Cornell University Library Dthaca, New Pork BOUGHT WITH THE INCOME OF THE SAGE ENDOWMENT FUND THE GIFT OF HENRY W. SAGE 1891 | v.2 | nN psc Lal gsto¥ an Woe Ea. 2 ity Library fin 05 Cornell University Library The original of this book is in the Cornell University Library. There are no’ known copyright restrictions in the United States on the use of the text. http://www. archive.org/details/cu31924024787057 TEXT-BOOK OF EMBRYOLOGY MACMILLAN AND CO., LimiTED LONDON - BOMBAY CALCUTTA MADRAS MELBOURNE THE MACMILLAN COMPANY NEW YORK + BOSTON - CHICAGO DALLAS SAN FRANCISCO THE MACMILLAN CO, OF CANADA,, Lr. TORONTO TEA T-BOOK OF EMBRYOLOGY VOL. II VERTEBRATA WITH THE EXCEPTION OF MAMMALIA BY J. GRAHAM KERR, REGCIUS PROFESSOR OF ZOOLOGY IN THE UNIVERSITY OF GLASGOW MACMILLAN AND CO, LIMITED ST. MARTIN’S STREET, LONDON 1919 pe A 45a yo. COPYRIGHT TO THE MEMORY OF THREE CAMBRIDGE NATURALISTS CHARLES DARWIN, M.A., Curist’s CoLLEGE FRANCIS BALFOUR, M.A., TRINITY COLLEGE ADAM SEDGWICK, M.A., Triniry CoLiecy PREFACE THE object of this volume is to sketch in its main outlines the science of Vertebrate Embryology as disclosed by the study of the non-mammalian vertebrates. It is not meant as a work of reference as regards details. The facts of embryology are dealt with as illustrating general principles: large masses of data which have no particular bearing, in the present condition of knowledge, are deliberately omitted. It is believed that a volume upon the lines indicated is greatly needed—not merely for students intending to specialize in vertebrate morphology but also for students of medicine who desire to know something of the framework of morphological principles which serves to unite together the detailed facts of anatomy. The science of embryology, in fact the science of animal morphology as a whole, has suffered much through the patient but undiscriminating accumulation of masses of mere descriptive detail which have tended to obscure general principles and incidentally to smother interest in one of the most fascinating of sciences. It is hoped that the student who reads through portions of this book will have at least his suspicions aroused that behind the dull facts of anatomical structure there lies a very charming and living philosophy. It has again been one of the misfortunes of vertebrate embryology that its teaching has been dominated in great part by general ideas based upon insufficient data. In an evolutionary science like morphology the real fundamental principles are to be elicited by enquiry into the more archaic types of existing animal life. But the material for the earlier embryological investigations was chosen not for its archaicisim but rather for purely practical reasons such as vii viii EMBRYOLOGY OF THE LOWER VERTEBRATES accessibility or ease of investigation. It follows that at the present time when we have knowledge of the more archaic subdivisions of the vertebrata not accessible to the early builders of the science, it is necessary to regard the historical foundations of vertebrate embryology rather critically in the light of the fuller knowledge of to-day. In essaying the writing of this volume I have been fortunate in having at my disposal—for the first time in the history of embryology—developmental material of all three genera of Dipnoi as well as of Polypterus—in addition to the more accessible material of the other relatively archaic groups constituted by the Elasmo- branchs, Actinopterygian Ganoids, and Urodele Amphibians. This has rendered possible an all-round survey of the chief problems of vertebrate embryology which would otherwise have been quite impossible. As already indicated I do not intend this volume as a work of reference on the details of vertebrate embryology: that rdle is fulfilled by the wonderful and indispensable Handbuch edited by O. Hertwig—of which incidentally I have made constant use and to which I must express my sincere acknowledgments. Nor do I attempt to give full historical accounts of the development of various parts of the subject. The literature lists are merely guides to point the way to the student who desires to extend his reading to original papers. The dates given in these lists are as a rule the dates given on the title-page of the complete volume, and are merely to facilitate finding the particular paper: they must not be taken as giving the actual date of publication of the individual memoir. I have to express my grateful thanks to various friends. As regards the first three chapters I had the benefit of the wise counsel of Mr. Walter Heape, who unfortunately however found himself com- pelled by the exigencies of war work to withdraw from the Editorship. Various chapters have benefited by the help and advice of my friend and colleague, Dr. W. E. Agar. The entire volume has been read in proof by Mr. James Chumley and Dr. Monica Taylor, to both of whom I am deeply indebted. To Dr: John Love and to Dr. Jane Robertson I am indebted for helpful criticism in regard to special sections of the book. The illustrations which form a marked feature of the volume I PREFACE ix owe for the most part to the artistic skill, combined with high scientific accuracy, of Mr. Kirkpatrick Maxwell. Apart from the completely original figures it will be- noticed that there are many which have been worked up from illustrations in original papers, but which are practically new figures. In all such cases, however, T have thought it only right to make due acknowledgment of the author of the original figure. . For permission to borrow particular text-book figures I have to thank Professor Frank R. Lillie, Mr. John Murray, Messrs. Masson & Cie, and Messrs. Macmillan. The present unfortunate circum- stances of international strife call for a special acknowledgment of the generous way in which Protessor Alf. Greil entrusted to me the originals of his valuable unpublished figures illustrating the development of the heart in the bird. They are reproduced on pages 384 and 385. I have included the name of Charles Darwin in the dedication of this volume to emphasize the fact that Embryology is primarily a branch of synthetic evolutionary science. While the fashion of the day in evolutionary research favours rather experimental research into the phenomena of inheritance and more or less speculative enquiry into the ultimate mechanism of inheritance or into the possible causes of evolutionary change—morphology, and more especially embryology, is steadily at work all the while, mapping out the paths along which the evolution of organisms and their con- stituent organs has taken place, Working away in comparative seclusion, unadvertised, and for the most part unnoticed, embryology is thus building up an important part of the framework of what will be the permanent edifice of evolutionary science. J. GRAHAM KERR. February 3, 1919. CONTENTS CHAPTER I SEGMENTATION, GASTRULATION, AND THE ForMATION oF THE GERM LayErs CHAPTER II THE SKIN AND ITs DERIVATIVES CHAPTER III Tae ALIMENTARY CANAL CHAPTER IV THE CoELOMIC ORGANS CHAPTER V THE SKELETON CHAPTER VI VASCULAR SYSTEM CHAPTER VII Tue ExTeRNAL FEATURES OF THE Bopy CHAPTER VIII ADAPTATION TO ENVIRONMENTAL CONDITIONS DURING EARLY STAGES or DEVELOPMENT xi PAGE 69 144 197 288 360 429 455 xii EMBRYOLOGY OF THE LOWER VERTEBRATES CHAPTER IX PAGE SoME OF THE GENERAL CONSIDERATIONS RELATING TO THE Empry- OLOGY .OF THE VERTEBRATA a 2 . 484 CHAPTER X THe Practican Stupy oF THE EmpBryoLocy or tHE Common Fown 508 CHAPTER XI HINTS REGARDING THE PRactTicaL StupyY oF THE EMBRYOLOGY OF THE Various Types or LowER VERTEBRATES . 558 APPENDIX Tare GeneraAL MrtrHops or EMBrYOLOGICAL RESEARCH 573 INDEX : 583 CHAPTER I SEGMENTATION, GASTRULATION, AND THE FORMATION OF THE GERM LAYERS THE Vertebrate begins its individual existence in the form of a single _cell, the Zygote or fertilized egg, which in turn originates in the process of fertilization by the fusion or conjugation of two gametes.! Of these the microgamete or spermatozoon, derived from the male parent, is of relatively insignificant bulk as compared with the macrogamete or unfertilized egg. As a consequence the more obvious features of the Zygote, such as shape, size, and so on, are simply taken over from the macrogamete—in other words, they are of maternal origin. Such maternal features may remain obvious for some time during early stages of development, so long in fact as the maternal protoplasm remains predominant in bulk as compared with that elaborated under the control of the Zygote nucleus, but it seems unnecessary to assume that this fact has the important bearing upon questions connected with Heredity which has been claimed for it by some workers on Invertebrates. / The Zygote is a typical cell, composed, so far as its living substance is concerned, of cytoplasm and nucleus, the cytoplasm containing a lesser or greater amount of food-material or yolk. In shape it is in the vast majority of cases approximately spherical. In the Myxinoids it is elongated, almost sausage-shaped, and in a certain number of cases, for example Amia, its shape is literally “ oval.” The macrogamete—and therefore the Zygote—differs much in size in different Vertebrates, ranging from about ‘1 mm. in diameter in Amphiozus to as much as 85 mm., or more, in the, case of the African Ostrich. In some of the Sharks the size of the Zygote is also very great and this was doubtless the case too, with that of such extinct birds as Aepyornis.2 Such relatively huge Zygotes are of 1 A general account of the processes of gametogenesis and fertilization has already been given in Vol. I. and they are not further dealt with in this volume. : 2 Assuming that the Zygote of Aepyornis bore the same ratio in size,to its protective envelopes as does that of the Ostrich it would measure about 160 mm. in diameter. VOL. IL B 2 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. interest as being in bulk the largest single cells known in either the Animal or the Vegetable Kingdom. The adjoining list based upon data obtained by various observers —M‘Tntosh and Masterman, Bashford Dean, Boulenger, Budgett, Bles, Semon, Salensky and others—brings out in more detail the ditferences in the size of the Zygote amongst the lower Vertebrates. APPROXIMATE DIAMETER OF EGG (IN MILLIMETRES) OF VARIOUS FISHES AND AMPHIBIANS Amphioxus +1. Bothus maximus 1. Petromyzon 1. Pleuronectes platessa 1-8. Bdellostoma 14-29 x 7 — 10-5. P. microcephalus 1-3. Pristiurus melanostomus 16. P. limanda -75. ) Acanthias 35 — 40. Solea vulgaris 1-2. » Japanese Lamnid? (Doflein) 220. Clupea harengus -9— 1. ' Torpedo ocellata 20 - 25. C. sprattus -1. Polypterus 1-1. Ceratodus 2-8. Acipenser 2-2-8. Protopterus 3-5 - 4. Lepidosteus 3. Lepidosiren 6-5 - 7. Amit 2.5-3x 2-2-5. Axolotl 2. Boccus labrax 1-4. Salamandra maculosa 4. Mullus surmuletus +9. Triton 1-8. Cottus scorptus 1-75. Necturus 6. Trigla gurnardus 1-5. Hypogeophis alternans 4 — 5. Agonus cataphractus 1-8. HT, rostratus 7 - 8. Trachinus vipera 1-3. Xenopus laevis 1-5. Scomber scombrus 1-2. Pipa 6-7. Gobius minutus 1— 1-4. Alytes obstetricans 3 - 5, Cyclopterus lumpus 2-5. Pelobates fuscus 1-5. Anarrhichas lupus 5.5 - 6. Bufo lentiginosus 1, Pholis gunnellus 1-7. Hyla goeldic 4. Gadus collaris 1-4. Nototrema fisstpes 10. G. aeglefinus 1-4. Paludicola fuscomaculata 1. G. virens 1-1. Engystoma ovale 1-25. Motella mustela -7. Cornufer salomonis 5. Brosmius brosme 1:3. Rhacophorus reinwardtti 3. Ammodytes lanceolatus -76, Rana temporaria 2. Hippoglossus vulgaris 3-3-8. RR. opisthodon 6 — 10. Within the limits of a single genus different species may show marked differences in the size of their eggs, e.g. the Teleostean fish Arius australis has eggs a little over 3 mm. in diameter (Semon) while in the case of A. boekiz they measure over 10 mm. in diameter and in A. commersonit as much as 18 mm. Even within the limits of a single species quite measurable, though less conspicuous, differences in size exist between the eggs of different females, and the same holds also, though to a far less extent, for the individual eggs laid by a single female. The differences in size which have just been alluded to are correlated with the fact that the egg of the Vertebrate carries in its cytoplasm a less or greater amount of reserve food-material or yolk. The presence of a readily available supply of food within the egg carries with it the immense advantage of freeing the young 5 I YOLK 3 animal, during the early stages of its development, from the need of having to fend for itself. And, correlated with this, the necessity of developing more or less complicated adaptive fea- tures to fit it for survival as an independent free- living creature in these early stages is avoided. The yolk consists occa- sionally of fluid but more usually of rounded or | sub-angular granules of | highly nutritious material. The yolk granules are frequently of a character- istic colour, yellow or salmon colour or greenish, and these impart their colour to the egg as a whole. ae Where however the yolk A.—Section through egg of Amphioxus. becomes very finely subdi- After Cerfontaine, 1906.) vided we find, asin the case The small crosses mark the position of apical and abapical of ectotred “glnsg grunt) Se os meeiesl is 2 eerie ae ae , into powder, that the char- nuclei have not yet fused, and are seen in close proximity|to acteristic colour is replaced one another. by white. This fine subdivision of the , yolk with its accom- , panying white colour is commonly found in parts of the egg where metabolism is particularly active, for example those portions in which active growth or cell division is about to take place, the fine subdivision making the yolk readily assimilable and so available for meta- B.—Section through egg of Polypterus, showing more marked , i; di Wh tendency for the yolk granules to concentrate towards bolic needs. ere the abapical pole (Telolecithal condition). the yolk a coll n, nucleus. p, pigment. paratively small in The yolk granules are indicated in both figures by dark dots. amount, as in Am- —_ phioxus (Fig. 1, A), it may be distributed nearly equally through- out the egg substance; in other words there is an approach 4 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. to the isolecithal condition: but as a rule in the Vertebrate the yolk is large in amount and is concentrated towards the lower or abapical pole of the egg, the protoplasm towards the upper or apical pole being comparatively poor in yolk (Telolecithal condition). This segregation of the dead yolk and the living protoplasm towards opposite poles of the egg is well seen in the relatively huge ege of the bird where the protoplasm is concentrated in a germinal disc containing practically no yolk and forming a cap at the apical pole of an enormous mass of yolk practically free from protoplasm. It has already been indicated that the egg may have a character- istic coloration due to the colour of the yolk. Such yolk coloration may be looked upon as accidental and without any special biological significance in itself. Many eggs on the other hand especially amongst the Ganoid fishes and the Amphibians are given a dark colour by the presence within them of brownish-black pigments belonging to the melanin group. Such pigment appears to be of definite biological significance, providing as it does an opaque coat which protects the living protoplasm from the harmful influence of light. Eggs in which it occurs develop, as a rule, under conditions where they are exposed to intense daylight. The eggs of ordinary Frogs and Toads for example which are surrounded by clear trans- parent jelly have a well-developed pigment coat. On the other hand in the case of Frogs and Toads whose eggs are surrounded by light- proof foam (see Chapter VIII.) or are deposited in burrows under- ground they are commonly without pigment. In all probability this deposition of melanin pigment in the superficial protoplasm of the egg (normally in its wpper portion) is to be interpreted as having been originally a direct reaction to the influence of light, the metabolism being so affected as to bring about the formation of this particular iron-containing excretory pigment. It may be objected that the pigment is produced before the egg is laid (e.g. the Common Frog) and therefore before it is exposed to the action of light, but as a matter of fact the body-wall of the adult is by no means opaque to light rays and even while still in the ovary the eggs are exposed to the influence of faint light. If we may take it, as seems probable, that the influence of natural selection has gradually developed in such cases the particular type of sensitiveness to light which leads to the formation of melanin, on account of its protective value, then there is nothing surprising in the developing of this sensitiveness at earlier and earlier periods until at last it has resulted in the pigmentation of the still intra-ovarian egg in response to the feeble light rays which penetrate the body-wall. The other possible explanation of this precocious pigment forma- tion is that the production of the pigment though originally taking place as a direct reaction to light in the laid egg, has become so engrained in the constitution of the species that it now comes about even in the absence of the original stimulus. The objection to this explanation is that it postulates the inheritance of an “acquired I SEGMENTATION 5 character,” and that is unfortunately not justified by our knowledge so far as it goes at present. SEGMENTATION The first important steps in the evolution of the unicellular Zygote into the multicellular adult are seen in the process of Segmentation which is, in fact, a process of mitotic cell division showing special peculiarities in different groups of the Vertebrata. ‘During this process there appear in succession on the surface of the egg grooves which gradually deepen and eventually divide the egg incompletely or completely into dis- = tinct segments or Blastomeres. se Before entering into the details of this process it will be convenient to describe it in outline and define the various technical terms used in its description. The first phase of segmentation is commonly marked by the appear- ance of a superficial groove which ay may conveniently be designated by — the letter a, passing through both the upper and lower poles of the egg. Such a groove or furrow is termed meridional, as it marks @ great circle on the surface of the i egg corresponding to a meridian of abp longitude on a terrestrial globe. Fic. 2.—Diagram to illustrate technical The single nucleus of the Zygote terms used in describing the process of meanwhile divides by mitosis—a segmentation. daughter nucleus passing into each — a», apical pole ; ab.p, abapical pole ; e, equa- hemisphere. From the known facts, torial furrow i 1, latitudinal furrow ; m, meri- ‘ees % dional furrow ; ¥, vertical furrow. of fertilization we have reason to believe that the Zygote nucleus contains exactly equivalent amounts of chromatin from each of the two parents. In the process of mitosis this maternal and paternal chromatin is again shared equally between the two daughter nuclei. The first meridional furrow gradually deepens so that the egg becomes completely divided into two blastomeres or segments each representing a hemisphere of the Zygote. A second meridional furrow (f) now appears in a plane perpendicular to that of the first and by the deepening of this the egg becomes divided into four equal blastomeres. The next furrow to appear may be one running round the equator of the egg (equatorial). In eggs, however, which are not absolutely isolecithal—and this holds for all the lower Vertebrates— 6 EMBRYOLOGY OF THE LOWER VERTEBRATES cH. the third furrow appears, not at the equator but at a level nearer the apical pole, and is termed a latitudinal furrow, corresponding as it does with a parallel of latitude upon a terrestrial globe. The distance of this furrow from the equator, its degree of latitude so to speak, is roughly proportional to the degree of telolecithality of the particular egy, suggesting that the volume of living protoplasm may be roughly equivalent in amount upon the two sides of the division plane to which this furrow gives rise. When this third division is completed the egg consists of eight blastomeres, the four on the apical side of the division plane being smaller (micromeres) than those on the abapical side (macromeres). The next furrows to appear are two in number and in the simplest condition they are meridional, bisecting the angles between the two first furrows. More frequently however those furrows instead of traversing the pole of the egg are discontinuous at this point and each is displaced somewhat so as to join the first or second meridional furrow at a less or greater distance away from the pole. To such a furrow we apply the term vertical (Fig. 2, v., cf. also Figs. 14 and 16, C). It is as a rule noticeable that meridional or vertical furrows tend to become apparent first in their portions nearest to the upper or apical pole of the egg, their lower ends gradually extending down- wards towards the abapical pole. This phenomenon appears to be due to the retarding influence of the dead and inert yolk. The proportion of this to the living protoplasm becomes greater and greater as the distance from the apical pole is greater, and in correlation with this the retarding effect becomes more and more pronounced. After segmentation has reached the stage indicated its further progress tends to become irregular. New furrows make their appearance—latitudinal, and vertical or meridional—and the surface of the egg. takes on the appearance of a mosaic-work, while its substance becomes cleaved or split apart into corresponding blasto- meres as the superficial furrows gradually deepen into slits. At somewhere about this period there begins a new type of mitotic division in which the individual blastomere becomes split in a direction parallel to a plane tangential to its outer surface, so that it divides into an outer blastomere visible in surface view and an inner one concealed in the interior of the egg. With the further progress of segmentation the blastomeres divide over and over again, so that eventually the egg is converted into a very large number of small cellular elements. The rapidity with which the cells divide bears a rough inverse relation to the richness of their contents in yolk. Dead inert yolk tends to cause the cell to lag behind in the process of division, and the result of this in telolecithal eggs is that the difference in size between micromeres and macromeres becomes more and more marked as segmentation goes on—the lower and more richly yolked segments I SEGMENTATION i tending to lag, in their mitotic division, more and more behind the less yolky upper elements. This inequality is found at its maximum in the large eggs of Elasmobranchs, Reptiles, and Birds, where the main mass of the egg has its proportion of protoplasm reduced so nearly to vanishing point that it does not divide at all. It is only a small portion of the egg in the neighbourhood of the apical pole that is rich enough in protoplasm to carry out the process of segmentation into separate cells. This is known as the germinal disc or, later on, when it has segmented into a mass of cells, blastoderm. An egg of such a type, showing partial or incomplete segmentation, is termed meroblastic in contrast with the more primitive holoblastic type in which the egg segments as a whole. The blastomeres into which the ege divides being composed of protoplasm—a somewhat viscous fluid—tend under the physical laws of surface tension to assume a spherical shape except when flattened by pressure against their neighbours. There thus exist normally chinks between the blastomeres filled with watery fluid. As the process of segmentation proceeds this intercellular fluid increases in amount and the process normally culminates in the stage known as the blastula. The blastula consists of a more or less spherical mass of cells surrounding a relatively considerable volume of fluid which is for the most part no longer distributed in small chinks but collected together into a large space—the blastocoele or segmentation cavity. In the simplest case, that of Amphiowus, the wall of the blastula is composed of a single layer of cells—the cells towards one pole being larger and containing fine granules of yolk or food material. In holoblastic Vertebrates above Fishes it is however, as a rule, no longer composed of a single layer, the roof of the segmentation cavity being frequently composed of two layers while the floor is composed of a thick mass of large heavily yolk-laden cells. The details of the segmentation process may now be followed out as it occurs in the various types of lower Vertebrates. AMPHIOXUS Amphioxus is, of all the lower Vertebrates, that in which developmental processes are least interfered with by the presence of yolk, and for this reason the phenomena shown during its seg- mentation must form the basis for the comparative study of the corresponding phenomena in the Vertebrata in general. ; The process of segmentation in Amphioxus was described first in two works which are now amongst the classics of morphological science: the first by A. Kowalevsky (1867) and the second by - B. Hatschek (1881). ; The process begins (Fig. 3) with the appearance of a depression of the surface in the region of the apical pole. This depression takes an elongated groove-like form and extends outwards at each 8 EMBRYOLOGY OF THE LOWER VERTEBRATES CH. end until finally it forms a wide meridional valley encircling the entire egg (Fig. 3, B). This valley gradually deepens dividing the egg into two halves. Finally after about 5 minutes from the commencement of the process the protoplasmic bridge connecting the two halves snaps across and the egg is now completely divided Fic, 3.—Illustrating the process of segmentation of the egg of Amphioxus. (After Hatschek, 1881.) The apical pole is above in each figure. The second polar body.is seen in proximity to it. into two blastomeres, each of which assumes a spherical shape in response to surface tension. The two blastomeres become slightly flattened where they are in contact «ein the plane of the first meridional furrow (Fig. 3, C). The future course of development shows that this plane corresponds to the sagittal plane of the embryo (Cerfontaine, 1906): in other words the two blastomeres represent the right and left halves of the developing individual. I SEGMENTATION 9 After a resting period of about an hour a second meridional furrow develops in a manner similar to the first and in a plane perpendicular to the plane of the first furrow. This gradually deepens and each hemisphere becomes divided into two blastomeres, each of which as before assumes a spherical shape and then becomes flattened out slightly against the other. Of the four blastomeres which are now present two, shown by subsequent development to be anterodorsal in position, are according to Cerfontaine normally smaller than the other two. The two meridional furrows (a4 and #) are followed after an interval of about a quarter of an hour by.a latitudinal furrow slightly above the equator and this divides each of the four segments into two. The egg now consists of eight blastomeres—four smaller micromeres on the apical side of the latitudinal division plane, and Fic. 4.