their confluence so that the granular material, which in Flemming or vom Rath preparations is quite evenly distributed, as shown in the drawing, becomes aggregated in denser strands and may simulate protoplasmic strands extending between tube and myotome. However it does not stain like protoplasm, even when aggregated in the manner suggested,


MORPHOLOGY OF EYE MUSCLE NERVES 31

and for this reason has been universally ignored in all drawings. Paton ('07), for example, does not indicate its presence in the sections figured by him (figs. 1 and 2, pi. 23) although it is undoubtedly present in those stages. It may not be confused with the protoplasmic paths emphasized by him and Held ('06, '09).

The plasmodesms of Held ('09) appear to be formed by the union of the plasmoid material of the intercellular spaces with the amoeboid protrusions of the basal cells of the germ layers. At the stage represented in figure 1, protoplasmic processes are lacking in the case of the neural tube, the outside boundary of which consists of an imperforate basal membrane. The chorda is likewise without protoplasmic outgrowths. The median surface of the somite, however, shows a few short and inconspicuous amoeboid extensions. However, between the homogeneous, slightly stained protoplasmic outflows of the somitic cells and the unstained, granular material of the vacuolated plasmoid material between the germ layers, there is distinct contrast, although the two appear to connect with each other. In the drawings the plasmoid substance is greatly exaggerated in order to show it at all. As has been stated, it has been generally overlooked in most embryological studies.

Even when the films of plasmoid material coalesce as the result of reagents, they require intense stains and special illumination to make them visible. They form a most attenuated material for the production of nerves. It would seem most unlikely that growing nerves would trust themselves to such flimsy paths as guides to their destination. That students of nerve histogenesis should seriously consider such all-but-invisible films of non-protoplasmic material as the substance or path of a growing nerve suggests that followers of the Hen sen hypothesis are in desperate need of a material basis for their assumptions. In no true sense do the plasmoid films constitute a primary protoplasmic connection between tube and somite. The actual protoplasmic connections are effected secondarily as the following evidence shows.


32 H. V. NEAL

3. What cells contribute to the genesis of protoplasmic connections

between tube and so7nitef What cells form the first anlagen

of spinal somatic motor nerves?

In stages slightly later than those just described, protoplasmic processes are extended into the intercellular space between tube and somite by cells of neural tube and somite. In the regions in which nerve anlagen later make their appearance protoplasmic movement makes its first appearance sometimes from somitic cells and sometimes from cells of the neural tube (figs. 2, 3 and 4). In either case the ends of the projections appear connected with plasmoid films. But the first real protoplasmic connections between tube and somite are formed by the outflows of medullary cells (fig. 4).

In subsequent stages the reciprocal movement of medullary protoplasm and mesenchymatous cells (sclerotome) results in the close approximation of the two sets of protrusions and in exceptional cases, such as are shown in figures 8 and 9, the two appear indistinguishable. Such close approximation is rare and, when it occurs, appears temporary, so that the genetic relations of medullary cells to the protoplasmic bridges between tube and somite seem indisputable.

Those medullary cells, which by their protoplasmic outflow, effect the first protoplasmic connections between tube and somite usually stain more deeply than the adjacent cells, a fact to which attention was first called by His (79). The same peculiarity distinguishes the Rohon-Beard cells during analogous staged in histogenesis; that is to say, in the stages of neuraxon production. These 'neuroblasts' of the somatic motor nerves are bipolar in shape (figs. 4-11), and within them may be detected, in suitable preparations, a neuro-reticulum with intensely stained fibrils. The reticulum does not appear to be limited to one pole of the neuroblast cell but extends around the nucleus. To its presence may be attributed the deeply staining properties of the neuroblasts. The fact that neurofibrillae make their appearance in the extended processes of these cells sufficiently evinces their neuroblastic character.


MORPHOLOGY OF EYE MUSCLE NERVES 33

Kerr ('02) was led to doubt the hypothesis of the secondary connection of somite and tube on the ground that in the earhest stages, when somatic motor nerves are visible in Lepidosiren the tube and the somite are in immediate contact with each other : Lepidosiren thus affords a definite anatomical basis for the view that the nervous bridge between nerve center and endorgan exists from the beginning, and that the growth of the nerve is a drawing out of this bridge as the end-organ is pushed away by the development of the underlying mesenchyma." On the basis of such evidence Kerr concluded that the Hensen hypothesis is 'almost demonstrated.'

This inference seems to be a non-sequitur from the evidence presented, since Kerr has traced the anlagen of the somatic motor nerves only to those stages when the protoplasmic connections are already established. Neumayer ('06, page 54) has already called attention to the fact that the stages described by Kerr correspond to advanced stages of histogenesis. If tube and somite in Lepidosiren be normally in contact in early embryonic stages, it would seem to be a form little suited to the requirements of an investigation of the primitive connections of nerve and muscle. Squalus embryos, in which a space is normally present, would seem much better objects of research. The normal distance between tube and myotome in Squalus is so small, however, that the theoretical objection that it would be difficult to explain the growth toward the muscle as a tropism is not likely to be suggested. If the direction of growth of the neuraxon processes of the medullary neuroblasts were determined chemotropically by secretions of the myotome cells there would be little chance for the neuraxons to go astray through a mixing or diffusion of specific secretions. In such a case the theoretical necessity for predetermined paths emphasized so strongly by Held ('09) does not appear very convincing. The difficulty of explaining how the extended processes of Rohon-Beard cells reach their area of distribution by a chemotropic response is much greater, but Held can hardly expect that his speculations (page 274) as to the stimulation of the neuroblast process through predetermined plasmodesmatous paths will be regarded as pref JOURNAL OF MORPHOLOGY, VOL. 25, NO. 1


34 H. V. NEAL

erable. But this theoretical difficulty loses its force when applied to the connections of organs in such close proximity as are neural tube and mytome in the trunk region of vertebrate embryos.

To summarize the evidence that connection between neural tube and muscle segment is formed by protoplasmic movement of medullary cells, we have first the fact that in early stages no protoplasmic connections between tube and somite are found. A plasmoid substance fills the space between tube and myotome. Later, medullary cells in the ventro-lateral wall of the tube develop a neuro-reticulum and manifest deeply staining properties. This change is followed by a protrusion of amoeboid processes into the space between tube and somite. Later such protrusions become wider and more extensive, several adjacent cells adding to the size of the protoplasmic bridge thus formed between tube and myotome. The movement of medullary protoplasm is correlated with a migration of sclerotome cells into the space between tube and somite but the latter elements participate only temporarily, if at all, in the formation of the ectoderm-mesodermic connection.

In this manner are formed the anlagen of somatic motor nerves, which at first appear non-fibrillar and protoplasmic and entirely devoid of cells. These are the protoplasmic bridges of Paton ('07). Later the neuroblastic processes extend ventrad, along the median surface of the myotome and in close contact with it and the sclerotome (figs. 10-11). Medullary nuclei soon begin to wander into the anlage from the tube and the movement becomes so extensive that the form of the tube in cross section is greatly changed (fig. 12). All these stages may readily be seen in cross-sections of a Squalus embryo of Balfour's Stage I, by comparing sections beginning with the cloacal region and passing forward toward the head. Connections in the cloacal region at this stage are still unformed, while the movement of medullary cells has already begun in the more anterior metameres of the trunk.


MORPHOLOGY OF EYE MUSCLE NERVES 35

4. Have these protoplasmic connections a genetic relation to the

neurofibrils?

Paton decides adversely to the view that the protoplasmic bridges are formed by the processes of neuroblasts on the ground that they do not contain fibrillae. He does not deny, however, that under certain circumstances a nerve cell may throw out a process of very considerable length." But he thinks that "the greatest caution should be observed in assuming that mere length of process, without positive knowledge regarding the nature of the structures contained in it, is in any sense to be considered a criterion as to whether a cellular prolongation is or is not to be called a nerve." To "refer to an undifferentiated tract as a nerve would give rise to endless confusion."

Paton ('07, p. 560) finds that in sections fixed in sublimateacetic and stained by hematoxylin-eosin, the medullary cells appear 'faintly fibrillar;' but he thinks that this 'primitive fibrillation' has no connection with the development of the neurofibrils.

In his opinion, the conclusions of those who claim a connection between the primitive ventral root fibrils and medullary cells are based upon faulty technique. "No reliable method of staining has yet been employed that is capable of demonstrating the presence of processes in the vicinity of the distal ends of these primitive filaments." According to Paton, the first neurofibrils or 'primitive fibrils' make their appearance within the myotome at a point remote from the protoplasmic bridges, as shown in Paton 's figure 2, plate 23, shown in outline in textfigure A of this paper (p. 25).

The primitive neurofibrils are coarse, deeply stained structures appearing primarily in a locality where more than in any other place the ground substance, even after sublimate fixation, seems to be granular in character, while the more delicate and attenuated filaments only become visible at later stages. . . : . The apparent independence of these primitive neurofibrils in the ventral roots from cells is one of their distinguishing characteristics, but in the large cells of Beard a different arrangement exists. There the fibrils appear in the apical process of the cell close to its nucleus (compare figure 2 with figure 12, plate 23).


36 H. V. NEAL

Thus Paton not only doubts the neuroblastic origin of the connections between the tube and myotome but also the neuroblastic origin of the neuro-fibrillae. That he has been led into error in his inference of the independent origin of the neurofibrils will be suggested by a comparison of Paton's figure 2 with figure 12 of this paper. Paton is mistaken in his inference, partly because of his failure to trace carefully in earlier stages the genesis of the protoplasmic bridges; partly because he has not used a method suitable to demonstrate cell boundaries, and partly because the neuro-fibrillae in his preparations are incompletely stained, as has already been stated by Held ('09). A comparison of Paton's figure 2 with figure 12 of this paper shows that the 'primitive fibril' which Paton thinks arises within the myotome actually lies between the myotome and the sclerotome. Moreover, its position corresponds with the position of the termination of the cell processes of medullary cells. The two drawings strikingly show the difference in the histological results of the \-om Rath and the Bielchowsky-Paton methods. The great advantage of the former appears in the sharp definition of the cell boundaries, of the latter in the clear differentiation of the neurofibrils. The convincing proof given by Held ('09) of the endogenous origin of the neurofibrils within medullary neuroblasts certainly warrants the inference that the neurofibrils are incompletely stained by the Bielchowsky-Paton method. Later in his paper Paton admits the possibility that the neurofibrils are genetically related to the processes of the medullary cells. But he thinks that new technical methods are needed in order to solve the problem thus raised. To determine this, however, needs not so much a new method of neurological technique as a careful study of the successive stages in their differentiation by methods which we now possess, especially a method like the vom Rath which defines clearly the cell boundaries and relations. Paton's procedure admirably supplements such a method.

Held ( '06, '09) has made a most thorough and painstaking investigation of the histogenesis of the neurofibrils on the ground that, as stated by Max Schultze, von Kupffer and von Apathy, these are the distinctive structures of the nerve fiber. In agree


MORPHOLOGY OF EYE MUSCLE NERVES 37

ment with Apathy, Held is of the opinion that the solution of the problem of the devolopment of nervous tissue depends primarily upon the demonstration of the place and manner of origin of the neurofibrillar substance. He concludes that this important nervous element is differentiated within the neuroblasts of His, which are thus the essential nervous cell centers. These results are in harmony with the experimental results of Harrison, Lewis and Burrows.

Held finds that the first and surest histological characteristic of a cell of neuroblastic tendency lies in the occurrence of a specific network or neuro-reticulum which appears in a circumscribed region of the neuroblast in the vicinity of the nucleus. In the course of its development this network undergoes an extraordinary complication and extension of its substance. The observation of the primary appearance of the neuro-reticulum within the neuroblast had previously been made by Besta ( '04) .

Contrary to the opinion expressed by His, Held affirms that it is not the outflow of protoplasm into a cell process, but the definitely directed growth of a new and special cell substance which originates the first nerve trunks in the embryo. The process is not that the cell produces a protoplasmic elongation of the cell body in which subsequently and secondarily a fibrillar substance appears. On the contrary there are inner changes within the protoplasm of a neuroblast that lead eventually to the differentiation of a neurofibrillar substance which, while primarily of a minimal extension within the fibrillogenous zone of the neuroblast cell, later in the course of its special growth undergoes a mighty extension in the body after it has produced externally the pear-shaped form of neuroblast first observed by His. This fibrillar structure of the neuroblasts is 'neurofibrillar' in the von Apathy sense of the word.

Held's conclusions do not agree with those of von Apathy ('98) who holds that the ganglion cells produce no neurofibrils but are secondarily penetrated by them. According to Apathy, the neurofibrils are produced by other and special cells — the 'nerve cells' — which are not to be confounded or identified with 'ganghon cells.' The 'nerve cells' produce, ac


38 H. V. NEAL

cording to von Apathy, the conducting or neurofibrillar substance just as muscle cells produce muscle fibers. From these 'nerve cells ' the conducting substance grows on the one hand toward the center in the 'ganglion cell' and on the other hand toward the periphery into the sensory cell or muscle cell and so forth. The nuclei of the nerve cells lie in the course of the nerve fibers themselves and form in vertebrates the nuclei of the neurilemma sheath. The protoplasmic body of the nerve cell is, in general, spindle-shaped so that it may be called the nerve spindle. Its membranous boundary is the neurilemma.

For every motor nerve fiber of a vertebrate, on the other hand, according to Held, there may be distinguished a central and peripheral nerve stretch. The external limiting membrane of the neural tube divides the two. The neurofibrillae of the central stretch are formed before the peripheral. They arise through the unilateral growth of the neurofibrillar substance of a unipolar or bipolar neuroblast which proceeds basalwards in the direction of the Rabl's 'chief axis' and soon transcends the outer limits of the neural tube in its growth toward the terminal organ. The central stretch of the nerve varies with the position of the neuroblast cell. In its extent through the marginal zone the nerve fiber becomes secondarily surrounded and enclosed by glia cells.

Held distinguishes three chief stages in the development of the peripheral stretch of the nerves:

1. That of the outgrowing nerve itself, which has not yet reached its terminal organ with its specific substance, but nevertheless is connected with its terminal organ by means of a simple undifferentiated plasmodesmatous strand. Theoretically this stage is the most important. The His doctrine, according to which the amoeboid neuroblast processes extend through open spaces in the tissues, cannot be correct, since in the Anamnia the growing point of the motor nerve is connected with portions of the epithelium, that is by plasmodesmatous connections, and in the Amniota on the other hand with parts of the mesenchyma.

2. In the second stage the nerve has reached its muscle anlage, which it not only touches superficially, but cells of which


MORPHOLOGY OF EYE MUSCLE NERVES 39

it penetrates with its neurofibrils. At this stage the motor nerve anlage in Anamnia is non-cellular while in Amniota it is cellular. The cause of this difference is as follows : In case the nerve -anlage makes its appearance while the germ layers and the anlagen of the organs are still epithelial as in the Anamnia, the peripheral neurofibrillar tract lies in nuclear-free and net-like plasmodesms, which connect the related epithehal surfaces with each other. If, on the other hand, a connective tissue is already developed, then such nerve fibers extend through a variable number of cell-bodies. In the first case the nerve appears primarily without nuclei, while in the second it appears as a nucleated structure. The difference is due to a difference in the relative time of differentiation of the neuroblasts.

3. The third stage is a transition to the condition in the adult nerve with its cellular neurilermna sheath, and arises through the emigration of medullary cells into the fibrillar nerve and their differentiation as sheath cells.

Held summarizes his views regarding the histogenesis of motor nerves as follows:

The origin of motor nerves rests upon the peripherally directed growth of the specific cell substance of definite neuroblasts of the neural tube and which proceeds in the direction of the chief cell axis of the neuroblasts, and, for definite but unkown reasons, transcends the outer boundary of the embryonic neural tube. A stage of development precedes the formation and growth of the nervous substance itself, by means of which the neural tube and its motor neuroblasts, through net-like arranged paths of a simpler and not yet neurofibrillar substance, is brought into coimection with the peripheral muscle anlage by complicated processes of outgrowth.

Such connection paths, which in Anamnia are simple and epithelial and in the Amniota, on the contrary, are a complicated connective tissue, are in the Hensen's sense of the word 'used' by, and in some sort of an unknown manner fused with, the nervous substance growing from the neuroblasts. The growth of the nerves to the terminal organ does not proceed in the liquid-filled vacuoles of the intercellular spaces as affirmed by His, since the growing points of the nervous substance are connected both laterally and at the outer extremity with the fartherreaching plasmodesms of the surrounding tissues. Later, by means of complicated processes of multiplication and movement of medullary cells, which wander secondarily into the primarily non-nuclear, or nucleated, nerve-path as cell-elements organically connected with it, the motor nerve becomes the multinucleated strand supplied with


40 H. V. NEAL

neurilemma cells. The cell-chain hypothesis has entirely looked upon and conceived the beginning of this third stage as the actual beginning of the specific formation of nervous substance. Actually, however, this condition belongs to that period of the development which transforms the embryonic type of nerve into the adult and differentiated structure.

In presenting evidence of the secondary connection of nerve and muscle, and perhaps more convincingly, evidence of the endogenous origin of the neurofibrils and their genetic relation to the medullary neuroblasts, Held supports the essential points of the process theory of nerve development. His conclusions, however, differ from those usually held by supporters of the KupfferBidder theory in regard to the origin of protoplasmic connections of nerve center and end-organ. In this regard Held considers himself an advocate of the Hertwig theory, since he holds that protoplasmic coiinection of tube and myotome is formed, not by the neuroblasts which form the neurofibrils, but independently as plasmodesms of indeterminate and multicellular origm.

There is nothing in Held's monograph, however, to indicate that he has given as careful attention to the development of these plasmodesms as to the histogenesis of the neurofibrils. The evidence presented by the writer ('03 and in the present paper) supports the view of His, von Lenhossek, Cajal that the same cells w^hich form the neurofibrils also form the protoplasmic connections between tube and myotome. In other words, the neuroblasts themselves form the protoplasmic connections of tube and myotome. As shown in plates 1 and 2, the processes of medullary neuroblasts extend along the median surface of the somite between myotome and sclerotome and within these processes (w'hich may be traced to their connections with medullary cells as stated by Froriep, '04) neurofibrils make their first appearance as is clearly shown in preparations made by suitable methods of staining. In the light of Held's results, it seems probable that a neuro-reticulum is present wdthin the neuroblast cell, but no method sharpl^^ differentiates it in Selachians in early stages of histogenesis. Sometimes in Bielchowsky-Paton preparations a fibrillar network appears in certain medullary cells in stages before protoplasmic connection with the mvotome has been formed. That this is neu


MORPHOLOGY OF EYE MUSCLE NERVES 41

rofibrillar has not been demonstrated, however, although the analogy of the histogenesis of the Rohon-Beard cells would favor this opinion.

Although the evidence presented in this paper is not sufficient to demonstrate positively that the neurofibrils of somatic motor nerves arise endogenously within those medullary cells which form the protoplasmic connections of tube and myotome, the conclusion that such is the case appears justified on the following grounds:

• 1. The evidence goes to show that medullary cells, by a process of outgrowth, form the first protoplasmic connection of tube and myotome.

2. In subsequent stages these processes extend ventrad along the median surface of the somite between myotome and sclerotome.

3. Within these processes which in vomRath preparations may be traced to their connection with medullary cells, the neurofibrillae make their appearance.

4. Since, as shown by Paton ('07) the neurofibrils arise endogenously within the processes of Rohon-Beard cells, and Besta ( '04) and Held ( '06, '09) have demonstrated a similar histogenesis in somatic motor neuroblasts of Amniota, analogy would seem to support a similar origin in Selachians.

Coghill ( '13) has been able to demonstrate that in Amblystoma the neurones of the somatic motor column become well differentiated and typically oriented in the spinal cord before the ventral roots appear at the corresponding level. The earliest demonstrable root fibers arise as collaterals from these neurones, usually from descending processes.

5. Is the neuraxon of a soryiatic motor fiber multicellular in origin, or is it the process of a single medullary cell?

Since the begimiing of the twentieth century the SchwannBalfour 'cell-chain' theory of nerve development has been supported by Rafaelle ('00), Bethe ('00-'07), Oscar Schultze ('04- '07), Brachet ('05- '07), C'ohn ('05- '07) and possibly von Apathy ( '07) . From its inception the cell-chain hypothesis has depended


42 H. V. NEAL

for its support largely upon the well established fact of the cellular structure of embryonic nerves. It was this fact that caused Kupffer finally to abandon his process theory. Then also von Apathy's important discoveries of the intercellular relations of the neurofibrils have seemed to some to favor this view of neurogenesis, so that the theory appears to have a new lease of life. Bethe ('03) is of the opinion that he has established the fact of the dependence of the axis cylinder upon the 'nerve cell' — in the Apathy sense — because he finds that in his preparations the mitotic division of a nucleus interrupts the nerve fiber. Held ( '09) suggests that this evidence might warrant another inference not complimentary to the quality of Bethe 's preparations. Taken as a whole, the arguments in favor of this much discussed hypothesis, which has derived the larger part of its support from preparations unsuited to the requirements of neurological investigation, seem most unconvincing.

Held's ('09, p. 51) discovery of the polyneuroblastic origin of nerve fibers should not be taken as a confirmation of the cell-chain hypothesis, since it does not involve the idea of a cell-chain but asserts the formation of some neuraxons by the fusion of the processes of adjacent neuroblasts.

Bardeen ('03, p. 255) and the writer ('03) have advanced arguments against the claim that the cells of embryonic nerves participate in the formation of the fibers of the nerves. In this connection, Bardeen says that "in an early embryonic nerve of moderate size one finds many hundred fibrils enclosed by a sheath of flattened cells, but with no cells among them. In such nerves one can most easily see that the fibrils are not differentiated parts of cells lying in the nerve."

The phenomena of motor nerve histogenesis in Squalus affords no support to the cell-chain hypothesis, since somatic motor nerves in this animal acquire connection with the myotomes and a fibrillar structure before cells make their appearance within the nerve .anlage. The cellular structure emphasized by the advocates of the hypothesis appears only in somewhat advanced stages in histogenesis. Evidence that the cell elements present in the nerve anlagen have a genetic relation to the fibers or neu


MORPHOLOGY OF EYE MUSCLE NERVES 43

rofibrils is wholly lacking. Proof that they form the neurilemma has been given by the writer ('03) and by Carpenter

('06).

6. Hoiv does the elongation of the neuraxon take place?

Does the elongation of a nerve involve simply a migration of protoplasm from the neuroblast cell as the recent experiments of Harrison ('04, '11), Lewis ('06, '07) and Bm-rows ('11) seem to prove? Or does the neuraxon fiber grow by the progressive differentiation of a primary plasmatic connection between tube and myotome as Sedgwick ('94) and Held ('07) have suggested?

Are we to accept Cajal's evidence of the free termination of the neuraxon fiber in the intercellular spaces of the embryo? Or shall we agree with Held ( '09) that ' ' in the growing nerve we have to do with an organically advancing growth, and not with the pushing through of a cell process?"

Although the raising of these questions takes us back to the time of von Baer and the beginnings of embryology, it cannot be said that they have been finally answered. Indeed, the problem of how the elongation of the neuraxon process is effected remains one of the most vexed questions of histogenesis.

Cajal ('07) describes the nerve fibers as advancing and growing through the mesoderm, using the cellular interstices. To him the 'Leitzellen' and their anastomosing expansions are always situated on the sides of the axons which, as the result of the disturbing action of the reagents (pyridine, alcohol), attach themselves to, or partially unite with, the mesodermic framework. To Held ('09) such evidence of nerve terminations ending freely in the intercellular spaces is a result of the rupture of the plasmodesms, with which he affirms the growing tip is normally connected. The growth of the nerve substance he states (p. 274) is correlated with a resorption of a plasmatic path so that the latter passes over into the former and is utilized in its formation.

Harrison ('01) states that the sensory nerves of Salmo salar possess fine lateral branches which might easily be confused


44 H, V. NEAL

with coagulation threads, which are everywhere present," but, unHke Held, finds no genetic relations between such coagulation threads and the elongating neuraxon process. Later ('07) Harrison writes that "it is by no means certain that the plasmodesms are not artifacts — products of coagulation," and von Lenhossek ('06) offers the same criticism.

Such criticisms suggest to Held ('09) that Cajal and Harrison have not seen the true plasmodesms, which are not only ver}' easily ruptured but appear everywhere different from coagulation threads. They may be covered with coagulation bodies, or not; that varies with individual preparations. If, however, one has ever seen these fine plasmodesmata branch out in characteristic fashion from subdividing and branching cell processes, it is impossible longer to hold the view that these fine and definitely branching threads are coagulation products or artifacts.

Held emphasizes the delicacy of the plasmodesmata and the ease with which in fixation or imbedding and sectioning they may be ruptured and lost. While this is unquestionably true, it is also a fact that the plasmodesms become more conspicuous in preparations in which there has been excessive plasmolysis and vacuolation, as in many preparations made for special neurological purposes. On the other hand, the more faithful the cytological fixation, the less conspicuous becomes the coagulable su?jstance found in the intercellular spaces where the neuraxon processes of medullary neuroblasts and those of Rohon-Beard cells grow towards their end-organs.

While Held (p. 297) admits that Harrison's experiments prove that processes may grow out from neuroblasts in suitable fluids, that nerve elongation may occur exclusively by protoplasmic movement and therefore does not necessarily involve the participation and differentiation of plasmodesms, nevertheless he maintains that a normal nervous system oriented to the terminal-organ is formed only with the participation of plasma connections already present. If, however, it may be assumed that the experiments of Harrison ('04-'ll), Lewis ('06, '07), and Burrows ('11) throw any light upon normal processes of histogenesis — a conclusion which Held does not seem willing to admit


MORPHOLOGY OF EYE MUSCLE NERVES 45

— they tend to show that however important the plasmodesms may be as guides for the growing nerve, their share in its production is neghgible. They prove as convincingly as experimental data possibly can the logical fallacy of Held's inference, that because the growing tip of the neuraxon is connected with branched plasmodesmatous processes that therefore these processes have a genetic relation to the neuraxon. The significant fact in this connection is that Harrison's preparations show these same processes at the termination of the neuraxons growing freely in salt solutions! The genetic relations are here precisely the opposite . to those inferred — or assumed — by Held in the normal growth of the neuraxon. Even if it be admitted that Held's assertion is correct that the growing and more or less branched point of growth of the nervous substance at its extremity possesses, in advance of its momentarily attained length, a projecting extension which connects (as an undifferentiated plasmatic mass between the termination of the growing nerve and the end-organ), the neuroblast with the terminal-organ, his conclusion, already quoted, that the plasmodesm is utilized in the formation of the nerve is a logical non-sequitur. Nowhere in his excellent monograph does Held give adequate evidence to prove this assertion, and the experimental evidence seems to make the assumption unnecessary. Furthermore, the assertion that the growing termination of the neuraxon is connected by a strand of undifferentiated protoplasm with the end-organ has not been demonstrated.

As further evidence in favor of the view that neuraxons and neurofibrils are differentiated within plasmodesmatous or protoplasmic strands, and consequently in conflict with the idea that they grow 'naked' into the vacuolar spaces of the embryo, Held emphasizes the fact that from their first appearance the neurofibrils appear surrounded by a layer of granular protoplasm, a relation better seen in cross-sections than in longitudinal ones. On the other hand, von Lenhossek ('06) states, on the basis of observations on the histogenesis of spinal nerves in the chick, that first and foremost is it untrue that the young fibers are embedded in protoplasm" but that on the contrary, "the fibers


46 H. V. NEAL

lie definitely free, and in the spaces between them there is not the slightest amount of protoplasm to be seen. Held denies the accuracy of this description as based upon unsatisfactory methods of preservation and staining. Cajal ('06, '07) in Held's opinion has been led into error for the same reason. On the other hand, Ko Hiker ('05) finds the fibers of the embryonic trochlearis nerve in a calf surrounded by sheaths of an intermediate substance. Gurwitsch ('00) mentions a lamellar network which grows into the bundle of 'naked' fibers from the surrounding mesenchyma sheath and divides up the nerve bundle, while according to Kappers ('04) the nerve bundle is coarsely split or divided by fine processes of the neurilemma cells.

In his search for evidence to support the hypothesis of the plasmodesmatous origin of nerve fibers, it seems not to have occurred to Held that the protoplasmic envelope of an embryonic nerve fiber may be produced by the same neuroblast that formed the fiber. The simple fact of a protoplasmic sheath around the neurofibrillae no more proves the existence of nerve 'paths' than it does the hypothesis of neuroblastic outflow. The evidence is in harmony with either supposition. In Squalus, embryonic nerve fibers are 'naked' in the sense that primarily they have no cellular sheaths. But it is also a fact that they have granular, protoplasmic sheaths, visible, not only in sections, but in cover glass preparations of the living nerve fiber. The substance of the sheath stains lightly under the same treatment which stains the fiber intensely. For this reason the protoplasmic envelope is easily — and has been generally — overlooked. The thickness appears to vary in proportion to the length of the neuraxon process. In nerve anlagen, like the trochlearis, for example, in which the fibers are especially long and slender, the protoplasmic sheaths of the fibers are proportionally^ thin. In spinal somatic motor nerves the sheaths are relatively thick, becoming thinner as the nerve fibers increase in length. The ingrowth of processes of neurilemma cells described by Gurwitsch ('00) and Kappers ('04) occurs in later stages and adds to the interfibrillar protoplasmic substance.


MORPHOLOGY OF EYE MUSCLE NERVES 47

The intraplasmatic position of the nerve fiber or neurofibrillar bundle, therefore, instead of proving the doctrine of primary plasmatic nerve 'paths,' is equally in harmony with the process theory of nerve development. An examination of Squalus embryos shows that, sooner or later, within the plasmatic neuroblast process which effects the first connection between tube and somite, the neurofibrillar bundle is differentiated as an axial fiber surrounded by the undifferentiated protoplasm of the neuroblast process. The advancing end of the nerve consists of undifferentiated protoplasm and is amoeboid in appearance. The phenomena in sections of preserved embryos are essentially identical with those in vitro of the living nerve fiber. The elongation of the neuraxon, therefore, instead of involving the use and incorporation of primary plasmodesmatous paths into the growing nerve, on the contrary is effected by a movement of the protoplasm of a neuroblast cell and the endogenous differentiation of the neurofibrils. With especial clearness are the phenomena strikingly shown in the giant cells of Rohon-Beard.

a. The histogenesis of the cells of Rohon-Beard. The conclusions based upon the study of the histogenesis of spinal somatic motor nerves are greatly strengthened and confirmed by the phenomena presented by the development of the cells of Rohon-Beard. These phenomena are presented in figures 13 to 22. The close propinquity of the nerve center (neural tube) and the terminal-organ (myotome) in the case of the spinal somatic motor nerve makes it very difficult to find wholly convincing evidence of the extension of the neuraxon process, and of the secondary nature of the connection between the two organs.

■ In the case of the Rohon-Beard cells, however, the neuraxon process, in reaching its peripheral termination, grows into and through spaces where mesenchymatous cells are entirely absent. As shown in figures 13 to 16, the neuraxon process appears primarily as an amoeboid extension of a large, deeply stained cell, lying in the dorsal wall of the neural tube. As the neuraxon process becomes further extended, its peripheral termination shows many pseudopodia-like extensions. In some cases


48 H. V. NEAL

(e.g., fig. 21), as the neuraxon process reaches the apex of the myotome, some of the pseudopodial processes extend median and some lateral to the myotome. It appears to be undetermined whether, in its fm-ther extension, the neuraxon ma}^ grow between the myotome and the neural tube or between the myotome and the ectoderm. The significant fact presented b}- such a section and by all the sections which show the cells of RohonBeard with their greatly elongated processes, is that there is not the slightest suggestion of a direct primary connection between the cell and its later connection with the ectoderm in the extra-embrj'onic blastoderm.

Such relations as are shown in figure 20, afford still further evidence of the genetic relations of ganglion cell and neuraxon. The section shows a Rohon-Beard cell lying at the apex of a sensor}' ganglion, and which has reached its present position by migration from the dorsal wall of the neural tube. The continuity of the cell bodj^ and of the neuraxon and evidence of their genetic relation to each other is more clearly shown than in the case of sections where the Rohon-Beard cell lies within the wall of the neural tube. The objection to the inference that somatic neuraxons are formed as elongated. processes of medullary neuroblasts, on the ground that it is impossible to trace the protoplasm of the neuroblast through the wall of the neural tube into the growing nerve anlage is met by the evidence presented in this figure.

Doubt remains only as to the inference that the process is formed b}' a migration of protoplasm from the neuroblast into the neuraxon. ^^^ly not infer that the neuraxon is formed from intercellular bridges in situ? In answer to this question, it is sufficient to state that in the earlier stages when the neuraxons of the Rohon-Beard cells first appear, there are no protoplasmic bridges extending in the direction taken later by the growing neuraxons, but simpl}" a non-staining plasma. Furthermore, it is as logical to conclude that the bod}'- of the Rohon-Beard cell shown in figure 20 was differentiated in situ as that its neuraxon was formed in situ. On the contrary, if there be good reason for inferring that the Rohon-Beard c.ell has reached its present


MORPHOLOGY OF EYE MUSCLE NERVES 49

position as the result of protoplasmic migration from the neural tube, there is equally good reason for the conclusion that its neuraxon has been formed by the same method.

7. What is the source of the cells of somatic motor cells?

The cellular structure of embryonic somatic motor nerves has been emphasized by many embrj^ologists since Schwann ('39) advanced the cell-chain hypothesis of neurogenesis and since its subsequent revival by Balfour ('77). The question of the source of these cells has never been fully determined. Are they medullar}' or mesenchymatous in origin?

Balfour ('75) was the first to infer the migration of medullary cells into somatic motor nerve anlagen and this conclusion has been confii-med by many investigators including Dohrn ('88), Beard ('88j, Herrick ('93), von Kupffer ('94, '95), Hoffmann ('97), Harrison ('01), Bardeen ('03), Xeal ('03), Schultze ('04'07), Held ('06, '09), Carpenter ('06), Carpenter and Main ('07) and Kuntz ('10-'12). His ('89) and von Kolliker ('92), on the other hand, have denied the derivation of these cells from the neural tube. The former concluded that the cells which appear to be in the process of migration from the tube, are later enclosed b}' the 'Randschleier' which prevents their escape. If it were otherwise, he thinks, the process would lead to the formation of motor ganglia. To His it . appeared quite inconceivable that connecti^•e tissue elements could come from the nervous system. The mesenchjmiatous origin of the cells of somatic motor anlagen has also been assumed by Vignal ('83), von Lenhossek ('97), Kolster ('99), Gurwitsch ('00) and Bardeen ('03). Von Kolliker ('05) came finally to accept the probabihty of the medullary origin of the neurilemma. According to ^Marcus ('09), the cells of somatic motor nerve anlagen come from the neural crest.

The fact that sclerotome cells are in close proximity to the somatic motor anlagen has led some students to infer that the cells present in later stages within the nerve have had a mesenchymatous derivation. In fact, the mesodermal origin of all connective tissue cells has seemed so well attested that many

JOURXAL OF MORPHOLOGY, VOL. 25. XO. 1


50 H. V. NEAL

have taken for granted the mesenchymatous nature of all cells associated with motor nerve anlagen. Kerr ('04) seems to have made this assumption for the cells associated with the somatic motor anlagen of Lepidosiren. He says (p. 121) that "richly yolked masses of mesenchymatous protoplasm become aggregated around the nerve, which till now has been quite naked." Again ('02) he writes that "in tracing back the motor trunks of the spinal nerves I reached a stage in which the faintly fibrillated trunk was ensheathed in a nucleated mass of protoplasm of mesenchymatous origin." The mesenchymatous origin of these cells has seemed so self-evident to Kerr that the problem of tracing them to their source does not seem to have occurred to him, and in consequence he advances not a particle of evidence that he has traced their genesis.

That the cells of somatic motor nerve anlagen in Squalus are largely, if not exclusively, of medullar}^ derivation seems demonstrated by the following facts:

First, when cells make their earliest appearance in the nerve anlagen they lie partly in and partly outside of the neural tube. Then later, when cells are found definitely' within the anlagen, as seen in figure 12, they appear at the base of the anlagen and near the tube. As cells grow more numerous in the anlagen in later stages more nuclei appear to be in the process of migration from the tube. As a result of the migration, the contour of the tube as seen in cross-section becomes changed and the nerve anlage greatly thickened. Furthermore, the nerve anlage shows a limiting membrane continuous with that of the tube, which makes it possible to distinguish the boundary of the nerve anlage and to infer its independence of mesenchymatous cells in the immediate vicinity. Moreover, the boundaries of the cells of the anlagen retain their smooth epithelial contour, while the mesenchymatous cells assume characteristically irregular, branched outlines. Whereas at first the cells of the sclerotome from which the mesenchyma of the region is derived, are closely apposed to the nerve anlagen, as seen in the figures on plates 1 and 2, they soon lose contact with the nerve and with each other and become metamorphosed into a loose connective tissue, but still retain


MORPHOLOGY OF EYE MUSCLE NERVES 51

connection with each other by fine plasmodesmatous threads. Thus the study of serial sections of Squalus embryos in closely related stages of histogenesis permits no doubt that the cells of somatic motor nerve anlagen are in large part migrant medullar}^ elements.

From an estimate of the number of migrating medullary cells and a comparison with the estimated number in adult nerves the writer ('03) concluded that the neurilemma receives accessions in later stages from the mesenchyma and that the mesenchymatous participation in the formation of the sheaths of the adult nerve is greater than is the medullary contribution. The reasons for this inference are more convincing if, as has been suggested by Harrison ('01) and others, some of the migrant medullary elements form the sympathetic anlagen. The writer agrees with His, Jr. ('97) that at least a part of the cells of somatic motor anlagen have a mesenchymatous derivation.

8. What is the fate of the cells found in somatic motor yierve anlagen?

Three views have been advanced as to the fate of the cells of somatic motor nerve anlagen:

1. They are 'nerve-forming cells' and they secrete the neurofibrillae.

2. They form the neurilemma cells.

3. They become the ganglion cells of the sympathetic. According to the first view the cells of somatic motor nerves

anlagen form the nerves, either by fusion into cell-chains as Schwann ('39) and Balfour ('78, '81) suggested and has since been maintained by Marshall ('78), Van Wijhe ('82, '86, '89), Beard ('85, '88, '92), Beraneck ('87), Goette ('88), Dohrn ('91), Miss Piatt ('94, '96), Sedgwick ('94) with modifications, Hoffmann ('96), Kupffer ('90, '91, '94), Capobianco e Fragnito ('98), Rafaelle ('00), Bethe ('00-'07) with modifications, Brachet ('05, '07), Cohn ('05-'07) and Oscar Schultze ('04, '07); or as the 'nerve cells' which secrete the neurofibrillae and which attain connections with ganglion cells and muscle fibers along plasmodesmatous paths without the participation of cell-chains.


52 H. V. NEAL

By his discovery of a 'ventral neural crest' in Ammocoetes Kupffer was led to abandon the cell-process theory of nerve development and to adopt the cell-chain hypothesis. According to Sagemehl ('82) and Held ('09), however, somatic motor nerves in Petromyzon are primarily fibrillar. Consequently, von Kupffer's inference may be regarded erroneous as the result of defective observation.

Schultze ('04-'07) has adopted a modified form of the cellchain hypothesis from his observation of the growth of nerves in amphibian embryos, in which he finds the nerves primarily cellular. He consequently regards the nerves as a syncytium of peripheral neuroblasts, on the assumption that the cells are the formative ones which secrete a nerve. Thus Schultze regards the neurilemma as ectodermal in origin.

According to Held ('09), however, Schultze has failed to see the earlier stages of nerve histogenesis before the anlagen have acquired a cellular character. Held ('09) regards the cells which find their way into embryonic nerves as 'Leitzellen,' and holds that they have nothing to do with the formation of nerve fibers more than to furnish the paths in which the neurofibrils are driven forth from the neuroblasts. But Schultze claims that the fibril-free cells (Held's Leitzellen) are as much neuroblasts as muscle-forming but primarily fibril-free cells are myoblasts.

The writer ('03) argued on the following grounds that the cells of somatic motor nerve anlagen have nothing whatever to do with the formation of neuraxons:

1. In the earliest stages of histogenesis, when the number of fibers increases most rapidly in the nerve anlage, the cells of the anlage are distinctly peripheral in relation to the fibrillar bundles (Bardeen, '03, has also emphasized this point).

2. None of the ventral nerve cells at any stage of histogenesis , show the deeply-staining properties of neuroblast cells such as appear in the ganglia of the dorsal (somatic sensory) nerves. Without exception the cells of the somatic motor nerve anlagen are vacuolated, granular and lightly stained.

3. While marked changes in size and shape (correlated with the growth of the neuraxons) appear in the neuroblasts of the


MOEPHOLOGY OF EYE MUSCLE NERVES 53

sensory ganglia, no such changes appear in the cells of the somatic motor nerves.

4. While the long axes of the neuroblasts of the sensory nerves correspond, and are parallel with the direction of growth of the neuraxons, no such relation is seen in the cells of the somatic motor nerves in those early stages when the fibers increase in number most rapidly.

5. The neurofibrils do not lie within, nor are they directly connected with, the protoplasm of the cells of the somatic motor nerve anlagen. The cells in the earUer stages of histogenesis are clearly peripheral to the neurofibrillae. On the basis of such considerations, it appears extremely doubtful if the cells of somatic motor nerve anlagen have genetic relations to the nerve fibers. The cells must have some other fate.

On the other hand, Cajal ('08) thinks that he has been able to discover nerve-forming cells in a bipolar phase in the somatic motor nerves of the chick.

9. What is the histogenesis of the neurilemma?

A second, and practically unanimous opinion is that the cells of embryonic nerves form the neurilemma sheaths. Indeed, this has beien the opinion of investigators, whatever their creed as to the histogenesis of nerve. According to supporters of the cell-chain theory (e.g., Dohrn, '91), that portion of the peripheral protoplasm of the nerve cells which does not enter into the formation of the neuraxon becomes the neurilemma sheath. The same opinion is shared by supporters of the process and of the plasmodesm theory. Difference of opinion centers chiefly about the problem of the origin of the cells.

Harrison ('06), on experimental grounds, concluded that the neurilemma cells of the frog are exclusively derived from the' neural crest and this opinion is shared by Held ('09) who thinks that the neurilemma cells have a similar origin in Axolotl and Triton. According to Held the neurilemma cells are peripherally emigrated glia cells which have secondarily followed neurofibrillar tracts aheady laid down. Their source varies in different vertebrates. In Petromyzon and the Selachii the source


54 H. V. NEAL

is exclusively medullary. In the trout some of the cells possibly come from the dorsal ganglia, while in the Amphibia they come exclusively from this source. In reptiles, birds and mammals they appear to come chiefly from the sensory ganglia.

According to the writer ('03) the cells of somatic motor nerves of Squalus have a medullary origin. Figures 12 to 32 (pis. 23-24) show the gradual transformation of cells into the neurilemma of the adult. Since the fact is unquestioned there appears no reason for repeating the evidence upon which this conclusion is based. Bardeen ('03), Carpenter ('06), Carpenter and Main ('07), also drive the neurilemma from migrant medullary cells. Kuntz ('10) confirms this conclusion.

But is it not possible that the medullary elements contained in somatic motor nerve anlagen enter into the formation of the sympathetic? Such a fate has been claimed for them by more than one competent investigator.

10. To ivhat an extent do the emigrated medullary elements enter into the formation of the sympathetic ganglia?

In recent j^ears the Remak-Kolliker view that the sympathetic system is of mesodermal origin has been shared by few investigators. Onodi ('86) considered such an origin as possible for the peripheral sympathetic plexuses and both Paterson ('90) and Fusari ('90) have held the same view of the genesis of the elements of the sympathetic.

Balfour ('78) and Schenk and Birdsall ('78) led the way to the present understanding of the ectodermal origin of the sympathetic. Balfour did not decide whether the sympathetic cells were derived from dorsal or ventral roots, but Schenck and •Birdsall inferred their origin from the dorsal ganglia, a conclusion since drawn by Onodi ('86), VanWijhe ('89), His ('90), His, Jr. ('91, '97), von Kollicker ('94), C. Rabl ('97), Kohn ('05, '07), Lillie ('08), Held ('09) and Marcus ('09).

Hoffmann ('00) inferred a double source of origin of the sympathetic elements as derivatives both of the dorsal and ventral roots, an opinion shared by Neumayer ('06) and Kuntz ('11).


MORPHOLOGY OF EYE MUSCLE NERVES 55

An exclusively medullary origin of the sympathetic elements has been inferred as a possibility by Harrison ('04) on experimental grounds, and Froriep ('07) draws the same inference. Kuntz ('10, p. 250) who had stated in an earher paper that Froriep 's conclusion "is probably correct with regard to the neurones in the sympathetic trunks and the prevertebral plexuses," in a later paper ('11) supports the view of the double source of the sympathetic elements.

Held ('09) draws the conclusion that the cells of the motor roots have nothing to do with the formation of the sympathetic system, on the following grounds: First, the anlage of the sympathetic ganglion in Selachii lies entirely in the axis of the sensory nerve root and not in that of the motor root. Second, the motor root has fewer cells,' is nowhere thickened, and is nowhere united with the sensory nerve by means of cell-strands. Moreover, the ventral nerve extends along the myotome parallel with the sensory nerve, but its cellular material is not mixed with that of the latter. Also, the motor root lies lateral to the sensory root, nearer the myotome and farther from the sympathetic. The two are not united by cell-strands. Finally, the two unite ventral to the sympathetic anlage.

With Kohn and Froriep, Held admits the principle of a free cell migration of purely nervous cell clusters, in the formation of the sympathetic, but thinks that the sympathetic ganglia arise by the budding of chain-like connected cell masses and strands containing both ganglion cells and sheath cells.

Kuntz ('11) is the most recent student of the histogenesis of the sympathetic in Selachii, and his is the latest presentation of the case in favor of the medullary origin of some of the elements of the sympathetic. Kuntz 's argument that the cells of the ventral roots as well as those of the dorsal participate in the formation of the sympathetic appears to consist of the following considerations: In sections of earlier stages (11 mm. Squalus embryos) scattered cells of large size and deeply staining nuclei appear in the mesenchyma between the myotome and the dorsal aorta. Because of their staining properties and evidence of cellular migration along the nerve anlagen, Kuntz assumes these


56 H. V. NEAL

cells to be derived from the dorsal ganglia and the ventral nerve roots. In sections of later stages (15 mm. embryos) the anlagen of the sympathetic ganglia may be seen as clusters of similar cells attached to the nerves at the level of the aorta. The deeply staining cells of the mesenchyma are now less numerous. He therefore infers that they have entered into the formation of the ganglia. Proof that the sympathetic anlagen are not the product of the continued migration of cells from the dorsal and ventral roots is not presented. As to the evidence of migration, Kuntz is able to obtain such kinetoscope effects as the following: '^ numerous cells (in a 10 mm. embryo) push out from the ventro-lateral angles of the neural tube and migrate peripherally along the paths of the motor nerve-roots." That such cells enter into the formation of the sympathetic ganglia he thinks is proved by the following considerations (p. 206) :

As my observations on the early stages of embryos of Acanthias have shown, numerous cells migrate peripherally both from the neural crest and from the ventral part of the neural tube before the fibers of the ventral nerve-roots can be traced peripherally to the level at which the sympathetic trunks arise. These cells later become aggregated to form the anlagen of the sympathetic trunks. In view of these facts it can not be doubted that in embryos of Acanthias many cells which have their origin in the ventral part of the neural tube enter the anlagen of the sympathetic trunks.

While there seems little reason to doubt that cells of somatic motor nerve anlagen in Squalus participate in the formation of the neurilemma, convincing evidence that they migrate into the anlagen of the sympathetic is wanting. The assertions of Kuntz in this connection appear quite unconvincing. In the numerous papers by this investigator which have appeared in rapid succession in recent years, covering the development of the sympathetic in all classes of vertebrates except amphibia, observations and conclusions, sense impressions and inferences are indiscriminately advanced as 'observations.' Kuntz, for example, does not bring forward the slightest evidence to prove the assertion ('11, p. 183) that the cells which migrate peripherally from the neural crest and the neural tube .... become scattered in the mesenchyma of the region lying between the lateral mus


MORPHOLOGY OF EYE MUSCLE NERVES 57

cle plates, the notochord and the aorta." The fact that some nuclei found in the mesenchyma have a more deeply staining quality certainly does not prove such an unqualified assertion. That emigrated medullary elements enter into the sympathetic is possible, but Kuntz has presented no facts that make this inference seem more certain.

So far as the evidence from sections of Squalus embryos goes, it seems rather to favor the view that the sympathetic anlagen receive their cellular elements largely if not exclusively from the sensory ganglia, as inferred by investigators upon all classes of vertebrates from Schenck and Birdsall to Held and Marcus. In the first place there appears no reason to doubt the ventral movement of the cells which enter into the formation of the dorsal ganglia. The erltrance of cells into the sympathetic anlagen would involve only the continuation of this migration on the part of some of these cells. The masses of cells which constitute the anlagen of the sympathetic appear somewhat more closely connected with the sensory bundle of the embryonic nerve than with the motor bundle. Such relations appear in frontal sections at the level of the ganglionic anlagen (fig. 30) .

In Squalus, however, the separation between sensory and motor bundles in the mixed spinal nerve, on the median side of which the sympathetic anlagen first make their appearance, does not appear as distinct as described by Held for other Selachii. In such sections cells lie between the sensory and motor bundles, and protoplasmic strands • connect the motor bundle with the ganglia, so that the possibility that medullary elements enter the ganglia from the motor root does not seem excluded. Connection, on the other hand, by no means proves a migration. Under the circumstances inferences appear most uncertain.

But if one admit an analogy between the development of cranial and spinal somatic motor nerves, there are facts connected with the development of the oculomotor and the trochlear which favor the inference of the predominantly ganglionic origin of the sympathetic anlagen. For the first clusters of cells associated with the anlagen of these nerves are derived from the neural crest. These cell clusters, in their relations and — in the case of


58 H. V. NEAL

the ciliary — in their adult structure, appear to be sympathetic. Their derivation from the neural crest favors the inference that the sympathetic anlagen of the trunk have a similar origin.

It may be further urged in favor of the ganglionic origin and against the medullary derivation of the sympathetic anlagen in Squalus that the medullary migration into somatic motor nerves, like the abducens and the hypoglossus which have no sympathetic anlagen, is as striking as in those nerves associated with sympathetic ganglia. The migrant cells of the hypoglossus and abducens can have no other destination than the neurilemma. On the other hand, sympathetic cells do migrate along the vagus nerve, the ganglionic nerve of the hypoglossus segments. So the evidence seems to favor the opinion that in Squalus the sympathetic anlagen are derived from neural crest elements and not from emigrated medullary cells.

11. Summary of the histogenesis of spinal somatic motor nerves

The more important features of the histogenesis of spinal somatic motor nerves in Squalus are shown in figures 23 to 30, and may be briefly summarized as follows:

Nerve and muscle are not primarily connected with each other in Squalus embryos. Previous to the establishment of protoplasmic connection, the space normally found between somite and neural tube is filled by a vacuolated, non-staining, nonprotoplasmic liquid containing a relatively small amount of coagulable material.

Protoplasmic connection of somite and tube is established by an amoeboid protoplasmic extrusion from cells in the ventrolateral wall of the neural tube forming the 'protoplasmic bridges' of Paton ('07j or 'plasmodesms' of Held ('06). The cell processes which form these connections extend gradually along the median surface of the somite between myotome and sclerotome.

Within these processes the neurofibrils soon make their appearance, as shown in Cajal and Bielchowsky-Paton preparations. This evidence demonstrates the neuroblastic nature of the cells


MORPHOLOGY OF EYE MUSCLE NERVES 59

which form the protoplasmic connections between tube and somite and their processes are therefore to be regarded as neuraxons.

The cell-chain hypothesis of neurogenesis receives no support from the evidence presented in sections of Squalus embryos. The neurofibrillar structure appears in the nerve anlagen before any cells are present in them.

The growth of a nerve fiber toward its terminal-organ does not involve the use and resorption of primary plasmatic paths but simply the movement and differentiation of the protoplasm of the medullary neuroblast. The most convincing demonstration of the truth of this assertion is afforded by the growth and extension of the processes of Rohon-Beard cells.

The numerous cells which, in somewhat advanced stages of histogenesis, make their appearance in the ventral nerve anlagen are not of mesenchymatous but of medullary origin; exclusively so in the earlier stages of development. That mesenchymatous cells are added to the growing nerve in more advanced stages to form the connective tissue sheaths seems probable.

The cells of the motor nerve anlagen have no genetic relations to the neurofibrils or neuraxons. In other words they are not 'nerve cells' in the von Apathy sense, nor do they unite in chains to form the neuraxons or neurofibrils with their sheaths. Whether or not they participate in the formation of the sympathetic is an open question. The evidence on the whole favors, but does not prove, the conclusion that most of the cells of the sympathetic have their source in the dorsal ganglia. That the cells of the motor nerve anlagen in Squalus for the most part form the neurilemma cells can be convincingly demonstrated.

Thus the phenomena of spinal motor nerve histogenesis in Squalus support the Kupffer-Bidder-His theory of nerve histogenesis, which has recently been strongly confirmed by the brilliant experiments of Harrison and Lewis.


60 H. V. NEAL

HISTOGENESIS OF THE EYE MUSCLE NERVES

Does the histogenesis of the eye-muscle nerves resemble that of spinal somatic motor nerves, and are they therefore morphologically comparable? Upon the answer to this question depend in very large degree our views of the morphology of the vertebrate head.

1. General description of the histogenesis of the eye muscle nerves

The three eye-muscle nerves make their appearance at different stages in the development of the embryo; the oculomotor in a 9.5 mm. embryo; the abducens in a 10 mm. embryo; and the trochlearis in a 19 mm. embryo. They arise like spinal somatic motor nerves as plasmatic protrusions from the wall of the brain. Their terminations are amoeboid and remain so until connection with the muscle is established. The primary anlagen possess a peripheral layer of clear protoplasm of somewhat granular structure, within which may be detected an axial deeply-stained fiber (fig. 32). The thickness of the peripheral protoplasmic sheath varies in inverse ratio with the length of the fiber. As the nerve anlage elongates the granular sheath becomes exceedingly thin. The trochlear, from the time of its first appearance, in contrast with the abducens -and oculomotor, appears fibrillar rather than protoplasmic, owing to the delicacy of the plasma film surrounding the axial fibers (figs. 49 and 50). This difference may be attributed to the extended course of the fibers of the trochlear within the brain wall.

Evidence of a primary plasmatic path extending from the amoeboid terminations of the eye muscle nerve anlagen to the myotomes with which they are later connected is lacking. The demonstration of their presence or absence is rendered difficult, however, as a result of the presence of a loose mesenchyma in the spaces traversed by the elongating nerve anlagen. The anlagen assume close relations with mesenchyma cells and their processes, but so far as may be observed these connections are secondary and are independent of the process of elongation of the nerve anlagen. In all cases where it is possible to identify


MORPHOLOGY OF EYE MUSCLE NERVES 61

the growing tips of the nerve fibers the amoeboid terminations have no intimate relations with plasmodesms extending toward the muscle anlage, although they indisputably have more or less indirect connections with the processes of adjacent mesenchyma cells. Evidence that these connections are utilized in the extension of the nerve anlage has not been discovered, and the burden of proof rests upon those who assert the absorption of these plasmodesms into the nerve path.

The relations of the nerve fibers with bipolar neuroblasts in the somatic motor column of the brain are as easily demonstrable in Squalus embryos as in the adults and are especially clear in Cajal preparations. Even in Vom Rath preparations it is possible to follow with certainty the deeply stained processes of the neuroblasts into the roots of the nerve anlagen. There appears no good reason to doubt the genetic relation of the neuroblasts to the nerve fibers. The fact that the anlagen of the eye-muscle nerves are primarily fibrillar and not cellular, and that the cells are secondarily added to the anlagen and remain for some time distinctly peripheral in relation to the bundle of fibers, and that none of these cells assume the appearance of neuroblastic cells, constitutes evidence quite irreconcilable with the cell-chain hypothesis.

The cells which make their appearance in relation with the anlagen of the eye-muscle nerves are partly of mesenchymatous and partly of medullary derivation, like those associated with spinal somatic motor nerves. Cells are also added to the distal portion of the trochlear and oculomotor anlagen from the superficial and profundus branches of the trigeminal nerve, forming the anlage of the ciliary ganglion in the case of the former, and of a transient (?) ganglion in the case of the latter. Thus the relations to sympathetic ganglia of these two nerves seem to be the same as those of spinal somatic nerve anlagen. The abducens assumes no connections with ganglionated nerves and (therefore?) is associated with no sympathetic ganglion. "V\Tiether any of the medullary cells associated with the oculomotor and trochlear anlagen participate in the formation of the sympathetic ganglia is uncertain, as no criterion of distinction between the


62 H. V. NEAL

various cell elements derived from different sources has been found. The clusters of ganglion (?) cells associated with the trochlear anlage, which the writer interprets as homologous with sympathetic ganglia, are clearly derivatives of the superficialis nerve and not migrant medullary elements. Most of the cells associated with the anlagen of the eye-muscle nerves may be traced into the neurilemma cells of the differentiated nerves. Their gradual penetration into the fiber bundles of the anlagen, beginning with their proximal and distal extremities, may be easily followed in successive stages.

The nidulus of the oculomotor is limited to the somatic motor column of the midbrain vesicle (neuromere II) and presents no evidence of subdivision into two or more niduli. The nidulus of the trochlear lies in the somatic motor column of the first hindbrain neuromere (neuromere III). That of the abducens is post-otic and extends through two post-otic neuromeres (VII and VIII). These relations afford important clues to the primitive segmental relations of the e3^e muscle nerves, and present an insuperable difficulty to the hypothesis that the eye muscles have migrated into their present relationships from a post-otic source.

2. Does the histogenesis of the ocvlomotorius resemble that of a somatic motor spinal nerve?

a. Are the muscle and nerve, that is, midbrain and somite 1 of VanWijhe, connected with each other ab initio? Among the opinions in regard to the histogenesis of the oculomotor nerve no one has maintained upon an observational basis that the nerve is ab initio connected with the first somite of VanWijhe. Not even Sedgwick ('94), who regarded himself as a supporter of the Hensen theory of the primary connection of nerve and muscle, was able to maintain this view by the histogenesis of the oculomotor; for he stated that the oculomotor is differentiated from the ciliary ganglion to the floor of the midbrain." It appears, therefore, justifiable to infer — in view of the failure of all students of this much investigated nerve to demonstrate a


MORPHOLOGY OF EYE MUSCLE NERVES 63

primary protoplasmic path connecting its nidulus and its endorgan — that in this respect the histogenesis of the oculomotor resembles that of a spinal somatic motor nerve as described above. The evidence about to be presented goes to show that the midbrain vesicle becomes connected with the first headcavity by means of the movement or continuous extension of the protoplasm of neuroblasts situated in the somatic motor column of the midbrain.

There is, however, a slight difference in the conditions under which connection of tube and somite are effected as compared with those which obtain in the growth of spinal somatic motor nerves. As a result of the late appearance of the oculomotor anlage, relatively to the formation of a loose mesenchyma in the head, the region between midbrain and somite, which had been earlier filled with an unstaining plasmoid substance, becomes filled with a loose mesenchyma of uncertain origin, and protoplasmic connection of tube and somite is effected by growth in the midst of loose mesenchyma. Since it is almost as difficult to find the oculomotor anlage in the midst of this connective tissue, especially in preparations made by the older non-neurological methods, as to find the proverbial needle in a haystack, it is not surprising that the discovery of the true manner of growth of the nerve has been delayed and that difference of opinion has arisen.

Since, however, today all investigators, whose observations are sufficiently extended to cover this point, agree that midbrain and the first head cavity become connected with each other secondarily, and since the opposite view is, for this nerve, an assumption without a single drawing or demonstration to support it, there seems to be no good reason why an affirmative answer should not be given to the question: Are protoplasmic connections between somite 1 and midbrain primary or secondary? Held's ('09) assertion, that the neurofibrillae of the oculomotor alw^ays appear as intra-plasmatic fibrils, affords no foundation for inductions regarding the genesis of these protoplasmic paths. Held himself admits that he does not know whether the paths are primary or secondary.


64 H. V. NEAL

Gast ('09, p. 428) expresses himself as agreeing with Dohrn that primary paths do not exist for the oculomotor. His observations confirm those of Harrison ('07) that the first fiber anlagen are amoeboid and those of the oculomotor are the processes of medullary neuroblasts.

In disproof of the view that the oculomotor is primarily connected with its muscle, Belogolowy (10, p. 366) mentions the fact that in embryo chicks a transient branch of the oculomotor extends to the external rectus muscle, later innervated by the abducens. Belogolowy is unable to reconcile the presence of this transient and aberrant relationship with the doctrine of the fixed and unchangeable relationship of nerve and muscle.

According to Filatoff ('07) the oculomotor anlage appears in Emys primarily as a cellular strand extending from the myotome of VanWijhe's first somite toward the base of the midbrain. Filatoff, therefore, infers the primary connection of nerve and muscle. Johnson ('13), however, finds that in reptile embryos the anlage of the pculomotorius becomes secondarily connected with its myotome. " Filatoff 's conclusions are not applicable to Chelydra."

b. What cells contribute to the formation of the protoplasmic connections between midbrain and somite 1 f According to Miss Piatt ('91), Mitrophanow ('93) and Sedgwick ('94) protoplasmic connection between midbrain and somite 1 is initiated by a proliferation of cells from the mesocephalic ganglion toward the brain. According to Sedgwick this migration is accompanied by a differentiation of a pre-existing syncytial strand. Ziegler ('08) also infers the centripetal growth of the oculomotor, on the basis of the evidence that in a 25 mm. embryo of Chlamydoselachus the oculomotor "has not extended its growth as far as the floor of the midbrain." Since the oculomotor nerve makes its first appearance in a 9-10 mm. selachian embryo and connection between brain and somite is already established in a 10 mm. embryo, Gast is undoubtedly correct in inferring that the oculomotor anlage in Ziegler's specimen was ruptured through shrinkage and that his- inference is consequently fallacious. Miss Piatt, Mitrophanow, and Sedgwick were also dealing with ad


MORPHOLOGY OF EYE MUSCLE NERVES 65

vanced stages in the histogenesis of the oculomotor and they were consequently led astray in their inductions.

Filatoff's ('07) assertion that the oculomotor is differentiated from the myotome toward the brain in Emys has not been confirmed by the more recent investigation of reptile embryos by Johnson ('13). According to a second view — upheld by His ('88), Chiarugi ('97), Neal ('98), Held ('09), Gast ('09), Carpenter ('06), Belogolowy ('10) — the first protoplasmic connection between midbrain and somite 1 is effected precisely as has been found for spinal somatic motor nerves, by the continuous extension of the processes of medullary neuroblasts. Sooner or later, according to Dohrn ('88, '91), C. L. Herrick ('93), Goronowitsch ('93), Chiarugi ('97), Carpenter ('06) and Gast ('09), this protoplasmic movement is accompanied by a migration of medullary cells from the base of the midbrain. Neither Neal nor Held were able to find convincing evidence of nuclear migration into the oculomotor roots, but they did not deny the fact of migration.

The evidence presented in sections of Squalus embryos and represented in the figures of plate 6 of this paper, strongly confirms the opinion of those who have maintained the genetic relations of medullary neuroblasts in the ventral wall of the midbrain to the first protoplasmic connections between midbrain and the first somite. The anlage of the oculomotor makes its first appearance in Squalus embryos of 9 to 9.5 mm. (fig. 58), as a short, deeply-staining fiber, formed by the union of the processes of medullary neuroblasts. The evidence that this is the anlage of the oculomotor consists in the fact that the point of attachment to the brain, the direction of the long axis of the process, and the histological appearance correspond with those of the oculomotor in a 10 mm. embryo when the connection of the nerve with the myotome is already established. The nerve anlage already consists of two roots, each containing a deeply staining fiber surrounded by an envelope of granular protoplasm. At the point of union of the two roots a mesenchymatous cell appears in close proximity to the anlage. Evidence of the migration of medullary cells is wanting at this stage. The deeply staining processes of neuroblasts at the base of the midbrain

JOURNAL or MORPHOLOGY, VOL. 25, NO. 1


66 H. V. NEAL

extend toward the nerve anlage and may in some cases be traced to a point outside the external limiting membrane of the brain wall.

In later stages (fig. 59) a larger number of processes of medullary cells may be traced toward the nerve anlage and a larger number of deeply stained fibers may be seen within the roots of the anlage. In this, as in all subsequent stages, mesenchymatous cells may be seen in close proximity to the nerve fibers. Some earlier stages, such as are represented in my ('98) figures G and i/, pages 222-223, show no cells whatever in relation to the oculomotor anlage; the majority of sections of these early stages show cells more or less closely associated with the nerve anlage. That these cells are mesenchymatous seems indicated by the fact that there is in these stages no evidence of medullary migration and that the possibility of a derivation from the mesocephalic ganglion is excluded, since the nerve has as yet no connection with that ganglion. No evidence of a genetic relation of these cells to the nerve is discoverable. In all respects they resemble the branched granular cells of the surrounding mesenchyma. Comparison of these with later stages in the development of the oculomotor (figs. 61, 64, 70) favors the inference that the fibers of the oculomotor are of neuroblastic origin and that they have their nidulus of origin in the base of the midbrain.

The oculomotor therefore attains connection with its terminal organ secondarily, in precisely the same manner as a spinal somatic motor nerve.

c. Have these protoplasmic connections genetic relations to the neurofibrils of the oculomotor? Most embryological investigators of the oculomotorius have paid no attention to the genesis of the neurofibrils. This is true of the latest research upon the histogenesis of the oculomotor — that of Gast ('09) — as of the earlier ones. In fact, no investigation of the histogenesis of the nerve by the use of a specifically neurofibrillar stain has been made. The vom Rath method, as apphed by Neal ('98) and Carpenter ('06), comes the nearest to a truly neurofibrillar stain of any which have been used upon the nerve. Of the histological


MORPHOLOGY OF EYE MUSCLE NERVES 67

results of this method Carpenter ('06, p. 195) says that "a heavy black precipitate along the neuraxons differentiates these clearly against the less darkly colored stroma in which they appear to be imbedded." The appearances presented in sections prepared by this method resemble those in Cajal and Bielchowsky-Paton preparations in demonstrating, in the very earliest stages of histogenesis, deeply staining fibers within the nerve anlagen. And, while the results are not as specific and differential in vom Rath preparations, the staining is more complete than in either of the other two, so that the connection of the fibers with medullary neuroblasts is more clearly demonstrated. Now, since in vom Rath preparations there is no stage in the development of the oculomotor when such deeply stained fibers are absent from the nerve, and since the connection of these fibers with processes of medullary neuroblasts can be readily traced in many sections, and since there is not the slightest evidence that the cells associated with the fibrillar bundle of the nerve have any genetic relations with the fibers, it appears legitimate to conclude that the neurofibrils are differentiated within the processes of the midbrain neuroblasts, which, by their peripheral extension, form the protoplasmic connection between brain and somite.

Vignal ('83), Bardeen ('03), Carpenter ('06) and Paton ('09) have called attention to the coarseness of the primary fibers of embryonic nerves as compared with those seen in later stages of histogenesis and have inferred a process of splitting of the primitive fibers into true fibrils, an inference that seems to have been convincingly demonstrated by Miss Dunn ('02). The present paper makes no contribution to the discussion of this important histogenetic problem.

In the neuroblastic origin of its fibrils the- oculomotor resembles histogenetically a spinal somatic motor nerve.

d. Are the neuraxoyis of the oculomotor multicellular in origin or are they the processes of medullary neuroblasts? The evidence that the neuraxons of the oculomotor are formed as the result of protoplasmic movement of neuroblasts in the somatic motor column of the midbrain consists, first, of the fact that the growth of the nerve is centrifugal as stated for various vertebrates by


68 H. V. NEAL

His ('88), Dohrn ('91), Chiarugi ('97), Neal ('98), Carpenter ('06), Gast ('09), Belogolowy ('10). Appearing at first as protoplasmic outgrowths from the base of the midbrain, the anlage of the nerve elongates in later stages in the direction of their terminal organ — the first somite — with which it quickly becomes connected.

In the second place, the nerve, in all stages of differentiation in embryonic material, manifests a fibrillation in the form of coarse, deeply-staining fibrils which appear connected with the deeply staining processes of midbrain neuroblasts. On the basis of analogy with phenomena seen in the histogenesis of RohonBeard cells and in preparations of the living nerve fiber in coverglass preparations, such evidence points toward the neuroblastic origin of the neurofibrils. In no sections prepared by any of the neurological methods used, is evidence to be found of the fusion of centripetally growing fibers with processes of medullary cells in the manner suggested by Paton ('07) .

Finally, the inference is further strengthened by the absence of any positive evidence in favor of the cell-chain hypothesis. It is true that cells appear along the nerve anlage from very early stages in histogenesis, but none of these cells show the familiar form and staining properties of neuroblasts. On the basis of this evidence the conclusion appears warranted that each of the neuraxons of the oculomotor is the product of one or more medullary neuroblasts and not the product of the fusion of a chain of cells. The analogy with the histogenesis of spinal somatic motor nerves is clear.

Burckhardt ('92) describes medial and lateral niduli of the oculomotor in Protopterus, confirming Ahlborn's ('84) description of the central, relations of the nerve in Petromyzon. Ahlborn's assertions, however, are not confirmed by Johnston ('05) who finds the nidulus of the oculomotor to be that of a somatic motor nerve. All investigators agree that the nidulus of the oculomotor in selachians belongs in the somatic motor column.

e. By what means does increase in the length of the neuraxon take place? The oculomotor develops under conditions which differ somewhat from those which have been described for spinal


MORPHOLOGY OF EYE MUSCLE NERVES 69

somatic motor nerves, since the region through which it grows is filled with a loose mesenchyma before the growing tip of the nerve emerges from the base of the midbrain. Under such conditions it is difficult to ascertain that processes of mesenchymatous cells make no contribution to the growth of the nerve. The burden of proof, however, falls upon those who claim that the nerve grows by the differentiation, in situ, of plasmodesmatous paths. Such proof has not been given.

As has been suggested, the presence of a granular envelope around the fibers of the oculomotor is not to be taken as evidence of primary protoplasmic paths. Were a primary protoplasmic path of the diameter of the anlage of the oculomotor present, it could easily be found and traced in the sections. The fact is, however, that in all the stages in which the nerve anlage may be traced to its peripheral termination in the stages before it reaches the myotome, the distal end shows an amoeboid character, precisely resembling the distal extremity of nerves growing in cover-glass preparations as drawn and described by Harrison, Lewis and Burrows. No direct relation with the processes of adjacent cells nor with a direct plasmodesmatous path extending toward the myotome can be detected. All of the phenomena advanced by Held ('09) in his attempt to reconcile the Hensen with the Kupffer-Bidder-Harrison theory and to establish the assumption of primary plasmodesmatous paths are fully in harmony with the process theory of the free outgrowth of nerve fibers, while the Hensen-Held theory of primary plasmodesmatous paths finds as little support in the phenomena presented in the histogenesis of the oculomotor as in the growth of spinal somatic motor nerves.

/. What is the source of origin of the cells of the oculomotor anlage? Dohrn ('91) was the first to affirm the medullary origin of the cells of the oculomotorius anlage and his conclusions have been repeatedly confirmed by later investigators, among them Goronowitsch ('93), C. L. Herrick ('93), Chiarugi ('97), Carpenter ('06) and Gast ('09). Neal ('98) and Held ('09) were unable to find convincing evidence of, but they did not deny, medullary migration into the oculomotor. Dohrn's argument


70 H- '^'- NEAL

for his conclusion was that cells of the midbrain migrate peripherally and then send out processes, which unite in a network, just outside the base of the midbrain, to form the stem of the nerve. Immediately at the beginning of the plasma outflow, cells are seen half in and half ovit of the wall of the tube, and later, before the oculomotorius shows any connection with the mesocephalic ganglion, large deeply-staining nuclei, like those which appear in the process of emergence from the neural tube, make their appearance in the network of the roots of the nerve. In later stages the nuclei, in increasing numbers, appear to emerge from the midbrain into the nerve anlage. The phenomena of medullary migration into spinal somatic motor nerves add support to Dohrn's contention.

Evidence similar to that presented by Dohrn ('91) may be found in sections of Squalus embryos in stages from 10 to 15 mm., and confirms the accuracy of his description. The cells closely associated with the oculomotor in the earliest stages of development, appear to be mesenchymatous rather than medullary in origin, but it seems not unlikely that these, like similar mesenchymatous cells associated with spinal somatic motor nerves in the very earliest stages of their appearance, are not permanently associated with the nerve. Their relations may be purely topographic and transient. But there is no reason to doubt that some of the cells of the oculomotor anlage have, like those of spinal nerves, a medullary origin.

According to Hoffmann ('85), Ewart ('90), Miss Piatt ('91), Mitrophanow ('93), Sedgwick ('94), Chiarugi ('94, '97), Neal ('98), and Gast ('09) cells migrate into close relations with the oculomotorius anlage from the mesocephalic ganglion. I was not sure that these cells came into permanent relations with the nerve, while Gast inferred the derivation of some of the sheath cells from them. Harrison's demonstration, through experiments upon amphibian embryos, of the derivation of neurilemma cells from the neural crest might appear to support Gast's inference, although nerve histogenesis in amphibia is not necessarily analogous in all features with that in selachians. Gast's evidence of the forward and back — centripetal and centrifugal —


MORPHOLOGY OF EYE MUSCLE NERVES 71

migration of cells derived from the mesocephalic ganglion along the oculomotor nerve is not very convincing.

Carpenter ('06) finds evidence of free cell migration from the mesocephalic ganglion into the loose mesenchyma and towards the oculomotor anlage in the chick. Guthke ('06, p. 43) on the other hand, asserts that the ganglion of the profundus nerve (which he erroneously calls the ciliary ganglion) has absolutely nothing to do with the oculomotorius which has no connections with the profundus ganglion in Torpedo.

Belogolowy ('10, p. 275) infers, in agreement with Carpenter ('06), that the cells, which in the chick migrate from the brain, probably participate in the formation of the motor neuroblasts of the ciliary ganglion; and he is of the opinion that neural crest cells play an important part in the formation of the accompanying cells.

Several have inferred the participation of the mesenchyma in the addition of cells to the oculomotor anlage. Indeed that has always been the orthodox assumption of the derivation of the neurilemma of the nerve. Held ('09) was in doubt as to the actual origin of the neurilemma nuclei of the oculomotor, but regards them as belonging to mesenchymatous cells of unknown origin. There can be no doubt of the close association of mesenchymatous cells in all early stages of the histogenesis of the nerve, and it seems not unlikely that they contribute to the formation of the sheath cells of the nerve. In fact Gast ('09) confirms the assertion of His, Jr., and Romberg ('90) that neurocytes migrate freely through the mesenchyma toward the nerve anlage.

Thus, in the derivation of its cellular elements, the oculomotor of Squalus resembles a spinal somatic motor nerve. Part at least of its component cells are of medullary origin ; some of them maj^ be mesenchymatous; and, finally, the nerve has close association with cells derived from a cerebro-spinal ganglion, the mesocephalic or profundus ganglion.

g. What is the fate of these cells? The fate of these cells appears even less certain than their derivation. The opinions have been expressed that at least some of the cells — whether derived


72 H. V. NEAL

from the midbrain or from the mesocephaUc ganglion — are neuroblast cells which form the fibers of the oculomotor; that the cells of the anlage form the neurilemma; that they contribute to the formation of the ciliary ganglion; and that those which migrate from the midbrain, if not also those that are derived from the mesocephalic ganglion, have both fates.

Taking up these views in turn, we find that Dohrn ('91) is the only one who has attempted to support the cell-chain hypothesis on the basis of evidence derived from the study of the development of the oculomotor nerve, but as he has wholly repudiated ('07) the considerations which earlier led him to advocate the cell-chain hypothesis, it seems inadvisable to discuss his arguments in favor of the former theory.

Gast ('09) asserts that some of the migrant medullary cells become located in the 'root ganglion' of the oculomotor, where they become differentiated as neuroblasts, but the evidence upon which he bases his statement is far from convincing. It would seem necessary, even were one prejudiced in favor of such a view, to be able to demonstrate in its favor more convincing evidence than the presence of spindle-shaped or even multipolar cells in the cell mass which Gast calls a ganglion, especially in preparations made according to the usual embryological methods which do not differentiate the distinctive histological elements of ganglion cells. In a section made by usual embryological methods, how can Gast tell whether a spindle-shaped cell is a nerve cell or a sheath cell? Even multipolarity is no decisive criterion of a ganglion cell. Therefore, while there may be no theoretical reasons why the migrant medullary elements may not be expected to form ganglion cells, it would seem that some more critical evidence than form is to be taken as proof that such is their fate. None of the methods used in the present investigation upon Squalus embryos demonstrate the presence of neuroblasts or ganglion cells among the cells in the roots of the oculomotor. Whatever multipolar cells appear in close proximity to the nerve appear to be similar to adjacent cells of the mesenchyma.


MORPHOLOGY OF EYE MUSCLE NERVES 73

On the other hand, Carpenter ('06) gives convincing reasons for thinking that some of the migrant medullary cells from the midbrain participate (as a group of smaller cells easily distinguishable from the larger ganglion cells derived from the mesocephalic ganglion of the chick) in the formation of the ciliary ganglion. If this conclusion be confirmed — and this has been done by Belogolowy ('10) — it appears, that at least some of the medullary cells in the oculomotorius anlage are neuroblastic.

Gast ('09) is not able to confirm Carpenter's conclusion that some of the migrant medullary elements enter into the formation of the neurilemma, biit is of the opinion that the neurilemma is formed by migrant cells from the mesocephalic ganglion which travel centrad along the nerve anlage as far as the roots by which it arises from the brain and thence they migrate back again into the nerve to become the neurilemma cells. As has been stated above. Cast's evidence of this forward and back migration does not appear very convincing. Analogy with the derivation of the neurilemma of motor roots in higher vertebrates from the sensory ganglia with which they are connected does not prove such a derivation in selachians.

That the migrant medullary elements participate in the formation of the neurilemma seems to Carpenter evinced by the following considerations: '^Many of these nuclei, once out on the nerve, become elongated as they move away from the neural tube," and "maintain throughout development their close proximity to the nerve fibrils." Again, the absence of evidence of the intrusion of cells from the mesenchyma is further indication of the ectodermal origin of the neurilemma. Furthermore, analogy with the differentiation of the neuroglia elements within the central nervous system points in the same direction. Finally, Carpenter was able to trace these cells through successive stages until they were, in structure and relations, demonstrably the neurilemma elements of the differentiated nerve. I reached similar conclusions ('03) for the derivation of the neurilemma of spinal somatic motor cells. The evidence appears to warrant the inference. Even Kolliker, who for years main


74 H. V. NEAL

tained that the neurileniina was of mesenchymatoiis origin, finally became convinced COo) of its ectodermal origin.

By tracing the histogenesis of the nerve through closely connected stages until adult conditions are established in Squalus it is possible to demonstrate that a large number, if not all, of the cells present in the anlage of the oculomotor become differentiated as neurilemma cells. Some of these stages are shown in the figures of plate 8 of this paper.

The cells, which are at first peripheral to the bundle of fibers (fig. 65), soon begin to penetrate among them (fig. 68) as stated by Vignal ('83), Gurwitsch ('00), Bardeen ('03), Carpenter ('06), and Gast ('09). Gast is probably correct in asserting the centrifugal migration of the group of cells clustered about the roots of the nerve anlage (figs. 64, 70) and their penetration into the bundle of fibers. In fact the penetration of the fiber bundle appears to occur largely in this manner, as stated by Gast.

Altogether the evidence goes to show that the neurilemma of the oculomotor has a similar derivation and differentiation to that of somatic motor spinal nerves.

h. What is the histogenesis of the neurilemma? Gast ('09, p. 428) summarizes the histogenesis of the neurilemma of the oculomotor as follows: In its proximal portion the oculomotor passes through four different developmental stages: (1) the naked processes of the central neuroblasts extend to the mesocephalic ganglion; (2) from the mesocephalic ganglion neurocj^tes (Kupffer '90) wander centrad along the fibrillar bundle; these lie upon and between the fibers and form about them, especiallj^ among the roots, a loose plasmatic network; (3) the neurocytes arrange themselves wholly on the surface of the nerve anlage, which at this stage consists of a central fibrillar bundle with a cellular sheath; (4) the neurocj'tes migrate back into the central fiber bundle and are transformed into sheath cells.