—Apical view of Amphioxus eggs at the eight-blastomere stage. (After E. B. Wilson, 1893.) A, ‘‘Radial” type; B, ‘‘Spiral” type; and C, ‘“ Bilateral” type. four larger macromeres upon its abapical side. Each micromere lies, according to Hatschek, exactly over the corresponding macromere so that the apical side of the egg as seen from above looks like A in Fig. 4. Wilson (1893), followed by Samassa (1898), has however drawn attention to the fact that in a considerable proportion of cases the cap of four micromeres is, as seen from above, rotated in a clockwise direction through an angle varying from 0° to 45° (Fig. 4, B) thus conforming to Wilson’s “spiral” type of segmentation or cleavage. Again in a still smaller percentage of eggs at this stage the blasto- meres are arranged according to Wilson’s “ bilateral” type (Fig. 4, C) the eight blastomeres being arranged symmetrically on each side of the first division-plane but either two or all four macromeres being displaced outwards somewhat from this plane. Fourth division. After another short interval (less than a ‘quarter of an hour) a new set of furrows appear bisecting each of the already existing blastomeres so that the embryo comes to consist of sixteen blastomeres arranged in two tiers, eight micromeres above and eight macromeres below (Fig. 3, F). Hatschek described this fourth set of furrows as being meridional (Fig. 3, F) while according to Cerfontaine (1906) the division planes are when first 10 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. indicated meridional in position but become displaced somewhat so as to be in the case of the micromeres perpendicular to the first (sagittal) division-plane or in that of the macromeres slightly oblique. Fifth division. Each blastomere divides again, the smaller blastomeres towards the apical pole dividing rather earlier than the others, and the result is that there are now thirty-two blastomeres in all, arranged in eight meridional rows of four cells each, the cell at the lower (abapical) end being decidedly larger than the others. Between these four large elements a wide opening is present (Fig. 3, G) leading into a space which made its appearance as a little chink between the blastomeres of the four-cell stage but which has since then increased greatly in size. This space in the interior of the egg is the blastocoele or segmentation cavity. From this period onwards the segmentation process becomes less regular. There has already shown itself a tendency for the larger blastomeres towards the lower pole to lag behind somewhat. And the arrangement of the blastomeres becomes less regular as they become smaller and fit more closely together. In particular the bilateral symmetry in the arrangement of the blastomeres which is conspicuous in most of the eggs during the earlier stages (Cerfontaine) ceases to be apparent. To summarize the later phases of segmentation it may simply be said that the blastomeres go on dividing, the segmentation cavity increases in size, its communication with the exterior closes up, and there is formed eventually a blastula of approximately spherical shape. The wall of the blastula is composed of a single layer of cells those towards the apical pole being smaller and less rich in yolk than those on the opposite side (Fig. 3, I). RANA In the case of the Frog we have an egg in which as compared with that of Amphioxus there is present a much greater proportional amount of yolk and which in consequence serves well to illustrate the nature of the influence of yolk upon segmentation. The process of segmentation begins with the appearance, in the region of the apical pole, of a small dimple on the surface of the egg which gradually lengthens out to form the first meridional furrow (a). The furrow gradually extends downwards over the surface of the egg (Fig. 5, A) and becomes completed by reaching the lower pole after about an hour and a quarter.? It also extends inwards from the surface and finally bisects the egg into two hemispheres. The second furrow (8) is also meridional and is in a plane 1 As there are marked discrepancies between the accounts given by different observers we may take it as probable that there is considerable variability in the details of segmentation about this stage. 2 See, however, later for caution in reference to the time factor in development. I SEGMENTATION 11 perpendicular to that of the first. It appears about three-quarters of an hour after the latter and, like it, extends downwards and inwards so that the egg becomes divided into four approximately equal segments. The third furrow is latitudinal in position being situated (Fig. 5, C) roughly about 20° above the equator. It extends inwards and the egg is now converted into eight blastomeres, four micromeres towards the apical pole and four macromeres towards the lower pole. Closer study of these first three cleavages in the case of the Frog brings out a number of im- portant points. It will be noticed in Fig. 5 that the circular area of the egg-surface which is free from pigment is placed somewhat eccen- trically so that at one edge it approaches the equator of the egg much more nearly than it does at the opposite edge. It will be noticed further that the egg as judged by the distribution of pigment is arranged symmetric- ally about the plane of the first furrow. This furrow seems to cor- respond, under normal conditions, with the sagittal plane of the embryo, and therefore the two hemispheres separated by the first furrow correspond to the right and left halves of the embryo. The tn study of later stages will bring out fo the fact that the point in the Fic. 5.—Ilustrating segmentation of Frog’s boundary of the unpigmented egg. (After Schultze, 1899.) portion which lies nearest to the equator marks what will become the posterior end of the embryo. From the time of appearance of the third furrow onwards wide differences occur between different eggs. Occasionally one may be found in which matters proceed with diagrammatic regularity. Two new meridional furrows appear intersecting the angle between « and £ and like the latter they gradually extend downwards, halving each of the existing blastomeres and thus giving rise to sixteen blastomeres—in two tiers of eight, micromeres above, macro- meres below. Then a latitudinal furrow appears dividing the micromeres, and later a similar furrow dividing the macromeres; so that there are now four tiers of eight blastomeres each. Commonly however there is no such regularity either in the arrangement or in the time of appearance of the furrows. The meridional furrows in particular tend to be replaced by vertical 12. EMBRYOLOGY OF THE LOWER VERTEBRATES cu. furrows which intersect a or 6 at some distance from the poles. As regards the variation in order of development of the various furrows a good idea will be got from Fig. 6. Whatever be the case with the divisions immediately succeeding Fic. 6.—Illustrating the varia- tion in the order of appear- ance of the first cleavage furrows in Rana palustris. (After Jordan and Eycles- hymer, 1894.) The sequence, in time, of the appearance of the furrows is in- dicated as follows :—l, ===; a 3 4, 5 5, fi several cells thick (Fig. the eight cell stage, from now onwards there is little regularity. All that can be said is that each individual blastomere goes on dividing over and over again, the length of time elapsing between successive divisions bearing a rough relation to the amount of yolk present in the particular blasto- mere. i Already at the third cleavage the eight blastomeres have a distinct chink—the com- mencing blastocoele—between their inner ends and as segmentation goes on this space becomes larger. The thirty-two-cell stage is a, blastula which in a meridional section (Fig. 7, A) is seen to correspond in its general character with the blastula of Amphioxus but to differ from it in three features: (1) it is of larger size, (2) it is composed of fewer cells and (3) the difference in size between the less richly yolked cells towards the apical pole and the more heavily yolked cells towards the opposite pole is more marked. As development proceeds a farther differ- ence becomes apparent. In the various mitotic divisions during the preceding phases of segmentation the axis of the spindle has been arranged more or less tangentially but now divisions begin to take place in which the spindle axes are arranged radi- ally and the division-planes tangentially. When this happens one of the two result- ing daughter cells is nearer the centre, the other nearer to the surface of the blastula and the effect of repeated divisions + of this type is that the blastula-wall loses * its original character of being composed only of a single layer of cells and becomes vases ELASMOBRANCHS The egg of any ordinary Elasmobranch such as a Dogfish, Skate, or Torpedo, illustrates the type of segmentation that takes place I SEGMENTATION 13 in an egg in which the proportion of yolk present approaches the maximum. In this case the zygote nucleus commonly undergoes two mitotic divisions before there is any external symptom of segmentation of the cytoplasm. Usually a single furrow makes its appearance first, incising the surface of the germinal disc but not extending to its periphery (Fig. 8, A). Occasionally a second regular furrow makes its appearance intersecting the first at right angles Fic. 7.—Vertical (meridional) sections through blastulae of Frog. (From Morgan, 1897.) AR, commencing invagination ; SG, segmentation cavity. and it is a curious point that it is sometimes this second furrow which corresponds to the first nuclear division. These first two furrows apparently represent the first two ieridional furrows of the holoblastic egg though in the Elasmobranch the first to appear may be either a or 8. More usually, in place of a second regular furrow developing, irregular branches of the first furrow, or even independent furrows, appear and an arrangement of somewhat radiating furrows is brought about which gradually becomes converted into a network (Fig. 8, B, C, D). It should be noticed in regard to these segmentation furrows 14. EMBRYOLOGY OF THE LOWER VERTEBRATES cu. that the first latitudinal furrow cannot be identified in the Elasmo- branch and further that the study of sections shows that the furrows sometimes cut into the germinal disc obliquely instead of being perpendicular to the egg surface. The nuclei of the blastoderm divide synchronously and after four divisions have taken place, when there are sixteen nuclei in place of the original single zygote nucieus, the segmentation furrows (Fig. 8, C) form a network dividing up the blastoderm into smaller central and larger peripheral blastomeres. These blastomeres are, however, not completely isolated from one another but are still in continuity Fic. 8.—Surface views of the blastoderm of Elasmobranchs illustrating the process of segmentation. (After Riickert, 1899.) A, B, C, Torpedo ; B*, Pristiurus ; D, EB, Seyllium. Fig. B* shows an abortive segmentation which may often be observed transitorily round the margin of the germinal dise [Merocyte segments of Rickert). at their bases, and, in the case of the peripheral blastomeres, at their outer ends. Up to the fifth mitotic division the axes of the mitotic spindles have been approximately parallel to the surface but now blastomeres begin to divide with their spindle axes perpendicular to the surface so that a set of superficial segments becomes separated off. Beneath these fluid accumulates intercellularly and a segmentation cavity arises (Fig. 9, B). During the sixth division some of the blastomeres forming the floor of the segmentation cavity become separated off trom the I SEGMENTATION 15 underlying, unsegmented, yolk (Fig. 9, C) and in surface view the blastoderm assumes the appearance shown in Fig 8, D. Up to and including the seventh division mitosis takes place Fic. 9.—Vertical sections through Elasmobranch blastoderms illustrating the process of segmentation. (After Riickert, 1899.) A, C, D, E, Torpedo ; B, Scyllium; F, G, Pristiurus. EH, F, and G are sagittal sections with posterior edge of blastoderm to the right. practically synchronously throughout the blastoderm. In Lorpedo Riickert (1899) found that even in the ninth division the majority of the nuclei still divided synchronously and that in some eggs the same was the case with the tenth division but in any case approxi. y I 16 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. mately about this period individual nuclei lag behind others and the regular rhythm becomes lost. This rhythm of nuclear division is of interest in relation to the size of the individual blastomeres. It is often noticeable in an Elasmobranch blastoderm that the blastomeres are somewhat smaller in what is shown by later development to be its posterior half «.e. the half next the side on which the embryonic rudiment makes its appearance later. It would be natural to suppose that the smaller size of the blastomeres is due to their having gone through a greater number of divisions but this explanation is rendered less satisfac- tory by the synchronism of the mitotic divisions. Apparently the inequality is at least to some extent due to the zygote nucleus, and, later on, the first segmentation furrows, being not quite central in position in the germinal disc but situated slightly towards its posterior edge (Rickert). The stage up to which mitosis remains synchronous varies amongst individuals of one species and a fortiort amongst those of different species and genera. Thus in Pristiurus it is, commonly, regular only up to the fifth mitosis acvording to Riickert. While segmentation has been proceeding, important changes have been taking place in the segmentation cavity. About the time of the seventh division the rounded inner blastomeres fill up most of the cavity so that it becomes reduced to chinks between the individual blastomeres. These chinks are filled with fluid secreted by the egg substance, and in the yolk beneath the blastoderm the activity of this process of secretion is indicated by the appearance of fluid vacuoles. As development goes on the amount of fluid increases greatly and about the tenth division it begins to collect especially between the blastoderm and the,yolk, forming the “germ cavity” of Riickert (Fig. 9, D, E, F). This cavity is best marked towards the posterior side of the blastoderm and in ground-plan is crescentic in shape. It varies greatly in its degree of development in different individuals. Whether it is advisable to use a separate name for this cavity is very questionable. When a broad view is taken of the relations of blastomeres and segmentation cavity in the Elasmobranchs these seem to be similar in kind to those which hold in the case of the Lung fishes. In these fishes, as will be shown later, the blasto- meres which originally formed the floor of the segmentation cavity become later on shifted in position towards its roof but the resultant change in the topographical relations and form of the segmentation cavity would clearly afford no valid reason for giving it a new name. The Yolk Syncytium.— The layer of substance immediately underlying the blastoderm and segmentation- or germ-cavity is distinguished from the main mass of yolk upon which it in turn rests by the finer grained character of its yolk granules, and by its I SEGMENTATION 17 greater richness in protoplasm. This layer shows no division into cells and is therefore termed the yolk-syncytium! (H. Virchow: Riickert’s term “merocytes” is synonymous). The marginal portion round the edge of the blastoderm is sometimes termed the germ-wall. . Functionally the yolk-syncytium is apparently concerned especi- ally with the digestion and assimilation of the yolk. Scattered about in it are nuclei, often of enormous size and irregular form. Concerning the origin and fate of these nuclei much discussion has raged and the matter cannot yet be regarded as satisfac- torily settled. The question is complicated by the fact that, as shown by Riickert (1890), polyspermy appears to be a normal occurrence In Elasmobranchs. In addition to the single micro- gamete which takes part in the formation of the zygote-nucleus a variable number of extra spermatozoa make their way into the egg and give rise to accessory sperm-nuclei. Where such sperm-nuclei are situated in the coarse yolk they apparently soon degenerate but when, on the other hand, they are within the protoplasm of the germ-disc they remain during the early stages of development in a living and apparently healthy condition, even undergoing mitosis synchronously with the nuclei derived from the zygote-nucleus up to the fourth or even fifth or sixth division in the case of Z'orpedo. The importance of this fact should be noted in connexion with our ideas of the reciprocal physiological relations of nucleus and cytoplasm. It is fully recognized that the nucleus governs and controls cell metabolism: it is not always so fully recognized that conversely the cytoplasm exerts an important influence over the nucleus. Clearly the fact that the accessory sperm-nuclei “keep step” in their mitotic divisions with the embryonic nuclei must be due to some influence exerted on the former nuclei through the cyto- plasm. It should, in fact, never be forgotten that cytoplasm and nucleus are merely locally specialized portions of the same common living substance or protoplasm. At first the accessory sperm-nuclei are clearly distinguishable in the germinal disc from the true embryonic nuclei by their smaller size and reduced (haploid) number of chromosomes. After the zygote-nucleus has undergone two mitoses however—or even before the second mitosis—the accessory sperm-nuclei wander—or become transported by cytoplasmic movements—outside the limits of the germ-dise. They continue their mitotic rhythm for a time so that, for example, at the 8-nuclear stage of the blastoderm they may be seen in groups of eight lying in the yolk-syncytium. During early stages of segmentation numerous such obviously accessory sperm- nuclei may be seen in the syncytium but as time goes on the nuclei 1 Although Haeckel originally defined the term syncytium (Die Kalkschwamme, Bd. I. p. 161) as a protoplasmic mass formed by the fusion of originally separate cells the word has come into such general use for a multinucleate mass of protoplasm which shows no subdivision into cells, whatever its origin may have been, that there seems no serious objection to the use of the term yolk-syncytium as suggested by Virchow. VOL. II Cc 18 EMBRYOLOGY OF THE LOWER VERTEBRATES — cu. of the syncytium are seen to be of a different character. They are now of enormous size and of peculiar lobed appearance. The lobing becomes more complex as time goes on and appears to be due to incomplete and irregular attempts at amitotic division. The discussions, alluded to above, have centred round the mode of origin of these highly characteristic giant nuclei. Balfour, who first described them (1874), did not express any opinion as to their origin. Riickert in his first paper (1885) on Elasmobranch develop- ment looked on them simply as specialized embryonic nuclei and gave the masses of protoplasm in which they are embedded the name “merocytes.” Latterly however Riickert, after his discovery of polyspermy in Elasmobranchs, has taken the view that the yolk- nuclei are really the accessory sperm-nuclei before alluded to which have altered their character in correlation with the altered environ- ment in which they find themselves after leaving the germinal disc. In spite of Riickert’s more recent observations and conclusions, and in spite of their being supported by Samassa, Beard and others, A B Cc Fic. 10,—Views of the segmenting germinal dise of Bdellostoma stoutt. (After Bashford Dean, 1899.) it must, I think, be admitted that the sperm-origin of the yolk-nuclei, is by no means demonstrated. And all general considerations are in favour of Riickert’s earlier view being the correct one, namely that the nuclei of the yolk-syncytium are genetically of the same order as the ordinary embryonic nuclei. Such general considerations render it extremely improbable that accessory gamete nuclei should really play an important physiological part in the developing embryo : it is far more probable that such nuclei simply persist for a time, undergo mitosis a few times and then degenerate and disappear. The variations in the process of segmentation are well illustrated by the three cases just described and # will be convenient now to summarize the general characteristics of the process in the various remaining groups. Lampreys.—In ,the Lamprey the phenomena of segmentation agree closely with those observed in the frog and need not be further described. Myxinoips.—In the Myxinoids the somewhat sausage-shaped egg is heavily yolked and possesses a germinal disc situated close to one pole. A few segmentation stages of Bdellostoma (Fig. 10) I SEGMENTATION 19 have been described by Bashford Dean (1899) and as might be expected the segmentation is meroblastic. Apparently the first two furrows (a and 8) have the normal meridional arrangement the specimen figured by Dean (Fig. 10, A) showing a displacement at the intersection of the two furrows. These latter do not reach the edge of the germinal disc. The third set of furrows (Fig. 10, B) appear to be vertical and in the next stage figured (Fig. 10, C) the. furrows have become joined up to form an irregular network which still barely reaches the edge of the blastoderm. CROSSOPTERYGIANS. — Our knowledge rests entirely on the observations of Budgett (Graham Kerr, 1907). These, necessarily fragmentary, observations suffice to show that the process of segmentation is of great interest. In the earliest stage observed, but not figured, by Budgett the egg was “segmenting in four equal Fra. 11.—Segmentation and gastrulation in Polypterus. (Drawings by Budgett. Graham Kerr, 1907.) A, represents a view of the apical pole: the remaining figures are side views. portions, the constrictions being deeper than in the frog.” A second egg (Fig. 11, A and B) is in the eight-blastomere ‘stage. The blastomeres are practically equal in size and it may be inferred with considerable probability that in Polypterus two meridional furrows are succeeded by a latitudinal one which is very nearly equatorial. The nearness of the latitudinal furrow to the equator is remarkable in view of the fact that the egg of Polypterus, as shown by the study of sections (Fig. 1, B, p. 3), is not by any means nearly isolecithal. ACTINOPTERYGIANS.—The typical Teleost is characterized by the fact that its richly yolked eggs show a more complete segregation of protoplasm and yolk than do those of any other Vertebrata. In correlation with this the segmentation is here the most markedly meroblastic in character. These featuresisuggest that in the ancestral ‘Teleost the yolk was large in quantity and that the egg as a whole was of great size. Amongst present-day Teleosts however it is only, comparatively speaking, a few forms mostly inhabiting fresh water, 20 EMBRYOLOGY OF THE LOWER VERTEBRATES | CH. ‘which produce eggs of very large size eg. Gymnarchus niloticus (Budgett, 1901; Assheton, 1907) where they measure about 10 mm. in diameter, or the Salmon or Trout where they measure from 4 to 5 mm. The majority of fishes produce eggs in enormous numbers, amounting in some cases to several millions, and correlated with this the size of the individual egg has become much reduced. The average diameter of a Teleostean egg may be taken as about 1 mm. In an egg of this size segmentation of so markedly meroblastic a character would be puzzling except on the hypothesis that the meroblastic condition had arisen in ancestral forms in which the eggs were much larger. The larger part of the egg consists of a spherical mass of practically pure yolk. On the surface of this is a thin layer of protoplasm containing droplets of oil, and this layer of protoplasm is more or less distinctly thickened in the region of the apical pole to form a germinal disc in which is contained the nucleus. Irregular prolongations of the superficial protoplasm may sometimes, especially in immature eggs, be traced inwards into the substance of the yolk. A characteristic feature of many teleosts is the tendency for the yolk to assume a liquid form. This is particularly marked in many pelagic eggs where it is not merely liquefied but runs together at the time of spawning or of fertilization to form a sphere of glassy transparency. There may further be, interspersed amongst the ordinary yolk, droplets of oily looking fluid often with a distinctive colour. These may unite into a few droplets or into a single larger drop forming a conspicuous, often coloured, sphere in the midst of the ordinary yolk. The colour and size of such drops frequently afford an easy means of recognizing the species to which a particular egg belongs. They may also have a characteristic position and may be surrounded by a special condensation of protoplasm or, on the other hand, they may simply float freely in the main mass of fluid yolk. Although these droplets may, as already indicated, exhibit peculiarities characteristic of particular species they do not seem to give indications of genetic affinity in regard to genera or larger groups: nor do they show any definite relation to the conditions, pelagic or otherwise, under which the egg develops (Prince, 1886). » rhe yolk of teleosts is also characterized by a diminution of its Specific gravity which causes the egg to assume a reversed position with the apical pole below, and which further, in the case of a vast number of marine fishes, causes the egg as a whole to float freely suspended in the sea water. Seeing that the Teleostei as a group is above all characterized by specialization for a swimming existence, independent of a solid substratum, we are perhaps justified in assuming that the freely floating pelagic mode of development above mentioned was originally present throughout the group. The demersal type of development, where the eggs are deposited on the, solid substratum, would then I SEGMENTATION 21 be regarded as a secondary reversion to, rather than a persistence of, a pre-teleostean habit. Possibly the reversed position of the egg is to be regarded as a means of protecting its more sensitive apical portion from injury by contact with the surface film of the water in which it floats. ; When fertilization takes place the most conspicuous immediate result is the onset of a gradual concentration of the protoplasm in the germinal disc—the dise becoming at the same time more heaped up, its vertical diameter in- creasing and its horizontal diminishing. The segmentation of the germinal disc in A teleostean fishes is usually of a very regular and characteristic kind. It is illustrated as seen in surface view by Fig. 12. The germinal disc lengthens out into an elliptical shape. The first furrow to appear (A) is meridional B and occupies the shorter diameter of the ellipse. The second furrow is also meridional and in a plane perpendicular to that of the first. The third and fourth sets of furrows (B, C, D) are vertical and they become arranged so as to be practically parallel to the first and second, with C the result that the blastoderm as seen in surface view assumes a very characteristic arrange- ment of sixteen segments arranged in four rows (Fig. 12, D). The internal phenomena of segmentation D may be described from what occurs in the Trout (Kopsch, 1911). In the first place it has to be noted that the early furrows do not extend 10) pepmantation right through the substance of the germinal in the blastoderm of a disc but leave a continuous basal stratum of teleostean fish (Ser- protoplasm next the yolk. The blastoderm Ha le grer assumes a two-layered condition by the 3rd and _ Wilson, 1891.) 4th furrows curving round in their deeper portions so as to intersect the preceding division-planes which were throughout perpendicular to the surface (Fig. 13, B). Up to the 16- cell stage all the segments remain connected by broad protoplasmic bridges apart from the continuous basal layer of protoplasm which connects the deepest cells together. In the 32-cell stage (Fig. 13, C) the cells of the superficial layer have become completely isolated while the deep cells are stall connected together. With the next division the blastoderm becomes three layered, the cells of the intermediate layer being derived some from the superficial, some from the deep layer, as is shown by the evidence of broad bridges of protoplasm which persist here and there between sister cells. With the next division (128-cell stage, Fig. 13, D) the four-layer 22 EMBRYOLOGY OF THE LOWER VERTEBRATES — cu. condition is reached, the cells of the basal layer being still connected by a thick continuous stratum of protoplasm. By this time it 1s found that the nuclear divisions of the basal layer are clearly lagging behind those of the other layers. As segmentation proceeds further the continuous basal sheet of protoplasm decreases relatively in thickness. For a time (Fig. 13, E) bulgings of its upper surface indicate that it is giving off cells into the overlying layer but as the thinning process goes on these become less and less numerous. H. Virchow distinguished three zones-in the basal sheet of protoplasm—marginal, intermediate and central, although the latter Fig. '13.