The results of a former ('98) and of the present study are in essential agreement with this description of Gast. The evidence derived from a study of Squalus embryos, however, seems hardly to justify the inference of a forward and backward migration of neurocytes along the nerve anlage. Moreover, it


MORPHOLOGY OF EYE MUSCLE NERVES /5

appears to favor the conclusion that the cell clusters at the roots of the nerve anlage are of mesenchjaiiatous and medullary derivation. The inference that cells derived from the mesocephalic ganglion migrate centrad as far as the roots of the nerve anlage does not appear to the writer to be well established. In agreement with Gast, however, I find that penetration of the fibrillar bundle begins at proximal and distal extremities of the nerve anlage, where it divides into roots and peripheral branches, and the penetration proceeds in opposite directions towards the middle stretch of the nerve. In the phenomena presented in the histogenesis of the neurilemma there appears to be no essential difference between that of the oculomotor and a spinal somatic motor nerve. In both cases the derivation of the neurilemma cells appears to be the same.

i. To what an extent do the emigrated medullary elements go to form the sympathetic? That a ciliary ganglion is associated with the oculomotor nerve in selachians seems proved by the consensus of opinion of investigators on the basis of anatomical, histological, and embryological evidence. Anatomical evidence of its presence has been presented by Schwalbe ('79), Jegorow ('86-87), and Allis ('02); on the basis of histological structure by Haller ('98); and on embryological grounds by VanWijhe ('82), Beard ('85), Ewart ('90^, Dohrn ('91), Miss Piatt ('91), Hoffman ('99), Gast ('09). Physiological evidence of a functional ciliary ganglion in selachians is wanting.

Furthermore, that the ciliary ganglion belongs morphologically and physiologically with the sympathetic system has been held by many investigators since its discovery in man by Schacher (1701) and the suggestion of its sympathetic character by Arnold ('31). Later, doubt as to its sympathetic nature was raised by Schwalbe ('79), who became an exponent of the view that it is a cerebro-spinal ganglion and who initiated the long controversy regarding its nature which has not yet been ended. Carpenter ( '06) has so admirably reviewed the literature dealing with this problem that it appears unnecessary to enter into an extended discussion of the various arguments for and against its cerebro-spinal character. That the ciliary ganglion is, in part


76 H. V. NEAL

at least, sympathetic in all forms from fishes to man seems, on the basis of anatomical, histological, embryologidal and physiological evidence, indisputable. But the presence of bipolar ganglion cells, like those of cerebro-spinal ganglia, in the ciliary ganglion of many vertebrates has led to the general acceptance of the view of its double nature, a view first advanced by Krause ('82).

But the presence of bipolar ganglion cells within the ciliary ganglion by no means demonstrates its morphological comparability with cerebro-spinal ganglia, since ontogenetic evidence shows that the entire ganglion — whether derived from the midbrain or the mesocephalic ganglion — has not a genetic but merely a secondary relation to the anlage of the oculomotor, a fact emphasized by Gast ('09).

While Dohrn ('91) claimed an exclusively medullary derivation of the ciliary ganglion in Selachii, Gast ( '09) , working upon the same material — in fact the same sections — concludes that they are exclusively derived from the mesocephalic ganglion. These diametrically opposite conclusions indicate how obscure the phenomena of migration are in sections of embryonic material. Carpenter ( '06) , working upon chick embryos, infers a double derivation of the cells of the ciliary ganglion.

Von Kupffer ('95), Johnston ('05), and Belogolowy ('10b) have attempted to associate the ciliary ganglion genetically with the 'thalamic nerve. ' Johnston says (p. 244) that since the profundus ganglion is distinct from the ciliary and is formed from a different part of the neural crest, it seems altogether probable that the ciliary ganglion permanently represents the N. thalamicus. " Belogolowy 's discovery of an anastomosis between the oculomotor nerve and the transient ganglion of the ' thalamic nerve ' likewise seems to him to prove the homodynamy of the latter with the ciliary ganglion of vertebrates. In drawing this conclusion Belogolowy fails to take into account the fact that the ciliary ganglion of vertebrates is a derivative of the mesocephalic or profundus ganglion. It is not at all clear that the complicated anastomoses between the branches of the eye-muscle nerves in reptile embryos and their peculiar relations with branches of the trigeminal nerve


MORPHOLOGY OF EYE MUSCLE NERVES 77

throw any light upon the primitive relations of these nerves in vertebrates. Johnston's inference is fallacious since his premise is incorrect. The cells of the ciliary ganglion of Squalus are proliferated from the medial surface of the profundus ganglion. It may be granted, however, that the demonstration that the ' thalamic nerve ' represents the primitive root of the profundus nerve, would associate this so-called nerve with the ciliary ganglion as its sympathetic visceral sensory component.

Sections of Squalus embryos give convincing evidence of cellular migration, both from the midbrain and from the mesocephalic ganglion, into or towards the oculomotor anlage, but it is impossible to feel convinced that the migration of the medullary cells extends as far as the ciliary ganglion. In fact the evidence of the participation of medullary cells in the formation of the sympathetic is as doubtful in the case of the oculomotor as in the case of spinal somatic motor nerves. The possibility of such a derivation can not be denied, but positive demonstration on the basis of evidence obtained from serial sections has not yet been given. In this respect, that is, in the derivation of its proper sympathetic ganglion, the oculomotor resembles a spinal somatic motor nerve.

j. Summary of the histogenesis of the oculomotor and of the ciliary ganglion. In all essential respects, the oculomotor nerve resembles in its histogenesis a spinal somatic motor nerve, i.e., in its appearance as a product of the protoplasmic movement of medullary neuroblasts situated in the somatic motor column of the neural tube ; in its secondary connection with a mesodermic somite. Van Wijhe's first; in the growth and extension of its fibers by the continuous movement of the protoplasm of medullary neuroblasts; in the differentiation of its neurofibrils within the protoplasm of these processes; in the migration of medullary cells into the nerve anlage to form — at least in part — the neurilemma; in the probability of the participation of mesenchymatous cells in the formation of the neurilemma; in its union with cells of a cerebro spinal ganglion ; and finally in its association with a sympathetic ganglion, the ciliary, derived mainly if not exclusively, from a cerebro-spinal ganglion as a result of cellular migration.


78 H. V. NEAL

3. Does the histogenesis of the trochlearis resemble that of a somatic motor spinal nerve?

a. Are protoplasmic connections between myotome 2 and hindbrain primary or secondary? The distance between the point of emergence of the trochlearis fibers and the myotome which this nerve innervates is, in Squalus embryos, the greatest traversed by any motor nerve. This fact lends especial interest to the development of this nerve in relation to the problem of the existence of primary paths or connections between nerve center and the end-organ. This nerve would seem to afford a critical test of the truth or falsity of the Hensen-Held hypothesis of nerve histogenesis. If any nerve in the vertebrate body needs a primary path so that it may not go astray, that nerve is the trochlear. In view of this fact it is somewhat surprising that Paton ('07), who supports this hypothesis, concedes that the trochlearis (and also the oculomotor) acquires its connection with the superior oblique muscle without the participation of plasmodesmatous bridges (Paton '07, p. 556).

Some years ago ('98, p. 237) I expressed the opinion that the possibility of a primary connection between muscle and nerve appears excluded in the case of the superior oblique muscle and the trochlear nerve" and Dohrn, later ('07, p. 410), said quite as emphatically that "Jede Moglichkeit eines uranfanglichkeit Zusammenshangs zwischen Muskelzelle und Nervenzelle ist dabei ausgeschlossen. " Belogolowy ('10a, p. 380) also agrees in affirming the absolute impossibility of a primary connection between nerve and muscle in the case of the trochlearis. While Sewertzoff ('98) and Fiirbringer ('02) are unable to agree with this assertion of mine, they advance no arguments in its rebuttal. It is a significant fact that none of the supporters of the Hensen hypothesis have been able to demonstrate a primary protoplasmic path connecting the isthmus or the hindbrain with the myotome of Van Wijhe 's second somite.

On the contrary, and in support of the view that the trochlear nerve develops in accordance with the process theory of neurogenesis and that the connection of hindbrain and superior oblique


MORPHOLOGY OF EYE MUSCLE NERVES 79

muscle is secondary, Dohrn ('07, p. 416) advances the following argument :

The development of the trochlearis forms a crux for the HeiisenGegenbaur-Furbringer hj^jothesis of the primary connection of the nerve and its terminal organ. So long as we have to do Avith the connections of spinal somatic nerves with somites already formed, one may believe in the existence of plasma bridges and be willing to recognize in them at least the paths that guide the growing nerve. The separation in this case between tube and somite is the least possible and few cells lie between them. But if one extend this view to the eye muscles he at once meets insuperable difficulties. All these have a considerable distance to traverse in reaching their end-organ, and the abducens, oculomotorious and trochlearis solve the problem each in its own way. I have already pointed out ('01) that this theory must be tested, not on the spinal somatic nerves, but on the eye muscle and splanchnic motor nerves, if it is to be regarded as tenable, and I must now put the question to the defenders of the Hensen theory how they conceive of the connection of the trochlearis through plasma bridges "with its terminal organ in the other antimere.

In embryos of 12 to 18 mm. the actual trochlearis fibers remain either completely within the neuroblasts from which they are to arise or they have at most made their way through the medullary wall to the chiasma, but are not yet protruded from the wall. The superior oblique muscle, on the other hand, has developed from the mandibular somite as the most dorsal of the eye muscles and has already begun to differentiate muscle fibers. What and how many plasmodesmatous connections must be assumed in order to establish the most complicated of all paths traversed by any motor nerve? How is one to demonstrate that first plasmodesms of the neural tube, then others of the mesenchyma, then — in case ' Kettenf asern ' are present before the trochlearis fibers extend through their region of distal distribution — ■ plasmodesms of the 'Kettenfasern,' and finally those of the muscle fibers of the superior oblique muscle are likewise prepared to furnish neurofibrils through further differentiation? Would this be in the least more easily conceivable than the outgrowth of the nerve from its neuroblasts to its terminal organ? It is true that the outgrowth points to problematical forces, but shall we be mthout a riddle if we put another in its stead — ^one that presents a number of unfounded assumptions?

Held ('06), who admits the insufficiency of the hypothesis of plasmodesmatous paths for the trochlearis nerve, thinks that he may assume as the direct path-determining impulse for the outgrowing neurofibrils the principle of the axial position of the neuroblast and that of the shortest distance. These obviously fail


80 H. V. NEAL

in their application to the growth of the trochlearis. The chief axes of the neuroblasts which form the trochlearis fibers do not point toward the end-organ, neither do they follow the shortest possible course, but practically the most indirect route possible.

In Squalus embryos, previous to the first appearance of the trochlear anlage, the region of its future path is filled with a loose mesenchyma derived at least in part from the neural crest. It is a most important fact, demonstrated by the observations of Kastschenko ('88), Neal ('98) and Dohrn ('07), that, from the time of its first appearance to the time it attains connections with the superior oblique muscle, the trochlear anlage in some selachians shows no close associations with mesenchymatous cells, but extends as a loose bundle of fibers through vacuolar spaces (fig. 49). The trochlearis development, says Dohrn ('07, p. 413) "affords a proof that an elongated motor nerve can grow through a considerable extent of tissue without becoming attached to a single cell of the surrounding mesoderm." Evidence of a granular protoplasmic envelope surrounding the fibers, however, is not lacking, but there is no evidence that surrounding mesenchj^matous cells contribute to the production of this. The descriptions of Froriep ('91) and Miss Piatt ('91) relate to more advanced stages in the histogenesis of the trochlear than those described, so that their assertions of the participation of cells in the formation of the trochlearis anlage have no relevancy in connection with the problem of the formation of the primary connection between tube and myotome.

b. What cells participate in the formation of these connections? The fact that in the adults of all classes of vertebrates the fibers of the trochlearis may be traced in suitable preparations to a nidulus in the somatic motor column, posterior to the midbrain leads to the supposition that these cells are genetically related to the fibers. The further evidence that during the period of extension of the fibers toward their terminations in the superior oblique muscle, no cells which may be regarded as neuroblasts or 'nerve-cells' in the Apathy sense, may be seen in connection with the peripheral stretch of nerve anlage, points in the same direction. The connections with cell-clusters or fragments of


MORPHOLOGY OF EYE MUSCLE NERVES 81

the neural crest as described by Hoffman ('89), Oppel ('90), Dohrn ('91), Miss Piatt ('91), and Froriep ('91) are secondary and do not appear in the primary fibrillar stretch of the nerve anlage. Filatoff ('07, p. 358) agrees with the writer ('98) that the so-called primary trochlearis (Miss Piatt '91) has no genetic relation with the definitive trochlearis, which develops wholly independently of the neural crest and of the trigeminal nerve, although it requires secondary connections with the latter. According to Belogolowy ( '10a) no ganglion is associated with the trochlear in bird embryos, while in the turtle the so-called ganglion of the trochlearis has no connection with the trochlearis but is connected with the superficial branch of the trigeminal (p. 424). The same is true also of the relations of the ganglion in the sturgeon. The superficialis nerve is, therefore, to be regarded as the dorsal nerve of the trochlearis metamere.

Schwalbe ('79, '81), Johnson ('05), and Carpenter ('06) have asserted that the neuroblasts which form the nidulus of the trochlear are situated in the somatic motor column of the midbrain and not in the hindbrain. Such divergent opinion may possibly be explained by the fact that in the ventral portion of the brain in this region no constriction is present such as that which divides midbrain and hindbrain regions dorsally, so that it is difficult to distinguish the limit of these two brain divisions. The majority of investigators find the nidulus of the trochlear, either in the hindbrain or in the isthmus. The assertion of Martin ('90), that the nidulus of the trochlear is primarily in the visceral (lateral) column, and that it secondarily migrates mediad, has not been confirmed and, like many of the other amazing assertions of this writer, should be placed in quarantine as suggested by Dohrn ('07). In Squalus the nidulus of the trochlear is clearly ventral and, while the fibrillar bundle of the nerve emerges dorsally from the brain wall in the region of constriction between midbrain and cerebellum, the central fiber-tract with which it is connected may be traced posteriad and ventrad to the nidulus in the somatic motor column in the anterior portion of the cerebellar neuromere (neuromere iii). On the other hand. Van Walkenburg ('10) finds two trochlear niduli in human embryos.

JOURNAL OF MORPHOLOGY, VOL. 25, NO. 1


82 H. ■ V. NEAL

Belogolowy ('10a, p. 367-8) describes the histogenesis of the trochlear, which may be summarized as follows : The first developmental stages of the nerve take place in the brain. Here the fibers of the brain are gradually differentiated as processes of cells in the somatic motor column of the hindbrain. The fibers grow in the direction of the dorsal side of the brain and toward the constriction between second and third brain vesicle, that is toward the basis of the cerebellum anlage. The fiber bundles may be easily traced through the brain wall and usually number three.

At the time when these bundles have reached a point about two-thirds the height of the neural tube a nidulus of large ganglion cells makes its appearance at the place of the future point of emergence of the fibers. The evidence of their ganglionic character consists in their larger size and in the outgrowth of fibers from them. (Belogolowy 's figures, however, do not bear out the latter assertion.) This dorsal nidulus is gradually enclosed by the spongiosa and penetrated by the coarse fibers of the trochlear until only isolated cells remain scattered in the region of the chiasma. The bundle of fibers of the nerve anlage is at first free of accompanying cells, which appear first at about the level of the middle of the brain wall.

Belogolowy then goes on to present evidence that the accompanying cells can have nothing to do with the genesis of the fibers of the trochlear. Notwithstanding the divergence in his account in regard to the existence of a dorsal nidulus, Belogolowy does not hesitate to compare the trochlear with typical somatic motor nerves. He agrees with the conclusion that connection between hindbrain and superior oblique muscle is secondary, an inference which — in the light of the evidence now in our possession — there seems no good reason to doubt.

c. Have these protoplasmic connections a genetic relation to the neurofibrilsf As first stated by Kastschenko ('88, p. 465) the trochlearis, from the time of its first appearance, is a fibrillar structure. In preparations of Squalus embryos by the VomRath method these fibers are deeply-staining and highly refractive threads, contrasting strongly with the protoplasm of the ad


MORPHOLOGY OF EYE MUSCLE NERVES 83

jacent mesenchymatous cells between which they lie. In cross sections of the nerve anlage (fig. 52, p. 181) the fibers appear as black dots or granules scattered in a granular and vacuolated protoplasmic envelope. Their appearance is essentially the same in VomRath preparations as in preparations by the Cajal or the Bielchowsky-Paton methods. The identity of this bundle of fibrils with the trochlearis in advanced stages of differentiation is established, not only by its fibrillar structure, but by its relations with tube and myotome at this and all subsequent stages of development. As this anlage has been traced in unbroken continuity until the differentiation corresponds essentially to that of the adult nerve there is no good reason to doubt the identity of the fibers of the anlage with those of the adult and that the neurofibrillae make their appearance within them. It is true that the splitting of the fibers of the nerve anlage into the neurofibrils of the fully differentiated nerve is a matter of inference rather than one of direct observation, yet students of nerve histogenesis have not hesitated to make this inference which has been drawn by Bardeen ('03), Carpenter ('06) and Paton ('07), on the ground that, while the primitive fibrils of the nerve anlage are relatively coarse structures, the neurofibrillae of later stages are much finer and more numerous. There is no reason for inferring that the neurofibrils have any other origin than the coarser fibrils of the nerve anlage, which make their appearance in the nerve from its first inception in the stages when the nerve anlage is wholly free from cells. Furthermore, at no time in the histogenesis of the nerve is there any indication of a genetic relation of the cells, which later make their appearance in the nerve, to the fibers. No neuroblasts have been discovered within the nerve anlage. So that there seems to be no reason why the histogenesis of the neurofibrils of the trochlearis may not be regarded as in every respect similar to that of the neurofibrils of spinal somatic motor nerves. In other words they are to be regarded as the product of differentiation within the neuraxon processes of medullary neuroblasts situated in the somatic motor column of the neural tube.

Miss Piatt ('91a, p. 96) inferred the neuroblastic origin of the fibers of the proximal portion of the trochlearis on the basis of


84 H. V. NEAL

the evidence that "the fibrous root of the nerve cannot be composed of prolongations from the distal cells (derived from the R. ophth, sup. V), because the fibrous or proximal part of the trochlearis arises before the distal or cellular part; further, the fibrous root is the thickest as it comes from the brain, becoming gradually attenuated as it proceeds into the mesoderm. " Belogolowy ('10a) likewise finds the trochlear of the chick fibrillar from the time of its first appearance. There appears no good reason to doubt that the trochlear anlage is not only 'fibrillar' but 'neurofibrillar' in the Apathy sense of the word.

d. Is a neuraxon of the trochlear anlage multicellular in origin or the process of a single 7nedullary cell? While it is practically impossible for any particular neuraxon of any nerve to give an answer to this question upon the basis of direct observation (and for the trochlear nerve the difficulty is increased by the extended intra-medullary course of the fibers of the nerve) it is easy to disprove the assertion that the trochlearis anlage in its peripheral stretch is the product of the fusion of a chain of 'nerve' cells, since in Squalus, as in Pristiurus (Dohrn '07), the nerve anlage may be traced as a bundle of fibers, devoid of cells or nuclei, from its point of emergence from the brain to its terminal organ. If such evidence could convert Dohrn ('07) from the advocacy of the cell-chain hypothesis of neurogenesis which on the basis of evidence partly derived from the study of the same nerve he had defended for years, it would seem that little argument would be needed to convince unprejudiced students that the trochlear nerve develops by the free outgrowth of medullary neuroblasts. It appears unlikely that any student of neurogenesis will demand that a neuraxon be traced from neuroblast to muscle fiber in order to convince him of the applicability of the process theory to the phenomena of histogenesis of the trochlear. So that, while in the case of the trochlear nerve it is impossible to deal with individual cells and fibers, the fact that, from the time of its first appearance as a fiber bundle emerging from the dorsal wall of the brain, these fibers may be traced centrally to their connection with a group of deeply stained neuroblasts lying in the somatic motor column of the hindbrain, and that these are also the adult re


MORPHOLOGY OF EYE MUSCLE NERVES 85

lations, will seem to most students sufficient evidence to demonstrate their neuroblastic derivation.

The evidence that in early stages of histogenesis the distal fibers of the trochlear are associated with cell clusters or ganglia, as shown by Dohrn and others, affords, as stated above, no support to the cell-chain hypothesis, since these connections with the trochlearis fibers are secondary. Their relations with the trochlear anlage therefore must receive another than a genetic interpretation. Such another interpretation is given below. • e. By what means does the increase in length of the neuraxon take place? The histogenesis of the trochlearis affords no evidence in support of the Hensen-Held contention of the differentiation in situ of plasmodesmatous paths in the formation of a nerve anlage. The loose distal brush of fibers (fig. 49), in which the nerve anlage terminates in early stages, show no such relations to cells or cell processes as Held's position would require. Furthermore, while the existence of a loose brush of fibers may readily be interpreted as an advantage to a bundle of neuraxons which effect connection with their terminal organ by a process of free outgrowth, there would appear to be no meaning to the phenomenon were the Hensen-Held hypothesis correct. As soon as the connection with the myotome is effected by the fibers, they quickly unite into a compact bundle and lose their brush-like form. Therefore, while it is difficult to demonstrate the freegrowing terminations of the fibers within the mesenchyma, the positive evidence, as far as it is known, is in harmony with the process theory. The assertion that the trochlear is differentiated in situ through the use and incorporation of primary plasmodesmatous paths is wholly unfounded.

/. What is the source of the cells of the trochlearis? Somatic motor spinal nerves derive their cells from the neural tube by migration along the nerve anlage, possibly from the mesenchyma, and come into close association with cells masses derived from the spinal ganglia. Do the cells associated with the trochlearis anlage have a similar derivation?

That medullary elements enter the trochlear anlage has been held by Dohrn ( '07) on the basis of evidence similar to that pre


86 H. V. NEAL

sented in figure 53, which shows nuclei half in and half out of the brain wall at the point of emergence of the trochlearis root. Dohrn (p. 293) also called attention to the fact that the nuclei are not peripheral in position but lie between the fibers of the nerve root. Furthermore, medullary nuclei lie nearer the periphery of the brain wall in the region where the fibers of the trochlear enter the brain than they do elsewhere in the vicinity of the root, and, as development goes on, an increasing number of nuclei appear among the fibers of the nerve root. More convincing evidence of migration would be difficult to find in sectioned material.

Belogolowy ('10a, p. 375) suggests that the absence of accompanying cells in the fiber bundle of the trochlear anlage in its intracerebral and in its proximal extent proves that the accompanying cells are not medullary in origin. However, had Belogolowy studied somewhat more advanced stages with this problem in mind, it is posssible that he would have found evidence of medullary migration. The writer ('98), on grounds similar to those advanced by Belogolowy, concluded that medullary elements were absent from the trochlear anlage of Squalus, but renewed investigation of more advanced stages proves this inference to be erroneous.

On the other hand, evidence that mesenchymatous cells attach themselves to the nerve anlage is diflficult to obtain and positive proof is altogether wanting. Belogolowy ('10a) infers the participation of mesenchymatous cells in the formation of the neurilemma of the trochlear. The most convincing evidence that may be secured is possibly that mesenchymatous cells lie in various degrees of proximity to the nerve anlage (fig. 51). In other words the evidence of the participation of the mesenchyma in forming the cellular elements of the nerve anlage are quite as obscure as in the case of spinal somatic motor nerves.

Miss Piatt ('91) was the first to observe evidence of migration of cells from the anlage of the superficial branch of the trigeminal into or towards the trochlear anlage. The evidence is similar to that seen in the relations of the oculomotor to the profundus ganglion. Miss Piatt, after describing (p. 95) the first appearance of the trochlearis as a "small fibrous nerve growing


MORPHOLOGY OF EYE MUSCLE NERVES 87

from the constriction between midbrain and hindbrain, " says that soon after the appearance of this small nerve, which is the root of the permanent trochlearis, cells are proliferated to meet it from the ganglion cells that lie above the superior oblique muscle." These cells cannot rise from the brain, since "no cells are found in the root of attachment." Thus the permanent trochlearis arises from two sources from the brain and from the ganglion cells."

Dohrn ('07) also inferred that the first cells connected with .the trochlear anlage in Pristiurus are derived from the ramus superficialis V, while in those forms in which, as in the Torpedinidae, the superficial branch is not differentiated, but is represented by fragments of the neural crest, the trochlearis anlage acquires relations to these cell clusters, similar to those in Squalus and Pristiurus, and presents similar evidence 'of cellular migration toward the nerve anlage.

The evidence of cell migration from the R. superficialis V into the trochlear anlage in Squalus consists of the fact that cells appear primarily, not in the proximal root of the nerve nor in the bundle of fibers, but in the distal portion of the nerve in the region where its fibers cross those of the superficialis. The phenomena presented are practically identical with those described for the oculomotor in its cellular relations with the R. prof. Trig. As development proceeds, cells increase in number in the region between the two nerve anlagen, although the proximal portion of the nerve remains free from cells. Finally, cells accumulate between the trochlear and superficialis anlagen in a mass precisely comparable to the anlage of the ciliary ganglion, forming the cell cluster to which Miss Piatt ('91, p. 100) refers as the ganglion of the trochlearis lying above the superior oblique muscle anlage.

Further evidence of migration into the trochlear anlage from the superficiahs nerve is found in spindle-shaped cells closely applied to the fibers of the trochlear anlage in stages when cells are wanting in the proximal portion of the nerve and when evidence of additions from the mesenchyma is lacking. Similar spindleshaped cells show connections, on the one hand with the ramus


88 H. V. NEAL

superficialis, and on the other with the fibers of the trochlear anlage.

Thus, in the derivation of its cellular elements, the trochlear resembles the oculomotor and spinal somatic motor nerves. There is no reason to doubt the medullary derivation of some of its cells; mesenchymatous participation is somewhat doubtful; and, finally, the migration of cells from a ganglionic nerve in relation with its anlage appears unquestionable.

g. What is the fate of these cells? Dohrn ('07) makes the interesting suggestion that the manner of development of the fibers of the trochlearis in Pristiurus is a natural experiment, confirming Harrison's ('04) experimental demonstration of the origin of the neurilemma in amphibian somatic motor nerves from the neural crest. That is, he concludes that the cells which migrate from the superficialis into the trochlear anlage form the cells of the neurilemma. The possibility that they may be compared with the cells of sympathetic anlagen does not seem to have occurred to Dohrn. Miss Piatt ('91) speaks of these cells as ganglion cells although she has made no attempt to follow their fate in later stages.

However, the similarity of the cell-cluster, associated with the trochlearis anlage and derived from the superficialis, with the ciliary anlage in its relations to the oculomotor and the profundus is so striking that the suggestion that in this mass we have a sympathetic ganglion seems permissible. The main objection to such a view, and one that may appear insuperable, is that the trochlear appears in the adult to be associated with no sympathetic ganglion. The ganglion of the trochlear anlage — not to be confused with the fragments of the neural crest associated w^th the trochlear anlage in the Torpedinidae and reptiles — is conspicuous in Squalus embryos of 25 mm. but has disappeared as a mass of cells in 45 mm. embryos. Since, however, many of the sympathetic anlagen of spinal somatic motor nerves similarly disappear in these stages and do not appear hi the adult, the disappearance as a distinct cell-mass of the hypothetical sympathetic ganglion of the trochlear is not to be regarded as a serious objection to the hypothesis. Moreover, the possibility


MORPHOLOGY OF EYE MUSCLE NERVES 89

is by no means excluded that sympathetic cells are associated with the adult trochlear nerve. This point needs investigation. The important fact remains that a cell cluster, associated with the embryonic trochlearis, has the same relations and derivation as the sympathetic anlagen of the trunk and as the ganglion of the oculomotor and that in this respect as in others the nerve conforms in its histogenesis with that of a typical somatic motor nerve.

h. What is the histogenesis of the neurilemma of the trochlea!'? More convincingly than any other nerve in the vertebrate body the histogenesis of the trochlear anlage shows that a nerve may become connected with its terminal organ without the slightest indication of the participation of a single peripheral 'nerve cell' or chain of nerve-forming cells. Dohrn ('07, p. 411) states that it was the study of the histogenesis of the trochlearis which converted him from a supporter of the cell-chain hypothesis to one of the process theory of His. There appears in Pristiurus (Dohrn) and in Squalus the decisive evidence which, in normal embryos, effectually invalidates the cell-chain hypothesis in whatever modified form it may present itself. The evidence is wholly convincing that the neurilemma cells of Scjualus have a secondary and exogenous derivation. Primarily, instead of a cellchain extending from nerve center to end-organ the trochlear nerve appears as a fibrillar bundle. Into this fibrillar anlage cells migrate from the neural tube and from the ramus superficialis V. Some cellular additions may come from the adjacent mesenchyma. These cells, generically independent of the fibers, soon assume an elongated form and a relation to the nerve fibers which marks their destination as neurilemma cells.

i. To luhat an extent do the emigrant cells go to form the sympathetic? The absence of any positive evidence of sympathetic cells in the adult trochlearis would appear to afford presumptive evidence against such a fate for any of the cells of the trochlear anlage. But the fact that sympathetic anlagen disappear ontogenetically in many spinal sor^iatic motor nerves makes the supposition that the trochlear possesses a transient sympathetic ganghon seem not unreasonable. Moreover, until it has been


90 H. V. NEAL

demonstrated that sympathetic cells are wanting in the adult trochlear, the possibility that the cell cluster associated with the nerve in 25 mm. embryos forms a diffuse sympathetic along the nerve in the adult is not excluded. As far as it goes, the evidence favors the opinion that a transient sympathetic ganglion, derived from the ramus ophthalmicus superficialis trigemini is associated with the trochlear anlage as in the case of certain anterior spinal nerves. There is no evidence whatever that migrant medullary elements enter into this hypothetical transient sympathetic anlage.

j. Summary of the histogenesis of the trochlear nerve. In the manner of its growth and extension to form a secondary connection with its myotome through the protoplasmic movement of medullary neuroblasts; in the source of derivation of its neurilemma; and in its relation to a ganglionic nerve and to a sympathetic anlage, the histogenesis of the trochlear resembles in all essentials that of spinal somatic motor nerves. To this extent its serial homology with them is demonstrated.

4. Does the histogenesis of the abducens resemble that of a somatic

motor spinal nerve?

a. Are protoplasmic connections betiveen hindbrain and myotome 3 primary or secondary? No investigator has attempted to support the Hensen hypothesis of the primary connection of nerve and muscle on the basis of the histogenesis of the abducens nerve. On the contrary all who have made a careful study of its histogenesis— Dohrn ('90, '91), Neal ('98), Carpenter ('06), and Belogolowy ('10) — agree that, at first, the abducens is not connected with the posterior rectus muscle, but in its early stages of development terminates freely in the mesenchj-ma at the base of the brain. No evidence of the participation of plasmodesmatous paths in the forward extension of the nerve has been advanced. And, while no thorough investigation of its development from the standpoint of the Hensen hypothesis has been made, it does not seem likely that demonstrable plasmodesms would ha\'e remained unperceived.


MORPHOLOGY OF EYE MUSCLE NERVES 91

The mere presence of a loose mesenchyma in the region later traversed by the nerve anlage gives no presumption in favor of view of Hensen and Held. The possibility of the utilization of such material in the elongation of the growing nerve may readily be granted, but as yet no demonstration of such use of protoplasmic bridges has been made. On the contrar}^, it is possible to demonstrate in sections the free growing end of the nerve anlage and to ascertain, on the basis of actual observation, that no plasmodesm or undifferentiated protoplasmic path connects the end of the nerve anlage with the myotome. Such a growing end of the abducens is given in fig. 35, which shows on what slight evidence the hypothesis of the utilization of protoplasmic paths depends. The amoeboid terminations of the nerve fibers (neuraxons) show attenuated connections by means of fine granular threads — as seen in sections — with amoeboid processes of mesenchymatous cells. But such evidence, so far as it permits any inferences at all, favors the opinion that the fine amoeboid processes of the growing tip are derivatives of the nerve anlage itself. For, if the processes of the mesenchymatous cells are genetically related to them, there is equally good ground for thinking that the processes which extend from the end of the nerve anlage in all directions are derivatives of the nerve itself. Harrison's experimental results verify this assumption, since, in his preparations of the living nerve fibers, similar delicate extensions of the amoeboid termination of the nerve make their appearance. In such preparations there can be no question of the genetic relations of the fine threads. To call them paths as Held ('09) has done is obviously a misnomer for structures which radiate in all directions. Were a growing structure to depend upon such flimsy paths for the material for its growth, its extension would be neither fast nor far. Furthermore, neither this section nor others, in which the observer may feel confident that he is dealing with the actual termination of the growing nerve anlage, is the relation of the termination to adjacent mesenchymatous cells such as to suggest the utilization of their substance in the elongation of the nerve.


92 H. V. NEAL


In view of the unanimous agreement of all investigators of the histogenesis of the abducens that the nerve becomes secondarily connected with its myotome, and in view of the entire absence of evidence of primary plasmatic paths within which the neuraxon might be differentiated; furthermore, in view of the fact that it is possible to demonstrate the growing tip of the abducens anlage as an amoeboid structure similar to those seen in Harrison's cover-glass preparations, there appears little reason to doubt that the abducens acquires a secondary connection with its myotome. This inference is strengthened by Belogolowy's discovery that some fibers of the abducens anlage of the chick extend in their growth farther than the anlage of the posterior rectus muscle and later atrophy. It seems easier to harmonize such evidence with the hypothesis of the free outgrowth of nerve fibers than with the hypothesis of primary connection of nerve and muscle. Moreover, the fact that the nidulus of the abducens is about equidistant from post-otic and pre-otic myotomes and that, in early stages of development, its fibers grow in both directions — that is, both anteriad and posteriad — suggests that the direction of growth is chemotropically determined. The extension posteriad is, however, transient, the posterior process is soon withdrawn, and the protoplasmic movement anteriad alone continues.

h. What cells participate in the formation of the protoplasmic connections? If the possibility of the development of the abducens by the progressive differentiation of plasmodesmatous paths be excluded, the possibility remains that the abducens develops, either by the differentiation of a chain of cells, or as the product of protoplasmic movement of medullary neuroblasts. Dohrn ('90 a, '91) alone, of all of the students of the histogenesis of this nerve, has attempted to demonstrate the former mode of histogenesis. The fact, however, that later ('07) be admitted the inadequacy of the grounds upon which that opinion was based makes it unnecessary to call attention to the fallacy of his argument. Dohrn's premises were weak, not because he was a poor observer — no man has done more to enrich our knowledge of the histogenesis of nerves in selachians — but because his


MORPHOLOGY OF EYE MUSCLE NERVES 93

preparations were not suited to demonstrate the histogenesis of the neurofibrillae.

Filatoff ('07, p. 343) finds that the anlage of the abducens in reptile embryos appears primarily as a strand of mesenchymatous cells extending from the posterior rectus muscle towards the base of the medulla and as yet unconnected with the brain. The cellular strand is not fibrillar. How connection with the brain is effected, Filatoff is not able to state.

All other students of the histogenesis of the abducens agree that the cells of the abducens anlage have no genetic relation to it but are secondary and accessory. The fibers of the anlage arise as processes of medullary neuroblasts situated in the somatic motor column of the hindbrain, posterior to the otic vesicle. All have noted the greatly elongated nidulus, which, according to the writer ('98) extends through the first and second post-otic neuromeres (neuromeres VII and VIII). Belogolowy ('10 a, p. 270) makes the interesting discovery that in a 2.2 mm. embryo chick the roots of the abducens extend through as manj* as five neuromeres." Later they arrange themselves into three groups, which however show no regular relation to the neuromeres. By a comparison of the relations of the abducens anlage to the neuromeres, Belogolowy concludes that the most anterior roots of the abducens of the chick take their origin from a neuromere next anterior to the one from which they originate in the dogfish (Squalus). Posteriorly the roots of the embryonic abducens in the chick are connected with those of the hypoglossus. . The posterior roots of the abducens show a marked tendency to grow backwards as if to join those of the hypoglossus which lie immediately behind them. Still later some of the roots degenerate. Another important discovery made by Belogolowy is that of the existence of a transient somatic motor nerve, ventral to the ramus maxillaris trigemini and uniting peripherally with the oculomotor with which the abducens also unites. The inference of the medullary derivation of the neuraxons of the abducens is based on the evidence of the connection between deeply staining cells of the medulla and the fibers of the nerve anlage, from the time of the first appearance of the


94 ' H. V. NEAL

anlage (plate 8). The earliest stage represented (fig. 31) shows that the anlage, which already has two roots of origin, contains deeply staining fibers surrounded by an envelope of granular protoplasm. Mesenchymatous cells crowd thickly around the roots of the anlage so that it is impossible to distinguish their boundaries from the protoplasmic envelope. In this and later stages (figs. 32, 41, 37) the continuity of the processes of the medullary cells with the fibers of the anlage is unmistakable.

In later stages the number of roots increases to the number of five or six and there is a correlated increase in the number of medullary cells, the processes of which may be traced towards or into the roots of the anlage. In the meanwhile cells make their appearance in the nerve anlage (figs. 33, 36), but the comparison of closely connected stages indicates that these cells secondarily attach themselves to the anlage or, in advanced stages, migrate from the neural tube (fig. 39). None of them ever has the form or staining properties of neuroblastic cells, such as appear within the ganglionic nerves or within the wall of the medulla. Also, during the stages during which the fibers most rapidly increase in number these cells are distinctly peripheral in relation to the bundle of fibers. Furthermore, the fact that the fibers in successive stages grow centrifugally toward the myotome of somite 3, and that, in the earlier stages of growth, the number of fibers is greater in the proximal portion of the anlage than in the distal portion (figs. 42-44), accords with the supposition that the fibers have a medullary origin in central neuroblasts.

c. Have these protoplasmic connections a genetic relatio7i to the neurofibrils? Students of the histogenesis of the abducens have paid little or no attention to the differentiation of the neurofibrils. There seems, however, little reason to doubt the genetic connection between the deeply stained fibers of the nerve anlage and the finer fibrils of the fully differentiated nerve. Their most important histogenetic change appears to consist in the splitting of the fibers into finer fibrils as the diameter of the fiber increases during growth. The granular envelope, so conspicuous in the anlage, gradually disappears and is replaced by


MORPHOLOGY OF EYE MUSCLE NERVES 95

the neurilemma cells as they penetrate among the fibers in advanced stages of histogenesis.

There appears to be no reason for doubting the identity of the fibers differentiated by the Vom Rath method with those which appear in Cajal and Bielschowsky-Paton preparations. They have practically identical appearance, form and relations, both in cross and longitudinal sections. So that if it be granted that the coarse fibrils seen in preparations made by the latter methods are genetically related to the neurofibrils of the adult nerve, the reasons apply with equal force to those which appear in Vom Rath preparations.

d. Is the individual neuraxon of the abduceris anlage multicellular in origin or is it the process of a single cell? Reasons have aheady been stated why the abducens fibers may be regarded as products of protoplasmic movement and not of cell chains, and it seems unnecessary to repeat them. All the evidence favoring this conclusion for spinal somatic motor nerves and for the oculomotor and trochlearis may be advanced for the abducens. ,The proof is even more convincing in the case of the abducens, since development is simplified by the absence of connection with a ganglionated nerve.

e. By what means does the increase in length of the constituent neuraxons take place? This question may be answered in the same way as in the case of the other somatic motor ne^-ves described, and has already been stated above. The evidence upon which the hypothesis of the plasmodesmatous origin of the nerve anlage is based is no less equivocal and unconvincing than that offered by its most recent exponent. Held ('09).

/. What is the source of origin of the cells of the abducens anlage? Since the abducens makes no connection with a ganglionic nerve and therefore receives no cells from that source, the cells of the anlage must either be derived from the mesenchyma or from the medulla or from both. Neal ('98) and Belogolowy ('10) were unconvinced of the medullary origin of the cells, which both derived from the mesenchyma. But Dohrn ('91) was undoubtedly correct in inferring the medullary origin of some of the abducens cells. The evidence of migration is quite as convinc


96 H. V. NEAL

ing as that seen in spinal somatic motor nerves. A section through a root of the abducens is shown in figure 39 and the appearances are seen to resemble closely those of early stages of the development of ventral spinal roots. The evidence that mesenchymatous cells attach themselves to the anlage is equivocal but seems probable. Filatoff' s ('07) observation of a migration of mesenchymatous cells from the anlage of the posterior rectus muscle is interpreted by the present writer as subsequent to the acquisition of connection between brain and myotome and as a part of the process of formation of the neurilemma. Filatoff advances no evidence that these cells have a genetic relation with the fibers of the definitive nerve.

g. What is the fate of the cells of the abducens anlage? As there is no sympathetic associated with the abducens in any vertebrate, the cells of the abducens can have only one fate if they persist in the adult nerve. In a non-ganglionic nerve they must form the neurilemma. With this conclusion all investigators, whatever their views regarding nerve histogenesis, agree.

h. What is the histogenesis of the neurilemma of the abducens? At the outset all of the cells of the abducens are peripheral in relation to the fiber bundle. Penetration of cells into this bundle is slow and has scarcely begun in a 25 mm. embryo. As in the case of the oculomotor and trochlearis, penetration begins in the region of the proximal roots, that is, in the oldest part of the nerve anlage. By the time the embryo has reached a length of 45 mm., however, cells have penetrated all parts of the fibrillar bundle and have begun to assume the characteristic form and relations of neurilemma cells.

{. To what an extent do the emigrated medullary elements go to form the sympathetic? The one essential respect in which the abducens differs from the other somatic motor nerves described is in the absence of the sympathetic in relation to the nerve anlage. As in the case of the hypoglossus, however, this feature appears correlated with the absence of connection with a ganglionic nerve, a correlation emphasized by Hoffmann ('00).

j. Summary of the histogenesis of the abducens. Like typical somatic motor nerves, the abducens acquires secondary connec


MORPHOLOGY OF EYE MUSCLE NERVES 97

tion with the posterior rectus muscle by the continuous outflow of the protoplasm of cells in the somatic motor column of the hindbrain. Its fibers are directly connected with the medullary neuroblasts from which they arise, while its cellular elements are partly medullary and possibly, in part, mesenchymatous. From them develops the neurilemma of the differentiated nerve. No evidence that they are genetically related to the neurofibrils has been obtained. No sympathetic ganglion is associated with the abducens anlage or with the adult nerve.

5. What light does the study of histogenesis throw upon the question of the homology of pre-otic and post-otic metameres?

The demonstration of the similar histogenesis of the eye-muscle and spinal somatic motor nerves creates a strong presumption that pre-otic and post-otic divisions of the vertebrate body are fundamentally alike. It has been found that in no essential respect does the histogenesis of eye-muscle and spinal somatic motor nerves differ. The dorsal chiasma of the trochlearis and the lack of sympathetic connection with the abducens in no way invalidates the comparison. The hypoglossus nerve differs from the typical somatic motor nerves in precisely the same way as the abducens, but its comparability with spinal somatic motor nerves is unquestioned. In fact, if the relations of somatic motor and sensory nerves in Amphioxus and Petromyzon may be regarded as primitive, those of the abducens and of the hypoglossus are more primitive than those of other somatic motor nerves in Squalus. The absence of a sympathetic is likewise a primitive character. The study of the histogenesis of the eye-muscle nerves favors the prevalent conception of the vertebrate head as once like the trunk.

But the proof of the morphological similarity of segmental nerves depends, not on evidence of histogenetic similarity alone or central relations with the motor nidulus, but equally upon the peripheral distribution. So that the question arises whether or not the myotomes innervated by the eye-muscle nerves are serially homologous with those innervated by spinal somatic motor nerves.

JOURNAL OJ MORPHOLOGY, VOL. 25, NO. 1


98 H. V. NEAL

The study of histogenesis has shown that connection between muscle and nerve is not primary but secondary. This suggests the possibihty that new segmental relations may be acquired by somatic motor nerves and that a nerve may invade territory foreign to it. In this way some peculiarities in the relations of eye-muscle nerves may be explained.

RELATIONSHIPS OF THE EYE MUSCLE NERVES

1. Is the musculature innervated by the eye-muscle nerves somitic

musculature?

The eye-muscle nerves innervate muscles derived from Van Wijhe's first, second and third somites. The oculomotor innervates muscles differentiated from the first, the trochlear innervates that derived from the second, and the abducens is connected with a muscle formed from the third and a part of the second (Dohrn '07) . Are these muscles serially homologous with those innervated by spinal somatic motor nerves? This obviously depends upon the somitic nature of Van Wijhe's three 'somites.'

They have been regarded as true somites by many morphologists including Van Wijhe ('82), Oppel ('90), Miss Piatt ('91, '97), Hoffmann ('94), Neal ('96, '98), Furbringer ('97), Koltzoff ('01), Killian ('91), Boecke ('04), FilatofT ('07) and their presence has been demonstrated in such diverse groups as Selachii (Van Wijhe '82, Miss Piatt '91, Hoffmann '94, Neal '96, Braus '99, Johnston '09); Cyclostomes (Koltzoff '01); Teleosts (Boecke '04) ; Amphibia (Miss Piatt '97) ; and Reptiles (Oppel '90, Filatoff '07).

On the other hand the eye muscles have been regarded as splanchnic muscles by Stannius ('51), Langerhans ('73), Balfour ('78), Marshall ('81, '82), Dohrn ('85, '87, '04, '07), Houssay ('90), Hatschek ('92), Rex ('97, '05) and Sewertzoff ('98), and the presence of true somites in the pre-otic region has been denied by Kastschenko ('88), Rabl ('92) and McMurrich ('12). Sewertzoff ('98 b) bases his objection to the inclusion of the second myotome in the series of somitic muscles upon the con


MORPHOLOGY OF EYE MUSCLE NERVES 99

elusion of Froriep and Miss Piatt that the trochlearis is a mixed (splanchnic motor) nerve, and not upon the basis of direct evidence.

Dohrn ('04) concluded that the superior oblique muscle is splanchnic and not somitic on the evidence that a portion of the splanchnic (ventral) mesoderm of the mandibular cavityshifts during development to a dorsal position as a result of the cephalic flexure of the brain. From this displaced mesoderm, according to Dohrn arises the masse ter muscle, the superior obhque muscle and a portion of the posterior rectus muscle.

Filatoff ('07, p. 354) takes exception to this inference of Dohrn on the ground that it is difficult to distinguish dorsal from ventral mesoderm in the mandibular cavity and more especially to determine what is dorsal and what is ventral during the process of shifting relations in successive stages. Filatoff finds that the superior oblique muscle in Emys arises from the dorsal segmented portion of the mandibular cavity.

Against the somitic value of the pre-otic mesodermic segments it has been argued that the head of the vertebrate ancestor was, like that of Tunicate larvae, unsegmented; that the supposed segmentation is not actually metameric but is the result of mechanical influence of other organ systems; that the mesodermic divisions are irregular in size and inconstant in number; that they are discontinuous with those of the trunk; that they do not differentiate sclerotome and myotome, at least in the typical manner; that the topographic relations to nerves are different from those of trunk somites. But, since renewed investigation has disproved many of these assertions, the argument in favor of the somitic value of Van Wijhe's somites seems on the basis of the following evidence, much the stronger of the two alternatives.

That Van Wijhe's somites are serially homologous with those of the trunk seems sufficiently established by the repeated confirmation of their presence in diverse groups of vertebrates; and on the ground that their segmentation is independent of the visceral segmentation; that Van Wijhe's somites form a continuous series with those in the trunk; that they are dorsal in relation to chorda and dorsal aorta; that their development is


100 H. V. NEAL

progressive, beginning with the neck region; that they differentiate into myotome and sclerotome; that their muscles are primarily differentiated from the median wall; and that they correspond numerically, as in the trunk region, with the neuromeric divisions of the neural tube. They, therefore, constitute the best established evidence of the primitive metamerism of the vertebrate head and of its comparability with the trunk. The somitic nature of the muscles derived from them seems therefore, indisputable.

2. Are the relations of the oculomotorius comparable with those of a somatic motor spinal nerve f

a. How is the relation of the oculomotor to the ramus profundus to he interpreted? The oculomotor nerve resembles a somatic motor nerve not only in its histogenesis but also in its central and peripheral relations.

Like a spinal somatic motor nerve, it arises from a nidulus in the somatic motor column of a neuromeric segment of the neural tube and innervates muscles derived from a somite serially homologous with those of the trunk. Furthermore it becomes connected with a ganglionic nerve after the manner of typical spinal somatic motor nerves.

This connection with the ramus profundus V, however, needs closer scrutiny. True, it is a ganglionic nerve, but is it certain that all ganglionic nerves are morphologically comparable? Our present knowledge of nerve components and their differences in the various cranial nerves, as demonstrated by the investigations of Strong ('95), Herrick ('99) and Johnston ('05 a), discourages the indiscriminate comparison of nerves on such superficial grounds as the possession of a ganglion. Therefore one, who maintains the similarity of the relations of the oculomotor with spinal somatic motor nerves, is bound to demonstrate the morphological similarity of the ramus ophthalmicus profundus V with spinal somatic sensory nerves. That they are similar seems proved first by their similar histogenesis from neural crest cells; second, by their similar peripheral distribution as general cuta


MORPHOLOGY OF EYE MUSCLE NERVES 101

neous nerves; and third by their common classification as somatic afferent nerves.

Three objections, however, may be raised against this homology; first, that while spinal somatic sensory nerves grow median to the myotome, the ophthalmicus profundus, like other typical cranial nerves, grows lateral to the somites; second, that cellular elements enter the ganglion of the ophthalmicus profundus from the skin, while the cellular elements of spinal sensory nerves are exclusively^ derived from the neural crest; and third, that the profundus is not an independent segmental nerve but a sensory branch of another nerve.

The first of these objections may be met by calling attention to the fact that in some cases, for example in the post-otic region of Ammocoetes, typical cranial nerves lie partly lateral and partly median to the myotomes, indicating that no sharp line of demarkation can be drawn on this basis between cranial and spinal somatic sensory nerves. Such differences appear to be correlated with differences in the relative size of the somite. Furthermore, there are comparative anatomical grounds for thinking that somatic sensory nerves were primarily inter-myotomic in position. From such primitive relations the present modified relations of cranial and spinal somatic sensory nerves may readily have been derived.

The second objection is more serious, but if the comparability of the somatic sensory nerves of Amphioxus and Squalus be granted, it will be seen that the cranial somatic sensory nerves have retained the primitive relations of the former, while the absence of direct contact with the skin in the case of spinal somatic sensory nerve ganglia, may be regarded as a secondary modification. In this respect, as in respect to the retention of a mixed function, typical cranial nerves appear more conservative than spinal nerves. Therefore, while it must be admitted that the cranial somatic sensory ganglia are more complex in their derivation than are spinal ganglia and that the serial homology of the two is incomplete, nevertheless their partial homology appears demonstrable.


102 H. V. NEAL

Finally, the fact the fibers of the • ophthalmicus profundus enter the medulla in common with those of the trigeminus will not seem to morphologists a serious objection to the comparison of the profundus with spinal somatic sensory nerves, since it would seem a matter of indifference whether somatic sensory fibers enter the brain by one path or another. In the case of the union of the roots of once independent ganglia we seem to have to do with a particular case under the general principle of the centralization of function in the region of the medulla. Analogous instances may be found in all organ systems.

Taking all the facts into consideration there appears to be no insuperable objection to the view that the ophthalmicus profundus is serially homologous with spinal somatic sensory nerves.

h. How is the relation of the oculomotorius to the ciliary ganglion to he interpreted? The comparison of the profundus nerve with spinal somatic sensory nerves is still further strengthened by the evidence of the relations with the ciliary ganglion which have been found above to be those of a somatic motor nerve to a sympathetic ganglion. The facts which prove the sympathetic character of the ciliary anlage have already been stated above and need no restatement. The ciliary ganglion of Squalus is to be regarded as partly, if not exclusively, a sympathetic ganglion. So that in its relations with a sympathetic ganglion the oculomotor forms no exception in the series of morphologically similar sortiatic motor nerves.

c. Hoiv may the relation of the oculoynotorius to four eye muscles be best interpreted? The distribution of the oculomotor to four muscles may appear to require interpretation. The fact, however, that all of these muscles are derived from a single myotome by a process of splitting and that this splitting is correlated with the considerable enlargement of the eye-ball; furthermore, that the innervation of more than one muscle by a single somatic motor nerve is by no means exceptional, brings these relations also into line with those of spinal somatic motor nerves. The oculomotor, therefore, in its histogenesis and in its relations to a somatic sensory ganglion, to a sympathetic ganglion and to


MORPHOLOGY OF EYE MUSCLE NERVES 103

somatic musculature may be considered serially homologous with spinal somatic motor nerves.

d. Conclusions regarding the morphology of the oculomotor. It is scarcely necessary to say that the conclusion that the oculomotor is serially homologous with spinal somatic motor nerves has not been reached by all morphologists. It is, however, the opinion of the majority, including Van Wijhe ('82), Beard ('85), Hoffmann ('86-'00), His ('88), Martin ('90), Dohrn ('90, '91), Zimmermann ('91), Kolliker ('96), Neal ('96, '98), Koltzoff ('01), Carpenter ('06), Filatoff ('07), Belogolowy ('08) and also, on the basis of anatomical evidence. Muck ('15), Bell ('30), Stannius ('49), Bonsdorff ('52), Budge ('55), Huxley ('74, '75), Schneider ('79), Gaskell ('86, '89), Strong ('90), Fiirbringer ('97), Wiedersheim ('98), Gaupp ('99).

On the other hand, it has been regarded as a splanchnic motor nerve homologous with the trigeminal by Balfour ('78), Marshall ('81, '82), Dohrn ('85, '87), Houssay ('90), Hatschek ('92), Sewertzoff ('98), and as a somatic sensory or mixed nervfe with coenogenetic atrophy of the sensory element by Marshall ('82), Rabl ('89), Martin ('90), Miss Piatt ('91), Mitrophanow ('92, '93), Sedgwick ('94), Sewertzoff? ('99), Gast ('09), and on anatomical grounds by Schwalbe ('79, '81) and Gaskell ('89.)

The most recent argument in favor of the latter opinion is that given by Gast. He admits that in its first anlage, the oculomotor develops in the same way as the anlage of a somatic motor nerve. On page 420 he states that naked neuroblast processes grow from the neural tube, unite to form a naked fiber bundle, grow to a segmental ganglion, and receive their sheath cells from it. Medullary neuroblasts also migrate into the nerve anlage. Sheath and sympathetic cells ^^'ander from the mesocephalic ganglion into the nerve anlage but retain connection with their source by a ramus communicans, a phenomenon quite in line with that of typical somatic motor nerves.

On the other hand, he says, there are indications that lateral horn elements are combined with these ventral root elements. In the first place, the close relations with the mesocephalic ganglion through which the oculomotor fibers grow in some instances


104 H. V. NEAL

confirm this view. In other cases the oculomotor assumes very close relations with the ganglionic placode connected with the mesocephalic ganglion. Again, there is evidence of the anlagen of sensory nerves in connection with the oculomotor. The proof of this (p. 401) consists in the fact that a short cell-chain, free from fibers, extends from the mesocephalic ganglion toward the oculomotor anlage. The absence of a fiber in the cell-chain appears to Gast to exclude the possibility that the cells are in the process of migration centrad along a motor fiber of the oculomotor. On the other hand, Gast (p. 402) regards the evidence given by Mitrophanow ('93) and Sedgwick ('92) in favor of the mixed function of the oculomotor as fallacious. But Miss Piatt ('91) may have seen what he regards as evidence of sensory elements in the oculomotor.

While Gast admits that this conclusion appears to conflict with the evidence of the position of the nidulus of the oculomotor, he asks if it is not conceivable that Dohrn's suggestion is correct — that the lateral and ventral niduli have united together. Then there is the possibility that the separation of ventral and lateral niduli in the head region is a coenogenetic separation and that they were primitively united. Gast, however, is of the opinion that the union of lateral and ventral niduli assumed for the oculomotor is a secondary one, as Dohrn suggested.

On the basis of this supposition, Gast (p. 421) indulges in the following speculation:

The segmentally arranged mesocephalic-oculomotorius system with its sensory and motor roots becomes secondarily separated into sensory and motor elements, whereby the motor neurones of the oculomotor retain their central connections, while the sensory neurones of the mesocephalic ganglion acquire a new root. In the case of the Selachii this root formation came to pass in such a way that a commissure was formed between the individual ganglia of the anterior head region (trigeminal, trochlear, mesocephalic and the ganglia anterior to these). This commissural nerve gradually assumed the character of a root, while fibers of the segmental sensory roots proportionally degenerated. To-day, indications of these primary sensory roots are found in the oculomotor as well as in the trochlear.


MORPHOLOGY OF EYE MUSCLE NERVES 105

The foundation for Gast's speculation consists, primarily, in the evidence of close or intimate relations with a cerebro-spinal ganglion. But if such logic were rigidly followed, it would be necessary to regard every spinal somatic motor nerve as a lateralhorn (splanchnic motor) nerve, since close and intimate relations with a cerebro-spinal ganglion is characteristic of spinal somatic motor nerves in Squalus.

The supposed demonstration of the participation of sensory elements in the genesis of the oculomotor is one that would satisfy only on the basis of a strong presumption in its favor. The position of the nidulus of the oculomotor and its peripheral distribution create a strong presumption against the assumption. Spindle-shaped cells lying in the mesenchyma between the profundus ganglion and the oculomotor nerve are not necessarily neuroblasts. Spindle-shaped cells may be found almost anywhere in the mesenchyma. Even if it be admitted that the evidence that these cells are in the process of migration toward the oculomotor anlage is convincing, Gast does not know their fate. They may form neurilemma or they may enter the sympathetic or what-not. Their later histogenesis is wholly unknown. Gast has not used neurofibrillar stains in order to ascertain their neuroblastic character. It would be surprising if morphologists accepted Gast's conclusion upon the basis of the slight evidence he is able to present in its favor.

The evidence in favor of the view that the oculomotor nerve is a mixed nerve homologous with typical cranial nerves such as the trigeminal is so unconvincing, while the evidence of its histogenesis and its central and peripheral relations so strongly support the supposition that it is a somatic motor nerve, as the majority of morphologists have beheved, that the acceptance of the latter seems unavoidable. This view at least does not require support from such an unproved assumption as the secondary fusion of lateral and ventral motor niduli.


106 H. V. NEAL

S. Are the relations of the trochlearis comparable with those of a somatic motor spinal nerve?

There is practically a consensus of opinion that in its most essential relations, namely, in its relation to a nidulus in the somatic motor column of the hindbrain, the relations of the trochlear are comparable with those of spinal somatic motor nerves. There is less general agreement, and yet a considerable majority of morphologists agree, that the trochlear innervates somitic musculature. The main objection which has been advanced against this conclusion has been that raised by Stannius ('51) and Langerhans (73) of the different histological structure of the superior oblique muscle, which, according to these investigators, resembles splanchnic rather than somitic musculature. This objection may be met by denying the truth of the assertion as a generalization for all vertebrates. Differences in size and in detail between fibers of the eye muscles and those of the lateral trunk muscles may exist — especially in those forms in which the eye muscles are differentiated from a loose mesenchyma, but in forms like Squalus, there is no important histological difference between the muscles of the eye and those of the trunk. Such slight differences as do obtain may be ascribed to differences in environment. That they are not due to difference in genesis appears demonstrated by the evidence of the somitic origin of the eye muscles. Therefore, in histological structure as well as in its histogenesis and in relations, both central and peripheral, the trochlearis conforms to the type of spinal somatic motor nerves. But there are other relations of the trochlear that may profoundly affect our views of its morphology.

a. How may the 'ganglion' of the trochlear be interpreted/ Gast ('09) thinks that the trochlear 'ganglia' described by Dohrn ('85), Hoffman ('89), Martin ('90) and Froriep ('91), afford very clear proof that the trochlear is a 'complete segmental nerve.' Dohrn ('07, p. 396) regarded the evidence that in Torpedo embryos rudimentary centripetal ganglionic nerve fibers unite with the trochlearis as supporting the view that the trochlear is a lateral-horn nerve. I ('98) had interpreted the same evidence


MORPHOLOGY OF EYE MUSCLE NERVES 107

as proof of the derivation of the neurilemma from the neural crest. Dohrn, however, thinks that the evidence of cellular migration into the trochlear anlage from the ramus superficialis indicates that genetic relations between the trochlear anlage and rudiments of ganglia still persist today, even though they are merely transient.

Dohrn's thorough investigation of the histogenesis of the trochlear in elasmobranchs, however, indicate that the so-called ganglia of the trochlear are irregular fragments of the neural crest lying in the region through which the trochlear grows; further, that the trochlear anlage in such forms as Pristiurus attains connection with the myotome without any relation whatever with these fragments, which develop in inverse ratio with the development of the ramus superficialis V, of which therefore they appear to be the equivalent. In other words, the relations of the trochlear with these 'ganglia' of the Torpedinidae in which they appear seem similar to its relations with the ramus superficialis V in the Squalidae. This equivalency is recognized by Gast ('09) who says that the sensory elements of the trochlear which appear as ganglia in the Torpedinidae are represented by the ramus ophthalmicus superficialis V in the Squalidae.

Whether or not this equivalency be admitted, the relations of the trochlear to the 'ganglia' of the nerve resemble those of a spinal somatic motor nerve to the ganglionic or nervous derivatives of the neural crest. Such relations disprove the somatic motor character of the nerve in question quite as little in one case as in the other.

While in Squalus the trochlear anlage has no such relations to irregular fragments of the neural crest as in the Torpedinidae, the nerve does have relations to cell masses which precisely resemble those of the oculomotor to the anlage of the ciliary ganglion, or those of a spinal somatic motor nerve to sympathetic anlagen. Such relations of the trochlear anlage are represented in figures 54 and 55. With a strong presumption in favor of the view that the trochlear is a somatic motor nerve as evidenced by its histogenesis and by its central and peripheral connections the most reasonable interpretation of the mass of cells


108 H. V. NEAL

at the place of union of the anlagen of the superficiahs and the trochlear is that it is sympathetic. Like the cells of all sympathetic ganglia those of the trochlearis appear to come from a sensory ganglion. In both instances they collect in the region where sensory and motor fibers unite. Against this view may be urged the fact that there is no sympathetic in relation to the adult trochlear nerve. But the objection loses much of its force when it is remembered that transient sympathetic ganglia are not uncommonly found in the trunk region of elasmobranchs.

This supposed sympathetic ganglion of the trochlear anlage must not be confused with those fragments of the neural crest which Dohrn ('85) and others have called ganglia of the trochlearis. Such 'ganglia' appear, as suggested by Dohrn ('07) and Gast ('09), to be comparable with degenerated portions of the superficial nerve and not with sympathetic ganglia. Belogolowy ('10 b) finds in reptile embryos a ganglion present at the point of anastomosis of the trochlearis anlage with the ramus ophthalmicus superficiahs trigemini. Belogolowy expresses the opinion (p. .70) that the so-called ganglia of the trochlearis are really the ramus ophthalmicus superficiahs trigemini which is represented in some 'forms by scattered clumps of cells. With this opinion the writer is in full accord.

b. How may the innervation of a muscle derived from a somite {Van Wijhe's second) also innervated by the abducens be interpreted? Another problem presented by the relations of the trochlear is the fact that it innervates a muscle derived from a somite also supplied by the abducens nerve. This comes about in the following manner: A portion of the myotome of Van Wijhe's second somite unites with the myotome of the third somite to form the posterior rectus muscle. Miss Piatt ('91), with characteristic accuracy, observed this connection, but was led to infer its later degeneration and to confirm the conclusion of Van Wijhe ('82) that the posterior rectus muscle is derived exclusively from the third somite. Lamb ('02) reached the same conclusion. But Dohrn ('07), after a careful reinvestigation of the development of the superior oblique muscle, asserted the


MOEPHOLOGY OF EYE MUSCLE NERVES 109

persistence of that portion of the second myotome in connection with the myotome of the third somite. Dohrn has affirmed its persistence correctly. There is no doubt whatever that it persists in the superior obhque muscle of Squalus. Its relations are diagrammatically shown in figure 81. Thus it comes about that the trochlear and abducens nerves innervate portions of the same somite. In this way a problem in nerve relations is presented which will be fully discussed in connection with the problem of the relations of the abducens nerve. Suffice it to say here that this relationship of two somatic motor nerves to a single somite does not in any way affect adversely our views of their real morphology. As a matter of fact, emphasized by Bardeen ('04), all typical somatic motor nerves have a bimeric distribution to two adjacent myotomes. The most perplexing problem raised in this connection is that there is reason to think that the trochlear and abducens nerves do not belong to successive metameres. This question will be taken up later. But, even more difficult for one who attempts to demonstrate the somatic motor character of the trochlear, is the problem of its dorsal chiasma. In this feature the trochlear is the most peculiar nerve in the vertebrate body.

c. How may the dorsal chiasma of the trochlear be best explained? Van Wijhe ('86), in order to explain the dorsal emergence of the fibers of the trochlearis, assumed that, as a result of the extension of the anterior column to the olivary body and to the loop extending behind the corpus quadrigeminum, the root of the trochlear was drawn over the loop into its present position.

According to His ('88) the peculiar relations of the trochlear may possibly be explained as the result of the flexure of the neural tube in the region of the isthmus, a condition which he thinks is favorable to the sagittal growth of the neuraxon processes of the neuroblasts at the base of the cerebellum.

Rabl ('89), however, on the basis of observations on the selachii, birds and mammals, concluded that the roots of origin, both of the oculomotor and of the trochlear were primitively dorsal, but that gradually, through the enlargement of the peduncular paths, the root of the oculomotor was shifted to the ven


110 H. V. NEAL

tral surface of the brain, while the trochlear has retained its primitive position. It may readily be seen, however, that such considerations have to do with the dorsal emergence of the trochlear fibers rather than with the chiasma of the nerve.

Martin's ('90) unconfirmed and incredible assertions regarding the transmigration of the trochlear nidulus appear to deserve no restatement, although they were accepted tentatively by both Minot ('92) and Kolliker ('96) so far as they relate to a central origin of the trochlear chiasma. Kolliker, however, finds it difficult to conceive of a transmigration of a nidulus, although von Lenhossek and Ram6n y Cajal have inferred a migration of neuroblasts from the ventral column of the tube into the sensory roots. An extensive migration of neuroblasts therefore appears possible.

Fiirbringer ('02, pp. 134-136) has given the most thorough consideration to the problem of the chiasma of the trochlear and his hypothesis, although affected by his views of the primary and unalterable connection of nerve and muscle, seems to be the most elaborate of all that have been advanced. Following an idea advanced earlier by Hoffmann ('89), Fiirbringer suggests that the trochlearis may have innervated musculature belonging to the parietal eye. According to Fiirbringer the parietal eye (or pau' of eyes) was situated primarily near the paired lateral eyes, which were then slightly differentiated and probably occupied a more dorsal position than later. The aberrant musculature connected with both kinds of eyes may have very early separated itself — as the course of the nerve indicates — from the dorsal portions of the neighboring myotomes and may, in correlation with the primitive condition of the eyes, have been in a very slightly differentiated and — so to speak — fluid condition, with its fibers extending in various directions. Those portions of the muscle which extended somewhat diagonally or transversely had the tendency, like other muscles extending toward the median line of the body, to migrate over into antimeric territory. There was nothing hke a median fin in the region to hinder this migration,, which could occur freely.

With the degeneration of the parietal eye the musculature which became associated with it disappeared. The muscles of


MORPHOLOGY OF EYE MUSCLE NERVES HI

the lateral eyes, however, which at that time were in close proximity to one another did not degenerate but became attached to, and at the same time migrated with newly differentiated fibers, to the eye of the opposite side where they increased in size while the original muscles of that side atrophied. In this way they formed the present superior oblique muscle, the nerve of which with its dorsal chiasma — primarily peripheral — still discloses the early history of the transmigration of its muscle. Then with the higher development of the retina of the lateral eyes there came an enlargement of the centers within the brain, especially a greater development of the roof of the midbrain, which extended backward over the place of emergence of the trochlear fibers and covered it, so that a part of its root as well as its chiasma was shoved backward and at the same time enclosed within the wall of the brain. The lateral eyes were not restricted to that dorsal muscle derived from the second myomere of the opposite side but very soon acquired connections with more ventral musculature innervated by the oculomotor and abducens, and, as the eyes moved into a ventral position, the muscles became enlarged and differentiated.

In support of this hypothesis, Fiirbringer advances the following considerations: First, in defense of the assumption that the parietal eye once possessed a musculature of which today there is no evidence (with possibly the exception of the unconfirmed case in mammals mentioned by Nicholas ('00) who claimed to demonstrate rudimentary striated muscles in the pineal region of the ox), he argues that the absence of parietal muscles today by no means proves that they never existed. Many skeletal structures have existed in fossil species without leaving a trace in modern vertebrates. Typhlichthys has no eye muscles but its ancestors must have had them. The assumption of an antimeric transmigration of the hypothetical ancestral eye muscles has met both opposition (Dohrn '01) and support (Gaskeil '01). In reply to the opposition, Fiirbringer cites many cases of the extensive migration of muscles in all directions within the vertebrate body. Many instances of the transmigration of musculature on the ventral side of the body are known, as for example the muscles innervated by the facialis, vagus and hypoglossus,


112 H. V. NEAL

as well as some parts of the M. sternalis. Since there is no median fin in the head region to prevent transmigration, it may have occurred there as readily as in the ventral trunk and head region.

The vagrant nature of the superior oblique muscle is evinced by the considerable variation in the origin and insertion of the muscle as well as in its extended migration in the embryo.

Several fairly obvious objections may be raised against Fiirbringer's ingenious hypothesis: First, the absence of any direct evidence from comparative anatomy and embryology that any vertebrate muscle has migrated in toto from one side of the body to the other; Second, the complete failure of ontogenesis to support the hypothesis. Ontogenesis is hardly so discredited, even by Fiirbringer, that this lack of ontogenetic support can be wholly ignored; Third, the total lack of evidence of an epiphysial musculature. Nicholas ('00) did not demonstrate the actual connection of rudimentary muscles with the epiphysis; Fourth, the improbability of such a swapping of muscles as is assumed in the hypothesis. Why the lateral eyes should lose muscles which they already possessed and adopt those of another degenerating organ is not perfectly obvious. Still other objections might be mentioned but it seems unnecessary to multiply them.

Of course it may be said that no hypothesis dealing with phylogenetic changes so remote as the origin of the trochlear chiasma may be advanced which will seem so compelling as to preclude criticism. To many, Fiirbringer's hypothesis will appear a reasonable one, on the assumption that nerve and muscle are phylogenetically inseparable. The relations of the trochlear and superior oblique muscle have always seemed a stumbling block to the supporters of that assumption, the truth of which does not seem more certain as the result of the difficulty of explaining the chiasma of the trochlear.

Fewer difficulties, it might appear, would meet an hypothesis which assumed the secondary connection of nerve and muscle by means of the free outgrowth of the nerve fibers.

Johnston ('05, p. 210) suggests that:


MORPHOLOGY OF EYE MUSCLE NERVES 113

The course of the root fibers dorsally through the brain wall may be due to the course of the fiber tracts through which they run. The position of the nucleus of the nerve relative to the tracts which form the ansulate commissure in typical fishes suggests strongly that the axones from the cells of the trochlearis nucleus may have followed some of these bundles as the path of least resistance. The tracts between the tectum opticum and the base of the oblongata, the tracts between the inferior lobes and the cerebellum, and others, all running more or less dorso-ventrally in the side wall of the brain and decussating ventrally at the level of the trochlearis nucleus — ^these bundles, which lie ectal to the nucleus of the trochlearis, may have constituted an effective barrier to the axones of the trochlearis in their attempt to reach the ventro-lateral surface of the brain. The axones may then have turned upward along the ental surface of these bundles until they reached the dorsal surface of the brain. If the fibers were thus directed in their course they would be carried to the mid-dorsal line before gaining an exit from the brain and if they then grew straight on they would pass to the opposite side.

Filatoff ('07, pp. 366-7) thinks that certain phenomena in the development of the brain may help to explain the peculiar origin of the trochlearis. At the time of the formation of the cephalic flexure the roof of the midbrain, which up to that time formed a part of the thin upper wall, thickened. The production of this thickening may be explained in the following way. By the development of the flexure the cells of the floor of the midbrain become most strongly compressed, since they come to lie directly in the place of flexure, and they seek to elongate themselves in the most direct way where the pressure is a little less strong, namely towards the upper wall. The point at which the trochlearis arises is directly determined by this elongation of the cells. The point of emergence of the trochlearis fibers is shoved from the ventral to the dorsal side.

Figure 82 of this paper suggests two possible phylogenetic .stages in the development of the trochlear chiasma. An earlier stage is represented on the left of the diagram and a later stage, corresponding essentially to that seen in some elasmobranch embryos, is shown on the right. It is assumed that originally the trochlear, as a somatic motor nerve, was distributed to the myotome of the second somite of its own side after the fashion of typical somatic motor nerves, and that this myotome, when

JOURNAL OF MORPHOLOGY, VOL. 25, NO. 1


114 H. V. NEAL

fully developed, extended dorsally to form a union or interdigitation with the antmieric myotome. As Flirbringer has correctly stated, no skeletal structures would prevent the extension of fibers across the median plane. It may be imagined that this extension of muscle fibers across the median plane was correlated with the muscular development of the prostomial region. Petromyzon still shows (fig. 77) the extension of myotomes into this region. Under such conditions shght variations in the length of the nerve fibers which grow to connect with these muscles might bring about a peripheral chiasma. The possibility that such muscles had connections with the epiphysis is not excluded, but such a supposition does not seem necessary. Changes in the extension and direction of growth of muscle and nerve fibers in this region may have been correlated with the development of the cephahc flexure which would seem to require some adjustment of the muscles since the forebrain and midbrain regions were flexed into a more \;entral position. The final result of the flexure, however, appears to have been a shifting of those portions of the musculature which persisted in this region into a more ventral position and a separation of the muscles which had been apposed in the median plane above the brain wall.

The growth and great enlargement of the lateral eyes also brought about changes in the (Van Wijhe's) second myotome, which became split into dorsal and ventral moities {my. 2 v.L, my. 2 m., 7ny. 2 d.l.) in precisely the same way as occurs ontogenetically in the post-otic ' muscles of Petromyzon as a result of the growth of the otic vesicle (figs. 78, 79).

It may be assumed that, as in the latter case, the median portion of the myotome degenerated, together with its branch of the somatic motor root (trochlear nerve), while the lateral moiety became innervated by a branch of the abducens (fig. 82, abd.). The dorsal moiety, however, retained its connection with the trochlear nerve, and possibly also with fibers from both sides of the brain, by means of a dorsal, peripheral chiasma. Then, when later this dorsal moiety degenerated with the exception of that portion which became attached to the eye-ball to form the external oblique muscle (crossed hatched in the dia


MORPHOLOGY OF EYE MUSCLE NERVES 115

gram) the result would be essentially the conditions that obtain today in those forms of selachian embryos in which a peripheral chiasma persists (fig. B, p. 25) , and are shown in the right hand side of the diagram. It is necessary, however, to assume that phylogenetically the chiasma came to be more and more central and that gradually the fibers of the trochlear nerve became exclusively crossed fibers. What factors determine the survival of the crossed fibers and the elimination of the direct is no less mysterious than those which have produced the ventral chiasmae of the eye and the pons. Possibly purely mechanical conditions of the sort suggested by the writer ('98) and by Johnston ('05) are responsible. All who have discussed the chiasma agree in one essential point — namely, that the chiasma of the trochlear is secondary and that it constitutes a coenogenetic modification of a somatic motor nerve. Therefore, its existence does not affect our views of its morphology. All other questions are subordinate to this, and it does not appear greatly to matter whether or not the true histor}^ of the origin of the chiasma has been or ever will be told.

d. Conclusions regarding the morphology of the trochlearis. The trochlear has been regarded as a somatic motor nerve on the basis of anatomical evidence by Stannius ('49), Huxley ('74, '75), Schneider ('79), Gaskell ('86, '89), Osborn ('88), Strong ('90), Ftirbringer ('97), Wiedersheim ('98), and Gaupp ('99) and on the basis of embryological evidence by Van Wijhe ('82), His f^'88), Martin ('90), Dohrn ('90, '91), Zimmermann ('91), Hoffmann ('94), von Kolliker ('96), Neal ('96, '98, '12), Koltzoff ('01), Filatoff ('07), Belogolowy ("08, '10).

On the other hand it has been regarded as a splanchnic motor nerve on the basis of anatomical evidence by Bell ('30), Hatschek ('92), Haller ('98), Furbringer ('02). Stannius ('51) and Langerhans ('73) reached the same conclusion on the basis of the histological structure of the superior oblique muscle; while Balfour ('78), Marshall ('81, '82), Dohrn ('85, '90, '04, '07), Beraneck ('87), Houssay ('90), Hatschek ('92), von Kupffer ('94) and Sewertzoff ('98) supported this view on the basis of ontogenetic evidence.


116 H. V. NEAL

Then too we have the view that the trochlear was primitively a dorsal nerve, innervating visceral musculature, advanced by Schwalbe (79, '81), Dohrn ('85) and von Kupffer ('95); innervating somitic musculature, by Miss Piatt ('91) and Hoffmann ('97, '99, '00); and finally that it was primarily purely sensor}^ and possibly secondarily mixed but eventually losing its sensory components, by Rabl ('89, possibly), Martin ('90), Oppel ('90), Froriep ('91), Piatt ('91), Mitrophanow ('92, '93), von Kupffer ('95), Hoffmann ('89, '97, '99, '00), and Sewertzoff ('98).

This brief summary of the views held regarding the morphology of the trochlearis wiil possibly suffice to show that opinion is about equally divided for and against the view that the trochlear is a somatic motor nerve. Since a detailed statement of the arguments, for and against, has been given by both Fiirbringer ('02) and Dohrn ('07) it appears to be unnecessary to discuss slightly divergent individual opinions.

The chief argument in favor of the comparability of the trochlear with dorsal ganglionic nerves appears to be, not the evidence of its dorsal origin, since the point of emergence of its fibers is ultra-dorsal rather than dorsal and in this respect it differs as much from a dorsal nerve as from a ventral one, but the evidence of its relations with neural crest cells, either aggregated as the ramus superficialis V, or as scattered clumps of neural crest cells called ganglia by several investigators. Dohrn ('07) is right in asserting in contradiction to Neal ('98) that the relations of the trochlearis with the ramus superficialis or with transient clumps of neural crest cells have a phylogenetic significance. But the fallacy of regarding the trochlear as a dorsal nerve on the basis of such evidence has ah-eady been pointed out above. By a similar argument every somatic motor nerve of the body which becomes associated with a ganglionic nerve must be regarded as a dorsal nerve.

Dohrn is also correct in asserting that the evidence of the formation of the trochlear as a bundle of neuraxon processes of medullary neuroblasts proves that the nerve is a somatic motor nerve no more and no less than it proves the trochlear to be a


MORPHOLOGY OF EYE MUSCLE NERVES 117

splanchnic motor nerve. It is not a little surprising, however, that Dohrn seems to have forgotten that the motor nidulus of the trochlear has the relations of a somatic motor nerve and not those of a splanchnic motor nerve. In the heat of the attempt to prove the trochlear a dorsal nerve Dohrn seems also to have forgotten that somatic motor nerves have relations with ganglia, both dorsal and sympathetic, comparable with those of the trochlear. Of his special views regarding the segmental relations of the trochlear consideration will be given later.

On the other hand it has been shown that in every essential detail of central and peripheral relationship, of histogenesis and of adult histological structure the trochlear must be regarded as a somatic motor nerve. Its peculiarity consists in one special feature — its chiasma, but in this respect it differs quite as much from a typical dorsal nerve as from a ventral somatic motor one. The interpretation of the mode of genesis of this chiasma is quite as easy upon the assumption that it is morphologically a somatic motor nerve as upon the assumption that it is a dorsal nerve, the opinion of Dohrn ('07) to the contrary notwithstanding. ^

If. Are the relations of the abducens comparable with those of a somatic motor spinal nerve?

As a purely motor nerve, with a ventro-lateral nidulus and with a distribution to somatic musculature, there seems to be no good reason for questioning the serial homology of the abducens with spinal somatic motor nerves. The absence of a sympathetic ganglion is one striking point of difference, however, which is correlated with the absence of any connection with a dorsal nerve. In both these respects the abducens resembles the more primitive somatic motor nerves of Amphioxus and Petromyzon, and the absence of a sympathetic ganglion is to be regarded as a case of the retention of an ancestral character.

1 Dohrn ('07, p. 410) thinks that the writer's view of the trochlear as a somatic motor nerve makes the solution of the problem of the chiasma more difficult.


118 H. V. NEAL

The second and more important point of difference between the abducens and spinal somatic motor nerves is its distribution as a post-otic nerve to pre-otic myotomes. Instead of innervating musculature of its own metamere, it is distributed to myotomes of anterior metameres (fig. 76). This relation, therefore, demands interpretation.

a. How may the innervation of pre-otic musculature, the posterior rectus muscle, by a post-otic nerve, the abducens, be interpreted? Are the myotomes innervated by the eye-muscle nerves postotic myotomes which have migrated into pre-otic territory? In the attempt to solve the problem presented by the distribution of the abducens, namely, the problem of the distribution of a postotic nerve with a post-otic nidulus to pre-otic muscle, two alternative hypotheses suggest themselves.

According to the first hypothesis, the posterior rectus muscle is to be regarded as a post-otic muscle which has migrated in the course of phylogeny into pre-otic territory, carrying with it the associated nerve, the abducens. As would be expected the nidulus of the abducens has retained its primitive position in the medulla, posterior to the otic capsule. One of the conditions which has brought about the migration of the posterior rectus muscle may have been the reduction and final disappearance of the pre-otic nmscles, through the development and hypertrophy of the sense organs, cranial ganglia and cartilage cranium. Then, after the atrophy of the pre-otic muscles, post-otic myotomes invaded the territory in the same way as occurs ontogenetically in the case of the anterior trunk somites of Petromyzon (figs. 77 and 78). This evidence from Petromyzon meets an objection which may be raised against the theory, namely, that it is unreasonable to suppose that if the pre-otic region became too crowded to retain its own muscles it would be. able to contain muscles from elsewhere, since this is precisely what seems to occur ontogenetically in this animal. In Petromyzon the pre-otic somites break up into loose mesenchyma and in later stages post-otic myotomes invade the territory. A comparison of figure 77 with figure 81, shows how similar the relations of the associated nerve in Petromyzon {nv. 2, fig. 77) and of the abducens nerve of Squalus are.


MORPHOLOGY OF EYE MUSCLE NERVES 119

McMurrich ('12) has recently advanced a theory (p. 175) which involves the assumption of this migration of post-otic myotomes into pre-otic relationships. Johnston ('02 and '05, pp. 230-233) has in my opinion completely refuted Fiirbringer's argument in favor of the phylogenetic migration forward of trunk nerves and myotomes into the occipital region. To all who hold the view of the primary continuity of nerve and muscle the hypothesis will seem the only tenable one, in spite of obvious difficulties and objections. Belogolowy ('10 a, pp. 380-384) has recently advanced an extended argument against the hypothesis of the primary continuity of nerve and muscle.

The most obvious objection to this hypothesis is the entire lack of ontogenetic evidence in its support. Had the phylogenetic migration of muscle assumed by the hypothesis actually occurred, we should expect to find some ontogenetic evidence of it. But there is as little ontogenetic evidence that the somites from which the posterior rectus muscle develops have migrated from behind the ear as that the mandibular and hyoid arches with which they are associated topographically have migrated from a post-otic position into their present location in the embryo. Furthermore, the hypothesis cannot be reconciled with the fact that the abducens — a post-otic nerve — innervates a myotome (Van Wijhe's 2nd) of which a dorsal moiety is innervated by the trochlearis, a nerve with a pre-otic nidulus (fig. 81). The niduli of these two nerves lie in widely separated neuromeres of the brain — one of them pre-otic and one-postotic — and yet they innervate muscles derived from the same somite. If the posterior rectus muscle were once post-otic, it is difficult to explain how the somite from which it is in part derived (Van Wijhe's 2nd) is innervated also by a pre-otic nerve with a pre-otic nidulus. If the hypothesis were true, it would be necessary to assume a migration of the nidulus of the trochlearis from behind the ear into its present position. Of such a migration of a motor nidulus from one metamere into another several segments removed, there is neither comparative anatomical nor embryological evidence. The careful comparative anatomical investigation of the nidulus of the abducens by Kappers ('10) discloses no such migration of the nidulus as McMurrich's


120 H. V. NEAL

hypothesis requires. On the other hand, there is much evidence of the distribution of motor nerves — such as the abducens — into metameres other than those in which they have their nidulus. This fact suggests a second hypothesis of the origin of the relations of the abducens to the posterior rectus muscle.

According to the second hypothesis, the abducens has attained its present relations by a process of substitution or nerve piracy; that is to say, a post-otic nerve has, in the course of phylogeny, usurped the area of distribution of a pre-otic one. The clue to the process by which this substitution has taken place is afforded by the anterior post-otic nerves in Petromyzon.

Ontogenetically, the two anterior post-otic somites of Petromyzon divide into median and lateral divisions, of which the former lie median to the otic capsule and adjacent to the notochord, while the latter lie lateral to the otic capsule and to the large ganglia of the ninth and tenth cranial nerves (figs. 78 and 79) . The lateral portions of the myotomic divisions divide (completely, in the case of the first post-otic myotome) into a dorsal division {my. 4 d.L), above the otic capsule, and a ventral division, below the ear, the division occurring along the line of lateral line sense organs. The relation between the series of lateral line sense organs — including the otic capsule — and the line of cleavage of the myotomic divisions seems more than merely topographic, and there is little reason to doubt that the development of the ear and the sense organs and their related ganglia has been one of the conditions — if not the essential condition — of the splitting of the myotomes. That is to say, the splitting of the myotomes may reasonably be regarded as an adaptation to the conditions brought about by the enlargement of the sense organs and cranial ganglia. The median division of the anterior post-otic myotomes develops embryonic muscle fibers, which degenerate and disappear in relatively early embryonic stages, for they are entirely absent in a 50 mm. embryo. With them disappear, it may be inferred, the associated somatic motor nerves; although, in spite of much pains, I have been unable to demonstrate the existence of embryonic nerves associated with these myotomic divisions. The reason for this fail


MORPHOLOGY OP* EYE MUSCLE NERVES 121

ure may perhaps be ascribed to the small size of lamprey embryos and the great difficulty of identification of embryonic nerves. The differentiation of muscle fibers, however, as Harrison has shown, does not imply the presence of nerves.

The portions of these myotomes lying lateral to the ear persist into the adult and form the anterior segments of the lateral trunk musculature (fig. 77; my. 4 d.l., my. 4 v.l.). They are innervated by branches of nerves of posterior myotomes, namely, those of the fourth and fifth post-otic myotomes. That each of the anterior myotomes was at one time innervated by its own segmental somatic motor nerve seems indisputable, and some explanation of the present modified relationships seems required. We may assume that the nerve {rx.v.) {;n'v. 1 and n'l;. 2, fig. 77) which innervates the five most anterior post-otic myotomes consists of the combined nerves of these five segments, each of which retains its primary connection with its related myotome. This assumption harmonizes with the hypothesis of 'the primary continuity of nerve and muscle. Were this assumption correct, however, we should expect to find the nidulus of the nerve extending anteriorly as far as the otic capsule. Of this there is no evidence. The only neuroblasts, the processes of which can be traced into the roots of this nerve, lie solely in the region of the roots.

As an alternative explanation of the nerve and muscle relations under discussion, it may be assumed that the nerves associated with the three anterior myotomes have degenerated, while their area of distribution has been usurped by the nerves of posterior myotomes. The degeneration of the median division of the first post-otic myotome may have been one of the conditions which lead to this substitution in the case of this myotome. Such an assumption of nerve substitution is in harmony with the rapidly increasing evidence in favor of the process theory of nerve development and with the limitation of the nidulus of the nerve as stated above.

Furthermore, the evidence presented by Johnston ('08) and his conclusions, support this assumption. In his discussion of the segmental relations of ventral nerves in Petromyzonts John


122 H. V. NEAL

ston writes (p. 584): There seems to be no definite or constant arrangement of these motor fibers. They pass in a haphazard fashion to one or two myotomes, branch once, twice or three times, et cetera. In studying the peripheral nerves of Amphioxus with methylene blue I gained the general impression that the nerves in that animal showed still less regard for segmental relations." The point of these statements in this connection is, not the fact* of irregularity of nerve relationships in these primitive chordates, but the evidence against the view of the inseparability of muscle and nerve, afforded by them. Such facts point unmistakably toward the possibility of changed innervation under changed conditions. Moreover, the more recent conclusions regarding the phylogeny of the nervous system (Parker '10) are against the view of the primary continuity of nerve and muscle.

Still further in confirmation of the view that the abducens has acquired its present relations by a process of substitution, is the fact to which my attention has been called by Dr. W. H. Lewis, that when the digastricus first arises in the human embryo it is innervated by the facial nerve. Later in development, the muscle divides and the anterior belly becomes innervated by a branch of the trigeminal. If such a process of nerve piracy occur ontogentically, it is clearly possible that a similar process may have taken place phylogenetically in the case of the abducens.

h. How may the innervation of musculature derived from two somites by a single nerve — the abducens — be best interpreted? The majority of investigators have confirmed Van Wijhe's statement that the musculature innervated by the oculomotor is derived from the first or pre-mandibular somite; that innervated by the trochlear is differentiated from the second or mandibular somite; while the abducens musculature is developed from the third somite. That is, each eye-muscle nerve is distributed to a single somite. With the exception of Dohrn ('01, '04) all students of the genesis of the eye muscles including Kastschenko ('88), Miss Piatt ('91), Hoflmann ('97), Neal ('98), Sewertzoff ('99) and Lamb ('02), agree upon the monomyotomic distribution of these nerves. Belogolowy ('10 a, p. 252) finds that the anlage of the


MORPHOLOGY OF EYE MUSCLE NERVES 123

posterior rectus muscle in the chick consists of two or three masses, but he does not suggest that this division indicates a metamerism of the muscle.

Upon the basis of observations upon Torpedo embryos Dohrn ('90) inferred that each of the eye-muscle nerves has a polymeric origin and distribution, the abducens representing three or four metameres, the oculomotorius possibly a larger number, and the trochlear a single metamere. A year later, however, he concluded that the m. obliquus superior is developed from two myotomes. JRegarding the number of somites in Torpedo, however, Killian ('91) and Sewertzoff ('99) reached conclusions divergent from those of Dohrn, and Dohrn ('91, '01, '07) has repeatedly revised his conclusions, and in his latest paper ('07) decides that the second somite of Van Wijhe represents three or four myomeres. In this connection the statement of Sewertzoff ('98) that the somites of Dohrn ('90) and Killian ('91) in Torpedo secondarily merge into those of Van Wijhe and that the myomeric segmentation of Torpedo and Pristiurus is identical, is important.

Miss Piatt ('91) had noticed that just anterior to the anlage of the posterior rectus muscle there appears the rudiment of a large muscle derived from the posterior portion of the second (mandibular) somite. This muscle, according to Miss Piatt, soon degenerates and her conclusion has been confirmed by Lamb ('02). Johnson ('13) thinks that he is able to identify the same rudiment in Chelydra embryos. But Dohrn ('01, '04) denies its degeneration and affirms its persistence as an integral part of the definitive posterior rectus muscle, the major portion of which is derived from the third somite of Van Wijhe.

Sewertzoff ('99) derives the posterior rectus muscle of the selachii from Van Wijhe's third somite, which he regards as the first true somite, and which he states is formed by the confluence of two primary somites. Sewertzoff makes the interesting discovery in Torpedo that the third and fourth somites unite to form the posterior rectus muscle. This divergence of opinion regarding the metameric relationships of the posterior rectus muscle is important and the whole problem should be reinvestigated in embryos of both Squalidae and Torpedinidae.


124 H. V. NEAL

After a careful reinvestigation of the genesis of the posterior rectus muscle the writer is able to affirm with positiveness that Dohrn ('01, '04) is correct in claiming the persistence of the rudimentary muscle which Miss Piatt ('91) thought was transient in the embryo. On the contrary, it persists and forms, as Dohrn stated, the anterior portion of the posterior rectus muscle, and is innervated by the abducens nerve.

The problem presented by the fact of the distribution of the abducens nerve to two myotomes — Van Wijhe's second and third ■ — is more apparent than real, since, as has been shown by Bardeen ('98) for mammals and by Johnston ('08) for Cyclostomes, a bimeric distribution is the rule for somatic motor nerves. Therefore, the distribution of the abducens to two myotomes, instead of presenting a difficulty, constitutes still further evidence of its morphological similarity with spinal somatic motor nerves, c. Is the abducens of Gnathostomes homologous with the most anterior spinal nerves of Petromyzon? The cyclostomes have no posterior rectus muscle and no abducens nerve (Johnston '05 a). Assuming the primitive character of the cyclostomes, in this feature the conclusion seems inevitable, therefore, that in the course of phylogeny one or two of the anterior spinal nerves of the Cyclostomes have been converted into the abducens of the Gnathostomata. The apparent similarity between the ventral ramus of the anterior spinal nerves of Petromyzon (n'v. 2, fig, 77) and the abducens gives sufificient grounds for raising the question whether there may not be an homology — partial or complete — between them. The position attained by the musculature {my. 4 V.I., fig. 77) innervated by the nerve n'v. 2 is so close to the eye as to make possible, through slight variations in the course of phylogeny, its attachment to the eye-ball. The nidulus of origin of both nerves is post-otic and somewhat extended, while the distribution is pre-otic. But here the resemblance ceases.

Against the exact homology of these nerves, it may be urged; first, that the myotome innervated by the abducens is a preotic one, while that innervated by the nerve n'v. 2 is a post-otic myotome; second, that the nerve n'v. 2 is the nerve of the fourth


MORPHOLOGY OF EYE MUSCLE NERVES 125

and fifth post-otic myotome while the fourth and fifth post-otic myotomes of Squalus (Van Wijhe's 7th and 8th) have their own somatic motor nerves, namely, the anterior roots of the embryonic hypoglossus (fig. 81), which must therefore be the exact homologues of the nerve n'v. 2.

The relations are such, however as to indicate that the abducens is the intrinsic nerve of Van Wijhe's 4th and 5th somites. The transient ramus recurrens of the abducens (VI, rec, fig. 76) may possibly indicate the former relations of the abducens to the 6th somite or at least to posterior myotomes. If this conclusion be correct, then it follows that Petromyzon has, in the course of phylogeny, lost the homologue of the abducens with the loss of the median portions of the post-otic myotomes.

It will be recalled that the abducens has a ramus recurrens, observed by Dohrn ('90, '01), Miss Piatt ('91), Neal ('98) and Belogolowy ('10), and interpreted by them as evidence of a former posterior distribution of the abducens. If this conclusion be correct, the homology of the abducens with the lost nerves of Petromyzon is rendered more certain.

One of Belogolowy 's ('10 a) most important discoveries is the fact that in chick embryos the roots of the abducens and of the hypoglossus form members of a continuous series of anastomosing roots. By such evidence their serial homology seems clearly demonstrated. Moreover, he also finds roots of the abducens arising from the second hindbrain neuromere — a pre-otic neuromere — and infers that this root is a remnant of the primitive nerve belonging to the third somite. Belogolowy lays considerable stress upon the fact that anastomoses are formed between the abducens and the oculomotor fibers, resulting in relations resembling those between the abducens and the hypoglossus. Altogether the evidence seems overwhelming in favor of the view that all of these nerves form members of a continuous series of homologous nerves, and that the assumed distinction between post-otic and pre-otic regions of the head is arbitrary and artificial.

d. Conclusions regarding the morphology of the abducens. That the abducens is a somatic motor nerve innervating somitic mus


126 H. V. NEAL

culature has been held on anatomical grounds by Bell ('30), Stannius ('49), Huxley ('74, '75), Schneider (79), Gaskell ('86, '89), Strong ('90), Fiirbringer ('97), Wiedersheira ('98), Gaupp ('99), Kappers ('10) and on embryological grounds by Van Wijhe ('82), Beard ('85), His ('88), Dohrn ('88, '90), Martin ('90), Oppel ('90), Zimmermann ('91), Miss Piatt ('91), Hatschek ('92), Hoffmann ('94-'00), von Kupffer ('94), von Kolliker ('96), Neal ('96, '98), Sewertzoff '('98, '99), Carpenter ('06), Koltzoff COl), Filatoff ('07), Belogolowy ('08, '10).

Relatively few morphologists have regarded the abducens as a splanchnic motor nerve. These are Stannius ('51) and Langerhans ('73) on the basis of the resemblance of the histological structure of the posterior rectus mucle to visceral musculature and Balfour ('78), Marshall ('81), Dohrn ('85) and Von Kupffer ('94) on embryological grounds.

Therefore on the basis of the strong preponderance of morphological opinion and on the ground that in its histogenesis, in its relations to a somatic motor nidulus and to somitic musculature, and finally in its histological structure the abducens resembles a somatic motor nerve no alternative view of its morphology seems possible. In its two divergent characters, namely, its lack of connections with sensory and sympathetic ganglia, the abducens shows primitive features which do not affect our conception of its morphology.

If, on the basis of the considerations presented above, there seem good reasons for thinking that the pre-otic and post-otic regions of the head were primitively alike and segmented in correspondence with a somitic musculature, the question naturally arises whether the eye muscle nerves and their relations throw any light upon the vexed question of the number of cephalic segments. The demonstration of a segmented somitic musculature with associated somatic motor nerves in the head region would seem to warrant an optimistic view of the possibility of a definite answer to the problem with which morphology has wrestled without cessation for over a century. We may therefore turn to the following question:


MORPHOLOGY OF EYE MUSCLE NERVES 127

5. How many metameres are represented by the three eye-muscle

nerves?

It was the undisputed opinion of the earher morphologists that each of the eye-muscle nerves represented a single metamere. As long as vertebrate morphology was largely based upon anatomical data, or at least the leading morphologists were comparative anatomists, there appeared little reason for assuming a polymerism of these nerves. Each was considered a segmental nerve or the ventral root of a segmental nerve and that conception met all intellectual demands.

When, however, comparative embryology developed, the discovery of a cephalic coelom and an independent mesodermic segmentation in the head region of elasmobranch embryos drew investigators away from anatomy, and the history of the head became written largely in terms of ontogenesis. Competition arose among embryologists to determine who could discover the largest number of ancestral head segments; and somites, neuromeres and epibranchial placodes became successively the favorite objects of investigation. The need of motor nerves to supply these segments soon became apparent and all evidence of polymerism of nerves was eagerly sought.

Dohrn ('90), upon the discovery of more numerous mesodermal segments in Torpedo embryos than had been discovered elsewhere, became a strong advocate of the polymerism of the eye-muscle nerves. For it was evident that the morphological importance and segmental value of his mesodermic segments in large measure depended upon the demonstration of a corresponding segmentation of other organ systems. However, objections were quickly raised to throw doubt upon the real metameric value of Dohrn's mesodermic segments. It was soon found that some of the microcoelic cavities which Dohrn had called somites were merely transient vesiculations of the n;iesoderm of the mandibular arch ventral to the somitic mesoderm ; that the segments did not correspond upon the two sides of the body (an objection to which Dohrn replied that it was to be expected in degenerating structures) ; that the segments are not constant, as evinced


128 H. V. NEAL

by the divergence in the results of different investigators — Dohrn ('90), KilHan ('91), and Sewertzoff ('98, '99); that similar microcoeles are seen in the trunk region but that they soon become confluent to form the definitive somites, just as in the head region they unite to form the somites of Van Wijhe (Sewertzoff '98, '99); that they lack numerical correspondence with other metameric structures; that they appear only in a divergent and highly modified group of elasmobranchs — the Torpedinidae; and that they may be considered of metameric value only by ignoring the evidence from comparative anatomy.

Notwithstanding all the objections raised against their metameric worth, Dohrn regarded them as the essential criteria of the primitive segmentation of the vertebrate head, and he considered this opinion supported by evidence of the polymerism of the eye-muscle nerves. Gast ('09) shares with him this conception of their value on the ground (p. 424) that the polymerism of the hyoid and mandibular arches is attested not only by the different cavities of their mesodermic segments but by the strong evidence of the polymerism of the nerves associated with them." Brohmer ('09, p. 39) objected to Dohrn's conclusions on the ground that although it is necessary to assign a nerve to each 'primitive segment' of the head, Dohrn had left this anatomical standpoint out of consideration. The truth of this assertion is emphatically denied by Gast ('09) who affirms that Dohrn did not disregard the innervation of the mesodermic segments. Gast accuses Brohmer of being only superficially acquainted with Dohrn's argument for the polymerism of the trochlear and the abducens. Brohmer, however, is not the only morphologist who is skeptical of the real existence of the 'nerves' which Dohrn associates with his numerous mesodermic segments. As a matter of fact the foundation for Dohrn's assumption of the polymerism of the eye-muscle nerves, namely, the polymerism of Van Wijhe's somites, is denied.

Dohrn bases his inference of the polymerism of the trochlear on the assumption that the mandibular cavity is polymeric; that the trochlearis divides into two branches; that there are two hindbrain neuromeres corresponding to the two trochlear nerves;


MORPHOLOGY OF EYE MUSCLE NERVES


129


and that there are in Torpedo marmorata two transient gangha associated with the trochlear anlage. Text figures C and D illustrate Dohrn's argument. Corresponding with the four mesodermic segments which he claims are represented in the mandibular cavity Dohrn discovers two 'trochlear neuromeres' {tr'ch. 1; tr^ch. 2) and two trigeminal neuromeres (IV and V). The



VII opt suy^^Vjj^^ '^ U'i'V-.-O;/ V J?'



Text fig. C Horizontal section of a 9 mm. embryo of Torpedo marmorata, after Dohrn ('07); cl.crs.n. 1, 'chiasma' group of cells; cl. crs.n. 2, anterior 'trochlear ganglion;' cl.crs.n. 3, second 'trochlear ganglion;' cl.crs.n. 4, Gasserian ganglion; VII.opt.su., ganglion of the r. ophth. sup. facialis; ///, IV, Neal's neuromeres.

Text fig. D A parasagittal section of an embryo of Torpedo, after Dohrn ('07); I-VII, neuromeres according to Neal ('98); tr'ch. 1, tr'ch. 2, Dohrn's trochlear neuromeres; IV, V , Dohrn's trigeminal neuromeres; VI, Dohrn's facialis neuromere.

two 'trochlear ganglia' found in Torpedo are seen in the frontal section represented in figure 3.

In support of the contention that the second myotome is polymeric and corresponds to at least two myotomes, Dohrn mentions the fact that from this myotome are differentiated two


JOURNAL OF MORPHOLOGY, VOL. 25, NO. 1


130 H. V. NEAL

muscles innervated by two independent nerves, the abducens and the trochlear. This evidence alone, however, does not necessarily warrant the conclusion drawn by Dohrn, since each myotome is typically innervated by two somatic motor nerves. Further, the two myotomic divisions of the mandibular cavity are not antero-posterior in their relations, as Dohrn's hypothesis would require, but are dorso-ventral in their relations to each other.

Since it would be difficult to find any somatic motor nerve which does not divide into at least two branches, this fact does not materially strengthen Dohrn's polymeric assumption.

The second of Dohrn's trochlear neuromeres is a late secondary subdivision of the most anterior hindbrain neuromere. From the evidence of a purely topographic relation with a clump of disintegrating neural crest cells, Dohrn infers that the trochlear includes the splanchnic motor niduli of two metameres. Unfortunately for this supposition, however, Dohrn is unable to demonstrate the presence of these two niduli. He states that no splanchnic motor fibers persist in connection with the second trochlear neuromere. Here again he is in error, since the splanchnic motor fibers of the ramus mandibularis trigemini have their nidulus in this portion of the hindbrain. These fibers enter the brain as the minor root of the trigeminal. If Dohrn's scheme of segmental relations were correct, two splanchnic motor nerves — his 'second trochlear' and the ramus mandibularis trigemini would have their niduli within the same neuromere — Dohrn's 'second trochlear' neuromere. These nerves actually do have their niduli in a single neuromere (neuromere IH — Dohrn's first and second trochlear) but one has a nidulus in the somatic motor column while the other has a splanchnic motor nidulus. The weakest point in Dohrn's argument is its failure to take into consideration the central nidular relations. The argument in favor of the polymerism of the trochlear seems unconvincing, although it is the strongest argument yet advanced with this purpose in view. The argument in favor of the polymerism of the oculomotor is even less adequate.


MORPHOLOGY OF EYE MUSCLE» NERVES 131

In objection to Dohrn's conclusions Belogolowy ('10 b, p. 23) advances the following considerations:

Dohrn, on the basis of the plurality of the ganglionic clumps connected with the trochlearis anlage and the subdivision of the nerve into single fibers, inferred the plurality and polymerism of the trochlearis. The discoveries of Dohrn appear to me quite insufficient to draw such a conclusion from, since it is impossible to use as a basis for such inferences the plurality of position and number of such indefinite elements. Were we to follow this method, it would be easy as the final result to consider the maximum number of branches which might be met by chance as the number of segmental nerves which have participated in the formation of the trochlearis; and the presence of clumps of neural crest cells which accompany these branches might be considered as sufficient proof of our assumption. But it appears to me at least risky to admit as a decisive criterion for our inferences the accidental occurrence of this or that number of anastomoses or ganglionic clumps.

On the other hand, there appear good reasons for regarding the abducens as a nerve representing more than a single metamere. In the first place its roots arise from at least two hindbrain neuromeres (neuromeres VII and VIII) and its nidulus is equally extensive; its transient and rudimentary ramus recurrens is suggestive of an earlier distribution to posterior myotomes; and lastly, it is distributed to myotomes other than those of its own metamere.

It might seem at first thought as if the most convincing evidence bearing upon the question of the polymerism of the abducens and the other eye-muscle nerves is the simple fact that they are distributed to three successive myotomes, Van Wijhe's first, second and third. And were this the only relation to be taken into consideration, such a conclusion would seem unavoidable. This conclusion however conflicts with the evidence that at least three neuromeres separate the niduli of the trochlear and abducens; and, further, the evidence that the abducens innervates muscles of other metameres than those of the post -otic metamere to which it belongs. Therefore the problem of the number of metameres represented by the three eye-muscle nerves cannot be solved by ignoring the neuromeric relations.


132 H. V. NEAL

Ziegler ('08) and his pupil Brohmer ('09) have recently attempted to determine the number of cranial metameres, ignoring the neuromeric segmentation. Is neuromerism a matter of no consequence to the student of head morphology?

6*. Are neuromeres satisfactory criteria of the primitive metamerism

of the head?

By his discovery of symmetrical cross furrows on the widely expanded neural plate of Salamandra atra embryos von Kupffer became the first exponent of the view that the nervous system manifests a 'primary metamerism,' independent of the mesodermic segmentation. He was the pioneer — as he was in many other lines of morphology — among those who lay stress upon the segmentation of the nervous system as the best preserved manifestation of the primary metamerism of the vertebrate body. Repeated confirmation of the presence of neural segments prior to the appearance of mesodermic segmentation has been given by Froriep ('91, '93), Locy ('95), Hill ('00), Johnston ('05), Wilson and Hill ('07) and Griggs ('10).

At the same time it is a truism of morphology that the segmentation of the nervous system is a secondary one, determined by and dependent upon the segmentation of the muscular system. This has been especially emphasized by Mihalkowitz ('77), Ahlborn ('84 b), Froriep ('92), Kingsley ('12) and Coghill ('13). Observation, however, seems to have warranted the generalization that the nervous system is most conservative. When once it has acquired a segmentation adapted to whatever peripheral system, the segmentation is retained, even after the associated structures, sensory or muscular, have disappeared. So that when it was demonstrated beyond a doubt that the hindbrain neuromeres manifest a segmentation which could not be interpreted upon purely mechanical grounds and which appear independently of the mesodermic segmentation, the conclusion seemed inevitable that in these hindbrain neuromeres is preserved the indisputable remnants of the primary segmentation of the head. Finally when Locy ('95) claimed to have been able to trace the


MORPHOLOGY OF EYE MUSCLE NERVES 133

'primary' neuromeres of the open neural plate of selachian embryos into the neuromeres of the closed neural tube, it seemed as if neural segmentation were the court of last appeal in all questions relating to the ancestral metamerism of the head. Few morphologists — Mihalkowitz (77), Broman ('95), and Filatoff ('07) — have held the neuromeres to be purely the results of mechanical pressure or of growth in confined space and devoid of phylogenetic significance. This interpetation has been discredited since the demonstration that the hindbrain neuromeres of many vertebrate embryos are local thickenings of the medullary wall and not defined merely by foldings of the wall.

Johnston ('05, p. 234) probably expresses the general attitude of morphologists when he says that "nervous structures represent more segments than have been preserved in the mesoderm. In other words, there are preserved vestiges of nerve structures belonging to segments whose entodermal and mesodermal organs have disappeared for the most part. But the division of the brain wall into neuromeres gives a clue to the number of segments." The same implicit confidence in the value of the subdivisions or segments of the nervous system as criteria of the primitive metamerism is expressed by Griggs ('10, p. 434) in the conclusion that if, as the most recent investigation seems to show, the nervous system, appearing first, presents a simpler and more unaltered condition than the other two systems, then it may well serve as a basis for the study of the segmentation of the head; and other organs should be shown to correspond to it rather than vice versa." It may well be doubted, however, if the truth of the premises of either Johnston or Griggs may be admitted.

Notwithstanding the faith inspired in these morphologists that, through the study of neuromerism the primary metamerism of the head will be ascertained, the conflict in their observations and conclusions seems hardly to justify their confidence. In the Urodeles, for example, Kupffer ('85) finds eight primary neuromeres in the region where Froriep ('91, '93) finds three, or four, or five, and an anterior unsegmented region large enough to include three or four more. Froriep, however, denies their seg


134 H. V. NEAL

mental value. Eycleshymer ('95) also finds only a few large segments of no metameric significance. Locy ('95) found four or five in the region and agreed with Froriep and Eycleshymer that they were of no segmental importance. Griggs ('10) recognized one neuromere in the forebrain region, two in the midbrain and one or more in the hindbrain region and regards these as the only true neuromeres. On the other hand, there are "in the closed neural tube of Amblystoma a series of swellings, extending from the anterior end of the brain to the otic pit, but since these divisions are of varying morphological significance they cannot rightly be called neuromeres." By what criteria shall the real neuromeres be determined? Is Griggs correct in denying neuromeric value to the forebrain or midbrain vesicles or their secondary subdivisions? Are the true neuromeres those of the open neural plate of Amphibia, as Griggs maintains, or are the real neuromeres those of the closed tube as held by the majority of morphologists? The difficulty of determining the real metamerism of the nervous system is increased, if morphological opinions are influenced at all by the segmentation of the brain of Bdellostoma.

Dr. Bashford Dean, in a letter which he kindly permits me to use, makes the following statement of the neuromeric conditions in Bdellostoma embryos:

If we believe that a 'neuromere' is represented objectively by a specially and definitely dilated spot in the central nervous system, expressed either in the inner lumen or in the outer wall of the medullary axis, we certainly cannot interpret the conditions in the hag-fish in terms, for example, of the shark or the amphibian. In the first place, the hag-fish embryos may show a great number of these dilated regions, as many indeed as twenty-seven or twenty-eight in the midbrain and hindbrain. They may indicate also that these dilated areas are indefinite in number; and that in relatively the same age these areas may be either very obscure, absent, or barely visible, or may be reckoned with almost mathematical precision. They are moreover, rarely symmetrical; it is curious also that their symmetry is never expressed in the same way in the large series of embryos examined. Neuromeres are difficult to distinguish in the forebrain of Bdellostoma. Occasionally there appear to be two, three, or four present. These can be discerned faintly on one side or the other side of the medullary axis, never paired.


MORPHOLOGY OF EYE MUSCLE NERVES 135

In the midbrain as many as eight may be reckoned, and in other eml)ryos none at all. A s^^mmetrical number sometimes occurs. In the cases just examined, I find five specimens out of fourteen showing the same number of midbrain neuromeres on both sides. In one case I counted three right, three left; in another eight right, eight left; in another three right, three left; in another four right, four left, and finally one in Avhich but a single neuromere appeared on either side. In the hindbrain, the number of neuromeres varies between three and twenty-four, and they differ in number on different sides, a difference of ten having Ijeen noted in the same individual in right and left sides of the body.

It may be objected of course, (1) That in these cases of asymmetry the neuromeres were present on the 'off' side but were 'suppressed,' and therefore could not be counted. (2) That the change in the number of neuromeres might be due to different stages in development, the neuromeres appearing most completely at a definite date of development. The fact remains, nevertheless, that asymmetry and variable numbers are present, and to an extraordinary degree, a state of affairs which does not make in the dii'ection of clearing up our knowledge of these 'segmental' structures. (3) That the neuromeres of Bdellostoma are artifacts. Against this criticism I note that I have seen them clearly in living embryos.

But it may be objected that the foldings of the Bdellostoma brain are not the sjrmmetrical foldings which have been recognized as neuromeres in other forms; that no one would be likely to interpret the variable and asymmetrical structures which Dean describes in Bdellostoma as of phylogenetic or morphological value; that real neuromeres are symmetrical and permanent thickenings of the brain wall. In reply to such objections it should be remembered that morphologists have not always insisted upon the constancy or the symmetry of problematic structures. In fact Dohrn ('04), in reply to this objection to his microcoelic mesodermic segments, replied that inconstancy and asymmetry would be expected in degenerating metameric structures like these. Out of hundreds of Squalus embryos examined by the writer in order to confirm Locy's results only two or three showed symmetry or regularity in the segmentation of the edges of the neural plate (figs. 2 and 3, Neal '98). Yet morphologists have not refused to accept Locy's conclusions because of this lack of confirmation of his results on selachians.


136 H. V. NEAL

The main objection to the use of neuromeres as essential criteria of metamerism is not so much their variabiUty in different vertebrates or the diversity of opinion among morphologists regarding them, as it is the difficulty of finding criteria by means of which the coenogenetic may be distinguished from the palingenetic, especially in the regions of the forebrain and hindbrain. In an earlier paper ('98)* the writer protested against the uncritical acceptance of all sorts of foldings of the central nervous system — dorsal, ventral or lateral — as evidence of the primary metamerism of the nervous system. Special protest was raised against the claim of Locy ('95) that he had been able to trace the ^primary' neuromeres of the open neural plate into the neuromeres of the closed neural tube. Eycleshymer ('95) and Kingsley ('97) were likev/ise unable to accept Locy's assertions.

Hill ('00), however, confirmed Locy's results by his observations on teleost and chick embryos and considered Neal's objections as 'negative.' Wilson and Hill ('07, p. 147) on the other hand cannot admit that Hill has fully and adequately met Neal's objections to Locy's interpretation of the early crenation of the margin of the cephalic plate, for example, in Squalus. The weightiest part of Neal's contention, as it appears to us, is not merely negative, as C. Hill represents it, but resides in the positive statement that the beaded thickenings found are not only asymmetrical but are quite variable in different specimens."

Johnston ('05), ignoring Neal's objections, accepts the results of Locy and Hill on the ground that "the work of these last two authors is evidently most painstaking and their results are so complete and so far in agreement that they may be taken to represent the present state of knowledge of the neuromeres." Yet von Kupffer ('06, p. 164), working on chick embryos, finds that Hill's 'astonishing pictures' of the neuromeres of the chick give the impression that "the subjective motive of the investigation had influenced too much the completion of the drawings." Kupffer states (p. 248) that in spite of equally extended observations he was not able to confirm Hill's results. Graper ('13) also has been unable to find Hill's neuromeres in the chick.


MORPHOLOGY OF EYE MUSCLE NERVES 137

Therefore, in view of the disagreement among students of neuromerism regarding the nature sind number of true neuromeres, and the persistent doubt regarding the results of Locy and Hill, morphologists may still feel skeptical regarding any scheme of metamerism based upon the segmentation of the nervous system and uncontrolled by the evidence of mesodermic segmentation.

The writer finds himself in agreement with Belogolowy ('10 a, p. 510) in the opinion that the neuromeres have only a secondary importance as criteria of the primitive segmentation of the head. The latter states (p. 515) that:

Without ha\Tiig any organic relations to the functional activities of the nervous system, and presenting merely form changes of the neural tube, the neuromeres in my opinion, can serve at best merely as topographic landmarks in the stud}^ of the nervous centers. The complete lack of any satisfactory explanation of their appearance and the indications of the possibility of their purely secondar}^ formation under the influence of this or that mechanical factor acting on the nervous system — ^as in the case of the constrictions of the spinal cord under the pressure of the somites — •limits to the utmost their employment as criteria of metamerism.

Johnston's assertion ('05, p. 234) that "nervous structures represent more segments than have preserved in the mesoderm;" that in other words, there are preserved vestiges of nerve structures belonging to segments whose entodermal and mesodermal organs have disappeared for the most part" begs the entire question. The fact that the brain shows a larger number of divisions — whether the problematical marginal headings of Locy or the secondary subdivisions of the differentiated tube — than does the mesoderm, does not prove that the nervous divisions, are ancestral or primitive. Moreover, there is no reason for assuming that mesodermic segments have disappeared in the head region of Squalus. As a matter of fact, the number of somites in Squalus corresponds with the number of primary brain vesicles. In this numerical correspondence we have the strongest proof of the metameric value of these two segmental structures. Moreover, this inference accords more fully with the conclusions


138 H. V. NEAL

of comparative anatomy than does the assumption of more numerous pre-oral segments.

Assuming on such grounds the metameric value of those primary brain vesicles or neuromeres which correspond numerically with the somites of Van Wijhe, more metameres are found anterior to the ear than are admitted in the schemes of metamerism of Ziegler and Brohmer, but fewer than in those of Johnston and Belogolowy. In figure 76 are diagrammatically expressed the primitive segmental relations based on the assumption of a correspondence of primary neuromeres and Van Wijhe's somites. That the abducens may be regarded as the somatic motor nerve of a metamere posterior to the one whose myotome it innervates seems indicated by its relations. Neuromeric relations therefore afford an important clue to the primitive metameric relations of the head. For this reason neuromeric segmentation may not be disregarded in any attempt to define the metamerism of the head.

7. Are these meta7neres serially hofnologous ivith those of the trunkf

A summary of the evidence gathered by the last two generations of vertebrate morphologists convincingly demonstrates that the vertebrate head possesses a metamerism comparable with that of the trunk. First and most important, a true somitic segmentation occurs in pre-otic and post-otic regions alike. An identical segmentation characterizes selachian (Van Wijhe '82), amphibian (Miss Piatt '97) and cyclostome fKoltzoff '01) embryos. That is to say, these craniotes pass ontogenetically through an acraniate stage, comparable to the adult form of Amphioxus.

The discovery by Koltzoff ('01-'02) that in embryos of Petromyzon a series of somites — exactly homologous with those discovered by Van "Wijhe ('82) in selachian embryos — occurs, not a single one of which disappears in ontogeny so that the muscular metamerism is unbroken as in Amphioxus, appears to the writer to be one of the most important made during the present generation. The recent rehabilitation of Amphioxus as an ancestral type by Delsman ('13) seems to justify the hope that


MORPHOLOGY OF EYE MUSCLE NERVES 139

the ancestral history of the head may yet be known and general agreement among morphologists be attained.

In both pre-otic and post-otic regions of the body the mesodermic corresponds numerically with the neuromeric segmentation. While the topographic alternation is not clear in the head as in the trunk, the relations, both nervous and numerical, indicate a primitive correspondence.

The somatic motor column continues uninterruptedly from post-otic into pre-otic regions. In both regions nervous connection with somitic musculature is effected similarly by the movement of the protoplasm of the neuroblasts lying in that column. The secondary connection of nerve and muscle in both regions affords the possibility of the acquisition of new metameric relations such as appear in the case of the abducens nerve. The relations of this post-otic nerve to pre-otic myotomes indicates that no fundamental difference distinguishes the two regions.

The somatic motor nerves acquire relations with somatic sensory nerves and with sympathetic anlagen in the pre-otic region in precisely the same manner as do spinal somatic motor nerves. The misinterpretation of these relationships has long obscured the perception of their true morphology and delayed the acceptance of the conclusion that head metameres are comparable with those of the trunk.

Typical pre-otic metameres, represented by the midbrain-oculomotor-premandibular and by the hindbrain-trochlear-mandibular segments, possess all of the essential components of typical trunk metameres, namely, myotome, sclerotome, neuromere, somatic motor and somatic sensory nerves, and sympathetic anlagen. Their morphological comparability can be doubted only by doubting facts which have been repeatedly confirmed. Neither comparative anatomy nor embryology justify the speculation that these elements are of exogenous, post-otic origin.

Over against such evidence, we have differences between preotic and post-otic regions such as would be expected in highly differentiated regions. But the considerations advanced above indicate that these differences are differences of detail and are not fundamental. They would appeal more strongly as objec


140 H. V. NEAL

tions to those not fully conversant with the embryos of selachians and cyclostomes, or those whose views are affected by some divergent view of the ancestry of vertebrates. Ontogenesis strongly favors the view that the metamerism of the vertebrate body extends throughout its length and that the metameres of the head are morphologically similar to those of the trunk.

8. The primary segmentation of the pre-otic region: Conclusions

The evidence presented in the present paper scarcely justifies the attempt to draw conclusions regarding the maximal number of pre-otic metameres. Admirable attempts in this direction have recently been made by Johnston ('05 a) and Belogolowy ('10 a). The writer is convinced, however, in the light of the evidence now at hand that we may make positive assertions as to the minimal number of pre-otic segments, and as to their essential constituent elements.

The vertebrate head anterior to the ear consists of at least five metameric divisions, diagrammatically represented in figures 73 to 76. Their more important constituents may be summarized as follows:

The most anterior pre-otic metamere contains a well defined mesodermic segment, the anterior somite of Miss Piatt TQO), the early degeneration of which is correlated with the absence of a somatic motor nerve in this metamere. The homology of this somite with the adhesive organ of Amia (Reighard '02) does not invalidate the comparison with a trunk somite, however sm*prising such a modification of a somite may appear. The absence of a sympathetic ganglion is to be expected in a metamere devoid of a motor nerve. The nervus terminalis appears to be the somatic sensory element of the segment. The writer agrees with Belogolowy ('10 a) in regarding this as the primary nerve of the olfactory apparatus, and its ganglion the primary ganglion of the olfactory nerve. ^ Burckhardt's assertion that the nervus terminalis of selachians contains motor fibers needs con 2 Brobkover has recently confirmed this opinion on the basis of observations upon Amia embryos.


MORPHOLOGY OF EYE MUSCLE NERVES 141

firmation. The neuromere of the first cephaHc metamere is the primary forebrain vesicle. None of the secondary subdivisions

of the vesicle are morphologically comparable with hindbrain neuromeres. Notwithstanding the fact that morphologists, including Dohrn, von Kupffer, Johnston, and Belogolowy, assume a larger number of segments in this region of the brain, the writer is unable to accept their conclusions as well-founded. They rest upon evidence of equivocal neuromeres, doubtful nerves, problematic microcoelic cavities, the walls of which Dohrn has called somites, upon the far-fetched homology of the eye with dorsal ganglia and other debatable grounds. The so-called thalamic nerve is simply a persistent strand of neural crest cells which never shows neuroblast nor fiber. If it have any phylogenetic significance at all, it is not to be regarded as evidence of an additional member of the series of somatic sensory nerves, but as the old cellular root of the profundus nerve, the fibers of which now enter the brain along the profundus commissure and through the Gasserian ganglion. An optic neuromere is recognized by Johnston and Belogolowy on the debatable ground of the homology of the optic vesicle with a portion of the neural crest. In view of the doubt regarding the phylogenesis of the paired and the pineal eyes and the great uncertainty whether the former were primitively dorsal or ventral, the comparison of these structures with the anlagen of ganglionic nerves and with each other appears decidedly premature.

The assumption of Hoffmann ('94) that the anlagen of the ganglionic nerves are hollow outpocketings of the neural tube is a concept rather than a percept. No one has ever seen in sections of well preserved embryos the neural crest appear as hollow outpocketings with a lumen continuous with that of the tube. The basis for such a conception as that of Hoffmann consists of the doubtful evidence of two layers of cells in the nerve anlagen. The lumen is a product of the imagination.

Moreover, it is not so certain as would be desirable for a confirmation of the hypothesis that the paired eyes were primarily dorsal structures. The lowest vertebrates — including Amphioxus in that category — show the eye as a ventral or lateral structure


142 H. V. NEAL

and not as a dorsal one (Parker '08). Granting the fact that vertebrates have existed — and still exist — with eyes near each other and the median plane, the evidence that this was the primitive relation is wanting.

In the light of the evidence now in our possession, all that may be affirmed with assurance with regard to the metamerism in the forebrain region is that in this region we have at least a single metamere, serially homologous with those of the trunk. The morphologist who goes farther than this and affirms the polymerism of the forebrain segment is skating on extremely thin ice.

The elements of the second metamere are shown in the diagrammatic cross-section represented in figure 73. The myotome is the premandibular and the neuromere the midbrain. The somatic sensory nerve is the ophthalmicus profundus and the sympathetic ganglion is the ciliary. That the ophthalmicus profundus trigemini was once an independent segmental nerve seems evinced by its relations in cyclostomes. The secondary splitting of the premandibular myotome into dorsal and ventral moieties is evidently correlated with the development of the eyeball (fig. 81). The facts do not warrant the assumption of some morphologists that the oculomotor nerve--the somatic motor nerve of this metamere — has a bimeric distribution. No one. who has made this assumption has been able to demonstrate the required two motor niduli. The premandibular somite is a single somite. The slight ventral fold in the wall of the midbrain is not sufficient evidence to establish the existence of two nem'omeres. The large size of this neuromere, as well as that of the forebrain, is correlated with the functional importance of these portions of the brain. Their later subdivisions may be best interpreted as coenogenetic.

The third metamere consists essentially of the elements shown in figures 75, 76 and 81. Its myotome is the mandibular and its neuromere the cerebellar (neuromere III), within which lies the nidulus of the trochlear nerve, which is therefore the somatic motor nerve of the segment. The trochlear nerve becomes connected with the ramus ophthalmicus superficialis trigemini, the somatic sensory nerve of the metamere. There is evidence of


MORPHOLOGY OF EYE MUSCLE NERVES 143

a transient sympathetic aniage. While the chiasma of the trochlear is an anomaly, it may be regarded as coenogenetic and its existence does not invalidate the comparison of this metamere with a trunk segment. The ramus mandibularis trigemini appears to be the splanchnic motor element of this metamere.

The fourth metamere contains the third or hyoid myotome and the fourth neuromere (second hindbrain neuromere). To this segment may be assigned as the somatic sensory nerve the major root of the trigeminal in part. Since no neural crest is proliferated from this neuromere, however, this assignment must be made with a question mark, although the major root of the trigeminal is attached to this neuromere. The neuroblasts in the somatic motor column of this neuromere do not produce a nerve. The transient nerve seen in this region in chick embryos (Belogolowy '10) may be the somatic motor nerve of this metamere which has disappeared phylogenetically. The myotome of the metamere, however, is innervated by the nerve of a postotic metamere, the abducens. An attempt has been made above to explain this anomalous relationship which does not appear to vitiate the comparison with a. trunk metamere.

The fifth and last pre-otic metamere includes the fifth neuromere and the fourth somite which is partly sub-otic, a position to which it presumably owes the loss of its myotome. To the degeneration of the myotome may be attributed the loss of the somatic motor nerve of this metamere. No sympathetic aniage develops in this segment and the somatic sensory components are also lost. But the proliferation of the cells of the facialis nerve from this neuromere justifies the inference that they once have been present in this nerve. The loss of the myotome of this and of the following somite, a loss in all probability due to the enlargement of nerve ganglia and sense organ in this region, tends to show that the preservation of the myotomes of the first, second and third somites is due to their functional relation with the eye-ball. The eye muscles are the last remnants of the lateral trunk musculature anterior to the ear. Their earlier relations with post-otic myotomes are diagrammatically expressed in figure 81.


144 H. V. NEAL


LITERATURE CITED


Ahlborn, F. 1884 a Ueber den Ursprung und Austritt der Hirnnerven von Petromyzon. Zeit. wiss. ZooL, Bd. 40.

1884 b Ueber die Segmentation des Wirbelthierkorpers. Zeit. wiss. Zool., Bd. 40.

Allis, Jr., E. p. 1889 The anatomy and development of the lateral line system in Amia calva. Jour. Morph., vol. 2.

1897 The cranial muscles and cranial and first spinal nerves in Amia calva. Jour. Morph., vol. 12.

1902 The lateral sensory canals, the eye muscles, and the peripheral distribution of 'certain of the cranial nerves of mustelus laevis. Quart. Jour. Micr. Sci., vol. 45.

Altmann, R. 1885 Bemerkungen zur Hensen'schen Hypothese von der Nervenentstehungs. Arch. Anat. u. Phys. 1885.

Anderson, H. K. 1903 Reflex pupil-dilatation by way of the cervical sympathetic nerve. Journ. Phys., vol. 30.

VON Apathy, S. 1889-1890 Nach welcher Richtung hin soil die Nervenlehre reformirt werden? Biol. Centralbl., Bd. 9.

1897 Das leitende Element des Nervensystems und seine topographische Beziehungen zu den Zellen. Mitth. Stat. Neapel., Bd. 12

1898 Ueber Neurenfibrillen und iiber ihre nervos leitende Natur. Proc. Inter. Cong. Zool., 1897.

1907 Bemerkungen zu den Ergebnissen Ramon y Cajals hinsichtlich der feineren Beschaffenheit des Nervensystems. Anat. Anz., Bd. 30.

Apolant, H. 1896 Ueber die Beziehung des Nervus oculomotorius zum Ganglion ciliare. Arch. f. mikr. Anat., Bd. 47.

Arnold, F. 1831 Der Kopftheil des vegetativen Nervensystems beim Menschen, in anatomischer und physiologischer Hinsicht bearbeitet. Dissertation, Heidelberg.

Ayers, H. 1892 Vertebrate cephalogenesis. II. A contribution to the morphology of the vertebrate ear, with a reconsideration of its functions. Jour. Morph., vol. 6.

Bach, L. 1896 Ueber die Localization der Oculomotoriuskerne. Neur. Centralbl., Bd. 15.

VON Baer, K. E. 1828 Ueber Entwicklungsgeschichte der Tiere. I. Konigsberg.

Balfour, F. M. 1877 a On the development of spinal nerves in elasmobranch fishes. Stud. Physiol. Lab. Univ. Cambridge. Also in Philos. Trans. Roy. Soc, 1875.

1877 b Development of the spinal nerves and the sympathetic nervous system. Jour. Anat. and Physiol., vol. 11.

1878 A monograph on the development of elasmobranch fishes. London. 1881 A treatise on comparative embryology. London.

Bardeen, C. R. 1898 The development of the musculature of the body-wall in the pig, including its histogenesis and its relations to the myotomes and to the skeletal and nervous apparatus. Johns Hopkins Hosp. Reports, vol. 9.


MORPHOLOGY OF EYE MUSCLE NERVES 145

Bardeen, C. R. 1903 The growth and histogenesis of the cerebro-spinal nerves

in mammals. Am. Jour. Anat., vol. 2.

1904 The bimeric distribution of the spinal nerves in elasmobranchii

and urodela. Am. Jour. Anat., vol. 3. Barker, L. F. 1899 The nervous system and its constituent neurones. New

York. Beard, J. 1884 On the segmental sense organs of the lateral line and on the

morphology of the vertebrate auditory organ. Zool. Anz., Bd. 7.

1885 The system of branchial sense organs and their associated ganglia in Ichthyopsida. A contribution to the ancestral history of vertebrates. Quart. Jour. Micr. Sci., vol. 26.

1888 Morphological studies. II. The development of the peripheral

nervous system of vertebrates. Quart Jour. Micr. Sci., vol. 29.

1892 The histogenesis of nerve. Anat. Anz., Bd. 7.

1896 The history of a transient nervous apparatus in certain Ichthyopsida. Zool. Jahrb., Abt. f. Anat., Bd. 9. Bell, C. 1830 The nervous system of the human body. Papers delivered to

the Royal Society of London. London. Belogolowy, J. 1908 Zur Entwickelung der Kopfnerven der V"ogel. Ein Bei trag zur Morphologie des Nervensystems der Wirbeltiere. Bull. Soc.

Imp. Moscou, Hft., 3 und 4.

1910 a German reprint, Moscou.

1910 b Une branchie prespiraculaire chez le triton. Sur la question

de I'origine de la bouche des vertebres. Biol. Zeitschr., Bd. 1.

1910 c Studien zur Morphologie des Nervensj'stems der Wirbeltiere.

Die Entwickelung der Nervus trigeminus Gruppe bei Reptilien. Bull.

Soc. Imp. Moscou. Bd. Beraneck, E. 1884 Recherches sur le developpement des nerfs craniens chez

les lizards. Rec. Zool. Suisse, torn. 1.

1887 a. Etude sur les replis medullaires du poulet. Rec. Zool. Suisse,

tom. 4.

1887 b Histogenese des nerfs cephaliques. Arch. sci. nat. Geneve,

tom. 17. Bernheimer, S. 1897 Ein Beitrag zur Kenntniss der Beziehungen zwischen

dem Ganglion ciliare und der Pupillarreaction. Arch. f. Ophthmol.

(Graefe's), Bd. 43. Besta, C. 1904 Richerche int. al modo cui si. stab, i rapp. mutui tra gli ele menti nerv. embr. e suUa form, del retic. int. della cellula nerv. Riv.

sper. di fren., tom. 30. Bethe, a. 1903 AUgemeine Anatomic und Physiologic des Nervensystems.

Leipzig.

1904 Der heutige Stand der Neuronlehre. Deutsch. Med. Wochen schr., Bd. 33.

1906 Bemerkungen zur Zellkettentheorie. Anat. Anz., Bd. 28.

1907 Neue Versuche liber die Regeneration der Nervenfasern. Pfliiger's Archiv, Bd. 116.

Bidder, F., tjnd von Kupffer, C. 1857 Untersuchungen ueber das Ruckenmark. Leipzig.

JOURNAL OF MORPHOLOGY, VOL. 25, NO. 1


146 H. V. NEAL

BoEKE, J. 1904 Beitrage zur Entwickelungsgeschichte der Teleostier. 2. Die Segmentirung des Kopfmesoderms, die Genese der Kopfhohlen, das Mesectoderm der Ganglienleisten und die Entwickelung der Hypophyse bei den Muranoiden. Petrus Camper, Bd. 3.

BoNSDORFF, E. J. 1852 Symbolae ad anatomiam comparatam nervorum animalium vertebratorum. I. Nervi cerebrales corvi cornicis. 2. Nervi cerebrales Gruis cinereae. Acta soc. sci. Fennicae, Bd. 3.

BovERi, Th. 1885 Beitrage zur Kenntnis der Nervenfasern. Abh. phys.-math. Kl. d. Acad. Wiss. Miinchen, Bd. 15.

Bracket, M. A. 1905 Sur I'histogenese et la signification morphol. des fibres nerveuses peripheriques. Bull. Soc. roy. sci. med. et nat. Bruxelles. 1907 Recherches sur I'ontogenese de la tete chez les amphibiens. Arch, de Biol., torn. 23.

Brauer, a. 1904 Beitrage zur Kenntniss der Entwicklung und Anatomic der Gymniophionen. IV. Die Entwickelung der beiden Trigeminus-Ganglien. Zool. Jahrb., Suppl. 7.

Braus, H. 1899 Beitrage zur Entwickelung der Muskulatur und des peripheren Nervensystems der Selachier. 1. Teil. Die metotischen Urwirbel und spinooccipitalen Nerven. Morph. Jahrb., Bd. 27.

1905 Experimentelle Beitrage zur Frage nach der Entwickelung peripherer Nerven. Anat. Anz., Bd. 26.

Brock, G. 1906 Untersuchungen tiber die Entwicklung der Neurofibrillen des

Schweine-foetus. Monatschr. f. Psych, u. Neurol., Bd. 18. Brodman, K. 1907 Bemerkungen iiber die Fibrillogenie und ihre Beziehungen

zur Myelogenic mit besondrer Beriicksichtigung des Cortex cerebri.

Neurol. Centralbl. Brohmer, p. 1909 Der Kopf eines Embryos von Chlamydoselachus vmd die

Segmentirung des Selachierkopfes. Jen. Zeit. f. Naturwiss., Bd. 44. Broman, I. 1895 Beschreibung eines menschlichen Embryos von beinahe drei

mm. Lange mit spezieller Bemerkung iiber die bei demselben befind lichen Hirnfalten. Morph. Arbeiten., Bd. 5. Budge, J. 1855 Ueber die Bewegung der Iris. Burckhardt, R. 1892 Das Centralnervensystem von Protopterus annectans.

Eine vergleichende anatomische Studie. Berlin. Burrows, M. T. 1911 The growth of tissues of the chick embryo outside the

animal body, with special reference to the nervous system. Jour.

Exp. Zool., vol. 10. Cajal, Ram6n y, S. 1894 Les nouvelles idees sur la structure du systdme ner veux chez I'homme et chez les vertebr&s. Rev. gen. sci. pures et

appliques, tom. 5.

1906 Genesis de la fibros nerviosos del embryon, observaciones contr. a la teoria catenaria. Trab. labor, inv. biol. Univ. Madrid, tom. 4.

1907 a Die histogenetischen Beweise der Neuronentheorie von His und Forel. Anat. Anz., 30.

1907 b Nouvelles observations sur revolution des neuroblasts avec quelques remarques sur I'hypothese neurogenique de Hensen-Held. Trav. d. lab. de rech. biol., Bd. 5. Also 1908, Anat. Anz., Bd. 22.


MORPHOLOGY OF EYE MUSCLE NERVES 147

Capobianco, F. 1900 Delia prima genesi delle cellule nervosa dcUa midoUa e

dei gangli spinali. Verhandl. Anat. Gesellsch., torn. 14. Pavia.

1902 Lo svilluppo della cellule nervose midoUo spinnale di polio.

Annali di Neurologia, torn. 20.

1905 Rich, ulterieures sur la genese des cellules nerveux. Ann. di

Neurol., torn. 23. Capobianco, F., e Fragnito, O. 1898 Nuove richerche su la genesi e i rapport) mutui degli elementi nervosi e nervoglici. Ann. neurol., Milano,

torn. 2-3. Carpenter, F. W. 1906 The development of the oculomotor nerve, the ciliary

ganglion, and the abducent nerve in the chick. Bull. Mus. Comp.

Zool. Harv. Coll., vol. 48. Carpenter, F. W., and Main, R. C. 1907 The migration of medullary cells

into the ventral nerve-roots of pig-embryos. Anat. Anz., Bd. 31. Chiarugi, G. 1891 Sur les myotomes et sur les nerfs de la tete posterieure et

de la region proximale du tronc dans les embryons des amphibies

anoures (resume). Arch. Ital. de Biol., torn. 15.

1894 Lo sviluppo dei nervi oculomotore e trigemello. Nota preliminare. Monit. zool. ital., tom. 5.

CoGGi, A. 1895 II cordone gangliare anteriore in Torpedo e Pristiurus. Richerche fati nel labor. Anat. Univ. Roma ed in altri labor, biolog.,

tom. 5. CoGHiLL, G. E. 1913 The primary ventral roots and somatic motor column of

Amblystoma. Jour. Comp. Neur., vol. 23. Cole, F. J. 1898 Reflections on the cranial nerves and sense organs of fishes.

Trans. Liverpool Biol. Soc, vol. 12. Corning, H. K. 1899 Ueber die Entwickelung der Kopf- und Extremitaten muskulatur bei Reptilien. Morph. Jahrb., Bd. 28.

1900 Ueber die vergleichende Anatomic der Augenmuskulatur. Morph.

Jahrb., Bd. 29. Dean, B. 1899 On the embryology of Bdellostoma stouti. Festschr. zum

70-ten Geburtstag von C. von Kupffer. Delsman, H. C. 1913 a Der Ursprung der Vertebraten. Zool., Anz., Bd. 41.

1913 b 1st das Hirnblaschen das Amphioxus dem Gehirn der Kranioten

homolog? Anat. Anz., Bd. 44. D'Erchia, F. 1894 Contributo alio studio della struttura e delle connessioni

del ganglio ciliare. 1. Sulla struttura del ganglio ciliare. Monit. zool.

Ital., tom. 5.

1895 Contributo alio studio della struttura e delle connessioni del ganglio ciliare. II. Connessioni del ganglio ciliare. Monit. zool. Ital., tom. 6.

Dixon, A. F. 1895 On the development of the branches of the fifth cranial nerve in man (abstract). Proc. Roy. Soc. London, vol. 57.

1896 Full account of above paper. Trans. Roy. Soc. Dublin, vol. 6. DoHRN, A. 1875 Der Ursprung der Wirbelthiere und das Princip des Functions wechsels. Genealogische Skizzen. Leipzig.

1885 Studien zur Urgeschichte des Wirbelthierkorpers. X. Zur Phy logenese des Wirbelthierauges. Mitt. Zool. Stat. Neapel, Bd. 6.


148 H. V. NEAL

DoHRN, A. 1888 a Studien, u. s. w. 13. Ueber Nerven und Gefasse bei Ammocoetes und Petromyzon Planeri. Mitt. Zool. Stat. Neapel., Bd. 8. 1888 b Studien, u. s. w. 14. Ueber die erste Anlage und Entwicklung der motorischen Riickenmarksnerven bei den Selachiern., Bd. 8. 1890 a Bemerkungen fiber den neuesten Versuch einer Losung des Wirbelthierkopf-Problems. Anat. Anz., Bd. 5.

1890 b Studien, u. s. w. 15. Neue Grundlagen zur Beurtheilung der Metamerie des Kopfes. Mitt. Zool. Stat. Xeapel, Bd. 9.

1891 a Studien, u. s. w. 16. Ueber die er.ste Anlage und Entwicklung der Augenmuskelnerven bei Selachiern und das Einwandern von Medullarzellen in die motorischen Xerven. Mitt. Zool. Stat. Neapel., Bd. 10.

1891 b Studien u. s. w. 17. Xervenfaser und Ganglienzelle. Histogenetische Untersuchungen. Mitt. Zool. Stat. Xeapel., Bd. 10.

1892 Die Schwann'schen Kerne der Selachierembryonen. Anat. Anz., Bd. 7.

1901 Studien, u. s. w. 18-21. Die Schwann'schen Kerne, ihre Her kunft und Bedeutung. ^litt. Zool. Stat. Xeapel, Bd. 15.

1904 a Studien, u. s. w. 23. Die Mandibularhohle der Selachier. Mitt.

Zool. Stat. Xeapel, Bd. 17.

1904 b Studien, u. s. w. 24 Die Praemandibularhohle. Mitt. Zool.

Stat. Xeapel, Bd. 17.

1907 Studien, u. s. w. 25. Der Trochlearis. Mitt. Zool. Stat. Xeapel,

Bd. 18. Dunn, Elizabeth H. 1902 On the number and relation between diameter and

distribution of the nerve fibers innervating the leg of the frog. Jour.

Comp. X'eur., vol. 12. Edgeworth, F. H. 1907 The development of the head muscles in Gallus do mesticus, and the morphology of the head muscles in the sauropsida.

Quart. Journ. Micr. Sci., vol. 51. Edixger, L. 1896 Vorlesungen ueber den Bau der nervosen Centralorgane des

Menschen und der Thiere. Leipzig. Emmert, J. 1900 Beitrage zur Entwickelungsgeschichte der Selachier, insbe sondere nach Untersuchungen an jungeren Embryonen von Torpedo

mannorata. Arch. f. mikr. Anat., Bd. 56. EwART, J. C. 1890 On the development of the ciliary or motoroculi ganglion.

Proc. Roy. Soc. London, vol. 47. Eycleshymer, a. C. 1895 The early development of Ambh^stoma, with observations on some other vertebrates. Jour. Morph., vol. 10. FiLATOFF, D. 1907 Die Metamerie des Kopfes von Emys lutaria. Zur Frage

liber die korrelative Entwicklung. Morph. Jahrb., Bd. 38. Fragnito, O. 1899 La cellula nervosa rappresenta un'unita embry.? Congr.

della Soc. ital. fren. Xapoli. Fritz, K. W. 1899 L^ntersuchungen iiber das Ganglion ciliare. Diss. Marburg. Froriep, A. 1885 a Zur Entwickelungsgeschichte der Kopfnerven. Verh. Anat

Ges. Munchen.

1885 b Ueber Anlagen von Sinnesorganen am Facialis, Glossopharyn geus und Vagus, iiber die genetische Stellung des Vagus zum' Hypo glossus, und iiber die Herkunft der Zungenmuskulatur. Arch. Anat.

u. Physiol., Anat. Abth.


MORPHOLOGY OF EYE MUSCLE NERVES 149

Froriep, a. 1891 Zur Entwickelungsgeschichte tier Kopfnerven. Verh. Anat. Gesellsch. Munchen.

1892 a Entwickelungsgeschichte des Kopfes. Ergebnisse Anat. u. Entw'g., Bd. 1.

1892 b Zur Frage der sogenannten Neuromerie. Verhdl. Anat. Gesellsch. VVien.

1894 Entwickelungsgeschichte des Kopfes. Ergeb. Anat. u. Entwickl., Bd. 3.

1901 Ueber die Ganglienleisten des Kopfes und des Rumpfes und ihre Kreutzung in der Occipitalregion. Arch. f. Anat. u. Physiol., Anat. Abth.

1902 a Zur Entwickelungsgeschichte des Wirbelthierkopfes. Verhandl. Anat, Gesellsch. Halle. •

1902 b Einige Bemerkungen zur Kopffrage. Anat. Anz., Bd. 21.

1907 Ueber Entwicklung und Bau des autonomen Nervensystems. Med. naturwiss. Archiv, Bd. 1.

1908 Ueber die elemente des peripherischen Nervensystems und ihre Entwickelung, in Edinger's Vorlesungen liber den Bau der nervosen Centralorgane, Bd. 2, Leipzig.

FtJBBRiNGER, M. 1897 Ucber die spino-occipitalen Nerven der Selachier und Holocephalen und ihre vergleichende Morphologie. Festschr. zum 70-ten Geburtstag. von Carl Gegenbaur, Bd. 3.

1902 Morphologische Streitfragen. 1. Nervus trochlearis. 2. Rabl's Methode und Behandlung der Extremitaten-frage. Morph. Jahrb., Bd. 30.

FxJSARi, R. 1890 Contribution a I'etude du developpement des capsules surrenales et du sympathetique chez le poulet et chez les mammiferes. Archivio per le scienze mediche, torn. 16.

Gaskell, W. H. 1899 On the relation between the structure, function, distribution and origin of the cranial nerves; together with a theory of the origin of the nervous system of Vertebrata. Jour. Physiol., vol. 10.

Gast, R. 1909 Die Entwickelung des Oculomotorius und seiner Ganglien bei Selachierembryonen. Mitt. Zool. Stat. Neapel, Bd. 19.

Gegenbaur, C. 1871 Ueber die Kopfnerven von Hexanchus. Jen. Zeitschr. Natwiss., Bd. 6.

1872 Untersuchungen zur vergleichenden Anatomie der Wirbelthiere. 3. Das Kopfskelett der Selachier. Ein Beitrag zur Erkenntniss der Genese des Kopfskelettes der Wirbelthiere. Leipzig. 1887 Die Metamerie des Kopfes und die Wirbeltheorie des Kopfskelettes. Morph. Jahrb., Bd. 13.

1898 Vergleichende Anatomie der Wirbelthiere mit Beriicksichtigung der Wirbellosen. Leipzig.

Gierlich, N. 1906 Ueber die Entwickelung der Neurofibrillen in der Pyramidenbahn des Menschen. Neurol. Centralbl.

1907 a Kurze Bemerkungen iiber Fibrillogenie im Zentralnervensystem des Menschen zur Arbeit Brodmann's 'Bemerkungen iiber die fibrillogenie und ihre Beziehungen zur Myelogenie mit besonderer Beriicksichtigung des Cortex cerebri.' Neurol. Centralbl.


150 H. V. NEAL

GiERLiCH, N. 1907 b Ueber das verschiedene Verhalten der Neurofibrillen in den

Fortsatzen und ini Zelleib der motorischen Ganglienzellen. Xeiirol.

Centralbl. GiERLiCH, N., UND Hi^RZHEiMER, G. 1907 Studicn iiber Neurofibrillen im Zen tralnervensystem. Wiesbaden. GiGLio-Tos, E. 1902 Sulla cellule germinative del tubo midoUare embryonale

deU'uomo. Anat. Anz., Bd. 20. GoETTE, A. 1875 Die Entwickelungsgeschichte der Unke (Bombinator igneus)

als Grundlage eine vergleichenden Morphologie der Wirbelthiere.

Leipzig.

1888 Ueber die Entwicklung von Petromyzon fluviatilis. Zool. Anz.,

Bd. 11.

1890 Entwickelungsgeschichte des Flussneunauges (Petromyzon fluviatilis). Hamburg und Leipzig. Goldberg, M. 1891 Ueber die Entwickelung der Ganglien beim Hiihnchen.

Arch. f. Mikr. Anat., Bd. 37. GoLGi, C. 1893 Intorno all'origine del quarto cerebralo (patetico o trocleare)

e di una questione di isto-fisiologia generate che a questa argumentosi

collega. Atti Accad. Lincei, Rendiconti, tom. 3. Also in French;

Arch. Ital. Biol., tom. 19. GoLowiNE, E. 1890 Sur le developpement du systeme ganglionnaire chez le

poulet. Anat. Anz., Bd. 5. GoRONOWiTSCH, N. 1888 Das Gehirn und die Cranialnerven von Acipenser

ruthenus. Ein Beitrag zur Morphologie des Wirbelthierkopfes. Morph.

Jahrb., Bd. 13.

1892 Die axiale und die laterale (A. Goette) Kopfmetamerien der

Vogelembryonen. Die Rolle der sog. Ganglienleisten im Aufbaue der

Nervenstamme. Anat. Anz., 7.

1898 Untersuchungen iiber die erste Anlage der Cranialnerven bei

Salmo fario. Nouveaux Mem. Soc. Imp. Natural. Moscou, Bd. 16. Graper, L. 1913 Die Rhombomeren und ihre Nervenberichungen. Arch.

mikr. Anat. Bd. 83. Griggs, L. 1910 Early stages in the development of the central nervous system of Amblj'stoma punctatum. Jour. Morph., vol. 21. Gurwitsch, A. 1900 Die Histogenese der Schwann'schen Scheide. Arch. f.

Anat. u. Physiol., Anat. Abth. GuTHKE, E. 1906 Embryologische Studien iiber die Ganglien und Nerven des

Kopfes von Torpedo ocellata. Jen. Zeitschr., Bd. 42. Haller, B. 1898 Vom Bau des Wirbelthiergehirns. I Theil. Salmo und Scyl lium. Morph Jahrb., Bd. 26. Harrison, R. G. 1901 a The histogenesis of the peripheral nervous system in

Salmo salar. Biol. Bull., vol. 2.

1901 b Ueber die Histogenese des peripheren Nervensystems bei Sahno

salar. Arch. mikr. Anat., Bd. 57.

1904 Neue Versuche und Beobachtungen iiber die Entwickelung der

peripheren Nerven der Wirbeltiere. Stizungsber. Niederrhein. Ges.

Bonn.


MORPHOLOGY OF EYE MUSCLE NERVES 151

Harrison, R. G. 1906 Further experiments on the development of peripheral nerves. Am. Jour. Anat., vol. 5.

1907 a Experiments in transplanting limbs and their bearing upon the problems of the development of nerves. Jour. Exp. Zool., vol. 4. 1907 b Observations upon the living developing nerve fiber. Amer. Jour. Anat., vol. 7.

1910 The development of peripheral nerve fibers in altered surroundings. Arch. f. Entwickelungsmech. d. Organismen., Bd. 30.

1911 The outgrowth of the nerve fiber as a mode of protoplasmic movement. Jour. Exp. Zool., vol. 9.

Hatschek, B. 1881 Studien tiber Entwicklung des Amphioxus. Arb. Zool. Inst. Wien, Bd. 4.

1892 Die Metamerie des .Amphioxus uiid des Ammocoetes. Verh. Anat. Gesellsch., Wien.

1893 Zur Metamerie der Wirbelthiere. Anat. Anz., Bd. 8.

Held, H. 1897 Beitrage zur Structur der Nervenzellen und ihre Fortsatze. Arch. f. Anat. u. Physiol., Anat. Abth.

1904 Zur weiteren Kenntniss der Nerven, u. s. w. Abh. d. Kgl. sachs. Ges. d. W., Bd. 29.

1906 Zur Histogenese der Nervenleitung. Verh. Anat. Ges., Rostock.

1907 a Kritische Bemerkungen zu der Verteidigung der Neuroblastenund Neuronenlehre durch R. Cajal. Anat. Anz.

1907 b Ueber den Begriff der Wirbeltierganglienzelle. Verh. der. Ges deutsch. Naturf. und Aerzte, Dresden.

1907 c Zur Kenntnis der marg. Neuroglia. Verh. d. Ges. deutsch. Naturf. u. Aerzte, Dresden.

1909 Die Entwicklung des Nervengewebes bei den Wirbeltieren. Leipzig. Hensen, V. 1864 Zur Entwickelung des Nervensy stems. Arch. f. path. Anat. u. Physiol., Bd. 30.

1876 Beobachtungen iiber die Befruchtung und Entwicklung des Kaninchens und des Meerschweinschens. Zeit. Anat. u. Entwickl., Bd. 1.

1903 Die Entwicklungsmechanik der Nervenbahnen im Embryo der Saugethiere. Kiel u. Leipzig.

1908 Ueber das Auswachsen der Nerven im Embryo. Miinchen. Med. Woch., Bd. 55.

Hensen, V., und Volkers, C. 1868 Experimentaluntersuchungen iiber den

Mechanismus der Accommodation. Kiel. Herms, E. 1884 Ueber die Bildungsweise der Ganglienzellen im Ursprungsge biete des Nervus acustico-facialis bei Ammocoetes. Sitzb. Miinchen.

Akad., Math.-phys. 01., Bd. 14. Herrick, C. J. 1899 The cranial and first spinal nerves of Menidia: A contribution upon the nerve components of the bony fishes. Jour. Comp.

Neur., vol. 9.

1903 The doctrine of nerve components and some of .its applications.

Jour. Comp. Neur., vol. 13.


152 H. V. NEAL

Herrick, C. J. 1909 The criteria of homology in the peripheral nervous system.

Jour. Comp. Neur., vol. 19. Herrick, C. L. 1892 Embryological notes on the brain of the snake. Jour.

Comp. Neur., vol. 2.

1893 The development of meduUated nerve-fibers. Jour. Comp. Neur., vol. 3.

Hertwig, O. uxd R. 1878 Das Nervensystem und die Sinnesorgane der Medusen, Monographisch dargestellt. Leipzig.

Heymans, J. F., UND Van der Stricht, O. 1898 Sur le systeme nerveux de I'Amphioxus et en particulier sur la constitution et la gcnese des racines sensibles. Mem. cour. par. 1, Acad. Roy. Sci. Belgium, Bd. 56.

His, W. 1879 Ueber die Anfange des peripherischen Nervensystems. Arch. f. Anat. u. Physiol., Anat. Abth.

1888 a Die Entwickelung des ersten Nervenbahnen beim menschlichen Embryo. Uebersichtliche Darstellung. Arch. f. Anat. Physiol., Anat. Abth., Bd. 6.

1888 b Die morphologische Betrachtung der Kopfncrven. Eine kritische Studie. Arch. f. Anat. u. Physiol., Anat. Abth., Bd. 6.

1888 c Zur Geschichte des Gehirns, sowie der centralen und peripherischen Nervenbahnen beim menschlichen Embryo. Abh. sachs. Gesellsch. Wissensch., Math.-phys. CI. Bd. 4.

1889 Die Neuroblasten und deren Entstehung im embryonalen Mark. Abh. sachs. Gesell. Wiss. Leipzig., Math.-phys. CI., Bd. 1.5. Also in Arch. f. Anat. u. Physiol., Anat. Abth. *

1890 Histogenese und Zusammenhang der Nervenelemente. Arch. Anat. u. Physiol., Anat. Abth., Suppl.

His, Jr., W. 1897 Ueber die Entwickelung des Bauchsjinpathicus beim Hi'mchen und Menschen. Arch. Anat. u. Physiol., Anat. Abth., Suppl.

Hi.s, Jr., W., and Romberg, E. 1890 Beitriige zur Hirninnervation. Fortschr. d. Medicin., Bd. 8.

1891 Die Entwickelung des Herznervensystems bei Wirbelthieren. Abhandl. Math.-Phys. Classe Konig. sachs. Gesellsch. Wiss., Bd. 28.

Hoffman, C. K. 1885 Weitere Untersuchungen zur Entwicklungsgeschichte der Reptilien. Morph. Jahrb., Bd. 11.

1889 Ueber die Metamerie des Nachhirns und Hinterhirns, und ihre Beziehung zu den segmentalen Kopfnerven bei Reptilienembryonen. Zool. Anz., Bd. 12.

1894 Zur Entwickelungsgeschichte des Selachierkopfes. Anat. Anz., Bd. 9.

1896 Beitriige zur Entwickelungsgeschichte der Selachii. Morph. Jahrb., Bd. 24.

1897 Beitrage zur Entwickelungsgeschichte der Selachii. Morph. Jahrb., Bd. 25.

1899 Beitrage zur Entwickelungsgeschichte der Selachii. Morph. Jahrb., Bd. 27.

1900 Zur Entwickelungsgeschichte des Sympathicus. Die Entwickl. des Sympathicus bei den Selachiern. Verh. Akad. Amsterdam, Bd. 7.


MORPHOLOGY OF EYE MUSCLE NERVES 153

Hoffman, C. K. 1902 Zur Entwickl. des Sympathicus. II. Die Entwicklung des Sympathicus bei den Urodelen. Verhdl. d. Konigl.. Akad. v. Wetensch. te Amsterdam, Bd. 8.

HoLTZMANN, H. 1896 Untersuchungen liber Ciliarganglion und Ciliarnerven. Morph. Arbeit., Bd. 6.

HousEB, G. L. 1901 The neurones and the supporting elements of the brain of a selachian. Jour. Comp. Neur., vol. 11.

HoussAY, F. 1890 Etudes d'embryologie sur les Vertebres. Archiv de Zool. exp. et gen., torn. 8.

HuBER, G. C. 1897 Lectures on the sympathetic nervous system. Jour. Comp. Neur., vol: 7.

1899 A contribution on the minute anatomy of the sympathetic ganglia of the different classes of vertebrates. Jour. Morph., vol. 16.

Huxley, T. H. 1858 The Croonian lecture: On the theorj' of the vertebrate skull. Proc. Roy. Soc. London., vol. 9.

Jegorow, J. 1886-1887 Recherches anatomo-physiologiques sur le ganglion ophthalmique. Archir slav. biol., tom. 2, 3.

Johnson, Alice, and Sheldon, Lillian 1886 On the development of the cranial nerves of the newt. Proc. Roy. Soc. London, vol. 40.

Johnson, C. E. 1913 The development of the prootic head somites and eye muscles in Chelydra serpentina. Amer. Jour. Anat., vol. 14.

Johnston, J. B. 1902 a An attempt to define the primitive functional divisions of the central nervous system. Jour. Comp. Neur., vol. 12. 1902 b Das Gehirn und die Cranialnerven der Anamnier. Ergebnis. d. Anat. u. Entwg., Bd. 11.

1905 a The morphology of the vertebrate head from the viewpoint of the functional divisions of the nervous system. Jour. Comp. Neur., vol. 15.

1905 b The cranial nerve components of Petromyzon. Morph. Jahrb., Bd. 34.

1906 The nervous system of vertebrates. Philadelphia.

1908 Additional notes on the cranial nerves of Petromyzonts. Jour. Comp. Neur., vol. 18.

1909 The morphology of the forebrain vesicle in vertebrates. Jour. Comp. Neur., vol. 19.

Jones, W. C. 1905 Notes on the development of the sympathetic nervous system in the common toad. Jour. Comp. Neur., vol. 15.

JoRis, H. 1904 Histogenese du neurone. Bull, de I'acad. royale de Med. de Belgique.

JuiiN, Ch. 1887 Le systeme nerveux grand sympathetique de I'ammocoetes (Petromyzon pi.). Anat. Anz., Bd. 2.

Kallius, E. 1900 Ergebniss. Ergeb. d. Anat. u. Entwckl., Bd. 9.

Kappers, C. U. a. 1904 Recherches sur le developpement des gaines dans le tube nerveux. Petrus Camper, tom. 2.

1910 The migrations of the abducens nucleus and the concomitating changes of its root-fibers. Psych, en Neurol. Bladen, Bd. 4.

Kastschenko, N. 1888 Zur Entwicklungsgeschichte des Selachierembryos. Anat. Anz., Bd. 3.


154 H. V. NEAL

Kerr, Graham J. 1902 The early development of muscles and motor nerves in Lepidosiren. Rep. Brit. Assoc. Adv. Sci., Belfast.

1903 The development of Lepidosiren paradoxa. 3. Development of the skin and its derivatives. Quart. Jour. Micr. Sci., vol. 46.

1904 On some points in the early development of motor nerve trunks and myotomes in Lepidosiren paradoxa (Fitz.). Trans. R. Soc, Edinburgh, vol. 41.

1910 The development of the peripheral nerves of vertebrates. Presidential address. Proc. Roy. Soc. Phys., Edinburgh, vol. 18.

KiLLiAN, G. 1891 Zur Metamerie des Selachierkopfes. Verh. Anat. Gesell. Miinchen.

KiNGSLEY, J. S. 1897 The segmentation of the head. Biol. Lect., Marine Biol. Lab., Woods Hole, 1895. 1912 Comparative anatomy of vertebrates. Philadelphia.

Klinkhardt, W. 1905 Beitrage zur Entwickelungsgeschichte der Kopfganglien und Sinneslinien bei Selachiern. Jena. Zeit. f. Naturwiss., Bd. 40.

VON KoLLiKER, A. 1879 Entwicklungsgeschichte des Menschen und der hoheren Thiere. Leipzig.

1885 Histologische Studien an Batrachierlarven. Zeit. wiss. Zool., Bd. 43.

1892 Ueber die Entwickelung der Elements des Nervensystems, contra Beard und Dohrn. Anat. Anz., Bd. 7, Erganzungsh. 1894 a Ueber die feinere Anatomie und die physiologische Bedeutung des sympathischen Nervensystems. Verh. Gesell. deutsch. Naturf. u. Aerz, Versam, Bd. 1.

1894 b Ueber das Vorkommen von Nervenzellen in den vorderen Wiirzeln der Riickenmarksnerven der Katze. Verh. Gesel. deutsch. Naturf. u. Aerz., Bd. 2.

1896 Handbuch der Gewerbelehre des Menschen. Leipzig. 1900 Gegen die Entstehung von Nervenfasern aus Zellenstrangen. Anat. Anz., Bd. 18.