—Vertical sections through the blastoderm of a Teleost (Salmo JSario) illustrating the process of segmentation. (After Kopsch, 1911.) A, end of second division. Section perpendicular to plane of first furrow which is therefore seen cut across. B, commencement of fourth division. Plane of section as in A. The division surfaces of the third division are seen to curve inwards so as to meet the first division surface. As a result the latter has become distorted and no longer forms a plane. C, middle of sixth division. D, be- ginning of eighth division. E, beginning of tenth division. F, 62-hour blastoderm. (The dark portion at the top of Fig. B represents the free surface bounding the second furrow: the dark tone at the lower edge of each figure represents yolk.) is not quite central but situated rather towards the posterior or embryonic edge of the blastoderm. The intermediate zone is marked off from the others by the fact that the thinning process has there progressed farther. Up to about the twelfth division the nuclei all through the blastoderm divide practically synchronously except those of the basal layer which as already indicated lag behind. Soon after this however (from about the 41st hour—Kopsch) the divisions become irregular. The basal layer becomes the yolk-syncytium: the cell limits visible on its upper side become obliterated and it becomes more and more flattened out. Although its nuclei undergo repeated mitosis I SEGMENTATION © 23 there is no longer any budding off of cells, the nuclei simply lying within the substance of the syncytium. . As they increase in number nuclei from the central and marginal regions spread into the intermediate zone, which up to now contained very few nuclei, while others pass outwards into the peripheral protoplasm (Periblast— Agassiz and Whitman, 1885) lying outside the limits of the blastoderm. Towards the end of the second day the syncytial nuclei begin to increase markedly in size and they begin to undergo abnormal multipolar mitoses. During the third day they complete the assumption of these peculiarities which are characteristic of the nuclei of a yolk-syneytium—enormous size,curiously lobed appearance, and the tendency for the lobes to become nipped off irregularly so as to give rise to groups of small nuclei. During these later stages of segmentation the blastoderm becomes flattened somewhat and instead of bulging out over its attached base all round, its surface passes into the extrablastodermic surface by a slope very much as it did before segmentation began (Fig. 13, F). ACTINOPTERYGIAN Ganorps.—The ganoid fishes are of special embryological importance because, so far as actinopterygians are concerned, they appear to be the least. modified descendants of those ancestral forms from which the Teleostean fishes have been evolved. Study of their developmental phenomena is desirable in order to see to what extent they throw light upon the peculiarities of development which characterize the Teleostean fishes. It will be necessary therefore to review the segmentation processes so far as they are known in each of the three types—the Sturgeon, Amia and Leprdosteus. 2 The only Sturgeons of which anything is known regarding their early development are the commod sturgeons of the genus Acipenser. Polyodon, Psephurus and Scaphirhynchus are so far completely unknown, though it is highly desirable that their development should be investigated. In both . mes. VEZ LO aN NMI Noe ii Soe Spysaty Fic. 37.— Transverse sections through embryos of (A) Zriton and (B) Rana temporaria showing continuity of endoderm and mesoderm on each side of the notochord. (After O. Hertwig, 1882 and 1883.) end, endoderm ; m.p, medullary plate ; mes, mesoderm ; N, notochordal rudiment. be sufficient reasons for regarding these connexions as primitive rather than those more ventrally situated. The balance of probability appears rather to favour the view that of the two connexions it is the ventral one which is the persistent original one, and that it is the dorsal which is to be interpreted as due to secondary fusion. The second bearing is at least equally important. It rests on the occurrence of active cell-proliferation on each side of the noto- chordal rudiment. For in some of the meroblastic vertebrates (Amniota)—correlated with the more and more complete segregation of yolk from protoplasm—this zone of proliferation becomes ap- parently the main source of the mesoderm. ELASMOBRANCHIL—In the Elasmobranch, while there are still traces of formation of mesoderm by a process of delamination from the main mass of endoderm or yolk (Fig. 38, A), a more con- spicuous mode of formation is provided by active proliferation of T ORIGIN OF THE MESODERM 63 the endoderm cells along the inner and outer edges of the sheet of mesoderm. In early stages and in the anterior part of the embryo this pro- liferation process may alone be in evidence, so that in place of a broad continuous sheet of mesoderm there are found two narrow strips— one (Fig. 38, C, mes’) arising from the endoderm just external to the notochordal region and the other (mes") arising from the endoderm Fic. 38.—Three transverse sections through an embryo of Pristiurus (Stage B, Balfour), illustrating the origin of the mesoderm. (After C. Rabl, 1889.) Section A, through the posterior half of the embryo; B, through the iniddle ; C, through the anterior half. ect, ectoderm ; end, endoderm ; mes, mesoderm ; y.n, yolk nuclei. peripherally. The two strips are known respectively as the axial (Riickert ; or Gastral, Rabl) and the peripheral mesoderm (Riickert ; or Peristomal mesoderm, Rabl). Much discussion has centred round this double origin of the mesoderm and attempts have been made to distinguish axial and peripheral mesoderm in holoblastic forms including even Amphiowus, thus infringing one of the chief canons of embryological science— that developmental phenomena in the higher forms are to be ex- plained by those of the lower and not vice versa. Sania 64 EMBRYOLOGY OF THE LOWER VERTEBRATES — cu. In the formation of axial mesoderm we recognize a zone of fusion of mesoderm and endoderm accompanied by proliferation of mesoderm entirely analogous with that which occurs in Amphibians but which had not yet made its appearance in lower holoblastic forms. Whether it is justifiable to regard the outer zone of mesoderm formation in the Elasmobranch as equivalent to the region of de- lamination (a process which necessarily involves cell-proliteration) is doubtful. It is indeed doubtful to what extent there is justification for drawing any morphological distinction between axial and peri- pheral mesoderm, seeing that the two regions of proliferation are on the protostoma hypothesis morphologically closely related to one another (see Chap. LX.). If we look at the matter from the point of view of physiology rather than of morphology we may probably recognize in the active formation of axial mesoderm an expression of the general tendency in the meroblastic egg for all processes of growth and cell prolifera- ect. Fic. 39.—Transverse section through the blastoderm of a snake (Z'ropidonotus) illustrating the origin of the mesoderm, (After O. Hertwig, 1906.) ect, ectoderm ; end, endoderm ; mes, mesoderm. tion to become concentrated towards the mesial plane dorsally and to slacken off peripherally and ventrally. REPTILES.— According to the view taken in this book the meso- derm in the holoblastic Craniates at one period spread outwards by a process of delamination from the yolk-laden endoderm. In the Amphibians we have seen that a new source of addition to the mesoderm had made its appearance in the form of a zone of pro- liferation on each side of the notochord, in which region cells are budded off into the mesoderm. In the Reptiles—admittedly descendants from Amphibian-like ancestors—in correlation with the concentration of developmental activity towards the mid-dorsal line brought about by the accumula- tion of the yolk ventrally, this parachordal source of mesoderm has become predominant while the lateral source has become greatly reduced. In Fig. 39 is represented the typical mode of mesoderm forma- tion as seen in a transverse section through the trunk region of a reptilian embryo. The mesoderm is seen to be spreading out asa wing of cells towards either side from the notochordal or primitive streak region between the two primary cell-layers. Brirps.—In the Birds also the method of first mesoderm formation I ORIGIN OF THE MESODERM 65 appears to be closely comparable with that of Reptiles and Amphib- ians. Here, at the time when the mesoderm begins to make its appearance, the position of the notochord is occupied by the primitive streak. The mesoderm forms a loose sheet of irregularly shaped cells spreading out on each side and added to from two distinct sources : on its inner side by proliferation from the primitive streak and on its outer side by delamination from the endoderm of the germ wall. It will facilitate comprehension of the evolutionary changes which ent. mes. end. Fic, 40.—Semi-diagrammatic transverse sections through the embryos of various vertebrates to illustrate the origin of the mesoderm. A, Amphiorus; B, Petromyzon; C, Lepidosiren ; D, Amphibian ; E, Elasmobranch. eet, ectoderm ; end, endoderm; ent, enteric cavity; mes, mesoderm; N, notochord; n.r, neural rudiments. The small crosses indicate regions in which active extension of the mesoderm is taking place. the writer believes to have taken place in the mode of development of the mesoderm within the phylum Vertebrata if the main steps are summarized in a diagram. In Fig. 40, A shows the primitive condi- tion where the mesoderm segments are in the form of enterocoelic pockets (Amphiozus). In B, with increasing amount of yolk, the hollow pocket is represented by a solid block in which the cavity will develop secondarily (Petromyzon). In C the condition is similar but the dorsal portion of the embryonic body is more flattened out, the bulk of the yolky endoderm over which it is spread having VOL. II F 66 EMBRYOLOGY OF THE LOWER VERTEBRATES CH. become greater (Lung-fish). In D the secondary continuity of the mesoderm with the endoderm just outside the notochord is present and proliferation of mesoderm cells has commenced in this region (Amphibia). Finally, in E, with the very great increase in the bulk of the yolk, the dorsal part of the embryo is still more flattened out, and the addition to the mesoderm by proliferation of endoderm cells into it close to the notochord has now become conspicuous (Elasmo- branch). THE MESENCHYME The fate of the mesoderm whose origin has just been traced is to give rise directly to the peritoneal epithelium which lines the body cavity and covers the organs lying within it, and also to the muscular system. Indirectly it, however, also plays a great part in the forma- tion of what is known as the mesenchyme. Whereas for a time the Vertebrate body is composed of compact masses or layers of cells, it is a general characteristic that, as develop- ment goes on, individual cells detach themselves and wander away through the body, multiplying by fission accompanied by mitosis, and behaving in fact very much as if they were independent organisms. In the course of the many generations of these cells which arise during the process of individual development, they become divided into various strains which show marked differentiation for the per- formance of different functions. Some retain a relatively primitive amoeboid form and undertake such functions as the transport of food material, the absorption of moribund tissues in regions where shrinkage in volume or atrophy is taking place, and the ingestion and destruction of attacking organ- isms such as disease germs. Some, their protoplasm laden with insoluble excretory products as a result of their active metabolism, wander towards the light and settle down near the surface of the body as pigment cells or chromatophores which serve on the one hand to protect the underlying tissues from the light and upon the other to give distinctive coloration to the animal. Others again settle down in an abundant jelly-like intercellular matrix to form connect- ive or packing tissue, which in turn shows evolution in various directions in accordance more particularly with different developments of the intercellular matrix. Of special importance are these types in which the matrix becomes hard and rigid so as to form skeletal tissues such as bone and cartilage. Another important strain of these cells is characterized by the fluidity of the matrix and the independence of the individual cells which float in it. This liquid connective tissue forms the blood which, pumped through an elaborate system of vessels, serves on the one hand for the transport of food and oxygen to the tissues, and on the other for carrying away the waste products of metabolism to the special excretory organs the duty of which is finally to remove these harmful substances. I ‘MESENCHYME 67 The sum of these amoeboid cells, which proceed along the various evolutionary paths above indicated, were, by O. Hertwig, given the ‘name Mesenchyme—to distinguish them from the mesothelium, or mesoderm in the restricted sense, in which the cells remain in the form of continuous layers or epithelia. The original mesenchyme cells arise by emigration from the pre-existing cell layers. Possibly all three layers give rise to mesenchyme cells. It is the mesoderm however which does so most conspicuously. In an Elasmobranch embryo, for example, active budding off of mesénchyme cells is seen over large areas of the somatic mesoderm and similarly from the inner surface of the splanchnic mesoderm. Most active of all is the production of mesen- chyme cells from the splanchnic mesoderm near the lower end of the mesoderm segment, where the proliferating mesenchyme cells may form a conspicuous mass projecting towards the mesial plane and termed the sclerotome.t The special consideration of the sclerotome and of the mesenchyme in general will come most conveniently after the other derivatives of the mesoderm (Chaps. IV., V., VI.). 1 The use of the word sclerotome in this restricted sense has come to be practically universal in embryological literature and is therefore followed in this volume. The word was invented by Goodsir and defined by him, at the British Association meeting in 1856, as meaning a segment of the supporting tissue or framework (whether ‘fibrous ” or cartilaginous or osseous) in a segmented animal. LITERATURE Agassiz and Whitman. Mem. Mus. Comp. Zool., xiv, 1885 and 1889. Assheton. The Work of John Samuel Budgett. Cambridge, 1907. Balfour. Quart. Journ. Micr. Sci., N.S. xiv, 1874. Brachet. Arch. Biol., xix, 1903. Brauer. Zool. Jahrbiicher (Anat. Abt.), x, 1897. Budgett. Trans. Zool. Soc. London, xv, 1901. Cerfontaine. Arch. de Biol., xxii, 1906. Dean, Bashford. Journ. Morph., xi, 1895. Dean, Bashford. Quart. Journ. Micr. Sci., xxxvili, 1896. Dean, Bashford. Kupfiers Festschrift, 1899. Goette. Abhandl. zur Entwickelungsgesch. d. Tiere, v, 1890. Goodsir. Edinburgh New Philosophical Journal, v, 1857. Goronowitsch. Morph. Jahrb. x, 1885. Hatschek. Arb. zool. Inst. Univ. Wien, iv, 1881. Hertwig, O. Jenaische Zeitschrift, xv and xvi, 1882 and 1883. Hertwig, 0. Handbuch d. Entwicklungslehre, Jena, 1906. Jablonowski. Anat. Anzeiger, xiv, 1898. Jenkinson. Vertebrate Embryology. Oxford, 1913. Jordan and Eycleshymer. Journ. Morph., ix, 1894. Kerr, Graham. Phil. Trans., B, cxcii, 1900. Kerr, Graham. The Work of John Samuel Budgett. Cambridge, 1907. Kopsch. Internat. Monatsschrift f. Anat. u. Phys., xix, 1902. Kopsch. Arch. mikr. Anat., lexviii, 1911. Kowalevsky. Mém. Acad. Sci. St. Pétersbourg, Ser. 7, xi, 1867. Mitsukuri. Journ. Coll. Sci. Tokyo, x, 1896. Morgan. The Development of the Frog’s Egg. New York, 1897. Patterson. Journ. Morph., xxi, 1910. Prince. Ann. Mag. Nat. Hist., Fifth Series, xviii, 1886. Rabl, ©. Morph. Jahrb., xv, 1889. Rothig. Arch. mikr. Anat., lxx, 1907. 68 EMBRYOLOGY OF THE LOWER VERTEBRATES cui.1 Rickert. SB. Ges. Morph. Phys. Miinch., 1885. Rickert. SB. Ges. Morph. Phys. Miinch., 1890. Rickert. Kupffers Festschrift, 1899. Salensky. Memoirs of the Society of Naturalists of Kagan, vii, pt. 3, 1878. (In Russian, A French translation of the more important parts will be found in Arch. de Biologie, ii, 1881.) Samassa. Arch: Entwickelungsmechanik, vii, 1898. Sarasin, P. and F. Ergebnisse naturwiss. Forsch. auf Ceylon. Wiesbaden, 1887— 1893. Schultze, O. Kollikers Festschrift. Leipzig, 1887. ! Schultze, O. Arch. mikr. Anat., lv, 1899. Schwink. Die Entwicklung des mittleren Keimblattes und der Chorda dorsalis der Amphibien. Munich, 1889. Semon. Zoologische Forschungsreisen, i. Jena, 1893. Will. Zoolog. Jahrbiicher (Anat.), vi, 1892. Will. SB. Berl. Akad., 1895. Wilson, E. B. Journ. Morph., viii, 1893. Wilson, H. V. Bull. U.S. Fish-Commission, ix, 1891. Ziegler. Entwickelungsgesch. d. niederen Wirbeltiere, Jena, 1902. CHAPTER If THE SKIN AND ITS DERIVATIVES THE skin of the vertebrate consists of the epidermis—the persistent and less or more modified ectoderm—resting upon the superficial layer of mesenchyme—the dermis—which in the higher forms becomes strengthened by the formation of numerous tough inter- lacing fibres. In studying the development of the skin in the various types of vertebrate we find that the ectoderm undergoes characteristic modifications to fit it for the carrying out of special functions. In the fishes it becomes converted into a highly glandular mechanism concerned with the production of slippery mucus for the diminution of what the naval architect calls “skin-friction,” in other words the friction between the surface of the body and the water in contact with it. Local or general specializations of this glandular apparatus lead to the development of cement organs by the secretion of which the young animal is able to attach itself to solid supports, to the production of digestive ferments by which the eggshell is softened or, in the case of the portion of ectoderm which lines the buccal cavity, the digestion of the food initiated, or to the production of poisonous defensive or offensive secretions. In the case of the terrestrial amphibians the glandular apparatus serves to keep the skin moist, while in the Birds it develops arrangements for oiling the feathers. Again the ectoderm develops important protective functions. It becomes hardened and toughened to give mechanical protection : it becomes more or less loaded with opaque pigment to prevent the penetration of light rays, while in those highest vertebrates, in which, correlated with intensely active metabolism, the body is kept at a higher temperature than its surroundings, the superficial horny layer becomes as it were frayed out into a fluffy coating of feathers or hair which with its entangled air retards loss of heat from the surface of the body. Finally the ectoderm forms the great mechanism for the reception of impressions from the external world. It develops sensory cells which may become crowded together to form organs of special sense 69 70 EMBRYOLOGY OF THE LOWER VERTEBRATES — cu. while from its deeper layers arise the main portions of the central nervous system. THE EPIDERMIS The ectoderm covering the surface of the embryo becomes converted, normally, into the epidermis of the fully developed individual. Very usually the embryonic ectoderm consists of two layers of cells, the lower layer composed of actively living cells, the superficial of flattened plate-like protective cells. This outer layer has been termed by Krause the periderm : its superficial protoplasm is commonly hardened to form a cuticle in the strict sense of the term. Normally it plays no active part in development and is shed at an early period. The deep layer of the ectoderm on the other hand is active. Its cells multiply so that it becomes several layers thick: the outer layers become cornified to form the horny stratum of the epidermis while the deeper cells, composed of active living protoplasm, form the stratum of Malpighi. The outer layer of ectoderm cells may be for a time ciliated. This is well seen in young Amphibian embryos (Assheton, 1896). In Rana temporaria the 6-mm. embryo possesses ciliated cells scattered thickly over its surface, the movement of the cilia being such as to drive a current of water tailwards over the surface of the embryo. When the external gills develop, a specially strong ciliary current sweeps backwards over them and it is noteworthy that this current passes over the olfactory organ en route to the external gills so that the olfactory organ possibly plays an important part in testing the quality of the water going to the respiratory organs. The ciliary apparatus is sufficiently powerful at the stage in question to cause an embryo of this stage when laid on the bottom of a flat glass vessel to slide along at the rate of a millimetre in from four to seven seconds. As development proceeds the ciliation becomes less and less prominent and in a 20-mm. tadpole it has almost disappeared except on the surface of the tail which remains richly ciliated until the time of metamorphosis. This persistence of the tail cilia is doubtless correlated with the fact that the skin of the tail plays an important part in the process of respiration. HORNY DEVELOPMENTS OF THE EPIDERMIS Scales.—In many terrestrial Vertebrates the horny layer of the epidermis becomes so thickened and hardened as to become practically rigid. In such cases the flexibility of the skin as a whole is retained by the thickened areas of epidermis being separated from one another by lines along which thickening does not take place. The thickened portions now form epidermal scales of the type seen in Reptiles.) They may take the form of simple rounded projecting bosses or tubercles as in Chameleons, or they may be flattened II THE SKIN AND ITS DERIVATIVES 71 horny plates arranged edge to edge—as in Chelonians or as on the ventral side of the body in Crocodiles or the dorsal surface of the head in Snakes and Lizards—or, finally, they may overlap like slates on a roof as is the case on the bodies of Lizards and Snakes. Occasionally, as in certain Lizards, individual scales may become greatly thickened and assume a conical spike-like form. The individual scale arises in development (Fig. 41) as a slight elevation of the surface beneath which the dermal connective tissue is somewhat concentrated. The epidermis covering the projection develops a well-marked cuticle. As development goes on the epidermis increases much in thickness and the cells of the outer layers become entirely cornified so as to form a horny plate or scale —supported by the underlying tough condensed portion of the dermis. It will be borne in mind that such typical reptilian scales have to be sharply distinguished from the morphologically quite different Fic. 41,—Early stage in the development of the scale of a snake as seen in a longitudinal section perpendicular to the surface of the skin. scales developed in the dermis in fishes. The ordinary reptilian scales serve mainly to protect the body from mechanical violence and from desiccation. Feathers.—In the homoiothermic Birds, where the body is kept at a constant temperature usually higher than that of the surround- ing atmosphere, the scales have become for the most part replaced by fluffy feathers which with the air entangled in their interstices form an admirable non-conducting envelope to retard the loss of heat by radiation, or convection, from the surface of the body. The rudiment of the feather begins (Fig. 42, A) as a slight thickening of the epidermis resting upon somewhat condensed dermis. The rudiment in fact differs little from that of a normal scale. The rudiment comes to project backwards (B) and then in- creases in length (C), projecting freely tailwards while its now relatively narrow base of attachment becomes sunk below the general surface into a pit or follicle. The rudiment now consists of a core of dermis surrounded by thick epidermis. The epidermis becomes incised along its axial surface by deep longitudinal grooves which divide its deeper portions into longitudinally arranged masses (Fig. 42, D, 0), the rudimentary barbs, while leaving the superficial portion as a continuous sheath (sh.). The grooves in question do not reach to the base of the rudi- 72 EMBRYOLOGY OF THE LOWER VERTEBRATES Cu. ment—the unincised basal portion forming the quill of the feather. The horny sheath becomes strongly cornified and then breaks open and the longitudinal thickenings of the epidermis, now also strongly cornified, break-away from the sparse cornified dermal tissue of the axis and form the fluffy barbs of the down feather. _ _ ; In the basal quill portion of the feather the epidermis immedi- ‘ 3 "2 eee RTE A Pera aan SE aces ee ao ce. Fic. 42.—Iustrating the development of feathers. (After Davies, 1889.) A, B, C, longitudinal sections; D, E, F, transverse sections (D, E, down feather ; F, flight feather) ; G, longitudinal section through barb rudiment showing developing barbules ; H, longitudinal section through base of feather. b, barb; bb, barbule; e, horny septa; c¢.e, layer of cylindrical epithelium ; f, feather rudiment; g, germinal region ; p, pulp; 4, quill; 7, rachis ; sh, sheath. ately covering the outer end of the axial dermal tissue or pulp forms a thin strongly cornified superficial layer which separates off as a septum cutting across the cavity of the quill. This process being repeated periodically gives rise to a series of horny caps fitting one over the other, in the interior of the quill (Fig. 42, H, c). The flat feathers, found as contour feathers arranged in patches over the general surface and as Remiges and Rectrices in the wings II THE SKIN AND ITS DERIVATIVES 73 and tail, originate from the basal portions of down feathers which undergo a great increase in length. The basal part of the rudiment in this case increases much in diameter. The epidermis here again becomes incised on its inner surface to form barb rudiments. These however are much more numerous (Fig. 42, F) than in the typical down feather and, further, instead of being arranged strictly longi- tudinally they are arranged somewhat spirally, starting from a con- tinuous epidermal thickening (7) which runs along the outer side of the feather rudiment. This thickening is the rudimentary rachis or shaft and the barb rudiments run from it spirally round the feather rudiment until their tips meet along its inner side. The feather is thus in early stages curled into a cylindrical form round the central dermis or pulp—the whole being enclosed in a continuous sheath which disintegrates sooner or later setting free the elastic barbs and allowing them to flatten out to form the vexillum or vane. As is well known the barbs are united together in the fully- developed feather into a functionally continuous web, through the agency of the barbules which project from the two sides of the barbs much as the barbs do from the rachis. The mode of origin of the barbules is seen in a longitudinal radial section through a barb such as that shown in Fig. 42, G, where the outer portion of the barb rudiment is seen to be splitting up into barbules (bb) while its inner portion remains continuous to form the definitive barb (0). Traced downwards, towards the base of the feather, the rachis increases in width so as to extend round the whole periphery of the feather rudiment. Its outer layer assumes a translucent character and forms the cylindrical guill (calamus), the basal end of which becomes somewhat narrowed, bounding the umbilicus, the opening through which the dermal pulp extends up into the interior of the quill. The pulp of the feather undergoes a gradual shrinkage leaving behind it the series of cornified caps (H, ¢c) formed on its apical surface as already mentioned and which eventually le loose within the quill. The lips of the umbilicus are continued (Fig. 42, H) into a deep rim of uncornified epidermis (7). This with the dermal papilla pro- jecting into the feather base remains inactive until the period of moulting when it springs into activity, grows rapidly, and becomes converted into a new feather which pushes the old one out and takes its place. ; The scales which frequently occur upon the legs and feet of birds are probably not, as might at first sight be supposed, to be looked upon as having persisted from the Reptilian condition. They fre- quently bear feathers in the young condition and are probably secondary developments replacing an earlier feathery covering. __ In view of the convincing evidence offered by comparative anatomy and palaeontology we are compelled to believe that Birds have been evolved out of Reptile-like ancestors. Accepting this 74 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. view and having regard further to the fact that Reptiles are typi- cally covered with a coating of scales, we may safely also accept the view that feathers are to be looked upon as highly specialized and modified scales. : While the mode of development of the feather fully substantiates this hypothesis, perhaps the most interesting point that emerges fe B Cc Ss., Fra. 43.—Tllustrating the neonychium or claw-pad in the developing Bird. : (From Agar, 1909.) A, median longitudinal section through the claw of a chick of 19 days’ incubation. B, claw of a chick taken in the act of hatching. Tle neonychium is seen beginning to break away from the rest of the claw. C, section similar to A, but from a chick 12 hours after hatching. c¢.p, claw-plate ; cs, sole of claw ; n, claw-pad (neonychium). from its study is that the successive sets of feathers—the down feathers of the nestling, and the annual or other sets of feathers in the adult—are not to be looked on, as has been customary, as suc- cessive series of independent individual feathers. On the contrary the down feather and the definitive feathers, which succeed it in the series of moults, are all simply portions of a single greatly elongated and basally growing structure—the first down feather being its tip, and the succeeding feathers being successive portions of it. The moult consists not in the shedding of the whole feather but merely in the breaking off of its projecting portion. Claws, which make their first appearance in Anura (Xenopus), ul THE SKIN AND ITS DERIVATIVES 75 arise as special developments of the horny layer ensheathing the tip of the digit. To produce and retain a sharp edge or point by differ- ential wear the claw is normally of denser consistency and harder on the dorsal side and laterally, forming the “ claw-plate ” (Boas) ( Fig. 43, C,¢.p), while on the ventral side it forms the softer “sole” of the claw (Boas) (Fig. 43, C, ¢.s). Neonychia or Claw-pads.