1905 Die Entwickelung der Elemente des Nervensystems. Zeitschr. f. wiss. Zool., Bd. 82.

KoHN, A. 1905 Ueber die Entwickelung des peripheren Nervensystems. Verh. Anat. Ges., Jena.

1906 Ganglienzelle und Nervenfaser. Munch, med. Wochenschr.

1907 Ueber die Entwickelung des sympathischen Nervensystems der Saugethiere. Arch. f. mikr. Anat., Bd. 70.

KoLSTER, R. 1899 Beitrage zur Kenntniss der Histogenese der peripheren

Nerven nebst Bemerkungen liber die Regeneration derselben nach

Vorletzungen. Beitr. z. path. Anat., Bd. 26. KoLTZOFF, N. K. 1901 Entwicklungsgeschichte des Kopfes von Petromyzon

planeri. Ein Beitrag zur Lehre iiber Metamerie des Wirbeltierkopfes.

Bull. Soc. Nat. Moscou, Bd., 15; German trans. 1902. Kraxjse, W. 1881 Ueber die Doppelnatur des Ganglion ciliare. Morph. Jahrb.,

Bd. 7. KuHL, H. UND VON Hasselt. 1820 Beitrag zur Kenntnis des Kopfes der

Knorpelfische. Kuhl's Beitr. z. Zool. u. vergl. Anat. Bd. 2.


MORPHOLOGY OF EYE MUSCLE NERVES 155

KuNTZ, A. 1910 The development of the sympathetic nervous system in mammals. Jour. Comp. Neur., vol. 20.

1911 The development of the sympathetic nervous system in certain fishes. Jour. Comp. Neur., vol. 21.

VON KuPFFER, C. 1885 Primare Metamerie des Neuralrohrs der Vertebraten. Sitzungsber. Math, physik. Kl. Miinchen.

1890 DieEntwickelung vonPetromyzonplaneri. Arch. f. mikr. Anat., Bd. 35.

1891 Die Entwickelung der Kopfnerven der Vertebraten. Verh. Anat. Gesell. Miinchen.

1894 Studien zur vergleichenden Entwickelungsgeschichte des Kopfes der Kranioten. 2. Die Entwicklung des Kopfes von Ammocoetes planeri. Miinchen und Leipzig.

1895 Studien zur vergleichenden Entwicklungsgeschichte u. s. w. 3. Die Entwicklung der Kopfnerven von Ammocoetes planeri. Miinchen und Leipzig.

1906 Die Morphogenie des Centralnervensy stems. Handbuch vergl.

u. exp. Entwicklungslehre d. Wirbelthiere, Bd. 2. Lamb, A. B. 1902 The development of the eye muscles in Acanthias. Am. Jour.

Anat., vol. 1. Reprint in Tufts Coll. Stud., no. 7. Langendorff, O. 1894 Ciliarganglion und Oculomotorius. Arch. f. ges. Physiol., Bd. 56. Langley, J. N., AND Anderson, H. K. 1892 The action of nicotin on the ciliary

ganglion and on the endings of the third cranial nerve. Jour. Physiol.,

vol. 13. Legros, R. 1898 Developpement de la cavite buccale de I'Amphioxus lanceo latus. Contribution a I'etude dela morphologie de la tSte. Arch.

Anat. Micr., torn. 1 et 2. VON Lenhossek, M. 1890 Ueber Nervenfasern in den hinteren Wlirzeln, welche

aus dem Vorderhorn entspringen. Anat. Anz., Bd. 5.

1891 Zur Kenntnis der ersten Entstehung der Nervenzellen und Nervenfasern beim Vogelembryo. Verh. 10 Inter, med. Kong., Bd. 2.

1892 Beobachtungen an den Spinalganglien und dem Riickenmark von Pristiurusembryonen. Anat. Anz., Bd. 7.

1897 Nervensystem. Ergeb. Anat. u. Entwickl., Bd. 7.

1906 Frage nach der Entwickelung der peripherischen Nervenfasern. Anat. Anz., Bd. 28.

Lewis, W. H. 1906 Experimental evidence in support of the outgrowth theory of the axis cylinder. Proc. Am. Ass. Anat., Ann Arbor, 1905. Am. Jour. Anat., vol. 5.

1907 Experimental evidence in support of the theory of outgrowth of the axis cylinder. Amer. Jour. Anat., vol. 6.

LiLLiE, F. R. 1908 The development of the chick. New York.

LocY, W. A. 1895 Contribution to the structure and development of the vertebrate head. Jour. Morph., 11. 1904 A new cranial nerve in selachians. Mark Anniv. Vol.

London und Pesker 1906 Ueber die Entwicklung des peripheren Nervensystems bei Saugetieren (weissen Mausen). Arch. f. mikr. Anat., Bd. 67.


156 H. V. NEAL

Marcus, H. 1909 Ueber den Sympathicus. Sitzungsber. Gesell. Morph. u.

Physiol. Munchen. Marina, A. 1898 II neurone del ganglio ciliare ed i centri del movimenti pupil lare. Riv. di patolog. nerv. e ment., torn. 3.

1899 Das Neuron des Ganglionciliare und die Centra der Pupillen bewegungen. Eine experimentelle Studie. Deutsch. Zeitschr. f. • Nervenheilk., Bd. 14. Marshall, A. M. 1878 The development of the cranial nerves of the chick.

Quart. Jour. Micr. Sci., vol. 18.

1881 On the head cavities and associated nerves of elasmobranchs. Quart. Jour. Micr. Sci., vol. 21.

1882 The segmental value of the cranial nerves. Jour. Anat. Physiol., vol. 16. Also, Stud. Biol. Lab. Owens Coll., vol. 1, 1886.

Marshall, A. M., and Spencer, W. B. 1881 Observations on the cranial nerves

of Scyllium. Quart. Jour. Micr. Sci., vol. 21. Also in Stud. Biol. Lab.

Owens Coll., vol. 1, 1886. Martin, P. 1890 Die erste Entwickelung der Kopfnerven der Katze. Oesterr.

Monatsschr. Thierheilkunde, Bd. 15.

1891 Die Entwickelung der neunten bis zwolften Kopfnerven bei der

Katze. Anat. Anz., Bd. 6. McClure, C. F. W. 1882 The primitive segmentation of the vertebrate brain.

Zool. Anz., Bd. 12.

1890 The segmentation of the primitive vertebrate brain. Jour.

Morph., vol. 4. McMtjrrich, J. Playfair 1912 The problem of the vertebrate head in the light

of comparative anatomy. Proc. 7th Internat. Cong., Boston, 1907. Meek, A. 1909 The encephalomeres and cranial nerves of an embryo of Acan thias vulgaris. Anat. Anz., Bd. 34.

1910 The cranial segments and nerves of the rabbit with some remarks

on the phylogeny of the nervous system. Anat. Anz., Bd. 36. VON MiHALKOwiTZ, V. 1877 Eutwickeluugsgeschichte des Gehirns. Leipzig. von Miklucho-Maclay, N. 1868 Beitrag zur vergleichenden Anatomic des

Gehirns. Jenaische Zeitschr. f. Naturwiss., Bd. 4. MiNOT, C. S. 1892 Human embryology. New York.

1896 Die friihen Stadien und die Histogenese des Ner\-ensystems. Ergeb. Anat. u. Entwickg., Bd. 6.

1897 Cephalic homologies. A contribution to the determination of the ancestry of vertebrates. Amer. Nat., vol. 31.

1910 Laboratory text-book of embryology. Philadelphia. MiTROPHANOW, P. 1892 Note sur la signification metamerique des nerfs cra niens. Congres Internat. Zool. Muscou, Bd. 1.

1893 Etude embryogenique sur les selachiens. Arch. Zool. Exper.,

tom. 1. MoNCKEBERG, G., UND Bethe, A. 1899 Die Degeneration der markhaltigen

Nervenfasern der Wirbeltiere unter hauptsachlicher Beriicksichtigung

des Verhaltens der Primitiv-fibrillen. Arch. micr. Anat., Bd. 54. Nawrocki, F., UND Przybylski, J. 1891 Die pupillenerweiternden Nerven der

Katze. Arch. f. ges. Physiol., Bd. 50.


MORPHOLOGY OF EYE MUSCLE NERVES 157

Neal, H. V. 1896 A summary of studies of the segmentation of the nervous system of Squalus acanthias. Anat. Anz., Bd. 12.

1897 The development of the hypoglossus musculature in Petromyzon and Squalus. Anat. Anz., Bd. 13.

1898 a The segmentation of the nervous system in Squalus acanthias. A contribution to the morphology of the vertebrate head. Bull. Mus. Comp. Zool. Harvard Coll., vol. 31.

1898 b The problem of the vertebrate head. Jour. Comp. Neur.,

vol. 8.

1903 The development of ventral nerves in Selachii. I. Spinal ventral nerves. Mark Anniv. Vol.

1912 The morphology of the eye muscle nerves. Proc. 7th Internat.

Cong., Boston, 1907; Separate pub. in 1909. Neumayer, L. 1906 Histo- u. Morphogenese des peripheren Nervensystems

der Spinalganglien und des Nervus sympathicus. Handbuch d. vergl.

u. exp. Entwickelungslehre der Wirbeltiere, Bd. 2. Nicolas, A. 1900 Note sur la presence de fibres musculaires strias dans la

gland pineale de quelques mammiferes. Compt. Rend. Soc. Biol.,

tom. 52. NoRRis, H. W. 1891 Recent studies of the vertebrate head. Amer. Nat., vol.

25. Onodi, a. D. 1884 Ueber die Entwickelung der Spinalganglien und der Ner venwiirzeln. Internat. Monatschr. f. Anat. u. Histol., Bd. 1.

1886 Uber die Entwicklung des sympathischen Nervensystems.

Arch. f. mikr. Anat., Bd. 26. Oppel, a. 1890 Ueber Vorderkopfsomiten und die Kopfhohle bei Anguis fra gilis. Arch. f. mikr. Anat., Bd. 36. Orr, H. B. 1887 Contribution to the embryology of the lizard. Jour. Morph.,

vol. 1. Osborn, H. F. 1888 A contribution to the internal structure of the amphibian

brain. Jour. Morph., vol. 2. OsTROUMOFF, A. 1889 Regio occipitalis. Eine morphologische Studie. Arb.

Naturf. Ges. an d. Univ. Kasan., Bd. 21. Parker, G. H. 1900 The neurone theory in the light of recent discoveries.

Amer. Nat., vol. 34.

1908 The origin of the lateral eyes of vertebrates. Amer. Nat.,

vol. 42.

1910 The phylogenetic origin of the nervous system. Anat. Rec,

vol. 4. Paton, S. 1907 The reactions of the vertebrate embryo to stimulation and the

associated changes in the nervous system. Mitt. Zool. Stat. Neapel,

Bd. 18. Paterson, a. j\I. 1890 Development of the sympathetic nervous system in

mammals. Phil. Trans. Roy. Soc, vol. 181.

1898 Preliminary note on the development of the sympathetic system

in elasmobranchs. Brit. Med. Jour. Patten, W. 1912 The evolution of the vertebrates and their kin. Philadelphia.


158 H. V. NEAL

La Pegxa, E. 1904 Su la genesi ed i rapp. recipr. de-gle elementc nervose nel

midollo spinali di polio. Ann. di Neurol., torn. 22.

1905 Sulla trasformazione delle radice spinale e suUa prima com parsa di fibrille nelle cellule del midollo. Riv. spar, di freniatria,

torn. 31. PiGHiNi, G. 1904 Sulla sviluppo della fibre nervose periferichi e centrali spinali

e ganglione cefalici nell'embrione dell polio. Riv. sper. di freniatria,

torn. 30.

1905 Sur I'origine et la formation des cellules nerveuses chez les em bryons de selaciens. Bibl. anat., tom. 14. PiNKUs, F. 1894 Ueber einen noch nicht beschriebenen Hirnnerven des Pro topterus annectans. Anat. Anz., Bd. 9. PiTZORNo, M. 1910 Sulla stnittura dei gangli simpatici nei Selaci. Monit.

Zool. Ital., tom. 21. Platt, Julia B. 1890 The anterior head-cavities of Acanthias. Zool. Anz.,

Bd. 13.

1891 A contribution to the morphology of the vertebrate head, based

on a study of Acanthias vulgaris. Jour. Morph., vol. 5.

1891 b Further contribution to the morphology of the vertebrate head. Anat. Anz., Bd. 6.

1896 Ontogenetic differentiations of the ectoderm in Necturus. 2. On the development of the peripheral nervous system. Quart. Jour. Micr. Sci., vol. 38.

1897 The development of the cartilaginous skull and of the branchial and hypoglossal musculature in Necturus. Morph. Jahrb., Bd. 25.

Pollard, H. B. 1891 On the anatomy and phylogenetic position of Polypterus.

Zool. Jahrb., Bd. 5. Prenant, a. 1889 Note sur I'existence des replis meduUaires chez I'embryon

du pore. Bull. Soc. Sci. Nancy, tom. 9. Prentiss, C. W. 190^ The nervous structures in the palate of the frog; the

peripheral networks and the nature of their cells and fibers. Jour.

Comp. Neur., vol. 14. Rabl, C. 1885 Bemerkung iiber die Segmentirung des Hirnes. Zool. Anz.,

Bd. 8.

1889 a Theorie des Mesoderms. Morph. Jahrb., Bd. 15.

1889 b Ueber die Prinzipien der Histologie. Verh. d. Anat. Gesellsch.

Berlin.

1892 Ueber die Metamerie des Wirbeltierkopfes. Verh. Anat. Gesellsch. Wien.

1901 Gedanken und Studien iiber den Ursprung der Extremitaten.

Zeit. f. wiss. Zool., Bd. 70. Rai'faele, F. 1900 Per la genese dei nervi da catene cellulari. Anat. Anz.,

Bd. 18. Ra.vsom, W. B., and Thompson, D'A. W. 1886 On the spinal and visceral nerves

of cyclostomata. Zool. Anz., Bd. 9. vOiM Rath, O. 1895 Zur Conservirungstechnik. Anat. Anz., Bd. 11. Reighard, J. 1902 On the anterior head-cavity of the elasmobranchs. Rep.

Mich. Acad. Sci., 1901.


MOKPHOLOGY OF EYE MUSCLE NERVES 159

Remak, R. 1837 Weitere mikroskopische Beobiichtungcn tiber die Primitivfasern des Nervensystems der Wirbeltiere. Froriep's Notizen, Bd. 3. 1847 On the independent alimentary nervous system. Berlin. 1855 Untersuchungen iiber dieEntwickelungder Wirbelthiere. Berlin.

Retzius, G. 1881 Untersuchungen iiber die Nervenzellen der cerebro-spinalen Ganglien und der iibrigen peripherischen Kopfganglien mit besonderer Rlicksicht auf die Zellenauslaufer. Arch. f. Anat. u. Physiol., Anat. Abt.

1893 Zur Kenntnis der ersten Entwickelung der nervosen Elemente im Riickenmark des Hlinchens. Biol. Unters., Bd. 5.

1894 Ileber das Ganglion ciliare. Anat. Anz., Bd. 9.

Reuter, K. 1897 Ueber die Entwickelung der Augenmuskulatur beim Schwein

Anat. Hefte, Bd. 9. Rex, H. 1897 Ueber das Mesoderm des Vorderkopfes der Ente. Arch. f. mikr.

Anat., Bd. 50.

1900 Zur Entwicklung der Augenmuskulatur der Ente. Archiv fiir

mikr. Anat., Bd. 57.

1905 .Ueber d. Mesoderm der Vorderkopfes d. Lachmose. Morph.

Jahrb., Bd. 33. RoHON, V. 1884 Zur Histiologie des Riitckenmarks der Forelle. Sitz. Ber. Akad.

Wiss. Mvinchen, Bd. 1. Sagemehl, M.' Untersuchungen iiber die Entwickelung der Spinalnerven (Diss.)

Dorpat. Schaeppe, Th. 1904 Ueber den Zusammenhang von Muskel und Nerv bei den

Siphonophoren. Ein Beitrag ziu' Neuromuskeltheorie. Mitth. d. Nat urwiss. Ges. Winterthur.

1909 Kritische Bemerkungen zur Frage nach der Entstehung der

Nerven. Anat. Anz., Bd. 35. ScAMMON, R. E. 1911 Normal plates on the development of Squalus acanthias.

Normentafeln zur Entwicklungsgeschichte der Wirbelthiere, heraus gegeben von F. Keibel. ScHLACHER, A. 1801 De cataracta. ScHAFER, A. E. 1881 Note on the occurrence of ganglion cells in the anterior

roots of the cat's spinal nerves. Proc. Roy. Soc. London, vol. 31. ScHAPER, A. 1894 Die morphologische und histologische Entwicklung des

Kleinhirns der Teleostier. Morph. Jahrb., Bd. 21.

1897 Die friihesten. Differenzirungsvorgange im Centralnervensystem.

Kritische Studie und Versuch einer Geschichte der Entwickelung ner voser Substanz. Arch. f. Entwickl. Mech., Bd. 5. ScHENCK, S., UND BiRDSALL, W. R. 1878 Ueber die Lehre von der Entwickelung

der Ganglien des Sjinpathicus. Mitt, aus d. embryol. Inst. Wien.,

Bd. 3. ScHiEFFERDECKER, P. 1906 Neurone und Neuronenbahnen. Leipzig. Schneider, A. 1879 Beitriige zur vergleichenden Anatomic und Entwicklungsgeschichte der Wirbelthiere. Berlin.

1880 Ueber die Nerven von Amphioxus, Ammocoetes, und Petromy zon. Zool. Anz., Bd. 3.


160 H. V. NEAL

ScHULTZE, O. 1904 a Ueber die Entwickelung des peripheren Nervensystems Verh. Anat. Ges., Bd. 18. Vers.

1904 b Nachtrag zu meinem auf der Anatomenversammlung in Jena gehaltenen Vortrag ueber die Entwickelung des peripheren Nervensystems. Anat. Anz., Bd. 25.

1905 a Beitrage zur Histogenese des Nervensystems. 1. Ueber die multizellulare Entstehung des peripheren sensiblen Nervenfaser, u. s. vv., Arch. f. mikr. Anat., Bd. 66.

1905 b Die Continuitat der Organizationseinheiten der peripheren Nervenfaser. Pfliiger's Archiv, Bd. 108.

1905 c Weiteres zur Entwicklung der peripheren nerven, u. s. w. Verhandl. d. phys.-med. Ges. zu Wiirzburg, Bd. 37.

1906 Zur Histogenese der peripheren Nerven. Verh. d. anat. Gesell. Rostock.

1908 Zvu' Histogenese des Nervensystems. Sitzungsber. d. kgl. pr.

Ak. d. Wiss., Bd. 6. ScHWALBE, G. 1879 a Das Ganglion oculomotorii. Ein Beitrag zur vergleich.

Anatomie der Kopfnerven. Jena. Zeitschr., Bd. 13.

1879 b Ueber die morphologische Bedeutung des Ganglion ciliare.

Sitzungsber. Jena. Gesell. f. med. u. Naturwiss., 1878.

1881 Lehrbuch der Histologie. Erlangen. Schwann, Th. 1839 Mikroskopische Untersuchungen liber die Uebereinstim mung in der Structur und dem Wachstum der Thiere und Pflanzen.

Berlin. Scott, W. B. 1887 Notes on the development of Petromyzon. Jour. Morph., 1. Sedgwick, A. 1892 Notes on elasmobranch development. Quart. Jour. Micr.

Sci., vol. 33.

1894 On the inadequacy of the cellular theory of development, and

on the early development of nerves, particularly of the third nerve

and of the sympathetic in Elasmobranchii. Quart. Jour. Micr. Sci.,

vol. 37. Sewertzoff, a. N. 1895 Die Entwickelung der Occipitalregion der niederen

Vertebraten im Zusammenhang mit der Frage iiber die Metamerie des

Kopfes. Bull. Soc. Imp. Nat. Muscou., Heft. 2.

1898 a Die Metamerie des Kopfes von Torpedo. Anat. Anz., Bd. 14.

1898 b Studien zur Entwicklungsgeschichte des Wirbeltierkopfes. 1.

Die Metamerie des Kopfes des elektrischen Rochen (Fortsetzung).

Bull. Soc. Imp. Nat. Moscou, N. S., T. 12. Shipley, A. E. 1887 On some points in the development of Petromj^zon fluvi atilis. Quart. Jour. Micr. Sci., vol. 27. Shore, T. W. 1889 On the minute anatomy of the vagus nerve in selachians

with remarks on the segmental value of the cranial nerves. Jour.

Anat. and Physiol., vol. 23. Shorey, M. Louise 1909 The effect of the destruction of peripheral areas on

the differentiation of the neuroblasts. Jour. Exp. Zool., vol. 7. Stannius, H. 1849 Das peripherische Nervensystem der Fische, anatomisch

und physiologisch untersucht. Rostock.


MORPHOLOGY OF EYE MUSCLE NERVES 161

Steinach, E. 1893 Ueber die motorische Innervation des Darmtractus durch die hinteren Spinalnervenwtirzeln. Lotos, N. F., Bd. 14. 1895 Motorische Functionen hinterer Spinalnervenwiirzeln (in collaboration with H. Wiener). Arch. ges. Phys., Bd. 60.

Streeter, G. L. 1905 The development of the cranial and spinal nerves in the occipital region of the human embryo. Amer. Jour. Anat., vol. 4. 1908 The nuclei of origin of the cranial nerves in the 10 mm. human embryo. Anat. Rec, vol. 2.

Strong, 0. S. 1890 The structure and homologies of the cranial nerves of the amphibia, as determined by their peripheral distribution and internal origin. Zool. Anz., Bd. 13.

1895 The cranial nerves of amphibia. Jour. Morph., vol. 10. 1903 The cranial nerves of Squalus acanthias. Science, vol. 17.

VON SziLY, A. 1904 Zur Glasskorperfrage. Anat. Anz., Bd. 24.

Tretjakoff, D. 1909 Das Nervensystem von Ammocoetes. 2. Das Gehirn. Arch. f. micr. Anat., Bd. 74.

Van Valkenburg, C. T. 1910 Nucleus facialis dorsalis, nucleus trigemini posterior, nucleus trochlearis posterior. Konink. Akad. Wetensch. te Amsterdam.

Van Wijhe, J. W. 1883 Ueber die Mesodermsegmente und die Entwickelung " der Nerven der Selachierkopfes. Naturk. Verb. d. K. Akad. Wiss. Amsterdam, Bd. 22.

1886 a Ueber Somiten und Nerven im Kopfe von Vogel- und Reptilienembryonen. Zool. Anz., Bd. 9.

1886 b Ueber die Kopfsegmente und die Phylogenie des Geruchsorgans der Wirbelthiere. Zool. Anz., Bd. 9.

1889 a Ueber die Mesodermsegmente des Rumpfes und die Entwickelung des Excretionssj'stems bei Selachiern. Arch. mikr. Anat., Bd. 33.

1889 b Die Kopfregion der Kranioten beim Amphioxus, nebst Bemerkungen liber die Wirbeltheorie des Schadels. Anat. Anz., Bd. 4.

ViGNAL, W. 1883 Memoire sur le developpement des tubes nerveux chez les embryons des mammiferes. Arch, de physiol. norm, et path., tom. 1. 1888 Recherches sur le developpement des elements des couches corticales du cerveau et du cervelet chez I'homme et les mammiferes. Arch, physiol. norm, et path., tom. 2.

Waldeyer, W. 1891 Ueber einige neuere Forschungen im Gebiete der Anatomie des Zentralnervensj'stems. Deutsche med. Wochenschr., Bd. 17.

Waters, B. H. 1891 Some additional points on the primitive segmentation of the vertebrate brain. Zool. Anz., Bd. 14.

1892 Primitive segmentation of the vertebrate brain. Quart. Jour. Micr. Sci., vol. 33.

Weber, A. 1900 Contribution a I'etude de la metamerie du cerveau anterieur chez quelques oiseaux. Arch. d'Anatomie Microscop.

WiEDERSHEiM, R. 1880 Das Gehirn von Ammocoetes und Petromyzon planeri u. s. w. Morph. Stud., Bd. 1.

WiLLEY, A. 1894 Amphioxus and the ancestry of vertebrates. New York.

JOURNAL OF MORPHOLOGY, VOL. 25, NO. 1


162


H. y. NEAL


Wilson, J. T., and Hill, J. P. 1907 On the development of Ornitlinrhynchus.

Phil. Trans. Roy. Soc, London, vol. 199. Wlassek, R. 1898 Die Herkunft des Myelins. Ein Beitrag zur Physiologie

des nervosen Stutzgewebes. Arch. f. Entwick.-Mechan., Bd. (i. Workman, J. S. 1900 The ophthalmic and eye-muscle nerves of the catfish

(Ameiurus). Jour. Comp. Neiu*., vol. 10. ZiEGLER, H. E. 1908 a Die phylogenetische Entstehung des Kopfes der Wirbel thiere. Jena. Zeitschr. f. Naturw., Bd. 43.

1908 b Ein Embryo von Chlamydoselachus anguineus Garm. Anat.

Anz., Bd. 33. Ziehen, Th. 1906 Die Histogenese von Hirn- und Riickenmark. Entwickelung

der Leitenbahnen und der Nervenkerne bei den Wirbelthiere. Hand buch d. vergl. u. exp. Entwickelungslehre der Wirbeltiere, Bd. 2. ZiMMERMANN, W. 1891 Ueber die Metamerie des Wirbelthierkopfes. Verh.

Anat. Gesellsch. Miinchen.

1899 Ueber Kopfhohlerudimente beim Menschen. Arch. f. mikr.

Anat., Bd. 53.


ABBREVIATIONS


I to VIII, neuromeres or encephalomeres

1 to 9, somites of VanWijhe

l'^ to 6*^, dorsal divisions of myotomes lto6

1"^ to 6'^', ventral divisions of myotomes 1 to 6

V.md., ramus mandibularis trigemini

V.7nx., ramus maxillaris trigemini

V. opt.su., ramus ophtlialmicus superficialis trigemini

V.opi.p'fnd., ramus ophthalmicus profundus trigemini

Vlrec, ramus recurrens abducens

Vllac, ramus acusticus facialis

V lib lie, ramus buc calls facialis

VII hoi., ramus hyoideus facialis

Vllopt.au., ramus oi)hthalmicus superficialis facialis

a., 'anterior' somite of Miss Piatt

abd., abducens nerve anlage

ao.d., dorsal aorta

arc.vsc, visceral arch

ax., axone process

cd.d., chorda dorsalis

ch.dorfi., dorsal chiasma


cl.crs.n., neural crest cells

cl.n'bl., neuroblast cell.

cl.R.B., Rohon-Beard cell

d.pro., pronephric duct

ec'dryn., ectoderm

en'drm., entoderm

fis.vsc, visceral clefts

gls.phy., glossopharyngeus

gn.dors., dorsal ganglion

gn.opt.pj'nd., ganglion of the r. ophthalmicus profundus

gn.sym., sympathetic ganglion

gn.vag., vagus ganglion

hyp.,^ first, or transient, root of hypoglossus

I., lens

la.ct., cutis plate

la.niu., muscle plate

m.., mouth

ms'ec'drrn., mesectoderm

77is'ench., mesenchyma

7ns'en'drm., mesentoderm

mb.cl., cell membrane

mb.lim., limiting membrane

mu.hyp., hypoglossus muscidature

mu. obi. sup., superior oblique muscle


MORPHOLOGY OF EYE MUSCLE NERVES


163


vnj., myotome

my .^•'^•^'^' myotomes of VanWijhe's somites 1 to 8 my.'^^., dorso-lateral division of the

second myotome my.-^^., ventro-lateral division of the

second myotome. my.-™^., median division of the second

myotome my.'^^^., dorso-lateral division of the

first post-otic myotome my.'^^^., ventro-lateral division of the

first post-otic myotome my.'*™., median division of the first

post-otic myotome /(., olfactory pit nidi., nidulus, or 'motor nucleus' of

nerve nidl.abd., nidulus of abducens nerve nidl.oc, nidulus of oculomotor nerve nidl.tr'ch., nidulus of trochlearis nerve n'v., first post-otic nerve in Petromy zon


n'r.-, second post-otic nerve in Pctromyzon

oc'mot., oculomotorius nerve anlage

oL, otic vesicle

pi., plasma

rx.d., fiber bundle of dorsal nerve

rx.v., fiber bundle of ventral nerve

scL, sclerotome

so., somite

sp., spiracle

subch., sub-chordal rod

ib.n., neural tube

thyr., thyreoid

tr'ch., trochlearis

tr'ch.dors., dorsal branch of the trochlearis nerve

tr'ch vent., ventral branch of the trochlearis nerve

tr.art., truncus arteriosus

vac, vacuole

vag., vagus nerve

vn.crd., cardinal vein

vs. opt., optic vesicle


PLATE 1


EXPLANATION OF FIGURES

All the figures of this plate were drawn with Abbe camera, one-twelfth homogeneous oil immersion objective and No. 6 compensation ocular of Zeiss. In reproduction the magnification has been reduced by one-third. The series of drawings illustrates the stages just preceding and following the appearance of the anlagen of somatic motor nerves in the trunk region of Squalus embryos.

1 A portion of a cross-section (DK 2-3-10) of a Squalus embryo with eight somites (Stage D of Balfour) in the middle trunk region showing the relations of neural tube, somite, and chorda. The absence of plasmodesmata or protoplasmic strands connecting neural tube and somite is to be noted. A vacuolated plasma fills the intercellular space between the neural tube and the myotome.

2 A portion of a cross-section (IK 3-2-37) of a Squalus embryo of 6 mm. (Balfour's Stage I) in the middle of the trunk region, showing the conditions just previous to the appearance of a ventral nerve anlage. The section is taken from an embryo of about twice the length of the one from which figure 1 was drawn. The section shows no indication of protoplasmic or nervous connection between the neural tube and the somite. The outflow of processes from the sclerotome cells {scl.) is the beginning of the movement of mesenchymatous cells


PLATE 1: EXPLANATION OF FIGURES (CONTINUED)

into the region between myotone and neural tube. The anlagen of the somatic motor nerves of the myotomes immediately anterior to the one shown in the section have already made their appearance as protoplasmic outflows from the neural tube, an outflow which is correlated with a movement of the sclerotome cells.

3 A portion of a cross-section (II 5-3-12) of a Squalus embryo of 7 mm. (Balfour's Stage I) in the cloacal region, showing the relations of neural tube and somite immediately before the appearance of a somatic motor nerve. Nervous, or protoplasmic, connection of neural tube and myotome has already been effected in the metameres just anterior to the one from which this section was taken. The deeper staining properties of the cell (cl.n'bl.) may indicate that it is a neuroblast about to extend a process toward the somite. The migration of the mesenchyma cells from the sclerotome has already begun. This and figures 1 and 2 show that before outflows of neuroblastic cells make their appearance there is no protoplasmic connection between the nervous and muscular systems. In the intercellular space between the two, however, may be demonstrated a dilute, plasmoid substance containing a minimal amount of vacuolated coagulable substance more resistant to stains than the cellular protoplasm. To show it at all in a drawing greatly exaggerates its visibility.

4 A cross-section (IL 4^3-11) of a Squalus embryo of 7 mm. (Stage I of Balfour) in the trunk region just anterior to the cloaca, showing a very early stage of protoplasmic connection between neural tube and myotome, established by an amoeboid outflow of a single neuroblast cell of slightly deeper staining properties than those of the surrounding cells. The pseudopodial processes extend in various directions toward the somite and show finer branches which have connections with the vacuoles of the intercellular plasmoid substance. That the processes are genetically related to the neuroblastic cell, however, is evinced by their staining properties.

6 A portion of a cross-section (IL 4-2-1) of the same embryo as the one shown in figure 4 in the second metamere anterior to it. At least two cells appear to participate in the protoplasmic outflow from the neural tube. Focussing brings out the fact that the amoeboid processes extend antero-posteriorly along the surface of the myotome as well as dorso-ventrally. The neuroblast shows the characteristic deeper staining qualities of the neuroblast cell.

6 A portion of a cross-section of the same embryo as figures 4 and 5. The section (IL 4-2-14) is through the metamere immediately posterior to that in figure 5 and anterior to that in figure 4. As compared with figure 5, the amoeboid processes seem further extended and the limiting membrane of the neural tube seems interrupted for a greater extent than in the sections anterior and posterior. The outer boundaries of at least three cells are extended beyond the limiting membrane of the neural tube.


164


MORI'IIOLOGY OF liYE MUSCiE NERVKS

n. V. NEAL









PI.ATK 1








■ "^:g%^*-;;- \



JOURNAL OF MORl'HOLOGY, VOL. 25, NO. 1


my.


y

^^.



^■


1G5


PLATE 2


EXPLANATION OF FIGURES


All the figures of this plate were drawn with the same lenses, etc., as those of plate 1. In reproduction the magnification has been reduced one-third. The series illustrates stages in the development of the anlagen of spinal somatic motor nerves in Squalus embryos up to the time of migration of medullary cells into the nerve.

7 A portion of a cross-section (IL 4-2-17) of a Squalus embryo of 7 mm. (Stage I of Balfour) of the same series as those shown in figures 4, 5 and 6, showing the anlage of the somatic motor nerve of the same metamere as that in figure 6 but of the opposite side of the body and three sections posterior. The evidence that the protoplasmic connection between the neural tube and somite is a derivative of the neural tube is the same as that shown in the earlier figures. The section also shows a protoplasmic outflow from a cell in the lateral wall of the neural tube. Such outflows are exceptional and do not appear to persist in later stages.

8 and 9 Portions of two adjacent sections (II .5-2-17 and II 5-2-18) in the cloacal region of a 7 mm. embryo, showing an early stage in the development of a plasmodesm or anlage of a somatic motor nerve. The protoplasmic connection between the neural tube and the myotome consists of two elements, one derived from the neural tube, that is, a portion composed of the processes of neuroblasts (cl.n'bl.) labeled ax. in the figure, and the other consisting of migrant cells from the sclerotome of the somite (scl.). The two elements are readily distinguishable in the sections, since the processes of the neuroblastic cells stain more deeply than the sclerotome cells and the limiting membranes of the cells are shown distinctly.

10 A portion of a cross-section (IL 4-1-.30) of a 7 mm. embryo in the middle trunk region where the myotome and sclerotome lie somewhat nearer the neural tube than in the caudal region. The protoplasmic bridge between neural tube and myotome consists of the processes of medullary neuroblasts which extend between the myotome and sclerotome along the median surface of the myotome. A comparison with earlier stages shows that the bridge is not a primary one, as thought by Held ('06) and Paton ('07), but secondary.

11 A portion of a cross-section (II 5-1-13) of a 7 mm. embryo, showing a somatic motor nerve anlage in a stage of development somewhat more advanced than that shown in figure 10. Processes of medullary neuroblasts may be traced for some distance along the median surface of the myotome, between the sclerotome and the myotome. The more deeply staining properties of the distal portions of the axone processes are noteworthy, in connection with the problem of the origin of the neurofibrillae. The stage corresponds essentially with Paton's ('07) figure 2. The difference in the phenomena and resultant difference in interpretation may be ascribed largely to the difference in the methods of staining ami of preservation. What Paton regards as a neurofibril arising in the myotome independentlv of the nervous system is actually, as shown in sections of embryos preserved by the vom Rath method, the distal portion of the neuraxon process of a medullary neuroblast. Instead of appearing within the myotome as stated by Paton, the actual position, as seen in figure 11, is between the myotome and sclerotome. The results obtained by Paton's excellent method of staining the neurofibrillae need to be controlled by a comparison with methods which, like that of vom Rath, bring out the cell boundaries.

12 A portion of a cross-section (IL 3-2-42) of a 7 mm. embryo, showing an early stage in the process of migration of medullary cells into a somatic motor nerve anlage in the middle trunk region. Evidence has been given in a former paper ('03) that these cells are chiefly, if not entirely, concerned in the process of formation of the neurilemma of the somatic motor fibers. As shown in the figure, the breaking up of the sclerotome into loose mesenchyma has already begun. The relation of the fibrillar portion of the nerve anlage to neuroblasts in the neural tube is not shown in the section, a result of the bending of the axones upon their emergence from the neural tube.

16G


MOUrilOI.OtlV OK KYE MUSCLE NERVES

11. v. SEAL


I'l.AI E 2



12


JOUKNAI. OK MORl'HOLOOV, VOL. 25, NO. 1


167


PLATE 3


EXPLANATION OF FIGURES


All the figures of this plate were drawn in the same way as those on plates 1 and 2. Magnification reduced one-third in reproduction. The series of figures on this and the following plate illustrate stages in the extension of the neuraxon processes of the Rohon-Beard cells. The phenomena appear analogous with those shown in the formation of the anlagen of ventral motor nerves (plates I and II).

13 A portion of a cross-section (IJ 4-1-22) of a 6 mm. embryo in the middle trunk region. A portion of the ectoderm and of the dorsal wall of the neural tube are shown. The amoeboid process (pi.) of a medullary cell is strikingly similar to those which form the ventral nerve anlagen. Finer branches of the protoplasmic processes show definite relations to the vacuolated intercellular plasma, but no evidence that this relation is a genetic one. The phenomena are entirely in harmony with the supposition that the amoeboid protoplasmic process has a genetic relation to the medullary cell. Earlier stages show no evidence of protoplasmic continuity between neural tube and ectoderm.

14 A portion of a cross-section (IK 2-6-37) of a 6 mm. embryo in the middle trunk region, showing a portion of the dorso-lateral wall of the neural tube and a Rohon-Beard cell with deeply stained neuraxon process extending beyond the limiting membrane of the neural tube. Similar outflows of adjacent cells may also be seen, and the phenomena strikingly resemble those presented in the formation of the somatic motor nerves (plates 1 and 2).

15 A portion of a cross-section (sections IK 2-3-22 and IK 2-3-23 combined) of a 6 mm. embryo in the middle trunk region, showing a part of the dorso-lateral wall of the neural tube and the greatly elongated neuraxon process of a RohonBeard cell which may be traced through the wall of the neural tube into the plasma-filled space between the neural tube and the ectoderm. The pseudopodia-like extensions of the distal extremity of the neuraxon show relations to the vacuolated plasma, but such evidence does not prove any genetic relationship between the two, nor a primary protoplasmic relationship between the neuroblast cell and its peripheral distribution.

16 A portion of a cross-section (sections II 5-1-31, II 5-1-32 and II 5-1-33 combined) of a 6 mm. embryo in the middle trunk region, showing the dorsolateral wall of the neural tube and the neuraxon processes of two adjacent RohonBeard cells, one of which extends into the space between myotome and ectoderm the other between myotome and neural tube, the neuroblast cell of the latter not shown. The neuraxon processes appear distinctly fibrillar only in the proximal portion. Since the distribution of these neuraxon processes in later stages is essentially the same, viz., the extra-embryonic ectoderm, the different relations to the dorsal part of the myotome appears to favor the view that the path by which a neuraxon finds its way to its terminal organ is a matter of chance rather than a predetermined one along an intercellular bridge. Moreover, the amoeboid processes do not extend in one direction, as would seem demanded by the latter assumption, but irregularly in different directions.

17 to 20 (See p. 170).


168


MORiMior,o(;v of eye muscle xei{\i;s

H. V. NF.AL


PLATE 3


ec'drm



gn. dors.




ec'drm. c|. crs. n

15 '^ ' '•


cl. R. B.


?C--^-^^^


13


'^^^^,

M


tb. n.



r


my.




pl.


ec'drm. v,,*'-^


ec'drm.




%Yy^



19


JOURNAL or MORI'HOI.OrFY, VOL. 25, NO. 1


109


PLATE 3: EXPLANATION OF FIGURES (CONTINUED)

17 A portion of a cross-section (sections IK 2-4^15 and IK 2-4-16 combined) of a 6 mm. embryo in the middle trunk region, showing the processes of two Rohon-Beard cells, one of which lies out of the plane of the section. The conditions resemble those shown in figure 16, except that the processes appear confluent instead of divergent. The neurofibrillae are well differentiated in the proximal portion of the neuraxon.

18 and 19 Portions of cross-sections (IK 2-5-6 and IK 2-5-37) of a 6 mm. embryo in the middle trunk region, showing the distal termination of the growing neuraxon process of a Rohon-Beard cell as it extends into the space between the myotome and the ectoderm. In the case of the process in figure 18 the growing end lies nearer the myotome, while in figure 19 the growing end lies nearer the ectoderm. Comparison with earlier stages shows that in this region there are no protoplasmic bridges in the region shown in the drawing — the intercellular spaces are filled with an unstained vacuolated plasma. The brush-like termination is a characteristic feature of the neuraxon process.

20 A portion of a cross-section of a 7 mm. embryo in the middle trunk region, showing a Rohon-Beard cell on the top of a spinal ganglion. Its deeply staining neuraxon process contrasts strongly with the surrounding mesenchyma, and the evidence of the genetic relation of ganglion cell and neuraxon process seems more convincing than in the case of the cells which are imbedded in the wall of the neural tube. The process may be traced to a point lateral to the myotome. The section figured however does not show the peripheral termination of the process, which bends out of the plane of the section.


PLATE 4

EXPLANATION OF FIGURES

21 and 22 Portions of cross-sections of 6 mm. embryos in the middle trunk region, showing Rohon-Beard cells in somewhat advanced stages of the extension of the neuraxon processes. Figure 21 is a combination drawing from three sections (II 4-4-26, 27 28). Figure 22 is a combination of two sections (IK 2-6-29, 30). The presence of neurofibrillae in the protoplasmic processes evinces their nervous character. The distal extremity of these neuroblast processes, however, is granular, vacuolated and e.xtended into pseudopodial processes. Figure 21 does not appear to harmonize with the assumption of predetermined paths for the extension of the nervous outgrowths, but rather with the view that neuraxons are pseudopodia-like outgrowths of neuroblasts. The variation in the form of the amoeboid termination of the neuraxon processes shown in the two figures is striking and significant.


170


M<)UIMK)I.()(JY OK EYK MUSCLK NKItVKS

H. V. NKAL


'^1 "■:


1. R. B.


'\'


^^'/'j^"^


v:r)


ec'drm.



mb. iim.


my.




22


f3^-^


ec'drm



cl. R. B.


tb. n.


jou


RNAL OF MORl'llOI-OGV, VOL. '2r,, No. 1


171


PLATE 5


EXPLAXATIOX OF FIGURES


All the figures are based on camera drawings. Figures 23 to 28 inclusive are a series of semi-diagrammatic drawings, illustrating the more important stages in the development of spinal somatic motor nerves and summarizing the essential features of the histogenesis of spinal ventral nerves. Figures 29 and 30 are camera drawings of spinal nerves with one-twelfth homogeneous oil-immersion objective, No. 6 compensation ocular of Zeiss, and camera lucida of Abbe. Reduction one-third in reproduction.

23 A semi-diagrammatic drawing of a portion of a cross-section (IH 1-6-9) of a 6 mm. embryo in the region of the cloaca. The anlage of the somatic motor nerve is seen as a structure consisting of the protoplasmic processes of neuroblasts in the ventrolateral wall of the neural tube. The migration of the mesenchyma of the sclerotome has already begun.

24 A portion of a cross-section (IE 2-1-last), of a 9 mm. embryo in the caudal region, showing a somewhat more advanced stage in the development of a ventral nerve. The neuraxon processes of the medullary neuroblasts have extended farther along the median surface of the myotome and the mesenchyma cells have advanced to a point nearer the neural tube.

25 A semi-diagrammatic drawing of a cross-section (IE 2-7-27) of a 7 mm. embryo in the middle trunk region, showing the beginning of the migration of medullary cells into the anlagen of somatic motor nerve and neural crest. The relation of the fibrillar portion of the somatic motor nerve to neuroblasts is shown in the figure.

26 A semi-diagrammatic drawing of a portion of a cross-section (IH 2-1-20) of a 9 mm. embryo in the trunk region showing a somewhat more advanced stage in the development of a somatic motor nerve. Medullary cells are still in the process of migration from the neural tube along the nerve anlage. The ganglion of the dorsal nerve has been formed by the migration and aggregation of neural crest cells between myotome and neural tube. A Rohon-Beard cell in an early stage of formation of a neuraxon process appears in the dorsal wall of the neural tube.

27 A semi-diagrammatic drawing of a cross-section (KJ 3-4-middle) of an 11 mm. embryo in the middle trunk region, showing a more advanced stage in the differentiation of a somatic motor nerve. The emigrated medullary cells at this stage form a sheath enclosing the fibrillar portion of the nerve. Neuroblasts have become differentiated within the dorsal ganglion. A Rohon-Beard cell with neuraxon process extending between myotome and ectoderm is shown.

28 A semi-diagrammatic drawing of a cross-section (LAB 5-3-17) in the middle trunk region of a 17 mm. embryo, showing a stage in the development of a somatic motor nerve, when the anlage of a sympathetic ganglion has made its appearance at a point opposite the dorsal aorta and median to the trunk of the spinal nerve of the metamere. At this stage fibrous connection of the dorsal ganglion with the neural tube has been established and both dorsal and ventral rami have reached a considerable extent in their development. The migration of medullary cells into the somatic motor root has ceased. The cells that have

173


PLATE 5: EXPLANATION OF P'IGURES (CONTINUED)

wandered into the nerve anlage form a sheath around the fibriUar part of the nerve. The section shows a Rohon-Beard cell with neuraxon enterinjo; the root of the dorsal nerve — a rather exceptional relation.

29 A cross-section of a spinal nerve (MM 21-2-4) of a 46 mm. Squalus embryo, showing the fibrillar portion of the nerve consisting of deeply stained fibrils and a surrounding layer of sheath cells, presumably of medullary origin. The neuraxon fibers of the sensory bundle are somewhat larger in diameter than those of the motor bundle. Both appear as highly refractive granules, separated by a somewhat vacuolated intercellular plasma.

30 A cross-section of a spinal nerve in a 17 mm. embryo, showing sensory and motor components at the level of a sympathetic ganglion. The motor bundle lies nearer the myotome than does the sensory bundle, and each is surrounded by a single layer of sheath cells. The sympathetic ganglion anlage lies median to and somewhat between the two.


174


MORPHOLOGY OF EYE MUSCLE NERVES

H. V. NEAL


PLATE 5



175


MORPHOLOGY OF EYE MUSCLE NERVES

H. V. NEAL


PLATE 6


d, nbl ml^^S? £)



17G


PLATE G


EXPLANATION OF FIGURES


A series of camera drawings illustrating the histogenesis of the abducens nerve With the exception of figures 37. 38 and 40, all were drawn with one-twelfth homogeneous oil-immersion objective and compensation ocular 6 or 8 of Zeiss.

31 A portion of a parasagittal section (KP 2-1-6) of a 10 mm. embryo, showing an early stage in the development of the abducens anlage, which makes its appearance as a product of the union of the processes of neuroblasts in the ventrolateral wall of the medulla immediately posterior to the otic capsule. The presence of loose mesenchyma in the region of the nerve anlage complicates the picture, but the processes of deeply stained neuroblasts may easily be traced into the nerve anlage. The mesenchyma cells are granular, much vacuolated, and stain lightly, showing in some cases protoplasmic connections with the abducens anlage. In every essential respect the picture resembles an early stage in the histogenesis of a spinal somatic motor nerve. The chief differences consist in the relatively early appearance of a loose mesenchyma and in the remoteness of the myotome from the point of emergence of the neuraxon processes of the nerve anlage. The presence of a lightly staining granular protoplasm surrounding the deeply stained processes of the neuroblasts is noteworthy, but it is difficult to determine whether this ensheathing protoplasm is of mesenchymatous or medullary in origin. Analogy with spinal nerves would favor the conclusion that it is an outflow from medullary cells.

32 A portion of a parasagittal section (KR 1-3-4, 5, 6) — strictly, a combination of three sections — of an 11 mm. embryo, showing the anlage of the abducens nerve, when it possesses two roots and when its growing tip extends for some distance anteriorly in the mesenchyma lying at the base of the medulla. Deeply stained processes of medullary neuroblasts may be traced into the nerve anlage with its sheath of granular protoplasm. The highest powers of the microscope fail to resolve the neuraxon processes into constituent neurofibrils. As no evidence of the migration of medullary elements into the nerve anlage appears in this or in preceding stages, the cells in the vicinity of the nerve anlage are presumably mesenchymatous.

33 A portion of a parasagittal section (KS 2-2-7) of an 13 mm. embryo, showing a portion of the abducens anlage in a later stage of development than in figure 32. The nerve anlage consists at this stage of a compact bundle of deeply stained neuraxons surrounded more or less completely by a sheath of mesenchymatous cells. The section is somewhat exceptional in showing two isolated neuraxons which join the nerve anlage and each of which consists of a deeply stained fibril surrounded by a very thin, lightly stained granular protoplasmic sheath. Each of the neuraxons is also associated with a nucleus presumably of mesenchymatous origin.

34 A portion of a parasagittal section (LAA 2-1-3) of a 19 mm. embryo, showing a peripheral portion of the abducens anlage, consisting of a bundle of highly refractive neuraxons surrounded by a sheatli of mesenchymatous cells and granular protoplasm.

■ 35 to 44 (See pp. 178 and 179).

177

JOURNAI- OF MORPHOLOGY, VOL. 25, NO. 1


PLATE 6: EXPLANATION OF FIGURES (CONTINUED)

35 A portion of a parasagittal section (LB 2-2-6) of a 17 mm. embryo, sliowing the peripheral termination and growing tip of the abducens nerve anlage at a stage when it has nearly reached the posterior rectus muscle. The amoeboid termination closely resembles those of somatic motor spinal nerves and of the Rohon-Beard cells, and the section thus strongly favors the outgrowth theory of nerve histogenesis and the opinion that the abducens in its histogenesis resembles the spinal somatic motor nerves.

36 A portion of a parasagittal section (LK 2-2-10) of a 19.25 mm. embryo showing a portion of the abducens anlage in the vicinity of its point of emergence from the medulla. The section is of interest chiefly because it shows a single deeply stained fibril surrounded by a sheath of spindle-shaped cells. It was such evidence as this that led Dohrn ('91) at one time to infer that the spindle-shaped cells were genetically related to the neuraxon fiber. The evidence presented in this paper, however, tends to show that there is no genetic relation between the fiber and its surrounding cellular sheath. At this stage of development of the nerve anlage the process of migration of medullary cells has already begun and the ensheathing cells shown in the figure may be of medullary origin.

37 A graphic reconstruction of the anlage of the abducens nerve as seen in parasagittal sections of a 19 mm. embryo. The nerve arises from the base of the seventh and eighth brain neuromere (Neal), four roots emerging from the seventh and one only from the eighth. Its connection with these two distinct segments of the brain suggests the inference that the abducens is the nerve of at least two metameres. The posteriorly directed ramus recurrens of the abducens also suggests that the nerve was once distributed to musculature posterior to that which it now innervates. This branch soon atrophies. The roots of the nerve show a tendency to form a network or plexus.

38 A semi-diagrammatic drawing based on a parasagittal section through the head of a 19 mm. embryo, intended to show the position of the niduli of the eye muscle nerves. The niduli of the oculomotorius and of the trochlearis belong to adjacent metameres, while the nidulus of the abducens is several neuromeres removed.

39 A portion of a cross-section (LF 3-5-19) in the region of the nidulus of the abducens, showing evidence of the migration of medullary cells into the anlage of the nerve. The phenomena are similar to those which appear in somewhat advanced stages of histogenesis of spinal somatic motor nerves. As in the latter, the nerve anlage contains both fibrillar and cellular components, the fibrillar portion showing genetic relations with neuroblast cells in the ventro-lateral wall of the neural tube. Evidence that the cellular portion of the nerve is medullary and not mesenchymatous in origin is found in the fact that the cells first appear in the proximal portion of the nerve anlage and that, in successive stages, tliere appear to be more and more nuclei in this portion of the nerve, and a larger number apparently in the process of migration.

40 A portion of a cross-section (LAD 1-7-17) in the region of the nidulus of the abducens of a 19 mm. embryo, showing the ventral half of the medulla as seen under the low power microscope. The section shows the position of the nidulus of the abducens and its relation to the bundle of deeply stained neuraxon fibers which traverse the marginal zone and enter the nerve root.


178


PLATE 6: EXPLANATION OF FIGURES (CONTINUED)

41 A portion of the same section (LAD 1-7-17) enlarged to show the structure of the neuroblasts. These are pear-shaped, deeply stained cells with protoplasm extending into the fibrillar neuraxon process.

42, 43, 44 A series of cross-sections of the anlage of the abducens nerve as seen in a 19 mm. embryo. Figure 42 is the most posterior of the sections. The nerve anlage at this stage consists chiefly of a bundle of deeply stained neuraxons irregularly grouped together among vacuolar spaces and partly enclosed by a layer of mesenchymatous cells. In figure 44, the relations of the neuraxons of the nerve as it subdivides into a branch are shown. The highly refractive granules seen in cross-section correspond with the deeply stained fibers appearing in longitudinal sections.

PLATE 7

EXPLANATION OF FIGURES

Most of the figures are camera drawings with one-twelfth homogeneous oilimmersion objective and No. 6 compensation ocular of Zeiss showing stages in the histogenesis of the trochlearis nerve. Figures 55 and 56 are graphic reconstructions from parasagittal sections as seen under low power objective.

45 and 46 Portions of cross-sections (LAG 1-6-1, 2) of a 19 mm. embryo through the region of constriction between midbrain and hindbrain vesicles, showing portions of the trochlearis anlage as it passes through the mesenchyma lateral to the wall of the brain. The growing tip of the nerve is not shown. At this stage the anlage is wholly fibrillar and its deeply staining fibrils contrast strikingly with the granular vacuolated protoplasm of the surrounding mesenchyma. A thin layer of granular protoplasm may be seen covering portions of the nerve anlage. The fibrils of the nerve may be traced dorsally into the chiasma at the anterior boundary of the cerebellar neuromere, but the nerve has not yet become connected with the myotome of VanWijhe's second somite.

47 and 48 Proximal and distal portions of the trochlearis anlage in crosssections (AIBB 1-4-17 and MBB 1-5-2) of a 21 mm. embryo in the region of the cerebellar neuromere. The nerve anlage differs from the previous stage only in the larger number of fibers composing the nerve bundle. No cells closely associated with the nerve anlage are to be found. The growing tip of the nerve is not shown.

49 A group of neuraxon fibers near the peripheral termination of the trochlearis nerve anlage as seen in a parasagittal section (ME 5-1-4) of a 21 mm. embryo, showing the nerve anlage as a loose brush of neuraxons without any closely associated cells. The highly refractive fibrils show a thin sheath of granular protoplasm.

50 and 51 Portions of the trochlearis nerve anlage as seen in parasagittal sections (MD 2-2-5 and MD 2-2-8) of a 22 mm. embryo, showing the fibrillar character of the nerve. Figure 51 shows a distal and figure 50 a proximal portion of the nerve anlage. In its proximal portion, the trochlearis consists of a compact bundle of neuraxon fibers, naked or with a thin coating of granular protoplasm and devoid of closely associated cells. Distally, however, where the nerve anlage breaks up into a brush of loose fibers, mesenchymatous cells are more closely associated with the individual cells, but there is no evidence of a genetic relation between these cells and the neuraxon fibers.

52 to 56 (See p. 180).

179


PLATE 7: EXPLANATION OF FIGURES (CONTINUED)

52 A cross-section of the two roots of the trochlearis anlage near the point of emergence from the neural tube as seen in a parasagittal section (MD 5-1-3) of a 22 mm. embryo. At this stage the nerve anlage consists exclusively of deeply stained neuraxon fibers devoid of nuclei. A sheath of granular protoplasm, however, may be distinguished and mesenchymatous cells show loose relationships with the distal portion of the nerve.

53 One of the roots of the trochlearis as seen in a parasagittal section OA 3-1-8) of a 45 mm. embryo, showing evidence of the migration of nuclei from the neural tube into the roots of the nerve. Proximally the nerve root is cut longitudinally, while distally it is cut transversely. Nuclei appear within the nerve anlage among the neuraxon fibers and also partly within and partly without the neural tube. No mesenchymatous cells are closely associated with the nerve anlage near its point of emergence from the neural tube, and there is no evidence that the nuclei lying within the roots of the nerve have any other than a medullary origin.

54 A portion of a cross-section (MAG 1-6-7) of a 25 mm. embryo in the region of the optic vesicle, showing a peripheral portion of the trochlearis anlage at a stage when it has reached its termination in the superior oblique muscle. The section shows the relation of the trochlearis to the median surface of the muscle and to the ramus ophthalmicus superficialis trigemini which extends along the dorsal surface of the muscle, appearing in cross-section in the figure. Between the two nerve anlagen lies a mass of cells (gn.sym.) presumably derived from the r. ophth. sup. trigemini (Miss Piatt '91). Both of the distal branches of the trochlearis anlage show a similar aggregation of cells where they cross the ramus ophthalmicus superficialis trigemini. Their derivation and relations show that such groups of cells are to be regarded as sympathetic ganglia and not as primitive sensory ganglia of the trochlearis. They afford as little evidence of the sensory character of the trochlearis as do the sympathetic ganglia associated with somatic motor spinal nerves.

55 A graphic reconstruction of the trochlearis anlage and the structures associated with it made from a series of parasagittal sections (MAA 3, etc.) of a 24 to 25 mm. embryo as seen from the right lateral aspect under low power microscope. At its anterior extremity the ramus ophthalmicus superficialis trigemini terminates in a group of cells derived from the neural crest. The group of cells (gn.sym.) associated with the same nerve at the point of crossing of the trochlearis have a similar origin and their relations to the trochlearis anlage are secondary.

56 A graphic reconstruction of the head of a 25 mm. embryo as viewed from the left lateral aspect, based upon parasagittal sections (MAH 3, etc.), showing the relations of the cranial nerves at this stage. In order not to complicate the figure, only two of the eye muscles are shown.


180


MORPHOLOGY OF EYE MUSCLE NER\TES

H. V. NEAL


PLATE 7



V opt su


gn. sym

trch.


181


PLATE 8


EXPLANATION OF FIGURES


A series of camera drawings, showing the histogenesis of the oculomotorius nerve. Most of the drawings were made with one-twelfth homogeneous oilimmersion objective, No. 6 compensation ocular, and Abbe camera.

57 A parasagittal section (KQ 1-4-18) of a 9.5 mm. embryo viewed from the right side, showing the relations of the oculomotorius anlage at the time of its first appearance. The nerve anlage passes directly from the base of the midbrain vesicle towards the first somite of VanWijhe, and closely associated with the pre-cardinal vein.

58 A portion of the same section as in figure 57 (KQ 1-4-18) as seen under a 4 mm. apochromatic objective and No. 6 compensation ocular, showing the oculomotorius anlage as a product of neuroblasts in the base of the midbrain vesicle. Shrinkage has resulted in breaking the roots of the nerve anlage, but there seems no reason to doubt the genetic relation of the nerve anlage with deeply stained neuroblasts lying in the ventral wall of the midbrain, the deeply stained processes of which may be traced to, and beyond the limiting membrane of the brain wall. The nerve anlage shows a deeply stained fibril surrounded by a sheath of granular protoplasm. At the point of union of the two roots of the nerve anlage is a mesenchymatous cell. There is no evidence that this cell has any genetic relation to the nerve anlage, nor is there any evidence of a migration of medullary elements from the neural tube at this stage.

59 A portion of a parasagittal section (KH 2-1-1) of a 9.5 mm. embryo, showing the oculomotorius anlage at a somewhat more advanced stage of histogenesis than that shown in the previous figure, as evinced by the greater number of neuroblasts and neuraxons. There is no evidence of migration of cells from the neural tube at this stage, but several mesenchymatous cells are somewhat closely associated with the nerve anlage. The highly refractive fibrils of the latter are in sharp contrast with the lightly staining vacuolated and granular protoplasm of the mesenchymatous cells.

60. A combination of four sections (KL 2-2-1, 2, 3, 4) of a 10 mm. Squalus embryo in the head region, showing the relations of the oculomotorius anlage, which extends from the midbrain to the first myotome, as seen under the low power microscope.

61 A combination of four sections (KL 2-2-1, 2, 3, 4) showing a part of the oculomotorius anlage in a 10 mm. Squalus embryo. The association of the fibrillar portion of the nerve anlage with medullary neuroblasts in the ba^e of the midbrain is clearly seen. Distally, near the myotome, the nerve breaks up into a brush of separate fibrils, more or less closely associated with mesenchymatous cells. As there is no evidence up to this stage of the migration of medullary cells from the neural tube into the nerve anlage, and as the cells associated with the nerve anlage are more abundant in its peripheral portion near the myotome and the profundus ganglion, the evidence favors the conclusion that these cells are either mesenchymatous or that they are derived from the profundus ganglion, or from both sources.

62 to 71 (See pp. 184 and 185).

182


MORPHOLOGY OF EYE MUSCLE NERVES

H. V. NEAL


PLATE 8



PLATE 8: EXPLANATION OF FIGURES (CONTINUED)

62 A single neuraxon of the oculomotorius anlage in a parasagittal section (KL 1-5-12) of a 10 mm. embryo, showing its relations with mesenchymatous cells at the base of the midbrain vesicle. The fibril shows the characteristic granular sheath and the evidence which has been interpreted by some investigators in favor of the theory that the neuraxons are differentiated from a protoplasmic reticulum. Such a conclusion, however, ignores the evidence of the peripheral growth of the neuraxons from medullary neuroblasts. The picture is strikingly similar to Cast's ('09) figure 3, plate 13.

63 A portion of the oculomotorius anlage in a parasagittal section (LA 4-2-8) of a 16 mm. embryo, showing the axial fibrillar bundle and the cellular sheath. The conditions are the same as those seen in spinal somatic motor nerves and in the other eye-muscle nerves at similar stages of development. There is no doubt as to the neuroblastic origin of the fibrillar portion of the nerve anlage, but it is not possible to determine positively whether the cellular sheath is ectodermal or mesodermal in origin.

64 A portion of a parasagittal section (LC 2-1-10) of a 16 mm. embryo in the region of the roots of the oculomotorius anlage, showing the relation of fibers of the nerve anlage to neuroblasts lying in the base of the midbrain vesicle. No evidence of cellular migration from the brain into the nerve anlage can be seen, and the source of the cells grouped at the roots of the nerve is doubtful.

65 A cross-section (LG 2-2-9) of the oculomotorius nerve anlage in its peripheral portion in a 15 mm. embryo, showing the bundle of neuraxon fibers with the sheath of mesenchymatous cells. The structure is essentially the same as that in similar sections of the anlagen of spinal ventral nerves. The nerve figured shows approximately two hundred highly refractive granules in the axial bundle, evincing approximately that number of neuraxons in the nerve anlage.

66. A longitudinal section of the oculomotorius anlage in the distal portion of the nerve of a 17 mm. embryo, showing the protoplasmic sheath — no nuclei appear in the portion figured — and the axial bundle of fibers.

67 The oculomotorius anlage as frontal sections (LAG 1-4-24 and 25) of a 16 mm. embryo, showing the relations of the nerve to the ganglion of the ramus ophthalmicus profundus trigemini and to the first somite of VanWijhe. The group of cells lying between the ganglion of the profundus and the oculomotorius may be traced in later stages into the ciliary ganglion of the adult and thus form the anlage of that sympathetic ganglion. The evidence in Squalus seems to favor the view that these cells are derived from the profundus nerve; the possibility, however, that they are in part derived by migration from the brain is not excluded. Of this, however, I am able to find no evidence.

68 A cross-section (LAB 8-13-20) of the oculomotorius anlage in a 17 mm. embryo, showing an early stage in the penetration of the bundle of neuraxon fibers by neurilemma cells. As development continues the number of such cells within the fibrillar bundle steadily increases, beginning at the proximal and distal ends of the nerve anlage.

69 A part of the pro.ximal portion of the oculomotorius anlage in a parasagittal section (MC 4-1-4) of a 21 mm. embryo, showing the penetration of the neuraxon bundle by neurilemma cells.


184


PLATE 8: EXPLANATION OF FIGURES (cONTINUED)

70 The nidulus and roots of the oculomotorius anlage in a parasagittal section (AIE 3-1-10) of a 21 mm. embryo, showing the increase in the number of roots and the relation of the fibers of the nerve to deeply staining neuroblasts in the ventral wall of the midbrain. With the growth of marginal zone of fibers, the neuroblasts have receded from the limiting membrane of the wall of the midbrain, but they still retain their deeply staining properties and their neuraxon processes may be easily traced through the marginal zone of fibers into the roots of the nerve.

71 A single neuroblast of the nidulus of the oculomotorius in a 46 mm. embryo, showing the beginning of multipolarity and the relation of the neuraxon process.

PLATE 9

EXPLANATION OF FIGURES

. A series of diagrams designed to show the metameric relations of the e3'e muscle nerves and their comparability with spinal somatic motor nerves.

72 A diagram showing the topographic relations of dorsal and sympathetic ganglia to a somatic motor nerve in the trunk region of a Squalus embryo of 15 mm.

73 A diagram showing the relations of the oculomotorius nerve to a sympathetic (ciliary) ganglion and to the ganglion of the ramus ophthalmicus profundus trigemini in a Squalus embryo of 11 mm. A comparison of these relations with those of a spinal somatic motor nerve as shown in figure 72 shows that the}^ are essentially the same. The differences are such as would be brought about, following a reduction of the myotome. The evidence favors the view of the similar mode of formation of the s>anpathetic ganglion. The reduction of the myotome of VanWijhe's first somite allows the precocious connection of the profundus nerve with the skin and brings about the slightly different topographic relations of nerve, ganglia and myotome shown in the diagram.

74 A diagram showing the relations of the abducens nerve in a Squalus embryo of 13 mm. Like the primitive somatic motor nerves of Amphioxus, the abducens does not become connected with a dorsal nerve and (therefore ?) is not associated with any sjonpathetic ganglion. Its nidulus, that is, its central relations, and its distribution to somatic musculature, however, sufficiently attest its serial homology with spinal somatic motor nerves. The nidulus of the abducens is elongated, extending through two brain neuromeres (VH and VIII), from both of which the cells which form the vagus ganglion are proliferated.

75 A diagram showing the relations of the trochlearis in a Squalus embryo of 25 mm. Except for the dorsal chiasma, the relations are similar to those of the oculomotorius (fig. 73). The nidulus of the trochlearis lies in the same zone as that of the oculomotorius and immediately behind it in the floor of the cerebellar anlage (third brain neuromere). In precisely the same way that the oculomotorius is associated with the profundus nerve and the ciliary ganglion, the trochlearis is associated with the ramus ophthalmicus superficialis trigemini and with a transient (?) sympathetic ganglion.

76 to 82 (See p. 186).

185


PLATE 9: EXPLANATION OF FIGURES CcONTINUED)

76 A diagram showing the greatly modified metameric relations of the eye muscle nerves, based upon the conditions found in Squalus embryos. The diagram is a composite of different stages of development. The relations of the oculomotorius nerve to the ramus ophthalmicus profundus trigemini are those which appear in a 12 mm. embryo, while those of the trochlearis to the ramus superficialis are such as appear in a 25 mm. embryo. The ramus recurrens of the abducens appears in a 19 mm. embryo. The neuromeres and somites are as they appear in much earlier stages of development. The diagram is not intended to show any hypothetical stage in the phylogeny of vertebrates, but merely to present gra])hically the highly modified metameric relations of the eye muscle nerves to the primary neuromeres and the somites of VanWijhe.

77 A diagram of the left lateral aspect of the head of a Petromyzon embryo of 50 mm., showing the relations of the post-otic myotomes and their associated ventral nerves. The innervation of the first three post-otic myotomes (4, 5, 6) by the ventral nerves of the ventral nerves of the fourth and fifth post-otic myotomes (7, 8) is especially to be noted, as is, also, the splitting of the first post-otic myotome (4) along the dorso-lateral line into a dorsal and ventral division. The homology of the first post-otic somite of Petromyzon with the fourth somite of Squalus is based on the evidence given by Koltzoff ('02).

78 A diagram based upon a reconstruction of a 3.5 mm. embryo of Petromyzon by KoltzofT ('02) showing the splitting of the anterior post-otic myotomes into median and lateral divisions. In order to explain the present relations of the eye-muscle nerves (fig. 76) to the pre-otic myotomes in Squalus, a similar subdivision of the myotomes of Van Wij he's first, second and third somites may be assumed.

79 A diagram based upon a cross-section of a Petromyzon embryo of 3.5 mm., showing the three divisions (my. 4 m., my. 4 d. 1., my. 4v. 1.) of the first post-otic myotome, and their relations to the otic capsule. The hypothetical primitive connection of a somatic motor nerve with the median (transient) division of the myotome is shown. The two divisions of the myotome lateral to the otic capsule are permanent and become innervated by the ventral nerves of the fourth and fifth post-otic myotomes. By a similar substitution, it may be inferred, the rectus posterior muscle — a pre-otic muscle — has become innervated by the abducens — a post-otic nerve.

80 A diagram based upon a frontal section in the occipital region of a Petromyzon embryo of 4 mm., showing the relations of the anterior most post-otic ventral nerves to the myotomes. The division of these myotomes into a median portion and a portion lateral to the otic capsule is shown.

81 A diagram of the left lateral aspect of a Squalus embryo, showing the hypothetical primitive relations of the eye muscles to the lateral trunk musculature. Those myotomic divisions which are not functional in the adult Squalus are indicated by broken lines. Only those pre-otic myotomes which have acquired connection with the bulbus oculi and have become functional as eye muscles have persisted. The homologues of myotomes 4, 5 and 6, however, are functional in Amphioxus and Petromyzonts.

82 A diagram intended to show the conditions under which the dorsal chiasma of the trochlearis has been acquired. The left side of the diagram shows an hypothetical stage when the myotome of VanWijhe's second somite was split into three moieties. The right side of the diagram corresponds in all essential respects with the conditions found in a 25 mm. Squalus embryo. For an interpretation of the figure, see p. 114 of the text.

186


MORPHOLOGY OF EYE MUSCLE NERVES

H. V. NEAL,


PLATE 9



187


THE GENESIS OF THE PLASMA-STRUCTURE IN THE EGG OF HYDRACTINIA ECHINATA

CORA JIPSON BECKWITH

From the Zoological Laboratory, Columbia University

SIXTY-SIX FIGURES (eIGHT PLATEs)

CONTENTS

I. Introduction 190

II. Review of Schaxel's work 191

ril. Protoplasmic granules and mitochondira in Hydractinia 193

A. Material and methods 193

B. Protoplasmic structure of mature egg 196

C. History of young egg (bouquet-stage) 198

D. Early growth-period 199

1. Nuclear reconstruction 199

2. Pseudochromatin-granules 199

3. Basic staining globules 200

E. Growth and differentiation of adult structures 202

1. Nucleus 202

2. Protoplasmic changes 202

a. Development of simple yolk 202

b. Development of compound yolk 203

c. Development of mitochondria 204

d. Development of oil 205

F. Discussion and conclusions (nature of the extra-nuclear granules) 206

1. Morphological evidence 206

2. Staining tests and experiments 207

a. Preserved material 207

b. Intra vitam stains 208

3. Digestion experiments 212

4. Various other tests 213

G. Other accounts of chromatin-emission in hydroids 214

H. Mitochondria 216

1. Experiments 216

2. Discussion and comparisons 218

a. Origin 218

b. Relation to yolk and yolk-nucleus 219

c. Relation to protoplasm 219

IV. Maturation phenomena and amitosis in Hydractinia and Eudendrium. . 220

189

JOURNAL OF MORPHOLOGY, VOL. 25, NO. 2 JUNE, 1914


190 CORA J. BECKWITH

A. Hydractinia 221

1. Nucleus of the growth period 221

2. Maturation 222

3. Fertilization 225

4. Fertilization-membrane and mitochondria 225

5. Nucleolus 226

6. Chromosomes and continuity 228

B. Eudendrium 228

1. Maturation 228

2. Fertilization 228

V. Summary 230

Bibliography 232

I. INTRODUCTION

This paper is offered as a contribution to the discussion concerning the origin and relationship of the protoplasmic^ granules, especially the chromidia and mitochondria. Although this complex and difficult problem has received much attention, it is still far from a satisfactory solution. A considerable group of observers, including Goldschmidt and his followers, have endeavored to extend Richard Hertwig's conception of chromidia in Protozoa, to certain of the protoplasmic granules in the metazoan cell, attributing a nuclear origin to the latter (chromidia) and regarding them as only temporary elements of the protoplasm. A second group of observers, led especially by Meves and Duesberg, are highly skeptical in regard to this hypothesis and have reached the conclusion that the most constant and characteristic of the granules (mitochondria, plastochondria) are purely protoplasmic in origin. According to them, the granules in question form an essential constituent of the protoplasm and are permanent cell-elements which exhibit a genetic continuity, and are the bearers of protoplasmic heredity ('plastochondrial germ-plasm') . The present paper is concerned especially with the conclusions of Schaxel, which are more or less intermediate between the two foregoing conceptions. He finds both types of protoplasmic granules present, 'extra-nuclear chromatin-granules' (corresponding to chromidia) which control differentiation, and mitochondria of purely protoplasmic origin.

The term 'protoplasm' is used throughout the paper instead of 'cytoplasm', according to the usage of the earlier writers on the subject.


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA 191

II. REVIEW OF SCHAXEL'S WORK

Since Schaxel's results are rather complicated, it may be well to review them briefly here. His interesting conception has been derived from observations on the eggs of animals from widely varying groups (Hydrozoa, 'lib; Sc5^hozoa, '10; echinoderms, '11a; annelids, '12; ascidians, '09) in all of which hedescribes essentially the following conditions. After the last oogonial division the nucleus is reconstructed in the typical fashion, that is, the chromosomes elongate as smooth threads to form a bouquetstage. These threads then lose their smooth contour, become granular, and, by further diffusion, soon form a typical granular reticulum, the threads of which as a rule center in the nucleolus. Only a. small layer of protoplasm (which at this time takes an acid or 'plasma' stain) surrounds the nucleus. This is the 'preemission stage' and the protoplasm is said to be in a state of 'primary achromasie.'

This stage merges gradually into the next or 'emission' stage, characterised by the accumulation of chromatin-granules on the nuclear net, especially at those points where the threads of the net touch the nuclear membrane. At the same time, by filtration through the membrane, groups of granules collect on the outside of the nucleus directly against the wall, at the ends of linin-threads which are poor in granules. The stage of actual emission of material is of rather brief duration and at its close the protoplasm is in a state of 'chromasie.' This 'extra-nuclear chromatin' stains with basic dyes.

The 'post emission' stage follows rapidly and is characterised by further reconstruction of the nucleus, spreading of the extranuclear chromatin throughout the protoplasm ('complete chromasie') and rapid growth and difl"erentiation of the yolk. The latter is formed indirectly from the extra-nuclear chromatin, as a small island of yolk in a 'nest'- of granules, at the expense of which it increases. In his earliest paper on ascidians, Schaxel ('09) describes a 'secondary achromasie' of the protoplasm of the mature egg, any extra granules not used in the formation of the yolk being absorbed by the phagocytic action of the test


192 CORA J. BECKWITH

cells. In the other forms described, Schaxel emphasizes the fact that that part of the extra-nuclear chromatin, not used in the formation of the yolk, remains between the yolk-spheres as 'intra vitelline chromatin.' The latter has been traced in the echinoderms and annelids to the end of cleavage, when it has completely disappeared, having, presumably, been used in the process of differentiation. The cells of the blastula are again in a state of 'achromasie.' In Aricia, in which later stages of development were studied, a 'secondary chromatin-emission' occurs in the cells of the gastrula. The granules of this emission are used in the differentation of cells into body tissues. In later papers^ Schaxel describes mitochondria in the cell, in addition to the extra-nuclear granules. He considers the two elements distinct, since the mitochondria are present in the egg before the chrbmatinemission occurs and also remain as constant cell-constituents from cell getieration to generation, while the protoplasmic chromatin disappears from the cell in the process of differentation.

Schaxel's conclusions regarding the chromatic nature of the extra-nuclear granules are based on both staining reactions and morphological evidence. He is well aware that not too great emphasis can be placed on the fact that the granules stain as chromatin, yet this fact combined with the morphological evidence, has some weight. The morphological argument seems stronger, since granules outside the nucleus at the end of threads having few or no granules, strongly suggests the passage of those granules through the membrane. His figures are most convincing.

My immediate purpose has been to obtain more accurate evidence for or against these interesting conclusions by testing the staining reaction of the extra-nuclear granules in order to determine if possible to what extent such reactions can be relied upon as an indication of their nature; but I have also endeavored to study the origin and ultimate fate of the granules, their relation to the yolk and mitochondria, and the origin and fate of the

2 Mitochondria as such are not described in Schaxel's early papers. They are described carefully in his later ones on echinoderms and annelids to correct the mistaken conception that the nuclear granules represent the chromatic origin of the mitochondria.


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA 193

latter. Since Hydractinia echinata shows both mitochondria and protoplasmic granules which take 'nuclear' stains, it forms a favorable object for study of the origin and relation of these elements, the results of which form the first portion of the paper. In a second part a few observations are briefly presented which are intended to supplement previous descriptions of maturation and fertilization in Hydractinia echinata and Eudendrium ramosum.

This investigation was undertaken at the suggestion of Prof. E. B. Wilson, to whom I wish to express sincere thanks for his kindly direction of the work. For suggestions concerning certain experimental aspects of the problem I am indebted to Prof. T. H, Morgan. I wish also to express my thanks to Prof. F. R. Lillie for his generosity in putting a room at my disposal during several summers at the Marine Biological Laboratory at Woods Hole.

III. PROTOPLASMIC GRANULES AND MITOCHONDRIA IN HYDRACTINIA

A. MATERIAL AND METHOD

The material for this work was obtained at Woods Hole during the summers of 1910, 1911 and 1912. The egg of Hydractinia is very favorable for staining tests, since all stages of development, from the early egg in the entoderm to the mature egg in the gonophore, are present on the same stalk, making it certain in staining tests that all stages receive the same treatment. Since a change in staining reaction occurs during development of the egg, this is of much importance. One possible difficulty with this material is the well-known fact that the nucleus early loses its affinity for basic stains and may take acid stains but slightly. If care be used in determining the staining reaction of the chromatin in the very early stages, no confusion need result from this condition.

Various killing fluids were used, the results of which for the sake of brevity and clearness have been tabulated (table 1). Material fixed in fluids containing chromic and osmic acids is best for the study of both nucleus and protoplasm, since the integrity of the elements is much better preserved than in other


TABLE 1 Young egg


STAIN


KILLING FLUID


NUCLEUS


NUCLEOLUS


PS. CHR.


LARGE







GLOBULES



Meves" Meves


brownish yellow


violet or brown<


brownish yellow






violet*


Bensley**^


. Bensley'


purplish


clear red






red



gray



No stain*'^ . .













Bleached (Irdn-hem)


Bensley


gray or black


black


gray


black



Ftemming's


gray


brownish


dark' gray


black



Meves'


gray


brownish


gray


black ,


Iron-hematoxyiin and


sub-acetic


blue-black



black


black









pic-acetic


blue-black


black


black


black



formalin^


black


black


black


black



100 pel cent ale.


blue-black


black


black


black



alc.-acetic


blue-back


black


black


black



hot water'


black


black


bla€k


black



Flemming's


red, blue,


blue


blue


blue



or Meves'


young






sub-acetic


red, blue


blue


blue


purplish


Thlonin and eosin



young




blue



pic-acetic


red, blue, young


blue


blue


purplish blue



formalin


blue


blue


blue


blue


Lithium carmine and


Suba- [separate


red


red


blue


blue


Lyons blue ■


cetic\ together


red


red


red


red



pio-acetlc


red


red


red


red



sub-acetic


red, green,


blue


greenish


greenish




young



blue


blue


Auerbach'2 '


pic-acetic


red, green


blue


greenish blue


greenish blue



ale. 100 per cent


red, green


blue


green


greenish blue



Flemming's


purpUsh or


red


purplish or


red


Saffranin and Licht



red



red



GrUn


Meves'

1


purplish or red


red


purplsh or red


red


Saffranin and Methyl- \ violet /


Flemming's


violet


red


violet


red


1 The color of the globules is given if they take a different stain from the ground-substance of the sphere. In that case the ground-substance always takes the 'plasma' stain.

2 Meves' killing fluid for mitochondria is a modification of Flemming's fluid, i.e., the acetic acid content is much reduced (Erg. d. Anat. u. Entw., Bd. 12; Lee, last ed.).

' Benda's stain for mitochondria is a double stain of sulphaUzarinate of soda and crystal- violet. (Erg. d. Anat. u. Entw., Bd. 12, or Lee, last ed.)

The color of the oil depends on the degree of extraction. With long extraction the violet is removed , leaving the blackening caused by the osmic acid in the fixative.

' Benaley's mitochondrial methods (modification of Altman's methods) :

a. Killing fluid:

Osmic acid 4% 2 cc.

Potassium bichromate 2.5% 8 cc. 1-24 hours

Acetic acid 2 drops J

b. Stains:

Altman's anilin fuchsin 6 min.