—To the embryo of an Amniotic Verte- brate, enclosed within its delicate membranes, the possession of sharp claws on the digits would obviously be a source of considerable danger during the later stages of development when the embryo moves its limbs, because of the liability of such structures to tear the foetal membranes. This danger is obviated by a beautiful adaptive arrangement which has been described by Agar (1909). In the embryo, the concavity on the lower side of the claw is completely filled up by a soft rounded pad or cushion (Fig. 43, A, n) formed by a thickening of the horny layer of the epidermis superficial to the sole of the claw. Agar has given the name Neonychium to this structure. In addition to mammals, which do not concern us here, Agar has studied these claw-pads in the Fowl and in the Lizard Tejus and there can be no doubt that the expanded : claw-tips observed by Rathke (1866), Voeltzkow Pu f4— Right pectoral (1899) and Goeldi (1900) in Crocodilian (Fig.44) — govtile about suet embryos are the same structures and it seems after oviposition, show- probable that they will be found to occur in ie eel claw-bearing Amniotic Vertebrates generally. 1899.) : The neonychia are purely foetal structures which become detached soon after hatching (Fig. 43, B and C) leaving behind the functional claw. Jaws and Oral Combs of Anuran larvae.—Amongst the most interesting developments of the horny layer are the jaws and oral combs of frog tadpoles. The buccal opening is bounded by an upper and lower horny jaw, and external to and roughly parallel with these are rows of little horny denticles which form the oral combs and are used for fraying out the food. The number and arrange- ment of these rows of denticles—“ upper labial” and “lower labial” —differs in different Anura and they afford useful characters for the identification of tadpoles (see Boulenger, 1897). The horny jaw is composed simply of a row of denticles so closely apposed as to be in contact. The terminal functional portion of each denticle is seen in longitudinal section (Fig. 45, A and B) to be com- posed of a series of hollow cones of hard horny material which closely ensheath one another. The terminal cone as it undergoes wear and tear eventually drops off, its function being taken over by the cone which it previously ensheathed. 76 EMBRYOLOGY OF THE LOWER VERTEBRATES _ Cu. These cones form simply the terminal members of a series which extends inwards in the form of a curved column nearly to the inner surface of the epidermis. Only the terminal members are strongly cornified, the other members showing less and less cornification until at a little distance down the series the cone is seen to be composed of unmodified protoplasm containing at one side, near its base, a Fic, 45.—Illustrating the development of the larval teeth of Tadpoles. A, B, C, Paludicolu fuscomaculata ; D, Rana temporaria. (D after Gutzeit, 1889.) nucleus. The cone is in fact simply a cornified cell. Traced towards the base of the column the cells are seen to be composed of more granular protoplasm and to have not yet developed a hollow, while the extreme base of the column is formed by an initial cell of com- paratively small size and flattened shape. The whole column is seen, then, to be composed of a sequence of conical teeth forming a replacement series, each tooth being a single cornified cell. . I THE SKIN AND ITS DERIVATIVES 77 e The jaw, composed of a closely set row of such columns, is sup- ported by the neighbouring parts of the epidermis which also under- go a certain amount of cornification. Thus just internal to the jaw is a cushion-like mass of large slightly cornified cells which forms an efficient backing to it (cf. Fig. 45, A and B) while external to the jaw the surface of the epidermis is composed A sl ot flattened much ” cornified cells (Fig. 45, A). The oral combs consist of a pallisade- like arrangement of similar denticles which however in this case are not in contact. Fig. 45, C, shows a longitudinal section through the posterior labial comb of Palu- dicola. Here again we see a succession- column of epidermal cells commencing with a small initial cell near the inner sur- face of the epidermis. From the initial cell outwards the cells increase in size, beeome gradually cornified and each one fits closely into the base of the next one which be- comes more and more deeply excavated as the tip is approached. Two conspicuous differences distinguish the denticle of the ; ; oral comb from that of *” upilia; s fmetional spine; sy spingeadinents the jaw: (1) instead of being regularly conical in shape it is claw-shaped with serrated edges (Fig. 45, D) the tip being recurved, and (2) the hollow base of the cornified cell is not entirely occupied by its successor in the series: it also accommodates an indifferent cell of the epidermis (supporting cell of Gutzeit) which bulges into it. I have described the development of these interesting structures / Fic, 46.—Vertical section through lingual spine of Petromyzon. (After Warren, 1902.) 78 EMBRYOLOGY OF THE LOWER VERTEBRATES CH. e as they occur in a South American tadpole (Paludicola)* but the description fits quite well the mode of development as it occurs 1n Fic. 47.—Illustrating the development of the cement -organ of Lepidosiren as seen in sagittal sections. A, stage 23; B, stage 25; C, stage 31; D, stage 35. In A the rudiment of the cement-organ is seen to be a thickening of the deep layer of the ectoderm; in B and C the superficial layer has disappeared over the thickened glandular area ; in D the organ is commen- cing to shrivel and crowds of phagocytes are collected in its neighbourhood. Tadpoles generally (Keiffer, Gutzeit), the differences between different species and genera, though of systematic import- ance, being differences in detail such as shape and arrangement of the individual teeth of the comb. “Teeth” of Cyclostomes.— The horny teeth of cyclosto- matous fishes, though they would naturally fall to be treated in the next chapter, situated as they are within the buccal cavity, may con- veniently be considered now owing to their resemblance— on a much larger scale and with multicellular structure—to the horny denticles of the tadpole. The tooth-like spines of the cyclostome are cones of highly cornified epidermal cells. ach tooth develops in the substance of the epidermis (Fig. 46, A) being strikingly like a hair- rudiment during early ‘stages. Successional spines develop be- neath the bases of the func- tional ones as shown in Fig. 46. GLANDULAR DEVELOPMENTS or THE Eripgrmis.—In the Anamnia it is usually the case that scattered cells of the epi- dermis take on a glandular function and serve to form a slimy secretion which amongst other functions serves to dimin- ish the “ skin-friction ” which is the main resistance to movement through water. Such unicellular glands may become collected together to form multicellular glands. Of these the most conspicuous examples are found, outside the Mammalia, in the Lung-fishes and Amphibians—where they form the flask-glands and the cement-organs, 1 Probably P. fuscomaculata according to Boulenger. ia THE SKIN AND ITS DERIVATIVES ft) The flask-glands of Lung-tishes and Amphibians develop in the first instance as solid local proliferations of the deep layers of the epidermis which grow down into the subjacent connective tissue of the dermis and form a lumen by secondary excavation. The fully developed flask- gland is ensheathed in a coat of smooth muscle-fibres and it is an interesting fact that these are believed to be developed from the ectodermal cells of the gland-rudiment. Cement-organs of apparently ectodermal origin occur in two out of the three surviving lung-fishes—Protopterus and Lepidosiren— and form conspicuous features during late embryonic and larval life (see Fig. 200, F, Chap. VII). In an embryo of Lepidosiren three days before hatching the cement-organ forms a crescentic structure stretching across the mid- Fic. 48.—Embryos and larvae of Bufo vulgaris to show the cement-organ upon the ventral surface. (After Thiele, 1887.) ventral line with its concavity forwards, just behind the position in which the mouth will appear later. About stages 32-34 the organ is at its maximum development and forms a large prominently pro- jecting structure ventrally below the opercular region. Towards the end of larval life the cement-organ commences to atrophy, the process being furthered by its invasion by crowds of phagocytes, and in a short time the organ has completely disappeared. The cement-organ is a derivative of the deep layer of the epi- dermis. It commences as a slight thickening of this layer (Fig. 47, A) the cells assuming a tall columnar form. These columnar cells become the secretory cells while the superficial layer of the ectoderm breaks down over them so as to expose their outer ends (Fig. 47, B and C). It is to be noted that there is no trace whatever of any connexion of this cement-organ with the endoderm: it is ontogenetically a purely ectodermal structure (see, however, p. 181). Cement-organs of very similar appearance are found in the larvae 80 EMBRYOLOGY OF THE LOWER VERTEBRATES Cu. of many Anura. In this case they appear very precociously in development, being indeed in some cases the first definite organs to become visible on the surface of the embryo. Fig. 48 illustrates the development of the cement-organ in the common toad (Bufo vulgaris) from the time of its first appearance up to the time of its atrophy. The organ is seen to be at first crescentic as 1n Lepidosiren then to become V-shaped and finally to become paired by the atrophy of its median portion. When at the height of its development, the cement-organ shows characteristic differences in form and position in different species of Anura and is consequently of use in identifying the species of Tadpoles. : The general appearance of the Anuran cement-organ as observed in sections is illustrated by Fig. 49. The glandular layer is com- monly said to belong to the superficial layer of the ectoderm but this does not seem by any means certain. Pigment of the Skin.—One of the con- Fie. 49.—Section through the cement-organ of a Frog spicuous features of the Tadpole (Ran temporaria) 8 mm. in length. (From gin in the majority of sais iditaieitis me Ls Vertebrates is the fact €.0, beacon es ieee ence that it is coloured by the deposition within it of excretory matter in the form of pigments. This is of physio- logical importance to the organism in two different ways, firstly in that it gives to the particular species its characteristic coloration, and secondly it serves to protect the underlying living tissues from the harmful influence of light rays. A certain amount of pigment may be formed within the proto-— plasm of the ectoderm cells. For example in frog tadpoles of about an inch in length, numerous fine granules of melanin are crowded together near the surface of the outer layer of ectoderm cells, just beneath the cuticular superficial layer. But by far the most important part of the pigmentary system of Vertebrates consists of mesenchyme cells with pigment-laden cyto- plasm which are positively heliotropic during life and creeping by the extrusion of slender pseudopodia, like those of Foraminifera, crowd together immediately beneath the ectoderm and form there a light-proof layer, some of them even wandering into the substance of the ectoderm between its constituent cells.! The chromatophores, during the process of development, commonly become specialized in different directions so that in the fully developed 1 The interpretation of the branched chromatophores as mesenchymatous in origin appears to the author to accord best with observation but it should be mentioned that some regard them as modified ectoderm cells, as for instance Winkler (1910). rat THE SKIN AND ITS DERIVATIVES 81 condition several distinct types may be recognized. Thus in Lepido- siren the most abundant type of chromatophore is characterized by the stout projections of the cell-body which carry the finer pseudo- podia and by the somewhat brownish pigment granules. A less abundant type has long slender, less richly branched and often vayri- cose pseudopodia with dense black and opaque pigment granules. Still another type of chromatophore has its protoplasm charged with bright yellow pigment. The melanin pigments are probably to be looked upon as waste products of cell metabolism. They are iron-containing pigments and during at least the later periods of development their production is commonly associated with the breaking down of that other great iron-containing pigment—haemoglobin. Their production is also related to the degree of activity of the cell metabolism. Thus, in the male Lepidosiren, at the close of the breeding season, when the moribund remains of the richly vascular respiratory projections of the hind limb are being devoured by crowds of voracious phagocytes, there takes place an active formation of melanin. Again melanin apparently tends particularly to be produced when the cell metabolism of comparatively unspecialized cells is interfered with by the prolonged action of light-rays. Thus as already indicated the layer of protoplasm in the egg which is turned towards the light frequently develops melanin granules. Again in developing eyes it commonly holds that comparatively unspecialized mesenchyme cells wandering into the zone of exposure to the light deposit melanin granules in their cytoplasm. Cells then which become chromatophores may be regarded as cells which are specially sensitive to light-stimulus and whose meta- bolism is liable to be so modified thereby as to produce pigment. Although it is reasonable to suppose that melanin-formation is primarily related to the influence of light it must not be forgotten that, as indicated in the preceding chapter, the actual laying down of pigment in the case of species where it has become a specific character may take place under circumstances in which the light- stimulus is incomparably more feeble than that which probably originally brought about pigmentation in the course of phylogenetic evolution, as ¢.g. in the case of the ovarian egg of the frog. That pigment - formation during individual development still remains linked up with exposure to light is shown by the frequency of the unpigmented condition in Anuran larvae which develop in water rendered opaque by fine clay held in suspension (Wenig, 1913). To illustrate the dependence of pigment-formation upon light during individual development the case of young flatfish (Pleuron- ectidae) is sometimes quoted, where the shading of the upper and the illumination of the lower surface during development brings about a reversal of the ordinary colouring (Cunningham, 1893). It is possible however that this reversal of colouring is due merely to the strongly VOL, II G 82 EMBRYOLOGY OF THE LOWER VERTEBRATES — cu. heliotropic tendencies of the chromatophores which lead them to migrate actively towards the illuminated side and there to remain. The chromatophores of Vertebrates often display their sensitive- ness to light very markedly by movement reactions. Such are well seen in the young stages of many fishes and Amphibians. In the young Lepidosiren for example the chromatophores during the day have their pseudopodia extended in all directions and their bodies flattened out into a plate-like form so that they constitute a light- proof coat giving a rich purplish - black effect. At dusk the pseudopodia become slowly withdrawn so that a few hours after darkness has set in the chromatophores have shrunk into minute spheres so wide apart as to have no influence on the general colouring. The young fish is then practically colourless except for the large yellow chromatophores here and there which remain expanded. Fic. 50.—Section through epidermis of Lepidostren D uring the course of larvae. development in many fishes, A, fixed under conditions of dull daylight; B, under a DUTOUS Amphibians, and a conditions of darkness during the night. (The chroma- few Reptiles such as the tophores in the dermis, which crowd together under the chameleon the compara- inner surface of the epidermis, are not shown in A.) tivel ees 1 e ti t p, chromatophore ; v, blood-vessel. Ive y sump e€ reac 1ons 0 light such as have just been indicated develop into reactions of a much more complex type in which the central nervous system is involved. Research into the development of these more complex reactions is highly desirable for at present little is known regarding them. NERVOUS SYSTEM The nervous system, which has to do with the receiving: of, and the reacting towards, impressions from the outer world, appears to have arisen in evolution, as might have been expected, from the outer layer or ectoderm. The first steps in the evolution of the Vertebrate nervous system are not within the scope of direct obser- vation but the view is probably correct that it arose from a diffuse subepidermal network or plexus of the type still persisting in some of the more primitive invertebrates. In the development of the Vertebrate embryo the main parts of the nervous system may still be seen to take their origin from the ectoderm. Il NERVOUS SYSTEM 83 ORIGIN OF THE CENTRAL NERvous SystEM.—The first obvious trace of the central nervous system consists of a thickening of the ectoderm of the dorsal surface of the embryo. This thickening extends forwards from the anus, or from a point slightly behind the anus, to the head region, and is termed the medullary plate. The thickening of the medullary plate is due primarily to the ectoderm cells, or where two layers are present, to the deep-layer cells of the ectoderm taking on a tall columnar form. There is also growth in area of the medullary plate and this, in conjunction with the binding down of the medullary plate along the line of the notochord and primitive streak, causes it to become curved from side to side so as to form a gutter or groove—the medullary groove. The, usually conspicuous, lips of this groove are known as the medullary folds. As the medullary plate keeps on increasing in width it bulges downwards and laterally into the surrounding mesenchyme and assumes the form of a longitudinally placed tube with a slit along its dorsal wall representing the original opening of the groove. Finally the lips of this slit grow towards one another and under- go fusion, so as completely to close in the neural tube, which now separates off from the ectoderm of the outer surface. The closing in of the neural tube commonly commences in the hinder head region and spreads from this point forwards and backwards. An interesting modification of this normal mode of origin of the neural tube is found in the case of Lampreys, Lepidosiren and many of the teleostomatous fishes. In this modification of the typical process the increase in bulk of the medullary plate leads to its growing downwards into the underlying tissue as a solid keel. In the middle of this at first solid rudiment a cavity appears secondarily either by the development of a fine intercellular split or by the cells along the axis breaking down. The cavity so formed gradually dilates and eventually there is a neural tube agreeing with that of normal forms. The neural tube which has originated in the way described is the rudiment of the central nervous system, its anterior portion becoming relatively enlarged to form the brain while the remainder forms the spinal cord. The central nervous system during the period of its development gradually attains to a condition of the greatest complexity and all that will be attempted here is to give an outline sketch of the more conspicuous changes which take place in its general form and in the arrangement of its parts without going into minute detail. SPINAL Corp.—The spinal cord remains throughout life in the form of a tube the lumen of the tube becoming relatively insignifi- cant while the walls become greatly thickened especially laterally. The relatively small size of the lumen (central canal) is not, as a rule, due merely to its retaining its embryonic dimensions while the walls of the tube are growing in thickness. On the contrary actual 84 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. occlusion of part of the lumen takes place in the great majority of the lower Vertebrates. The side walls of the tube approach one another so as to convert the rounded lumen into a vertical slit and finally they come into contact and fuse so as completely to obliterate the cavity except in its ventral portion which remains open as the definitive central canal. In the case of the Bird—in which the process has been worked out in detail (see Ramoén y Cajal, 1909) the increase in thickness of the wall of the neural tube is due primarily to the cells composing it taking on a tall columnar form—the individual cell extending right from the central canal to the outer surface. The cell-body becomes very attenuated, with a marked dilatation containing the nucleus. The nuclei become necessarily situated at different levels and this in an ordinary transverse section obscures the fact that the wall is still composed only of a single layer of cells. With subsequent development the cells become differentiated into those which are actually nervous and those which remain relatively indifferent and fulfil a mainly supporting function. The latter con- tinue for a considerable period to traverse the whole thickness of the wall. They increase greatly in length: their form becomes more and more attenuated the greater part of their length becoming prac- tically filamentous with small irregular projections and varicosities, while the portion nearer the central canal, in the course of which the nucleus is embedded, remains somewhat stouter. The presence of such supporting cells traversing the whole thick- ness of the wall is only temporary: in later stages they are replaced by the greatly branched neuroglia cells. While many authors have taken the view that these latter are to be regarded as immigrant mesenchyme cells—a view that has weighty general considerations in its favour—Ramon y Cajal and others have adduced strong evi- dence to show that they are simply the original supporting cells which have withdrawn, or lost, their internal and external portions and assumed a complicated branched form. In addition to the comparatively indifferent supporting cells which have just been mentioned there are present in the wall of the neural tube the numerous elements which are destined to become actual neurones or nerve-cells. Such embryonic nerve-cells have been termed by His neuroblasts in contradistinction to the non- nervous elements or spongioblasts. At first isodiametric these cells like their neighbours take on a tall columnar shape stretching throughout the thickness of the wall: their terminal portions become more and more attenuated and they present a spindle-like (bipolar) appearance. Later their shape becomes pearlike the stalk being prolonged into a nerve fibre (neurite, axon) while finally the development of branched projec- tions (dendrites) brings about the definitive multipolar condition. These developing neurones lie in the spaces between the in- different cells and from an early stage (third day in the case of the II NERVOUS SYSTEM 85 fowl embryo) the use of appropriate methods reveals the presence of neurofibrils in their protoplasm. The tail-like prolongation of the neurone which forms the neurite or axon is still believed by the great majority of workers to arise as an actual outgrowth of the cell body as was taught by His. Others regard the appearances upon which this belief is based as being probably deceptive, as will be explained later, The longitudinal axons of the spinal cord are concentrated in its outer layers forming the “ white substance” of the early anatomists. This makes its appearance as a rule in the more primitive Vertebrates as a continuous layer and in the higher forms as a sharply separated dorsal and ventral portion upon each side. The enclosure of the axons in the insulating medullary sheaths commences only within a few days of the end of incubation, in the case of the bird, and at similarly advanced stages of development in other Vertebrates. The sheath is generally believed to be secreted by the protoplasm of the axon. Its formation tends to take place approximately synchronously in all the axons belonging to a particular group. This fact, in conjunction with the use of specific stains for the insulating substance, facilitates the mapping out of the various groups of axons. The spinal cord, like the rest of the central nervous system, becomes invaded during ontogeny by immigrant mesenchymatous tissue. This provides the capillary network which traverses the nervous tissue, in addition doubtless to many other elements of a less conspicuous kind. A curious detail which is noticed in studying sections of develop- ing spinal cord (or brain) is that the active cell-multiplication is confined to the layer next the central cavity, in other words to what was originally the superficial region of the ectoderm. This is in striking contrast with the general ectoderm of the surface of the body where cell-multiplication is confined to the deep (Malpighian) layer. Brain.—The anterior portion of the neural tube becomes en- larged and dilated to form the brain and this gradually becomes so modelled as to present the various regions seen in the brain of the adult. The general course of this process will first be sketched as it occurs in Lepidosiren one of the lower gnathostomatous Vertebrates in which the egg is holoblastic. DIFFERENTIATION OF THE MaIN REGIONS OF THE BRAIN IN LEPIDOSIREN.—The brain rudiment becomes apparent as a slight enlargement of the neural tube. The first sign of differentiation is the appearance of a constriction marking off the primitive fore-brain or cerebrum (archencephalon of Kupffer) from the primitive hind- brain or rhombencephalon. As development goes on this boundary becomes specially marked ventrally where the floor of the brain bulges upwards into the cavity as a transverse fold ' (see Figs. 52 and 53, /). 1 This fold may in other vertebrates make its appearance before the medullary groove is covered in. This is shown clearly in Polypterus—Fig. 80, B, p. 146. 86 EMBRYOLOGY OF THE LOWER VERTEBRATES Cu. The side wall of the hind-brain now comes to project outwards as a prominent knob while the side wall of the fore-brain also bulges out- Fic. 51.—Brain of young Lepidosiren as seen from the dorsal side. A, stage 81+; B, stage 35- ; C, stage 38; D, adult. ¢.H, cerebral hemisphere ; ”.a, neural arch ; pin, pineal body; rh., rhombencephalon ; thal, thalamencephalon ; t.o, tectum opticum. I, II, ete., cranial nerves. [Figs. A and B are more highly magnified than C, and Fig. C than D.] wards—the bulging in this case being the rudiment of the cerebral hemisphere (Fig. 51, A, ¢.H). The portion of the primitive fore-brain lying just in front of the transverse fold of the brain-floor is the infundibulum. Farther forwards the inner surface of the brain-floor forms a transverse II BRAIN, 87 groove bounded behind and in front by a slightly projecting ridge— the rudiments of the optic chiasma and of the anterior commissure respectively (Fig. 53, ch, «.c). About stage 31 a little pocket-like diverticulum of the roof of the primitive fore-brain makes its appearance (Fig. 53, D, pin). This is the pineal body and its appearance is of topographical importance as serving to demarcate the primitive fore-brain roof into thalamencephalic and mesencephalic portions. cH. thal. Zo ot: Fic. 52.—Brain of young Lepidosiren as seen from the left side. A, stage 26; B, stage 31; C, stage 35- ; D, stage 39. ¢.H, cerebral hemisphere; f, primitive fold of brain-floor; inf, infundibulum; o0.b, olfactory bulb; 0.t, olfactory tubercle; pin, pineal body ; thal, thalamencephalon ; t.0, tectwm opticum. The lateral bulgings of the fore-brain have become more pro- minent and now project forwards beyond the limit of the rest of the fore-brain. In the mesial plane between the two hemispheres there projects upwards and forwards a little pocket of the anterior wall of the fore-brain. This is the rudiment of an organ of unknown significance—the paraphysis (Fig. 53, D, par). Soon after the appearance of the pineal body the roof of the primitive fore-brain becomes divided into a posterior portion belong- ing to the mesencephalon and an anterior portion belonging to the thalamencephalon (Fig. 51, B, 4.0, and thal). As development goes 88 EMBRYOLOGY OF THE LOWER VERTEBRATES _ chi. a st ac. G Fic. 53.