1% methyl green 00 dip

Wash in 95% alcohol

c. Bleach to be used after the above killing fluid:

30 sec 1% potassium permanganate

30 sec 5% oxaUc acid

Wash in water

Dip in 2.5 % potassium bichromate

194


PLASMA-STEUCTURE IN EGG OF HYDRACTINIA


195


TABLE 1— Continued Old egg


NUCLEUS


nucleolus"


PBOTOPLASM


MITOCHONDRIA


COMPOUND TOLK


FINE TOLK


OIL


"""----



light brown




brown


brownish


violet or


brownish


deep violet


violet


brown and


violet or


yellow


browni


yellow 1




violet


brown<


reddish


red (clear)


reddish purple


red (purplish)


red (clear)


purpUsh red


greenish or

brown dark brown light brown


gray


black


gray


black


black


black


black


green


black


green


black


black


black


brownish black


green


black


green


black


brownish black


brownish black


brownish black


blue-black


black


green



green and black


green and black


black


blue-black


black


green


9


green


green


black


green


black


green


black


black


green


black


blue-black


black


green


10


black


black


black


blue-black


black


green


10


black


green


black


green


black


green


black


black


green


black


red


blue


reddish


reddish


bright blue


pale blue


brownish blue


red


blue


red



red


red


red


red


blue


red


« 


red


red


red


red


blue


red


u


blue


pale blue


blue


red


red


blue


11


blue


blue


blue


red


red


red


11


bluish


bluish


bluish


red


red


red


11


bluish


bluish


bluish


bluish red


blue


red


9


red


red


red


bluish red


blue


red


9


red


red


red


bluish led


blue


red


10 '


red


red


red


green


ied


green


bright red


bright red


light red


brownish red


green


red


green


bright red


bright red


light red


brownish red


light violet


red


violet


violet


red


reddish violet


1 deep purple


6 Five per cent formalin was made in normal salt solution and the formic acid neutralized with sodium carbonate (Mann,'02).

' Hot water proves an excellent fixation both for nuclear and protoplasmic structures.

' The pseudochromatin-granules lie in a protoplasmic net taking the 'plasma' stain.

' The mitochondria are absent, probably dissolved by the acetic acid of the fixative.

1" The mitochondria cannot be identified because of the poor fixation given by alcohol.

n The mitochondria cannot be differentiated because they stain exactly like the other protoplasmic elements.

'2 Auerbach's stain, used as a chemical test, is a mixture of acidulated methyl-green and acid fuchsin. Jenaische Zeitschrift, Bd. 30, 1890.


196 CORA J. BECKWITH

fluids. Chemical tests were made on material fixed in indifferent killing fluids, such as alcohol, hot water and formalin in which the formic acid is neutralized. A large number of stains were also used, the results of which are again shown in table 1. Some experiments with 'intra-vitam' stains gave most useful results, which are recorded in table 2.

Other methods to determine the chromatic nature of the protoplasmic granules have been tried, such as artificial digestion of fresh or alcohoHc tissue, and tests for nucleo-histone and phosphorus (Mann '02), The possible proteid nature of the granules was also investigated by the use of Millon's reagent. Experiments on the staining reaction of egg-albumen which had been fixed with various ones of the same killing fluids were also made.

B. PROTOPLASMIC STRUCTURES OF THE MATURE EGG

It is essential to difl"erentiate the protoplasmic structures of the mature egg in order to make clear their relation to the extranuclear granules which I prefer to call 'pseudochromatin-granules.' The nucleus will therefore not be considered at present. The egg of Hydractinia is large and filled with various granules and spheres which fall into four groups :

1 . Small simple yolk-spheres which vary greatly in size and occur throughout the egg.

2. Compound yolk-spheres, the largest elements in the egg. They form in general a crescentic layer around the egg, a little below the surface and are found to some extent throughout the center of the egg.

3. Oily bodies, about the size of the small yolk-spheres, which darken in osmic acid and are often irregular in shape. These pervade the whole egg.

4. Mitochondria, small bacillus-like rods, slightly more abundant at the surface of the egg, but rather evenly distributed throughout the protoplasm.

These elements of the mature egg lie in a finely granular protoplasm, which stains entirely with plasma stain, that is, there are


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA 197

at this stage no basic-staining granules that correspond to Schaxel's intra-vitelHne chromatin'* (fig. 14).

No one staining method differentiates these various types of structure, since, as a rule, all stain the same way with certain dyes used after some killing fluids. For example, as shown in table 1, if iron-hematoxylin is used on material in Flemming's or Meves' kiUing fluids, all the elements take the hematoxylin stain in slightly varying degrees of intensity (fig. 16). By the use of various killing fluids and stains, however, they can be shown to be distinct elements in the following way (table 1).

The oily bodies are most easily distinguished from the yolk since the former darken somewhat in osmic acid. After staining in saffranin and methyl-violet they take a deep violet color, while the yolk is red. Also, the oil can be completely separated from the other cell-elements by centrifuging the egg.

The two kinds of yolk, simple and compound (often indistinguishable from each other when iron-hematoxylin is used after Meves' or Flemming's fixation) show their individuality when stained either with Benda's stain or with saffranin and methylviolet. These stains reveal, in certain of the spheres, drops or globules which stain differently from the ground-substance of the sphere. The geometrical regularity of these is shown in figure 15, in which young and old compound spheres are represented. After the use of Benda's stain the mature simple yolk is uniformly violet in color, while the compound spheres show brilliant violet globules in a yellow ground-substance. When stained in saffranin and methyl- violet, the globules of the compound sphere are brilliant red and the ground-substance light lavender, while the simple yolk is reddish violet. The most striking differentiation appears in material fixed in picro-acetic killing fluid and stained in iron-hematoxylin and light green, for the simple

^ These elements are diagrammatically represented in figure 14 in the following manner: Young yolk is represented by solid gray circles; mature simple yolk is represented by hollow circles; mature compound yolk is represented by hollow circles containing small circles; oil is represented by hollow circles containing dots; mitochondria are represented as black rods; the protoplasm is uniformly gray.


198 CORA J. BECKWITH

yolk takes the plasma stain, while the globules of the compound spheres are the only bodies in the egg which take the hematoxylin.

Since the globules of the compound yolk-spheres are of similar size and may take the same stain as the mitochondria (the violet of Benda's stain) they at first suggest nests of dividing mitochondria. This is an impossibility however, since the compound yolk-spheres are the most conspicuous structures in eggs fixed in killing fluids which dissolve the mitochondria (picro-acetic).

The mitochondria are so identified because of their rod-like shape and their typical mitochondrial behavior; that is, they dissolve in fixatives containing acetic acid, do not dissolve in alcohol, are darkened by osmic acid, and give the typical response to the so-called mitochondrial stains (Benda, Bensley or Altman, and iron-hematoxylin) . In living material they are highly refractive bodies which take Janus green as a vital stain.

C. HISTORY OF THE YOUNG EGG, BOUQUET-STAGE

The youngest eggs studied were found in the proliferating area of the stalk of the gonophore. Here the entoderm cells are smaller than those of other regions, their protoplasm is not vacuolated, and their nuclei are more deeply staining. The eggs, when differentiated from the surrounding entoderm cells, are somewhat larger and have nuclei in the bouquet-state; a nucleolus is present in which the chromatin-loops center (figs. 1 b, 2,). A slightly later stage shows the nucleolus in which the chromosomes center, pressed against one end of the nucleus (fig. 3). The origin of these cells was not determined. Since they are evidently the result of a recent division (they are usually found in pairs) it seems probable that they are the product of an oogonial division (fig. 1). The egg cells are conspicuous in this stage, since the chromatin and nucleolus take the basic stains intensely. The protoplasm, which is a very thin homogeneous layer around the nucleus, stains but lightly with plasma stains and thus contrasts strongly with the nucleus. This stage corresponds to Schaxel's 'pre-emission stage' The structure is the same, whatever fixative is used. No attempt to study synapsis was made, although many cells showed some evidence of double threads (fig. 1 a).


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA 199

D. EARLY GROWTH PERIOD

1 . Nuclear reconstruction

The growth-period begins directly after the preceding stage, both nucleus and protoplasm increasing in size. Nuclear changes appear, which result in the reconstruction of 'the nucleus; that is, the densely staining smooth chromatin-loops break up into a coarse, open, granular net, the threads of which center in the nucleolus and radiate to the periphery of the nucleus (figs. 4, 5, 6). The granules of chromatin are conspicuous in this period at the nodes of the net, while a few fine granules he against the inner wall of the membrane. By the time this stage is reached the chromatin takes most basic stains somewhat less intensely than previously, and may even fail to stain with most basic dyes or take the plasma stain instead. The latter condition is striking when eggs fixed in either sublimate-acetic or picro-acetic killing fluids are stained with thionin and eosine or Auerbach's fluid. As the egg grows, the radial arrangement of the nuclear net is lost (figs. 7, 8, 18) until by the time the egg has reached the gonophore the chromatin is coarsely granular, with little evidence of the netlike arrangement left. At the same time, its affinity for basic dyes has continued to diminish (figs. 19, 20, 22, 25, 26, 27, 28). A more complete accoimt of the disappearance of the net is given later.

2. Pseudochromatin-granules

When but a slight increase in the size of the egg has occurred and when the above nuclear changes are occurring, there appears throughout the protoplasm a fine granular precipitate ('Emission stage' of Schaxel) which, as a rule, takes the basic stains (figs. 4, 5, 6). The granules of which this substance consists I will call 'pseudochromatin-granules.' Striking differences appear in the form and staining reactions of these granules as a result of different modes of fixation. After Meves' method of fixation, they are fine and evenly distributed throughout the egg. If stained with iron-hematoxylin they appear gray in color while the chromatin is deep gray or black. If stained according to


200 CORA J. BECKWITH

Benda's method, the protoplasmic granules and the chromatin both take the yellow color of the alizarine. Flemming's killing fluid produces a slightly coarser but still evenly distributed precipitate which stains somewhat more intensely with iron-hematoxylin than that killed in Meves' fluid. Several other fixatives (sublimate-acetic, picro-abetic) give very striking pictures, since they produce a strong coarse precipitate, which is not evenly distributed, but more or less massed in the region of the nucleus, and which has a great affinity for basic stains. If stained with ironhematoxylin, the pseudochromatin-granules are intensely black. If stained with Auerbach's fluid or double stained with thionin and eosin, a striking contrast is produced since the nucleus here takes the 'plasma' stain (fuchsin or eosin), thus emphasizing the fact that the protoplasmic granules stain with the basic dyes (methyl-green or thionin) .

In my material I can discover no such striking picture as Schaxel finds of groups of granules on the outside of the nuclear wall at the ends of nuclear threads, regarded by him as centers of distribution and diffusion into the protoplasm. Poor fixation, resulting from the use of sublimate-acetic and picro-acetic fluids, may give rjse to some such appearance, but in good fixations I find a uniform distribution of the granules from the beginning (figs. 4, 5, 6, 7, 17). Furthermore, the egg is constantly increasing in size and yet the granular mass which increases with it, retains at all times its uniform distribution. By the time the egg has reached the gonophore, it is of considerable size and is completely filled with these densely staining granules (figs. 8, 9). In the figures they are always represented by gray granules.

3. Basic staining globules

In the young egg, in addition to the pseudochromatin-granules, a second element, consisting of basic-staining globules, appears against the nuclear membrane (fig. 17). Since these globules ordinarily stain as the granules do (iron-hematoxylin) , and since they appear very much like similar bodies figured by Schaxel in Pelagia, which he interprets as centers of dispersal of the 'extra


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA 201

nuclear chromatin,' it is important to determine their nature. They are differentiated from other cell elements diagrammatically in figures 4, 5, 7, 8 and 10 by the use of circles containing parallel lines. Although they are so closely applied to the nuclear membrane, again I find no evidence of a direct nuclear origin. Their distinction from both pseudochromatin-granules and chromatin is apparent for several reasons. (1) These globules darken in osmic acid while the chromatin and granules as seen in unstained preparations, do not. (2) After Benda's stain, they take a deep violet color, while the chromatin and granules stain yellow. (3) Also in material stained with saffranin and methyl-violet the pseudochromatin-granules are violet, while the globules stain a brilhant red. It is evident therefore that the globules are identical, neither with the nuclear chromatin, nor with the pseudochromatin-granules. Since I find, as shown in the figures, no grouping of the granules around the globules, they can hardly be dispersion-centers. The time when globules first appear in the egg is variable, for they may be present as soon as the pseudochromatin-granules appear, or not until later. They are found up to the time the yolk begins to appear, sometimes lying a short distance from the nucleus, but never far away (figs. 7-10) .

To sum up: There are in the early growth-period two protoplasmic elements, one a fine granular precipitate (pseudochromatin-granules) which is scattered throughout the protoplasm and takes basic stains, the other, large drop-like masses which appear near the nuclear wall and which are also probably not chromatin. Neither of these elements appear as such in the mature egg, both being completely used during development.

During this period the egg is migrating from its original position in the entoderm of the stalk toward the gonophore, where it becomes established in the ectoderm. The increase in the size of the egg is considerable during this process, but still further growth takes place after the egg has reached the gonophore before any change in the condition described occurs, or any evidence of the structures characteristic of the adult egg appear.


202 CORA J. BECKWITH

E. DIFFERENTIATION OF THE EGG

1. Nucleus

The nuclear structure will be described in detail in a later section, but may be considered briefly here. After certain killing fluids (Flenuning, Meves, neutral formalin, hot water) the nucleus early loses its affinity for all basic stains and shows a very fine, nearly homogeneous structure, in which little or no evidence of a fine net or chromosomes appears. It now takes even the plasma stains very slightly (figs. 21, 28). After some killing fluids (subhmate-acetic, picro-acetic, alcohol), a heavy net-like precipitate, which takes basic stains strongly, is formed (fig. 22) . The nucleus, which reaches a relatively enormous size, lies at first in the center of the egg (fig. 19) and moves to the periphery only a short time before growth of the egg is completed (fig. 23). A basic staining nucleolus, which contains one or several vacuoles staining with acid dyes, is constantly present.

2. Protoplasmic changes

a. Development of the simple yolk. The first elements to appear in the protoplasm are small yolk-spheres which develop into the the simple yolk and possibly the compound yolk-spheres of the mature egg. It seems evident that these spheres develop directly from the pseudochromatin-granules as described by Smallwood ('09). Material fixed in Flemmings or Meves' fluid or formalin is best for these observations, since the individual yolk-spheres are kept distinct. In eggs about one-fourth grown, which have been stained according to Benda's method, the extra-nuclear granules can be seen to have enlarged slightly and uniformly (fig. 9). They still take the yellow-brown color which Benda's stain gives the granules. In a slightly older egg (about one-third grown) some of the spheres are seen to have grown more rapidly than the others, so that the uniformity in size is lost (fig. 10). A few now increase in size very rapidly, giving yolk-spheres of very unequal sizes (fig. 11). It is important to notice that the spheres still stain like the pseudochromatin-granules (yellow


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA 203

brown). In my material, as shown in the figures, there are no nests of 'extra-nuclear' granules within which yolk-material with a different staining reaction is secreted, as described by Schaxel. There is, rather, a graded series from the smallest spherical granules to the largest spheres; and these latter do not as yet differ in staining reaction from the pseudochromatin-granules, a condition subsequently seen.

A little before the egg is half grown, however, a change of staining reaction begins. A few of the largest yolk-spheres change their reaction in the Benda stain, the brown of the alizarin being replaced by violet. Figure 19 shows such an egg under low magnification, while figure 12 represents this change diagrammatically. As the egg grows, more and more of the yolk-spheres take the violet color until the whole egg is dominated by violet, a few small yellow ones remaining among the large violet spheres (figs. 13, 14, diagrammatic) . This change in staining reaction is very striking on a slide in which all stages in the growth of the egg are present, when obviously they have all received the same treatment in the staining process.

These points are less strikingly but well shown when some of the same material is stained in iron-hematoxyhn, since the pseudochromatin-granules are grayish, the young yolk gray, the nature yolk black (figs. 16). Although picro-actic and sublimate-acetic fixations are poor for determining the development of the yolk, the change in staining reaction is even more strikingly shown when iron-hematoxylin counterstained with light green is used, since the granules of the young egg take an intense black, while the yolk-spheres of the mature egg are green. This fixation also makes it evident that the pseudochromatin-granules as such are not present in the mature egg between the yolk-spheres ('intravitelline chromatin' of Schaxel) for there are no deeply staining granules present in such eggs, while young eggs after the same treatment are dominated by black granules.

b. Development of the compound yolk. Since the young stages of the simple yolk-spheres and the compound spheres stain alike, they cannot be distinguished from one another until they are nearly grown, when internal globules appear in the compound


204 CORA J. BECKWITH

spheres. These are differentatied rather suddenly about the time that the staining reaction of the simple yolk begins to change (fig. 13). The Benda stain shows one or several small globules staining with crystal-violet, within a large compound yolk-sphere, the ground-substance of which is yellow brown (figs. 14, 15). The size of these spheres varies considerably, some of them forming the largest elements in the egg. The globules are arranged symmetrically, and increase in size until, in the mature egg, they may merge and practically fill the sphere. Usually a notched edge indicates such an origin (fig. 15). Certain stains always differentiate the compound from the simple yolk. Thus, saffranin followed by methyl-violet gives the inner globules a bright red color and the external ground-substance a pale violet color. Some of the spheres never reach this state of development but appear in the mature egg with small distinct globules within. It seems probable from staining reactions that certain of the simple yolkspheres are utilized for the storing of a different material in the form of globules.

c. The appearance of the mitochondria. The mitochondria are not seen in the egg until after the yolk is well formed, but before the staining reaction of the latter has changed. As is usual for mitochondria, special fixatives (Meves' fluid, Flemming's fluid, Bensley's fluid, formalin, hot water) are essential. When Benda's staining method is used after fixation with Meves' killing fluid, a few small rounded bodies, violet in color, appear here and there among the yellow spheres, uniformly distributed throughout the egg (fig. 12). These are larger than the pseudochromatin-granules, about the size of some of the smaller yolk-spheres, and of uniform size. They are comparatively few at first, but gradually increase in number and size until they are so numerous in the mature egg that they fill all the spaces between the yolk-spheres. From their first appearance they take the typical mitochondrial stains, that is, violet in Benda's stain; intense red in Bensley's stain; and deep black in iron-hematoxylin. Since the question of the origin of the mitochondria has been variously answered, it is of interest to determine that point in Hydractinia. As they are scattered throughout the


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA 205

egg from the beginning, they are certainly not of direct nuclear origin. Again I find no evidence for their origin from the pseudochromatin-granules; for, when the mitochondria first come into view, they are larger than the granules and also stain entirely differently if the Benda method be used. Also they can hardly originate from certain of the yolk-spheres which are of similar size, since again the staining reaction is different and yolk-spheres much larger than the mitochondria are yellow in this stain. These facts all point to their formation de novo in the protoplasm.

d. Oil globules. These appear very early while the egg is still in the entoderm of the stalk; that is, before the yolk formation has begun and even before all the pseudochromatin-granules are formed (figs. 7, 8, 9, 10). They lie scattered amoung the granules, the number varying much in different individuals. They are in some cases few in number, but in others are frequently very numerous even in this early stage (fig. 8). At this time they are easily confused with the globules which appear against the nuclear wall, since they usually take the same stain. That they are distinct from the nuclear globules is apparent by their greater blackening in osmic acid. Also, in double staining with saffranin and methyl-violet the oil is violet while the globules are red. Again, after Benda's stain, the oil is brownish and the globules violet. Auerbach's stain distinguishes the oil from the pseudochromatin-granules, for the latter are green while the former are red. In later stages the oil globules increase greatly in number until the egg is profusely dotted with them (figs. 13, 14). The oil is also formed de novo throughout the growth-period in the protoplasm of the egg.

Since none of the elements of the mature egg correspond to the large globular masses which appear against the nuclear wall in the early growth-period, it is possible that they form a source of elaborated food which is used during the process of growth.


JOURNAL OP MORPHOLOGT, VOL. 25, NO. 2


206 CORA J. BECKWITH

F. THE NATURE OF THE PROTOPLASMIC GRANULES

Because of the theoretical interest connected with this question, it is essential to determine, if possible, whether the extra-nuclear granules are chromatin, extruded from the nucleus as such, or whether they are of protoplasmic origin. Certainly, the first impression given by these granules, staining intensely in basic dyes, favors the conclusion urged by Schaxel, that they are chromatin sent into the protoplasm. The picture is so striking that after many staining tests I was still convinced that such was the case; for all ordinary dyes show identical staining reactions for the granules and the chromatin. It was only after an extended study of the effect of many dyes, both on fixed and living material, that I finally reached a different conclusion, based on the following facts :

1. The uniform distribution of the basic-staining granules, as described above, makes their direct nuclear origin doubtful. In my material no accumulation of the granules against the nuclear wall, no corresponding accumulation of chromatin-granules within the nucleus, is to be seen at any time. This is especially noticeable in the nucleus of late stages when the basic granules are rapidly increasing, for the nucleus is now uniformly homogeneous with an affinity for plasma-stains only.

2. The amount of granular material in Hydractinia can hardly be accounted for if it is all given off in the early growthperiod as Schaxel finds in his forms, since this period is of too short duration. An enormous increase in amount takes place after the stage corresponding to Schaxel's 'emission' stage which cannot be explained by simple separation and distribution of pre-existing granules. And since the granules are of the same size in all stages, they cannot be formed by repeated separation without the addition of further material.

3. If further material is added, there is no evidence that it arises from the nucleus, since at the time when the pseudochromatin-granules are increasing rapidly, the nucleus is changing into the so-called resting state, during which time it takes the plasmatic stains while the granules are colored by the nuclear stains. It


PLASMA-STRUCTUEE IN EGG OF HYDRACTINIA 207

seems no more probable that an emission should occur with the chromatin in a diffuse state than when it is in the condensed condition of the early growth-period, at which time no emission could be proved.

4. It is true that globular masses appear against the nuclear wall which after certain stains may be colored like both the pseudochromatin-granules and the chromatin. That they are not the same was shown by Benda's stain and double staining with saffranin and methyl- violet (table 1). This similarity of staining reactions in some cases makes confusion of the three elements easy. The globules also, as stated above, do not form dispersion centers.

5. Although in the majority of cases the pseudochromatingranules take the basic stains, there is still strong evidence from staining reactions against their chromatin character. Paradoxical as it may seem, the most convicing evidence against their being extruded chromatin is, in fact, from staining tests.

As can be seen from table 1, which presents the results of staining tests on sections after many fixations, the granules usually stain with the basic stains like the chromatin. For example, the result is the same when iron-hematoxylin is used after any killing fluid, whether it contains osmic acid, a heavy metal, or is an indifferent fluid such as neutral formalin, alcohol, or hot water. In fact, in these indifferent fluids, the granules stain as intensely as in material fixed in acid fluids, such as sublimate-acetic and picro-acetic fixatives, while they stain much more strongly than in eggs killed in the usual osmic acid containing fixatives. Again, in material fixed in Meves' fluid and stained according to Benda's method, the granules and the chromatin stain alike (i.e., yellow). Also, staining with thionin or Auerbach's fluid gives the same results, the granules take the basic stain. This is the result with the large majority of stains.

There are, however, some exceptions to this condition. The first fact to come to my notice was the variable results obtained with saffranin and light green when used after Meves' and Flemming's killing fluids. In the early stages the nucleus stains with


208 CORA J. BECKWITH

the saffranin very intensely while the protoplasm is colored green. The color of the granules differs, however, according to the length of time the slide is left in the solution of light green, and consequently to the degree of extraction of the saffranin, while the chromatin stains in the same way under all circumstances. For example, if the saffranin is but slightly extracted, the granules are bright red like the chromatin, the protoplasm green. If the saffranin is somewhat further extracted, the granules appear purplish (i.e., a combination of the two colors), while the chromatin remains bright red. If the extraction is more complete, the granules become entirely green while the chromatin still appears bright red.

Since the above experiments suggested that the granules may not respond to all chromatin-stains, further tests were made. Several investigators (Crampton '96, Foot '96) have used solutions of lithium carmine and Lyons blue to distinguish the yolk-nucleus from the chromatin ; the yolk-nucleus is stained blue, the chromatin stains with the carmine. Obviously, if the protoplasmic granules in Hydractinia are chromatin, they should stain with carmine. I find, however, that here again the granules stain differently from the chromatin, since they take the blue stain and the nucleus the red, this last even in late stages when the nucleus has usually lost its affinity for basic dyes. Further differences between the granules and the chromatin are shown by slight variations in their staining reactions when iron-hematoxylin is used after different killing fluids. As previously indicated, the granules are gray after Meves' killing fluid, slightly darker after Flemming's fluid, and intensely black after fixation in sublimate-acetic, formalin, alcohol, and hot water. The staining tests on preserved material indicate then, that the granules are not necessarily the same in their reaction as chromatin.

More striking and decisive results are given by experiments with stains on fresh material (table 2). I first tried a dilute solution of methyl-green slightly acidulated with acetic acid, as recommended by Lee ('03) for a chromatin stain. Here the results are as in the majority of experiments on fixed material, i.e., the granules take the basic stain (methyl-green) strongly


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA 209

while the nucleus is unstained. Auerbach'sstain, which combines acidulated methyl-green and acid fuchsin, also gives similar results since the granules take the green basic stain and the nucleus stains with the red plasma stain.

Since the above stains contain an acid, several of the more usual basic intra vitam stains were now tried (neutral red, methylene-blue and dahlia) . Different results appeared immediately, since both the granules and the chromatin are stained intensely with the neutral red or the methylene-blue. A still different result was obtained with dahlia, for while the nucleus is stained intensely purple, the granules were but slightly tinged.

Since Lee states that these intra vitam stains are harmful to the cell and the results not trustworthy, some further tests were made with some comparatively new stains (footnote 4, table 2), which have been found by other workers to be perfectly harmless to the cell and to stain chromatin (Kite '13). These give the most striking results. Three stains were used; new methyleneblue G.G., new methylene-blue R., and diamond-fuchsin. A very dilute solution was made by adding a small amount of the stain to sea-water containing Hydractinia eggs. After a few hours the eggs were mounted in glycerine. Methylene-blue G.G. and R. gave the most striking results, although diamond-fuchsin gives convincing preparations. The nucleus in every case takes the basic stain strongly while the pseudochromatin-granules are left colorless. A noticeable difference in staining reaction occurs between fixed and living material, for in living material the nucleus stains in all stages with these basic stains while in fixed material the nucleus in late stages stains only in acid stains. Since in certain cases the granules in living material do not take the chromatin-stains, while the nucleus does, they cannot be the same as the chromatin in the nucleus and are therefore not formed chromatin extruded from the nucleus. And since with a stain that is acid, the staining reaction is reversed, i.e., the granules stain with the basic dye and the nucleus is either non-staining or stains lightly with acid stains, it seems probable that the presence of the acid is the determining factor in this reaction. The same explanation would hold for the behavior after many killing fluids.


210


CORA J. BECKWITH


In indifferent killing fluids (alcohol, neutral formalin, hot water) the reversed staining reaction of the granules must be due to some other cause than the presence of the acid. It is possible that mere precipitation changes the chemical composition of the granules, giving it an affinity for basic dyes.

Similarity in staining reaction for the identifying of materials has long be questioned. The physical and chemical nature of staining processes has been ably discussed by many writers (Fisher '09, Hardy '99, Lilienfeld '93, Heidenhain '11, Nemec '10, Mathews

TABLE 2

Vital stains and digestion tests Young eggs


KILLING FLUID



STAIN


NUCLEUS


NUCLEOLUS


PS. CHR -GR.


LARGE GLOBULES


[■


ControU


iron-hem. and


blue black


black


black


black


95 per cent


Digested


light green


black


black


black


black


or 100 per : cent alcohol


ControU



green(young)


blue


greenish


greenish



Digested


Auerbach




blue


blue





green


green


green


green


TTnt water J


Control!


iron-hem. and


black


black


black


black


±X\J\i >\ClfcCl <


Digested


light green


black


black


black


black



Control!


acidulated


no stain


5


intense green


5



Digested


meth-green


no stain



green




Control!


Auerbach's


no stain



intense green




Digested


fluid


no stain



green




Control!


New methylene blue


intense blue



no stain




Digested


G. G. (4)


no stain



green



Control!


New meth

intense blue



no stain





ylene blue






ja


Digested


R. (4)


no stain



green



o


Control!


Diamond


intense pink



no stain



f^



fuchsin







Digested



no stain



pink




Control!


Millon's reagent


no stain



brick red




Digested



no stain ^


no stain





Neutral red


deep pink | deep pink


deep pink


i




Methylene blue


deep green deep green


deep green


5




Dahlia


deep purple deep purple


deep lavender


5


! Sections or fresh material were digested in artificial gastric fluid (1 per cent pepsin in 0.2 per cent HCl) at body temperature from two to three hours. In the adult egg the yolk and mitochondria are digested, leaving a framework of protoplasm. Millon's reagent shows that some proteid from the protoplasm of the young egg is removed while the pseudochromatin-granules are not affected.

2 The fixation is so poor that the mitochondria cannot be identified .


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA


211


'98, Mann '02, Prenant '10 a, Lundegard '12, et al.). The consensus of opinion on the subject is that even though staining may be chemical it cannot be rehed upon as indicative of chemical similarity of cell materials. This is clearly shown by Mathews ('98) and Zacharias ('97) who have demonstrated that previous acid or alkaline treatment determines what stain is selected by any cell element. The above experiments confirm the conclusion that similarly staining elements can hardly be considered identical. Again, experiments with egg albumen fixed in the various killing fluids according to Mann's method, confirm his results in that an


TABLE 2— Continued Mature eggs


NUCLEUS


NUCLEOLUS


PROTOPLOSM


MITOCHONDRIA


COMPOUND YOLK


SIMPLE YOLK


OIL


blue black


black


green


2


black


black


black


greenish gray


black


greenish gray





<


bluish red


blue


red


purplish


red


red


red


greenish red


green


red






green


black


green


black


black


green


black


green


black


green






blue green


5


light gray green


no stain


no stain


no stain


5


blue green


e


gray






purplish pink


6


grayish pink


no stain


no stain


no stair


S


purplish pink


b


grayish pink


no stain


no stain


no stain


s


intense blue


5


pale gray


no stain


no stain


no stain


6


no stain


6


pale gray


no stain


no stain


no stain


5


intense blue


6


pale gray


no stain


no stain


no stain


i


no stain


5


pale gray


no stain


no stain


no stain



intense pink


5


pinkish gray


no stain


no stain


no stain


5


no stain


5


pinkish gray


no stain


no stain


no stain


&


no stain


5


no stain


no stain


no stain


no stain


5


no stain


i


no stain


no stain


no stain


no stain


5


pink


pink


pale pink


no stain


no stain


no stain


'■>


pale green


pale green


very pale green


no stain


no stain


no stain


5


deep purple


deep purple


pale purple


no stain


pale purple


pale purple


6


3 The oil globules were not identified in any of these tests.

These stains furnished by the Caaella Color Company, New York, were found to be perfectly harmless used in dilute solution (Kite '13). ' The nucleolus, large globules and oil have not been identified in fresh material.


212 CORA J. BECKWITH

apparently differential stain may be obtained on coagulated egg-albumen which may depend on purely physical differences or differences in density. Experiments were also made to test Heidenhain's contention that selective staining is most successful when the stains are applied simultaneously and progressively. Since different results are obtained when stains are used successively, sumultaneously, progressively, or regressively, a true selective value can hardly be maintained. The evidence from all sides indicates that staining reactions are unreliable as chemical tests.

6. Artificial peptic digestion tests were tried to determine if possible, whether the granules are chromatin (table 2). If one accept the non-digestion of any material in the cell as proof of its chromatic character, then thife test supports the nuclear origin of the granules. Material killed in alcohol and in boiling water as well as fresh material, was used in these tests. The fixed material was sectioned and digested on the slide in a 1 per cent solution of pepsin in a i per cent solution of hydrochloric acid. Such digested sections together with an undigested control were stained with iron-hematoxylin and light green or with Auerbach's fluid. The fresh material was stained both before and after digestion with acidulated methyl-green, Auerbach's fluid, new methylene-blue G.G., new methylene-blue R., and diamond fuchsin, and mounted in glycerine. In all cases, after two to three hours of digestion at body temperature, the yolk, mitochondria and the bulk of the protoplasm were digested in the adult egg, leaving a slight framework or net which stained only with plasma stains. In the young eggs the protoplasmic granules remain undigested and. take the basic stains, while the nucleus is non-staining. This reversal in staining of digested material as compared with fresh material when stained with the same neutral vital stains may well be due to the acid in the digestive medium, a point which supports the view that staining depends on previous treatment.

Even though the granules in fixed material, whether digested or undigested, stain with basic dyes, a slight difference between the two occurs. If two slides, one containing digested, the other undi


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA 213

gested sections, are run back to back through staining jars to ensure similar treatment, the granules in the digested sections after staining with iron-hematoxylin, take the stain much less intensely than those of the control slide. Also after Auerbach's stain the granules are grayish in the digested material,, rather than green as in the undigested sections.

Still another point suggests that there may be other materials than chromatin in the cell which may not be digested by peptic digestion. In sections placed in a digestive medium in which the acid content is strong 0.5 HCl and digested either a long time (18 to 20 hours) at room temperature, or a short time (2 to 3 hours) at body temperature, all the material in the nucleus is digested while the granules remain undigested. This confirms the impression that the granules are of a different nature from chroniatin. Again, the failure of a certain material to be digested by peptic digestion is not necessarily a proof of its chromatic character, since peptic digestion depends on the degree to which a substance is penetrated and therefore on the density of the material. In the above case it is possible that the acid aids in the penetration of the nuclear material but is unsuccessful in penetrating the granules. On this basis peptic digestion is no test for chromatin. Since pancreatic digestion digests the whole cell, it is evident that an alkaline medium is essential for the digestion of the pseudochromatin-granules.

7. Millon's proteid test also is in harmony with the possible chromatin nature of the granules. Before digestion, young eggs containing granules (as well as mature eggs) give the typical brick red reaction very strongly, while no proteid test is obtained after digestion, although the granules are still present as stated above. The proteid reactions in the undigested egg must then be given either by the protoplasm or some proteid associated with the granules, rather than by the granules themselves.

8. Attempts to locate chromatin by testing for histone as described by Mann ('02) were only partially successful. The results, however, support the view that the granules are not the same as the chromatin. The strongest color reaction indicating the presence of histone occurred in the nucleus, while the proto


214 CORA J. BECKWITH

plasm of the young eggs containing the granules, although responding to the test slightly, did so no more strongly than the adult egg in which no granules are present.

9. Attempts to locate chromatin derivatives in the protoplasm by tests for phosphorus as described by Mann were also tried. These again are useless since Bensley ('06) has recently shown that the standard tests for phosphorus are unreliable.

To sum up : The balance of the evidence in Hydractinia decidedly indicates the nonchromatic nature of the granules in question. In all cases which seem to indicate the contrary conclusion (some staining and digestive tests and tests for proteid) the result can be interpreted in some other way. When to this are added the definite results from staining reactions in both fixed and living material and the morphological evidence that has been given above, we are, I think, forced to the conclusion that the granules are not chromatin extruded as such from the nucleus.

The above conclusions differ decidedly from Schaxel's, which appear to be based on very careful and detailed observations. Whether these conflicting views are due to the different forms studied, or to the fact that the large number of methods of fixation and staining used on Hydractinia has made the nature of the granules more evident, it is impossible to say. Since Hydractinia behaves, when Schaxel's methods are employed, in the same way that his material does, it seems probable that if the above methods were used on the forms studied by Schaxel they would yield similar results.

G. OTHER ACCOUNTS OF CHROMATIN-EMISSION IN HYDROIDS

Chromatin in the protoplasm has recently been described by several observers in a number of Hydroids. C. T. Hargitt ('13) finds a chromatin-emission in Campanularia brought about by the fragmentation of the nucleolus. The yolk develops from this extruded chromatin. Stschelkanowzew ('06) has described in Cunina a similar chromatin-emission through chromatin-nucleoli (secondary nucleoli). This condition is not found in Hydractinia, as the description of the nucleolus in a later section indicates. A


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA 215

still different method of chromatin-emission is described by Smallwood ('09) in Hydractinia. Soon before maturation small particles of chromatin leave the nucleus and wander into the protoplasm. Smallwood says, "The reason for regarding them as chromatin is because they give the same color reactions as similar shaped bodies in the nucleus, and for the further reason which is obvious in figure 1, namely, the actual migration of the chromatin from the nucleus." An exactly similar chromatin-emission is described by Smallwood ('07) for Pennaria, taking place here, however, after maturation rather than before. Also in this form the chromidia are stated to arise from both male and female germ-nuclei. In these two Hydroids the protoplasmic chromatin is connected with neither yolk formation nor differentiation. As will be shown in the next section, I have traced the nucleus in Hydractinia through the growth-period, the reappearance of the chromosomes and maturation stages, and in no case are bodies of the type described to be found. From Smallwood's figures, it is evident that the fixation is defective, since a very strong coagulation net is present which does not appear with good fixation. It seems to me the bodies in question may well be artifacts.

Trinci ('07) has described basophiUc granules in the oocytes of a number of Hydroids, which he considers as a differentation of the protoplasm brought about by the influence of the nucleus and as belonging in the same category as chromidia, mitochondria or plastosomes. The granules disappear during the development of the egg.

Van Herwerden ('13) has recently tested Schaxel's hypothesis of the nuclear origin of the basophylic granules by the use of nuclease. After treating the mature echinoderm egg (in which mitochondria are visible in life) with a preparation of nuclease, he finds that the mitochondria (basophilic granules) have disappeared. He concludes that the mitochondria are a nucleinic acid compound and therefore properly chromidia. Since in the young, living egg he can see none of the basophilic granules which appear in fixed material, he believes the latter are artifacts. Unlike Schaxel, however, he holds the mitochondria to be developed from this basophilic substance, since it is also a nucleinic acid


216 . CORA J. BECKWITH

compound as shown by its digestion with nuclease. Direct observation on young, hving eggs failed to show a direct migration of material from the nucleus into the protoplasm. Diffusion currents present in the egg, coincident with a slight nuclear shrinkage, he feels, favors the diffusion of a soluble substance through the nuclear wall. That he is dealing with a substance that differs from the granules of Hydractinia is apparent, since vital stains give different results in the two cases. His granules are stained with dahlia, those of Hydractinia are not. With methylene-blue and neutral red his granules do not stain, while those of Hydractinia do. Also the fate of the granules differs in the two cases. They can hardly then be homologized. Although he has been unable to furnish any more definite proof of a direct migration of nuclear material into the cytoplasm, the nuclease digestion indicates nucleinic acid present in the granules and mitochondria.

H. MITOCHONDRIA

1. Experimental

The presence and behavior of the mitochondria as seen in sections have been sufficiently described above. Further investigation of the function and behavior of these bodies was carried on by means of some centrifuging experiments. In previous experiments of this sort, no attempt has been made to locate the mitochondria after centrifuging the egg and to determine their further behavior in development. Hydractinia eggs within 5 minutes after fertilization were placed in a water-centrifuge and revolved at a moderate speed for Ij to 1^ hours, or until the control eggs had divided once. The cell-materials are separated into three layers (fig. 61, a, b,). Sections of eggs killed as soon as removed show the oil at the small end of the pear-shaped egg, forming an oil-cap. A clear protoplasmic layer lies below this, while the broad end of the egg is filled with a mingled mass of yolk and mitochondria (fig. 62). If such eggs were removed to sea water and allowed to develop it was found that the first cleavage-plane cuts the egg without reference to the stratification


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA « 217

(fig. 63). The majority of the eggs cleave so that the different materials are equally distributed in the two blastomeres; often, however, the distribution is unequal. Complete separation of the kinds of material may occur, as when the cleavage plane comes in through the protoplasmic layer between the ends (figs. 63, b; 65, g; 66, a, b, e, /). Sections show this to be due to the position of the nucleus in regard to the stratification. Since the eggs are not centrifuged sufficiently to move the nucleus, it may lie in any relation whatever to the egg-materials. It is, however, usually found in the protoplasmic layer (fig. 64, a, d) in which case the egg divides so as to distribute the materials equally (fig. 63, /). The nucleus may also lie in the yolk-end of the egg (fig. 64, c) in which case the cleavage-plane comes in through the yolk (fig. 63, a). The nucleus not infrequently appears at one side of the protoplasmic layer (fig. 64, d), in which case the first cleavageplane separates the two kinds of material, one blastomere receiving in addition to part of the protoplasm, yolk and mitochondria and the other the oil mass.

Individuals showing various distributions of the materials were isolated and their development was followed, apparently normal swimming larvae resulting (fig. 65). The yolk and mitochondria are confined to one region of the resulting planula, the oil to another. Also eggs which cleaved so that the different materials segregated in the two blastomeres were cut apart, separating the blastomere containing yolk and mitochondria from the one containing oil (fig. 66) . These were isolated and the development followed, and although many died, I succeeded in getting a considerable number of these half larvae. If the protoplasmic area were nearly evenly divided between the two blastomeres, both were apt to live. So far as I could tell those that lived were perfectly normal, except in size, the one being small and white from its oil content, the other large and greenish from the yolk content. Both were ciliated planulae. Sections of these planulae which were killed in Meves' killing fluid and stained with Benda's method, show the mitochondria apparently unchanged in one, while the other contains none. It is apparent, then, that up to this point of development the mitochondria are not essential for


218 . CORA J. BECKWITH

development. Hydroid planulae are hard to carry beyond this point of development in the laboratory, since any disturbance prevents the planula from attaching, which is essential for further development. Even so slight a disturbance as changing the water in the dish is sufficient to prevent attachment, so that further development after so great a disturbance as separating the blastomeres is impossible.

The above experiments are of value only in indicating that the mitochondria are not essential for differentation as far as the planula-stage and that they can hardly be vital constituents of the protoplasm since they may be centrifuged out of the protoplasm like any metaplasmic body, such as yolk.

2. Discussion

Mitochondria in Hydractinia do not agree in a number of points with descriptions of these bodies in other forms. An extensive review of the subject will not be attempted, however, since it has been so well discussed and reviewed by many recent investigators (Faure-Fremiet '10, Prenant '10 b, Montgomery '11, Duesberg '11). Duesberg presents a monumental review of the literature on the subject in which he brings together in classified form and in the most exhaustive way, all papers concerned with mitochondria and chromidia. The opposing views as to the nuclear or protoplasmic nature of these bodies have already been stated. The above writers assume the identity of mitochondria and chromidia. Schaxel, on the other hand, considers them quite distinct, as has been sufficiently indicated above. It is also apparent from the aforegoing description that Hydractinia corresponds with Schaxel's observations in this respect, that is, that two distinct elements are present, mitochondria of undoubted protoplasmic origin and basophilic granules in the protoplasm. The latter, however, are not chromatic in Hydractinia and therefore not chromidia (extra-nuclear chromatin of Schaxel).

I also agree with Schaxel in finding that the pseudochromatingranules first appear in the early growth-period of the egg, thus giving no evidence of continuity from cell generation to generation.


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA 219

Again, Schaxel finds mitichondria already present in the early growth-period of the egg before his 'chromatin-emission' occurs. In this respect results with Hydractinia are different, since the mitochondria appear only after the egg is a third grown. Hydractinia is an exception to the rule in this, since in most forms the mitochondria are either aheady present or appear in the early growth-period. The possibility of direct nuclear origin of the mitochondria in Hydractinia is also excluded, since they arise de novo throughout the egg and are not collected in a definite body against the nucleus (yolk-nucleus) as described in many forms.* The mitochondria in Hydractinia do no contribute to the formation of the yolk, as described for other forms by a number of workers.^ I also find no grouping of the mitochondria around an idiozome, indicating their origin through the retrogressive development of the' 'centroplasm' as suggested by Vedjovsky. Again, Hydractinia gives no evidence of the mitochondria forming a part of the architecture of the protoplasm (Faure-Fremiet and others), since in centrifuged eggs they are carried to one pole of the egg together with the yolk, leaving a free layer of protoplasm which has the usual protoplasmic structure. Since the blastomere of a centrifuged egg containing no mitochondria develops into a swimming larva, they can hardly be vital units of the protoplasm. I find no indication of their multiplication by division

■* A yolk nucleus of mitochondria is described in the eggs of the stint (Lams '04, Arch. Anat. micr., T. 6); Rana (Lams '07, Arch. Anat. micr., T. 9); Proteus (Schmidt '04, Anat. Hefte, Bd. 27, and M. Jorgensen '10, Festschr., R. Hertwig); Testudo (Loyez '05, '06, Arch. Anat. micr., Bd. 8) ; chick (D'HoUander '04, Arch. Anat. Micr., T. 7); some birds (Loyez '05, .'06, Arch. Anat. micr., T. 8); human egg (Van der Stricht, '05, Bull. Acad. Belgique); cat, (Russo, '09, '10, Arch. f. Zellfor., Bd. 4-5); bat and guinea-pig (Van der Stricht '05, Compt. Rend. Assoc. Anat., Geneve); Ascaris (Schoonjans '09., Bull. Soc. Roy. Sci. Med., Brussels); Ciona intestinalis (Loyez '09, Assoc. Anat., Nancy).

^ Yolk is described as being formed directly from the mitochondria by Russo ('09, '10) in cat; Loyez ('09), ascidians and human egg, Compt. rend. Assoc. Anat., Paris; Faure Fremiet ('10, Arch. Anat. micr., T. 11); in Lithobius; Zoja ('91, Mem. del R. Inst. Lomb. di Sci., vol. 16) etc. Yolk is described as formed indirectly under the influence of mitochondria by Van der Stricht ('05) in the bat; Lams et Devorene ('08, Arch, de Biol., T. 23) in some mammals; Van Durne ('07, Ann. Soc. Med. de Gand, T. 88); Schooonjans ('09) Ascaris; Bluntschli ('04, Morph. Jahrb., Bd. 32) etc.


220 CORA J. BECKWITH

at any time during the growth-period or cleavage, as suggested by Duesberg ('10) and Faure-Fremiet ('10 a). Their origin and behavior in Hydractinia indicate that they may be either precociously differentiated portions of the protoplasm (Vedjovsky' 07) or metaplasmic bodies.

Scepticism as to the identity of the bodies described as plasmosomes, chondriosomes, chromidia, ergastroplasm, etc., has been expressed by a number of observers (Veratti '09, Penas '11, Lundegard '10, Gurwitsch '10 and others) who believe that they have nothing in common but their name. Such an impression is certainly gained in reviewing the literature. The experiments with staining tests lead the writer to join these investigators in the belief that structures which have the same staining reactions may have been confused. If the standard tests for mitochondria are to be relied on for identifying them, then the mitochondria of Hydractinia, which respond to these tests, do not conform in many respects to the conditions found in other forms.

IV. MATURATION PHENOMENA AND AMITOSIS IN HYDRACTINIA AND EUDENDRIUM

When this work was begun it was my purpose to re-examine the eggs of several Hydroids which had been described as showing no mitotic figures during maturation, a nuclear disintegration or fragmentation occuring at the time of the disappearance of the germinal vesicle. The suggestion was made that reduction phenomena of maturation may well be accomplished without any of the complex and spectacular processes of mitosis" (Hargitt '06). Nuclear reconstruction was described as occurring later through the collection of these fragments in several 'nuclear nests' throughout the egg (C. W. Hargitt, Pennaria, '04, Eudendrium '04, Clava '06; Allen, Tubularia crocea, '00); and the cleavage of the egg as frequently amitotic. Smallwood ('09) and G. T.' Hargitt ('09), have since established the occurrence of typical maturation phenomena and mitotic cleavage in Pennaria and Tubularia, while independent studies by the writer ('09) gave the same result both in Pennaria and in Clava. Since


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA 221

Eudendrium, among the above forms, has not been re-examined, a brief account of the maturation stgfges occurring in this form will be given here. This section is concerned chiefly with a brief account of the maturation stages in Hydractinia to supplement Smallwood's account which as he states is incomplete because of lack of material. I have been fortunate in finding the stages lacking in his description.

A. HYDRACTINIA

1. Nucleus of the growth-period

As described in the previous section, in the early growth-period the nuclear net is centered in the nucleolus from which it radiates (figs. 2, 3, 4). As the egg grows this radial arrangement is lost, the chromatin assuming a reticular form (figs. 7, 8, 25). While these early stages show the same nuclear structure after preservation with any killing fluid, later stages are profoundly modified. When preserved in sublimate-acetic or picro-acetic killing fluids, a coarse, deeply staining, granular reticulum appears in a colorless ground-substance (fig. 12). That this reticulum is a coagulation phenomenon is suggested bj^ comparing this nucleus with those . shown in figures 19, 20, 21 and 23, which are sections of eggs killed in Meves' killing fluid. Here the nuclear net, which has changed into a fine net, stains more Hghtly in basic dyes than after the former fixatives, and lies in a finely granular, homogeneous ground substance which stains hghtly with plasma stains (figs. 18, 19, 25). Figures 20 and 28 show a shghtly older stage in which a process of diffusion of the chromatin net, previously begun, has proceeded until there is just a suggestion of the net in the homogeneous ground-substance. This leads directly to the condition shown in figure 21, in which all trace of the net has disappeared and only the ground-substance is left. The question of the method of disappearance of the chromatin will be taken up in detail a little later. The net completely disappears before the egg is one-third grown. The nucleus lies at the center of the egg until near the end of the growth-period when it moves to the free surface of the egg.

JOURNAL OF MOHPHOLOGY, VOL. 25, NO. 2


222 CORA J. BECKWITH

2. Maturation stages

The maturation phenomena take place while the egg is still in the gonophore as Smallwood ('09) states, and not after leaving it (Bunting '94). The lightly staining, homogeneous condition of the nucleus, which has long been recognized as a characteristic of the hydroid egg, and which has led to much confusion concerning the maturation stages, persists until the nucleus breaks down to form the chromosomes. The first indication of reappearing chromosomes occurs in gonophores killed from 20 to 30 minutes before eggs from the same colony are deposited. Since the reformation of the chromosomes is best seen in material preserved in fluids which do not cause a heavy precipitate, the following description is based on material fixed either in Meves' or Flemming's fluid, neutral formalin, or hot water. The deeply staining nucleolus is usually still present at the inner border of the nucleus ; its history will be described more fully later.

Out of the apparently homogeneous ground-substance of the resting nucleus, there appear very lightly staining threads on which are groups of granules staining a little more intensely than the threads. These threads appear in pairs, either parallel or X like in form (figs. 30, 31). In either case many fine branches merge from the main threads into the general homogeneous ground-substance. The number of these groups of threads in a single nucleus corresponds to the haploid number of chromosomes (12 or 14). The exact haploid number has not been determined but since 14 such pairs of threads is the number most frequently found in a nucleus and since 14 tetrads is the usual number found in the later stages (figs. 4, 46), it seems certain that this represents the haploid number and that they are bivalent chromosomes. The chromosomes now condense rapidly into tetrads which are very much smaller than the crosses. The initial stage in this process consists in a shorting of the arms of the X and the collection of the granules in a mass at the center, the latter taking a slightly deeper stain than before (fig. 32). In a later stage (fig. 33) the ends of the arms of the X are still visible, although the bulk of the granules appear at the center. The condensation


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA 223

consists apparently in the migration of the granules along the linin-threads toward the central point, leaving the lightly staining net merging into the ground-substance. The chromosomes now condense rapidly into small compact tetrads which stain intensely. A number of stages in the formation of a tetrad are shown in figure 24. The nuclear membrane breaks down about this time and the chromosomes which have been scattered through the nucleus collect at the center (fig. 24).

Up to this point there has been no evidence of a spindle. When it does appear, the chromosomes are already collected in the center of the nuclear area (fig. 35) and the spindle apparently arises in connection with them from the achromatic portion of the nucleus. The spindle, which is many times smaller than the nucleus, differs entirely in structure according to the fixation. The general topography of such a spindle in the center of the nuclear area after Meves' fixation is shown in figm-e 37. The much enlarged spindle shown in figure 38 makes it clear that no centrosomes or astral radiations are present, the blunt spindle lying free in the nuclear area. If material is fixed in sublimateacetic solution, in addition to the spindle, small asters appear, which are continuous with the coarse net present in the nuclei of such eggs (fig. 36) . They give every appearance of being part of the coagulation phenomena caused by the killing fluid. In some cases a small centrosome-like body occurs at the center of the aster. But since it is not constant and may be asymmetrically placed in regard to the spindle, it also seems to be a result of the coagulation. In fact, I have not been able to find a true divisioncenter, either outside or within the nucleus. The spindle fibers arise, apparently independently of an aster or centrosome, directly out of the nuclear ground-substance, for the spindle appears directly in the center of the large nuclear area, a considerable layer of the nuclear plasma surrounding it. The tetrads are now drawn on to the spindle and become arranged in an equatorial plate (figs. 37, 38). As stated above, the number of tetrads in the equatorial plate has not been definitely determined, the number being between 12 and 15. Since 14 was more commonly present and it is the number appearing in the polar body, it seems probable


224 CORA J. BECKWITH

that this is the haploid number (figs. 39, 40, 46). A small element near the center of the plate is characteristic.

The spindle, still without asters or centrosomes, now rotates 95° until it is perpendicular to the surface, and then moves out of the nuclear area toward the surface, where the first polar bodyis formed (figs. 43, 44, 45). As seen in some of the figures, a single or double granular mass — the remains of the nucleus — is left behind in the protoplasm, where it is absorbed. Since the manner in which the tetrads are formed is not determined, it is impossible to interpret this division in terms of reduction. The second polar spindle is formed immediately, as shown in figure 46. The egg is now shed from the gonophore, a small female germ nucleus being very rapidly reconstructed from the remaining chromatin (fig. 47) . The very great difference in size between this nucleus and the germinal vesicle has been sufficiently emphasized by previous writers on Hydi'oids. The egg, which has been somewhat flattened in the gonophore, rounds up when shed.

This description of the formation of the first polar spindle differs from that given by Smallwood, who finds it appearing with asters in the protoplasm in connection with a very small nucleus which is many times smaller than the typical germinal vesicle of Hydractinia. The small size of the nucleus and the position of the spindle outside of the nucleus makes it probable that he has figured the first cleavage-nucleus and spindle. Since I have traced consecutive stages in the breaking down of the germinal vesicle and the formation of the chromosomes and spindle, (stages which Smallwood lacked) it seems conclusive that a blunt spindle minus asters and centrosomes lying in the center of the large nuclear area, is typical for Hydractinia. This conclusion is supported by the condition found in other forms, since a blunt spindle minus asters has been described in a number of Hydroids (Gonothyrea, Wulfert, '02; Clava squamata, Harm '02; Clava leptostyla, Beckwith '09; Eudendrium, present paper; Linerges, Conklin '08; Cordylophora, Morningstein '01; Cunina, Stschelkanowzew '06). Further the formation of the spindle within the nuclear area itself also find support in the condition described


PLASMA-STRUCTUKE IN EGG OF HYDRACTINIA 225

for Clava by Harm ('02) who finds the first polar spindle very small and formed within the germinal vesicle from the achromatic portion, and the same condition in Cordylophora, as described by Morningstein ('01).

3. Fertilization

The egg is fertilized as soon as it is shed, the spermatozoon entering at any point on the surface. A f^ilization-membrane is formed (fig. 42). The method of union of the two germ-nuclei depends on the point at which the spermatozoon enters. If, as often happens, it enters near the female germ-nucleus (fig. 47) the sperm head may en,ter the egg-nucleus bodily, without expanding (fig. 49). Here the chromatin of the egg-nucleus is already collected in masses to form the chromosomes of the first cleavagespindle. An aster accompanies the sperm head in this case as in others, no spindle as yet having formed, however. If the sperm enter at some distance from the female germ-nucleus, it enlarges as 'Usual before union and a spindle develops in connection with it. The 'degree of enlargement varies, as shown in figures 48 and 50. Complete fusion of the two nuclei may take place before the breaking up into chromosomes but this is not essential.

The entrance of the sperm head directly into the egg nucleus without expansion is 'described in some other Hydroids (Wulfert, '02, Gonothyrea; Harm, '02, Clava squamata ; I have also observed it in Pennaria) . Also as the figures show, I do not find the sperm head in Hydractinia forming a group of vesicles as figured by Smallwood ('09) in Hydractinia and Smallwood ('09) and Hargitt ('09) for Pennaria. -It expands directly into a single vesicle.

4- Fertilization -membrane and mitochondria

The hydroid egg is usually spoken of as naked (Wilson '00, Hargitt '04). Smallwood ('09), however, has described a membrane for Pennaria, formed at fertilization. I have found this membrane easy to demonstrate in fresh material if intra vitam staining methods described by Kite ('12) are used. In Hydrac


226 CORA J. BECKWITH

tinia a membrane is also present. Sections of an unfertilized egg show a yolk-free area at the surface of the egg in which mitochondria are scattered irregularly (fig. 41) . No distinct membrane can been seen at this time. Sections of fertilized eggs show the mitochondria arranged in a distinct layer directly at the surface of the egg, a more or less free space being left between them and the yolk-spheres. Outside the layer of mitochondria a very thin, transparent layer or membrane appears, which as a rule clings closely to the surface of the egg (fig. 42) . The method of formation of the membrane was not determined.

5. Nucleolus

As stated earlier, a nucleolus is already present in the egg after the last oogonial division, when the chromatin is in the bouquetstage (figs. 1, 2, 3). Its origin was not determined and it stains intensely in basic dyes. As the chromatin breaks up into a net, it is still pressed against the nuclear wall and may already show vacuoles (fig. 4). When the radial arrangement of the net ^ is lost, the nucleolus is no longer flattened against the nuclear wall, but may be well toward the center of the nucleus (figs. 7, 18). In general, however, it retains its excentric position (figs. 19, 21, 22). During the growth-period the nucleolus increases in size and becomes vacuolated, the skeleton retaining- its intense staining capacity for basic dyes while the vacuoles stain with acid dyes. One large vacuole may occupy the center, leaving a rim of basic staining material (fig. 29 b), or many fine vacuoles may appear, making the nucleus more or less spongy (figs. 27, 28, 29 a). Both conditions are typical for Hydroids. The increase in size continues throughout the greater part of the growth-period, a point in which Hydractinia differs from Pennaria, as described by Hargitt, since in the latter form growth stops and nucleolar disintegration begins as soon as the spireme is completely broken up. About the time of maturation — usually before the breaking down of the nuclear membrane — the nucleolus dwindles and disappears in the substance of the nucleus by a process of dissolution and not by fragmentation as described by Hargitt for Tubularia. It may have


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA 227

disappeared before the reappearance of the chromosomes (fig. 23), or, as is more common, it ma} be disappearing as the chromosomes are reforming (fig. 29,, a).

Because of its staining reaction, the nucleolus was at first thought to be chromatic. Since the nucleus, throughout the greater part of the growth-period, is non-staining with basic dyes while the nucleolus takes these stains intensely, it at first seemed evident that the chromatin which .is to form the chromosomes is stored in the nucleolus during this stage, a condition described for a number of forms.® This conclusion proves to be unfounded after further staining tests, since the nucleolus does not always take the characteristic chromatin-stain (table 1.) For example, in Hermann's saffranin methyl-violet stain, the nucleolus is red while the chromatin of the 'resting' nucleus is violet. Also, in Bensley's acid fuchsin methyl-green stain the nucleolus stains with the fuchsin or plasma stain. Again, in x^uerbach's stain, which ordinarily stains the chromatin green, the nucleolus is stained blue. The strongest evidence that the nucleolus is not chromatin is given by Benda's stain, after which the chromatin is yellow-brown and the nucleolus ^dolet. This point is confirmed by Dublin ('05) who finds decisive proof with Auerbach's stain that the basic staining- nucleoli of Pedicellina are not chromatin. The nucleolus in Hydractinia has at no time any direct connection with the chromosomes as described by Guenther ('03) and Dublin ('05). There is no^ evidence then that the disappearance of the nucleolus at the time of the reappearance of the chromosomes bears any relation to the same. From the above description of the nucleolus in Hydractinia, it is evident that no fragmentation of the nucleolus contributes to the formation of basic granules (pseudochromatin-granules) which lie in the protoplasm, as Hargitt finds in Campanularia.

^ Chromatic nucleoli have been described among hydroids in Forskalia and Agalma by Schaxel ('11); Gonionemus, Bigelow ('07); Campanularia, Hargitt ('13); Cubomedusa, Conant ('98) Hydra, Downing ('09); Gonothyrea, Wulfert ('02). Gtinther ('03) finds the nucleolus in the echinoderm egg forming out of the nuclear net, and the chromosomes reappearing from the nucleolus.


228 CORA J. BECKWITH

6. Chromosomes and continuity

My results contribute little to the solution of this question, but since some evidence in its favor appears in the nucleus of Hydractinia eggs, it can hardly be passed by without comment. As previously stated, when the chromatin becomes diffuse, the netlike arrangement of the chromatin is lost. This is brought about by the threads becoming gradually arranged in groups of two either parallel or cross-like threads (figs. 25, 26). These crosses or parallel threads gradually fade out completely, leaving the characteristic lightly staining nucleus (figs. 21, 27, 28). That the chromosomes reappear in exactly the same way in which they disappear, that is, as X's and parallel threads, is evident by comparing the figures just described with those of reappearing chromosomes (figs. 29, 1, 30, 31). In fact, they are so exactly alike in form, in many cases, that whether they are disappearing or reappearing can only be decided by other conditions in the egg and nucleus. It is hard to resist the impression that this identity in their method of disappearance and reappearance is significant of some sort of continuity.

B. EUDENDRIUM RAMOSUM

Maturation and fertilization

The complete history of the chromatin in the maturation of Eudendrium has not been worked out, but such stages as I have establish its regular character. Maturation, fertilization and development of the egg up to the planula stage take place in the gonophore. The position of the nucleus after it leaves the center of the egg differs from most other hydroids, since it does not lie at the free surface of the egg as it ordinarily does, but at the inner end which leads to the cavity of the gonophore (fig. 51). The maturation and fertilization stages take place at this point. I have sections only of material fixed in sublimate-acetic killing fluid, so that there is always present in the resting nucleus the heavy net-like structure caused by this fixation.


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA 229

During the growth-period the nucleus consists of a fine net with a deeply staining nucleolus present (fig. 52), the whole very similar to a corresponding stage of Hydractinia after the same killing fluid (fig. 22). The nucleus grows enormously as is shown by comparing figures 52 and 53, the latter being ready for maturation. Because of the coagulation phenomena, I have been unable to trace the reappearance of the chromosomes in this form. Clumps of chromatin lying in the nuclear net give the first indication of reappearing chromosomes (fig. 53). The nucleolus here, as in Hydractinia, increases much in size throughout the growth-period, becomes vacuolated, and disappears before the polar spindle is formed. The origin of the spindle was not determined since the first spindles seen were completely formed and, in the equatorial plate state (figs. 52, 55), perpendicular to the surface. The spindle is of much smaller size than the nucleus and is also without asters and centrosomes. The chromosomes are not in the form of tetrads and, since their origin is not known, nothing can be said of their quadripartite condition (fig. 55). I have also too few sections of the equatorial plate to establish definitely the haploid number. Since 13 is the most constant number appearing, it is undoubtedly near the reduced number (fig. 56) . Two stages in the formation of the second polar spindle are shown in figures 57 and 58. A female germ-nucleus (fig. 59), characteristically smaller than the germinal vesicle (fig. 53), is reconstructed after the last polar division. The only cases of fertilization which I have observed show the two germ-nuclei of equal size (fig. 60), indicating that the sperm enters the egg early and so expands before meeting the egg nucleus. I have found no spindle in connection with the fusion nucleus, but the fact that the protoplasm killed at this period is not well fixed, may account for this. Development was carried no farther, the question of cleavage and amitosis not being studied in this form. The establishment of a single cleavage-nucleus makes nuclear fragmentation and subsequent reorganization in 'nuclear nests' impossible and amitotic cleavage improbable.


230 CORA J. BECKWITH

V. SUMMARY

1. The bulk of the evidence from staining reactions and morphological conditions indicates that those protoplasmic granules in Hydractinia which often take the chromatic stains are not extruded from the nucleus as such, and points to their formation de novo throughout the protoplasm. The granules are therefore not comparable to the chromidia of Hertwig and the term 'pseudochromatin-granules' is justified.

2. There is no evidence of formed material passing through the nuclear membrane into the protoplasm either early (Schaxel) or late (Smallwood) in the growth-period. Globules, which may be mistaken for such material, are formed during the growth-period, flattened against the nuclear wall, but their staining reactions under certain conditions differ from those of chromatin.

3. We are still unable to differentiate with certainty by any of the above methods, the nuclear derivatives in the protoplasm, i.e., staining tests cannot be relied on as tests for chromatin.

4. The yolk is formed in Hydractinia directly from the scattered pseudochromatin-granules and not from nests of granules.

5. The pseudochromatin-granules correspond in a general way to the yolk-nucleus of many other forms.. Here the granules are neve'r gathered into a distinct body.

6. The pseudochromatin-granules are completely used up in the formation of the yolk ; that is, none, as such, are left in the protoplasm between the yolk-spheres (intravitelline chromatin of Schaxel) to determine further differentiation.

7. The yolk is not formed from mitochondria in Hydractinia and there is no yolk-nucleus consisting of mitochondria.

8. The mitochondria are not of nuclear origin in Hydractinia, but arise de novo in the protoplasm after the formation of the yolk has begun. I find no evidence of multiphcation of mitochondria by transverse division, or of their genetic continuity from one cell-generation to another.


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA 231

9. The mitochondria and chromidia (extra-nuclear granules) are not identical in Hydractinia.

10. The mitochondria are not a vital part of the protoplasm in Hydractinia but are a highly differentiated product.

11. Maturation and fertilization are typical in Hydractinia and Eudendrium and cleavage is mitotic.


232 CORA J. BECKWITH

BIBLIOGRAPHY

Allen, C. M. 1900 The development of Parypha crocea. Biol. Bull., vol. 1.

Beckwith, C. J. 1909 Preliminary report on the early history of the egg and embryo of certain hydroids. Biol. Bull., vol. 16.

Benda, C. 1902-03. Die Mitochondria. Erg. d. Anat. u. Entw., Bd. 12.

Bensley, R. R. 1906 Methods for microchemical detection of phosphorus compounds other than phosphates. Biol. Bull., vol. 10.

BiGELovt', H. B. 1907 Studies on the nuclear cycle of Gonionemus murbachii. Bull. Mus. Comp. Zool. Harv., vol. 48.

Bunting, M. 1894 The origin of sex cells in Hydractinia and Podocoryne. Jour. Morph., vol. 9.

CoxNANT, F. S. 1898 The Cubomedusae. Mem. Biol. Lab. Johns Hopkins Univ., vol. 4.

CoNKLiN, E. G. 1908 The habits and early development of Linerges mercurius. Cam. Inst. Washington, no. 103.

Crampton, H. E. 1896 Early history of the ascidian egg. Jour. Morph., vol. 15.

Downing, E. R. 1909 The ovogenesis of Hydra. Zool. Jahrb., Bd. 28.

Dublin, L. I. 1905 On the nucleoli in the somatic and germ cells of Pedicellina americana. Biol. Bull., vol. 8.

DuESBERG, J. 1910 Sur la continuite des elements mitochondriens des cellules sexuelles et des chondriosome des cellules Embryonnaires. Anat Anz Bd. 35.

1911 Plastosomes, 'Apparato interno' und 'chromidial apparat.' Erg. d. Anat. u. Entw., Bd. 20.

Faur^-Fremiet, E. 1910 a La continuite des mitochondries a travers les generations cellulaires et le role de ces elements*. Anat. Anz., Bd. 36

1910 b Etudes sur les mitochondries des Protozoaires et des celles sexuelles. Arch. Anat. micr., Bd. 11.

Fisher, A. 1899 Fixirung, Fiirbung und Bau d. Protoplasm. Jena

Foot, K. 1896 Yolk-nucleus and polar rings. Jour. Morph., vol. 12.

GoLDSCHMiDT, R., and Popoff, M. 1907 Die Karyokinesis der Protozoen und

der chromidial Apparat der Protozoen und Metazoen Zelle. Arch. f.

Protistenkunde, Bd. 8.

GuNTHER, K. 1903 Ueber den Nucleolus im reifenden Echinodermen-Eie und seine Bedeutung. Zool. Jahrb., Bd. 19.

GuRWiTCH, A. 1910 Die Hauptstromungen in der Cytologic des verflossenen Jahrzehnts. Biol. Zeitsch., Bd. 1.

Hardy, H. W. 1899 On the structure of the cell protoplasm. Jour Phys vol 24.


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA 233

Hargitt, C. W. 1904 a The early development of Eudendrium. Zool. Jahrb., Bd. 20.

1904 b The early development of Pennaria tiarella. Arch. f. Entw'mech. d. Org., Bd. 18.

1906 The organization and early development of the egg of Clava leptostyla. Biol. Bull., vol. 10.

Hargitt, G. T. 1909 Maturation, fertilization and segmentation of Pennaria tiarella and of Tubularia crocea. Bull. Mus. Comp. Zool. Harv., vol.53.

1913 Germ cells of coelenterates. I. Campanularia flexuosa. Jour. Morph., vol. 24.

Harm, K. 1902 Die Entwicklungsgeschichte von Clava squamata. Zeit. wiss. Zool., Bd. 73.

Heidenhain, M. 1911 Plasma und Zelle. Fischer, Jena.

Hertwig, R. 1907 Ueber den Chromidialapparat und den Dualismus der Kernsubstanzen. Sitzb. Ges. Morph. u. Phys. Miinchen.

Kite, G. L. 1912 The nature of the fertilization membrane of the egg of Arbacia. Science, vol. 36.

1913 Studies on the physical properties of protoplasm. Am. Journ. Phys., 32.

Lee, a. B. 1903 Microtomist's vade mecum; 6th edition. Churchill, London.

Lilienfeld, L. 1893 Ueber die Verwandschaft der Zellelemente zur gewissen Farbstoffen. Arch. Anat. Phys.

LuNDEGARD, H. 1912 Fixieruug, Farbung und N omenklature der Kernstrukturen. Arch. mikr. Anat., Abt. Entw., 80

Mann, G. 1902 Physiological histology; methods and theory. Oxford, Clarendon Press. '

Mathews, A. P. 1898 A contribution to the chemistry of cytological staining. Am. Jour. Phys., vol. 1.

Montgomery, T. H. 1911 The spermatogenesis of an Hemipteron, Euschistus. Jour. Morph., vol. 22.

Morningstein, p. 1901 Untersuchungen liber die Entwicklung von Cordylophora lacustis. Zeit. wiss. Zool., Bd. 70.

Nemec, B. 1910 Das Problem der Befruchtungs Vorgangen. Leipsig.

Pensa, a. 1911 Ancore di alcune formazioni endocellulari dei vegetali. Anat. Anz., Bd. 31.

Prenant, a. 1910 a "M^thodes et resultats de la microchimie. Journ. de I'Anat. et Phys., T. 46.

1910 b Les mitochondries et I'ergastoplasme. Journ. de I'Anat. et Phys., T. 46.


234 CORA J. BECKWITH

ScHAXEL, J. 1909 Die Morphologic des Eiwachstums und FoUikelbildungen bei Ascidien. Arch. Zellforsch., Bd. 4.

1910 Die Eibildung der Medusa Pelagia noctiluca. Festschr. R. Hertwig.

1911 a Das Zusammenwirken der Zellbestandteil bei Eireifung, Furchung und ersten Organbildung der Echinodermen. Arch. mikr. Anat., Bd. 76.

1911 b Das Verhalten des Chromatins bei der Eibildung einiger Hydrozoen. Zool. Jahrb. Abt. Anat. u. Ont., Bd. 31.

1911 c Plasmastrukturen, Chondriosomen und Chromidien. Anat. Anz., Bd. 39.

1912 a Versuch einer cytologischen Analyse der Entwicklungsvorgange Erste Teil, Aricia. Zool. Jahrb., Abt. Anat. u. Ont., Bd. 34.

1912 b Weiter Untersuchungen iiber die Eibildung der Meduse Pelagia Jen. Zeit., Bd. 48.

Smallwood, M. W. 1909 A ree.xamination of the cytology of Hydractinia and Pennaria. Biol. Bull., vol. 17.

Stschblkanowzew, J. 1906 Die Entstehung von Cunina proboscidia. Mitth . Zool. Sta. Naples, Bd. 17.

Trinci, G. 1907 Studii suU' ovcyte dei Celenterati durante il periodo di crescita. Arch. Ital. Anat. Emb., T. 5.

Van Herwerden, M. 1913 Ueber die Nucleasewirkung auf tierische Zellen. Ein Beitrag zur Chromidienfrage. Arch. Zellforsch., Bd. 10.

Vedjovsky, F. 1907 Neue Untersuchungen liber 3ie Reifung und Befruchtung. Kon. Bohn. Ges. der Wiss. Prag.

Veratti, E. 1909 Sulla fina struttura della cellule di alcuni tumori. Boll. Soc. Med. Chir., Pavia.

Wilson, E. B. 1900 The cell in development and inheritance. Columbia Press, New York.

WuLFERT, J. 1902 Die Embryonale Entwicklung von Gonothyraea loveni. Zeit. wiss. Zool., Bd. 71.

Zacharias, E. 1896 Ueber einige mikrochemische Untersuchungs-methode. Ber. d. deutschen Bot. Ges., Bd. 14.


PLATES


235


PLATE V

EXPLANATION OF FIGURES

Hydractinia echinata* X 2625

1 Two egg cells resulting from the last oogonial division; n, diplotene stage of spireme-thread; h, early bouquet-stage.

2 Early bouquet-stage, showing loops centered in the nucleolus.

3 Late bouquet-stage, showing the nucleolus pressed against one end of the nucleus.

4 to 8 Five stages in the early growth-period, showing the accumulation of pseudochromatin-granules' throughout the protoplasm, the appearance of globules against the nuclear wall and oil in the protoplasm; semidiagrammatic. Granules are represented by gray, oil by circles containing dots, nuclear globules by circles containing parallel lines.

5 to 11 Three later stages of the growth-period, showing the direct development of the yolk-spheres from the pseudochromatin granules; semidiagrammatic. Young yolk-spheres shown in gray.

12 Egg about one-third grown, showing the appearance of mitochondria between the simple yolk-spheres, some of which are now^ mature and possess a different staining reaction ; semidiagrammatic. Mature simple yolk is represented by circles; mitochondria by black rods, young yolk and oil as above.

13 A nearly mature egg showing compound yolk-spheres in addition to the above elements, as well as more mature simple spheres ; diagrammatic. Compound spheres are represented by circles containing small circles.


' All figures except those of plate 8 and figure 61 were drawn with a camera.

All drawings were made from material fixed in Meves' or Flemming's killing fluid unless otherwise stated. • ,

236


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA

COR,V. J. BECKWITH


PLATE 1






^\.







©




©





jj^O . ,. ©









©


6'


o


rs




. &





^










®Q


, \ e






f.


©



,••;■•,*•••'• ••■•:'.>.:•-..!,'•••.


11




237




V -. '\> , • '/


"'^


' ^




hi


JOURNAL OF MORPHOLOGY, VOL. 25, NO. 2


PLATE 2

EXPLANATION OF FIGURES

Hydractinia echinata

14 Mature egg, showing the arrangement of mitochondria, simple and compound yolk, and oil; diagrammatic. Symbols as in plate 1. X 2625.

15 Types and development of compound yolk-spheres. X 2625.

16 Mature egg as shown after staining with iron-hematoxylin. X 2625

17 Young oocyte, showing radial arrangement of nuclear net. Meves' killing fluid. X 825.

18 to 20 Older eggs showing the gradual disappearance of the nuclear net in a homogeneous ground-substance. Meves' killing fluid. X 600.

21 Nucleus showing the complete disappearance of the nuclear net, a granular ground-substance only remaining. Meves' killing fluid. X 600.

22 The same, after sublimate-acetic killing fluid, showing a strong coagulation net in a colorless ground-substance. X 600.


238


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA

CORA J. BECKWITH


PLATE 2





^ h X



/.J .


■■.■•• • '• fir •;. * , V




V .■


r^-^FsSK-2S:


///




^;.».-"r-.:;v^;,




239





^(/


•fV*,' ' *.







PLATE 3

EXPLANATION OF FIGURES

Hydractinia echinata

23 Nucleus just before the reappearance of the chromosomes, showing the homogeneous ground-substance devoid of chromatin net and the nucleolus already much reduced in size. X 600.

24 Chromosomes reappearing in the nuclear area, the nuclear wall having broken down. X 600.

25 to 28 Four stages of nuclei in the early growth-period, showing the disappearance of the nuclear net by the chromatin arranging itself in crosses and parallel threads which become gradually fainter. X 1225.

29 a The reappearance of a chromosome, just before maturation, in the form of a cross, the nucleolus with many vacuoles, b, nucleolus with one large vacuole. X 1225.


240


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA

CORA J. BECKWITH


PLATE 3


•• c".. • • . •




- ^^


■■^*-::


.v^



s^^^^fe^


'"■•V' ■'■■


'«■■'•>•'■• -, -.



,,


Jl^


'%•>



""'

y *

. • * ^ 5^









'.'•..


- . - w .


.■^- •


■'


• r /••'1?!'

■-'.*^-:


"^


•'.♦- »♦".•, ..■.


■ " ' .:'


• *.o;, ,<



•■■.4r- :




'


».v« ?■





..- »^,>'-'



M


^


«^


\



X ■


¥■



k-^


■J %


^.,. -- ^


^^>/^^


« 


^7


•"I,


241


e


PLATE 4

EXPLANATION OF FIGURES

Hydractinia echinata

30 and 31 Two drawings from the same nucleus, showing the reappearance of the chromosomes in the form of crosses or parallel threads. X 1225.

32 and 33 Two stages in the condensation of a cross-shaped chromosome. X 1225.

34 Various stages in the late condensation of the crosses and parallel threads into tetrads, all found in the nucleus of figure 24. X 1818.

35 Partially condensed tetrads arranged in an equatorial plate. X 1818.

36 First polar spindle in the center of the nuclear area, showing astral radiations which form a part of the heavy nuclear net caused by sublimate-acetic fixation, X 600.


242


PLASMA-STRUCTURE IN EGG OF HYDRACTINLA

CORA J. BECKWITH


PLATE 4







243


PLATE 5

EXPLANATION OF FIGURES

Hydractinia echinata

37 The first polar spindle (having no asters) in the center of the homogeneous nuclear area, which is characteristic after Meves' fixation. X 600.

38 The first polar spindle, showing the tetrads becoming arranged at the equator. X 2625

39 and 40 Two equatorial plates of the first polar spindle, showing 15 and 14 chromosomes respectively. X 2625

41 Section to show the arrangement of cell-elements before fertilization. X 2625

42 Section to show rearrangement of mitochondria and fertilization membrane after entrance of the sperm. X 2625.

43 to 45 Three stages in the formation of the first polar body, the spindle having moved out of the nuclear area which remains as two homogeneous areas in the protoplasm. X 2625

46 The formation of the second polar body. The first polar body shows 14 chromosomes. X 2625


244


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA

CORA J. BECKWITH


PLATE 5







4 s




'* Q


•■•V




-/f>




'«*?-:*<


-^^



ttttA


_. *■'.•,••




\..:f^ .


- •: ^ ■'■^;*-*. .


,






>'„■■: ..« '«'■.. • •.«■ - ." ; ; •■■v. ■■■.-7:"



. > ,. . .■■ A**-;* ,....







' •. - '


■■■•^^'■.//■•"•■'•■♦—* *•■>■•■* ••.'■


^v^^'^••^••■•■^^^■*■'■•*^-•■•"■^• :


.





iy^,-;>*^;"





• • »'■.■•


44




•>-•■




.' *.*■' •"■- "-^•.


245


^o


PLATE 6

EXPLANATION OF FIGURES

Hydractinia echinata

47 Early fertilization stage in which the sperm has entered near the female germ-nucleus. X 2625.

48 Later fertilization stage showing the male germ nucleus somewhat expanded. X 2625

49 Fertilization stage in which the sperm head has entered the female germnucleus before expanding, the chromosomes of the egg nucleus being already formed. X 2625.

50 Fertilization stage in which the two germ nuclei are of more nearly equal size before union. X 2625.

Eudendrium ramosum^

51 Gonophore containing an egg showing the position of the nucleus at the inner end. X 75

52 Same nucleus much enlarged, growth period. X 2625.

53 Nucleus ready for maturation, showing the chromosomes forming. X 2625.


' All Eudendrium sections are of material fixed with sublimate acetic.

246


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA

CORA J. BECKWITH


PLATE 6


/




^■:.^.--'-'^'.:





^\




x;


52


' ..'.^



-JrU


X~'\.


•^.:^,.;Mv



i


1' ■ - • .'-^■"■■






y


- • . ' - .■ r-^ •."';


PLATE 7

EXPLANATION OF FIGURE.S

Eudendrium ramosiim

54 and 55 Two stages of the first polar spindle. X 2625.

56 Equatorial plate of first polar spindle, showing 3 chromosomes. X 2625.

57 and 58 Two stages in the formation of the second polar body. X 2625

59 The female germ nucleus. X 2625.

60 Fertilization stages, showing the two germ-nuclei of equal size. X 2625.

Hydractinia echinata

61 Two centrifuged eggs showing the egg materials separated into three layers. X 85.

62 Section of a centrifuged egg showing the distribution of the cell-elements, oil at the narrow end, a protoplasmic layer below, yolk and mitochondria at the broad end. X 338.


248


PLASMA-STRUCTURE IN EGG OF HVDRACTINIA

CORA J. BECKVVITH


PLu\TE 7



<),)



I


^?t"^^


\


U


-A)


f


.^




■<^


<r' ■■




"^^^- — -^




« 


3


C^'# •"


y.





^ #•



t. ■





V


^


^h"/'



.^^



61


62


m




"-»!'_'■*. V#,;j.>^;i.>^'


249


PLATE 8

EXPLANATION OF FIGURES

Hydractinia echinata. X 85

63 Centrifuged eggs, showing the direction of the first cleavage plane in reference to the stratification.

64 Sections of centrifuged eggs, showing the position of the nucleus in reference to stratification.

65 Centrifuged eggs which were isolated and development to planula followed : J and K show the resulting distribution of the materials in the planula.

G6 Centrifuged eggs in which the blastomeres were separated in the two-cell stage, isolated, and followed to the formation of the planula. F and G show several stages in this development.


250


PLASMA-STRUCTURE IN EGG OF HYDRACTINIA

CORA J. BECKWITH


PLATE 8




iT"



>h


(> 1



STUDIES ON CELL CONSTANCY IN THE GENUS EORHYNCHUSi

H. J. VAN CLEAVE From the Zoological Laboratory of the University of Illinois

FORTY-SIX FIGURES

CONTENTS

Introduction 254

1. Materials 254

2. Cell constancy and its relation to cell lineage 254

3. The fields of cell constancy 256

4. Methods 258

5. Cell constancy vs. nuclear constancy 259

Proofs of cell constancy in the genus Eorhynchus 260

1. Cuticula 260

2. Subcuticula .260

3. Lemnisci 266

4. Proboscis and proboscis receptacle 269

5. Proboscis wall and receptacle wall 271

6. Invertors of the proboscis 273

7. Retractors of the proboscis receptacle 273

8. Cement gland 274

9. Male genital apparatus 275

10. Female genital apparatus 277

11. Brain 278

12. Genital ganglion 279

13. Body musculature 279

Summary of cell constancy in the genus Eorhynchus, with table of results .... 280 General problems relating to cell constancy 281

1. Cell size vs. body size 281

2. Beginnings of cell constancy 285

3. Factors involved in the production of cell constancy 286

4. Cytomorphosis 287

5. Known extent of cell constancy in the animal kingdom 289

6. Significance of cell constancy 289

Bibliography 292

' Contributions from the Zoological Laboratory of the University of Illinois, under the direction of Henry B. Ward, No. 28.

253


JOURNAL OF MORPHOLOGY, VOL. 25, NO. 2


254 H. J. VAN CLEAVE

INTRODUCTION

1. Materials

During the fall of 1909 the writer began a study of the morphology of the Acanthocephala under the direction of Prof. Henry B. Ward. It happened that the form first taken up for study, Eorhynchus^gracilisentis (Van Cleave), displayed a marked uniformity in the number and arrangement of the nuclei in certain regions of the body. These observations opened the question of how far these phenomena of constancy existed in the other tissues and organs of the body. As additional species in the sariae genus came to the writer's attention, the analysis was carried on in a comparative manner until at length the present paper deals wi^ the results obtained from the study of five species, namely; Eo, gracilisentis (Van C. 1913), Eo. longirostris (Van C. 1913), Eo. emydis (Leidy 1852), Eo. cylindratus (Van C. 1913), and Eo. tenellus (Van C. 1913). Camera drawings of the probosces (figs. 1, 21, 29, 32 and 35), individual hooks (figs. 2, 30, 33 and 36) and embryos (figs. 3, 24, 34 and 37) of these species, are given in order to show what extreme differences of diagnostic characters may occur even though this study brings out the fact that in many regions of the body each individual nucleus of one species may be homologized directly with a nucleus in a member of a distinctly different species.

2. Cell constancy and its relation to cell lineage

While studies in cell lineage have added invaluable support to the idea of orderliness in vital processes they have failed to carry this idea to its legitimate conclusion. The development of the embryo through the cleavage stages of the ovum to the formation of the germ layers and even up to the beginning of differentiation in the anlagen of the organs of the adult body has been worked out within various species of animals, but most investigators have dropped the problem here. The more or less reg - The name Eorhynchus has been substituted for Neorhynchus, preoccupied, in a paper now in press.


CELL CONSTANCY IN THE GENUS EORHYNCHUS 255

ular procedure of cleavage and development up to this point has been very generally demonstrated in the forms whose cell lineage , has been determined, but it was left for such men as Weismann, Goldschmidt, and Martini to raise the question of a probable further continuation of an orderly process. It was already known that in many instances certain cells are very early set apart for the formation of definite groups of organs or tissues, but no one had actually determined how these organs arise, whether by mere indeterminate continuation of the process of cell division, or in an orderly, specifically predetermined manner. Might not the same forces at work upon and within the embryo during its cleavage stages continue to direct an accurately determined continuation of the cleavage throughout the course of development until the adult form is gained? The works of Goldschmidt and of Martini go a great way toward the confirmation of this hypothesis, at least within certain groups of animals. Their results indicate that in the forms studied mitosis has continued along very definite lines, resulting in a predetermined arrangement of the cellular elements of the adult organs and tissues. In his definition of cell constanc y Martini institutes a comparison, not alone of the numbers of n uclei and their relative positions, but also of the size, shape, and finer structure of the cells, and especially of the staining reactions of the nuclei of the cells.

The writer is surprised to find no reference in the more recent literature on cell constancy to Weismann' s exceptionally clear ideas as regards the phenomena of constancy. He writes on this point('93, p. 59):

In smaller and simpler organisms each individual cell may well be determined from the germ onward, and not only with the result that the number of cells is a definite one,' and the position of each definitely localized: the determination may also have caused individual peculiarities of each cell, in so far as they depend on changes in the germ plasm at all — i.e., are 'blastogenic,' — to appear in the corresponding cell in the next generation.

Such a statement of the fundamental principles underlying the theory of cell constancy coming from such a clear thinker as Weismann, even though at the time unsupported by any mass of


256 H. J. VAN CLEAVE

reliable data, is of extreme value in that it forshadows a field . of investigation from the viewpoint of the problem of the cytology of heredity before the exact facts in the case have been established.

3. The field of cell constancy

Martini has pointed out concisely the obstacles in the path of an investigation on cell constancy, in his comment upon the difficulty in determining either for or against any such position when extremely complicated relations exist between the component parts of a highly differentiated organ. It would be little less than folly to attempt a study of the cell constancy of an organ in one of the higher animals before the possibilities and the limitations of such a study in some of the simpler forms of life had been determined. As a similar instance, who would have considered the possibility of a regularity in the segmentation of the ovum of a vertebrate, had not the fundamental principles been marked out previously by the pioneers in the field of cell hneage, who chose for investigation such objects as displayed the phenomena of cleavage most clearly and with the least complication? Before going further it may be well to consider the nature of the limitations in the realm of cell constancy. Martini has very justly eliminated the cleavage stages of the ovum from consideration here, for the cells in these stages have not as yet arrived at a condition to which the term cell constancy could be applied properly, for in reality the field is limited to a study of such tissues and organs as have passed through the embryonic stage and are definitely differ- ' entiated as physiologically functioning organs. It must be kept in mind that though an organism may under normal conditions have portions displaying absolute constancy, yet in those regions where katabolic and secretory processes are proceeding at such a rate as actually to destroy the tissues, constancy, such as is found in the more stable regions which are practically free from radical metabolic changes, cannot exist. Likewise injury and mutilation of parts of the body with the accompanying excitation to renewed multiplication of cells, may profoundly alter the conditions even though the normal organism might display a


CELL CONSTANCY IN THE GENUS EORHYNCHUS 257

high degree of constancy in the number and the arrangement of its component cells.

Any tissue, the cells of which retain the power of continued multiplication after they have become differentiated, is said to possess the power of physiological regeneration. On the basis of the presence or of the absence of this property, Bizozzero has distinguished three different types of tissues within the human body. Morgan ('01, p. 128) quotes him as follows:

1. Tissues made up of cells which multiply throughout life, as the parenchyma cells of those glands which form secretions of a definite morphological nature; the tissues of the testes, marrow; lymph glands, ovaries; the epithelium of certain tubular glands of the digestive tract and the uterus; and the wax glands:

2. Tissues which increase in number of their cells till birth, and only for a short time afterward, as the parenchyma of the glands with fluid secretions, the tissues of the liver, kidney, pancreas, thyroid, connective tissue, and cariilage;

3. Tissues in which multiplication of cells takes place only at an early embryonic stage, as striated muscle and nerve tissues. In these there is no physiological regeneration.

There is no need of limiting this classification to the tissues of the human body for the general principle applies equally well to the tissues of any animal. The direct dependence of the idea of cell constancy upon the conditions set forth in the second and third groups of this scheme is strikingly evident without the need of further discussion. Those_ tissues which in most animals possess the power of physiological regeneration have been eliminated or profoundly modified in all forms displaying cell constancy. Two of the most obvious factors involved in accomplishing this reduction of inconstant tissues are modification of organs through degeneration and complete elimination of parts through adaptation to parasitism. Thus in the Acanthocephala adaptation to the parasitic habit has been so complete that all traces of the glands usually associated with the processes of metabolism are wanting. With the elimination of this great group of organs there is presented a condition most favorable for the development of a permanent, fixed, relation of the component parts of the body. In the genus Eorhynchus this finds expression in a surprisingly


258 H. J. VAN CLEAVE

marked cell constancy, the evidences of which will be brought out in this paper.

How strict an interpretation is to be placed upon this term ^constancy?' Is the term to be applied with absolute precision, utterly excluding the possibility of variation of any kind, or may some value be placed upon instances in which there is clearly a normal number and arrangement of the cells but in which a few individuals depart from this condition? The former of these two interpretations means the elimination of the possibility of variability and thereby precludes all possibility of evolution, for an absolute constancy could result in nothing else than an absolute fixity of species, since variability is one of the necessary or primary factors in evolution. The question of real importance is how far these phenomena may vary from the normal without entirely invalidating conclusions based upon the varying character in question. It is a very generally accepted fact that the application of stimuli or the action of internal enzymes may give rise to either an acceleration or a suppression of mitosis, thus producing abnormal numbers of cells and resulting in a confusing arrangement of the products. Since endoparasitism leads to a practical elimination of those varying environmental conditions which so profoundly influence the development of the freeliving forms, it would be quite natural to look among the endoparasites for materials adapted to the study of cell constancy. Such conditions have been found to exist in members of the genus Eorhynchus.

If. Methods

A method frequently employed in comparing the approximate number of cells in a given organ or region of the body of two individuals consists in making a count of the number of cells present in sections of the two animals at approximately the same plane. Such a method may be sufficient when the problem of relation of cell size to body size alone is being considered, but at best it can form the basis of but the crudest sort of a comparison and is valueless in determining the presence or absence of cell constancy. In order to avoid the possibility of misunderstanding regarding


CELL CONSTANCY IN THE GENUS EORHYNCHUS 259

the accuracy of the results set forth in the following study on cell constancy it should be understood that where a definite number of nuclei is recorded for a given organ that number has been obtained either by careful reconstructions of the part in question or by direct counts of all the nuclei in well cleared toto mounts. In no case has the writer based a strong claim of constancy upon the finding of a similar condition in as small a number as two individuals. In many instances even as high as two hundred individuals of the same species have been studied with the greatest care in order to establish the correctness of a finding. , For general results, killing and fixing in saturated aqueous solution of corrosive 'sublimate with about one per cent glacial acetic acid were by far the most satisfactory. Dilute aqueous solution of Ehrlich's acid hematoxylin proved best adapted for staining toto mounts. The chief advantage of this technic over all others consists in the remarkable translucence of the object after mounting in balsam. By this method the writer was enabled to study the cytology of even some of the larger forms in toto under a one-twelfth inch oil immersion lens. In the preparation of serial sections the same stain countered with eosin gave good differentiation, but for the finest work iron-hematoxylin was found most reliable.

5. Cell constancy vs. nuclear constancy

Since most of the cell walls in the Acanthocephala have disappeared, the question arises, whether the consideration of such syncitial structures may be included within the bounds of cell constancy or must be considered under an entirely separate heading which might be called nuclear constancy. To the writer the problem finds its solution in the known facts in the embryology of the group. Kaiser has recorded valuable information upon this question of the early development of the Acanthocephala. Regarding the determinate nature of the cleavage in the group he states ('13, p. 5): "Die Entwicklung der Echinorhynchen ist von der ersten Furchung bis zur Ausbildung des definitiven Wurmes vollig determiniert." Most writers have agreed that


260 H. J. VAN CLEAVE

the point wherein the Eorhynchi differ most essentially from other Acanthocephala lies in the fact that the nuclei of the mature worms in this genus are derived directly by the differentiation of the embryonic nuclei, while in the other genera there is introduced a period of indeterminate fragmentation of the nuclei at the time of the formation of the adult organs. In view of these facts the writer maintains that the term cell constancy is fully justifiable when dealing with the structure of the Eorhynchi, for each nucleus with the cytoplasm surrounding it represents a cell. Moreover the introduction of a wholly different term in the consideration of phenomena identical with those grouped under the head of cell constancy would but add confusion to the subject, for after all the wall is not the essential part of the cell.

PROOFS OF CELL CONSTANCY IN THE GENUS EORHYNCHUS

1. Cuticula

The cuticula of the Eorhynchi is a thin, homogeneous, layer covering the entire body. Even under a magnification of nine hundred diameters no specialized structure is visible. By virtue of its noncellular structure this body covering is readily eliminated from consideration in this connection.

2. Subcuticula

The subcuticula, which comprises by far the greater part of the body wall (figs. 4, 5 and 23 s), lies directly under the cuticula. While the general structure of this tissue in the Eorhynchi closely resembles that of other genera of Acanthocephala, the arrangement of the nuclei is decidedly distinctive of this genus. Another feature wherein these forms vary from their near relatives is that the component layers of the subcuticula are not as clearly defined as in many other genera. O. F. Miiller in 1780 (plate 61) figured Eo. rutih (Mtill.) and in the body wall showed five 'oscula,' as he called them later (1784), four of them on the dorsal side of the body and a single one on the ventral side near the anterior end. There can be no doubt that these structures are the giant nuclei


CELL CONSTANCY IN THE GENUS EORHYNCHUS 261

of the subcuticula, even though the actual discovery of the nucleus by Brown did not come until half a century later. Even as late a worker as Dujardin ('45) observed the swellings on the exterior of the body caused by these nuclei, though he did not understand their structure and significance. In regard to these same structures in the body of Eorhynchus agilis (Rud.), Dujardin ('45, p.538) writes: .... strie transversalementetpresentantdu cote convexe une serie longitudinale de cinq a six grands pores (?) ou disques orbiculaires,etunseuledu cote concave; . . . . " In speaking of the same species, Saefftigen ('84, p. 9) refers to four or five subcuticular nuclei in the dorsal canal and two in the ventral canal :

Bei Ech. clavaeceps haben die Subcuticulakerne einen machtigen

Umfang, ihr grosster Durchmesser iibersteigt oft 0.2 mm

Sie sind bei dieser Species nur in beschrankter Anzahl, ausschliesslich in den Kanalen, und zwar in den Hauptkanalen, deren ganzes Lumen sie einnehmen, vorhanden. Im Dorsalgefass finden sich ihrer vier bis fiinf, im Ventralgefass zwei, dessgleichen ein bis zwei in jedem Lemniscus.

Linton ('89, p. 490), in describing what he called Echinorhynchus agilis Rud., but was in all probability Eorhynchus cylindratus, notes large circular spaces in the vascular layer clearly defined by a circular thickened ring of connective tissue." Without further delay he accepts the name 'pores' or 'orbicular discs' introduced by Dujardin, to whose work reference has already been made. However, Linton's data differ from those of Dujardin, as indicated in the following: "In the specimens which I have examined," says Linton, "there does not appear to be either this regularity or proportion in the arrangement, e.g., one specimen had fom* nuclei on the concave side and two on the convex. In others they could not be made out definitely, but enough could be made out to show they were irregularly placed." However, in justification of modern development of histological methods it would scarcely be proper to leave this statement of Linton's position without noting the fact that his studies were upon living material under a compressor. While it is true that a good preparation of a stained toto mount does display the subcuticular nuclei with remarkable distinctness, yet observations of such nature must


262 H. J. VAN CLEAVE

ordinarily be subordinated to those upon carefully prepared serial sections if points of structure are to receive a final analysis. Hamann ('91, p. 28) refers to the condition of the subcuticular nuclei of Echinorhynchus clavaeceps, now Eorhynchus agilis, in the following terms:

Das Larvenstadium, in dem die Haut ein Syncytium mit wenigen Riesenkernen bildet, ist bei Echinorhynchus clavaeceps dauernd fixiert. Wie ich in dem systematischen Teil zeigen werde, ist diese Art auf dem Larvenstadium stehen geblieben, was ihre Haut, Muskulatur, Lemnisken anlangt, wir haben einen Fall von Phylo-Paedogenesis vor uns.

Eereits am lebenden Thiere sieht man die 0.2 mm. grossen, eiformigen bis kugeligen Riesenkerne in der Haut. Ich zahlte bei dem in Fig. 1 auf Taf. IX abgebildeten mannlichen Tiere acht Kerne in der Epidermis und je zwei in den Lemnisken.

It is interesting to observe that the foregoing statement as to the number of nuclei is not put in general form but only indicates that one individual possessed eight subcuticular nuclei. As the result of the examination of several hundred individuals of the genus Eorhynchus the writer has found but a single one in which the arrangement and number of nuclei varied from the typical condition of one ventral and five dorsal subcuticular nuclei. This, in itself, would seem to indicate the the specimen which Hamann was describing* was abnormal. Liihe ('04, p. 294), commenting upon the works of Mtiller and Hamann regarding the nuclei of the subcuticula, writes: "In der Figur Hamann's sind freilich 6 Kerne in der Haut gezeichnet. Aber wenn die Zahl dieser Kerne auch innerhalb gewisser Grenzen schwankt, so habe ich doch gerade die von 0. F. Miiller gezeichnete Fiinfzahl verhaltnismassig haufig beobachtet und alsdann auch stets in der von Mtiller gezeichneten Anordnung."

These records, while of importance historically, cannot be regarded as infallible, primarily because in each case they were made in the form of incidental observations rather than as accurate determinations. The following paragraphs present the results of my studies on subcuticular nuclei of five American species.

Eorhynchus gracilisentis. In this species the nuclei of the subcuticula conform to so distinct a pattern, as regards numbers and


CELL CONSTANCY IN THE GENUS EORHYNCHUS 263

arrangement, that variation beyond that which is capable of explanation on the basis of the degree of contraction is limited to a single instance. Each one of more than two hundred individuals had six nuclei in the subcuticula, arranged according to the following system: Five large nuclei occupy positions in the mid-dorsal line of the body. Three of these constitute a group toward the posterior end of the body while the remaining two are separated slightly farther from the rest and lie anterior to the group of three, but are included in the same sagittal plane as the posterior nuclei. In sexually mature individuals these nuclei averaged 0.078 mm. by 0.045 mm. The sixth nucleus of the subcuticula is located in the mid-ventral line, near the anterior end of the body, usually at a point between the anterior and posterior groups of the dorsal series. It is more elongated and larger than the other nuclei of this tissue, having a long axis of 0.125 mm. and a smaller diameter of 0.050 mm. These nuclei are irregularly ovoid in shape, though occasionally they are reniform as shown in figure 4, sn.5. The chromatin usually has a closely compacted arrangement, but sometimes takes the form of an irregularly branching network within the nucleus. The shape of the nuclei in this tissue of the Eorhynchi is in decided contrast to that described by Graybill ('02) for the nuclei of the subcuticula of Echinorhynchus thecatus Linton. According to that writer the nuclei of this last named species are of a wonderfully dendritic nature, each having a broad expanse of finely branching processes. The contrast is none the less striking if a comparison be made with the subcuticular nuclei of any other type of Acanthocephala. Three types of subcuticular nuclei are thus distinguishable in the Acanthocephala; the giant nuclei of the Eorhynchi, the dendritic nuclei of E. thecatus, and the numferous small nuclei scattered throughout the subcuticula as in most other Echinorhynchi and some of the other genera.

Only one specimen of all the Eorhynchi examined showed variation in the number and arrangement of these nuclei such as might be expected, judging from the statement of previous investigators. This individual was a small, immature female. Even here the variation was purely one of arrangement of the com


264 H. J. VAN CLEAVE

ponent nuclei, for numerical constancy remained undisturbed. In this case one nucleus normally in the anterior dorsal group had taken a position on the ventral side of the body. Before trying to explain this departure from the normal, attention must be called to the following points :

1. The nuclei in the dorsal line lie within the dorsal longitudinal canal, and the one nucleus on the ventral side of the body is similarly situated in the ventral canal. These two canals communicate directly through the circular canals of the lacunar system.

2. Hamann ('91, p. 21) ascribes to the nuclei of an Echinorhynchus in the embryonic state the power of amoeboid movement.

3. Since it is a well-known fact that some nuclei may wander to that part of the cell where they can best perform their functions, and may even penetrate heavy cell walls in order to reach a distant point of injury; may not the displacement of the nucleus in the abnormal specimen of Eorhynchus gracilisentis just cited be explained as the result of either active amoeboid movements which carried it along the lacunae to the opposite side of the body, or as a migration of the nucleus brought about by some local irritation or injury of the body on the ventral side?

In no other instance has the writer seen the parallel of Liihe's figure of Eo. rutili (copied as fig. 31) in which he portrays four dorsal and two ventral subcuticular nuclei (Liihe '11, p. 12, fig. 1).

Eorhynchus longirostris. The nuclei of the subcuticula of this species differ but little from those of the preceding species. In the first place the median dorsal row of five nuclei is not so clearly divided into an anterior group of two and a posterior group of three nuclei. The nuclei of this region are almost equidistant one from another. Moreover the terminal member of this series approaches nearer to the posterior tip of the body than does the terminal nucleus in Eo. gracilisentis. The swellings upon the surface of the body caused by the presence of these nuclei are more distinct in toto mounts of Eo. longirostis than in mounts of the preceding species. The dorsal nuclei have as an average, a long


CELL CONSTANCY IN THE GENUS EORHYNCHUS 265

diameter of 0.132 mm. and a short axis of 0.048 mm. while the ventral nucleus ranges in size near 0.250 X 0.057 mm. Instead of occupying a position posterior to the second dorsal nucleus, as in Eo. gracihsentis, the ventral nucleus of this species is typically opposite the foremost of the dorsal series. A slight variation does occur with regard to the relative location of the ventral nucleus but this probably is due to difference in degree of contraction of various individuals.

Eorhynchus emydis. Though the individuals of this species are very much larger than those of either of the preceding species, being from 3 to 32 mm. long, yet the number of nuclei in the subcuticula is the same. Using the arrangement of the nuclei of the subcuticula as a basis for argument it appears that the excessive size of the species is brought about by an elongation of the posterior region of the body rather than by a uniformly distributed elongation. The evidence for this lies in the fact that, while the two anterior dorsal nuclei remain relatively close together, the three of the remaining group are separated from each other by a marked interval. The ventral nucleus occupies a position opposite that of the foremost of the dorsal row, or, occasionally, a little anterior to it. No variations of number or of arrangement were observed in the numerous individuals of this species which were studied.

Eorhynchus cylindratus. Here, again, is found the typical number of subcuticular nuclei. The distribution of the dorsal nuclei closely approaches that found in Eorhynchus emydis (Leidy), but the ventral nucleus more often takes a position midway between the two anterior dorsals, approaching nearer to the posterior of these than to the anterior as in Eo. emydis. The distribution of the dorsals strongly suggests that found in the preceding species.

Eorhynchus tenellus. In this species, likewise, the five dorsal nuclei and one ventral nucleus in the subcuticula have an arrangement similar to that of Eo. emydis.


266 H. J. VAN CLEAVE

S. Lemnisci

These unique structures, while extremely simple in organization, have been the subject of considerable speculation among writers dealing with the genus Eorhynchus. Saefftigen ('84, p. 9) recorded two nuclei in each lemniscus of Eo. rutili (Mtill.). Hamann ('91, p. 140) corroborated this for the same species and later ('95, p. 30), described the same condition for the only other then known species of the genus, Eo. agihs (Rud.) . He mentioned an instance of an immature form of Eo. rutili in which only one nucleus was present in each lemniscus, while the second nucleus was just in the process of formation. He says, Ein junger Echinorhynchus clavaeceps, der mir vorliegt, zeigt gering entwickelte Lemnisken, die aber bereits den einen Kern enthalten, wahrend der andere noch an der Grenze ihrer Entstehung liegt." This condition finds ready explanation in the light of facts brought out later in this article. Still more recently Liihe ('11, p. 12) recorded two nuclei in each lemniscus of Eo. rutili. His figure 1, copied here as figure 31, shows but a single nucleus in each lemniscus.

It seems hardly necessary to mention that these observations as given by various writers on the Acanthocephala add nothing to the direct evidence for or against the problem of cell constancy. However, they do serve to indicate the ease with which one may be led to accept the assertion of an earlier worker without fully following out the evidence upon which his statements are founded. To one working on a pm'ely anatomical problem it is extremely plausible to decide that a form is immature when only a single nucleus appears where others say two should occur. Such seems to have been the condition when Hamann reported this abnormahty, but, in the light of the present study of the structure of the lemnisci of the numerous examples of Eorhynchi, it is difficult to see the source of this evidence of a second nucleus just at the beginning of its formation. All the facts gained from the study of five species of this genus point to a condition mentioned by none of these previous investigators, namely, while one lemniscus possesses the two nuclei so generally ascribed to each, the other


CELL CONSTANCY IN THE GENUS EORHYNCHUS 267

in no case has more than a smgle one. Figures 22 and 43 indicate clearly this relation of the nuclei in the lemnisci.

An obvious explanation of the cause of so many investigators going astray on the question of the number of nuclei in the lemnisci is found in the generally accepted conclusion that these are paired lateral organs. Therefore as soon as two nuclei were demonstrated in one of them it was the natural thing to expect a duplication of this condition in the organ of the other side of the body. Failure to recognize both nuclei in each of the two lemnisci could be accounted for by the interference of overlying parts. Or even in case serial sections were examined in this connection the lemnisci are usually so much twisted and coiled about one another that, unless reconstruction drawings are made, one is not sure of the interpretation of the numbers and relationships of the nuclei.

.Eorhynchus gracihsentis. The small size of the lemnisci in this species, and the fact that their usual condition is a mass of twists and turns in the anterior end of the body, together serve to render an accurate analysis of the exact structure very perplexing. The careful study of serial sections of numerous individuals fully demonstrated the inequality of the two organs. Invariably one lemniscus was found to be supplied with two large nuclei lying in the axis of the canal, while the other contained but a single nucleus in its canal. These nuclei so closely resemble those of the subcuticula that no special description seems necessary for them. In shape they are more elongated and display more of a tendency toward a pointed condition at either end.

Eorhynchus longirostris. In this species the lemnisci demonstrate the point in question most admirably (fig. 22). Here reconstructions from serial sections were needless, since the nature of the lemnisci was such that each stood out as a distinct, practically straight tube extending backward into the body cavity. An examination of toto mounts was sufficient to demonstrate that one of these organs possessed two nuclei of the subcuticular type while the other had but a single nucleus.

Eorhynchus emydis. It was from the study of the whole mounts of immature forms of this species that the writer first


268 H. J. VAN CLEAVE

gained the correct conception of structure of these organs, which has been carried in a comparative way to the other members of the genus. In these sexually immature stages the lemnisci appear as a pair of small sac-like projections into the anterior end of the body cavity, and in these sacs the nuclei are readily distinguished. It might be argued that as sexual maturity is reached changes occur in the structure of the body, resulting in either an increase in the number of the nuclei or a redistribution of them. Hence the study of these juvenile forms was supplemented by careful observations on the condition in fully mature individuals. These entirely confirm the results obtained for the much smaller, sexually immature members of the species. Figure 43 indicates the absolute clearness with which the nuclei appear in these organs.

Eorhynchus cylindratus from the black bass, Micropterus salmoides (Lacepede) has much longer lemnisci than the preceding species but the arrangement and number of its nuclei are identical with the conditions described for each of the preceding species of Eorhynchus.

Eorhynchus tenellus. Here again the same condition is met in the nuclei of these organs.

Hamann ('95, p. 9) said: Die Lemnisken sind als Fortsetzung der Haut der Korperwand anzusehen." In all the foregoing species the nuclei of the lemnisci so closely resemble those of the subcuticula in structure that this fact might well be used as an argument in support of the view that these organs arise as inpocketings of the body-wall. Moreover, the fact that their position in the canal of each lemniscus corresponds to the condition found in the subcuticula where the nuclei also occur in the longitudinal canals of the body, indicates a possible connection or relationship between these two sets of canals. No direct connection, however, has been made out by the writer in members of this genus, although such a connection is rather easily demonstrated in members of other genera of Acanthocephala.

One rather common variation in the form of the nuclei of this region of the body consists in the tendency to assume a rounded outline in immature forms as contrasted with the more elon


CELL CONSTANCY IN THE GENUS EORHYNCHUS 269

gated and less regular outline of each nucleus in the fully mature animal. At the same time this variation in shape is accompanied by just as striking a difference in the manner in which the chromatin material is distributed. While in the adult the chromatic substance is arranged in a more or less irregular branching figure in the midst of the nucleus, in the immature condition the chromatin is more evenly distributed throughout the nuclear mass.

Ij.. Proboscis and proboscis receptacle

While ordinarily the proboscis and its receptacle are considered as two separate organs or regions in the Acanthocephala, in the following it seems better to consider the two together, since certain structures are not limited to either but extend through both. Even though the proboscis shows practically no individual variation within the species, yet the different species may vary to a remarkable degree. On the other hand, several species of this genus present such slight variation in this organ of attachment that the only evident point of contrast is in the relative size of the various rows of hooks. In spite of the considerable range of shapes and structures in different species, the following analysis will show that in all probabiHty the same nuclei contributed to the formation of structures which are distinctly different in one group from those found in closely allied species of the same genus. In other words, it seems that during the processes of differentiation variations may have occmTed in the resulting structures but that the cellular elements entering into the differentation were the same in number in each species.

At the tip of the proboscis (figs. 6 and 38 po.) is a structure of questionable function and of variable form within the genus. This structure, which occupies the central region of the proboscis, is, in some species, an elongated cylindrical sac while in others it more nearly approaches a globular form. Various functions have been ascribed to this peculiar mass of tissue. Lespes ('64) considered a similar structm-e in Echinorhynchus clavaeceps as the indication of an alimentary canal. Of a similarly located structure made up of two cells in E. proteus Baltzer ('80, p. 26)

JOURNAL OF MORPHOLOGT, VOL. 25, NO. 2


270 H. J. VAN CLEAVE

made the statement " Wahrscheinlich stehen diese Zellen mit einem, hier vorhandenen Tastvermogen in Beziehung." Saefftigen ('84) thought this condition in E. clavaeceps resulted from the four retractors of the proboscis forming a space between them and in this the nuclei had come to he.

Hamann ('91, p. 51) has shown that in the ontogeny of E. proteus a group of three cells is laid down at the anterior end of the proboscis early in development. *'An der Spitze sind auf dem Schnitt drei der grossen Kerne mit ihrem kugeligen Kernkorperchen getroffen, deren Zellsubstanz den centralen Teil der Riisselanlage bildet," He has found such cells in all species of Echinorhynchus which he has investigated. As to their location and relations he adds ('91, p. 57) : '^Bei Echinorhynchus proteus und angustatus liegen sie nebeneinander, bei Ech. clavaeceps hintereinander .... Seine Abbildungen lassen das Verhaltnis dieser Zellen zur Haut (Subcuticula) nicht deutlich erkennen." Hamann considers the evidence insufficient to warrant the naming of this structure an alimentary tract. On the other hand he has offered the suggestion that the staining reaction of this tissue in E. clavaeceps is such that the granular structure with hematoxylin staining indicates a glandular function. For this reason he called the cells gland-cells. I do not consider the evidence sufficient to warrant ascribing a definite glandular function to this structure, but since there does not seem to be any close relationship between it and the other organs, might it not be the vestige of some organ which through disuse has degenerated since the Acanthocephala have acquired the parasitic habit?

Eorhynchus gracihsentis. This terminal organ of the proboscis in Eo. gracihsentis (fig. 6). contains three large oval nuclei lying in its long axis. It will be recalled that this arrangement of the nuclei is what Hamann ('91, p. 57) has described for Eo. rutili. The relative position of these is apparently influenced by the state of contraction of the proboscis muscles. In case the proboscis is fully extended these nuclei assume a regular oval form, the long axis of each coinciding with the long axis of the body, but in various stages of contraction of the proboscis these nuclei become slightly modified in their relations. When forced closer together, they, at times, come to lie with their long axes at right angles to the main


CELL CONSTANCY IN THE GENUS EORHYNCHUS 271

axis of the body. However it does not seem necessary to do more than mention that this abnormal or unnatural position is simply the result of the action of an external physical force and for this reason would not come under the consideration of abnormalities of location such as were discussed under the nuclei of the subcuticula.

A dense chromatin mass lies in the center of each of these nuclei. This mass at times is distinctly oval, again it may be irregularly branched, and finally it frequently appears as several smaller rounded masses scattered throughout the nucleus. In this same species a pair of small nuclei occurs in a clearly demarked area at the base of the terminal organ of the proboscis. These are usually almost perfectly spherical, but at times they assume a slightly ovoid shape (fig. 7).

Eorhynchus longirostris. In this species (figs. 22 and 26, po) the nuclei of the terminal proboscis organ have the same number and arrangement as just described for Eo. gracihsentis.

Eorhynchus emydis (Leidy). This form possesses a terminal proboscis organ of the type described by Hamann for E. proteus and E. angustatus. It is supplied with two large oval nuclei in the central enlarged portion and with two small spherical nuclei in a region at their posterior margin. The general relations of these parts are well shown in figure 45, a copy of Hamann's figure 28, plate 3, representing the condition found in Eo. agihs (Rud.).

Eorhynchus cylindratus. The terminal organ of the proboscis has the same structure and arrangement of nuclei as emydis. The two large oval nuclei lie side by side, while the two smaller ones occur posterior to them.

Eorhynchus tenellus. Here the organ at the end of the proboscis (fig. 38, Tpo) is of the type just described for the last two species, and the constancy of the cellular elements is identical.

5. Proboscis wall and receptacle ivall

Eorhynchus gracihsentis. The nuclei of the walls of the proboscis and its receptacle in this species have the following arrangement: The anterior end of the proboscis bears a circle of twelve nuclei, located in the region near the bases of the terminal hooks,


272 H. J. VAN CLEAVE

yet no direct connection could be traced between nuclei and hooks. Figure 10 shows the relations of these nuclei in a cross section of the proboscis. The nuclei of this group are small spheres of chromatin 0.005 mm. in diameter, each lying in the center of a 0.007 mm. clear space. Two pairs of nuclei are connected with the muscle sheath which effects the retraction of the hooks. These lie, one on the dorsal side of the proboscis and one on the ventral side, in the region between the second and basal rows of hooks. The location of these nuclei is shown in figure 6, which is a single sagittal section including but one nucleus of each of the two pairs.

The muscular wall of the proboscis receptacle proper has three pairs of nuclei. The most conspicuous of these is a pair of oval nuclei connected with the sheath musculature directly at the base of the brain (fig. 8). The other two pairs are located, one pair dorsally and the other ventrally, on the wall of the receptacle just anterior to the brain.

Eorhynchus longirostris. Due to the crowded condition of the receptacle in this species, brought about by the usual partial inversion of the proboscis, very little could be made out regarding the cellular structure of the proboscis receptacle. The two nuclei associated with the muscular wall at the base of the brain are very distinct.

Eorhjaichus emydis. The muscular wall of the proboscis rereceptacle is suppUed with three pairs of oval nuclei. One pair is located on the inner wall of the receptacle at the base of the brain. The other two pairs are located on the inner side of the dorsal and ventral walls, a short distance cephalad from the brain. In the proboscis two pairs occur, one on the dorsal side of the proboscis wall and the other pair on the ventral, in the region of the bases of the middle row of hooks.

Eorhynchus cyhndratus presents a nuclear constitution of the proboscis and its receptacle identical with that of Eorhynchus emydis.

Eorhynchus tenellus has the same arrangement of the muscular nuclei in the proboscis sheath and proboscis as the two preceding


CELL CONSTANCY IN THE GENUS EORHYNCHUS 273

species. This arrangement is shown in figure "38, from which one nucleus of each pair of dorsal and ventral receptacle wall nuclei has been omitted in the reconstruction.

6. Invertors of the proboscis

Eorhynchus gracilisentis. The proboscis invertors consist of two strips of tissue which extend backward along the dorsal and ventral surfaces of the terminal proboscis organ and pass through the base of the proboscis sheath. In the region anterior to the brain each invertor bears a pair of elongated or spindle shaped nuclei. In this species the invertors are so highly developed that they almost entirely fill the space within the walls of the proboscis and its sheath.

Eorhynchus longirostris. In this species the same structures are found. The invertors are somewhat reduced in size so that they stand out more clearly as separate structures. As in the preceding case each of them is supplied with two nuclei. Figure 26 shows these invertors cut in a slightly diagonal plane so that but a single nucleus shows in each.

Eorhynchus emydis. Figure 44 is a drawing from a toto mount showing the pair of nuclei associated with each invertor.

Eorhynchus cylindratus. The nuclei in the invertors are identical with those found in Eo. emydis. The invertors are in about the same stage of development.

Eorhynchus tenellus presents the same distribution of one pair of nuclei to each invertor, as shown in figure 38.

7. Retractors of the proboscis receptacle

In the genus Eorhynchus the retraction of the proboscis receptacle is accomplished through the contraction of two long fibrous bands which proceed from the posterior end of the proboscis receptacle and find attachment, one on the dorsal wall of the body and the other on the ventral body-wall. These retractors are direct continuations of the proboscis invertors.

Eorhynchus gracilisentis. The dorsal retractor of the proboscis receptacle in this species is supplied with two oval or spindle


274 H. J. VAN CLEAVE

shaped nuclei whose arrangement is shown in figure 16. The ventral retractor of this same species is similarly furnished with two nuclei. These are of the same type as those already described for the dorsal retractor. In no case was any evidence discovered indicating the presence of other nuclei of any type in these structures.

Eorhynchus longirostris. In members of this species the arrangement of the proboscis receptacle retractors is identical with that just set forth for Eo. gracilisentis. The nuclei are of the same number, the same appearance, and have identical relations within the retractors. Figure 28 shows the nuclei of the dorsal retractor of the proboscis receptacle.

Eorhynchus emydis (Leidy), Eorhynchus cylindratus, and Eorhynchus tenellus show the identical condition of the retractors of the receptacle as described for the other species of the genus.

8. Cement gland

In members of the genus Eorhynchus the cement gland presents a condition at wide variance from the type most commonly found in the other genera of Acanthocephala. Instead of consisting of a series of independent sac-like glands in the posterior region of the male, as in many genera, the cement gland is a compact mass. Bieler ('13) has suggested that this deviation in structure, together with the fact that a single covering encases the mass, is of value in differentiating the Eorhynchi from their near relatives. This characteristic was not mentioned by Hamann in his definition of the genus. In all of the species examined the nuclei of the cement gland are oval with a compact central chromatin mass.

Eorhynchus gracihsentis. In this species eight large oval nuclei lie imbedded in the mass of the gland (fig. 9, eg.) These are the only nuclei present in the entire structure. The study of toto mounts and of sagittal sections demonstrated that, though these nuclei are uniformly distributed through the gland, there is an evident arrangement in two lateral groups each of four nuclei.


CELL CONSTANCY IN THE GENUS EOEHYNCHUS 275

Eorhynchus longirostris. The gland in this species stands in marked contrast to the type preceding. While the nuclei are clearly of the same character as those of Eo. gracilisentis their number is just doubled. Instead of the eight large oval nuclei there are sixteen of the same general character (fig. 20, eg) .

Eorhynchus emydis. This species contains a long cylindrical cement gland bearing eight nuclei. The size of the gland is indicative of the degree of maturity of the individual. Figure 42 shows this gland in an immature male. The eight nuclei within thts small mass are arranged in two distinct groups. The same gland in a fully matured worm not only has increased considerably in size (fig. 41) but at the same time has undergone a complete rearrangement of the component nuclei, resulting in a single line of nuclei running the length of the gland.

Eorhynchus cylindratus and Eorhynchus tenellus both have cement glands of the type just described. In each case the eight large nuclei occur in an elongated cylindrical organ. The arrangement of the nuclei in each of these species is the same as that described for Eo. emydis.

9. The male genital apparatus

Eorhynchus gracilisentis. The cirrus in this form is supplied with nuclei of two types, and in some instances a clear distinction could be drawn between the two kinds of tissues connected with the respective types of nuclei. The portion of the cirrus lying nearest to the sperm duct has a pair of large lateral rmclei in the dorsal region (fig. 18, ci) in which the chromatin has a strikingly characteristic arrangement in small masses scattered through the nucleus, with a tendency toward collecting on the nuclear wall. These spheroidal nuclei are 0.007 mm. in diameter. Slightly anterior to this pair of nuclei and in the region of the cirrus farthest removed from the duct are two slightly larger spherical nuclei (fig. 18, ci) measuring about 0.009 mm. In these, the chromatin tends toward the formation of closely compacted masses lying in the center of the nuclei.


276 H. J. VAN CLEAVE

The walls forming the vas deferens have two nuclei just ventral to the cement gland. These nuclei have a diameter of 0.007 mm., and each contains a single chromatin mass 0.003 mm. in diameter. A slight ventral dilation of the vas deferens in that part of its course just opposite the posterior edge of the cement gland where it enters the sperm reservoir, is supplied with two small nuclei. These are bilateral in arrangement. Each has a long axis of 0.005 mm., and a short axis of 0.003 mm.

The muscular sac posterior to the reservoir of the cement gland contains two star-shaped nuclei which lie close to the anterior end of this structure. These have the appearance of lying in a very loose mass of tissue, sending out small projections to the more solid surrounding portions of the sac wall. The muscular wall of this sac contains a small pair of nuclei on its dorsal surface near the posterior end. These are of the type of small oval nuclei found in the body musculature of this genus. Figm*e 17, ms, shows the relations of this structure as presented in a reconstruction from sagittal sections. Very similar conditions have been found in the other species of Eorhynchus but the study has not been carried out in detail. Figure 25 indicates the appearance of the pair of nuclei within the muscular sac of Eo. longirostris.

The inturned portion of the posterior region of the male, extending from the exterior to the cirrus, is curiously modified in the region nearest the genital opening .when the copulatory apparatus is retracted. At this place there are formed a series of five closely fitted wedge-shaped structures, each of which bears three nuclei, as shown in figure 18, which is a reconstruction drawing from serial sagittal sections. The same condition has been demonstrated in well cleared toto mounts of this species. Connecting this structure surrounding the genital orifice with the arms of the copulatory bursa there is an extremely irregular, inconstant portion , the walls of which vary so in appearance with the various states of retraction or of protrusion of the genital apparatus that the writer has not been able to estabhsh a definite number or arrangement of the nuclei contained in it (fig. 18).


CELL CONSTANCY IN THE GENUS EORHYNCHUS 277

10. The female genital apparatus

Eorhynchus gracilisentis. The vagina presents a clear example of constancy of number and arrangement of the nuclei. Four large nuclei (figs. 11, 12 and 13, v) are found in its inner muscular layer between the genital orifice and the sphincter vaginae. Two nuclei are associated with the sphincter. The outer non-muscular layer of the vagina in the region of the sphincter contains three nuclei. Cephalad from the sphincter the vagina has a short rounded region consisting of a thick walled tube through which the embryos must pass from the uterus into the orifice of the sphincter. This heavy walled portion of the vagina bears four large nuclei. Anteriorly, the outer wall of the vagina passes over into the single-walled uterus with its large sac-like cavity. On the dorsal side of the uterus at its junction with the vagina there is an enlargement of the wall in which are found two nuclei, the only ones associated with the uterus. Figures 11, 12 and 13, u, V, and vs, show the relations of these parts, together with a representation of the nuclei. The selective apparatus (fig. 11, so), because of its extremely complicated structure, gives considerable trouble when an attempt is made to compare the nuclei of one individual with those of another. The eight nuclei shown in the drawing just cited have been found in every individual studied, but several other nuclei which occurred in this region in some individuals could not be found in others of the same species. The hard shelled embryos with which the body of a mature female is filled hide this structure almost completely in toto mounts, and in serial sections the embryos frequently cause it to be so torn that very little of the actual relations may be made out! Extending posteriad from the dorsal wall of this selective apparatus are two muscular strips whose function is evidently to support these organs since each finds a point of attachment on the dorsal wall of the body. One nucleus occurs in each of these (fig. 11, ssa). On either side of the genital orifice of the female there extends a broad fan shaped ligament which becomes attached to the dorsal wall of the body cavity (fig. 15, fl). In the broad portion


278 H. J. VAN CLEAVE

near the genital orifice each Hgament contains a single large star-shaped nucleus.

Eorhynchus longirostris. In this form the vagina, from the genital orifice to the sphincter, is supplied with four nuclei, as shown in figure 27. The two nuclei associated with the sphincter, which can be made out clearly in serial sections, are not shown in this drawing. Anterior to the sphincter are four vaginal nuclei. The uterus, as in all other species studied, has but two nuclei and these occur at its posterior extremity. As & consequence of the difficulties involved in studying the selective apparatus, as pointed out above, the writer has been unable to estabhsh any definite results regarding the occurence of nuclei in that organ.

Eorhynchus emydis. In immature individuals of this species the female tract presents a very diagrammatic arrangement of the component structures. Figure 40 shows the nuclei with which the vagina below the sphincter is supplied. These four nuclei are arranged in pairs, one at either end of the region. The sphincter muscle has a single pair of nuclei. Anterior to the sphincter the thick walled portion of the vagina has four nuclei arranged in two pairs as shown for Eo. gracilisentis. The uterus has a single pair of nuclei at its posterior extremity. In the selective apparatus six pairs of nuclei could be seen distinctly. Of the two muscles (fig. 40, ssa) for the support of this organ each bears a single nucleus.

Eorhynchus tenellus. Reference to the" figure will show that the female genital tract of this species (fig. 39) bears identically the same nuclei as have been described for Eo. emydis.

11. The brain

Eorhynchus gracilisentis. In the brain of Eorhynchus conditions are such as to render a careful analysis of the structure difficult. This organ (fig. 8, hr.) is a conical mass of ganglion cells having an extreme length of 0.1 mm. The cells of which this mass is composed are small rounded bodies of uniform character so that there are no regions demarcated by variations in the size or structure of the component cells. The small, 0.005 mm., cells


CELL CONSTANCY IN THE GENUS EORHYNCHUS 279

are so closely packed together that even in a most careful reconstruction from thin serial sections one is never absolutely certain that in some cases two adjoining sections do not contain parts of the same nucleus. Reconstructions of two individuals from drawings of serial sections magnified nine hundred diameters gave in one instance one hundred and eight cells in the brain, in the other one hundred and nine. While these facts are not given with the idea of establishing an absolute constancy they may be taken to indicate that, evien though the structure in this region is so extremely crowded, yet there is a close agreement in the results obtained from these two reconstructions.

Eorhynchus emydis. A reconstruction of the brain of a single individual of this species gave ninety-four nuclei as the total for that structure.

12. The genital ganglion

Eorhynchus gracilisentis. Eighteen cells are grouped together in the region of the cirrus in the male to form what is commonly spoken of as the genital ganglion (fig. 17 gg). As in the brain, these cells are all of the same type, giving no basis for a division of the ganglion into regions according to the structure of the component elements.

13. Body musculature

Eorhynchus graciUsentis. Two reconstructions of the body musculatiu-e each gave a total of thirty nuclei associated with that tissue. These were arranged in fifteen pairs, extending from the union of the body with the proboscis, caudad to the posterior extremity, one nucleus of each pair having a position slightly lateral to the dorsal canal. The distribution is such that a single cross section could not contain more than a single pair of nuclei. The difficulties involved in the determination of the number of nuclei present in this tissue consist chiefly in the discrimination between these nuclei and the embryos, which, in the female, are often so closely packed against the wall that only the most careful study will reveal whether a given structure is associated with the muscle layer or merely closely applied to it. The position of these nuclei in longitudinal section is shown in figure 5.


280


H. J. VAN CLEAVE


SUMMARY OF CELL CONSTANCY IN THE GENUS EORHYNCHUS, WITH TABLE OF RESULTS

Consideration of the facts presented in the foregoing parts of this study has led the writer to a firm behef in a remarkable degree of constancy in all the somatic structures in the genus Eorhynchus. Failure to demonstrate this beyond a doubt for all the organs is due to physical limitations, primarily, rather than to conflicting data, for in every instance where a positive count has been made under conditions precluding the possibility of error in manipulation, the same number and arrangement of the nuclei has been found in every individual of the same species. Table 1 summarizes the results found for the five species studied.


TABLE 1


OBQAN OR TISSUE


EO.GBACILI 8ENTIS


EO. LONQIROSTRIS


EO.EMTDIS


EO. CYLINDBATU8


EO.

TENELLUa


Subcuticula

Terminal proboscis ring of nuclei

Terminal proboscis organ

Leminisci, 1 and 2

Retractors of hooks

Wall of proboscis-receptacle

Proboscis invertors

Proboscis sheath retractors:

Dorsal

Ventral

Cement gland

Cirrus

Vas deferens.

Muscle sac:

Inside

Wall

Inverted part of male (posterior)

Vagina

Uterus

Vaginal sphincter

Selective apparatus

Stays of selective apparatus. . . . Ligaments at orifice of female . . .

Brain

Genital ganglion

Body musculature


6 12 5 3 4 6 4

2 2 8 4 2

2

2 15 11

2

2

8+

2

2 108 (?) 18 30


5 3

2+ 4

2

2

16


2

2

12

2

94(?)


8 2 2 12 2


€ELL CONSTANCY IN THE GENUS EORHYNCHUS 281 GENERAL PROBLEMS RELATING TO CELL CONSTANCY

1. Cell size vs. body size

A problem such as the relationship existing between body size and cell size is often approached only from a teleological point of view. Given as the end result the adult body form, that becomes the goal toward which the development of the organism is directed. Thus the possibility that this final body form and size may be a mere incidental result attained through precisely determined series of processes of development is entirely ignored. In other words, a common conception of body form and size might be crudely expressed in the idea that nature establishes a mold for each particular type of body form and that, given a zygote of a definite size, the development of the individual must consist in attaining the boundaries of size and form set for that species, regardless of the number of cell divisions that may be required for that purpose. According to this interpretation, the size of the body is the constant factor while the number of cells which go to make up the organism is purely a matter of circumstance.

The problem of the origin of the adult body and the relationship between body size and size of the component cells is one of long standing. WTiile the following discussion does not attempt to bring all the literature upon the subject into relationship, it will serve to give some idea of the nature of the contentions on the various sides of the problem. Conklin is one of the staunch supporters of the view that not only is cell size fixed but also that it bears no direct relation to body size. In Crepidula fornicata, in which large and small individuals occur, he has shown ('98) that the difference is due entirely to the larger number of cells in the larger individuals:

In Crepidula therefore the cell size is constant, and variations in the size of the body are due to variations in the number of cells present. The dwarfs are what they are by reason of external conditions and not because of inheritance; they are in short a physiological and not a morphological variety. In such a case the shape and size of the body as well as the number of cells in the entire organism, are greatly modified- by the direct action of the environment.


282 H. J. VAN CLEAVE

A general presentation of the recent results upon the relations between body size and cell size is given by Minot ('98, p. 65) :

Cells do not differ greatly from one another in size. The range of their dimensions is very limited. This is particularly true of the cells of any given animal. Recent careful investigations have been made upon the relation of the size of cells to the size of the animals, and it has been found that animals are not larger, one than another, because

their cells are larger, but because they have more of them

For example, a large frog differs from a small frog or a large dog from a small dog by the number of cells.

To the writer the foregoing statement has the appearance of too strong a generalization from a limited group of facts. In reality the cells of some animals do differ greatly from one another. Thus in Eorhynchus emydis, while the adult body form is perfectly presented in a small individual 1.7 mm. long, the maximum length of this species is over 30 mm., and both of these individuals have identically the same number of somatic cells. Even a most conservative calculation of the relative size of the subcuticular cells of the small and the large individuals shows that their ratio based on volume is 1 : 140. Since the subcuticula is a syncytium, this calculation is based on the ratio of the volumes of the subcuticula in the two individuals. The volume is determined by taking the product of the length, by the circumference, by the thickness of the subcuticula. Since the number- of nuclei present in the small and the large individuals is the same, the ratio of the volumes of the entire subcuticula of the two individuals is the same as the ratio of the individual cells composing it. It would seem that the range of their dimensions in this case is not very limited.

Closely associated with this problem of cell size and body size, and growing directly out of it, is the question of the relation between the number of cells present in the organs of the different individuals. In the analysis of any given organ four possibilities of this relation exist: (1) Body size and cell size are both fixed. Of necessity in this instance the number of cells is just as sharply fixed. No case of this sort is known to the writer; (2) Size of the cells is fixed but the body size may vary. Conk


CELL CONSTANCY IN THE GENUS EORHYNCHUS 283

lin's results on Crepidula exemplify this condition. The number of cells in any given organ may consequently vary within undetermined limits; (3) Body size is fixed while the size of the component cells may vary. There could be no precise limit to the number of cells in this instance; (4) Body size and cell size may both vary. This last possibility is still further divisible into two conditions : (a) if cell size and body size vary independently no necessary connection exists between the size of the body and the number of component cells, .while (b) if body size varies directly as the size of the cell the number of cells in any given organ becomes a fixed quantity, as in the case outlined under (1). It is in accordance with the last possibility that the explanation of the condition found in the genus Eorhynchus is to be sought. In general it is in organisms having this fixed correlation between body size and cell size that cell constancy is to be found.

Regarding the cell lineage of Nereis, Wilson ('92, p. 377) writes: "The entire ontogeny gives the impression of a strictly ordered and predetermined series of events, in which every cell division plays a definite role and has a fixed relation to all that precedes and fdllows it." How readily this conception of ontogeny lends a support to the theory of cell constancy. In another connection, Wilson, ('00, p. 390) has stated that the number of cells produced for the foundation of certain structures is often fixed. For example, in annelids and gasteropods the entire ectoblast arises from twelve micromeres segmented off in three successive quartets of micromeres from the blastomeres of the four-cell stage."

By way of explanation of this condition as indicated by Wilson and by Morgan, two possibilities present themselves, either (1) a constancy in numbers of cellular elements is the primitive condition of development which, as Conklin points out, is retained only in the early stages of ontogeny of the metazoa, or (2) cell constancy is a manifestation of a tendency toward fixity which is acquired only at a late stage in the development of a race.

That there is a tendency for the same number of embryonic cells to be used in the formation of the organs and structures of


284 H. J. VAN CLEAVE

the body is evidenced by the results obtained by Morgan in his experiments upon the embryos developed from isolated blastomeres of echinoderms and of chordates. In 1896 he found that the larvae of Amphioxus developed from isolated blastomeres of the two and four cell stages tend to utilize the same number of cells for the formation of the organs of the body as do the normal larvae. In the formation of the gastrula stage the normal larvae invaginate about one-tenth of the total number of cells present in the blastula. Thus in a one-half or a one-quarter embryo, while the number of cells in the blastula is not as great as the number in the normal blastula, the actual number of cells that is invaginated tends to equal the number invaginated in the whole embryo. In other words, it seems as though a definite number of cells were necessary for the formation of a given organ of the body. I have been personally most loath," he adds, 'Ho accept this conclusion, because it seemed a priori very improbable that a numerical question could enter into this problem, but I see no other alternative than to accept this view of the matter." It would seem that if this be the case in organ formation, the requirement is one rather of the number of cells entering into th5 composition of the organ, than a mere quantative regulation or partition of the amount of protoplasm.

The controversy over the relation between the cell size and body size may be taken too seriously. The writer has pointed out the various possible relations between body size and cell size. The problem does not involve a fundamental conception under which all data must be subsumed and made to agree. Different groups of animals may vary widely in respect to the possibility of a direct correlation between the two, the only problem of real concern is to what extent any given condition is fixed within a given group of organisms. In fact Morguhs ('11) has shown experimentally that the relation of cell size to body size may be altered in the individual through a change of physiological conditions. In connection with his observations on complete inanition he has said ('11, p. 259) : "In the case of Diemyctylus it was found that the volume of both the cells and nuclei of different tissues diminishes


CELL CONSTANCY IN THE GENUS EORHYNCHUS 285

as the animal becomes smaller, and that with a retm-n to normal diet, when the animal grows very rapidly, its cells again increase in volume."

2. The heginning of constancy

Throughout the entire animal kingdom observations, chiefly incidental, have been recorded indicating that an organism or a single organ contains a fixed number of cells or of nuclei. Originally this condition was looked upon as of uncommon occurrence. Various workers among the protozoa have recorded the tendency for a single individual during sporulation or colony formation to give rise to a fixed number of cells. Thus in Tillina four spores are usually formed from a single individual, but this condition has not been absolutely fixed, for occasionally but two spores are produced. This would seem to indicate a transition from a reproduction by simple binary fission to a higher type where a larger number of individuals is produced from the parent cell. A step higher in the scale is Colpidium which, though normally producing four individuals in its temporary cyst, not infrequently forms eight. Among the colonial protozoa a single colony is frequently composed of a definite, constant number of cells: as an example of this Gonium pectorale might be cited. This is a sixteen cell colony in which each cell of the association in turn gives rise to a sixteen cell aggregate while still within the parent colony. In Eudorina is found an example where thirty-two undifferentiated cells arise from a parent cell producing a colony, as also in Pandorina which contains sixteen undifferentiated cells associated together. Thus in the protozoa there is a transition from the condition where reproduction is accomplished solely by simple binary fission, through a series of stages to that in which a definite number of cleavages result in the formation of a colony having a fixed number of individuals.

In the Metazoa, Conklin ('98) has distinguished a difference between the earlier and the later cleavages, attributing to the former the greater morphological importance:

JOCRXAL OF MORPHOLOGY, \ OI,. 25, NO. 2


286 H. J. VAN CLEAVE

The difference in the number of cells offers no difficulty in the doctrine of cell homology unless we assume that all divisions are differential, a thing which we know is not true. After blocking out the protoblasts of various regions and organs an indefinite number of non-differential divisions may occur either before or after the complete differentiation of the parts, and this probably explains the larger number of cells in the embryo of Crepidula adunca and the smaller number in the adult. In fact after the complete differentiation of all the tissues and organs, the number of cells may vary greatly in the different individuals of the same species or in the same individual at different times. In adult Crepidulas the number of cells varies directly as the body size varies, the cell size remaining practically constant. These later divisions, in the main, are non-differential, and likewise it is probable that in the later stages of cleavage many non-differential and inconstant divisions occur. Not only is there greater variation in the number and size of cells in later as compared with earlier stages of cleavage, but there is also greater variation in the direction and time of division; all of which goes to prove that the earlier cleavages are more constant, more frequently differential, and therefore morphologically more important.

In citing these instances wherein among the Protozoa and in the embryology of the Metazoa constancy seems to have a beginning I do not wish to give the impression that from such isolated and fragmentary citations I would claim to have traced a probable actual rise of fixity in numbers of cells going to make up an adult individual. In fact I do not consider cell constancy as of simple origin, for, while it occurs in broadly separated groups of the animal kingdom, it has probably arisen independently in each from the primitive, more variable conditions through the processes which tend toward the elimination of variability during the phylogeny of the group.

3. Factors involved in the production of cell constancy

For the explanation of the causes and limitations of cleavage various factors have been suggested. Reference has already been made to the works of Morgan as indicating that the production of the anlagen of organs in individuals developed from isolated blastomeres tends to require the same number of cells as required in normal development, even though the abnormal individuals are smaller than the normal and have fewer cells. Conklin's work has also been cited, wherein he considers the earlier cleavages


CELL CONSTANCY IN THE GENUS EORHYNCHUS 287

morphologically of more importance, while the later divisions in .Crepidula are believed to be non-differential and inconstant, but so far no one has offered a satisfactory explanation of why this is true. I do not here attempt to suggest any new factors as determining the number and arrangement of the cleavages involved in organ formation, but from the facts observed in the development of colonial protozoa of constant numbers, and in the on-' togeny of the metazoa where constancy has been demonstrated, there must be some inherent factor determining that a certain number of cell divisions shall precede the formation of a given organ, and that when this number is realized the power of further nuclear division becomes lost. Moreover evidence gained from the study of the members of the genus Eorhynchus tends to discredit the possibility of cell size acting as such a factor, for in case size determines the time when a cleavage or division takes place, cell division would continue throughout the life of the individual, since these worms continue to increase in size from the time they reach the final host until they are expelled from it. There must be some factor within the original cell which determines every cell division, from the first cleavage to the establishment of the final adult condition. Whether this determining element is a 'force' which becomes dissipated in the processes of cleavage so that, when the definitive number of cells has. been attained, no further progress is possible on the account of the lack of this 'force,' or whether the cause is to be sought in the presence of definite substances within the cell in the nature of either 'determiners' regulating cleavage or materials that are used up in the processes of mitosis can not be decided in the present stage of knowledge of the subject.

4. Cytomorphosis

In the genus Eorhynchus is found an evident exception to the generally accepted view of celluhar changes accompanying advance of age in animal tissues. Probably the most important point of difference, or at least the one most easily observed, is the fact that as soon as the animal has reached the point


288 H. J. VAN CLEAVE

of development when it has tak n on the body form of the adult, cell division either mitotic or amitotic, ceases to play anyrole whatever in the later history of the somatic cells. Among the nmnerous specimens of five species of this genus which I have examined there has not been, in any instance, the least indication of further increase in the number of cells constituting the body. This is evidenced by the entire lack of mitotic figures and also by actual count of the cells present. This gives good ground for doubting the generally accepted view that life must be accompanied by cell division as expressed in the contention that the cell has lost its lease on life as soon as it can no longer divide.

Nothing is known regarding the changes in Eorhjnichus accompanying the differentiation of the embryonic cells into the tissues and organ.s of the adult, for no larval stages of any of the American species have been discovered, and none of the investigations upon the European species contain references to the changes in the structure of the cells at that time. In the later stages of development some shght changes have been observed in the structure of the nuclei but there are no strongly marked general cytological modifications such as usually accompany the advance of age. The nuclear change consists chiefly in the rearrangement of the chromatin. In the immature forms the nuclei, especially of the subcuticula and of the lemnisci, have their chromatic substance irregularly scattered throughout the nucleus, but with the advance to maturity this chromatic material assumes the form of a more compact solid mass lying in the center of the nucleus.

The conditions of existence for all of the cells of the body are so nearly uniform at all periods of life for the adult that it seems probable that death must follow as a consequence of the termination of the reproductive period of the individual or as the result of a combination of factors acting on the organism as a whole rather than as a gradual senescence of the individual cells.


CELL CONSTANCY IN THE GENUS EORHYNCHUS 289

5. Brief review oj known extent of cell constancy in the animal

kingdom

Lest the impression be conveyed that cell constancy is confined to .occasional forms within a small group of animals the following condensed survey is given of the important work upon this topic. Martini has shown that constancy is present in most and probably in all of the organs of a new-born nematode. The muscle cells of Oxyurias are constant in number throughout life. Goldschmidt has demonstrated cell constancy in the nervous system of Ascaris. Brandes has found a definite number of brain cells in Gigantorhynchus, but has not pursued his studies to other organs of the body, nor has he deduced any general conclusions. Looss found constancy in the oesophagus of several nematodes. Apathy, working on the central nervous system of Hirudinea, emphasized the constancy found there. Hirschfelder found constancy in the midgut of Rotifera, while Martini following him has carefully worked out constancy in the entire body of Hydatina senta, so that he can recognize nine hundred and fifty nine individual cells, each of which he is able to locate in every of individual of that species. Woltereck has shown a constancy in the cells of the larva of Polygordius. Even in the Chordata constancy has been established by Martini in the various organs of Oikopleura, and Fritillaria. Aside from these definite records given as proofs of constancy by these writers, there exist masses of isolated facts such as have been cited in connection with the reviews of the work upon Acanthocephala in this article, but no attempt has been made to catalogue all of these scattered references. Miss Erdmann ('12) has given a very general survey of the literature upon cell constancy and the problems closely associated with it.

6. The significance of cell constancy

Cell constancy has a most interesting relationship to the problems of comparative anatomy and evolution. Evidently it can occur only in those forms which have a determinate cleavage. In what way has this condition been brought about? Were all forms of development originally of this type or has this condition been


290 • H. J. VAN CLEAVE

developed only in certain groups of animals? If the latter be the case, then the question arises as to the factor or factors operating in widely separated groups, which could accomplish such an end. Up to the present time no direct attempt has been made to answer this question. There seems to have been a tendency, at least among some workers, to expect too great a future for the study of cell constancy. .This has probably grown out of the failure to realize fully the limitations of the field. There is the possibility of a comparative anatomy, such as Martini has predicted, in which two forms may be compared one with the other in a much more intimate way than has been undertaken heretofore. In this case individual cells rather than gross anatomical structures would be used as units for comparison. An examination of the facts brought out in the earlier parts of this paper will show that they constitute just such a study in comparative anatomy. Necessarily such a study is limited to those groups which have acquired constancy.

The fact that cell constancy has never been demonstrated in any marked degree in organisms which hold an unquestioned position in the main line of descent of the animal series seems to indicate that it is not the primitive condition. The Nematodes, the Tunicates, the Rotifers, and the Acanthocephala, forms which most clearly display the phenomena of constancy, stand either as side branches from the main line of descent of animal life or as highly modified groups which have lost most of their indications of close relationship with other forms. For instance the rotifers have reached a point in development but little higher than the trochophore stage found in the development of higher forms and have become permanently fixed. The tunicates, coming off from the stem which has given rise to the Vertebrates, have, in a similar manner, either failed to proceed farther or, in case at one time they stood on a higher level, have through degeneration regressed to a condition where they have become fixed as an aberrant group. Parasitism has so profoundly affected the somatic portions of the Acanthocephala that little is left to serve as an indication of earlier phylogenetic relations. Especially, in this group, species and even genera differ from each other in slight


CELL CONSTANCY IN THE GENUS EORHYNCHUS 291

degree when compared with the range of variation in the more plastic groups of animals. This tendency toward uniformity of body characters has led to, or possibly has resulted from, a corresponding stability in the numbers of cells which go to make np the organs of the body in the genus Eorhynchus. It is true that within this genus the parasitic habit has contributed toward the elimination of physical factors which might induce radical changes. The result of this is that variability is practically confined to minor variations such as those of body size, and in general those which concern slight rearrangements of the component cells, rather than those involving the production of entirely new types of structures. The future of such a group of organisms is very clear. The stability which has been produced serves as an effective barrier to evolution. Under present conditions those organisms displaying marked cell constancy are incapable of producing new creations in the lines of organic progress, for the elimination of variability has precluded the possibility of progressive evolution.

I wish to express my indebtedness to Prof. H. B. Ward for the many helpful suggestions and kindly criticism of methods and of general outline of the work, rendered throughout the period of its preparation.


292 H. J. VAN CLEAVE

BIBLIOGRAPHY

VON Apathy, St. 1893 tJber die Muskelfasern von Ascaris. Zeitschr. f. wiss. Mikr., Bd. 10, pp. 36-73, 319-361.

Balzer, C. 1880 Zur Kenntnis der Echinorhynchen. Arch. f. Nat., Bd. 46, pp. 1-40.

BiELER, W. 1913 tJber den Kittapparat von Neorhynchus. Zool. Anz., Bd. 41, pp. 234-236.

Brandes, G. 1899 Das Xervensystem der als Nemathelminthenzusammengefassten Wurmtypen. Abhandl. d. Naturf. Ges. Halle, Bd. 21, pp. 273-299.

CoNKLiN, E. G. 1898 Cleavage and differentiation. Biol. Lecture, Woods Hole, 1896-7.

1912 a Cell size and nuclear size. Jour. Exp. Zool., vol. 12, pp. 1-98. 1912 b Body size and cell size. Jour. Morph., vol. 23, pp. 159-188.

DujARDiN, F. 1845 Histoire naturelle des Helminthes. Paris; 483 pages.

Erdmann, R. 1912 Quantativ Analysis der Zellbestandteile bei normalem, experimentell veriindertem und pathologischem Wachstum. Ergebn. d. Anat. u. Entw-gesch., Bd. 20, pp. 471-566.

Graybill, H. W. 1902 Some points in the structure of the Acanthocephala. Trans. Amer. Micr. Soc, vol. 23, pp. 191-200.

Greeff, R. 1864 L'eber die Uterusglocke und das Ovarium der Echinorhynchen. Arch. f. Naturges., Bd. 30, pp. 361-375.

Hamann, O. 1890 Die Lemnisken der Nematoden. Zool. Anz., Bd. 13, pp. 210212.

1891 Die Nemathelminthen. Monographie der Acanthocephalen. Beitrage zur Kenntnis ihrerEntwicklung, ihres Baues, und ihrer Lebensgeschichte. I. Jen. Zeitschr. f. Naturwiss., Bd. 25, pp. 113-231.

1892 Das System der Acanthocephalen. Zool. Anz., Bd. 15, p. 195. 1895 Die Nemathelminthen. Monographie der Acanthocephalen. II. Jena. .

Hirschfelder, G. 1910 Beitrage zur Histologic der Riidertiere. Zeitschr. f. wiss. Zool., Bd. 96, pp. 209-234.

Kaiser, J. E. 1890-1892 Die Acanthocephalen" und ihre Entwicklung. Zoologica, Bd. 7.

1013 Die Acanthocephala und ihre Entwicklung, Beitrage zur Kenntnis der Histologic, Ontogenie und Biologie einiger einheimischer Echinorhynchen. Leipzig; 66 pages, 2 plates.

Lespes, C. 1864 Sur quelques points de I'organisation des Echinorhynques. Jour. Anat. et Physiol., Tom. 1, pp. 683-686.

Linton, E. 1889 Notes on entozoa of marine fishes of New England, witli descriptions of several new species. Rep. U. S. Fish Comm. for 1886, pp. 453-498.


CELL CONSTANCY IN THE GENUS EORHYNCHUS 293

LtJHE, M. 1905 Geschichte und Ergebnisse der Echinorhynchen-Forschung bis auf Westrumb (1821). Zool. Ann., Bd. 1, pp. 139-353.

1907 liber Cementbildung bei Nematoden und Acanthocephalen. Schr. physik-okonom. Gesell. Konigsberg, Bd. 47, pp. 88-89.

1911 Die Siisswasserfauna Deutschlands, Heft 16, Acanthocephalen. Jena; 116 pages; 87 figures.

Martini, E. 1908 Die Konstanz histologischer Elemente bei Nematoden nach Abschluss der Entwicklungsperiode. Verhandl. Anat. Ges., Bd. 22, pp. 132-134.

1909 Darwinismus und Zellkonstanz. Sitzungsber. u. Abhandl. d. Naturf. Ges, Rostock, Bd. 1, pp. 1-10.

1909 a tJber Eutelie und Neotenie. Verhandl. d. deuts. Zool. Ges., 1909. pp. 292-299.

1909 b Studien fiber die Konstanz histologischer Elemente, I, Oikopleura longicauda. Zeitschr. f. wiss. Zool., Bd. 92, pp. 563-626.

1909 c tJber Subcuticula und Seitenfelder einiger Nematoden, IV, Vergleichend-histologischer Teil. Zeitschr. f. wiss. Zool., Bd. 93, pp. 535-624.

1909 d Studien liber die Konstanz histologischer Elemente. II. Fritillaria pellucida. Zeitschr. f. wiss. Zool., Bd. 94, pp. 81-170.

1912 Studien liber die Konstanz histologischer Elemente, III., Hydatina senta, Zeitschr. f. wiss. Zool., Bd. 102, pp. 425-645.

MiNOT, C. S. 1908 The problem of age, growth, and death; a study of cytomorphosis. New York.

AI0RG.A.N, T. H. 1896 The number of cells in larvae from isolated blastomeres of

Amphioxus. Arch. f. Entw-m., Bd. 3, pp. 269-294.

1901 Regeneration. New York. MoRGULis, S. 1911 Studies of inanition in its bearing upon the problem of

growth. I. Arch. f. Entw-m. Bd. 32, pp. 169-268. Peter, K. 1909 Experimentelle Untersuchungen liber individuelle Variation

in der tierischen Entwicklung. Arch. f. Entw-m., Bd. 27, pp. 153-246.

1911 Neue "experimentelle Untersuchungen liber die Grosse der Vari abilitjit und ihre biologische Bedeutung. Arch. f. Entw-m., Bd. 31,

pp. 680-804. Sakfftigen, A. 1884 Zur Organisation der Echinorhynchen. Morphol. Jahrb.,

Bd. 10, pp. 1-53. Van Cleave, H. .1. 1913 The genus Neorhynchus in North America. Zool.

Anz., Bd. 43, pp. 177-190. Weismann, a. 1893 The germ-plasm, a theory of heredity. New York.

Wilson, E. B. 1892 The cell-lineage of Nereis. .Jour. Morph., vol. 6, pp. 361480. 1900 The cell in development and inheritance. New York, 483 pages.


294


H. J. VAN CLEAVE


ABBREVIATIONS


br., brain

bu., bursa

biv., body-wall

c, cuticula

cc, circular canals

eg., cement gland

ci., cirrus

cm., circular body muscles

cr., cement reservoir

/■/., fan ligament

gg., genital ganglion

go., genital opening

ht., terminal hook

hb., basal hook

km., middle hook

ip., proboscis invertor

I., 1.2, lemnisci

lor., longitudinal canal


lin., longitudinal body muscle

ms., muscle sac

po., terminal proboscis organ

prn., nuclear ring of proboscis

r., receptacle of proboscis

s., subcuticula

sa., selective apparatus

sn., subcuticular nuclei

sr., sperm reservoir

ssa., stay of selective apparatus

ta., anterior testis

tp., posterior testis

?/., uterus

v., vagina

vd., vas deferens

ve., vas efferens

vs., vaginal sphincter


EXPLANATION OF FIGURES A camera lucida was used in making all the drawings included in the figures.


Figs.

Fig.

Fig.

Fig.

Fig. X 310.

Fig. nuclei.


1 to 16 Represent Eorhynchus gracilisentis (Van C).

1 Proboscis of hematoxylin-stained specimen in balsam. X 97.

2 Hooks from the same specimen as shown in figure 1. X 310.

3 Embryos from body of mature female. X 310.

4 Portionof body wall showing a subcuticular nucleus. Sagittal section.


5 Sagittal section to one side of that shown in figure 4, showing muscle X 310.

Fig. 6 Sagittal section through proboscis showing terminal proboscis organ and hook retractors; hematoxylin. X 310.

Fig. 7 Sagittal section through proboscis receptacle of inverted specimen showing terminal organ of the proboscis. X 310.

Fig. 8 Sagittal section through base of proboscis receptacle showing location of brain. X 310.

Fig. 9 Cement gland of male; toto mount. X 97.

Fig. 10 Transverse section through tip of proboscis showing terminal ring of twelve nuclei. Iron hematoxylin. X 310.

Figs. 11 and 12 Reconstructions of female genital tract. X 310.

Fig. 13 Female genital tract in toto mount; hematoxylin. X 310.

Fig. 14 Sagittal section of same, in region of vaginal sphincter. X 310.

Fig. 15 Fan-shaped ligament at side of female genital orifice. X 310.

Fig. 16 Retractors of the proboscis receptacle, from di.ssection. X 310.



295


29() H. J. VAN CLEAVE


Figs. 17 to 19 Eorhynchus gracilisentis. (Van C).

Fig. 17 Genital organs in posterior region of male ; reconstruction from sagittal sections.

Fig. 18 Reconstruction of male copulatory apparatus. X 310.

Fig. 19 Male copulatory apparatus everted; toto mount. X 97.

Figs. 20 to 28 Eorhynchus longirostris (Van C).

Fig. 20 Cement gland. X 97.

Fig. 21 Proboscis of fully mature female; in balsam. X 97.

Fig. 22 Anterior end of body; proboscis retracted; toto mount. X 30.

Fig. 23 Portion of body wall in sagittal section. X 310.

Fig. 24 Embryos. X 310.

Fig. 25 Nuclei from muscle sac, posterior to cement reservoir. X 310.

Fig. 26 Sagittal section through anterior end of body, showing location of brain and terminal proboscis organ. X 97.

Fig. 27 Reconstruction of female genital tract, showing divisions and the nuclei. X 97.

Fig. 28 Dorsal retractor of proboscis receptacle; toto mount. X 97.

Fig. 29 Eorhynchus tenellus (Van C), proboscis of mature male. X 97.

Fig. 30 Hooks of Eo. tenellus. X 310.

Fig. 31 Eorhynchus rutili; copied from Liihe, 1911. About X 20.




26



2 3 im.


m f i


24




297


298 H. J. VAN CLEAVE


Fig. 32 Eorhynchus emydis (Leidy); proboscis of fully mature female. X 97.

Fig. 33 Hooks of Eo. emydis. X 310.

Fig. 34 Embryos of Eo. emydis. X 310.

Fig. 35 Proboscis of Eo. cylindratus (Van C.) X 97.

Fig. .36 Hooks of Eo. cylindratus. X 310.

Fig. 37 Embryos of Eo. cylindratus. X 310.

Fig. 38 Eo. tenellus; reconstruction from sagittal sections of anterior end of body. X 97.

Fig. 39 Eo. tenellus; reconstruction of posterior end of female, showing genital tract. X 97.

Fig. 40 Eo. emydis (juv.) ; genital tract of female; toto mount. X 310.

Fig. 41 Eo. emydis; Cement gland of mature male. X 97.

Fig. 42 Eo. emydis; Cement gland of immature male. X 97.

Fig. 43 Eo. emydis (juv.) ; lemnisci in toto mount. X 97.

Fig. 44 Eo. emydis (juv.); proboscis receptacle and adjacent organs; toto mount stained in hematoxylin. X 97.

Fig. 45 Eo. agilis; longitudinal section, showing especially the terminal proboscis organ. Copied from Hamann, 1895, plate 3, figure 28.

Fig. 46 Eo. rutili; anterior end of body. Copied from Saefftigen, 1884, plate 5, figure 6.


THE INNERVATION OF THE INTEGUMENT OF " CHIROPTERA^

JAMES EDWARD ACKERT

From the Zoological Laboratory of the University of Illinois

TWENTY-ONE FIGURES (FOUR PLATES)

CONTENTS

Introduction 301

Material and methods 302

Observations and discussion 303

General structure of the integument ■ 303

1. Integument of the body 303

2. Flying and interfemoral membranes 306

Nerve layers of the integument 308

1. Nerve layers of the body integument 308

2. Nerve layers of the flying and interfemoral membranes 308

Nerve endings in the integument 313

1. Free nerve terminations in the epidermis 313

2. Nerve endings on hairs 315

3. Special sensory end-organs 321

a. End-bulbs 322

b. Terminal corpuscles 324

4. Motor nerve endings on striated muscles : 325

5. Nerve endings on modified sweat glands 327

What sensory organs are concerned when blinded bats avoid obstacles

while on the wing 328

Summary 332

Bibliography 334

INTRODUCTION

As is well known, the skin of bats is very sensitive to tactile stimulation. These animals in captivity give vigorous responses when various parts of their bodies and membranes are touched. Even eighteenth century investigators thought the integument of bats was especially adapted for the perception of delicate

1 Contributions from the Zoological Laboratory of the University of Illinois, under the direction of Henry B. Ward, No. 29.

301

JOURNAL OF MORPHOLOGY, VOL. 25, NO. 2


302 JAMES EDWARD ACKERT

tactile stimuli, for Spallanzani and CXivier observed that bats deprived of sight avoided small objects with accuracy. Cuvier found the wings to be supplied with an enormous number of nerves, and thought that during flight the blinded bat, on approaching the object, sets up air currents, which, reacting on the sensitive patagium and external ears, enable the animal to avoid the obstacle. That the sense of touch is more highly developed in bats than in other mammals was asserted by Schobl ('71), who described 'Terminalkorperchen' at the bases of' the hairs. Moreover, Redtel, two years later, maintained that it is possible for these animals to perceive the slightest difference of external air pressure upon their wings.

The extreme sensitiveness of the integument of these animals and the possibilities of modern technique seemed to justify a further search for sensory structures in their skin. Moreover, at the time this work was begun, no investigator had made an extensive study of the innervation of the skin of bats since Schobl published in 1871 his account of the terminal corpuscles in the flying membrane.

The work has been carried on in the Zoological Laboratory of the University of Illinois under the direction of Prof. Frederic W. Carpenter, to whom I am indebted not only for his personal interest in the progress of the work, but also for his constant advice and helpful criticisms.

MATERIAL AND METHODS

The material for this investigation consisted of forty-one bats of which thirty-one were cave bats (Myotis lucifugus) from Indiana. The remainder, the common red bats (Myotis subulatus), were taken in the vicinity of Urbana, Illinois.