—Camera tracings of sagittal sections through the brain of Lepidosiren at successive periods of development. Figs. A to F are drawn under the same magnification, Fig. G under a lower magnification. A, stage 25; B, stage 28; C, stage 29+; D, stage 31; E, stage 35; F, stage 38; G, adult in second year. a.c, anterior commissure ; cer, cerebellum ; ¢.H, cerebral hemisphere ; ch, optic chiasma; f, primitive fold of brain-floor; h.g, ganglion habenulae ; h.c, habenular (superior) commissure ; i.g, infundibular gland ; inf, infundibulum ; par, paraphysis ; pin, pineal body; p.c, posterior commissure ; ly, choroid plexus of lateral ventricle ; t.o, tectum opticum ; * originally anterior end of brain-floor. Structures occurring not in the sagittal plane but in sections parallel to but some distance from it, are shaded with oblique lines. II NERVOUS SYSTEM 89 on the constriction between thalamencephalon and mesencephalon becomes more marked. The roof of the former remains thin and membranous, forming the cushion-like dorsal sac upon which the pineal body rests. The roof of the mesencephalon becomes slightly thickened on each side of the mesial plane forming the tectum opticum but correlated with the small size of the eyes in Lepidosiren the thickening never becomes so great as to produce projecting optic lobes such as are formed in most Vertebrates. In the hind-brain region the greater part of the roof, covering in the fourth ventricle, becomes thin and membranous. Across the anterior boundary of the hind-brain the roof does not undergo this secondary process of thinning but persists as a transverse thickened band—the rudiment of the cerebellum. SUBSEQUENT DEVELOPMENT OF THE BRAIN REGIONS RHOMBENCEPHALON or Hind-Brain.—The hind-brain, correlated perhaps with the fact that it contains nerve-centres of supreme importance to life, develops precociously and reaches a relatively enormous size during early stages (Fig. 51, A, 7h). The bulging inwards which marks its anterior limit is doubtless to be regarded as an expression of the active growth in length of its floor during these early stages. During later stages of development it forms a conspicuous pro- jecting restiform body on each side reaching forwards nearly to the anterior limit: of the mesencephalon but this becomes again less and less prominent as the adult condition is approached. The cerebellum retains through life its primitive condition as a simple transverse thickening of the hind-brain roof. MESENCEPHALON.—The roof, as already indicated, becomes thick- ened somewhat on each side (tectum opticum) but not to such an extent as to bulge outwards and form optic lobes. Close to its anterior limit a conspicuous bridge of transversely-running nerve- fibres makes its appearance at a late stage of development. This is the posterior commissure —an important brain landmark (Fig. 53, G, p.c). THALAMENCEPHALON.—The side wall of the thalamencephalon becomes greatly thickened to form the optic thalamus which bounds on each side the slit-like third ventricle. The roof becomes for the most part thin and membranous forming the dorsal sac. On either side of the pineal body however it becomes greatly thickened to form the habenular ganglion. As these ganglia develop a bridge of transverse nerve-fibres makes its appearance uniting them—the superior or, better, habenular commissure. The pineal body as development goes on enlarges somewhat and assumes a carrot shape. Its lumen becomes obliterated posteriorly so that it no longer opens into the third ventricle. The anterior isolated part of the cavity becomes eventually almost filled with t 90 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. granular material produced by the breaking down of the epithelial lining. The paraphysis forms for a time a conspicuous tube passing upwards and forwards in the space between the two hemispheres and ending blindly. In later stages of development it undergoes a relative reduction in size, and becomes irregularly twisted and mixed up with the choroid plexus of the ventricles. On either side of the paraphysis and just dorsal and posterior to its base, the wall of the brain becomes involuted into the third ventricle, the involuted portion being thin and membranous and enclosing an ingrowth of blood-vessels. This vascular ingrowth represents a structure which in most Vertebrates is continuous across the mesial plane with its fellow so as to form an unpaired structure the velum transversum. This is regarded by most writers on the brain as an important landmark in brain topography. On the floor of the thalamencephalon the optic chiasma and the anterior commissure form prominent bulgings into the ventricle. Each develops nerve-fibres in its substance, connected in the one case with the organs of vision and in the other with the cerebral hemi- spheres, especially those portions devoted to the sense of smell. In front of the optic chiasma lies a deep optic recess which is prolonged outwards by an outgrowth of the side wall of the brain, the optic outgrowth, which gives rise to a great part of the eye and will be described later. Behind the chiasma is the infundibulum, the tip of which at a late stage in development (about stage 38) sprouts out into narrow tubular diverticula. These increase in length, wind hither and thither, and partially penetrate into the substance of the pituitary body which lies immediately beneath. The epithelium of these tubular diverticula assumes a glandular appearance and together they constitute the “infundibular gland ”— often called the “nervous portion of the pituitary body.” The series of sagittal sections in Fig. 53 is of interest from its bearing upon a question which has excited some discussion, namely as to what point in the fully developed brain of the vertebrate corresponds to the morphologically anterior end of the brain rudi- ment in earlier stages of development. It has been held by many morphologists, such as von Baer, His, Sedgwick, that the tip of the infundibulum represents the anterior end of the primitive brain, the present condition having been brought about by the anterior portion of the brain becoming bent upon itself into a retort shape. As will be seen by an inspection of the figures the brain of Lepidosiren lends no support to this idea. On the contrary the tip of the infundibulum clearly corresponds to a point close to the letter A of Fig. 53, A. On the other hand, equally clearly the anterior tip (*) of the brain- floor of an early stage such as that shown in Fig. 53, B is represented in the adult by a point well up on the anterior wall of the thalamen- cephalon (lamina terminalis) and just ventral to the root of the paraphysis. II HEMISPHERES 91 ’ CEREBRAL HEMISPHERES.—The hemispheres arise as bulgings of the side wall of the fore-brain. As development goes on they increase in size and grow first dorsalwards and later on forwards until in the adult they are relatively very large. This increase in size is associ- ated with a corresponding growth in the thickness of the wall of the hemisphere—except at its hinder end next the thalamencephalon. Here the inner wall of the hemisphere facing the thalamencephalon remains relatively thin. About stage 35 a small rounded portion of this thin part of the hemisphere wall bulges into its cavity—the lateral ventricle, This ingrowth contains a vascular loop and is the rudiment of the choroid plexus of the hemisphere or lateral plexus. The plexus grows rapidly into the ventricular cavity, forming an irregular crumpled lamina which in the adult attains to great size and complexity traversing the whole lateral ventricle (Fig. 53, F and G, /.y). No doubt this, by diffusion between the blood in its vessels and the fluid in the lateral ventricle, helps to provide for the nutritive and respiratory needs of the hemisphere wall. During the later development of the hemisphere its walls become differentiated into regions in the manner described by Elliot Smith (1908). Most important from the point of view of general vertebrate morphology is the fact that a distinct cortex is developed in the form of a layer of ganglion-cells traversing the roof of the hemisphere parallel to its surface, and at about one-third of the distance from the surface to the ventricular cavity. This cortex extends on the one hand just on to the mesial face of the hemisphere and on the other to a point rather more than one-third of the distance from dorsal to ventral edge on the outer face of the hemisphere. Of this cortical formation, which constitutes the archipallium, the mesial portion corresponds to the hippocampus of higher verte- brates, and the outer portion to the pyriform lobe. The neo- pallium which in the higher forms becomes interposed between these does not appear yet to have become distinctly recognizable in Lepidosiren. Less important from the point of view of general morphology but more conspicuous in their structural expression are certain changes which take place in relation to the olfactory apparatus. The portion of hemisphere wall to which the first cranial nerve 1s attached—the olfactory bulb—is at first simply part of the lateral wall of the hemisphere but as development proceeds it is found to take the form of a sort of cap lying on the dorsal side or roof of the hemisphere at about the middle of its length as viewed from above. This change in position is brought about by an enormous hyper- trophy of the portion of the ventral wall of the hemisphere which lies in front of the optic chiasma—the olfactory tubercle. ; Later on, from stage 38, the portion of hemisphere roof lying posterior to the olfactory bulb undergoes active growth in length with the result that the bulb is gradually carried forwards and 92 EMBRYOLOGY OF THE LOWER VERTEBRATES = cu. eventually comes to lie right at the anterior end of the hemisphere (Fig. 52, D, 0b). At the same time the bulb comes to form a definite hollow projection of the brain surface immediately dorsal to the still greatly enlarged olfactory tubercle (0.t). DIFFERENTIATION OF THE BRAIN REGIONS IN ACANTHIAS.—The development of the brain of Elasmobranchs has been worked out by Kupffer (1906) for Acanthias and his account has been made use of in writing the following short summary. Figures of the early stages of the medullary plate as seen in surface view are given in Chap. XI. The medullary plate projects forwards from the posterior boundary of the blastoderm and is raised well above the general surface. As it increases in length its lateral edges become raised up so that the portion on each side slopes inwards and downwards into a kind of valley. Each half of the medullary plate extends hack into one of the “caudal lobes” which with growth in length come to project freely beyond the edge of the blastoderm. Another result of the increase in length is that the anterior end of the medullary plate comes to project freely forwards over the blastoderm forming a head-fold. Each side of the medullary plate arches inwards towards the mesial plane and the whole becomes converted into a neural tube in a perfectly normal fashion. As in the case of Lepidosiren, the first sign of differentiation of the brain into its parts is a division into primitive fore-brain (Arch- encephalon) and hind-brain. The demarcation is again most distinct ventrally where the brain-floor bulges into the ventricle (Fig. 54, B) as a prominent fold. Later on this fold spreads upwards on each side to the dorsal surface forming the rhombo-mesencephalic fissure which marks off the mid-brain from the hind-brain. It is only at a later stage in development that the mesencephalon becomes marked off by a constriction from the anterior portion of the archencephalon which forms the thalamencephalon. It is of interest to compare sagittal sections through the brain of the Elasmobranch with the corresponding sections already described for the holoblastic lung-fish. Neglecting small differences in detail there is seen to be a striking difference between the two brains—most marked in the middle stages figured—in relation to the longitudinal axis. In Fig. 54, C the Elasmobranch brain is seen to be as a whole strongly curved in a ventral direction: it shows a high degree of “cerebral flexure.” The corresponding stage of the Dipnoan brain is on the other hand almost straight, the superficial appearance of curvature being due mainly to the prominent fold of its floor which projects up into the cavity at the level of the mid-brain. This cerebral flexure, which is especially conspicuous not only in the brain of the Elasmobranch but also in the other types of Brain (Mammalian and Avian) that were first investigated development- ally, has played a large part in discussions on brain morphology. Thus the idea, already alluded to, that the tip of the infundibulum is the morphologically anterior end of the brain rests upon it. II BRAIN 93 That this idea is unsound seems clearly to be indicated by such series of sections as that shown in Fig. 53. Asa matter of fact it Fic, 54.—Sagittal sections through the brain of Acanthias. (After Kupffer.) A, 3°3 mm. embryo; B, 7-8 mm.; C, 10 mm.; D, 27 mm.; E, 70 mm. cer, cerebellum; ch, optic chiasma ; ect, external ectoderm; h.c, habenular commissure ; inf, infundibulum ; J7, cavity of mesen- cephalon ; N, notochord; n.p, anterior neuropore ; p.c, posterior commissure ; pin, pineal organ; Rh, cavity of rhombencephalon ; t.0, tectum opticum ; v, velwm transversum ; v.IV, fourth ventricle. seems probable that very pronounced flexure of the brain, such as is seen in the developing Elasmobranch, is to be regarded as a secondary 94 EMBRYOLOGY OF THE LOWER VERTEBRATES Cu. result of the heavily yolked condition of the egg. As a result of the concentration of yolk towards the ventral side in such heavily yolked Vertebrates the processes of growth are retarded upon that side. But it is clear that retardation of growth in length on the ventral side as compared with the dorsal would bring about a flexure towards the ventral side. That the cerebral flexure is due rather to such a general cause than to any inherent peculiarity in the brain itself is supported by the fact that the notochord is also strongly flexed (see Fig. 54, C). As a consequence of these considerations, we are inclined to take the view that the phenomenon of cerebral flexure is of much less fundamental morphological significance than is commonly supposed. Comparing the later stages figured for Acunthias with those of Lepidosiren, it will be seen that the brain shows the same elements although these differ in their relative size and other features in the two cases. Thus the cerebellum of the sharks—correlated with the active and complex movements of these fishes—becomes much more developed. It grows greatly in anteroposterior extent and that causes it to bulge outwards as shown in Fig. 54, E (cer). The pineal body is slender and elongated in form: the velum forms a conspicuous infolding of the thalamencephalic roof continuous across the mesial plane. The wall of the anterior portion of the primitive fore-brain under- goes a fairly uniform increase in thickness throughout with the excep- tion of a transverse band just in front of the velum which becomes thin and membranous. This portion of the brain increases somewhat in transverse diameter so that it is broad in shape as seen from above, but there is no definite bulging in its side wall to form a distinct hemisphere. The material that would ordinarily go to form the hemispheres remains here in the thickness of the wall. The olfactory bulb arises as a slight projection from the side wall of the fore-brain, but as development proceeds and the olfactory organ becomes removed from the brain by the interposition of mesenchyme the olfactory bulb remains in contact with the olfactory organ, its attachment to the brain becoming drawn out into a more or less elongated stalk the olfactory peduncle or olfactory tract. The salient features in the establishment of the topography of the Vertebrate brain have been illustrated in outline in the sketch which has just been given. It would be beyond the scope of this work to make any attempt to fill in the picture in detail but it is necessary to recall a few points which are of interest to morphologists apart from specialists in neurology. It should in the first place be borne clearly in mind that the brain—like indeed the whole of the nervous systein (see below, p. 118) is to be looked upon as a fundamentally continuous structure. The parts which compose the adult brain—medulla, cerebellum, mesen- cephalon and so on—are not to be regarded as constituent units which go to build up the complete brain, but rather as specialized portions of a once homogeneous whole. The process of specialization ae ae Il BRAIN 95 has probably been linked up more particularly with the processes of localization or centralization of particular functions in particular brain regions. When this has come about, increase in the number of ganghon-cells devoted to the particular function will cause an increase in bulk of that portion of the brain in which they are situated and it will assume definite characteristics of its own. The first step in the development of such a brain region consists in the mere thickening of the brain wall but with still greater increase in the number of cellular elements involved mere increase in thickness becomes insufficient for their accommodation and an increase in area comes about in addition. This necessarily causes a bulging of the particular part of the brain wall and some of the most characteristic differences between the brains of different types of Vertebrate depend upon whether the bulging takes place outwards or inwards. Thus in the majority of Vertebrates the cerebellum bulges out- wards as has been indicated in the case of Acanthias. In Teleostean fishes on the other hand this is the case with only the hinder part of the cerebellum: its anterior portion in these fishes bulges downwards and forwards underneath the roof of the mesencephalon forming the well-known valvula cerebelli. In the more primitive ganoid fishes on the other hand such as Polypterus (Graham Kerr, 1907) the hind. portion of the cerebellum also grows inwards, so as to form a structure projecting back into the fourth ventricle in just the same fashion as the valvula cerebelli projects forwards. A somewhat similar ditference appears to be present in the case of the hemispheres. These originate in most subdivisions of the Vertebrata as paired bulgings of the wall of the primitive fore-brain, and the present writer agrees with Studniéka (1896) in feeling com- pelled to accept on this ground the view taught by many of the older morphologists such as von Baer, Reichert and Goethe that the hemi- spheres are to be looked on as fundamentally paired structures, rather than the view, more fashionable of recent years, which regards the portion of the primitive brain lying in front of the velum and optic recess as forming together with the hemisphere region an unpaired complex (Telencephalon—His). The more complete knowledge that we now possess regarding the develop- ment of the brain in the more primitive Vertebrates with holo- blastic eggs, seems to the writer to make it clear that the reasons which have led to a departure from the older view can no longer be regarded as adequate. We take it then that the hemispheres are fundamentally paired projections of the side wall of the primitive fore-brain. Physiologically they are probably to be regarded as portions of the brain wall which have become specially enlarged in relation with the sense of smell, just as are the optic outgrowths in relation with the sense of sight. . Now whereas in the majority of Vertebrates the hemispheres bulge outwards, in the more primitive Teleostomes (e.g. Polypterus, 96 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. Graham Kerr, 1907) they bulge inwards. In the typical Teleosts what apparently corresponds to the hemisphere forms simply a solid mass projecting into the cavity of the fore-brain, the structure which is usually and probably erroneously spoken of as the corpus striatum in these fishes. A part of the brain which is of very special morphological interest is the thalamencephalon—which persists with singularly little change throughout the series of Vertebrates. Even in Amphiowus sagittal sections through the front end of the neural tube present appearances which vividly suggest the thalamencephalon of the more typical Vertebrates (Kupffer) and raise the question whether—as is probable enough on other grounds—the head region in Amphioxvus is degener- ate and once possessed a brain. Amongst the structures connected with the thalamencephalon special interest attaches to the pineal body.t| So far this has been alluded to merely as a comparatively simple diverticulum of the thalamencephalic roof. In the majority of Vertebrates it remains comparatively uncomplicated and its main function appears to be that of forming a peculiar internal secretion. In two sets of Vertebrates—the Lampreys on the one hand, and Sphenodon and many Lizards on the other—there becomes developed in relation to it an organ possessing a close resemblance to an eye, of the “camera” type, possessing a retina and in some cases a lens. The organ appears to be functional as the tissues overlying it are commonly free from pigment and its retinal cells on exposure to light show a change of position in their pigment granules similar to what is commonly found in visual organs. Though functional it does not follow that the organ serves for the detection of what we call light : it may be that its sensitiveness is rather towards radiant energy of other wave-lengths than that included within the range of the visible spectrum. There is a general tendency amongst those who have carried out researches upon the pineal eye to regard the eyelike condition as a relatively archaic condition of the pineal organ—a tendency which is encouraged by the evidence of palaeontology that certain ancient Tetrapods of the palaeozoic and mesozoic periods possessed a highly developed pineal organ—the skulls of these animals possessing a relatively huge parietal foramen, corresponding with the foramen in the roof of the skull of modern lizards in which the pineal eye lies embedded. The evidence of embryology indicates that the most archaic con- dition of the pineal organ was a simple diverticulum of the brain roof projecting towards the skin on the dorsal surface of the head. There is no clue whatever as to the original meaning of this diverti- culum. But we do know from the study of invertebrates that where tissue rich in nerve-elements comes to be exposed to light there is 1 An admirable account of the structure and development of this region of the brain by Studnigka will be found in Oppel (1905). ‘ II PINEAL ORGAN 97 frequently shown a well-marked tendency to the evolution of eye-like structure. In Molluscs for example we find eyes developing on the edge of the mantle (Pecten), round the tips of the siphons (Cardiwm sp.), on the dorsal surface of the body (independently in Chiton and Oncidiwm)—and similar instances might be quoted from other groups. Bearing such facts in mind one is compelled to acknowledge the possibility, if not probability, of such a projecting piece, of nervous tissue as the pineal diverticulum, lying close under the surface of the head on its dorsal side, in the position where light stimulus would be most pronounced, developing secondarily in some cases into an organ of the nature of an eye. Discovered by Leydig (1872), its structure investigated by Spencer (1886) and other workers, the development of the pineal eye has formed the subject of a number of excellent researches. It will be convenient to take as an example that of the common lizards of the genus Lacerta (Novikotf, 1910). The first indication of the organ appears in embryos of about 3 mm. in length in the form of a thickening of the thalamencephalic roof, in the region of the mesial plane, and divided by a transverse furrow on its outer surface into a smaller anterior and a larger posterior portion. This thickened part of the brain roof comes to bulge outwards and forms a prominent projection (Fig. 55, A) the groove dividing it externally into anterior and posterior portions being still visible though less distinct. The projecting pocket now grows forwards parallel to, and in close contact with, the brain roof (Fig. 55, B), its forwardly projecting portion becoming constricted off from the rest. The constriction in question deepens and the anterior portion (parapineal body) becomes nipped off to form a completely closed vesicle (Fig. 55, C)—the rudi- ment of the eye. As the external ectoderm recedes from the brain roof, with the increase in the amount of mesenchyme between the two, the parapineal vesicle remains close to the ectoderm and con- sequently retreats from the brain surface (Fig. 55, D). The eye is now ‘seen to be connected with the brain wall by a distinct optic nerve which, in full accordance with the view taken in this book with regard to nerve-trunks in general, is merely a primary bridge which already existed (Novikoff) at a time when the eye vesicle and the brain roof were still in immediate contact and which simply became extended in length as the gap between eye and brain became greater and greater (Fig. 56, p.m). Nerve-fibres develop in this optic nerve which pass at their cerebral end into the habenular commissure. Transverse sections through a 9-mm. embryo show that the fibres on entering the commissure bend away to the right, passing eventually to the right habenular ganglion. In this connexion with the right habenular ganglion Lacerta resembles the other lizard Jguana tubercu- lata (Klinckowstrém, 1894) but curiously differs from Sphenodon where according to Dendy (1899) the connexion is with the left habenular ganglion. Vols II . : H 98 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. Meanwhile the wall of the vesicle lying next the outer skin pu. Fig, 55.