Most of the material was prepared by an .intra vitam methylene blue method. The blood was washed out of the freshly etherized animal, and a 1 per cent solution of methylene blue in distilled water injected into the arterial system through the heart. After leaving the animal, with its vessels full of staining fluid, freely exposed to the air for oxidation, the fluid was washed out with


INNERVATION OF INTEGUMENT OF CHIROPTERA 303

Ringer's solution, and small pieces of tissue fixed in a cold 8 per cent solution of ammonium molybdate in distilled water. The tissues were then dehydrated, cleared in xylol, and imbedded in paraffine. Sections of 20^ were thick enough to enable one to follow the nerve fibers some distance, and sufficiently thin to admit ample light. For checking results other methods of preparing material were employed. The killing and fixing fluids used for this pvu'pose were corrosive sublimate and acetic acid, Zenker's fluid, ammoniacal alcohol, and 10 per cent formol. The stains used included silver nitrate (Cajal method for nerve fibrils), carmalum as a counter stain for methylene blue, Mallory's connective tissue stain, Heidenhain's iron hematoxylin, Hanson's hematoxylin and orange G, and Delafield's hematoxylin and eosin, the last of which proved the most satisfactory for general use.

OBSERVATIONS AND DISCUSSION GENERAL STRUCTURE OF THE INTEGUMENT

1. Integument of the body

The skin of the body of bats is covered with hair which, as Allen ('93) has found, varies in different regions in texture and amount. In general, the crown of the head, the neck, the sides of the under surface of the body, the rump and the pubis have a thick pelage, while the distal portions of the ears, the soles of the feet, the mammae and the external genitalia are almost naked. The snout is scantily clothed, but shows a limited number of vibrissae which arise from wart-like structures.

In different regions of the body the skin varies greatly in thickness. The integument of the face is the deepest ; that of other parts of the body diminishes in depth gradually in the following order: palmar region, plantar region, rump, ventral thoracic region, crown, and dorsal thoracic region.

As a rule, some difficulty is experienced in distinguishing all the layers commonly found in the human integument. In the epidermis the Malpighian or deeper stratum can be readily made out. Its deepest layer is made up of subcolumnar cells.


304 JAMES EDWARD ACKERT

The intermediate layer of polygonal cells is for the most part absent, though in places (e.g., the face) it appears as a single sheet of isolated, more or less flattened cells, whose nuclei are somewhat reduced in size (fig. 1). Numerous pigment granules are present in this layer (fig. 1, pg).

The stratum corneum is thickest in the palmar and plantar regions. It is made up of several layers of cornified epithelium, the outer ones of which are usually in the form of loose scales. The deeper layers are more compact, and appear to consist of flat, enucleate cells. In certain regions of the body (lining of the mouth, lumbar region) these layers resemble to some extent the stratum lucidum of the human skin, but the presence of this stratum can be made out definitely only in the palmar and plantar regions (fig. 2, si).

The surface of the epidermis is frequently interrupted by hairs, and also by' the openings of ordinary sudoriparous and of modified sweat glands as Diem ('07), Porta ('10) and others have shown. The ordinary sweat glands and the modified sweat glands may open into the hair follicle, or independently on the surface. The distribution of skin glands over the body is very variable. Though not numerous in the region of the rump, sweat glands are, however, present. This is in accord with Diem's results, but opposed to those of Hoffman ('98). The writer was unable to find sweat glands in the sole of the foot, and agrees with Toldt ('07) that these glands do not occur in the ball of the thumb. Toldt found numerous glands in the 'Saugescheibe,' and also large groups of glands in the region of the neck and of the external genitalia. The upper lip is more abundantly supplied with skin glands than any other part-of the body.

As is frequently the case, the superficial layer of the corium, the stratum papillare, is raised into ridges and papillae which project into the epidermis. These are most marked in the upper lip, where simple and compound papillae are present. The interlacing strands of connective tissue and the reticulum of elastic fibers which together form the ground work of the corium are comparatively fine and closely packed, thus causing this layer to be somewhat dense. Mallory's connective tissue stain


INNERVATION OF INTEGUMENT OF CHIROPTERA 305

shows that the general direction of these strands and fibers is parallel with that of the stratum Malpighii. While it is not possible to determine a boundary between the stratum papillare and the stratum reticulare, yet the deeper connective tissue bundles of the latter are obviously more loosely interwoven than those of the superficial layer of the corium. As in other mammals, the corium contains blood vessels, hair follicles, sebaceous glands, sudoriparous glands, striated and smooth muscle fibers, nerve trunks, medullated and non-medullated nerve fibers, tactile corpuscles, ajnd nerve endings. The last three structures mentioned will be described in detail later.

As has been noted, the upper lip of the bat is richly supplied with skin glands. One type of these, the modified sweat gland, differs somewhat from the typical sweat gland, so a description of its structure may not be out of place here. Compared with a hair follicle, this gland is enormous in size. It consists of a long, uncoiled secreting portion with an extended funnel-shaped duct. The secretory portion is lined by a single layer of columnar cells with finely granular protoplasm and round or oval nuclei (fig. 3, cc). Leydig, Schobl and Sabussow ('10) have called attention to the fact that these large modified sweat glands (in the flying and interfemoral membranes) have a coating of smooth muscle fibers, which, by their longitudinal course, cause a slight spiral striping of the gland. This coating of muscle fibers (fig. 3, mf) lies between the layer of columnar cells and an external covering or basement membrane (fig. 3, bm). The latter is homogeneous and without nuclei. The duct of the gland is lined throughout by short, somewhat irregularly cubical cells, arranged in a single layer, and surrounded by a delicate basement membrane. Not infrequently secretion products are found in the lumina of the glands. The products are more or less similar in appearance to what Wimpfheimer ('07) terms degeneration products ('detritus') found in uncoiled sweat glands in young moles.

It is worthy of remark that pigment cells occur in the corium both of the body integument and of the flying and interfemoral membranes (fig. 2, pc). In, the corium of the integument they


306 JAMES EDWARD ACKERT

are numerous, and appear to be scattered about promiscuously. Their form is very variable. They may be spherical, oval, elongate and slightly spiral, heart-shaped,' pear-shaped, raggedly lobulate, and with or without processes (fig. 5). In size they vary from 34)u in length and 25.5ju in width to 374iu in length and 60ju in width. The cell body is filled with fine brown granules. In hematoxylin-eosin preparations a few of the granules usually take the dull blue stain of the hematoxylin, while in the methylene blue material some or all of the granules may stain a bright blue. Figure 5, h represents a pigment cell containing stained and unstained granules.

2. Flying and interfemoral membranes

The flying membrane of bats is a skin duplicature formed by the lateral extension of the dorsal and ventral integument of the body. The proximal parts of the membranes are covered with fine hairs similar of those of the pelage, while over the distal areas extremely fine, more or less modified hairs occur sparsely. In the natural condition there is a manifold wrinkling and plaiting due to numerous elastic bands within the membrane (Schobl) .

Externally the flying membrane is made up of small, hexagonal, plate-like cells which form a continuous membrane. Each cell contains pigment granules which are collected into an intramarginal zone much as Schobl has described (p. 4). This investigator reports that the center and border of the cell (in Vesperugo serotinus) are free from pigment granules. In Myotis lucifugus and M. sublatus the writer found pigment in both of these regions, but in smaller quantities than in the intra-marginal zone. The cells of the outer (dorsal) surface of the flying membrane contain more and darker pigment granules than do those of the inner (ventral) surface (Schobl). In fact this surface in places contains almost no pigment.

As in the integument of the body the epidermis of the flying membrane stands out in the sections in contrast to the cutis. The Malpighian layer also can be readily distinguished from the stratum corneum. According to Schobl's studies the Mai


INNERVATION OF INTEGUMENT OF CHIROPTERA 307

pighian stratum is composed of two layers of scattered cells. The writer, however, finds that one layer of cells occurs quite as frequently as two. The nuclei of the deeper layers of both dorsal and ventral sides are slightly more oval than those of the more superficial layers, the latter being somewhat flattened. From the shape of the nuclei one would infer that when a single layer occurs it is the outer one. In the Malpighian stratum of the dorsal side of the patagium numerous pigment granules are present, while in this stratum on the ventral side very little pigment occurs. Aside from being somewhat thinner, the stratum corneum does not differ from the corresponding structure in the skin of the body.

The tissue enclosed between the dorsal and ventral Malpighian strata of the patagium constitutes the corium, which varies in thickness in different regions. In both the flying and interfemoral membranes it is thickest near the body, while in the more distal areas it gradually becomes thinner.

The corium is made up of three poorly defined strata of connective tissue, a central, somewhat loose one, corresponding to the stratum reticulare of the body integument, and two others — one on either side — of denser tissue, more or less similar to the stratum papillare. The chief arteries, which are accompanied by the larger veins and nerve trunks, cause this stratum to be much thicker in those regions where they occur than elsewhere. Although the outer surfaces of this stratum are thrown into folds to some extent, the writer has been unable to find papillae.

In the stratum reticulare are contained the larger blood vessels and nerves, and the striated muscle bundles and elastic bands (Balken) first described by Leydig, whose results were later confirmed by Schobl. Here also are found the secreting portions of sweat glands, and the proximal third of hair follicles. The outer stratum of the corium contains the central portions of the hair follicles, their sebaceous glands, and the sweat glands. Each follicle, with the sebaceous, sudoriparous and modified sweat glands associated with it, is surrounded by a capillary network.


308 JAMES EDWARD ACKERT

NERVE LAYERS OF THE INTEGUMENT

1 . Nerve layers of tha body integument

In the subcutaneous tissue and in the reticular stratum of the body integument, are large meduUated nerve trunks and branches which, for convenience, are called the first nerve layer. By dichotomous branching these nerves break up into a loosely intertwined meshwork consisting of an enormous number of medullated nerves. These interwoven nerves, which are not

actually united in a plexus, constitute the second nerve layer (fig. 4, snl). Arising from the latter are medullated nerves which pass toward the periphery. Near the outer surface of the corium they begin to divide. The resulting non-meduUated branches pass directly to the Malpighian stratum, forming the third nerve layer (fig. 4, tnl). Ordinarily so much pigment is present here that it is impossible to follow the fibrils to their endings. However, in places where the epidermis has accidentally been torn, one can readily trace the fibrils well into the Malpighian stratum, noting branching fibrils which pass outward and terminate in or between the cells of the stratum granulosum. As these can be traced more readily in the membranes where little or no pigment is present they will be considered more fully later (p. 313). Varicosities are numerous both in the second nerve layer and along the fibers which pass to the third layer (figs. 14, 15, 16, 17). The greater number of these enlargements, however, occurs on the smaller fibers. Varicosities in the third nerve layer, that is, on the surface of the corium or in the Malpighian stratum, have not as yet been observed by the writer. As the literature on nerve layers in the skin of bats deals almost entirely with these layers in the patagium, the brief historical survey will be given in the consideration of the flying membrane.

2. Nerve layers of the flying and interfemoral membranes

As early as 1796 Cuvier called attention to the abundance of nerves in the flying membranes of Chiroptera. Leydig, a halfcentury later, while differing somewhat with Cuvier as to the number of nerves present, admitted that these membranes are richly innervated.


INNERVATION OF INTEGUMENT OF CHIROPTERA 309

The first investigator, however, to make an intensive study of the arrangement of the nerves in the flying membranes was Schobl ('71), who states that the nerves of the patagimn are naturally divided into five layers. The first layer, situated in the innermost stratum of the flying membrane, contains the larger nerve trunks, the main blood vessels, the chief muscles and the elastic bands. The second nerve layer is double, one part lying above, and its duplicature below the first layer. The nerve trunks of this layer branch dichotomously again and again, forming an irregular network. The third layer of nerves, which is also double, lies external to the previous one, on a level with the finest blood vessels. As to the size of the nerve trunks of this layer, Schobl states that they consist usually of two, very rarely of four, non-meduUated fibers. The fourth nerve layer, likewise double, lies outside of the third. It consists throughout of an irregular net of single non-meduUated fibers. The meshwork in this layer, however, arises not by an interlacing of fibers, as is the case in the other two layers, but by direct anastomosis of single non-medullated fibers. On certain fibers of this layer Schobl noted a number of enlargements or swellings ('Krootenpunkten'), which were triangular, square, or polyhedral in form, having a fine granular appearance, but exhibiting no nuclei. He also occasionally saw more or less similar spindle-shaped enlargements in the course of a single fiber, especially the larger ones. The fifth and last nerve layer, also double, Hes immediately over the previous one on the surface of the cutis, ordinarily remaining attached to the deepest cells of the Malpighian stratum. The fibers of this layer are likewise non-medullated, and have a diameter ranging from O.Q/x to immeasurable fineness. This layer arises from the previous one by the division of the finest fibers of the latter. At the places of division of the fibers, the swellings which were found so frequently in the last nerve layer seldom occur in this one, and the spindle-shaped variety is lacking entirely. This layer of extremely fine non-medullated nerve fibrils lying immediately at the surface of the corium, partly between the lowest cells of the Malpighian stratum, Schohl holds as a terminal. He further


310 JAMES EDWARD ACKERT

states that in preparations in which the lowest cells of the Malpighian stratum remain undisturbed on the corium, no free endings of the finest non-medullated nerve fibrils are found, and that fibrils passing further toward the surface between the cells of the stratum granulosum, are never to be found either on the surface of the preparation or in cross section. He pointed out, however, that occasionally round or elliptical, swollen structures resembling fine nerve endings are to be seen, but these almost always prove to be nodal points of division of nerve fibrils. These minute swellings occurred so seldom that Schobl attributed their presence to faulty technique.

Sabussow ('10), working on the innervation of the flying membrane, did not wholly accept Schobl' s idea of the distribution of the nerves of this part of the body. This investigator found large nerve trunks in the innermost stratum of the patagium, but held that Schobl' s second nerve layer lay in the same plane as the first, and consequently could not be said to exist as a separate nerve layer. Concerning Schobl' s third nerve layer Sabussow simply stated that it is not double. But this investigator confirms the existence of the fourth nerve layer of Schobl, adding that, no matter how the membrane be torn, this layer can be seen to be double. He also confirms Schobl's fifth layer, which is non-medullated and double; but instead of the few 'swellings' which Schobl observed, Sabussow found numerous varicosities. The latter sums up the layers he found as follows: (1) a simple layer including the first two layers of Schobl; (2) a broadly meshed double network with triangular enlargements in it ; (3) a network of varicose fibers also double. Consequently according to Sabussow, there are five nerve layers in the patagium.

In transverse sections of my own preparations of the flying and interfemoral membranes there can readily be seen, here and there, regions which are approximately twice as thick as that of the remaining area of these membranes. It is in these thickened regions that the chief arteries, veins, nerve trunks and frequently the principal muscle bundles are found. These particular regions contain, as will be shown, one more layer of nerves than do the others.


INNERVATION OF INTEGUMENT OF CHIROPTERA 311

The main blood vessels, accompanied by the chief nerve trunks, pass out from the body through the flying and interfemoral membranes in the stratum reticulare, giving off, here and there, important branches, which, as stated, are frequently found with the muscle bundles. These blood vessels, partly because of their own size, and partly on account of the increased amount of connective tissue around them, cause the elongated thickenings or ridges already referred to in these membranes. The meduUated nerve trunks and their chief branches, both found in the innermost stratum (reticulare) and existing only in the aforesaid ridges, constitute the first nerve layer". The second, a double layer of nerves, arises from the first by repeated dichotomous branching, traverses the deeper part of the corium, and spreads throughout the entire area of the flying and interfemoral membranes. In methylene blue preparations this layer is seen to consist of a loose network of medullated nerve fibers, many of which contain comparatively large varicosities (fig. 13, va) . The third and last nerve layer is likewise double. Numerous medullated fibers arising from the second nerve layer pass toward the two external (dorsal and ventral) surfaces of the membranes. Many of these fibers, on approaching the Malpighian stratum, divide dichotomously ; others do so at the surface of the corium. Both lose their medullation. The forked branching continues to some extent in the Malpighian stratum, the larger fibrils giving off smaller ones, until finally delicate nerve threads end in minute enlargements, which will be described in detail later. These branchings of non-medullated nerve fibrils at the surface of the corium and in the stratum Malpighii constitute the third nerve layer. While varicosities of different sizes (figs. 13, 14, 15) appear in the nerve fibers leading up to this layer, the writer has not observed them in the latter.

According to the present observations, then, certain regions of the flying and interfemoral membranes are supplied with three layers of nerves, others with but two. Briefly stated their number and distributiofi are as follows:

1. A layer of medullated nerve trunks and numerous medullated branches, occurring in the stratum reticulare, but only in


312 JAMES EDWAED ACKERT

the elongated ridges containing the largest blood vessels and much connective tissue.

2. A double meduUated nerve layer in the deeper part of the corium, extending throughout the membranes.

3. A layer, likewise double, present in the entire Malpighian stratum, and consisting of numerous branches of non-medullated nerve fibrils.

A comparison of the foregoing results shows that the first nerve layers of Schobl and of the writer coincide; that Sabussow's first layer included Schobl's first and second layers and the writer's first, together with the innermost branches of his second nerve layer. A study of sections from different parts of the membranes has convinced the writer that Schobl's second, third and fourth nerve layers may well be considered as one layer, the writer's second. Close to the body, where these membranes are thick, and where Schobl probably made his observations, since he especially recommended this region for study, it is true that the writer's second layer is thicker dorso-ventrally than it is near the elbow, or in the region midway between the body and the tail. But at the periphery, between the elongated phalanges, and near the posterior border of the interfemoral membrane, where the skin duplicature is thin, this nerve layer is exceedingly compressed. The contention of Sabussow that Schobl's second nerve layer lay in the same plane as the first, and consequently could not be considered as a separate layer, is not supported by the present observations. Schobl's first layer is to be found in the stratum reticulare, while the second, arising by repeated dichotomous branching of the first, takes a position in the deeper part of the superficial layer of the corium. The writer's second nerve layer corresponds to Schobl's second, third and fourth layers, while Sabussow's second layer includes the third and fourth layers of Schobl, and the greater part of the writer's second. The third nerve layer of Sabussow and of the writer, respectively, corresponds to Schobl's fifth.


INNERVATION OF INTEGUMENT OF CHIROPTERA 313

NERVE ENDINGS IN THE INTEGUMENT

The nerve terminations in the integument of Chiroptera may be grouped into five classes as follows:

1. Free nerve terminations in the epidermis

2. Nerve endings on hairs

3. Special sensory end-organs

a. End-bulbs

b. Terminal corpuscles

4. Motor nerve endings on striated muscles

5. Nerve endings on modified sweat glands

1. Free nerve terminations in the epidermis

As was stated in an earlier part of this paper (p. 308), free nerve endings in the form of minute swellings were observed in the stratum Malpighii. These free nerve terminations or end-knobs can most readily be seen in sections of the ventral portions of the membranes, where little or no pigment is present. Especially desirable for this purpose are oblique sections, or those which contain small areas of the surface of the membrane (fig. 6). In such sections it is possible to focus down through the transparent stratum corneum, thereby obtaining distinct views of the deeply stained (blue) nerves of the third layer (fig. 6, n). The latter stand out in bold contrast to the weakly stained cytoplasm of the Malpighian stratum. In sections 20/x thick one can, b}^ focusing, trace non-medullated nerve fibers, from the point of branching near the surface of the corium, on out among the deeper Malpighian cells. The larger fibrils and the smaller ones given off from them are plainly visible. Finally, among the cells of the stratum granulosvmi (fig. 6, sgr), the ultimate branches terminate in minute round or oval end-knobs (figs. 6, 7, e). Similar structures were mentioned but misinterpreted by Schobl, who, observing a very limited, number of minute round swellings resembling fine nerve endings in the stratum Malpighii concluded that they were foreign particles due to faulty technique.

The end-knobs take a deep blue stain similar to that of ordinary axis cylinders, and appear to be homogeneous in structure. They are oval or spherical, their size varies from 0.5/x in length


314 JAMES EDWARD ACKERT

and 0.4,u in width to 0.9/i in length and O.S/x in width. In the sections studied, these end-knobs appear to be numerous. Ordinarily, one to a cell is observed, though occasionally even two are seen close to a single cell boundary. Sometimes a tiny fibril appears to end without any enlargement (fig. 7, x) . This, however, may be due to the failure of the methylene blue to differentiate the end-knob, as those who have worked with this stain will readily understand.

That these diminutive enlargements or end-knobs are real nerve terminations and not the nodal swellings sometimes seen where fibers divide has been satisfactorily proven. For example, in focusing on the surface of the transparent stratum corneum no knobs nor fibrils can be seen. A deeper focus brings into view end-knobs with a fine nerve fibril running into each. A still deeper focus shows tissue below the end-knobs and enables one to follow the nerve fibrils from the now indistinct terniinal swellings back to the branch from which the nerve fibrils are given off. Where Kttle or no pigment is present, these nerve terminations can be seen without difficulty.

For a time the writer was unable to determine whether the nerve end-knobs are situated in the stratum Malpighii or in the deepest layers of the stratum corneum. At length, however, a section was found in which a part of the ventral surface of the interfemoral .membrane curled up, permitting an oblique view. The methylene blue stain was deep enough to show the margins of a number of consecutive superficial cells of the stratum granulosum (fig. 6), and little pigment was present. By focusing upon this obliquely turned portion of the surface of the membrane, it was comparatively easy to distinguish the flat, elongate, scalelike cells constituting the stratum corneum from the more oval, clearly defined, superficial cells of the Malpighian stratum. By focusing upon the curved surface it was possible to see a number of nerve end-knobs on or near the surface of the stratum granulosum, but as yet no end-knobs have been seen by the writer in the stratum corneum.

The question of the exact position of the end-knobs in respect to the epithelial cells naturally arises. It is certain that a large


INNERVATION OF INTEGUMENT OF CHIROPTERA 315

number of the structures in question are situated on the surfaces of the cells (fig. 7, es) . Others appear to be within the cytoplasm. However, it is frequently possible by focusing to see that these end-knobs are after all on the borders of the cells. If all were completely stained, it is not improbable that the remaining endknobs could be shown to be intercellular.

So far as the writer has been able to ascertain the only refe*'ence to free nerve terminations in the epidermis of bats is that of Botezat ('08). The study of this investigator was made principally on the nerves in the epidermis of the dog's nose, but he mentioned the finding of intracellular end-knobs ('Endknopfen') in the skin of the nose of the bat. He held that the free nerve terminations in the epidermis, not only of bats, but of all classes of vertebrates, are intracellular, though none of his figures indicates it. Retzius ('92) showed free nerve terminations (intercellular) in the epidermis of the lip of the human foetus. Of free nerve endings in the epidermis of the mouse and of the rabbit he made the following statement: "Die feinen, varicosen Nervenfasern verzweigen sich und endigen im Rete Malpighii interzellular ohne jeden directen Zusammenhang mit Zellen." Van Gehuchten, in 1893, described free nerve terminations in the epidermis of the face, lip, ear, paw and tail of the white mouse and white rat. He likewise found the free nerve endings to be intercellular. He stated, "Partout nous avons trouve I'existence de fibres nerveuses intra-epidermiques se ramifiant et se terminant librement entre les cellules epitheliales." While Dogiel ('03) did not hold intracellular endings as out of the question, yet he was strongly of the opinion that the free nerve terminations are intercellular.

2. Nerve endings on hairs

While the innervation of hairs has for some time been a field of fruitful investigation, there still remain some unsolved problems in connection with the hair of Chiroptera. In recent years especially, more attention has been directed toward the innervation of tactile or sinus hairs than toward that of the hairs of the pelage. The writer's descriptions will be confined wholly to the latter.


316 JAMES EDWARD ACKERT

Observations made by different investigators on the innervation of the hairs of bats have been so conflicting that it seems advisable to give a brief review of the literature. Schobl (71) studied the innervation of the hair of the flying membrane, and set forth the following principal points: In the hairs of the bat, the nerves terminate in special corpuscles ('Terminalkorperchen') situated at the bases of the hair follicles. The hair receives a bundle of nerves which consists of from two to five medullated fibers. These twist many times in a spiral about the hair shaft forming a nerve wreath or ring. From this spiral ring two to four nerve fibers are given off, and these extend downward, ending in the terminal corpuscle beneath the hair follicle. A superficial nerve ring which consists of from one to two coils is formed by fibers from the fourth nerve layer. ^ Boll ('71), working on similar material, confirmed Schobl's observations.

The following year Stieda took exception to Schobl's account, especially in regard to his 'Terminalkorperchen.' This observer concluded that the structure in question was not a nervous apparatus but rather a differentiated part of the hair folhcle ('Haarkeime'). The nerve ring was not mentioned by Stieda. Beil ('71) also denied the existence of Schdbl's end-corpuscles, although he was able to see the nerve ring. Concerning the structure of the latter, its course, or the endings of its fibers, he could determine nothing definitely. Above the sebaceous glands, however, Beil noted the entrance t>f two or three bundles of nonmedullated nerves into the hair follicle.

Using the method described by Schobl himself, Veleeky ('72), investigating the flying membrane, likewise did not find the socalled end-apparatus; nor did the use of gold disclose these 'Terminalkorperchen.' By the latter method, however, he demonstrated non-medullated nerves which approach, from below, the cells of the epithelium of the outer root sheath of the hair, and spread into the intercellular spaces, forming a net.

In the same year appeared a more important piece of work by Jobert, in which he described in considerable detail the innervation of the hair of the bat's wing. The principal points brought out were as follows: (1) All the hairs of the skin are supplied


INNERVATION OF INTEGUMENT OF CHIROPTERA 317

with nerves and are perceptive. (2) The affirmation of Schobl that into each hair folhcle there enters one nerve is not true. Each hair is supphed with many nerve fibers, five to six or more, which approach the folhcle, together or separately, and from different sides. They may unite into two or three small trunks. On reaching the neck of the hair the nerves divide, lose their medullation, and are distributed on the hyaline membrane more or less like radiations, ending freely at about the same level. (3) The nerve ring of Schobl does not exist, neither do the 'Terminalkorperchen.' (4) At the level of the superficial subepithelial network of nerves, minute threads are seen which surround the follicle and disappear in the epithelial sheath.

Arnstein ('76) recognized two different kinds of nerve terminations on hairs: (1) The free endings on the hyaline membrane in the form of a 'palisade;' (2) The nerve network which occurs in the outer root sheath.

Bonnet ('78), who investigated the innervation of the hair follicles of a number of mammals including the bat, confirmed Arnstein's observations on the endings of nerves on the hyaline membrane. Bonnet's idea was that a nerve ring exists in connection with each hair. The small fibers which constitute this structure lie outside of the straight fibers, which terminate in a 'palisade,' and surround them much as hoops surround a barrel, in the form of a ring consisting of six or more pale fibers. Of the root sheaths in the region of the sebaceous glands Bonnet says, "This is a rendezvous of the various small medullated nerve fibers which come to the hair partly above and partly below the sebaceous glands. These fibers going to the follicle spread out forming a woven net of minute medullated fibers."

In describing the innervation of the hair of certain mammals, including bats, Szymonowicz ('01) pointed out that the medullated fibers approach the follicle below the sebaceous glands, divide, losing their medullation, and penetrate to the hyaline membrane, where some of the fibrils encircle the hair, while others end on the hyaline membrane. The latter fibrils branch regularly, and run parallel with the long axis of the hair. This investigator observed perceptive menisci in a strongly developed

JOURNAL OF MORPHOLOGY, VOL. 25, NO. 2


318 JAMES EDWARD ACKERT

outer root sheath of a common hair on the face of Vesperugo serotinus.

According to the observations of Sabussow ('10) the hair of the flying membrane of bats is supphed with several meduUated nerve fibers whose number is never less than two. These fibers approach the hair follicle, divide, spread around the hair spirally, more or less in the form of a ring, give off small fibrils from the latter, branch, and finally end in the form of a 'pahsade' on the hyaline membrane. The fibrils of the 'palisade' may contain varicosities along their courses, or their distal ends may be lanceshaped. The spiral ring around the 'palisade' consists of small varicose threads. This Sabussow holds as a second kind of nerve ending on the hair. He asserts that he never saw these two kinds of endings, namely, the 'palisade' and the varicose threads of the spiral ring, at the same time in the same hair. From this he concludes that there exist two kinds of hairs, each of which is supplied with one of these nerve terminations.

Parallel with the spiral ring just described and more superficial, Sabussow observed a broadly meshed network of fibers resembling a nerve ring, and apparently surrounding the hair above the sebaceous glands. This network or ring, which belongs to the second nerve layer, could by focusing be seen to give off more or less flattened fibers resembling the 'palisade.' Being unable to find any definite connection between the 'palisade' apparatus first described and this one which comes from the subepithelial network, Sabussow inferred that the two were independent.

In the writer's deeply stained methylene blue preparations of the bat's skin, both of the body and of the membranes, the hair follicles with their numerous nerves stand out in bold contrast to the surrounding, weakly stained connective tissue (figs. 8, 9). The nerves which supply the hairs arise from the second nerve layer, pass outward to approximately the level of the inner third of the hair follicle, where, at first, they appear to pass along from one hair to another. But upon close examination it is seen that nerves may be distributed in one of two ways: (a) The whole fiber may end directly in a single follicle (figs. 8, 9, ff) ; (b) Upon


INNERVATION OF INTEGUMENT OF CHIROPTERA 319

approaching hairs the nerve may divide, one or two branches going to a follicle, the others passing out to the epidermis (figs. 8, 9, fe). By far the greatest number of the nerves in question are distributed in the first way. The numerous fibers form a veritable network, which might justly be termed a nerve layer, but which for simplicity is not so considered by the present writer.

As to nerve endings on the hair, it may be said that they occur at three different levels and in three separate layers of the follicle : (1) A superficial nerve ring situated above the orifices of the sebaceous glands and giving off nerve threads in the connective tissue sheath (fig. 9, sn); (2) Fine varicose or flattened nerve fibrils which lie immediately below the sebaceous glands, and end on the hyaline membrane parallel to the long axis of the hair (fig. 9, eh) ; (3) Nerve fibrils at the level of the lower third of the follicle, which take a horizontal position in the outer root sheath (fig. 9, eo). A further consideration of these types of nerve endings follows.

1. Superficial nerve ring. Medullated nerve fibers approach the hair above the opening of the sebaceous glands. At the outer border of the connective sheath, they divide, spreading around the follicle and forming a loose ring of from two to six or more fibers. From the ring are given off non-medullated fibrils, some of which are interwoven into a delicate network, while others appear to end freely in the connective tissue sheath of the follicle. This ring doubtless corresponds to the "broadly meshed network resembling a ring" described by Sabussow ('10) above the sebaceous glands. As is seen in figure 9, /, the non-medullated fibrils show no tendency to pass downward to a nerve ring below.

2. Varicose or flattened nerve fibrils. Immediately below the sebaceous glands medullated nerve fibers, chiefly of type (a), enter the region of the hair follicle, penetrate the connective tissue layers, divide, losing their myelin, and encircle the hair in a nerve ring. The number of fibers constituting the nerve ring varies from two to eight or even more. From the inside of the ring fibrils are given oft" which divide dichotomously. The branched fibrils assume a position parallel to the long axis of the


320 JAMES EDWARD ACKERT

hair, and usually end in slight enlargements (fig. 9, eh), some of which are merely small varicosities, while others resemble the minute end-knobs seen in the free nerve terminations in the epidermis. In certain cases there are no enlargements, but in these instances the terminal fibers are flattened. This type of nerve ending undoubtedly corresponds to the well known nerve ring and 'palisade' described first by Arnstein ('76), and recognized since by Bonnet ('78), Szymonowicz ('01), and Sabussow ('10). Merkel ('80) described a similar end-apparatus on a common hair in the lip of a cat. The "termaisons en fourchette" of the Hoggans ('83) and the nerve rings of Retzius ('94), Van Gehuchten ('96) and Ostroumow ('00) are probably corresponding nervous structures.

3. Nerve fibrils in the outer root sheath. At the level of approximately the lower third of the root of the hair, medullated nerves penetrate the connective tissue layers of the follicle. When the hyaline membrane is reached they divide and run for a short distance on or near its surface. These nerve fibers give off a few strong non-medullated fibrils which pierce the glassy membrane, and end in the outer root sheath, usually taking a horizontal position in the latter (fig. 9, eo). The nerve endings of this type are found in a slight swelling of the root sheath, which may correspond to the superior swelling described in typical sinus hairs. So far as the writer has been able to ascertain, nerve endings of this type have not previously been described in the pelage hair of the bat. While such examples are not numerous, yet they seem to him to be genuine. Nerve endings in the form of tactile corpuscles were described by Szymonowicz ('01) in the outer root sheath of a hair in the face of Vesperugo serotinus. The same observer, in 1909, mentioned the finding of Merkel's corpuscles in this layer of the follicle in man. Retzius ('94) described a nerve fiber in the outer root sheath of a hair of a mouse, and Vincent ('13) found nerve fibrils in this layer of the sinus hair of the rat.


INNERVATION OF INTEGUMENT OF CHIROPTERA 321

3. Special sensory end-organs

The literature dealing with special sensory end-organs in general has recently been reviewed by a number of investigators : Szymonowicz ('95), Trejakoff ('02), Dogiel ('03), Schafer ('10). Therefore only a brief survey of the observations upon such endorgans in the skin of bats will be given here.

Arnstein ('76) found in the flying membrane of a bat an endbulb which he thought resembled the well known cylindrical endbulb of Krause. It was possible for him to trace an axis cylinder into the organ, but he was unable to make out the ending of the fiber. In one instance, however, he saw it break up into fine fibrils. Arnstein was of the opinion that these end-bulbs occurred in the flying membrane where no hairs were present.

Schumacher, in 1907, mentioned the presence of a large number of layer-like corpuscles ('Lamellenkorperchen') among the phalanges.

Sabussow ('10) investigated the flying membrane of two species of bats (Vesperugo noctula and Vespertiho daubentonii) . He stated that in weakly stained preparations he could see terminal bulbs which were divisible into two classes according to size. Some were so small that upon shght magnification it was difficult to see them; others were comparatively large, had a zigzag course, and could be recognized with ease. The latter could also be seen in material prepared by Apathy's after-gilding gold chloride method. The general characteristics peculiar to all end-bulbs which Sabussow observed were: (1) a longitudinal course of the fibers of the enveloping connective tissue membranes, apparent in gold preparations, and (2) a delicate wavy appearance of these membranes seen in methylene blue material. The connective tissue nuclei did not stain in methylene blue preparations, but on account of the difference of refraction, Sabussow thought that by focusing he could see them. He noted that the core of the bulb was narrow, but was unable to make out its finer structure.

The consideration that led Sabussow to classify these structures under the cylindrical t^-pe of end-bulbs was chiefly the


322 JAMES EDWARD ACKERT

way in which the nerve fiber ended in the interior of .the bulb. He observed that medullated nerve fibers divided at Ranvier's nodes, giving off several medullated branches. Occasionally, one of these branches entered an end-bulb, and passed through the whole interior of the organ to its opposite extremity. This naked axis cylinder in the bulb became slightly expanded, and ended either with a sharp point or in a thickening resembling a button.

Sabussow, in his figure 10, pictured an end-bulb stained with methylene blue, which he called a variation of the cylindrical end-bulb type. He described it as follows:

Within the bulb the axis cylinder expands, and, in the middle, broadens into a wide, paw-like plate with deep notches in its edges. From this plate there is given off a fiber which bends backward and upward, and in turn widens into a similar paw-like plate. The substance of the plate has a granular appearance, with here and there small masses of stain.

In my own methylene blue preparations of the integument two kinds of special sensory end-organs have been observed. (1) A small elongate end-bulb into which a single medullated nerve fiber enters, passes approximately to the opposite end, and terminates in a sHght enlargement (fig. 18) ; (2) A large, round, cellular corpuscle innervated by a single fiber which disappears among the cells of the organ (figs. 19, 20). A more detailed description of each type follows.

a. End-bulbs. These structures occur in the corium near hair follicles, but clearly outside of the root sheaths. Ordinarily they are found below the level of the sebaceous glands, parallel with the long axis of the hair. Their size is approximately 1.5m in length by 0.5m in width. In general appearance they are regularly club-shaped in outline, the interior being filled with a semi-fluid substance. The medullated nerve, on entering the bulb, loses its myelin, the sheath of which becomes continuous with the sheath of the end-bulb. After passing through half the length of the organ, the axis cylinder expands slightly into a flat plate (fig. 18, p) which gives off two or three short heavy branches, and terminates near the end of the bulb in a small enlargement


INNERVATION OF INTEGUMENT OF CHIROPTERA 323

(fig. 18, en). The deeply stained blue plate stands out in bold contrast to the weakly colored bulb about it. The distal branch arising from the thickened axis cylinder usually bends to one side, breaks up into an irregular, elongate, granular mass, and as such extends back through the expanded part of the bulb (fig. 18, hr). In the portions of the organ surrounding the plate and the recurrent granular mass no layers nor nuclei are visible. With the exception of the lack of nuclei the appearance of this end-bulb is practically identical with that of the structure which Sabussow ('10) showed in his figure 10, and which he termed a modified end-bulb of Krause. The absence in my preparations of the nucleated capsule characteristic of the cylindrical end-bulb of Krause can be explained by the fact that such structures do not ordinarily stain in methylene blue. Although these organs are somewhat smaller than the cylindrical end-bulbs in question, their location and structure are such that the writer is inclined to think that they are modified cylindrical end-bulbs of Krause. As has been shown, the writer confirms one type of Sabussow's end-bulbs, the first, which closely resembles Krause 's cylindrical end-bulb (fig. 18, Sabussow's fig. 10). But the Russian author's second type, namely, the large one containing a nucleated sheath,, and having a zigzag course, has not yet been seen in the preparations used for this study. However, there are present in this material structures (portions of medullated nerve trunks) which correspond so closely to the descriptions and pictures of his second type of end-bulb (Sabussow's plate 1, figs. 5, 6, 7, 8, 9) that it appears very probable that the two are identical. In his figure 8, he shows 'end-bulbs,' which, to the writer, seem clear examples of cut medullated fibers separating out from a common trunk. Sabussow himself points out that medullated fibers in this region branch repeatedly at the nodes of Ranvier. His figure 7, a picture of two so-called end-bulbs, represents apparently a portion of a medullated nerve fiber, which at one of these nodes, divides into branches. In his descriptions he states that the axis cylinder passes through the whole bulb ta its very end. This is precisely what is found in a portion of a medullated nerve (fig. 21) when cut obliquely. The 'ending*


324 JAMES EDWARD ACKERT

of the axis cylinder appears pointed or hooked, according to the position of the nerve when cut. The connective tissue nuclei, which did not stain in methylene blue, but which Sabussow thought he could make out by focusing, are in the opinion of the writer nuclei of the sheaths of Schwann or of the perineurium.

In the present investigation no gold chloride method was used, but material prepared by the Cajal silver nitrate method, Bielschowsky's method, or with methylene blue counter-stained with carm.alum, has failed so far to reveal the presence of this second type of 'end-bulb' as described by Sabussow.

b. Terminal corpuscles. These comparatively large spheroidal corpuscles are found in the stratum reticulare of the corium, usually at some distance beneath the lower level of the hair folhcles. In methylene blue preparations they stain deeply, exhibiting a cellular appearance, and frequently showing one or more nuclei (figs. 19, 20). Their size, which is fairly constant, is approximately 20m in length by lO/x in width. Each corpuscle is innervated by a medullated fiber which arises from the second nerve layer. The fiber passes into the deeper part of the corium, and after giving off a few branches, enters the corpuscle, where it disappears among the cells. Occasionally, the fiber, on approaching the organ in question, forms a shght spiral coil (fig. 19). Thus far it has been impossible to trace the course of the nerve fiber within the corpuscle. To estabhsh the identity of these organs with any of the known types of corpuscles is difficult. The layer-like capsules characteristic of Pacinian corpuscles are not apparent, but their absence may be due to incomplete staining. They more nearly approach the size of the type in question than that of any other type of corpuscles commonly found in the mammalian skin. Their location is identical with that of Pacinian corpuscles. From the fact set forth it seems possible that these spheroidal, cellular bodies may be Pacinian corpuscles.

To these two types of sensory end-organs may be added terminal varicosities, which are abundant in the region of hair follicles, outside of the root sheaths. Comparatively strong nerve fibers can be seen to enter these structures, where they break up into fine fibrils, and are surrounded by neuroplasm. It


INNERVATION OF INTEGUMENT OF CHIROPTERA 325

is possible that these organs are varicosities in which the fibers beyond the enlargements fail to stain. But they are found constantly in deeply colored preparations, and moreover, are somewhat greater in size than those ordinarily occurring in the course of a fiber. Arnstein ('76) described varicosities in the outer root sheath of sinus hairs in bats, and Sabussow mentioned the presence of these structures on the courses of nerves outside of the root sheaths, but it appears that terminal varicosities external to the hair follicles have not previously been observed in Chiroptera.

4. Motor nerve endings on striated muscles

Voluntary muscles in the integument of the face, especially in the upper lips of bats, are well developed. In sections stained intra vitam with methylene blue these muscles are ordinarily deeply colored, the cross striations being of a slightly darker hue. Under such conditions it is usually possible to make out only the muscle fibers and their nuclei; but in regions somewhat removed from the larger blood vessels, where the blue stain is weak, one can see medullated nerve trunks among the muscle bundles. Along the muscle fibers which are stained only sufficiently to see their outline, it is possible to trace medullated nerve fibers (fig. 11, nv) which give off a small number of branches. The latter in terminating, form motor end-plates (figs. 11, 12, mep) on the muscle fibers.

It is not the purpose of the writer to enter into a discussion of the literature on this important subject, but merely to describe his observations and to mention wherein they agree or disagree with those of a few recent workers. For a review of the literature on motor nerve endings see Boeke ('09), Dogiel ('06), Huber and DeWitt ('97).

As a rule, the medullated nerve fibers can be traced to the border of the muscle fiber. x4.t or near the edge of the latter, the medullation stops and the nerve fibers soon begin to separate into their component fibrillae, and finally end in a more or less regular end-net or arborization (figs. 11, 12, ea). This is in accord


326 JAMES EDWARD ACKERT

with the observations of a number of investigators, including Boeke, who described motor end-plates on muscles in the tongue of the bat.

At the point where the nerve fibers enter the enlarged motor end-plate there is a slight elevation of the surface of the muscle (fig. 12, el). The position of end-plates on muscle fibers has been in doubt for some time. In the preparations used in this study they appear to be beneath the sarcolenmia (fig. 11, so). This shows especially well in cross sections of muscle fibers in the tongue (fig. 12, so). Most investigators are now in accord in regarding this structure as under the sarcolemma.

In weakly stained preparations the branched endings can be seen to lie in more or less irregularly shaped matrices. The latter are of two kinds : (1) A. weakly stained area containing numerous deeply colored granules of various sizes; (2) A somewhat smaller area without granules.

In shape, the former are irregularly circular or even triangular. The granules, which vary greatl}^ in size, stain almost if not quite as deeply as the nerve fibrils themselves (fig. 11, mag). To structures corresponding to these Kiihne ('87) gave the name soles ('Sohlen'). The smaller areas or soles, which appear to be free from granules in this stain, are oval or pear-shaped, the axis cylinders always entering the narrowed end. Huber and DeWitt ('97), Dogiel ('90) and Retzius ('92) stated that the sole does not stain in methylene blue preparations, whether examined at once or fixed in aimnonium molybdate and studied in sections. The material, from which the present observations were made, was prepared according to the latter method. The irregularly shaped matrix in which the axis cylinder terminates is typically granular, but the nuclei seen by Huber and others in such preparations counter stained with carmalum or picro-carmine do not stain in methylene blue. In each of the two matrices or soles described above, the end-arborizations are nearly similar. The fact that Huber and the other observers mentioned failed to see soles whose granules stained in methylene blue preparations, may have been due to the inconsistency of the stain. This possibility, together with the fact that nerve terminations on motor plates


INNERVATION OF INTEGUMENT OF CHIROPTERA 327

are very similar in each kind of sole described, and that the size of the so-called second kind comes within the possible range of that of the first, leads the writer to think that perhaps the apparently different kinds of soles found by him in the striated muscles in the epidermis of the bat are in reality one and the same type. In the one, the granules have stained, in the other they have not.

5. Nerve endings on modified sweat glands

So far as the writer has been able to ascertain, the literature contains no reference to the innervation of sweat glands in bats. As was noted elsewhere (p. 305), the modified sweat glands (fig. 10) have a coating of smooth muscle fibers which are arranged longitudinally (fig. 10, mu) . In a weakly stained methylene blue preparation from the interfemoral membrane of Myotis lucifugus. such sweat glands have been observed with numerous stripes running at right angles to the smooth muscle fibers (fig. 10, fi). These stripes, which occur at comparatively regular intervals, extend hoop-like around the secretory portion of the gland external to the muscle fibers. The structures in question are much smaller than the muscle fibers, have a wavy course, and take the deep blue stain characteristic of nerve fibrils. A number (two to five) of delicate non-medullated nerve fibers (fig. 10, no) can be traced to the sides of these glands, but whether they connect with these circular stripes, the writer is at present unable to ascertain.

That sweat glands are under the control of the sympathetic nervous system is generally recognized. As is well known, preganglionic neurites leave the spinal cord through the ventral roots of the spinal nerves, and, after a shorter or longer course, terminate in some sympathetic ganglion in a very characteristic manner. Here the preganglionic neurites branch repeatedly, dividing into numerous small varicose nerve fibers, which interlace to form intracapsular plexuses around the cell bodies of the sympathetic neurones. It is likewise well known that in the sympathetic ganglia of Mammalia such intracapsular pericellular plexuses may be very simple, consisting of only a few varicose fibrils, as well as


328 JAMES EDWARD ACKERT

very complex. The general structure of these pericellular plexuses, the absence of definite observations upon endings of sympathetic neurites (post-ganglionic) on sweat glands, and the striking hoop-like arrangement of these fibrillar stripes around the glands, lead the writer to question whether, perhaps, the postgangUonic neurites may not form simple plexuses about the glands more or less similar to the pericellular plexuses about the cell bodies of the sympathetic neurones. Such an arrangement of the terminal fibers of a post-ganglionic neurite would be most effective. The nerve threads lying immediately upon the smooth muscle fibers and the bases of the gland cells could form functional connections with them. This view of the endings of post-ganglionic neurites on modified sweat glands seems somewhat more plausible from the fact that these circular fibrillar structures appear about the glands only in their secretory portions, and likewise only in the regions covered by the longitudinal smooth muscle fibers.

WHAT SENSORY ORGANS ARE CONCERNED WHEN BLINDED BATS AVOID OBSTACLES WHILE ON THE WING?

Although the present problem is primarily a morphological one, yet a discussion of the function of the integumentary sensory structures described may not be out of place here. Of especial interest is the question of the means by which blinded bats avoid obstacles while on the wing. A number of investigators, experimenting with living bats, have maintained that the organ concerned in these delicate reactions is the skin. A brief summary of their opinions follows.

As already noted, Cuvier thought the fljdng bat, on approaching an object, sets up air currents, which react on the patagium, enabling the animal to avoid the obstacle.

Jobert ( 72) observed that on pinching the skin, the animals responded faintly as compared with their vigorous reactions when hairs were pulled out. This led him to think that the sensitiveness of the flying membrane is due to the hairs. He inferred that currents of air affect the hairs and that each movement of the lat


INNER^ ATION OF INTEGUMENT OF CHIROPTERA 329

ter is transferred to the nerve ring in such a way that objects are perceived and avoided.

In 1873, Redtel liberated a blinded bat in a room in which had been placed numerous threads. The animal avoided the threads successfully. From the abundance of nerves found by Schobl in the flying membrane, and from his own experiments, Redtel inferred that it was possible for bats to perceive the slightest change of air pressure upon the wings.

Sabussow ('10), like Jobert, was of the opinion that air currents were set up between the object and the approaching blinded animal, and that by means of these currents the nerves of the hairs were affected.

A few investigators have held that the organs for the perception of the delicate stimuli, which bring about these avoiding reactions, are located in the internal ear. ♦

Jurin ('98), experimenting upon living bats, observed that when their organs of hearing were destroyed, they were unable to avoid obstacles placed in their way. A mutilation so severe as this, however, would certainly produce shock effects which might affect very considerably the results of the experiments.

Hahn ( '08) caused a large number of mutilated bats (blinded, ears cropped, etc.) to make a given number of flights in an enclosure through which numerous wires had been stretched. He agreed with Jurin that objects are perceived by the flying animals chiefly through sense organs located in the internal ear. The evidence for this was obtained by closing the external auditory meatus with plaster of Paris, whereupon he found that the percentage of 'hits' (collisions with wires) was much higher for this experiment than for that of any of the others. The fact should not be overlooked that the placing of a hard substance like the one used in the meatus, and possibly against the sensitive tympanic membrane, is likely to interfere seriously with the normal functioning of the nervous system.

It is thus seen that the weight of evidence favors the view that condensations (pressures) of the atmosphere set up between the obstacle and the blinded bat stimulate sensory structures in the integument.


330 JAMES EDWARD ACKERT

The question of what integumentary organs are concerned naturally arises. Organs capable of being stimulated by such condensations would have to meet certain requirements: (1) They must be distributed over the head and flying membranes at least, as' these parts are foremost in flight; (2) They must be superficially located, for stimulations from air condensations are doubtless very slight.

While no special nerve structures which appear to form the sole basis for the perception of air pressures have been observed by me, yet the presence of large numbers of free nerve terminations (end-knobs) near the surface of the epidermis seems significant. These structures comply with both of the requirements set forth. They are widely distributed over the body and membranes, and their superficial position among the outermost cells of the stratum Malpighii makes them especially well placed for the reception of light touch stimuli. Their number in the epidermis is enormous.

The superficial nerve rings (and their terminal fibers), though not located as near the surface of the integument as the nerve endknobs, are so situated about the necks of the follicles as to be affected by even the slightest movements of the hairs. These nervous structures are also widely distributed over the skin and their position is somewhat superficial. Von Frey ('96), in his researches on the sense of pressure in man, has shown that pressure nerve fibrils terminate in a ring surrounding the hair follicle, this form of termination serving as an end-organ. This writer states that on" account of the position of the ring, the fibrils are stimulated by any pressure exerted upon the hair. The other nerve endings on hairs of bats are farther from the surface, so that movements of the hair sufficient to stimulate them would probably have to be more pronounced than those produced by condensations of the atmosphere.

An examination of the anatomical evidence thus indicates that two types of sensory end-organs in the skin of Chiroptera meet the requirements mentioned for the perception of air pressures. These are the free nerve terminations, and the superficial nerve rings of hairs.


INNERVATION OF INTEGUMENT OF CHIROPTERA 331

End-bulbs and terminal corpuscles no doubt are tactile in function, but their depth below the surface of the epidermis precludes any probability that they aid in the perception of condensation pressures of the air.

In connection with the question as to which of the two sensory endings mentioned above functions to the greater extent in the perception of atmospheric pressures, it should be pointed out that the area of the integument supplied by superficial nerve rings is insignificant in comparison with the area supplied with nerve end-knobs. Likewise, the number of terminal fibers of the rings is not to be compared with the enormous number of end-knobs in the epidermis.

As is well known, the human cornea is very sensitive to delicate tactile stimuli. Cohnheim ( '67) has shown that the only type of perceptor to be found in the cornea is that of free nerve terminations.

Goldscheider, in 1886, determined by experimentation the location of tactile spots on his arm, and then removed for study pieces of skin containing them. Here, also, the only sensory structures revealed by a histological examination were free nerve terminations.

The evidence thus leads to the conclusion that free nerve terminations are more important in the tactile reaction than are the superficial nerve rings of hairs.

It is not, of course, to be inferred that all the free end-knobs function alone as pressure perceptors, for, as is well known, the sensory nerves of the human skin mediate at least four different qualities of sensations, namely, pressure, warmth, cold and pain. But the number of nerve end-knobs in the skin is so great, and the latter in the bat is so sensitive to delicate tactile stimuli, that the number of free nerve terminations in the epidermis functioning as pressure perceptors must necessarily be very large.

To sum up, then : The writer is i'nclined to think that the most reasonable explanation of the avoidance of obstacles by blinded bats involves the assumption that condensations of the atmosphere are set up between the object and the approaching bat, and


332 JAMES EDWARD ACKERT

that these condensations are perceived by the blinded animals chiefly by means of the free nerve end-knobs in the epidermis, but also in part by the superficial nerve rings of the hair follicles. As the flight of these animals, when close to objects, is tolerably slow, it seems probable that such sensory impulses could be transmitted to the central nervous system, and motor ones be carried back to the muscles of the wings in time for the bats to avoid obstructions in their way.

SUMMARY GENERAL STRUCTURE OF THE INTEGUMENT

1. The integument of Chiroptera has a general covering of hair, although the soles of the feet, the mammae, the external genitalia, and the distal parts of the ears and of the flying and interfemoral membranes are almost naked.

2. The skin consists of epidermis and corium. The epidermis is made up of a well developed stratum corneum (whose deepest layers, the stratum lucidum, can be seen distinctly only in the palmer and plantar regions) and of a Malpighian stratum. In the integument of the body the Malpighian stratum contains the three layers commonly found in the mammalian skin, while this stratum in the membranes consists at most of but two layers, and frequently of but one. The corium is composed of an external stratum papillare, containing both simple and compound papillae, and of an internal stratum reticulare.

3. Pigment granules are abundant in the Malpighian stratum, while in the stratum corneum they are much less numerous. In the flying and interfemoral membranes more pigment is present in the dorsal than in the ventral duplicature of the epidermis. Isolated pigment cells are of frequent occurrence throughout the corium.

NERVE LAYERS OF THE INTEGUMENT

1. In the integument and subcutis of the body three layers of nerves are found. The first (most internal) layer consists of medullated trunks in the subcutaneous tissue. By dichotomous branching these nerves break up into a loosely intertwined meshwork, consisting of an enormous number of medullated nerves,


INNERVATION OF INTEGUMENT OF CHIROPTERA 333

and foniiiiig the second nerve layer. Arising from the latter are medullated fibers which pass to the stratum JVIalpighii. Here they divide, forming a simple, non-medullated network, which constitutes the third nerve layer.

2. Certain regions of the flying and interfemoral membranes have three layers of nerves, others but two. These are (1) A layer of medullated nerve trunks with numerous medullated branches, occurring in the stratum reticulare, but only in the elongated ridges containing the largest blood vessels and much connective tissue; (2) A double, medullated nerve layer in the deeper part of the corium extending throughout the membranes; (3) a layer, likewise double, present in the entire Malpighian stratum, and consisting of a large number of branching nonmedullated nerve fibrils.

3. Numerous varicosities are found in the corium on branches from the second nerve layer.

NERVE ENDINGS IN THE INTEGUMENT

1. Free nerve terminations occur in the uMalpighian layer. Small medullated fibers from the third nerve layer can be traced out among the deeper Malpighian cells to the stratum granulosum, where they terminate in minute end-knobs, probably intercellularly.

2. Nerve fibers supplying the hair follicles may be distributed in two ways: (a) The whole fiber may end directly in a single follicle; (b) On approaching hairs a fiber may divide, one or two branches going to a follicle and the others passing out to the epidermis.

3. Nerves end on pelage hairs at three levels and in three different sheaths of the follicles. These endings are: (1) A superficial nerve ring situated above the orifices of the sebaceous glands, and giving off nerve threads in the connective tissue sheath; (2) fine, varicose or flattened nerve fibrils which lie immediately below the sebaceous glands, and terminate on the hyaline membrane parallel to the long axis of the hair; (3) Nerve fibrils at the level of the lower third of the follicle, which usually take a horizontal position in the outer root sheath.

JOUKNAL OF MORPHOLOGY, VOL. 25, NO. 2


334 JAMES EDWARD ACKERT

4. Two types of special sensoiy end-organs are found in the skin : (1) A small elongate end-bulb into which a single meduUated nerve fiber enters, passes approximately to the opposite end, and terminates in a slight enlargement; (2) A large, round, cellular terminal corpuscle innervated by a single fiber whose branches disappear among the cells of the organ.

5. Terminal varicosities are abundant in the region of the hairs outside of the sheaths of the follicles.

6. In the skin of the face, especially, striated muscles are well developed. Motor end-plates occur on these muscles. In the integument the end-plates appear to be beneath the sarcolemma, and in the muscles of the tongue these plates are clearly below the sarcolemma.

7. Small fibers resembling sympathetic post-ganglionicneurites extend hoop-like around the large modified sweat glands external to the longitudinally arranged smooth muscle fibers.

8. Blinded bats, when on the wing, probably perceive obstacles through the sense of touch by the effect of condensations of the atmosphere (produced on approaching the object) upon the free nerve terminations in the epidermis and the superficial nerve rings of the hair follicles.

BIBLIOGRAPHY

Allen, H. 1893 A monograph of the bats of North America. Bull. U. S. Nat.

Mus., no. 43. Arnstein, C. 1876 Die Nervender behaarten Haut. Sitzungsber. d. k. Akadem.

d. Wissensch., Abt. Heft Separat, pp. 203-232. Beil 1871 Ueber die Nervenendigungen in den Haarbiilgen einiger Tasthaare.

Inaugur. Dissert., Gott. Boeke, J. 1909 Die motorische Endplatte bei den hoheren Vertebraten, ihre

Entwickelung, Form und Zusammenhang mit der Muskelfaser. Anat.

Anz., Bd. 35, pp. 193-226. Boll, F. and Schobl, J. 1871 Die Flughaut der Fledermause namentlich

Endigung ihrer Nerven. Centralblatt f. d. med. Wissensch. Bonnet, R. 1878 Studien i'lber der Innervation der Haarbalge der Hausthiere.

Alorph. Jahrb., Bd. 4, pp. 329-398. BoTEZAT, E. 1908 Die Nerven der Epidermis. Anat. Anz., Bd. 33, j)p. 45-75. CoHMiKLM, J. 1867 Ueber die Endigung der sensiblen Nerven in der Horn haut. Archiv f. path. Anat., Berlin, Bd. 38. Diem, Fh. 1907 Beitrage zur Entwicklung der Schweissdriisen an der behaarten

Haut der Saugethiore. Anat. Hefte, Abt. 1, Bd. 34, pj). 189-236.


INNERVATION OF INTEGUMENT OF CHIROPTERA 335

DoBSON, G. 1873 On secondary sexual characters in the Chiroptera. Proc.

Zool. Soc. London, pp. 241-253. DoGiEL, A. 1890 Methylenblautinction der motorischen Nervenendigungen in

den Muskeln der Amphibion iind Reptilen. Archiv f. mikr-. Anat.,

Bd. 35.

1903 Ueber die Nervend-apparate in der Haul des Menschen. Zeitschr.

f. wissensch. Zool., Bd. 75, pp. 46-111.

1906 Die Endigungen der sensiblen Nerven in den Augenmuskeln

und deren Sehnen beim Menschen und den Siiugethieren. Archiv f.

mikr. Anat., Bd. 68. VON Frey, M. 1896 Untersuchungen liber die Sinnesfunctionen der mensch lichen Haut. Abhandlungen, k. siichsischen Gessellch. der Wissen schaften, Math.-Phys. Klasse, Bd. 23, pp. 175-265. GoLDscHEiDER, A. 1886 Histologische Untersuchungen fiber die Endigungs weise der Hautsinnesnerven beim Menschen. Archiv f. Physiol., pp.

191-228. Hahn, \V. 1908 Some habits and sensory adaptations of cave-inhabiting bats.

Biol. Bull., vol. 15, pp. 135-193. Hoffman 1898 Ueber Talg- und Schweissdriisen. Diss., Tubingen. HuBER, G. C., AND DeWitt, L. 1897 A contribution on the motor nerve endings in the muscle-spindles. Jour. Comp. Neur., vol. 7, pp. 169-230. JoBERT, C. 1872 Etudes d'anatomie compai'ee sur les organes du toucher.

Ann. d. Scienc. Natur., T. 16. KtJHNE, W. 1887 Neue Untersuchungen liber motorische Nervenendigungen.

Zeitschr. f. Biologic, N. F., Bd. 5. Leydig, Fr. 1859 Ueber die ausseren Bedeckungen der Silugethiere. Archiv

f. Anat. Physiol, und wissensch. Med., pp. 677-747. Merkel, F. 1880 Ueber die Endigungen der sensiblen Nerven in der Haut

der Wirbelthiere. Rostock. OsTROUMOW, P. 1900 Zur Nervenendigungen in den Haarbiilgen der Sauge thiere. Kasan. Porta, A. 1910 Sulle glandule facciale del Vesperugo noctula (Schreb). Zool.

Anz., Bd. 36, pp. 186-189. Redtel, a. 1873 Der Nasenaufsatz des Rhinolophus hippocrepis. Zeitschr.

f. wissensch. Zool., Bd. 23, pp. 254-288. Retzius, G. 1892 Ueber die sensiblen Nervenendigungen in den Epithelien

bei den Wirbelthieren. Biolog. Untersuch., N. F., Bd. 4.

1892 Zur Kenntnis der motorischen Nervenendigungen. Biolog.

Untersuch., N. F., Bd. 3.

1894 Ein Beitrage zur Kenntnis der intraepithelien Endigungsweise

der Nervenfasern. Biolog. Untersuch., N. F., Bd. 6. Sabussow, N. 1910 Zur Innervation der Flughaut von Chiropteren. Trans.

Soc. Nat. Univ. Kasan, Bd. 43, pp. 1-67. Schafer, E. 1910 The essentials of histology. New York, Lea and Febiger. Rchobl, J. 1871 Die Flughaut der Fledermiluse. namentlich die Endigung

ihrer Nerven. Archiv f. mikr. Anat., Bd. 7, pp. 1-31. Sc'HTTMArHER, S. 1907 Ueber das Glomus coccygeus des Menschen und die

Glomeruli caudales der Silugethiere. Anat. Anz., Ergiinzungsheft,

Bd. 30, pp. 172-178.


336 JAMES EDWARD ACKERT

Stieda, L. 1S72 Die aiigcblifheii Terminalkorjierchen an den Haareii einiger

Saugetbiere. Archiv f. mikr. Anat., Bd. S, pj). 274-278. SzYMUNOwicz, L. 1901 Lehrbuch der Histologic. Wiirzburg, A. Stuber's

Verlag. ToLDT, K. 1907. Ueber die Hautgebilde der Chiiopteren. Verb. k. k. zool. bot.

Ges., Wien, Bd. 57, pp. 83-91. Van Gehuchten, A. 1<S96 Les nerfs des poils. Mem. Cour. Acad. Belg., T.

34, pp. 24-38. Veleeky 1872 Die Endigungen der Nerven in den Haarbalgen der Saugetbiere.

St. Petersberg Soc. Nat., Bd. 3. Vincent, S. 1913 Tbe tactile bair of tbe wbite rat. Jour. Comp. Neur., vol.

23, pp. 1-38. WiMPFHEiMER, C. 1907 Zur Entwicklung der Scbweissdriisen der bebaarten

Haut. Anat. Heft, Abth. 1, Bd. 34, pp. 429-503.


EXPLANATION OF FIGURES

All figvires are from bat integument except figure 12, wbich is from bat's tongue. All, witb the exception of figures 2 and 3, were drawn witb tbe aid of the camera lucida. Figures 2 and 3 were drawn with the aid of the Spencer projection apparatus. The magnifications follow the descriptions of the figures.


PLATE 1

EXPLANATION OF FIGURES

1 Part of a transverse section of the skin of the face; py, pigment granules; mal, stratum Malpighii; .sc, stratum corneum. Fixed in corrosive-acetic, and stained in hematoxylin and eosin. X 375.

2 Portion of transverse section of integument at base of thumb, co, corium; mal, stratum Malpighii; pc, pigment cell; sc, stratum corneum; si, stratum lucidum. Fixed in corrosive-acetic, and stained in hematoxylin and eosin. X 250.

3 Transverse section of modified sweat gland from wing membrane; hm, basement membrane; cc, columnar cells; mj, smooth muscle fibers; mn, nucleus of smooth muscle fiber. Stained intra vitam with methylene blue, fixed in anunonium molybdatc, and counter-stained with Mayer's carmalum. X 750.

4 Part of a transverse section of skin of face; ep, epidermis; hj, hair follicle; s(jl, sebaceous gland; snl, second nerve layer; Inl, third nerve layer. Stained intra vitjiin witb methylene blue, and fixed in anuuoniniii inoiybdate. X 375.


INNERVATION OK INTEGUMENT OF CIIIHOPTERA

JAMES EDWAHD ACKERT


PLATli 1





-""^fiV^^"^




^^^'^wM^'^'^^^^'



PLATE 2


EXPLANATION OF FIGURES


6, 7 Portions of epidermis of interfemoral membrane; e, end-knob (free nerve termination); es, end-knob on surface of cell; n, nerve of third layer; sgr, cell of stratum granulosum; x, terminal fiber without end-knob. Stained intra vitam with methylene blue, and fixed in ammonium molybdate. X 575.

10 Portion of modified sweat gland from interfemoral membrane; fi, hooplike fibril resembling sympathetic nerve fibril; 7)iu, smooth muscle fiber; no, non-medullated nerve. Stained intra vitam with methylene blue, and fixed in ammonium molybdate. X 575.

11 Striated muscle from upper lip; ea, end arborization; mag, matrix of motor end-plate; mu, striated muscle fiber; nu, nucleus; nv, motor nerve fiber; sa, sarcolemma. Stained intra vitam with methylene blue, and fixed in ammonium molybdate. X 1500.

12 Transverse sections of striated muscles of tongue; ea, end arborization; el, elevation where axis cylinder pierces the sarcolemma; ma, matrix of motor end-plate without granules; mep, motor end-plate; .so, sarcolemma. Stained intra vitam with methylene blue, and fixed in ammonium molybdate. X 1500.


338


INNERVATION OF INTEGUMENT OF CHIROPTKUA

JAMES EDWARD ArKERT


PLATE 2




'Sdr



4



es



■ ff


'— fe



339


PLATE 3


EXPLANATION OF FIGURES


5 Pigment cells in the corium of skin of face; a, cell with granules unstained; b, cell containing both stained and unstained granules; c, cell with granules stained. Stained intra vitani with methylene blue, and fixed in ammonium molybdate. X 750.

8 Portion of transverse section of skin of back; ep, eindermis; fe, nerve fiber giving one branch to hair follicle and one to epidermis; .|f, whole fiber ending in hair follicle. Stained intra vitam with methylene blue, and fixed in ammonium molybdate. X 166.

9 Longitudinal section of hair follicle; drawn from a methylene blue preparation, but some features of the hair sheaths added from other preparations; cs, connective tissue sheath; eh, nerve endings on hyaline membrane; eo, nerve endings in outer root sheath; /, fibrils from superficial nerve ring; fe, nerve fiber giving one branch to hair follicle and one to epidermis ; ff, whole fiber ending in single follicle; h, hyaline membrane; hs, hair shaft; o, outer root sheath; sgl, sebaceous gland; s)i. superficial nerve ring. Stained intra vitam with methylene blue, and fi.xetl in ammonium molybdate. Other features from preparations fixed in Zenker's fluid, and stained in hematoxylin and eosin. X 750.

13 Portion of transverse section of wing membrane showing varicosity on fiber terminating in epidermis; mal, stratum Malpighii; sc, stratum corneum; ra, varicosity. Stained intra vitam with methylene blue, and fixed in ammonium inolvbdate. X 1500.


340


INNERVATION OF INTEGUMENT OF CHIROPTERA

JAMES EDWARD ACKERT


PLATE 3


sa

I



--\


mu


K^-no


mgyO


\ / e/


ea



ma^



341


PLATE 4


EXPLANATION OF FIGURES


14, 15, 16, 17 Varicosities on nerve fibers in skin of face. Stained intra vitam with methylene blue, and fixed in ammonium molybdate. X 1500.

18 End-bulb from upper lip; br, distal branch of axis cylinder; en, enlarged ending of axis cylinder; p, plate (expanded axis cylinder). Stained intra vitam with methylene blue, and fixed in ammonium molybdate. X 1500.

19, 20 Terminal corpuscles in skin of back; nu, nucleus. Stained intra vitam with methylene blue, and fixed in ammonium molybdate. X 750.

21 Section of nerve trunk from interfermoral membrane ; ax, axis cylinder Stained intra vitam with methylene blue, and fixed in ammonium molybdate. X 166.


342


INNERVATION OF INTEGUMENT OF CHIROPTERA

JAMES EDWARD ACKERT


PLATE 4



eo



nu


20


343


THE PRONEPHRIC DUCT IN ELASMOBRANCHS

GEORGE A. BATES

From the Tufts College Medical School

SIXTY-ONE FIGURES (fIVE PLATES)

HISTORICAL RESUME

The development of the pronephric duct in elasmobranchs has been studied by a number of observers, with very different conclusions as to its origin. These differences have arisen, in part, from the difficulty of demonstrating cell boundaries and limiting membranes in embryos preserved by the usual methods and, in part, from the difficulty of interpreting oblique sections. The theory that the vertebrates stand in phylogenetic relations with the annelids and the consequent attempt to homologize the nephridial system of the two is, in a measure, responsible for the discussion and the resulting conflict of opinions. The question in controversy is: Whether the duct takes its origin in whole or in part from the mesoderm, or whether it arises from or is contributed to by the ectoderm. At first there seems to have been no hesitancy in pronouncing the origin to be entirely mesodermic^ and all the early papers (Semper, et al) so described it.

In 1888 Johannes Ruckert ('88) stated that "the duct is developed (in the group) from the ectoderm." Van Wijhe ('89) claimed to have found evidence for the same conclusion. Several papers followed in rapid succession describing similar conditions in other groups; one author comparing the growth of the duct to that of the lateral line nerve which, he claimed, is developed from the ectoderm in the same way as the pronephric duct. The question raised by such comparison has been answered by Harrison ('03, '10). Within more recent years the statements have been very conflicting, some stating that it is, at least in part, ectodermal in origin, while others claim that the cells concerned in its development arise wholly from the mesoderm.

It has been generally assumed that in the phylogenetic history of vertebrates the pronephric " duct once opened directly to the exterior through the ectoderm, which assumption had its origin in the theory of the phylogenetic relationship of annelid and vertebrate. This is presented very clearly by Haddon ( '87) and also in his "Introduction to the study of embryology" ('87, p. 239). In these publications he recapitulates the discussion from the time ('75), when Hensen determined for the rabbit that the pronephric duct is of ectodermal origin, to Perenyi ('87) who claimed the same thing for the frog.

He quotes Spee ( '84) for the guinea-pig and Flemming ( '86) as confirming Hensen for the rabbit; also Van Wiihe ('86) as saying that the duct arises from the ectoderm in the thornback ray (Raja clavata). He says: "The origin of the segmental duct from the epiblast being now known to occur in Elasmobranchs, Anura, Lacertilia and Rodents we are justified in assuming that this is the general fact." Upon this assumption he proceeds to present an hypothesis by which Jie homologizes the pronephric duct of vertebrates with the nephiidia of the chaetopod worms. He says: "There can be little doubt that the segmental duct arises from the epiblast. This discovery will necessarily lead to a modification in our views concerning the morphology of the vertebrate excretory organs." He quotes Hatschek as having described a single nephridium in Amphioxus which is in all respects comparable with a vermian nephridium and then makes the following statement :

We have, then, only to assume that a pair of similar vermian nephridia occurred in each body segment of the ancestral vertebrate, and that the nephridia of each side of the body opened externally into a lateral groove [which he has previously described as developed in the epiblast much as the neural tube is developed]. It would further only be necessary for the groove to deepen and next to form a canal (just as does the neural groove) to bring about the vertebrate arrangement. Thus, in vertebrates, as in invertebrates, the nephridia open by epiblastic pores, but in the former the area into which they open is precociously converted into a canal which subsequently acquires a secondary opening to the exterior through the cloaca. On this hypothesis the nephridia of vertebrates always open by their original epiblastic pores, primitively directly to the exterior, secondarily into a canal separated from the epiblast.

Hertwig ('92, pp. 358-9) states the case very clearly and brings the discussion down to the date of his own publications. He speaks of the view entertained by almost all investigators that the duct arises from the mesoderm and grows by proliferation of its own cells as far as the hind gut (proctodaeum) and thus was constricted off from neither the outer nor the middle germ-layers nor yet derived from them cell-material for its increase." He claims, however, that this interpretation has become untenable, quoting various authors to substantiate the fact that, in several classes of vertebrates, the duct in process of growth is in close union with the outer germ-layer and that it is prolonged backward by means of cell-proliferation in that layer, while in front it is being constricted off from that layer or, as he terms it, the parent-tissue. He further states that the pronephric duct grows at the expense of the outer germ-layer and moves, as it were, along the latter, with its terminal opening behind, as far as the hind gut." This interpretation led, later, to the assumption by various authors of the view that not only the pronephric duct but the entire urinary system was derived from the ectoderm (Hertwig '92, p. 358).

Hertwig, later, shows that such views cannot be made to harmonize entirely with conditions found in the lower vertebrates and, making allowance for all observations, he summarizes the subject thus: "The pronephros is developed from the middle germ plate, and that then its posterior end comes into union with the outer germ-layer and in conjunction with" the latter gi'ows farther backward as the pronephric duct." He quotes Van Wijhe and Riickert in support of this explanation and then interprets the pronephric duct "at its first appearance as a short canallike perforation of the wall of the body, which begins in the bodycavity with one or several ostia and opens out upon the skin by a single orifice. Originally the outer and inner openings lay near together, later they moved so far apart that the outer opening of the canal united with the hind gut. "


It will be seen that these statements present the two phases of the problem, i.e., the origin of the duct from the mesoderm and the contribution of the ectoderm to its development. The phylogenetic theoretical consideration is also suggested.

Among the many observers who have discussed the development of the pronephric duct may be mentioned Semper ('74, 75), Balfour (78), Spee ('84), Beard ('87), Haddon ('87), Ruckert ('88), Van Wijhe ('86, '88, '89), Mitsukuri ('88), Rabl ('96), Gregory ( '98, '00). Of these Ruckert, Van Wijhe and Rabl have entered actively into the controversy, the two former being advocates of the theory of ectodermic origin or contribution to the duct, and the latter directly and radically opposed and an ardent supporter of the theory of its exclusive mesodermic origin.

Riickert ('88) seems to have taken the initiative in his very radical position, stating that the duct is developed wholly from the ectoderm. " In his figure 36, plate 16, he shows three mitotic figures at the point of union of the ectoderm and anlage of the duct. He states that two layers, ectoderm and anlagal, are fused together at the point of contact, making the thickness of the wall much greater, and this, in conjunction with the presence of the numerous mitotic figures in this region, gives much reason for his contention. He says: "Die auffallend zahlreichen Mitosen, welche sich stets in diezer Region des Ectoblast finden, geben einen weiteren Beleg fiir diese Auffassung. Ein in dieser Hinsicht ziemlich pragnanter Schnitt (Horizontal-schnitt) durch das hintere Ende des Vornierenganges ist in Fig. 36 dargestellt."

Van Wijhe claims that the ectoderm contributes to the structure of the duct, and compares its growth to that of the nerve of the lateral line organ which, according to his interpretation, is developed in part at least from cells contributed by the ectoderm. He admits the possibility of the duct growing by the division of its own cells, but considers it improbable, his general conclusion being that the ectoderm is in considerable measure responsible for its growth. He says:

Was nun seine Abstammung betrifft, so betheiligt sich das Ektoderm sicher an seiner Bildung, indem er in ahnlicher Weise wie die Nerven der Seitenorgane weiterwachst. Ebensowenig wie bei diesen Nerven mochte ich eine ausschliessliche Abstammung von der Haut behaupten, da die Mcglichkeit nicht ausgeschlossen ist, diss Zellen des Pronephros iinter fortgesetzten Theilungen den Gang in seiner ganzen Lange mitbilden helfen. Doch kommt mir Letzteres nicht wahrscheinlich vor.

He quotes Balfour ('87, p. 127). Here Balfour describes the pronephros as

. . . . arising as a sohd knob from the somatic layer of the mesoblast and growing outward toward the epiblast. The knob consists of from 20 to 30 cells agreeing in character with the neighboring cells of the intermediate cell-mass and are, at this period, rounded. It is mainly, if not entirely, derived from the somatic layer of the mesoblast. From this knob there grows backward a solid rod of cells which keeps in very close contact with the epiblast, and rapidly diminishes in size toward its posterior extremity. Its hindermost part consists in section of, at most, one or two cells. It keeps so close to the epiblast that it might be supposed to be derived from that layer were it not for the sections showing its origin from the knob above mentioned. We have in this rod the commencement of what I have called the segmental duct.

Van Wijhe, commenting on Balfour's statement, thinks that Balfour was working with the technical methods of his time, which were defective. The solid knob was the pronephros and did not arise from the middle, but from the lateral plate, and he conjectures that, because Balfour interpreted the knob as consisting of from 20 to 30 cells, he must have drawn his conclusions from a single section instead of from a series, as he himself had done. He states that his series cut the knob into 23 sections. He further states that the close contact of the duct with the epiblast, spoken of by Balfour, is not a mere touching or contact, but an actual fusion and quotes Beard, Rabl and Riickert as substantiating his statement. He quotes Riickert as having seen an invagination of the ectoderm on the duct and claims to have observed, in exceptional cases, the same thing. He states that the contribution of cells by the ectoderm to the structure of the duct is made certain through the occurrence of a mitotic figure in one of his sections, in which one daughter cell was situated in the skin and the other in the duct. He says: "Die Betheiligung des Ectoderms an der Bildung des Ganges wird sicher gestallt durch Kerntheilungsfiguren, bei welchen der eine Tochterkern in der Haut, der andere in der Anlage des Ganges liegt. Dies zeigt Fig. 5b, welche einem Schnitt durch einem Embryo von Scyllium catulus mit 37 Somiten entnommen ist."


OBSERVATIONS

The object of the present paper is to present what the writer believes to be conclusive evidence of the development of the pronephric duct, from its origin in the pronephros to its union with the cloaca, from cells derived originally from the mesoderm. No attempt has been made to describe the development of the pronephros, except in the merest outline which may be of use in introducing the account of the development of the pronephric duct itself.

The investigation was undertaken at the Harpswell Laboratory at South Harpswell, Maine, at the suggestion of Dr. H. V. Neal, who also kindly furnished the material, ready prepared, and to whom I gratefully render my thanks, not only for the material but for the kind advice and helpful criticisms. The material used was Acanthias embryos prepared by the vom Rath picro-osmoplatinic method. This renders cell outlines and limiting membranes very distinct, and makes it possible to differentiate between different cells and cell layers. In the writer's opinion, it is the sharp differentiation of cell outlines that has made it possible to distinguish with certainty, in all cases, the layer to which any individual cell belongs.

The pronephros arises from the middle plate, or mesomere, of a number of body somites — in Acanthias, six^ — by a proliferation of cells from the somatic wall. These proliferations soon fuse and the pronephric ridge (Vornierenwulst) thus formed extends laterally until it comes into contact with the ectoderm. At the line of contact the cells of the outer surface of the ridge form a cord, which later acquires a lumen and becomes converted into the pronephric duct. The cord of cells which is the anlage of the duct, rapidly grows caudally from the pronephl-ic body, between the somites and the ectoderm, until it reaches the cloaca.

During the process of backward growth the tip or growing end of the anlagal duct is at first entirely free from the ectoderm. As, however, the line of growth is obliquely backward and outward, it almost immediately comes into contact with the adjacent ectoderm. The relation of the two structures, duct and ectoderm, is very close at the point of contact, but as the former elongates, its more proximal portion gradually separates from the latter and lies free in the space between the ectoderm and the somites. It is only at the point where the growing tip of the duct impinges upon the ectoderm, if anywhere, that any contribution by the ectoderm to its growth can take place; and it. is here that the various observers claim to have found evidence of such contribution.

In an embryo Acanthias measuring 4.5 mm., corresponding nearly to No. 18 of Scammon's Normentafeln, the development of the pronephros is well advanced. The growth of the pronephric anlage toward the ectoderm is very irregular. The ridge does not extend outward with an even margin but seems to be thrust out in processes, sometimes massive and again in long, slender, pointed projections, as seen in figures 6, 7a and b, and 8. When the ridge comes into contact with the ectoderm it does not fuse with it but is merely pressed against it. This fact is demonstrated in figure 6, where the mass has been pulled away from the outer layer in the process of preparation, leaving the limiting membrane intact, but leaving its impress upon the ectoderm. The ectoderm at this stage is a one-celled layer and, while the pronephric anlage has made its impression upon it and it has become thinned out, yet it remains intact, as will be seen by its unbroken, limiting membrane. Dr. Gregory ('97) figures several instances where the separation of the two layers has occurred. She thinks that cells from the ectoderm have been torn away with the pronephric anlage. I have examined her figures and am convinced, by my own observations in similar cases, that the ectoderm is intact and that the mistake lay in her inability to follow cell outlines and limiting membranes by the method of preparation employed. "When the pronephric anlage has finally reached the ectoderm the outer cells arrange themselves so that the long axis of the nuclei coincide with the long axis of the anlage. This arrangement is preparatory to the formation of the anterior portion of the anlage of the pronephric duct which extends the entire length of the pronephric body on its outer margin. At its caudal extremity the pronephric body is prolonged into the space between the ectoderm and somites. At first free, as seen in figure 9, it later impinges upon the ectoderm, as shown in figures 1, 2 and 3, drawn from sections of an embryo of 5 mm., corresponding closely to No. 19 of Scammon's Normentafeln. The growing tip of the anlage of the duct extends along the ectoderm in close contact with that layer, and in this stage has reached a length corresponding to that of three somites. Within the pronephric body are numerous cells in mitosis and others are found also in the growing duct. In figure 1 the cavity of the somite (nephrocoele) has begun to bend toward the anlage of the duct, marking the beginning of the first pronephric tubule. The relation of the ectoderm to the anlage of the duct is very clearly shown in figure 3, which is drawn under a one-twelfth oil immersion objective in order to bring out definitely the existence of a limiting membrane between the two layers. Figures 1, 2 and 3 represent consecutive sections, figure 3 showing the growing tip and its relations to the somites as well as to the ectodermic layer. They show the extent of the growth of the anlage at this stage and demonstrate its mesodermic origin and also the fact that the contribution of cells from the ectoderm at this early period, when the anlage has just come into contact with it, is improbable, because a hmiting membrane is present between them. This will be made still more convincingly manifest by sections to be presented in subsequent figures.