—Sagittal sections through the pineal organ of embryos of Lacerta. (After Novikoff, 1910.) A, L. vivipara, 3mm.; B, ditto, 4mm.; C, L. muralis, 6 mm.; D, L. vivipwra, 9mm. ect, external ectoderm; J, lens; p.e, pineal eye; p.n, pineal nerve; p.s, pineal stalk; pin, pineal outgrowth ; thal, roof of thalamencephalon, ul PINEAL ORGAN 99 assumes a lenticular form, its cells becoming much elongated though remaining in a single layer. This lenticular thickening occasionally becomes lost during development—a fact which may be taken as forming a piece of evidence in favour of the view that the eye, at the present time, is in a retrogressive phase of evolution. Those parts of the vesicle wall which do not take part in the formation of the lens undergo histogenetic changes into retinal tissue. The cells undergo differentiation in two directions. The one set become pigment cells—tall columnar cells which traverse the whole thickness of the retina and have their nuclei towards the basal or outer end and which develop dark melanin granules in their proto- pir. plasm. pe. d Interspersed with these are the s percipient cells, shorter in form, their basal ends not reaching the outer surface, and carrying at their inner ends cilium-like structures which project into the cavity of the vesicle. The idea that these projec- tions correspond physiologically to rods appears to be negatived by the fact that they occur also on the inner ends of the cells forming the lens. At their basal ends these cells are continued into nerve-fibres, which form a distinct layer internal to the nuclei of the pigment cells jo are ee eventually continuous physiologi- : ; : cally with the fibres of the pineal eee cia ; nerve. Scattered amongst,and in the .b6,ubumlar commiware; mo, posn neighbourhood of, this fibrous layer p.s, pineal stalk ; par, paraphysis. ganglion-cells are present: they are about the first detinite elements to become recognizable during the histogenesis of the retina and appear first close to the point of attach- ment of the optic nerve. The cavity of the pineal eye is kept distended by a clear substance, the vitreous body, and this is colonized by a certain number of cells (see Fig. 55, C) which are most probably to he regarded as immi- grant mesenchyme cells. In Sphenodon, the sole survivor of the: only other existing group of Reptiles in which a pineal eye is present, the development of the organ according to Dendy (1899), who has worked it out in detail, agrees with that of Lacerta in its main features. ic In the Lampreys, also, somewhat eye-like developments occur 1n the pineal region. In the adult two vesicles—a dorsal (“ pineal ”) and a ventral (“parapineal”)—-are found overlying the roof of the 100 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. thalamencephalon. In each of these the lower wall shows histological characteristics of retinal tissue and each is in continuity with the brain—in the case of the parapineal organ directly and in the case of the pineal by an elongated stalk containing nerve-fibres. The parapineal organ lies in some cases (Geotria—Dendy, 1907) slightly to the left of the pineal and its nerve-fibres have been traced into the left habenular ganglion while those of the pineal organ have been traced to the right habenular ganglion. In neither case does the outer wall of the vesicle show any signs of thickening to form a lens—so that neither organ can form an image—but the overlying tissue is comparatively transparent so that diffuse heht stimulus can reach it. According to Studniéka the two organs develop as evaginations of the brain roof one (parapineal) in front of the other. The para- pineal evagination soon loses its lumen and becomes solid and it is noteworthy that at first it is continuous on each side with the habenu- lar ganglion of that side. Later on it becomes by differential growth shifted far forwards, away from the region of the habenular ganglia, and it loses its connexion with the right ganglion while it remains connected by nerve-fibres with the left. The two questions of special interest which present themselves in regard to the pineal and parapineal organs are (1) were they originally ocular in structure and function and (2) were they paired or unpaired ? (1) It is obvious that the presence of an eye-like pineal or para- pineal organ in certain Reptiles and in Lampreys, and of a large parietal foramen in the skull of various extinct Vertebrates suggests the possibility of these organs having had the form of visual sense organs in the ancestral Vertebrate. Against this however must be set the fact that in all other Vertebrates than those mentioned, including such relatively archaic forms as Elasmobranchs, Cross- opterygians, Dipnoans and Urodeles, there is no trace whatever of eye structure. It seems highly improbable that a well-developed visual organ once present on the dorsal side of the head in the ancestors of Vertebrates should have disappeared without leaving a trace in all the varied groups, with their very different modes of life, outside the limits of the Lampreys and Reptiles. To the present writer it does not appear that the evidence, so far as it exists at present, is anything lke convincing that the pineal eye 1s an ancestral feature of Vertebrates in general rather than a mere secondary development. (2) Various recent investigators of the pineal organs are inclined to look upon them as being originally paired structures, the pineal organ in the strict sense being the right-hand member of the pair and the parapineal organ the left. This is perhaps most clearly suggested by the Lampreys in which the parapineal organ is con- nected by nerve-fibres with the left habenular ganglion and the II NEUROMERY 101 pineal organ with the right (though also with the posterior com- missure). Again in various Vertebrates (Teleostomatous fishes— Hall, 1894: Amphibians, Birds—Cameron, 1903, 1904) the parapineal organ is in early stages slightly to the left of the pineal organ. On the whole it does not appear to the present writer that the evidence is sufficient to make the view probable. In the Lampreys the connexion of the parapineal body with only the left habenular ganglion appears, as indicated above, to be secondary : it is originally connected with both right and left. Again, to turn to the Reptilia, the eye is in Sphenodon connected with the left habenular ganglion and in the Lacertilia with the right, although it seems perfectly clear from the figures given by Dendy and Novikoff respectively that the eye is morphologically the same organ in the two types mentioned. Were it not so we should le driven to the position that of a pair of pineal eyes originally present one has disappeared entirely in Sphenodon and the other has disappeared equally completely in the lizards. The improbability from a physiological point of view of this having happened need not be accentuated. Here again, then, there seems to be up to the present no sufficient reason for departing from the view that pineal and parapineal organs were primitively median in position, one in front of the other. As to their original significance we have no obvious clue: the absence of convincing evidence that they were originally eyes does not of course preclude the possibility of their having been originally some kind of sense organ. NEvRoMERY.—It has been noticed in various Vertebrates, particu- larly Elasmobranchs, Amphibians and Birds, that the neural rudi- ment while still in the form of an open plate is sometimes divided up by numerous and regular transverse markings. Whether this appear- ance of segmentation is caused simply by the active growth in length of the medullary plate or whether on the other hand it has some deeper significance has not been conclusively determined. The name neuromere has been given to the apparent segments. That these are really primitive morphological segments as is believed by many and as is implied in the termination “-mere” seems improbable. The existence of a clearly marked segmentation of the nervous system where it occurs—in the phyla Annelida and Arthropoda—is brought about by the concentration of ganglion-cells in serially repeated masses, in correlation with the presence of serially repeated appendages (parapodia in Polychaeta), and there is no sufficient evidence to show that such were ever present in the ancestral Vertebrate. The tact that the longitudinal muscles are divided into myotomes would not be sufficient by itself to account for the external form of the central nervous system being segmented, for in that case the segmentation would be still clearly marked in the many fishes where the myotomes remain practically unmodified. During later stages, after the neural tube has become closed in, “neuromeres” are particularly conspicuous in the brain region. 102 EMBRYOLOGY OF THE LOWER VERTEBRATES CH. They are sometimes very distinct in the hind-brain of fowl em- bryos of about the fourth day (see Fig. 236 in Chap. X.). It is however an outstanding characteristic of the head region as compared with the trunk that the segmentation of its mesoderm has become blurred and in great part obliterated. It is under those circumstances improbable that a primitive segmentation of the central nervous system, which. is in its nature linked up to a segmentation of mesodermal structures, should have remained particularly distinct in a region where the mesodermal segmentation itself has become particularly obscure. : The appearances mentioned seem to be adequately explained by the active growth of the developing brain within its confined space, combined with the presence round it of mesodermal tissue with vestigial segmentation. It will be noticed in the figure referred to above that the dividing lines between the “neuromeres” are spaced out at exactly the levels where we should expect to see boundaries of mesoderm segments were the existing series prolonged forwards. Segments are no longer visible in this region but there is, as will appear later, convincing evidence that the series of segments did formerly extend through this region now occupied by continuous mesenchyme. It may well be that the individuality of the segments, no longer visible as such, is still expressed by a difference in consistency of the mesenchyme, sufficient to mould by its varying resistance the actively growing hind-brain as it presses against it. DEVELOPMENT OF THE PERIPHERAL NeERvES.—The development of the peripheral nerves of Vertebrates has been the subject of a large amount of investigation, partly on account of its intrinsic interest and partly on account of its bearing upon physiology and pathology. In spite of the labours of numerous investigators the problem—tfor we may take it that the mode of development is funda- mentally the same throughout—has not yet by any means been satisfactorily decided. ' While bearing in mind the undesirability of making use of modern facts merely to support, or to undermine, old hypotheses, it will be convenient to approach the question by stating shortly the three prevalent views as to the main features of the development of peripheral nerve-trunks neglecting differences in detail. For shortness these three views may be termed after their most prominent supporters (1) the His view, (2) the Balfour view and (3) the Hensen view. THE His View (Outgrowth theory).—This hypothesis may be said to have been founded by Kupffer in the embryological portion of Bidder and Kupffer’s work (1857) on the spinal cord. As however Kupffer later on gave up the view, in favour of that of Balfour, the hypothesis now under consideration is commonly associated with the name of His, who played the main part in building up the theory and who fully deserves to be regarded as its principal founder. II NERVE DEVELOPMENT 103 It is to be noted in passing that Kupffer’s original observations were made upon Mammals and those of His (1868) upon the Fowl. In other words, in both cases the embryos were such, in regard both to the minute size of their cell elements and to their high position in the Vertebrate scale, as to be unsuited to provide a reliable basis for the generalization that has been built upon them. The His view as expounded. by one of its most distinguished supporters S. Ramén y Cajal (1909) may be summarized as follows, the case of the motor nerves being taken for the sake of simplicity. Each motor nerve-fibre arises as an outgrowth from a neuroblast, or young ganglion-cell, lying within the spinal cord. The fibre sprouts out from the neuroblast, makes its way to the surface of the spinal cord, perforates that surface, and proceeds to grow freely through the mesenchyme. The free end of the fibre forms a “cone of growth,” commonly shaped somewhat like a grain of barley and with a pointed end. This “cone of growth” shows an active amoeboid movement, by which it insinuates itself through the interstices of the mesenchyme. Sometimes it may be seen to flatten or mould itself against obstacles in its path. In the Fowl these processes take place during the third and fourth days. Eventually (about the fifth day, in the Fowl, in most cases) the growing nerve-fibres reach their destination and become joined up to the muscle cells which form their definitive end-organs. The essential feature of the His view is that the nerve-fibre (which already shows the characteristic specific reactions of a nerve- fibre, e.g. on impregnation with silver salts, and is therefore not merely a strand of undifferentiated protoplasm) sprouts out from the central nervous system and grows through the intervening mesenchyme with a free end until it becomes joined up secondarily with the end-organ. The His view at the present day rests upon a large body of observed facts. In studying the embryology of almost any Verte- brate it is easy to find in sections what appear to be freely ending nerve-fibres sprouting out from the spinal cord. Some of the most beautiful preparations of this kind have been made by Ramon y Cajal and others by the use of silver impregnation methods. Perhaps the most striking evidence, which has recently been adduced in favour of the His view, has been obtained by experi- mental methods, especially by Harrison (1908, 1910). In one set of experiments which have been regarded as particularly convincing Harrison removed small pieces of embryonic spinal cord from Frog embryos at a period just before that at which the motor nerves became visible, and was able, by using ordinary antiseptic pre- cautions, to keep them in a living condition for relatively prolonged periods mounted in a drop of sterile lymph under a coverslip upon a slide. The lymph soon clotted and held the piece of spinal cord in position. Harrison now observed in many cases little projections 104 EMBRYOLOGY OF THE LOWER VERTEBRATES — cu. making their appearance from the pieces of spinal cord which he identified as rudimentary motor nerves. Any possible doubt as to the correctness of this identification was removed by Burrows (1911) repeating the work on the chick and obtaining the specific staining reaction of neuro-fibrils in the structures in question. These nerve-rudiments when kept alive under the conditions mentioned were observed to increase rapidly in length, the rate of growth being in one case as high as 56» per hour. The end of the rudiment (Fig. 57) was somewhat enlarged and projected into fine protoplasmic tags which showed active amoeboid movement. It is this amoeboid protoplasm at the free end of the fibre which, in Harrison’s belief, is the active agent in the extension of the nerve-fibre. As to the method by which it is, in the actual body, guided along the proper path to its destination, Harrison does not commit himself, but he appears to have a leaning towards the view held by Ramoén y Cajal that it is mainly a matter of chemiotaxy. In the words of their author (1908) these “ experiments place the outgrowth theory of His upon the firmest possible basis,—that of direct observation. The attractive idea of Hensen must be abandoned as un- tenable.” It should be added that the His theory fits in very well with current views in physiology and pathology— in particular with the fashionable neurone doctrine, according to which the cellular units which compose the nervous system are not in organic continuity with one another. Ob- viously this hypothesis and the — poe oe outgrowth hypothesis, according to oman live Taepandlon, B svenky- which the nerve-fibre is for a time five minutes later than A. (After Separated by a gap from its -end- Harrison, 1908.) organ, lend one another mutual support. 1 It must be borne in mind, however, that the histological basis of the neurone theory is not universally admitted to be beyond suspicion. Its main foundation consists of observations by the Golgi and similar methods of metallic impregnation. In preparations made in this way single cellular units are frequently picked out without the reaction taking place in neighbouring units arranged in series with them. A ganglion-cell A with its axon and terminal branches stands out deep black in the preparation, while the ganglion-cell B, next to it in the series, shows no reaction. Such an observation obviously suggests discontinuity. The possible fallacy in these observations lies in the fact that the stain used is not a true stain in the ordinary sense of the word but merely a precipitation of metal upon II NERVE DEVELOPMENT 105 The His view is concerned primarily with the actual functional nerve-fibres. As regards the primitive sheath (Gray Sheath of Schwann), in which these fibres are enclosed, the His view regards it as being formed by mesen- chyme cells which apply themselves to, and spread out over, the surface of the originally naked nerve-fibre. (2) Tue BaLrour View (Cell-chaintheory). —While Schwann(1839) long ago described the multicellular structure of nerve-trunks in the foe- tuses of mammals, it was F. M. Balfour (1876) who really founded the view that the nerve - trunk arises in development from a chain of cells. Fre. 58.—Section through the dorsal part of the trunk Balfour found in Elasmo- 0° 2 Torpedo embryo. (From Balfour’s Embryology.) branch embryos that the d.r, dorsal root; g, spinal ganglion ; my, myotome ; N, noto- nerve-trunk was repre- oe a n, nerve-trunk ; ne, cavity of spinal cord; v.7, ventral sented by a chain of cells in early stages (Fig. 58, v.r), and similar observations have been made by subsequent observers. According to this view the whole nerve- trunk is multicellular in origin, the cells not only forming the sheath of the nerve-trunk but also giving rise to the nerve-fibrils which come into existence traversing the cellular strand from end to end. On the question of the origin of the cells which constitute the nerve-rudiment opinions vary. Most supporters of this view have regarded them as having emigrated from the spinal cord (e.g. Balfour, van Wijhe, Dohrn): while others (Koélliker) have looked on them as mesenchymatous in nature. Sedgwick took this latter view and as he regarded the mesenchyme as a continuous syncytium, the bridges connecting the cells being primitive—persisting from the the surface of the cell and its processes. We know from the recognized unreliability of the method that the occurrence, or not, of this precipitation is liable to be decided by extremely delicate chemical differences. We know further that the axis cylinder, however it arose in development, is morphologically and physiologically a prolonga- tion of the cell-body (ganglion-cell), and therefore that its metabolism is under the control of the nucleus of that cell-body. The individuality of the cell and its pro- longation, due to the metabolic control by its own special nucleus, is probably quite enough, in itself, to account for a chemical character of its surface sufficiently different from that of its neighbours to influence the precipitation without there being, as the neurone theory assumes, any absolute discontinuity. 106 EMBRYOLOGY OF THE LOWER VERTEBRATES | cH. incomplete separation of the cells during the processes of segmenta- tion and cell division—the view in his hands came to approach the next view to be mentioned that of Hensen. ; (3) Tue HENSEN ViEW (Primitive continuity theory)—This view has found its strength in general physiological considerations rather than in convincing facts of observation. According to Hensen (1864, 1868, 1903) the nerve which connects centre with end-organ is a primary connexion which has been there from the beginning. It existed first as a simple bridge of protoplasm, such bridges being present between the various cells of the body owing to the fact that the processes of segmentation and cell division are not complete so as to lead to absolute isolation of the cells or segments from one another. According to this view the growth in length of a nerve-trunk is simply the extension of a pre - existing bridge, as the organs at its two extremities —centre and end- organ — are pushed apart from one an- other during the course of development. Fic. 59,.—TIllustrating Hensen’s view of the origin of peri- Hensen figures in pheral nerves. The section is taken from a 9-day Rabbit jg papers (Fig. 59) embryo, and passes through the trunk region. After Hana 1903.) a rae ( what he takes to be such nerve-rudiments, in the form of numerous fine filamentous structures passing across the space between spinal cord and myotome. There is however no evidence to show that these filaments have anything to do with nerve-trunks. Although for this reason it is impossible to accept the main observational basis of Hensen’s view, that does not neces- sarily invalidate the physiological considerations which may be held to give an a priort probability to the correctness of his general theory. The three views which have been outlined above were fashioned by their respective authors long ago as embryological science goes. Since then new facts have hecome known which have to be taken into account when considering their acceptability as working hypotheses at the present time. Some of these facts will now be touched upon. DEVELOPMENT OF PERIPHERAL NERVE-TRUNKS IN LEPIDOSIREN It is obvious enough from the diversity of statements by skilled observers that the investigation of the method of development of the peripheral nerves in Vertebrates is beset by technical ditticulties and resulting liability to error. In such a case it is of special importance to choose for investigation types of animal in which this lability to error is reduced to its narrowest limits. Such an animal should t I NERVE DEVELOPMENT 107 be, on the one hand, comparatively archaic—it should belong to one of the relatively more primitive groups of Vertebrates—-and, on the other hand, its histological texture should be as coarse as possible, its cell elements being of large size. Amongst Vertebrates investigated up to the present time in regard to nerve-development Lepidosiren (Graham Kerr, 1904) is unrivalled in its combination of these qualifications and a summary will now be given of the main features which’ have been made out from the study of the development of the motor nerves in this anunal. It will be convenient to commence with the fully formed nerve-trunk and then work backwards towards the earlier and more obseure stages. Fig. 61 represents a portion of nerve-trunk from a fully developed larva of stage 34. The nerve-trunk consists of a cylindrical bundle of nerve-fibrils, dotted over the surface of which are the numerous large nuclei of the protoplasmic sheath. The sheath itself is so thin as to be practically invisible even under a high-power immersion objective except in the angle close to a nucleus where it is distinctly visible. Fig. 60, D is taken from a larva ten days after hatching. At this stage the nerve-trunk, when examined superficially, has the appear- ance of a thick strand of protoplasm containing numerous nuclei ora chain of cells. Careful examination of well-fixed and well- Stained specimens shows however that this conspicuous mass of protoplasm is really only the sheath, the true nerve-trunk (m) being visible traversing it from end to end. Scattered about in the thick sheath of this stage there are still to be seen granules of yolk (black in the figure) which have not yet been used up. Fig. 60, C is taken from a larva at the time of hatching. At this- stage the nerve-trunk is a well-developed bundle of nerve-fibrils, just as in the later stages, but throughout the greater part of its length it is devoid of a sheath of protoplasm. In the section figured the sheath is visible as a mass of nucleated and heavily yolked protoplasm enclosing a portion of the nerve-trunk towards its outer end. This mass of protoplasm is obviously just a condensed part of the general mesenchyme which is scattered about in the form of irregular heavily yolked masses throughout the spaces between the main organ systems. The mass in question is identical with the rest of this mesenchyme in its various features and every here and there it is continued into it without a break. The section figured shows the whole length of the motor nerve-trunk from the ventral root to the myoblasts or muscle cells which form the myotome. Towards its outer end the trunk is seen to break up into numerous diverging strands which are directly continuous with the protoplasm of the myoblasts (see below, p. 204). Fig. 60, B is taken from an embryo about three days before hatching. At this stage the myotome has barely commenced to % 108 EMBRYOLOGY OF THE LOWER VERTEBRATES Fic. 60.—Portions of transverse sections through young Lepidosirens to illustrate the development of the spinal nerves (ventral roots). A, stage 24; D, stage 29+. m, myotome; 7, nerve-trunk ; s.c, spinal cord ; sh, sheath, CH. aa NERVE DEVELOPMENT Fic. 60a.—Portions of transverse sections through young Lepidosirens to illustrate the development of the spinal nerves (ventral roots). B, stage 25; C, stage 27. m, myotome; n, nerve-trunk ; s.c, spinal cord; sh, sheath 109 110 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. recede from the spinal cord, but yet each motor nerve is already present as a distinctly fibrillated trunk bridging across the narrow gap between spinal cord and myotome. A few mesenchyme cells _ have wandered into the gap but they have not yet begun to con- centrate round the nerve-trunk. Fig. 60, A is taken from an embryo of stage 24 at a time when myotome and spinal cord are still in close contact with one another. In specimens which were extended in one plane under normal salt solution while still glive and subjected to the action of the fixing agent in that position, it is found that the myotome is frequently pulled slightly away from the spinal cord (as in the specimen figured) and in such cases it is found that the nerve- trunk already exists in the form of a bridge of soft granular protoplasm (n) with- out any trace of fibrillation, connecting spinal cord and myotome. That these bridges are really the nerve-trunks is indicated by their occur- rence one to each myotome, apart from the fact that a con- tinuous series of stages have been observed between them and the fully developed nerve-trunks. In summing up we may Fie. 61.—Part of transverse section of Lepidosiren take the various stages in (stage 34), showing a portion of nerve-trunk. their proper chronological my, myotome; N, notochord; n, nerve-trunk ; n.S, nucleus sequence. of nerve sheath; $1, primary sheath of notochord ; XJ, 1 Th é - 2 lateral branch of vagus. ( ) e nerve-trun 18 already present as a proto- plasmic bridge at a period so early in development that spinal cord and myotome are still in contact with one another. (2) As the embryo grows and the myotome recedes from the spinal cord this protoplasmic bridge increases in length and becomes fibrillated. (3) As the nerve-trunk lengthens amoeboid masses of mesen- chymatous protoplasm collect round it and gradually spread out over its length to form the protoplasmic sheath. In stages later than those figured the sheath protoplasm insinu- ates itself in amongst the nerve-fibrils of the trunk, dividing them up into bundles or nerve-fibres. As the myotome resolves itself into the various muscles of the adult each piece retains its primitive nerve-strand, drawing it with it, as it becomes pushed about by the processes of differential growth, as its own special nerve. Tt should be mentioned that the most important point in the Ir NERVE DEVELOPMENT | 1 above description—-the existence of the motor trunk in the form of a bridge of protoplasm between myotome and spinal cord at a time when they are still in close proximity—has been confirmed for an- other very primitive group of Gnathostomes, the Elasmobranchs (Paton, 1907), as is shown in Fig. 62. It will now be convenient to review the facts just described for Lepidosiren in relation to the general theory of nerve-development. (1) It is clear in the first place that the His view is put out of court, seeing that before there is any development of nerve-fibrils the motor nerve-trunk already exists in the form of a bridge of proto- plasm connecting spinal cord and myotome. (2) It is equally clear that the Balfour view is inapplicable: the nerve-rudiment cannot in early stages by any possibility be re- SC. garded as a chain of cells, seeing that its total length is greatly less than the diameter of a single cell-nucleus. (3) While the nerve-rudiment forms a primary connexion be- tween spinal cord and myotome, in the sense that it is in existence before these organs begin to recede from one another, there is no evidence by which the connexion can be traced back to intercellular bridges or plasmodesms (Stras- Fic. 62. —Part of transverse section through burger 1901) of early eg. segmen- ad Pine mate of arian ison, Tani > ? ing e motor nerve-trun rudiment. tation, stages in the development (After Stewart Paton, 1907.) of the egg, as would be the case my, myotome ; n, nerve-trunk ; s.c, spinal cord. according to Hensen’s theory. (4) The primitive protoplasmic bridge gradually becomes fibril- lated but there is no means of determining with any degree of certainty how these fibrils are developed. It is suggested! that the development of the actual nerve-fibril is simply the gradual coming into view of a pathway produced by the repeated passage of nerve impulses over a given route. It is clear from the study of the simpler organisms that one of the most ancient properties of living protoplasm is that of the trans- mission of impulses through its substance. Although nothing is really known as to the precise nature of living impulses it is reasonable to suppose that they involve changes in the distribution of energy analogous to those involved in the passage of an electric disturbance. If this be the case their passage between two points will be determined by the relative potential, and the route along which the impulse passes will be that of least resistance. If the conductivity of the 1 Graham Kerr, 1904. It has been pointed out that similar suggestions in regard to the nervous system in general were made long ago by Herbert Spencer. 112 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. living substance were uniform the path would be a straight line joining the two points: if the conductivity were not uniform on the other hand the path would be diverted along routes of high conductivity where the total resistance would be at its minimum. Looking at matters from such a point of view we should regard a motor ganglion-cell at the moment of functioning as a centre of high potential and its muscle ending as of low potential, while a sensory cell at the moment of its functioning would be a centre of high potential and the central termination of its nerve-fibre as at relatively low potential. In early stages of evolution, whether phylogenetic or ontogenetic, we may take it that vital impulses flitted hither and thither in an indefinite manner within the living substance and that one of the features of progressive evolution has been the gradual more and more precise definition of the pathways of particular types of impulse, as well as of the transmitting and receiving centres between which they pass. We may then regard the appearance of neuro-fibrils within the protoplasmic rudiment of the nerve-trunk as the coming into view of tracks, along which, owing to their high conductivity, nerve impulses are repeatedly passing.! It may be that as each successive passer-by causes a jungle pathway to become more clearly defined so each passing impulse makes the way easier for its successors, and makes it less hkely for them to stray into the surrounding substance. The special physiological meaning of the differentiation of the fibril would simply be the increase of its conductivity — possibly towards one specific type of impulse—but correlated with this are optical and staining peculiarities which, though unessential in them- selves, make the fibril recognizable to the eye as a definite structure? The nerve-trunk in Lepidosiren is seen to be at first naked and later on to acquire a sheath formed by concentration of mesenchyme round it. This sheath is at first richly laden with nutriment in the form of yolk granules but these are gradually used up as the nerve- trunk goes on with its development, the products of digestion of the yolk being doubtless passed on to the developing nerve-trunk. This as well as the marked increase in the number of nuclei in the sheath seem to indicate that the main réle of the sheath is to look after the nutritional needs of the nerve-trunk.?® We have dealt, so far, only with the motor nerve-trunks. In regard to the general method of development of sensory nerves, there ' Paton (1907) shows that impulses are actually transmitted across the protoplasmic bridge at a very early stage in the case of Elasmobranchs. * The hypothesis here outlined in connexion with the embryonic development of nerve fits in well also with certain of the phenomena observed in the regeneration of nerves which have been severed and joined together again [see Trans. Roy. Soe er p. 126, also Mott, Halliburton and Edmunds in Proc. Roy. Soe., B, vol. 78]. * The medullary sheath of nerve-fibres is non-cellular and appears to : by the secretory activity of the protoplasm of the axon. e nears II - NERVE DEVELOPMENT 113 exists the same divergence of opinion as in the case of the motor nerves, and in endeavouring to decide which view has upon its side the balance of probability, it is well to bear in mind similar conditions to those alluded to on p. 106. Bearing these in mind, it is of interest to notice that in Lepidosiren (Elliot Smith, 1908) the process of development of the olfactory nerve takes place along exactly similar lines to that of the motor trunks. And it is significant that, in the opinion of those well qualified to judge (Retzius, Golgi, Ramén y Cajal, van Gehuchten, Kolliker, Elliot Smith), this nerve has advanced less from the primitive condition than has any other nerve, and in its general arrangements has undergone extraordin- arily little complication during ontogeny. Already at a time when the olfactory organ has not ‘yet com- menced to recede from the wall of the hemisphere the olfactory nerve exists as a stout protoplas- mic bridge (Fig. 63, J) which gradually increases in length as the olfactory organ recedes from the hemisphere. This observation seems to indicate clearly that the mode of development of the sen- sory nerve-trunk is fundamentally the same as that of the motor: that it develops out of a pre- existing protoplasmic bridge be- tween centre and end-organ. Fic. 63.—An early stage of the olfactory nerve of Lepidosiren. (From Elliot Smith, 1908.) REMARKS UPON THE GENERAL c.H, lateral wall of hemisphere; olf, olfactory PROBLEM OF NERVE DEVELOPMENT organ; J, olfactory nerve. The nuclei seen in the region where the olfactory nerve enters the hemi- It will be admitted by most sphere belong to the olfactory bulb. Zoologists that we are justified in believing that the process of nerve-development is probably fundamentally the same throughout the Animal Kingdom. It will also be clear, even from the short and imperfect statement which has been given here, that the detailed study of the phenomena of nerve-development has led,in the minds of different observers, to widely divergent conclusions as to the exact nature of the process. The subject is one to the discussion of which we may devote with advantage some further space. It is in itself of great embryological and physiological interest. It presents many problems still unsolved. And it may be taken as a type of biological controversy with which it will be to the student’s advantage to become acquainted. In approaching the question from the present-day standpoint it appears impossible to get round the fact that in two of the most VOL. II I 1l4 EMBRYOLOGY OF THE LOWER VERTEBRATES — cH. archaic groups of Vertebrates (Elasmobranchii and Dipnoi) the motor nerve-trunk is already present as a protoplasmic bridge at a time when myotome and spinal cord have not yet commenced to recede from one another. It does not seem possible to explain the appear- ances recorded in these cases by any conceivable errors of observa- tion. But if such bridges exist in these relatively archaic groups, the balance of probability is entirely on the side of their represent- ing the primitive mode of development of nerve-trunks in general, and of a fundamentally similar mode of development occurring in other Vertebrates though possibly in a modified and less distinct form. , On the other hand appearances of the kind which led to the original formulation of the His view, and which are still adduced in its support, and which are easily observed in series of sections through almost any type of Vertebrate embryo — nerve-trunks passing out from the spinal cord and ending freely amongst the mesenchyme—are peculiarly apt to be misleading. Such a misleading appearance is produced sometimes by compara- tively simple causes—by breakage of the nerve-trunk or by the nerve-trunk passing away out of the plane of a section and being unrecognizable when cut transversely in a neighbouring section. In other cases the appearance of a freely ending nerve-trunk is due to the portion of nerve-trunk which has received its protoplasmic sheath being distinctly visible in a stained section, while the delicate peri- pheral portion which is still naked is practically unrecognizable. On account of such liability to misinterpretation a very large proportion of the observational evidence which supports the His view is open to suspicion. A physiological difficulty which has been raised against accepting the His view is that involved in the idea that the free end of the growing fibre tracks down and finds its appropriate end-organ. It is pointed out that it never makes a mistake—never becomes joined up to a wrong end cell. And yet, if it be the case that nerve-fibres do grow outwards with free ends in the way involved in the His theory, certain experimental results show that such fibres do possess a very decided power of making mistakes. This is brought out clearly by the beautiful experiments of Braus (1905). In the experiments in question Braus made use of the method, invented by Born and developed by Harrison, Spemann and others, in which portions of one amphibian embryo are grafted upon the body of another, when the grafted portion (“parasite ”—Braus) pro- ceeds to develop as part of the individual (“ autosite”—Braus) upon which it has been grafted. In the experiments which are most important in their bearing on the point now under discussion the early rudiment of the pectoral limb was grafted upon a host in the region of the head. In this position the rudiment went on developing into a perfectly normal limb containing a normal arrangement of the limb nerves. Now the is NERVE DEVELOPMENT 115 implanted limb in such a case (Fig. 64) is situated in a region inner- vated by the facial nerve and the study of sections showed that the nerves in the implanted limb were continuous centrally with branches of the facial nerve. If we attempt to interpret this experiment on the outgrowth view we find ourselves compelled to admit that the facial fibres concerned made the serious mistake of growing into a limb rudiment and then continued on their mistaken course until finally they established the muscular connexions normal for the nerves of such a limb. Braus repeated this type of experiment in a number of cases and there appears to be no question as to the accuracy of his observations. If accurate, however, they provide a formidable, if not unsurmountable, difficulty for the outgrowth view—a difti- culty which is by no means got rid of by the suggestion (Harrison, 1908) that after arriving in the limb the nerves are “merely guided in their growth by the structures present in the transplanted part.” A similar difficulty is seen in post- embryonic nerve-development in the fact well known to surgeons that functional continuity can be established between the cut central stump of one nerve (e.g. spinal accessory) and the severed peripheral portion ric. 64.—Young Toad (Bom- of another (e.g. facial). binator) on bey oo And so again in the development of reg sen aie anastomoses between peripheral nerves such _ Braus, 1905.) as the well-known “ dialyneury ” of Gastero- pod molluses, or the short circuiting of the left pulmonary nerve over the dorsal side of the oesophagus which has come about in the evolu- tion of the Crossopterygians and Lung-fishes. ; All such cases present great if not insuperable difficulties to the His view. Again much of the evidence which is brought to the support of the His view is seen when looked at ctitically to be less convincing than it appears to be at first sight. Thus for example with the experiments of Harrison already described, which are regarded by their author as settling the whole question. Their true value will become more apparent if we bring Harrison’s results into correlation with the results described above for Lepidosiren. As has already been shown, in this animal the motor nerve-trunk , 1s represented at an early stage by a bridge of soft fragile protoplasm. These bridges require a very favourable object and very careful technique for their detection, and it is clear that one could not expect to see them in comparatively coarse preparations made by excising a piece of living unfixed spinal cord. There is therefore no guarantee that such protoplasmic nerve-rudiments were not already present in the pieces of spinal cord investigated by Harrison. 116 EMBRYOLOGY OF THE LOWER VERTEBRATES CH. Let it be assumed that such an experiment is repeated. upon Lepidosiren with a small piece of spinal cord rudiment with the protoplasmic bridge attached to it (Fig. 65, A). The piece of spinal cord is well supplied with food material in the form of yolk and, if kept under suitable conditions, it would go on developing. So also might the protoplasmic bridge, for every one agrees that the metabolic control of the motor nerve is exercised by the central ganglion-cell nuclei within the spinal cord. If this happened and the process went on quite normally we should get in succession stages such as those shown in B and C of Fig. 65. Now these would be inter- preted by Harrison presumably as demonstrating the outgrowth view, whereas all that they really show is that, given suitable con- ditions, the motor nerve increases = L in length—a fact which of course is obvious. What is needed as a demonstration of the His view is not merely to show that a nerve-trunk increases in length but to show (1) that it normally has a free end and (2) that it erows within the body at a greater rate than the tissues in which it is embedded, so that there is brought about a differ- ential movement in which the Fic. 65. — Drawings taken from the same free end pushes its A ay through preparations as those illustrated in Fig. 60, the tissues surrounding it. This showing a piece of spinal cord with the has not been shown by Harri- developing motor nerve but ignoring the goy’g experiments nor could it myotome which is in the actual embryo continuous with the outer end of the nerve. possibly be shown by this type of experiment. In Lepidostren the study of sections shows as has already been pointed out that, although the motor nerve-trunk grows actively in length with the increase in bulk of the body, at no period from the earliest stage figured has it a free end; it is throughout connected with its end- organ.1 In a word, it appears to the present writer that what are commonly regarded as the most convincing pieces of evidence in favour of the His view are by no means convincing. Views resembling that of His in that they also involve an out- 1 The actively moving pseudopodium-like tags which Harrison observed at the end of his outgrowing nerve-trunk are believed by the present writer to be mesen- chymatous in their nature—-possibly shreds of sheath protoplasm. It is a general feature of embryonic mesenchyme that its protoplasm shows active amoeboid moye- ment. Il NERVE DEVELOPMENT 117 growth of the motor trunk from the spinal cord, but differing from it in the essential feature that the outgrowth is simply protoplasmic and not fibrillar, have been enunciated by some modern workers such as Dohrn and Held. Dohrn (1888) describes the motor nerve-rudi- ment as arising by a “plasmatic outflow from the neural tube” but Paton later on finds that at the stage referred to by Dohrn the nerve-rudiment is already continuous at its outer end with the protoplasm of the myotome. Held (1909) also regards the motor trunk as arising by outgrowth from the spinal cord at a time when the myotome is still com- paratively close to it. It has to be borne in mind in interpreting such sections as Held figures that there is more liability to error in demonstrating the absence of continuity than in demonstrating its presence, owing to the extremely fragile character of the nerve- trunks during early stages in development and their consequent liability to rupture during the ordinary processes of preparation which precede section-cutting. It is sometimes said that the difficulty attaching to the His view involved in the idea of the nerve-fibre tracking down its own particular end-organ disappears if the view is taken that the out- growth takes place, at a stage so early as that indicated by Dohrn and Held. But as a matter of fact this involves, as indicated, a distinct departure from the view enunciated by His according to which not merely undifferentiated protoplasm but definite fibrillated trunks grow out from the spinal cord. Further if, as Held believes, the individual fibrils grow out in the substance of the protoplasmic outgrowth each one has still to seek out the particular portion of the myotome which will eventually be converted into its own proper muscle-cell—a view which, looking to the comparatively undiffer- entiated condition of the myotome cells at these early stages, is even more difficult to comprehend physiologically than the outgrowth towards a specialized muscle. The embryological evidence upon which the His view rests is seen, when submitted to critical examination, to be unconvincing. The same is the case with the observational evidence upon which the Balfour view rests. The nuclei and cell-bodies which commonly give a multicellular appearance to the nerve-rudiment are quite reasonably interpretable as sheath-cells, 7.2. mesenchyme elements which have collected round and it may be migrated into the, at first noncellular, nerve-trunk. In Lepidosiren, with its coarse and heavily yolk-laden mesen- chyme, it is comparatively easy to distinguish such elements from the actual nerve-trunk embedded in them, but in most Vertebrates this criterion is not available and there is no certain means of distinguishing in ordinary microscopic preparations the protoplasm of the nerve-trunk from that of the sheath-cells. The primitive protoplasmic bridge described in 1902 for Lepidosiren as representing the motor trunk at a time. when 118 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. myotome and spinal cord have not yet commenced to move apart,’ confirmed later in the case of the motor trunks of Elasmobranchs by Paton, and in the case of the olfactory nerve of various Vertebrates by Elliot Smith and others, seems to rest upon a secure basis of observation. It is difficult, therefore, to avoid the expectation that the progress of future research will show such a primitive protoplasmic bridge between centre and end-organ to be the normal forerunner of nerves in general. But, if this be so, we are faced by the question as to the actual mode of origin of such bridges and here we pass into a region where direct observation is either impossible or unreliable. Those who accept Hensen’s views in their entirety would look -upon them as representing intercellular connexions persisting from the earliest segmentation stages. Reasons have already been given (p. 37) for disbelieving in the persistence of such bridges between the cells of the segmenting egg. The connexion appears certainly to arise at some later period—but exactly when seems to be a question incapable of answer by direct observation. When considering these general problems regarding the nervous system it should be borne in mind that the nervous system has for its main purpose the keeping of the various parts of the body linked together into an organic whole, in spite of their increasing ditterentia- tion and specialization. It has for its function the providing of exquisitely specialized pathways by which the living impulses can traverse the whole length of a relatively immense body at least as readily as they originally did the minute blob of ancestral protoplasm. Bearing in mind this primary consideration will cause one to reflect that the evidence must be overwhelming before one is justified in believing that this organ system, whose most striking functional feature is continuity, has come in the course of evolution to be characterized by the structural discontinuity involved in the neurone theory of adult structure, or in the outgrowth theory of ontogenetic development. Again it is important to bear in mind the high degree of probability attached to the view, originated long ago by O. and k. Hertwig (1878), that the nervous system of the higher metazoa, including Vertebrates, has been evolved out of a sub- epithelial nervous network of the kind still seen in some of the more lowly organized groups such as Coelenterata and Echinodermata. We may suppose that such a plexus was present in the far back ancestors of Vertebrates over the basal surface of both ectoderm and endoderm cells (as in modern Actinians, Havet, 1901) and that nerve- trunks became evolved as local condensations of such a network, just as we still see in the nerve-strands of a Medusa or a Starfish. In this connexion, it is of interest to note that according to the protostoma theory of the fundamental structure of the vertebrate body, which will be found stated later, in Chapter IX., the points II NERVE DEVELOPMENT 119 represented by the two ends of the motor trunk were originally in close proximity, and a condensed strand of the network joining the two points would naturally be left as a bridge when they became separated by the deepening of the cleft between mesoderm and endoderm (cf. Fig. 66). To sum up, in regard to the mode of development of the nerve- trunks, it seems reasonable in the present state of our knowledge (1) to reject definitely that portion of the Hensen view which looks on the protoplasmic bridges as having persisted from the com- mencement of segmentation, (2) to regard the His view of free-ending fibrillated outgrowth as non-proven and for various reasons improbable, (3) to believe that the nerve- trunk already exists as a proto- plasmic bridge between centre and end-organ at a period when these are still in immediate contact, even although this has up to the present been definitely shown by actual ob- servation only in a few peculiarly favourable instances, (4) to leave the exact period at which the protoplasmic bridges come into existence an absolutely open question as being beyond the limit Fic. 66.—TIllustrating the structure of a hypothetical primitive Vertebrate at of reliable observation, (5) as regards the sheath of Schwann, to accept the view that it is derived from mesenchyme. - It will be noticed that little has been said so far regarding the mode of development of the actual neuro- fibrillae. Their origin is indeed unverifiable by direct observation, atime when the protostoma was still open. In the lower figure an entero- coelic pocket, the rudiment of a mesoderm segment, is becoming de- marcated from the rest of the endo- derm hy the downward spreading of a split between the points @ and 0. In the earlier stage shown in the upper figure this split has not yet begun to develop, and the points a and 6 are seen in close proximity to one another on the outer surface of with any certainty, owing to their _ the endoderm. minute size. They appear to spread outwards from the centre, and Held interprets this appearance by a kind of His theory on a minute scale, holding that each fibril grows out with a free end through the protoplasmic bridge. On the other hand if it be the case as suggested on p. 112 that the fibrils simply represent the specialized paths of nerve impulses there would be nothing surprising in their becoming visible first in the neighbourhood of the ganglion- cell from which the impulses start and from which also is exercised control over the metabolism of the nerve-trunk. Were this the m.p, medullary plate ; p.s, protostoma. 120 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. case we should get appearances which would closely simulate growth of freely ending fibrils—centrifugal in the case of motor nerves and centripetal in the case of sensory nerves. This view as to the meaning of the fibrils bridges over a good many of the difficulties in the way of accepting the outgrowth view, either as regards the individual fibrils or the nerve-trunk as a whole. Thus the secondary establishment of anastomoses between peripheral nerves becomes less surprising if it be the case that undifferentiated protoplasm is liable to develop nerve-fibrils as a reaction to the passage of nerve impulses through it, for wherever there are nerves there must be a certain amount of leakage of the particular form of energy which constitutes the nerve impulse. : So also with the joining up of the central and peripheral ends of a severed nerve or of the central stump of one nerve with the peri- pheral portion of another. In such cases we should assume that indifferent protoplasm accumulating between the cut ends gradually becomes fibrillated in response to the passage of more or less imperfect impulses through it, the newly developed portions of fibril being necessarily, from their mode of formation, continu- ous with both central and peripheral fibrils, leading respectively to the “high-potential” and the “low-potential” end of the nerve- fibre. Again it is known that a mass of embryonic ganglionic tissue implanted into some abnormal portion of another individual may establish nervous connexions with the surrounding tissue. On the outgrowth hypothesis this demonstrates “error” on the part of the outgrowing fibres: on the functional view it simply involves the gradual differentiation of paths along which impulses spread out- wards from the high potential ganglon-cells into the low potential surrounding tissue. On the whole, the present writer believes that this view, that the formation of nerve-fibrils is a response to functional activity, is at the present time the most plausible working hypothesis and also the one which is most likely to lead to fruitful research. Before leaving the subject it may be well to emphasize the fact that the solution of this general problem of nerve-development is to be sought in the study of Vertebrates of large-celled coarse histological texture, com- bined with a low degree of specialization of general structure. No amount of observations upon small-celled highly specialized Verte- brates will ever lead to a really convincing solution with the methods now at our disposal. Finally we would once more emphasize the fact that the kernel of the problem seems to centre round the origin of the fibrillae. Do they or do they not develop in a pre-existing bridge of protoplasm ? Assuming that they do, the possibility of such bridges dating back to the period of segmentation seems to be definitely excluded. The question at what precise moment they do become established seems to be of minor importance, u NERVOUS SYSTEM 121 While the present writer is inclined to believe that the junction is already in existence while end-organ is still in close apposition to the central nervous system there is no difficulty in principle in the way of admitting that the bridge may in certain cases be formed somewhat later, as Dohrn describes, provided always that the gap to be bridged over is small and the bridge itself protoplasmic and not fibrillar. It is probably along such lines that we may look for a reconciliation between the supporters of His (the outgrowth view) and those who believe in the protoplasmic bridge view but it will involve dropping what are essential features in the outgrowth view as enunci- ated by His himself—(1) that the outgrowth arises at a time when the end-organ has already retreated to a considerable distance from the nerve-centre and (2) that the outgrowth is already fibrillated dur- ing the outgrowing process and before it is united to its end-organs. SPINAL GANGLIA AND DorsaL Roots.—As has already been in- dicated, the.central nervous system of the Vertebrate consists in its most primitive condition of a specialized area of the ectoderm of the dorsal surface. It is further very characteristic of the Vertebrate that those ganglion-cells which belong especially to the sensory fibres have become concentrated into segmentally arranged clumps towards the margin of the nervous plate and have eventually come to lie out- side the limits of the actual plate, or tube into which the plate be- comes converted. These little detached pieces of the central nervous system are the ganglia of the dorsal roots or the spinal ganglia.’ During actual ontogeny the ganglion rudiments in some cases (e.g. Birds, Fig. 67, A) become distinctly apparent while the spinal cord is still in the form of an open medullary plate. They appear in the form of a continuous proliferation from the inner surface of the ectoderm in the angle between the medullary plate and the external ectoderm. In such cases the two rudiments become carried in towards one another, as the edges of the medullary plate curve inwards to form the neural tube, and undergo fusion across the mesial plane. There is thus formed a median unpaired plate or tract of cells lying just over the roof of the neural tube and between it and the external ectoderm. This is known as the neural crest (Marshall). More usually the ganglionic rudiment makes its first appearance after the closure of the neural tube and in such cases the paired stage of the rudiment is slurred over, the neural crest being formed by proliferation of the roof of the neural tube. This is well seen in the case of Elasmobranchs (Fig. 67, B, C). However it originated, the plate-like neural crest splits into two 1 There is reason to believe that this is an instance of a widespread tendency in evolution for groups’of ganglion-cells to undergo gradual shifting towards the direction from which their most frequent afferent impulses come. In other words there is a tendency to shorten the afferent path by shifting the cell-body. This principle of neurobiotaxis has been developed by Ariéns-Kappers in his various papers (¢.g. 1918). It is particularly conspicuous in the changes which have come about in the position of the ganglionic centres of the various cranial nerves within the brain in the different groups of Vertebrates. 122 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. lateral halves and then grows outwards on each side opposite each myotome, each outgrowth representing a single spinal ganglion. Eventually these break apart but in some of the more primitive Vertebrates the intervening portions of neural crest persist for a time in the form of a distinct longitudinal commissure (Fig. 68) linking up the series of spinal ganglia to one another (Elasmobranchii —Balfour: Dipnoi). The mode of development of the fibres forming the dorsal root, whether by outgrowth from the ganglion-cells of the spinal ganglion or by differentiation of an already existing protoplasmic bridge, comes Fic. 67.—Illustrating the mode of origin of the spinal ganglia. A, fowl embryo with four mesoderm segments (after Neumayr, 1906); B and C, Torpedo 4 mm. embryo (after Dohrn, 1902); ect, ectoderm ; g, rudiment of ganglion ; s.c, spinal cord. under the general controversy as to nerve-development and need not be specially discussed. CRANIAL NEeRVES.