A cross-section of an embryo of 4 mm. (Normentafeln No. 16) illustrates Balfour's description of the beginning of the development of the pronephros from the middle plate. Van Wijhe, in his discussion of Balfour's account, states that the soUd knob deiscribed by Balfour formed a part of the pronephros as demonstrated by his series of sections. It will also be recalled that the segmental duct of Balfour came into close contact with the ectoderm, and, but for its origin in the mesoderm, might be interpreted as being derived from it (the ectoderm). Van Wijhe claimed that this close contact was an actual fusion. It seems to the writer that the sections represented in figures 1, 3, 4 and 5 settle the question in favor of Balfour's interpretation. The demonstration of the limiting membrane between the growing tip of the anlage of the duct and the ectoderm, at the point of contact, shows beyond question the true relation between the two layers.

Figure 9 gives a comprehensive view of the entire embryo at the time when the anlage of the duct has a length of about three somites on one side and a little more than two on the other, showing the asymmetrical growth of the two sides and also the entire independence of the anlage at its first appearance. It will immediately come into contact with the ectoderm as shown in figure 3. It also shows the relation of the pronephros to the ectoderm. The appearance of the growing tip and its relation to the ectoderm present some peculiar and characteristic features at the point of contact. The ectoderm becomes thinner, apparently by the drawing out or extension of its cells. This modification of the ectoderm is very pronounced in many places, but particularly at the point where the growing tip of the duct impinges upon it. In many instances, where the two layers are closely attached, the ectoderm becomes very thin, while the extreme point of the anlage is reduced to a few cells, usually to only a single cell in diameter, but there is always to be distinguished a limiting membrane between the two. If this were not so the structure would present the appearance of thickening, as described by Riickert, and also, but for the limiting membrane, one would conclude that the thinning of the ectoderm was caused by the splitting off of cells to be contributed to the growing tip of the duct.

The pecuhar modification of the ectoderm seems to be the result of some influence, probably pressure, exerted by the growing tip of the duct. It often happens that in the process of preparation the two structures are separated, and then an indentation is left on the ectoderm where the anlage has pressed against it. Many instances of this kind, as well as cases where the same condition exists with the structures still intact, are clearly displayed in the series of cross sections. It will be observed, however, that no matter how close the contact may be, there is always to be demonstrated a limiting membrane separating the duct from the ectoderm. It will readily be appreciated that any method of preparation which obliterated cytoplasmic outlines and made limiting membranes invisible would show the two layers as a single mass, or even a syncytium, and the location of the Une of separation would be impossible.

It is obvious that if there be any contribution to the duct from the ectoderm it must take place at the point where the two layers come into contact, which is at the growing tip of the former. In order to cover this phase of the problem several series of cross sections have been drawn. The first series is from an embryo of 5 mm. and shows the terminal cell at figure 10. The section has cut through the cell at its extreme edge and left only a fragment of the cytoplasm adhering to the ectoderm. The fragment is composed of clear protoplasm and is quite thick, as may be seen from the fact that the next section has not cut entirely through it. This clear cytoplasm seems to be a constant feature in the structure of these cells and, in many instances, appears to be extra material forming the medium of attachment between the tip of the duct and the ectoderm. The next section has cut across the cell, including the nucleus.

Two features are especially striking in this series; first, the peculiar mode of attachment of the anlagal cells to the ectoderm by means of the clear cytoplasm above-mentioned, as observed in sections 10, 12, and 13, and, second, the thinning out of the ectoderm. In figures 10, 11 and 12 the contact of the anlage seems to have exerted no possible influence, as there is considerable space between the two layers.

As the sections begin to involve the larger portion of the anlage where several cells are included in each cross section of the duct, this space appears and disappears showing how loose the attachment between the layers is, except at the outer edge of the anlage where the clear cytoplasm seems to act as the direct medium of union. Particular attention is called to the distinctness with which, the line of separation between the ectoderm and anlage is shown in the. form of the limiting membrane. If the material had been prepared by any method which made this line invisible, the distinction between the two layers would have been well nigh impossible. It would have presented the appearance of a syncytium with the consequent difficulty in distinguishing to which layer the cells belonged. In order to be satisfied of the truth of this the writer examined many series of sections fixed and stained by the most common methods (e.g., corrosive-acetic, Zenker's fluid, picro-acetic, Bouih's fluid, for fixing; and boraxcarmine, Delafield's hematoxyUn, alum-hematoxyUn for staining), and found, in every instance that the difficulty of distinguishing cell outlines was so great that interpretation was practically impossible. Figures 16 to 24 are drawn from sections of an embryo of 6 mm., showing the terminal cell at figure 16 and the anlage of the duct entirely detached from the ectoderm at figure 24. The sections are consecutive and are intended to demonstrate the same conditions as the previous series, with the additional feature that the duct may be followed from its contact with the ectoderm behind to the point farther forward where it becomes entirely separated from that layer and lies free in the space between ectoderm and the somites. In this series the modification of the ectoderm is very conspicuous, as is also the peculiar mode of attachment by means of the cytoplasmic outgrowths from the cells of the tip of the duct. Particular attention is called to figure 22, where the ectoderm seems to be hollowed out for the reception of the anlage, and yet there is not only a limiting membrane, but also a considerable space between the two layers.

Figures 25 to 35 are from a series of cross sections of an embryo measuring 7 mm., beginning with the terminal cell at the growing tip of the duct and continuing forward until the anlage is practically free of the ectoderm. Prominent among the many interesting features in this series is the constant independence of the duct in its growth along the ectoderm. It will be noted that in the entire series there is constantly present a considerable space between ectoderm and duct. The modification of the ectoderm is a marked feature to be particularly noted in figures 28, 31 and 35, where the ectoderm is very much affected. Another feature, very clearly exhibited, is the way the anlage is attached to the ectoderm. The medium of attachment (already several times alluded to) is always the product of the mesodermic cells of the anlage and appears to be thrust out not unlike the pseudopodia of an amoeba. It seems to be the same in substance as that composing the processes thrown out by the pronephros before that structure reaches the ectoderm which were referred to when the growth of the pronephros was described. This substance seems to fasten itself to the ectoderm and thus unites that layer with the duct. It will be remembered that Van Wijhe claimed that the relation was one of fusion, not a mere contact. The presence of this substance acting as a medium of union between the two layers would, in a way, seem to bear out his statement, although not in the sense in which he intended it ; for it is entirely independent of the ectoderm, its relation to that layer being wholly secondary. It will be readily seen, however, that in places where the entire surface of contact is covered by this substance it might very easily be interpreted as a fusion, were the method of preparation such as to render the limiting membrane invisible. In some places it seems to exert tension upon the ectoderm. This is especially apparent in figure 35, where it forms a broad band and has apparently pulled the ectoderm toward itself, as if it had contracted and drawn the ectoderm out of line. The peculiar bend in the ectoderm brings the ectodermal cell, dorsal to the duct, into such relation with the latter that, but for the fact that the cell outlines and limiting membrane are so clearly manifest, the structures might be interpreted as being continuous, and the presence of a mitotic figure in the ectoderm at the point where this layer bends towards the duct would seem to present corroborative evidence of ectodermic contribution to the latter. But the independent character of the two layers is so clearly demonstrated in the section under discussion that no such inference is possible.

Figures 33, 34 and 35 present much the same conditions; except that in figure' 35 the mitosis is in metaphase and the cell is cut transversely so that the axis of the spindle is in the wrong direction for the daughter cell to be contributed to the duct. This was the case in 'the example drawn by Riickert in his figure 36, as cited above and pointed out by Rabl ('96). In figures 33 and 34 the mitotic figures are so placed that in division they would be parallel to the surface of the ectoderm and thus, but for the clearly demonstrated limiting membrane, the possibility of ectodermic contribution might be suggested. Another feature to be noted in this, as well as in the preceding series, is the irregular manner of the growth of the duct. This irregularity ig much more marked in earher than in later stages, and, in cases where the dividing line between it and the ectoderm is indistinguishable, would present conditions where the relative position of parts of the two layers would seem to warrant the conclusions reached by sonae observers. This is particularly obvious in the case of Riickert's figure 35, where the three sections present these conditions very markedly.

Figures 36 to 39 present a series of sections from an embryo of 10 mm. In this series the anlage is not so irregular in outline and the terminal cells are larger and not so closely attached to the ectoderm. It will also be observed that the anlage becomes freed from ectoderm much more abruptly, its attachment covering only four sections of 10 microns each. A change is also observable in the ectoderm; the cells being much thicker and more shortened, while the nuclei, in most cases, lie with their long diameter at right angles with that of the layer itself. It still, however, shows modification at the point of contact with the duct; at figure 36 profoundly so. Figure 36 is an example of the close contact between the ectoderm and anlage of the duct, and demonstrates how readily the conditions might be misinterpreted if the limiting membrane were invisible and the structures thus blended into a syncytium. The striking modification of the ectoderm and the relative position of the ectodermic cell dorsal to the duct might easily lead to the suggestion of ectodermic contribution. With the limiting membrane so clearly demonstrated such inference would seem to be unwarranted.

Figure 39a is a very interesting example of the close contact between the two layers. It shows a terminal cell of the duct and is very closely attached to the ectoderm. At either end of the cell the clear cytoplasm of attachment is very conspicuous. The whole cell seems, indeed, to be composed of it, the center around the nucleus being translucent. If the limiting membrane between this terminal cell and the ectoderm were invisible the distinction between the two layers, duct and ectoderm, would be impossible.

Figures 40 and 41 are drawn from sections of a 9 mm. embryo. The ectoderm has become very much thickened and the cells shortened. The nuclei are so placed that their long diameter is at right angles to the surface. The duct seems to lie in a groove in the ectoderm which has been hollowed out for its reception. The two sections are not consecutive, but are given to illustrate the change iij the character of the ectoderm and the close relation existing between it and the duct. Figure 40 is a section at the extreme end of the anlage and has cut through the growing tip at a point beyond the terminal cell into the clear cytoplasm which the cell throws out in its advance along the ectoderm. This peculiar feature of the growth of the anlage will be more clearly demonstrated in the frontal sections to be described later. It will be observed that the ectoderm seems to be grooved for the reception of the duct, and also that there is a well defined hmiting membrane between the two layers. These conditions are very clearly shown in figure 41 which cuts through the anlage more anteriorly and demonstrates more definitely the groove in the ectoderm, as also the arrangement of the ectodermic cells for its formation.

In figures 42 to 47, drawn from sections of an embryo of 11 mm., the duct is much larger than in the preceding series and comes into contact with the ectoderm only at the extreme tip. The ectoderm shows less modification and its cells have become much shorter and more thickened. In some places the connection between the anlage and the ectoderm is very close, as at figures 42 and 43. This is at the extremity of the growing tip, but elsewhere the two layers are widely separated and the connection between them has been completely severed. The ectoderm shows the usual evidences of modification, although the irregularity of its contour is much less great and seems to be in less degree dependent upon the presence of and contact with the anlage.

Thus far the study has been confined largely to cross sections and the evidence presented of the origin of the duct from the pronephros and hence from the mesodermic somites, seems convincing. This receives confirmation by the study of frontal sections where comprehensive views may be had of the two structures, ectoderm and duct, and their relation more extensively observed. Figure 48 is from a section of a 20 mm. embryo and shows the duct in contact with {he cloaca. It has not yet broken through the cloacal wall.

In figure 49, from an embryo of 4.5 mm., the anlage has just begun its growth from the pronephros and is pushing its way along the ectoderm. Figure 50 is the next section where the extremity of the growing tip of the anlage is shown. These two sections illustrate very clearly the origin of the structure from the pronephric body. The terminal cell is elongated into a finger-like process composed of the clear cytoplasm, before alluded to, and seems to be feeling its way, as if it were creeping along the ectoderm. The phenomena suggest the pseudopodia of an amoeba. This feature is very strikingly brought out in the subsequent figures of frontal sections, figures 51 to 54. These also show that the irregular character of the ectodermic layer is not always due to the presence of the growing tip of the duct, as the peculiar thinning of the wall seems to occur all along its course, independent of any influence that could be exerted by the pressure of the growing terminal cell. In figures 55 and 61, the duct is very large, compared with the thickness of the ectoderm, and this disproportion seems to preclude the origin of the thicker duct from the thinner ectodermic layer.

We come now to the consideration of the significance of the presence of mitotic figures in duct and ectoderm which were thought to be of so much importance by. Riickert and Van Wijhe. In his figure 36 Riickert shows a section with three mitotic figures in the region of the point of union between the ectoderm and duct. As will be recalled, he attributes much importance to the fact that the fusion of the duct with the ectoderm makes the tissue at the point of union much thicker and also lays stress upon the presence within the mass, of numerous mitotic figures. To this argument Carl Rabl long ago made what to the writer's mind is an entirely adequate answer. He claims that only one of the cells in mitosis is to be considered — the one in the ectoderm immediately contiguous to the duct. This cell is in metaphase (Miitterstern) and is so placed that one sees it in the direction of the axis, which is parallel to the surface of the ectoderm, not perpendicular to it, as it must be, if, in its division, it is to contribute a daughter-cell to the duct. He says:

Am hinter Ende des Ganges sind hier drei Theilungsfiguren zu sehen. Diejenige, welche sich unmittelbar an den Vornierengang anschliezt und auf die es daher zunachst ankommt, ist ein Miitterstern, der so gestellt ist, dass man auf ihn in der Richtung der Achse seiht; die Achse steht also parallel zum Ektoderm, nicht senkrecht, wie sie stehen mtisste, wenn die Figur beweisend sein sollte. Die beiden anderen Theilungsfiguren sind, meiner Ansicht nach, irrelevant.

Riickert, in his figures 35 to 41, shows several sections which display much the same conditions as are presented in the cross sections figured in the present paper, and in the light of the writer's own observations and the repeated instances where the same relative position of duct and ectoderm have existed, it is difficult to escape the conclusion that the same interpretation is inevitable in both cases. So far as the mitotic figures are concerned it seems certain that two of them are in the duct and one in the ectoderm. All three of the cells are in metaphase, and the two in the duct are so placed that the direction of division would be such as to preclude the possibility of exchange of cells between ectoderm and duct, even though the mitosis were in the ectoderm. This is what Rabl meant, it would seem, although he considered only one of the mitotic figures.

Numerous mitotic figures are illustrated in my drawings, several of which are of particular interest. Figure 35 shows a cell in mitosis in the ectoderm, which is thinned out at one place and the duct is attached to it just above this. The attachment is by means of the peculiar clear protoplasm, before alluded to. In this case it seems to have contracted and drawn the ectoderm out of line so that the mitotic cell is deflected toward the anlage. It will be readily observed that, if the method of preparation did not make it possible to demonstrate clearly the outline of the cell and the limiting membrane between the two layers, the structures would appear to be continuous. It is true that the mitosis is in metaphase, with the axis in the wrong direction for possible contribution of a cell to the duct, but suppose the same conditions to have existed in figures 33 and 34, might not the conclusion of such contribution have been reached? Indeed this is what did happen in the case figured by Van Wijhe. In this case, too, it seems to the writer, Rabl has given the right answer, where he claims that both parts of the dividing cell are in the anlage of the duct.

Figure 57 shows a mitosis which seems closely comparable to that figured by Van Wijhe. The section was stained with Delafield's hematoxylin and, as it originally appeared, figure 57 was very strikingly like Van Wijhe's figure. By accident it was broken apart (in the act of focusing a one-twelfth oil immersion objective upon it), and when the duct parted from the ectoderm, it left the latter intact. The drawing, figure 58, was made with the duct moved back to the ectoderm, where it so lies that the latter overlaps it, but it very clearly illustrates the fact that the two layers are entirely separate and that they were never fused, in the sense in which the term is repeatedly used; and also that contribution from the dividing cell in the ectoderm to the duct was impossible.

Figure 60 presents a similar instance. Here the terminal cell of the duct is very closely united to the ectoderm and particularly so with an ectodermic cell in mitosis. It is obvious that if the limiting membrane were not present and the structure were a syncytium, the interpretation would be very difficult and contribution of a daughter cell to the duct from the mitosis would seem more than probable. These are examples out of many instances and might be indefinitely multipHed. At figure 56 is shown a mitotic figure in the terminal cell of the duct and, at 55, in the anlage just above the terminal cell.

Other mitotic figures appear very frequently in the duct and in many instances in such positions and relation to the ectoderm as to lead to misapprehension as to their significance. One such is shown in figure 59. Here a mitotic figure occm's in the duct in such relation to the ectoderm that, but for the clearly distinguishable cell outline and the presence of a Umiting membrane, it would be difficult to determine to which layer the cell belongs. Under conditions as above presented the question could not be raised.


Figure 49 shows the caudal extremity of the pronephros in an embryo of 4.5 mm., with the duct just beginning, and here, just at the junction of the pronephros and duct, is a mitotic figure. Its presence at this point and the early stage of development seemingly can have but one meaning, namely, that the anlage is growing by the division of its own cells. This is not an isolated example as may be seen by reference to the figures.

All these examples, coupled with other evidence, seems to point inevitably to the conclusion that, not only is the anlage of the duct a direct outgrowth from the pronephros, and therefore mesodermic in origin, but its subsequent growth is accomplished by the division of its own cells and it in nowise receives contribution of cells from the ectoderm.

Respecting the theoretical phase of the problem, it seems clear that the relation of the anlage of the duct is secondary and that the duct is not developed from an ectodermic groove. If the nephridia of vertebrates ever opened to the surface through ectodermic pores, and later into a canal arising from the ectoderm, ontogeny has failed to repeat phylogeny in Acanthias, for there is no evidence of such arrangement in the embryological development of this form.

BIBLIOGRAPHY

Balfour, F. M. 1878 A monograph on the development of elasmobranch fishes.

London. Beard, J. 1887 The origin .of the segmental duct in elasmobranchs. Anat.

Anz., Bd. 2. Felix, W. 1890 Zur Entwicklungsgeschichte der Vorniere des Hiihnchens.

Anat. Anz., Bd. 5. Flemming, W. 1886 Die ectoblastische Anlage des Urogenitalsystems beim

Kaninchen. Arch, f . Anat. und Phys. Gregory, E.R. 1897 Originof the pronephric duct in selachians. Zool. Bull.,

Boston, vol. 1, pp. 123-129.

1898 The pronephros in Testudinata. Science, N. S., vol. 7.

1900 Observations on the development of the excretory system in turtles. Zool. Jahrb.^Anat. Abt., Bd. 13, pp. 683-714.

Haddox, a. C. 1887 Suggestion respecting the epiblastic origin of the segmental duct. Proc. of the Royal Dublin Society, vol. 5, February.

Harrison, R. G. 1903 Experimentelle Untersuchungen ueber die Entwicklung der Sinnesorgane der Seitenlinie bei den Amphibien. Arch. f. mikr. Anat., Bd. 63.

Harrison, R. G. 1910 The development of peripheral nerves in altered surroundings. Archiv f. Entwicklungs-Mechanik der Organismen, Bd. 30, Part 2, June.

Hertwig, O. 1892 Text-book of embryology.

MiTSUKURi, K. 1888 The ectoblastic origin of the Wolffian duct in Chelonia. Zool. Anz.,Bd.l2.

Perenyi, J. 1887 Die ectoblastiche Anlage des Urogenitalsystems bei Rana esculenta und Lacerta viridis. Zool. Anz., Jahrg. 10, p. 66.

Rabl, C. 1896 Ueber die Entwicklung des Urogenitalsystems der Selachier. Morph. Jahrb., Bd. 24, pp. 632-767, Taf. 13-19.

RucKERT, J. 1888 Ueber die Entstehung der Excretionsorgane bei Selachiern. Arch. f. Anat. Phys., Juliheft.

ScAMMON, R. E. 1911 Normal plates on the development of Squalus acanthias. Normentafeln zur Entwicklungsgeschichte der Wirbelthiere, herausgegeben von F. Keibel. 144 pp., 4 plates and 26 text-figures.

Semper, C. 1874 Segmentalorgane bei ausgewachsenen Haien (vorlalifige Mittheilung). Centralblatt f. die medic. Wiss., Bd. 7, November.

1875 Kurze Bemerkungen iiber die Entstehungsweise der Miiller'schen und Wolff'schen Gauge. Centralblatt f. die medic. Wiss., Bd. 26, Juni.

1875 Das Urogenitalsystem der Plagiostomen und seine Bedeutung fiir das der librigen Wirbelthiere. Arbeiten aus dem zool. Inst, in Wiirzburg. Bd. 2.

Spee, G. 1884 Ueber directe Betheiligung des Ektoderms an der Bildung der Urnierenanlage des Meerschweinchens. Arch, f . Anat. u. Phys.

Van Wijhe, J. W. 1886 Die Betheiligung des Ektoderms an der Entwicklung des Vornierenganges. Zool. Anz., Bd. 9, pp. 633-635.

1888 Ueber die Entwicklung des Exkretionssystems und anderer Organe bei Selachiern. Anat. Anz., Bd. 3, pp. 74-76.

1888 Bemerkungen zu Dr. Rlickert's Artikel ueber die Enstehung der Exkretionsorgane bei Selachiern. Zool. Anz., Bd. 11, pp. 539-540.

1889 Ueber die Mesodermsegmente des Rumpfes und die Entwicklung des Exkretionssystems bei Selachiern. Arch. f. mikr. Anat., Bd. 33, pp. 461-516.

1898 Ueber die Betheiligung des Ektoderms an der Bildung des Vornierenganges bei Selachiern. Verh. Anat. Gesellsch., 12 ten Versamml., pp. 31-37.


ABBREVIATIONS

Ect., ectoderm Pro.du., pronephric duct, or anlagal

G.tp., growing tip of the anlagal duct duct

Ldm.m., limiting membrane Pron., pronephros

A^f.c, notochord Sp.c, spinal cord, or neural tube

PLATE 1

EXPLANATION OF FIGURES

1 Frontal section of a 5 mm. embryo, showing the pronephros arising from six somites, the sixth participating but little in this section. The fifth curves outward toward the anlage of the pronephric duct, a, with which it is to unite, thus forming a complete tubule. The duct will arise on the lateral edge of the pronephric body, where the nuclei are beginning to arrange themselves in line, h. At the caudal end the anlage of the duct projects into the space between ectoderm and somites. The extent to which it has grown may be seen by comparing with figures 2 and 3.

2 The same parts as in figure 1, illustrating method of growth. Xote the numerous mitotic figures in the pronephric anlage and one in the anlage of the duct, indicating rapid growth; ectoderm a single cell thick.

• 3 Growing tip of figure 1 under high power, to show its relation to the ectoderm.

4 and 5 From cross sections of a very earh' embryo. They correspond to some figures in Balfour's Monograph and show a very early pronephros and its origin from the middle plate.

5 Same, more enlarged, showing contact with ectoderm at a.

6 and 7 From embryos of 4.5 and 5 mm., illustrating the manner of growth of pronephros toward ectoderm, the pronephros having been pulled away in figure 6, leaving its impress. This separation, caused by shrinking in preparation, demonstrates the independence of the two structures. Figure 7 illustrates the manner of growth of the pronephros; the clear translucent cytoplasmic processes at a and h are very characteristic.

8 Shows some isolated cells bearing such processes


EXPLANATION OF FIGURES


9 From an embryo of 5 mm., showing the anlage of the pronephric duct growing from the pronephric anlage at a. The ducts of the two sides are asymmetrical, one side extending over two and the other three somites. The independence of duct and ectoderm is seen at c. The two were in close contact in the living embryo but have separated in preparation. The drawing also shows the distance to be traversed by the duct to reach the cloaca.

10 to 15 Consecutive sections 6 n thick, of a 5 mm. embryo. In figure 10 the duct is attached to the ectoderm by clear protoplasm. Figure 12 shows the cell body and nucleus of the same cell. These sections show that the medium of connection of ectoderm and duct is derived entirely from the duct and also the limiting membrane between the two layers.

16 to 24 Consecutive sections, 4 n thick, of a 6 mm. embryo, showing the anlage of the duct from the terminal cell, a, to where it is free from the ectoderm. Note modification of ectoderm at point of contact with duct, the means of attachment of the two and the limiting membrane. The outline of the anlage of the duct is more regular than in last series.



EXPLANATION OF FIGURES


25 to 35 Cross sections, 6m thick, of a 7 mm. embryo. Note the medium of attachment of duct and ectoderm and its obvious origin from the former. In figure 30 the cytoplasm of the cell is a thick layer, attached to the ectoderm at only two points. This and figure 31 show the irregular form of the growing tip, which becomes more regular as it frees itself from the ectoderm. Figures 33 to 35 show mitotic figures in the ectoderm, those in figures 33 and 34 being in metaphase with the axis of the cell corresponding with that of the ectoderm ; that in figure 35 being at right angles to the same axis. The anlage of the duct is attached to the ectoderm by clear cytoplasm, which, in figure 35, seems to have contracted and pulled the ectoderm out of line.

36 to 39 Consecutive cross sections of a 10 mm. cmbi'yo, presenting much the same features as the last series, except that the duct is larger and the connection between it and the ectoderm is not so close and it becomes free more abruptly.



PLATE 4


EXPLANATION OF FIGURES


40 and 41 Selected sections of a Q.5 mm. embryo. In figure 40 the clear protoplasm at the extreme tip might be interpreted, but for the limiting membrane, as a part of the ectoderm, it conforms so perfectly to the contour of that layer; while the ectoderm seems hollowed out for its reception, the clearly defined line makes it easy to distinguish the layers. Figure 41 shows the ectoderm as if grooved for the reception of the duct. These sections show the ectoderm much thickened and the cells placed so that the nuclei have their axes at right angles with that of the layer.

42 to 47 Cross sections of a 11 mm. embryo, showing much the same features as the last, except that the separation of the duct from the ectoderm is more abrupt and the duct is rounded and more regular. The ectoderm is less modified and much thickened, with the cells shorter.

48 The anlage of the duct where it reaches the cloaca in a 20 mm. embryo. At this point the duct bends sharply inward to reach the cloaca, hence the section is oblique. The wall of the duct is of columnar epithelium. It has not yet broken through to the cloaca, but the walls are fused, and the drawing shows the duct as a triangular loop attached to the cloaca, c, by its base.


PLATE 5

EXPLANATION OF FIGURES

41) From a frontal section of a 4.5 mm. embryo (about 25 somites). The aniage of the duct is beginning to grow back from the pronephros and has just become attached to the ectoderm, which is much thinned and bent outward just above the point of attachment.

50 From a frontal section of a 5 mm. embryo, shows the close contact of ectoderm and the aniage of, the duct, and demonstrates the impossibility of distinguishing between the layers, were the limiting membrane invisible.

51 to 54 Frontal sections of 5.5 and 6 mm. embryos, demonstrating the character of the growing tip of the aniage of the duct .

55 to 60 ' Showing mitotic figures in duct and ectoderm. Figure 56 is in the terminal cells of the aniage of the duct. Figures 57, 58 and 60 are comparable to the figure of Van Wijhe. The mitotic cells in the ectoderm are about to divide and, but for the limiting membrane, they might be interpreted as contributing cells to the aniage of the duct. This is especially evident in figure 57. Figure 58 shows the same section after the aniage has separated from the ectoderm and has been moved back into place, showing the entire independence of duct and ectoderm; figure 60 is much the same.

61 From a frontal section of a 7.5 mm. embryo, shows the flisproportion in size between the aniage of the duct and the ectoderm. It also shows that the ectoderm may be modified at other points than at the growing tip. The close relation between duct and ectoderm and the limiting membrane are also shown.


Studies On Germ Cells

I. The History Of The Germ Cells In Insects With Special Reference To The Keimbahn-Determinants.

Ii. The Origin And Significance Of The Keimbahn-Determinants In Animals.

Robert W. Hegner

From the Zoological Laboratory of the University of Michigan, Ann Arbor,

Michigan, U.S.A.

Seventy-Four Figures

Contents

General introduction 376

I. The history of the germ cells in insects with special reference to the Keim bahn-determinants 379

1. Introduction 379

2. Diptera 380

A. Historical account 380

B. The Keimbahn in Miastor americana Felt " 387

C. The Keimbahn in Compsilura concinnata Meig 398

3. Coleoptera 4C0

A. Historical account 400

B. Nuclear division and differentiation in the eggs of Chrysomelid

beetles 408

C. The growth of the oocytes and development of testicular cysts

in Chrysomelid beetles 413

(1) The differentiation of nurse cells and oocytes in the ovary

of Leptinotarsa decemlineata 414

(2) The origin of the pole-disc granules in Leptinotarsa decem lineata 417

(3) The spermatogonial divisions and formation of cysts in the

testes of Leptinotarsa decemlineata 420

(4) Amitosis in the germ cells of Leptinotarsa 424

4. Hymenoptera 428


' Contributions from the Zoological Laboratory of the University of Michigan, and the Marine Biological Laboratory, Woods Hole, Mass.

375

JOURNAL OF MORPHOLOGY, VOL. 25, NO. 3 SEPTE.MBER, 1914


376 ROBERT W. HEGNER

II. The origin and significance of the Keimbahn-determinants in animals. . . . 432

1. The Keimbahn in the Crustacea 433

2. The Keimbahn in the Nematoda 442

3. The Keimbahn in Sagitta 446

4. The Keimbahn in vertebrates 449

5. The Keimbahn in other animals 456

6. The genesis, localization, distribution, and fate of the Keimbahn-;le terminants 460

A. The genesis of the Keimbahn-determinants 460

' a. From nuclear substances 462

b. From cytoplasmic or extracellular nutritive substances .... 469

c. From a differentiated part of the cytoplasm 474

B. The localization of the Keimbahn-determinants 478

C. The fate of the Keimbahn-determinants 482

7. Conclusions and summary 485

Literature cited r 489

GENERAL INTRODUCTION

Studies of the history of the germ cells in animals have proven that in many cases these cells originate in a perfectly definite way and at such an early embryonic period as to represent the first cellular differentiation that ta|ves place in ontogeny. In certain animals such a determinate segregation of the germ cells cannot be established with the data available without certain assumptions to which objections may be made. Limiting ourselves, therefore, to the instances where the germ cell cycle is completely known, it is possible to divide the history of the germ cells from one generation to the next into the following periods :

1. Primary cellular differentiation, i.e., the formation of one or more primordial germ cells during the segmentation of the egg;

2. A short period during which in some cases the primordial germ cells increase slightly in numbers by mitosis;

3. A long period of rest characterized by cessation of cell division, either active or passive change of position, separation of the germ cells into two groups which become the definitive germ glands, accompanied by the general growth of the embryo until the larval stage is almost attained;

4. Multiplication by mitosis of the primitive oogonia or spermatogonia to form a definite number (Miastor and perhaps


•STUDIES ON GERM CELLS 377

others) or indefinite number (so far as we know) of oogonia or spermatogonia;

5. In some cases the differentiation of oogonia into nurse cells and ultimate oogonia, and the spermatogonia into Sertoli cells and ultimate spermatogonia;

6. The growth of the ultimate oogonia and spermatogonia to form primary oocytes and primary spermatocytes;

7. Maturation;

8. Fertilization (if not parthenogenetic) .

This list of periods differs from the series usually recognized in that it starts with the beginning of the germ cell cycle instead of at a comparatively late stage, i.e., with oogonia and spermatogonia. Certain of these periods, especially those of maturation and fertilization, have been emphasized by investigators much more than others. Many of the fundamental problems of heredity and development are, however, concerned with the events which take place during the less known stages.

For a number of years the writer has been particularly interested, in the segregation of germ cells during embryonic development, and has studied especially certain visible substances which are present in the egg before cleavage begins, and later become part of the material contained in the primordial germ cells. In the eggs of certain Chrysomelid beetles this substance was termed the 'pole-disc' (fig. 7, A, g.c.d., p. 403), and the granules of which the pole-disc is composed were called 'germ-cell determinants' because they enable us to determine which cells will become germ cells. Since this term is Ukely to be misinterpreted, the granules of the pole-disc and other similar substances that have been found in the eggs of animals are, in this paper, called 'Keimbahn-determinants,' since they furnish the means of recognizing the germ-cell material in the undivided egg. or in cleavage stages, and thus make it possible for us to determine the 'Keimbahn' from one generation to another.

The following events may be listed in the history of the Keimbahn-determinants :

1. Localization of the Keimbahn-determinants in the oocyte or mature egg;


378 ROBERT W. HEGNER

2. Association of one or more cleavage nuclei with part or all of the Keimbahn-determinants to form one or more primordial germ cells;

3. The apparently equal distribution of the Keimbahndeterminants between the daughter germ cells at each mitotic division (Sagitta possibly excepted);

4. The disappearance of the Keimbahn-determinants in the oogonia and spermatogonia;

5. The reappearance of the Keimbahn-determinants in the oocyte or mature egg.

In the general history of the germ cells there may be two periods of differentiation :

1. The segregation of the primordial germ cells during cleavage stages.

2. The differentiation of nurse cells and ultimate oogonia, or Sertoli cells and ultimate spermatogonia in the germ glands. This second differentiation does not occur in Miastor and certain other animals, and even when it does occur it is doubtful whether the nurse cells and Sertoli cells should be considered as true somatic cells or simply as abortive oogonia or spermatogonia which have been unsuccessful in the struggle for development. A casual examination is Hable to delude one into thinking that the differentiations mentioned above are widely separated in the germ cell cycle, but a little closer study shows that they really occur during a relatively short period in the entire history. For example, in certain insects, where nurse cells arise from oogonia, this process takes place just before the growth period during which the Keimbahn-determinants became localized in preparation for the primary cellular differentiation.

It is evident from the general outline as stated above that the most important period in the germ cell cycle is that extending from the formation of the ultimate oogonia and spermatogonia to the complete segregation of the primordial germ cells. Our knowledge of events during the latter part of this period is comparatively great, whereas we know practically nothing about the early stages involving the genesis of the Keimbahn-determinants and their localization in the oocyte or mature egg.


STUDIES ON GERM CELLS 379

Embryological investigation has gradually progressed from the study of germ layers back to the study of the segmentation of the egg, and from this to the organization of the ovum, and from here to the genesis and localization of organ forming substances in the oocyte.

In the following pages the results of some investigations made by the writer are described, and a discussion of the results obtained by other investigators is given, in an attempt to determine the origin, nature, and significance of the Keimbahndeterminants.

I. THE HISTORY OF THE GERM CELLS IN INSECTS WITH SPECIAL REFERENCE TO THE KEIMBAHN-DETERMINANTS

1. INTRODUCTION

The Keimbahn in animals was first described by Metschnikoff ('65, '66) in the paedogenetic larvae of the fly, Miastor. Since that time various investigators have been able to trace the germ cells in many other species of insects, belonging to several different orders, from early cleavage stages to the definitive germ glands, and have discovered that a complete Keimbahn can also be demonstrated in species belonging to other classes and phyla, notably the Crustacea, the Nematoda, and the Chaetognatha. The writer ('09) has published an account of our knowledge of the origin and early development of the germ cells in insects up to the year 1908, but no complete account of the Keimbahn in other groups of animals has ever appeared.

The data regarding this phase of the germ cell cycle are widely scattered in the literature; frequently buried in treatises on general embryology, and less often contained in contributions devoted to this subject alone. The accounts found in current reference books and text books are for the most part obsolete or inaccurate. In the following account statements, with figures, of the more important discoveries of other investigators have been included in order to allow a general consideration of our entire knowledge regarding the Keimbahn-determinants.


380 ROBERT W. HEGNER

2. DIPTERA

A. Historical account

The segregation of the germ cells in the early embryonic stages of animal development was first discovered in certain Dipterous insects. In 1862 Robin described, in the nearly transparent eggs of Tipulides culiciformes the appearance of four to eight buds at one pole just previous to the formation of the blastoderm. He called these buds 'globules polaires' and thought that they were protruded at the anterior end. Weismann ('63) likewise discovered bud-like protrusions at a corresponding stage in the development of the egg of Chii'onomus nigroviridis and Musca vomitoria. He corrected Robin regarding their orientation by proving that they arise at the posterior end and not at the anterior end of the egg. Because of their position he applied to them the term 'Polzellen/ a term that has persisted until the present time. Weismann was unable to follow the history of the pole cells and so did not succeed in determining their true significance.

Metschnikoff ('65, '66) and Leuckart ('65) were the first to announce that the pole cells are really primordial germ cells, and the first to trace them from their initial appearance until they entered into the constitution of the definitive germ glands. Their results, obtained from the study of the eggs of Miastor and Simula, were confirmed by Grimm ('70) and Balbiani ('82, '85) in Chironomus.

Pole cells have also been described among the Diptera, in Musca by Kowalevsky ('86), Voeltzkow ('89), and Escherich ('00) ; in Calliphora by Graber ('89), and Noack ('01) ; in Chironomus by Ritter ('90), and Hasper ('11); in Lucilia by Escherich ('00); and in Miastor by Kahle ('08) and Hegner ('12).

Among insects belonging to other orders typical pole cells have been found only in parasitic Hymenoptera (Silvestri, '08), and in Chrysomelid beetles (Lecaillon, '98; Hegner, '08, '09; Wieman, '10a) although germ cells have been described at an early stage in the development of the butterflies, Euvanessa antiopa (Woodworth, '89) and Endromis versicolora (Schwan


STUDIES ON GERM CELLS 381

gart, '05); in the aphids (Metschnikoff, '66; Balbiani, '66-'72; Witlaczil, '84; Will, '88); in the honey-bee (Petrunkewitsch, '01-03); and in Forficula auricularia (Heymons, '95).

Since the original work contained in this paper was undertaken in order to determine the origin and significance of certain pecular inclusions in the primordial germ cells of various animals the writer has been particularly interested in any extra-nuclear substances visibly different from the general cytoplasm. One of the principal characteristics used for the purpose of identifying germ cells in the embryos of animals is the presence within their cytoplasm of yolk substance. Many of the authors cited above noticed yolk globules in the pole-cells. For example, in Chironomus, Weismann ('63) states that each pole-cell possesses ein oder zwei Dotterkornchen;" and Metschnikoff ('66) described dark yolk masses in the pole-cells of Simula and Miastor. These examples indicate the general presence of yolk-like substances in the primordial germ cells of the Diptera, but it remained for later more detailed investigations with finer methods to determine the origin and fate of these cytoplasmic inclusions. Five papers have appeared which contain information bearing on these problems; (1) Ritter ('90) on Chironomus, (2) Noack ('01) on Calliphora, (3) Kahle ('08) on Miastor, (4) Hasper ('11) on Chironomus, and (5) Hegner ('12) on Miastor.

Chironomus. As stated above, the pole-cells of Chironomus, were first described by Weismann ('63) who, however, did not recognize them as germ-cells. Grimm ('70) succeeded in tracing the pole-cells in Chironomus until they became surrounded by other cells, forming two germ-glands, thus confirming Metschnikoff's ('66) account in Miastor. Chironomus was later studied again by Weismann ('82), by Balbiani ('82, '85), by Jaworowski ('82), by Ritter ('90) and by Hasper ('11). Only the work of the last two needs to be considered here since that of the other writers mentioned was not carried on with modern methods nor in such great detail.

Ritter ('90) used the section method and was thus able to study the structure of the germ cells more carefully and to trace them more accurately during embryonic development. He found


382 ROBERT W. HEGNER

that the first pole-cell appeared at the posterior end of the egg when there were a large number of nuclei scattered about in the yolk. A second pole-cell was protruded close behind the first. Each carried out of the egg part of a flat mass of granules which, in section, formed a wreath around the nucleus. The two original pole-cells increased by division to four and then to eight. Two divisions of each pole-cell nucleus now occurred, resulting in eight quadrinucleated cells; these seemed to move of their own accord through the blastoderm which closed after them. They now lay at the posterior end of the germ-band from whence they were possibly moved anteriorly by the growing forward of the entomesoderm. The mass of pole-cells finally divided into two groups which occupied a position on either side of and dorsal to the hind-intestine; there they remained until after the larva hatched, when they became the definitive sex-organs.

Ritter was the first to determine the fact that the 'yolk masses' contained in the pole-cells of Chironomus are derived from a definite structure and are not chance acquisitions from the yolk granules in the egg. After giving a brief sketch of the polar bodies. and male and female pronuclei, he says:

In dem nachsten Stadium sind in dem Dotter keine Zellen mehr zu sehen; dagegen tritt an demjenigen Pol, an welchem spater die Polzellen erscheinen, also an dem hinteren, ein eigenthlimlicher wulstartiger Korper auf, welcher durch das Hamatoxylin sehr dunkel gefarbt wird. Er erscheint auf mehreren Schnitten und stellt eine etwas nach oben vorgewolbte Platte dar, welche vielfach runde Fortsatze zeigt und aus feinkornigem Protoplasma besteht. Er l^leibt bis zum Austritt der Polzellen an derselben Stelle.

Ritter then gives a fragmentary account of the early divisions of the cleavage nucleus, at the end of which, the two first polecells appear each containing a grossen Kern und um denselben herum kranzformig einen Theil des obengenannten dunklen wulstformigen Korpers." This darkly staining body he called 'Keimwulst.'

That the 'Keimwulst' played an important role in the segregation of the germ cells was quite obvious to Ritter, but he was in «rror when he stated that this body contained the first cleavage


STUDIES ON GERM CELLS 383

nucleus and dass nach der Theilung des Furchungskernes die Theilprodukte theils in dem dunklen wulstformigen Korper verbleiben, theils aus demselben herausriicken."

In 1911 Chironomus was again studied by Hasper, who published a complete description of the Keimbahn in Chironomus confinis and C. riparius. At the posterior pole of the eggs at the time of deposition is a disc-shaped mass of granules (fig. 1 A, kbpl) called by Hasper the 'Keimbahnplasma,' which is identical with the Keimwulst' of Eitter. Hasper characterizes this 'Keimbahnplasma'

. . . . als dichte, scharf konturierte, wurst- oder flaschenformige, gerundete oder auch in 2 Klumpen getrennte, mit wenigen Vacuolen versehene Masse prasentiert, die am hintern, im Ovarium nach hinten gekehrten Ende des Eies etwas unter der Oberflache leigt, in schaumigem Protoplasma eingebettet, zuweilen aber auch noch ganz von Dotter umgeben. Es ist cliese wichtige Differenzierung des Ooplasmas nichts anderes als jene spezifische Substanz, die bei der Determinierung des ersten von cler Entmcklung dargebotenen embryonalen Materials eine entscheidende Rolle spielt und die daher im Folgenden als Keimbahnplasma noch mehrfach Erwahnung finden wird (pp. 549-550).

Ritter ('90) advanced the idea that the cleavage nucleus of Chironomus divides within the 'Keimwulst' and that here the first cleavage division occurs, one daughter nucleus remaining in the 'Keimwulst' and becoming the center of the primordial germcell, the other giving rise to somatic nuclei. This is probably the basis for Weismann's ('04) statement regarding his conception of the germ-plasm that

If we could assume that the ovum, just beginning to develop, divides at its first cleavage into two cells, one of which gives rise to the whole body (soma) and the other only to the germ-cells lying in this body,

the matter would be theoretically simple As yet, however,

only one group of animals is known to behave demonstrably in this manner, the Diptera among insects ....

There is, however, nothing in the literature to warrant the above statement, since Ritter's hypothesis has been disproved by Hasper.

The primordial germ-cell is really recognizable as such in Chironomus at the four-cell stage (fig. 1, B, p.g.c). One of the


384 ROBERT W. HEGNER

first four cleavage nuclei migrates to the posterior end, and, separating from the rest of the egg together with the 'Keimbahnplasma' and the cytoplasm in which this substance lies, forms the 'Urgeschlechtszelle.' The primordial germ cell is undergoing division by mitosis at the time when it is protruded from the egg and during this process the 'Keimbahnplasma' is apparently equally divided between the daughter cells.

Wahrend die ersten Teilungen rasch aufeinander folgen, kommt die letzte gar nicht mehr sur Vollendung, d.h. sie erstreckt sich nur auf die Kerne, so dass schliesslich 8 zweikernige Genitalzellen im hintern Polraum liegen. Und damit ist die Entwicklung der Keimbahn fiir lange Zeit liberhaupt abgeschlossen; denn wahrend der nun folgenden Embryonalperiode ist sie durch ein durchaus passives Verhalten ausgezeiehnet (p. 553).

One of these binucleated germ-cells is shown in figure 1, C.

It is unnecessary to trace the history of these primordial germ-cells (pole-cells) since it has been shown repeatedly that they give rise to the oogonia or spermatogonia in the definitive germ-glands. Portions of the 'Keimbahnplasm' persist at least until the larva hatches (fig. 1, D). The origin and nature of the 'Keimbahnplasm' was not discovered by Hasper but the name applied to it and the fact that the author adopts my term 'germcell determinants' (Keimzelldeterminanten) in discussing it, indicate that he considers it of fundamental importance in the segregation of the germ-cells.

The possibility of determining the origin of the 'Keimbahnplasma' of Chironomus led the writer to study the oogonia in the terminal chamber and the various stages in their growth up to the time of deposition. Larvae were collected and allowed to develop in the laboratory and the ovaries were dissected out of the adults which were obtained from these larvae. However, the material procured has been found lacking in both the earlier stages of the development of the oocytes and the late stages in ' the formation of the ovum. It has been considered best, therefore, to reserve a study of this material until a complete series can be secured.


STUDIES ON GEKM CELLS


385


^Oo^o^^^



- Fig. 1 Chironomus (redrawn from Hasper, '11). A, longitudinal section • xrough the posterior end of a freshly laid egg. B, longitudinal section through egg during division of first four cleavage nuclei; at posterior end primordial germ cell is just being formed. C, one of primordial germ cells containing two nuclei and remains of 'Keimbahnplasma.' D, germ gland of the larva in which remains of 'Keimbahnplasma' still appear. Kbpl, 'Keimbahnplasma'; p.g.c, primordial germ cell.


386 ROBERT W. HEGNER

Calliphora. Noack ('01) found a dark granular layer, which he called the 'Dotterplatte' (fig. 2, Dpi) at the posterior end of the egg of Calliphora erythrocephala similar to the 'Keimwulst' discovered by Ritter in Chironomus. Each pole-cell took part of this layer of granules with it as it passed through the 'Keimhautblastem.' Concerning this process Noack says, "Im nachsten Stadium haben die Kerne eine runde Gestalt angenommen, die Platte hat sich in so viel Theile getrennt, als Kerne in ihren



P.gP:


Fig. 2 Calliphora (redrawn from Noack, '01). A, longitudinal section through posterior end of freshly laid egg showing 'Dotterplatte' (Dpi). B, longitudinal section through posterior end of egg at time of blastoderm formation showing protrusion of primordial germ cells (p.g.c).

Bereich eingetreten sind, und bildet nun um jeden dieser Kerne einen peripher gelegenen feinkornigen Halbmond. Hiermit ist die erste Zelldifferenzirung eingeleitet." Those cells which now contain granules from the "Dotterplatte' are recognized as pole-cells, while the remaining cells which have reached the periphery of the egg constitute the blastoderm. The 'Halbmond' of granules which surrounds the nucleus of each pole-cell now

. . . . schliesst sich allmahlich zu einem Kreise, welcher uin so mehr auffjillt, weil die von ihm eingeschlossene und den Kern einbet


STUDIES ON GERM CELLS 387

tende Protoplasmamasse fast farblos erscheint (fig. 2, B, p.g.c). Bei der Fortentwicklung der Polzellen schwindet allmahlich die scharfe Grenze zwischen Zellprotoplasma und Polplatte. Letztere lost sich auf und es entsteht eine gleichmassige Pigmentirung, welche den Polzellen noch auf lange Zeit ein ganz charakteristisches Aussehen verleiht.

Concerning the nature of this 'Dotterplatte' Noack says:

Dass die Platte am hinteren Pole des Musciden-Eies sich aus Dotterelementen zusammensetzt. Sie scheint den Zweck zu haben, das Wachsthum am hinteren Pol zu beschleunigen, ferner durch Eintritt in die Polzellen es diesen zu ermoglichen, sich auch weiterhin lebhaft zu vermehren, obgleich sie vom Dotter her keine Nahrung mehr erhalten. Schhesshch verursacht sie die charakteristische Pigmentirung dieser Zellen.

B. The Keimbahn in Miastor americana Felt

The paedogenetic flies of the genus Miastor furnish especially favorable material for the study of the germ-cell cycle. The process of paedogenesis was discovered by Wagner, and a short statement was published by him in 1862; three years later a more detailed account appeared by Wagner ('65) whose extraordinary discovery was confiirmed by Meinert ('64), Pagenstecher ('64) and Ganin ('65); but of the early authors Leuckart ('65) and Metschinkoff ('65, '66) have given the best descriptions of the developmental stages. In 1870 Grimm announced the occurrence of paedogenesis in a species of Chironomus. From that date until 1907 nothing new concerning this peculiar method of reproduction in insects was learned. Zavrel ('07) then reported paedogenesis in the genus Tanytarsus, and this has been confirmed for T. dissimilis by Johannsen ('10). In the meantime Kahle ('08) published perhaps the best account that has ever appeared on the 'Keimbahn' of any animal, using Miastor metraloas for this purpose. He was able to trace the germ cells from one generation to the next with remarkable clearness. Many of Kahle's results have been confirmed (Hegner, '12) for Miastor americana Felt and reference will be made to them more in detail in the following pages. Finally a short study of the life history of Miastor metraloas has been made by G. W. Muller('12).


388 ROBERT W. HEGNER

Metschnikoff's ('66) studies on Miastor indicated that the germ-cells of this fly are set aside very early in embryonic development, and led me in 1907 to attempt to obtain material of this genus. I was informed at that time by Prof. Samuel W. Williston that no paedogenetic Diptera were known to occur in this country. On October 5, 1910, however. Dr. E. P. Felt discovered great numbers of the larvae of Miastor americana Felt under the partially decayed inner bark and in the sapwood of a chestnut rail near Highland, New York, and kindly sent me an abundant supply of material.

Habitat and life history. Dr. Felt found the larvae of Miastor americana lining

. . . . in the moist, partly rotten inner bark and punky sapwood which has not been invaded to any considerable extent Ijy other Dipterous larvae or Coleopterous borers. They exhibit a tendency to occur in segregated masses, frequently between loose flakes of bark or in rather broad crevices. These colonies contain in autumn old empty skins of mother larvae; a number of yellowish mother larvae with approximately five to fifteen young within; very numerous, small yellowish larvae showing no trace of embryos; a number of white, various sized larvae, frequently white, sometimes semi-transparent;

and a few quiescent white larvae containing young embryos

The mouth parts of the larvae, though the anterior portion of the head

is strongly chitinized, appear to be comparatively weak

The ahmentary canal contains little that can be discerned with the aid of a compound microscope, and we are inclined to believe that a considerable portion of their nourishment is absorbed by osmosis after escaping from the mother larva, as well as before. It would appear as though the several types of larvae occurring in a colony are possibly only modifications due to the relative amount of nourishment obtained by the individual.

Normally, reproduction by paedogenesis occurs throughout the warm months of the year and even into late fall, and commences in early spring, the cold weather of winter simply causing a suspension

of activities The adults of Miastor and Oligarces occur

in midsummer, a season when midges of most of these forms are probably abroad (Felt, '11).

Methods. A number of fixing and staining methods have been employed in an endeavor to determine the origin and history of the 'polares Plasma' which plays such an important role in the differentiation of the germ cells of Miastor. Perhaps the easiest


STUDIES ON GERM CELLS 389

and most successful methods for studying the Keimbahn are fixmg in Gilson and staining in Mayer's acid hemalum followed by Bordeau-x red. Entire larvae may be fixed, sectioned and stained in this way. To bring out cytoplasmic details other methods were resorted to. The anterior and posterior ends of larvae were cut off and the middle part of the body containing the ovaries, and eggs and young were fixed in Meves' fluid. In other cases the eggs and young were dissected out and fixed in Meves' fluid. Still other larvae were fixed in Carnoy's solution. The best fixation was obtained with Gilson's Mercuro-nitric fluid. Besides acid hemalum the following stains were used: Heidenhain's iron hematoxylin followed by Bordeaux red or eosin, the iron hematoxylin method used by Rubaschkin ('12) in his studies of the mitochondria in the embryonic cells of the guinea pig, safranin followed by light green, Altmann's acid fuchsin differentiated in picric acid, Benda's method for the study of mitochondria, and the Erhlich-Biondi triple stain.

The morphological continuity of the germ cells. The ovaries of Miastor lie on either side of the body in the tenth or eleventh segment. They appear yellowish green in living white larvae, but are whitish transparent in the young yellowish larvae. Each ovary, when in the stage shown in figure 27, is surrounded by a thin cellular envelope (en) and contains typically thirtytwo oocytes (ooc.n) each with an accompanying group of mesodermal nurse cells (n.c), which are enclosed with it in the follicular epithelium (f.ep). The oocytes grow at the expense of the nurse cells, separate from the ovary, and are distributed throughout the body of the larva. Usually from five to seventeen embryos develop in one larva, but sometimes only one or two larvae are produced by a single mother-larva.

The nucleus of the oocyte (fig. 27, ooc.n) is large and clear, and the chromatin within it forms slender threads, rather evenly scattered about in the nuclear sap.

Figure 28 is a longitudinal section of an oocyte just before the maturation division. The germinal vesicle (g.v.) is large and clear and contains a great number of small scattered chromatin granules. It lies near one side of the oocyte in preparation for


390 ROBERT W. HEGNER

the formation of the first polar spindle. The nurse chamber (n.c.) contains a syncytium with about eighteen large nuclei, each of which possesses a large, centrally placed nucleolus surrounded by irregular chromatin granules. At the posterior pole of the egg is an accumulation of material (pPl) which stains deeply, and, as will be shown later, is intimately associated with the origin of the primordial germ-cell. This mass of material has been termed by Kahle 'polares Plasma'^ — a term adopted in the following description. A discussion of the origin and significance of the 'polares Plasma' will be reserved until later (p. 396).

There is nothing unusual in the process of maturation. The egg nucleus preparing for division is shown in figure 29, m.s. One polar body is formed in Miastor metraloas and the number of chromosomes could not be determined by Kahle but is probably from twenty to twenty-four. The number of chromosomes appears to be similar in M. americana but an accurate count could not be made. The first polar body divides by mitosis (fig. 30, p.b.) and the pronucleus (/.n.) moves over into the mass of cytoplasm (c), which is apparently elaborated by the nuclei in the nurse chamber, and becomes the cleavage nucleus. Here in this mass of cytoplasm the first cleavage takes place resulting in two apparently similar daughter nuclei (fig. 31, c.n.).

Reference must be made to Kahle 's figures and description for most of the events of early cleavage. The two nuclei of the two-cell stage divide by mitosis and the four daughter nuclei are apparently similar. They have been numbered i, ii, iii, iv, beginning at the anterior pole (fig. 3). The succeeding nuclear division is important, since the advance from the four-cell to the eight-cell stage witnesses the origin of the primordial germ-cell as well as a casting out (diminution) of chromatin by nuclei i, II, and III. In figure 3 nuclei i, iii and iv are shown in mitosis. Spindle iv is undergoing the ordinary process of mitosis, but spindles i and iii (and also spindle ii, which is not shown in the figure) are long and slender and a large portion of their chromatin (fig. 3, cMp) does not take part in the formation of the daughter nuclei but remains behind in the cytoplasm as a 'Chromosomenmittelplatte.' These masses of cast-out chromatin are never


STUDIES ON GERM CELLS 391

found in stages earlier than the four-cell stage, but are present in many of the later stages (figs. 32, 33, 34, 35, 36, cR) and are called by Kahle 'Chromatinreste.' One daughter nucleus resulting from the division of nucleus iv (fig. 3) becomes imbedded in the 'polares Plasma' and, with this substance, is cut off from the rest of the egg as the primordial germ-cell (fig. 32, p.g.c). The other daughter nucleus of spindle iv remains in the egg. These two nuclei are the only ones at this stage which contain a complete amount of chromatin.

During the next stage (viii-xv) the daughter nucleus of cleavage cell IV, which remains in the egg, undergoes a diminution process whereby it loses part of its chromatin, • and the other six nuclei within the egg pass through a second diminution process during which a second 'Chromosomenmittelplatte' is formed (figs. 32 and 4, cMp). At the fifteen-cell stage, therefore, one nucleus (that of the primordial germ-cell, fig. 32, p.g.c.) contains the full amount of chromatin; whereas all of the others (somatic nuclei) have lost a large portion of their chromatin. After the second diminution process, according to Kahle, the somatic nuclei possess only half the number of chromosomes present in the germ cells, that is "der Diminutionsprocess und der Reductionsprocess in derselben Karyokinese verenigt sind." My material did not contain enough of the early cleavage stages to enable me to confirm in detail Kahle's investigations, but one egg contained well marked mitotic figures which represent stages in the second diminution process (fig. 32) and a large number of sections were obtained which contained chromatin masses ('Chromatinreste,' figs. 32, 33, 34, 35, 36, cR). The details of the second diminution process are shown in figure 4.

The history of the germ-cells, from the time of the formation of the primordial germ-cell to the production of the sixty-four oocytes contained in the two ovaries, thirty- two in each, will now be described briefly.

The somatic nuclei divide rapidly, forming the blastoderm as shown in figures 33 and 34. Chromatin masses (cR.) representing chromatin cast out during the diminution processes are present in these early stages. The primordial germ-cell (fig. 32,

JOUR>fAL OP MORPHOLOGY, VOL. 25, NO. 3


392


ROBERT W. HEGNER




H-V"



^f^^r


Mp



Fig. 3 Miastor metraloas (redrawn from Kahle, '08). A section through an egg showing a dividing polar body (p.b); cleavage nuclei i and iii undergoing the chromatin-diminution process; and cleavage nucleus iv dividing normally. The daughter nucleus of the latter which enters the 'polares Plasma' {p. PI) becomes the nucleus of the primordial germ cell.

Fig. 4 Miastor metraloas (redrawn from Kahle, '08). Five stages in the second chromatin-diminution process. cMp, 'Chromosomenmittelplatte.'


STUDIES ON GERM CELLS 393

p.g.c.) divides by mitosis resulting in two oogonia (fig. 33, oogi) which he at the posterior end of the egg. Each of these divides again about the time when the blastoderm cells are cut off by cell walls. Four oogonia of the second order (fig. 34, oog^) are formed in this way. A third division results in the production of eight oogonia of the third order (fig. 35, 00^3). The germ-band then forms and segments, and the eight oogonia are passively carried around by the growth of the tail fold as shown in figure 36, ooQs). The embryo then grows broader and shorter until it entirely surrounds the yolk and the end of the tail fold coincides with the posterior end of the egg. During these developmental stages the oogonia remain undivided, but become separated into two groups of four each, which lie in two rows, one on either side of the embryo in the region of the eleventh segment (fig. 37, 00^3)Soon each row of four oogonia becomes enclosed by mesoderm cells, forming an ovary. The germ-glands then become almost spherical and soon the oogonia undergo a division by mitosis, thus forming eight oogonia of the fourth order in each germgland (fig. 38, ooQi). These divide again by mitosis (fig. 38, a) producing sixteen oogonia of the fifth order (fig. 39, 00^5) in each germ-gland. The final division of the oogonia takes place shortly before the larva hatches.

Typically, there are then thirty-two oogonia of the sixth order in each germ-gland, but in some cases certain of the oogonia of the fifth order are prevented from dividing. All of the oogonia do not produce embryos, since, as a rule, only from five to seventeen larvae are produced by a single mother-larva. ' The oogonia of the sixth order grow into oocytes (fig. 40, 00c.) ; each of these, together with a syncytium containing about twenty-four nursecells of mesodermal origin (fig. 27, n.c), becomes surrounded by follicular epithelium (fig. 27, f.ep.) also of mesoderm cells. During this process the nucleolus of the germ-cells disappears and the chromatin forms long slender threads (fig. 27, ooc.n). This completes the history of the germ-cells from one generation to the next. The accompanying diagram (fig. 5) shows graphically the germ-cell cycle in this animal.



PH"


Fig. 5 Miastor americana; diagram showing the entire germ cell cycle, cl.n., cleavage nucleus; ex.chr., extruded chromatin; oog, oogonium; p.b, polar body; p.g.c, primordial germ cell; p.o, primary oocyte; p. pi, 'polares Plasma'; st.c, stem-cell.

394


'STUDIES ON GERM CELLS 395

Summary. The principal points that should be emphasized are as follows:

1. Miastor americana Felt is truly paedogenetic and agrees with M. metraloas in regard to its reproduction as described by Kahle ('08).

2. One polar body, which divides by mitosis, is produced (fig. 30).

3. A diminution process takes place during the division of the first four cleavage nuclei in which a large part of the chromatin of three of these cleavage nuclei is cast out into the cytoplasm (fig. 3).

4. One daughter nucleus of the fourth cleavage nucleus, which does not lose any of its chromatin, passes into a deeply staining mass of material ('polares Plasma') situated at the posterior end of the egg, and is cut off from the egg by a cell wall. This cell which thus contains the ^polares Plasma' and a nucleus with the full amount of chromatin is the primordial germ cell (figs. 32, p.g.c).

5. A second diminution process takes place, during which each of the seven somatic nuclei loses part of its chromatin and emerges with one-half of the number of chromosomes. The primordial germ-cell does not undergo a diminution process (fig. 32).

6. The primordial germ-cell divides by mitosis until eight oogonia are produced. These separate to form two rows of four oogonia each. After a long period of rest further divisions result in the production of thirty-two oogonia in each germ gland.

7. The nurse cells are of mesodermal origin.

8. We have here for the first time a definite number of oogonial divisions, namely six, a definite and equal number of oogonia in each germ-gland, and a definite number of oocytes (sixtyfour) produced by the primordial germ-cell. It is no longer necessary, therefore, to express our ignorance by saying that there are n divisions during the period of multiplication of the oogonia, since in Miastor the number (n) is known positively to be six.


396 ROBERT W. HEGNER '

The differentiatio7i of the germ cells. We have seen from the foregoing account that in Miastor there are two features which distinguish the germ cells from the somatic cells; (1) the nucleus of the primordial germ cell is the only one in the egg which retains the full amount of chromatin and the complete number of chromosomes; and (2) the primordial germ cell contains, in addition to this nucleus, all of the 'polares Plasma' and apparently no other kind of material. In considering the differentiation of the germ cells we must therefore examine more in detail these two features.

Kahle does not discuss the origin of the 'polares Plasma.' In describing this structure in an oocyte just before the formation of the polar body, he says, Ganz auffallig ist eine Ansammlung von Protoplasma am hinterer Eipol. Sie wird durch Anilinund Karminfarben tiefer tingiert a^s das iibrige Plasma und macht den Eindruck einer ausserordentlich verdichteten Substanz" (p. 12). Further on the following statement is made:

Wie wir sahen, stammt der Kern der Urgeschlechtszelle in direkter Folge vom Furchungskern ab, das Protoplasma der Urgeschlechtsze le aber war schon lange vor ihm da. Es ist dasselbe, das ich als polares Plasma bezeichnet habe, das sich durch besonders intensive Farbung auszeichnet, dessen Vorhandensein in unveranderter Lage in alien aufeinanderfolgenden Stadien nachweisbar ist das im Fortgang der Entwicklung ein immer starkeres Wachstmn zeigt und sich l^is in die ungereifte Eizelle zuriichverfolgen lasst. Das polare Plasma ist infojge dessen also Keimplasma aufzufassen, das wahrscheinlich besondere Qualitaten enthalt und bereits in der ungereiften Eizelle praformiert wird, um spater seine Aufgabe als Geschlechtsplasma zu erfiillen. Erscheint uns also die Bildung der Urgeschlechtskerns schon als eine sehr frlihe, so ist che Differenzierung des Urgeschlechtsplasmas auf ein noch viel jiingeres Stadium verlegt. Dieses Plasma erwartet gewissermassen seinen Kern, um sich dann sofort mit ihm als Urgeschlechtszelle zu isolieren (p. 21).

If, as Kahle says, the 'polares Plasma' represents the 'Keimplasma,' it is of the greatest importance to determine its origin and fate. For this reason hundreds of young were preserved, sectioned and stained by the methods most likely to enable one to trace the history of this substance (p. 388). It must be confessed, however, that notwithstanding the efforts made with


' STUDIES ON GERM CELLS 397

this end in view, the problem is still unsolved. It is evident from the preparations that the oocyte is nourished by and grows at the expense of the nurse cells (figs. 28-32) . It is also absolutely certain that these nurse cells are not derived from the oogonia, as is true in so many insects, but, are modified mesoderm cells (figs. 38, 39, 27). At first, the growth of the oocyte takes place so slowly as to make almost no perceptible difference in the character of the cytoplasm contained within it. The oogonia are remarkably easy to distinguish during the embryonic development because (1) of their comparatively enormous nuclei, filled with large chromatin granules; and (2) the deeply staining quality of their cytoplasm, consisting of the corresponding deeply staining substance of the 'polares Plasma' of the mature egg (figs. 31-32). As the oocyte grows its cytoplasm becomes less deeply colored and presents a uniform appearance not distinguishably different from the cytoplasm of the other cells (fig. 40). Sometime before the oocyte is ready for maturation, however, deeply staining cytoplasm appears in the neighborhood of the nurse chamber and a substance begins to accumulate at the posterior pole which has a strong affinity for various dyes (fig. 28). The former is evidently elaborated under the influence of the nurse cells; the latter, which represents the 'polares Plasma,' may be derived from the nurse cells, but if it is, the process is so slow and its mass compared with the mass of the remaining egg contents so small that its passage from the nurse chamber to the posterior end of the oocyte is indistinguishable. We must conclude, therefore, that the 'polares Plasma' may originate from or under the influence of the nurse cells, but that this has not been demonstrated and probably never can be established.

A second hypothesis which may account for the presence of the 'polares Plasma' in succeeding generations is that of continuity and growth. Each oogonium is supplied with a portion (typically one sixty-fourth) of the 'polares Plasma' of the mature egg. Hence a certain amount of this substance, as well as a like amount of nuclear material, is passed on from one generation to the next. What is more probable than that this part, although


398 ROBERT W. HEGNER

minute when compared with the enormous contents of the mature egg, may become segregated at the posterior end of the egg and there bring about the development of a greater volume of similar substance, either by the division or budding of preexisting particles, or from the yolk or cytoplasm under its influence. A full discussion of this subject will reserved until the Keimbahndeterminants of other animals have been described (p. 460).

C. The Keimhahn in Compsilura concinnata Meig

Compsilura is a tachinid fly, introduced into this country in 1906 for the purpose of destroying gypsy and brown-tail moths. "Its eggs hatch in the uterus of the mother, and the tiny maggots are deposited beneath the skin of the host caterpillar by means of a sharp, curved 'larvipositor,' which is situated beneath the abdomen" (Howard and Fiske, 1912, p. 219). The maggot is ready for pupation in about two weeks; the pupal period is about one week; and the females require only about three or four days after their emergence to become sexually mature. I wish here to acknowledge my indebtedness to Dr. John N. Summers of the Gipsy Moth Parasite Laboratory, jVIelrose Highlands, Massachusetts, for an abundance of material.

The internal reproductive organs of a sexually mature female are shown in figure 6. Oocytes of various sizes are present within the ovarian tubules (o.). At a point near the union of the two oviducts (od) , the uterus is connected with two accessory glands {a.g.), and three seminal receptacles {s.r.). The mature eggs, which make their way down the oviduct and into the uterus, are here fertilized. They then gradually move down the uterus and are present to the number of about one hundred in a sexually mature individual. All stages from the maturation of the egg to the condition when the larva is ready to be deposited are passed through within the uterus of the mother, and most of these may be observed in a single specimen. Those eggs nearest the ovaries are of course the youngest. An attempt to trace the origin of the pole-disc granules in this species was unsuccessful, so only two illustrations are presented here to show that in this species


STUDIES ON GERM CELLS 399

there is a primary cellular differentiation similar to that already described in other Diptera. Figures 41 and 42 represent two stages in the formation of the primordial germ cells at the posterior end of the egg. The granules of the pole-disc are encountered by the cleavage nuclei which chance to reach the posterior pole; they surround these and are distributed about



Fig. 6 Compsilura concinnata; reproductive organs of a female, a.g, accessory gland; o, ovary; od, eviduct; s.r, seminal receptacle; u, uterus full of eggs in various stages of development.

within the cytoplasm of the germ cells when they are cut off by cell walls. The further history of the germ cells does not seem to differ in any way from that described in other Diptera.

Cecidomyia strobiloides. A pole-disc was also found in the eggs of the willow-cone gall just before deposition (fig. 43), and an attempt was made, as in Miastor, to trace this structure to its place of origin. The growth of the egg of this species resembles that of Miastor and all efforts to connect the poledisc with substances within the oocyte previous to its appearance at the posterior end were futile.


400 ROBERT W. HEGNER

3. COLEOPTERA

A . Historical account

The primordial germ cells of beetles have not received as much attention from investigators as have those of the Diptera, probably because they are not so conspicuous. The germ glands have been described in the embryo of Hydrophilus piceus by Heider ('89), in Leptinotarsa (Doryphora) decemlineata by Wheeler ('89), in Melolontha vulgaris by Voeltzkow ('89b), in Hydrophilus piceus, Melolontha vulgaris, and Lina tremulae by Graber ('91), in a number of Chrysomelid beetles by Lecaillon ('98), and by Friederichs ('06), in Tenebrio molitor by Saling ('07), and in Chrysomelid beetles by Hegner ('08, '09, '09b, '11a, 'lib) and Wieman (10a', 10b', '10c). Of these only the work of MTheeler, Lecaillon, Hegner, and Wieman needs to be considered. Although Wheeler ('89) failed to find the pole-cells in the very early stages of Leptinotarsa, he figures several of them (his fig. 82) in a sagittal section of an egg carrying a segmented germ-band. Here are shown three cells which are on the surface of the embryo in the amniotic cavity. They are very large and clear and the more anterior is apparently creeping in the manner of an Amoeba, along the surface of the abdominal ectoderm. These cells, the ultimate fate of which I have been unable to determine, probably escape from the anal orifice of the gastrula before it closes." This author also shows in transverse section (his fig. 87) a cell which, he says, is "about to wander through the blastopore into the amniotic cavity." He suggests that this may be the homologue of the Tolzellen.' That the cells thus described by Wheeler are really pole-cells was proved by my investigations on the same species.

The embryological development of the following species of Chrysomelidae was studied by Lecaillon ('98) ; Clytra laeviuscula, Gastrophysa raphani, Chrysomela menthastri, Lina populi, L. tremulae, Agelastica alni. In Clytra, the principal form examined, Lecaillon found that the first nuclei to arrive at the posterior pole of the egg became the centers of the primitive germ-cells; these could be distinguished from neighboring cells by their


STUDIES ON GERM CELLS 401

large size, larger nuclei, and more deeply staining cytoplasm. The germ-cells did not stop when they reached the surface of the egg, but passed outside and became separated from it; their number increased .... peu a peu par suite de Farrivee de nouvelles cellules peripheriques et aussi sans doute de la division des premieres cellules detachees du pole de I'oeuf." The germ-cells then started to re-enter the egg, retarding, by this migration, the formation of the blastoderm at this point. 'Tinalement, le blastoderme acheve de se former au pole posterieur de Toeuf, et alors les cellules sexuelles se trouvent groupees. . . . . entre le vitellus et I'enveloppe blastoderniique."

Several species of Chrysomelid beetles were also studied by Friederichs ('06), who discovered that the cleavage nuclei in Donacia crassipes reach the posterior later than the anterior end of the egg; the reverse is the rule in species of allied genera. After the blastoderm is formed "an der Ventralseite unmittelbar seitlich vor dem Pol, findet eine besonders lebhafte Zellvermehrung statt, so dass einzelne Zellen aus dem Blastodermverband heraus und ins Innere gedrangt werden." These, the primitive germ-cells, were not very different in Donacia from blastodermcells, but in Timarcha nicoeensis and Chrysomela marginata they could be distinguished by the larger size and darker color of their nuclei.

In a series of papers published within the last six years, the writer has given the results of morphological and experimental studies on the primordial germ cells of Chrysomelid beetles, particularly Calligrapha bigsbyana, C. multipunctata, C. lunata, and Leptinotarsa decemlineata. It has been possible to trace the entire Keimbahn in these insects, and to carry on experiments with the eggs and embryos without preventing further development. The reader is referred to the original papers for details, but a general account will be given here as an introduction to the. original work to be presented in the succeeding pages.

At the time of deposition, the eggs of the Chrysomelid beetles studied are not always in the same stage of development, although usually polar body formation is taking place. The egg


402 ROBERT W. HEGNER

figured (fig. 7, A) was fixed four hours after deposition. The polar bodies have already been produced and the male and female nuclei are in the act of conjugation. The egg consists of a large central mass of yolk and a comparatively thin peripheral layer of cytoplasm, the 'Keimhautblastem' of Weismann. The interdeutoplasmic spaces are filled with cytoplasm which is connected with the 'Keimhautblastem' by delicate strands of the same material. The enormous amount of yolk' contained in these eggs makes the identification of other substances extremely difficult. The yolk-globules range in size from large deutoplasmic spheres to small granules, and, as the dissolution of some of them is continually taking place, one is unable to determine where yolk ends and cytoplasm begins. The only accumulations of cytoplasm large enough for examination are those surrounding the nuclei within the yolk mass, and the peripheral layer, the 'Keimhautblastem.' No differences in composition or staining qualities were observed between the cytoplasm of these two regions. The 'Keimhautblastem' consists of a fluid ground substance in which are suspended very fine granules. It is a homogeneous layer of cytoplasm everywhere except at the posterior end of the egg. At this point there is a disc-shaped mass of larger granules imbedded within the inner portion of it. These granules stain deeply with haematoxylin. They are easily seen, not only in sections but also in eggs that have been properly stained in toto. Because of their ultimate fate I have called these granules the germ-cell determinants (fig. 7, A, g.c.d.).

The first cleavage divisions take place where the pronuclei fuse. The daughter nuclei move away from each other and as cleavage progresses a separation of the nuclei into two sections occurs. The nuclei of one group form a more or less regular layer equidistant from the periphery; these preblastodermic nuclei (fig. 7, B, jM.n) move outward and fuse with the Keimhautblastem. Cell walls now appear for the first time and a blastoderm is formed of a single layer of regularly arranged cells.

The genesis of the pole-cell is as follows: (1) four nuclei lying near the posterior end of the egg are recognized by their


STUDIES ON GERM CELLS


403




, , ¥1


>-■»- *.


Fig. 7 Calligrapha (from Hegner, '09a and '09b). A, longitudinal section through an egg of C. bigsbyana, four hours after deposition. B, longitudinal section through an egg of C. bigsbyana 14 hours after deposition. C, two germ cells just protruding from posterior end of egg of C. multipunctata. D, the poledisc in an egg of C. multipunctata. g.c.d, pole-disc; g.n, germ nuclei fusing; khbl, keimhautblastem; p, posterior end of egg; pbl.n, preblastodermic nuclei; v.7n, vitelline membrane; rt, vitellophags; y, j'olk.


404 ROBERT W. HEGNER

position as pole-cell antecedents; (2) these four nuclei divide producing eight daughter nuclei which move closer to the periphery of the egg; (3) these in turn divide resulting in sixteen nuclei, arranged in pairs, each of which separates entirely from the egg, carrying with it a portion of the Keimhautblastem containing pole-disc granules (fig. 7, C) ; (4) the sixteen primary pole-cells divide to form thirty-two secondary pole-cells; these divide resulting in sixty-four tertiary pole-cells which do not increase in number until a late period of embryonic life; (5) in mitosis the pole-disc granules are approximately equally distributed between the two daughter cells (fig. 8, B). After separation from the egg the pole-cells are (1) carried slightly forward on the ventral surface of the egg by the contraction of the ventral plate; (2) they sink into the posterior depression of the ventral groove, which is the beginning of the posterior amniotic cavity; (3) they are carried along by the developing tail-fold, which penetrates dorso-anteriorly into the yolk; (4) they migrate through a pole-cell canal into the embryo by means of amoeboid movements; (5) upon reaching the interior of the embryo they separate into two groups, which come to lie, one on either side of the body, in the last two abdominal segments; (6) these two strands become shorter by a crowding together of the germ-cells; (7) each of the two germ-glands thus produced acquires an epithelial covering of mesoderm-cells; (8) the germ-glands, situated as before in the last two abdominal segments, are carried, by the shortening of the embryo, to a ventral position on either side of the body; (9) by its lateral growth around the yolk, the embryo carries the germ-glands to a point near the dorsal surface on either side of the mid-gut; (10) the sexes can be distinguished at this time by the shape of the germglands, that of the male being dumb-bell shaped, while the female reproductive organ is pear-shaped, and shows the development of terminal filaments.

In all stages the germ cells may be distinguished easily from the surrounding somatic cells. Figure 8, A shows a pole cell shortly after separation from the egg. The pole-disc granules are quite conspicuous, and pseudopodia-hke projections are


STUDIES ON GERM CELLS


405



-pd-g.



R



-gca



H


PSC




X- p

I

Fig. 8 Calligrapha (from Hegner, '09a, '09b, '11a). A, a germ cell of C. multipunctata shortly after being cut off from the egg. B, division of a primordial germ cell. C, longitudinal section through egg of C. bigsbyana at blastoderm stage; the posterior end was killed with a hot needle just after deposition. D, longitudinal section through uninjured egg at same stage. E, two ectoderm cells (e), two mesoderm cells (m), and two germ cells (g.c.) from an egg three days old. F, germ cell during migration into the embryo (three days old). G.H.I, longitudinal sections through eggs centrifuged for one hour, two hours and four hours respectively, bl, blastoderm; g.c.d, granules of pole-disc; A-, killed portion of egg; khbl, keimhautblastem; p, posterior; pgc, primordial germ cells; v, vitellophags; t'.?, vesicular zone; y, yolk.


406 ROBERT W. HEGNER

plainly evident. Sixteen pole-cells were present at this time. After a mitotic division, during which the pole-disc granules are apparently approximately equally divided between the daughter cells (fig. 8, B), the pole cells are smaller, but, although no larger than the neighboring blastoderm cells, they may still be distinguished by the presence of the pole-disc granules and also by the larger nucleus containing a lesser number of chromatin granules. Pole-disc granules are still faintly visible at a later period when the germ cells are migrating into the embryo through the pole-cell canal, and still later, as figure 8, E shows, the germ cells can be distinguished easily from the ectoderm and mesoderm cells although the pole-disc granules have entirely disappeared.

The pole-disc. The pole-disc varies somewhat in compactness but in most cases appears in section as shown in figure 7, D. Nothing resembling it occurs in other parts of the egg. Its granules are very susceptible to stains and can be made visible by means of a number of different dyes. The 'Keimwulst' of Chironomus (Ritter '90), the 'Dotterplatte' of Calliphora (Noack, '01; fig. 2) and the 'Keimbahnplasma' of Chironomus (Hasper '11; fig. 1) all present a similar appearance. In these forms, as well as in the Chrysomelid beetles I have studied, all or nearly all of the granules (fig. 7, C) are taken out of the egg by the polecells. Wieman ('10a) however, gives a figure showing a section of the posterior end of the egg of Leptinotarsa signaticollis after the protrusion of the pole cells, in which there is still represented what he calls the pole-disc. The fact that the mass of granules described by Wieman does not resemble the pole-disc as I have found it, nor other similar accumulations in insect eggs (Keimwulst, Dotterplatte, Keimbahnplasma) and the statement that "the grarmles are not all taken up by the cells in their migration and the greater part of them remains behind after the cells have passed through" (Wieman, '10a, p. 186), a condition contrary to that described by every one of the writers cited above, lead to the conclusion that Wieman has confused something else for the pole-disc. This seems all the more probable, since the species studied by Wieman, namely Leptinotarsa signaticollis, is very


STUDIES ON GERM CELLS 407

closely allied to one of the species that I investigated (L. decenilineata) in which a typical pole-disc like that shown in figure 7, D occurs. Furthermore the cells which Wieman designates as pole cells have none of the characteristics of the pole cells described by other writers.

Several important results have been obtained by experiments that I have performed with the object of determining the character and significance of the pole-disc. When the freshly laid eggs of Leptinotarsa decemlineata are centrifuged with the posterior end toward the center of revolution the pole-disc is moved gradually toward the outer anterior end as shown in figure 8, G,H ,1, g.c.d. The movement en masse of the pole-disc granules proves that they are heavier than the oil globules of the vesicular zone (v.z.) and indicates that they do not form an adventitious accumulation but constitute a definite structure of sufficient rigidity to withstand the dispersing effects of a strong pull exerted during a period of at least four hours. It was hoped that by means of centrifugal force the pole-disc could be located in a part of the egg different from that normally occupied and that experimental proof of the necessity of their presence for the formation of germ cells might thus be obtained, but the abnormal development of the eggs prevented an accurate determin