—The development of the cranial nerves has been investigated by many workers and an immense amount of detailed observation has been accumulated. There is however great discrepancy in detail between the results obtained by different workers, and much of the observation seems to be perilously near the limit of probable error. Consequently the material seems hardly ripe for treatment in a text-book of a general kind and nothing of the sort will be attempted here beyond noting one or two points of particular importance.! 1A modern account of the development of cranial nerves will be found in Neuwmayr (1906). II CRANIAL NERVES 123 In the first place we find in the head region as in the trunk a tendency for the nerve-fibres to come off from the central nervous system in segmentally arranged clumps, and for the motor fibres to be situated more ventrally and the sensory more dorsally. In the head region however the dorsal root becomes reinforced by a large mass of motor fibres which have become shifted dorsalwards and incorporated with it. A neural crest develops resembling that of the trunk and in the Birds it can be seen similarly to have a paired origin, arising before the complete closure of the medullary tube. This neural crest of the brain region forms an anterior prolongation of that in the trunk: it is quite continuous with the latter, it develops outgrowths similarly, and the intervening portions here also persist for a time as a longi- tudinal commissure. A number of the most important cranial nerves wm i x Pg — Fic. 68.—Acanthias, stage 23, 9 mm. long, showing ganglia of cranial and spinal nerves. (Atter Scammon, 1911.) ant, intestine ; 2, lens; li, liver; pan, pancreas; sp.g, ganglia of spinal nerves; Th, thyroid ; V, ventricle ; v.c, visceral cleft ; y.st, yolk-stalk ; IV, V, etc., ganglia of cranial nerves. are simply prolongations of the outgrowths in question—V, VII, VIII, IX and X. A conspicuous feature in the development of the cranial nerves is that in portions of their length they receive components directly from localized thickenings (placodes) of the ectoderm (Kupffer, Beard) a possible reminiscence of the time when nerve-trunks became evolved out of a plexus in direct relation to the external ectoderm. I. The Olfactory nerve is unrivalled amongst all the sensory nerves of the Vertebrates as a subject for investigation on account of its large size, its short uncomplicated course and its retention of comparatively primitive conditions even in the adult. Research should. therefore be specially concentrated upon its mode of development. In the case of Lepidosiren, as already indicated, Elliot Smith has shown that the olfactory nerve is simply a drawn-out primary connexion between brain and olfactory organ, already present at a period before these organs have begun to move apart. In other vertebrates (Elasmobranchs—Holm, Teleosts, Amphibians 124 EMBRYOLOGY OF THE LOWER VERTEBRATES — cu. —Cameron and Milligan) there is evidence that the same mode of development holds. One of the important points to be settled is whether the nuclei which are seen scattered about in the young nerve-trunk and which give it a syncytial appearance are not really immigrant sheath nuclei. The conditions in Lepidosiren where it is easy to distinguish the heavily yolked sheath protoplasm make 1b appear probable that this will be found to be the case. . II. The optic nerve is not a true peripheral nerve comparable with the other cranial nerves but simply a narrow isthmus or stalk connecting the main brain with its outlying portion which forms the retina. Its development is mentioned in the description of the eye. III, IV and VI. The oculomotor, pathetic and abducent nerves appear to agree exactly in their main developmental features with ordinary motor nerves of the trunk (Neal, 1914). It is not proposed to say anything here regarding the topography of the cranial nerves but some points regarding it will be touched upon later on in connexion with the segmentation of the head. SYMPATHETIC.—The sympathetic ganglia, as was first shown by Balfour (1878) for Elasmobranchs, are derived directly from the spinal (or cranial) nerves. In its earliest recognizable stage the ganglion forms a swelling on the course of the nerve just ventral to and continuous with the spinal ganglion. With further development the ganglion bulges more and more pronouncedly towards the mesial plane at about the level of the dorsal aorta. The nerve-trunk in this region now splits longitudinally and the ganglion becomes shifted farther towards the mesial plane, lying immediately over the posterior cardinal vein and remaining connected by a slender bridge —the ramus communicans—with the spinal nerve from which it has become split off. In Sauropsida the sympathetic ganglia arise in similar fashion. In Amphibia and Sauropsida, where the sympathetic ganglia are in the adult connected by a longitudinal trunk, this latter is said to arise secondarily, the ganglia being at first quite separate. In view, however, of the difficulty of detecting such nerve-trunks in early stages of development it will be well not to dismiss altogether the possibility that the ganglia are after all in continuity from the beginning. From the basis of the sympathetic system so laid down extensions apparently sprout out into the various tissues which are eventually innervated by this system—but again there has to be admitted great possibility of error. The problem of the mode of development of these obscure portions of the nervous system will probably only be satisfactorily settled after we know with certainty the processes at work in the development of the main nerve-trunks and ganglia. THE ORGANS OF SPECIAL SENSE.—We may take it that in the early stages of the evolution of the nervous system, while this system was still a diffuse network, there were already present scattered sensory cells—cells specialized for the reception of im- nl OLFACTORY ORGAN 125 pressions from without. Local concentrations of such sensory cells and their further specialization for the better perception of some particular type of stimulus has led to the evolution of the various organs of special sense. The special sense organs of the vertebrate fall into two categories —(I.) the organ of vision, perhaps the oldest organ of special sense, which is developed within the limits of the central nervous system and (II.) the other organs of special sense which have probably arisen more recently from the sense cells of the skin outside the limits of the central nervous system. As the organs belonging to the second category have evolved less far from the primitive condition they will be considered first. They. appear to have become specialized functionally in two different directions, those in the neighbourhood of the mouth for the apprecia- tion of differences in chemical composition—the organs of taste and smell—and those on other parts of the body surface for the apprecia- tion of vibrations of the surrounding medium—the lateral line organs and the organ of hearing. OLFACTORY ORGAN.—The olfactory organ arises in the form of a localized thickening of the ectoderm on each side towards the anterior end of the head. Later this thickened ectoderm becomes depressed below the general surface so that it assumes a saucer- and later a cup-shape; its external opening eventually becomes comparatively narrow. In many of the Elasmobranch fishes the olfactory organ retains throughout life the condition of a simple inpushing of the skin opening to the exterior on the ventral side of the snout. In many Vertebrates on the other hand characteristic changes come about in the relations of the external opening. These will best be understood by considering first what happens in the lung-fish Protopterus as shown by Fig. 69. In A and B the olfactory organ is visible as a little rounded dimple on each side. In C the dimple has become a deep groove running obliquely from before backwards and outwards. It is further seen that this groove is becoming involved in the sinking in of the skin to form the buccal cavity. In D and E the groove has become a deep slit narrow in its middle part and dilated at its two ends. Finally, in the stage represented in F, the margins of the narrow part of the slit. have undergone complete fusion so that the continuous slit of the preceding stage is now represented merely by its terminal portions which form two widely separated rounded openings—the anterior and posterior nares (o/f' and olf”). Turning to the other Vertebrates we find various divergences from this simplest mode of origin of the external and internal nares as seen in Protopterus and Ceratodus. In the Actinopterygian fishes the phenomena are quite similar to those described, only here differential growth leads to the gradual shifting of the olfactory organ and its openings from the ventral side of the snout up to its dorsal side. The result is a topographical reversal of the positions of 126 EMBRYOLOGY OF THE LOWER VERTEBRATES — cua. anterior and posterior naris: the morphologically posterior naris coming to lie in front of that which was originally anterior. In the Amphibian and Amniote the upper lip, which completes the buccal cavity in front, is situated between the anterior and posterior narial openings, so that the latter opens into the buccal cavity E D Fic. 69.—Ventral views of head region of larva of Protopterus at stages 31 (A), 32 (B), 34 (C), 34 (D), 35 (E), and 36 - (F), to illustrate the development of the olfactory organ. ¢.0, cement organ ; e.g, external gill; olf, olfactory organ; olf1, anterior (‘external ”) naris; olf?2, posterior (“internal”) naris. In C the curved line running across the ventral side of the head is the posterior margin of the mouth: the darkly shaded grooves passing inwards and forwards from its outer ends are the olfactory rudiments. (internal naris), while the former remains outside (external naris). In the developing Amniote embryo (ef. Fowl, Chap. X.) the general arrangements, while essentially the same as those of Protopterus, are somewhat obscured by the modelling of the face region. The ridge which forms the upper lip, or anterior boundary of the buccal cavity, is cut across by the olfactory slit, here a wide and deep cleft, into a 1 OLFACTORY ORGAN 127 median portion (median nasal process) and a lateral portion (maxillary process). The ridge bounding the olfactory involution on its outer side remains for a time separated by a distinct groove from the maxillary process but as the latter grows forwards it obliterates this groove as wellzas the superficial portion of the cleft which separates it from the median nasal process, the deep portion of the cleft remaining as a definitive canal leading from olfactory organ to buccal cavity. In Amphibians, as in Lepidosiren among the lung-fishes, the posterior naris is frequently formed as a secondary perforation which Fic. 70.—Horizontal sections through the olfactory organs of Polypterus of stages 25 (A), 26 (B), and 27 (C). c.o, cement organ ; olf, olfactory rudimeut; opt, optic stalk ; Thal, cavity of thalamencephalon. breaks through from the posterior portion of the olfactory organ into the part of the buccal cavity lined by “endoderm.” This is a secondary modification of a type which will be discussed in ‘the next chapter in the description of the development of the buccal cavity in these forms. The first rudiment of the olfactory organ has been described as a thickening of the ectoderm. As in the case of other nervous or glandular developments of the ectoderm the superficial layer (Fig. 70) takes no part in its formation. Commonly it degenerates and disappears over the actual olfactory epithelium. Again, as frequently happens in the development of primitively hollow organs, the rudi- ment may be for a time solid, forming a simple solid downgrowth in 128 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. which a cavity makes its appearance secondarily, the actual involu- tion of the surface being delayed or reduced or absent (Fig. 71). Sometimes, as is well shown in the case of Polypterus (Fig. 70, A), the olfactory thickenings are at first continuous across the mesial plane and this fact, taken in consideration with the fact that in the Lampreys the olfactory organ of the adult is unpaired, obviously suggests the possibility that the olfactory organ of Vertebrates in general was originally unpaired (Kupffer). Though this must be admitted as a possibility the evidence does not appear to be sufficient to give the idea probability in view of the fundamentally paired character of the portions of the brain associated with the olfactory organ, even in the case of the cyclo- stomes where the organ as a whole has an unpaired appearance. After the olfactory involution has become definitely established it under- goes various complications of form, differing in detail in the various groups but consisting for the most part of bulgings outwards on the part of the lining epithelium so as to bring about an increase in its area. In the Elasmo- branch these outgrowths take the form of parallel grooves which gradually become converted into deep slits separated by thin partitions — the Fic. 71.—Longitudinal vertical sec- Schneiderian folds. In Crossopterygians tion through Polypterus (stages Jnstead of numerous folds with free 28-29), showing the olfactory rudi- edges complete septa are formed which ment as thickening of the deep radiate out from an axis formed by layer of the ectoderm in which a ee cavity has developed secondarily. the olfactory nerve and divide the cavity as seen in transverse section into distinct chambers, the lining of which in turn forms deep folds. In the higher forms the outgrowths of the olfactory lining are fewer in number and the projections left between them form the turbinals which have characteristic arrangements in the different groups. Amongst the Reptiles a conspicuous development of the olfactory apparatus is the Organ of Jacobson. This arises as a pocket-like outgrowth of the lining epithelium, on its mesial side and near its ventral edge, which becomes gradually constricted off from the olfactory organ and opens into the buccal cavity in the region of the posterior nares. In Chelonians, Crocodiles and Birds this organ has disappeared except for a possible vestige in the form of a transient bulging of the olfactory lining. A diverticulum which may correspond to Jacobson’s organ makes its appearance in Lung-fishes and Urodeles but in this case it becomes gradually displaced outwards until it lies external to the olfactory cavity. 11 OTOCYST 129 A curious, possibly adaptive, arrangement has been noticed in late developmental stages of certain Sauropsida, where for a time the external nares are plugged by a proliferation of ectoderm (Aplerya— T. J. Parker, Sphenodon—Dendy). Such temporary obliteration of a channel at a period of development where it is unnecessary or harmful is a phenomenon which occurs fairly frequently: examples of it will be met with later in connexion with the alimentary canal and the excretory organs. Orocyst.—The Vertebrate possesses a pair of otocysts situated one on each side of the hind-brain. Each arises in the typical fashion familiar in the invertebrates, by a sensory portion of ecto- derm becoming depressed below the surface of the skin and eventu- ally isolated as a closed vesicle. As in the invertebrates certain of the lining cells of the otocyst secrete otolithic masses of Calcium carbonate. / The otocyst of the Vertebrate however shows two developments which do not occur amongst the invertebrates. Firstly, in connexion with the primitive function of the organ, that of balancing, the wall of the growing otocyst becomes ioulded into the three semi- circular canals which are arranged in planes at right angles to one another. These canals have for their function the analysis of any rotatory movement into its components in these three planes. And secondly a special region of the otocyst wall becomes specialized in connexion with a new sense—that of hearing—and grows out into a curved horn-like pocket, the lagena, which may become greatly enlarged and spirally coiled, in Vertebrates in which the sense of hearing is very acute, forming the organ known as the cochlea. The development of the otocyst may be described as it occurs in the Fowl (Rothig and Brugsch,-1902). The otocyst begins to make its appearance during the second day of incubation as a thickened area of ectoderm on each side of the hind-brain. This thickened area becomes depressed below the general surface, forming a saucer- shaped depression which gradually deepens till it forms a deep pit. The lips of this pit, especially dorsally, grow inwards so as to constrict the opening which is finally completely obliterated, the original open depression being thus converted into a closed, somewhat pear-shaped, sac the otocyst. The wall of the otocyst remains for a time con- tinuous through a solid bridge with the outer ectoderm (Fig. 72, A) but as a rule this bridge persists merely for a very short time and only a small cellular tag attached to the otocyst remains to mark its position. As development goes on the otocyst increases in size by growth of its wall and this growth is especially marked ventrally and later- ally with the result that the point which was originally connected to the ectoderm. becomes displaced so as to be situated on the mesial side of the otocyst. This portion of the otocyst wall now comes to project upwards as a distinct pocket-like outgrowth the recess (Fig. 72, B, 7). External to this a wider bulging of the wall fore- VOL. II K 130 EMBRYOLOGY OF THE LOWER VERTEBRATES — CH. shadows the anterior vertical semicircular canal (a.v) and a little later a similar bulging more ventral in position—the horizontal canal (h.c). The lagena also is foreshadowed by a slight downward bulging of the floor of the otocyst. With further development the posterior vertical canal rudiment appears also as a bulg- ing of the otocyst wall continuous with that which will form the anterior canal (Fig. 72, C-F, p.v). The three canal rudiments come to project more and more prominently, the ‘ recess assumes the tubular shape of the endolymphatic duct and the lower portion of the otocyst (saccule) with its projecting lagenat pocket and en- dolymphatic duct be- comes more sharply marked off from the rest of the otocyst (utricle). The pouch-like rudi- ments of the semicir- cular canals, as they come to project more freely from the utricle, assume a flattened form and finally the central Fic, 72.—Ilustrating the development of the otocyst in the portion of the wall on Fowl. (A-G after Réthig and Brugsch, 1902; H after ‘ in- Retzins) mg , ; each side bulges in : wards and fuses with A and B, early stages; C, towards end of seventh day; D, to- + 4 wards end of eighth day; H, five days; F, towards end of ninth that on the other. day; G, towards end of twelfth day; H, adult. A is a view from In this way the in front, B from. behind, while C-H represent the left otocyst as central portion of the seen from the left side. a.v, anterior vertical canal; .c, horizontal ey canal; lag, lagena; p.v, posterior vertical canal; 7, recess; s.e, cavity of each pouch endolymphatic duct. becomes obliterated while the persisting peripheral part takes the form of a curved tube—the definitive canal. At first the spac subtended by the canal is traversed by a continu- ous septum formed out of the fused walls but this soon disappears leaving the canal as a freely projecting arch which opens into the utricle at each end. The ampulla appears at an early stage as a dilatation of the canal rudiment at one end. As will already have been gathered, the three canal rudiments do rat OTOCYST 131 not appear synchronously—the anterior vertical appearing first, then the horizontal and finally the posterior vertical. The same order is followed in subsequent development the anterior vertical canal keep- ing ahead of the other two—probably a foreshadowing of the greater importance of this canal, parallel to the sagittal plane, in the function of flight. It should be noticed that the posterior vertical canal assumes its Fic, 73.—Development of the otocyst in Lung-fishes as seen in transverse sections. (From drawings by M. C. Cairney.) A, Lepidosiren, stage 21; B, Protopterus, stage 23; C, Lepidosiren, stage 28. N, notochord; ot, otocyst; rh, rhombencephalon. In Fig. C the rudiment of the endolymphatic duct is visible as an out- growth from the otocyst wall dorsally and mesially. position at right angles to the planes of the other two secondarily. At first its rudiment is continuous with and almost in the same plane as the anterior canal but as it assumes the tubular form it swings outwards and forwards slipping, as it does so, over the horizontal canal in the way indicated in Fig. 72, F,G, H. The otocyst of Vertebrates in general develops along similar lines to those described for the Fowl while presenting modifications in detail. The sensory epithelium is a development of the deep layer of the ectoderm and in cases where the ectoderm is distinctly two 132 EMBRYOLOGY OF THE LOWER VERTEBRATES CH. layered during the early stages of otocyst development the super- ficial layer is seen to pass over the rudiment unaltered (Fig. 73, A). In Lung-fishes the cavity of the otocyst appears secondarily in the midst of an apparently solid downgrowth of the deep layer of the ectoderm (Fig. 73, B) but the examination of earlier stages (Fig. 73, A) shows that here also there is an actual involution of the deep layer although there is at first no patent cavity. In the Elasmobranchs the otocyst retains throughout life its con- nexion with the exterior, the connexion becoming drawn out into a slender tube. In Birds the recess, and therefore the endolymphatic duct, represents the remains of this original connexion, but curiously enough in certain other Vertebrates ¢.g. Lung-fishes (Fig. 73, C) the last connexion of otocyst with external ectoderm lies lateral of and somewhat anterior to the endolymphatic duct which latter here develops as an independent outgrowth of the otocyst wall. This is to be looked on as a secondary modification of the more primitive . arrangement seen in Elasmobranchs. The structure named endo- lymphatic duct in Teleosts also arises as a secondary outgrowth of the otocyst wall. The endolymphatic duct or recess commonly persists in the adult as a conspicuous blindly ending diverticulum of the otocyst wall. In Lung-fishes and Amphibians its wall proliferates actively giving rise to projections which in the Lung-fishes and some Urodeles, e.g. the Axolotl, meet to form an irregular sac over the roof of the fourth ventricle. In the Anura the irregular thin-walled sac formed in this way spreads forwards and also laterally until it becomes continuous ventrally so as completely to surround the hind-brain. An unpaired prolongation of this sac extends tailwards immediately dorsal to the spinal cord within the vertebral canal. Paired outgrowths of this extend outwards along with each spinal nerve and expand at their ends round the spinal ganglia to form the calcareous bodies so con- spicuous in the adult frog. The whole system of outgrowths is con- spicuous in the adult from the white otolithic particles in its interior. The vertebral portions become eventually broken up into a network of irregular tubes which is interpenetrated by a network of capillaries (Coggi, 1889). Somewhat similar outgrowths of the endolymphatic duct make their appearance in Sauropsida although in this case they do not undergo the wide extension that they do in the Anura. In the Geckos however they do become extended so as to form a large superficially placed irregularly lobed sac which covers over a great part of the neck region close under the skin (Wiedersheim). Lateral LINE OrGans.—These sense-buds (neuromasts), which are found arranged in rows on the head and body of fishes and aquatic amphibians, take their origin in linear thickenings of the deep layer of the ectoderm which spread along the surface of the head and body and eventually become segmented up into separate pieces. In correlation with the function of these organs, which II NERVOUS SYSTEM, 133° apparently is to detect slow vibrations in the water and is therefore closely allied to hearing, it is of interest to notice that the ectodermal rudiment from which they arise appears to be in some cases continuous at first, with that which gives rise to the otocyst. The sense organs are in correlation with their origin at first placed superficially but as development goes on they in most cases become depressed beneath the surface either in isolated pits or in continuous grooves. The latter may in turn remain open or may become covered in to form tubes except where at intervals openings remain leading to the exterior. This is the condition which is reached in the adults of the majority of fishes. The lateral line sense organs being correlated with an aquatic habit commonly degenerate on the assumption of a terrestrial existence. Various Anura however which remain purely aquatic after metamorphosis retain their full equipment of lateral line organs. ORGAN OF Dinkus.—In Lepidosiren and Protopterus a peculiar organ of special sense lies deeply embedded in the tissue on each side of the head in close contact with the wall of the auditory capsule. This organ, discovered in Protopterus by Pinkus, has been shown by Agar (1906) to be developed from the ectodermal ingrowth which forms the outer end of the spiracular rudiment. Eyr.—As the eye develops in the same general manner, differing only in detail, in the different subdivisions of the Vertebrata it will be convenient to describe first its development in the Fowl—the Vertebrate of which it is easiest to obtain material for practical study. The first obvious rudiment of the eye consists of a projection of the side wall of the thalamencephalon which juts out at right angles to the axis of the body and gives a characteristic hammer- shape to the fore-brain region (see Fig. 231, Chap. X.). A transverse section across the head near its front end in a chick about the middle of the second day of incubation shows (Fig. 234, D) the thalamencephalon extending out on each side as the optic outgrowth. As development goes on and mesenchyme accumulates between the brain-wall and the ectoderm the proximal part of each optic outgrowth becomes constricted, from above downwards, to form a relatively narrow optic stalk (Fig. 74, A, B, D, 0s). The optic outgrowth is closely apposed against the inner surface of the external ectoderm and a slight thickening of the latter soon becomes apparent just where it is in contact with the surface of the optic outgrowth in (Fig. 74, B, J). This thickening is the first rudiment of the lens. The lens-rudiment gradually becomes sunk below the general surface to form a saucer- and later a cup-shaped depression. As the rudiment becomes involuted in this way, the outer wall of the optic outgrowth also becomes 134 EMBRYOLOGY OF THE LOWER VERTEBRATES = CH. invaginated to form a cup-like structure—the optic cup (Fig. 7A, thal. hintuul A od [~~ .. hindi lnuhunl pe ce r pl Z., thal. luaaituul {ualunl e i BD Fic. 74.—Development of the eye as seen in transverse sections of Fowl embryos. A, latter half of second day of incubation ; B, end of second day; C, 24 days; D, 3 days; E, latter half of fifth day. ect, external ectoderm ; 1, lens; o.r, rudiment of eye; o.s, optic stalk ; p.l, pigment layer of retina; 7, retina; tha/, wall of thalamencephalon. B and CG). The cup-like lens-rudiment becomes gradually con- Il EYE 135 stricted: off and finally completely separated from the outer ectoderm (Fig. 74, C and D). In the meantime a marked difference becomes apparent between the two layers forming the wall of the optic cup. The layer next the cavity of the cup becomes greatly thickened its cells becoming tall and columnar: it forms the rudiment of the visual layer or retina in the strict sense. The outer layer of the cup-wall on the other hand degenerates, it becomes thinner and later it de- posits melanin pigment in its cells. 1t forms the pigment-layer of the retina. The invagination of one wall of the optic outgrowth within the other is not confined to that portion of the outgrowth in proximity to the external ectoderm as might be supposed from the description so far. The invagination involves also the ventral wall of the rudiment towards its outer end and for some distance along the optic stalk. The result is that the wall of the optic cup is interrupted by a gap ventrally the choroid fissure'—and that the optic stalk for some distance from the optic cup has a deep groove along its ventral side. The cavity of the optic cup, as is the case with cavities generally in the embryonic body, becomes filled with clear fluid secreted into it by the surrounding cells. This fluid becomes jelly- like later on and forms the basis of the vitreous body. As development goes on the eye increases greatly in size and assumes a spherical shape, the lens blocking up its opening towards the skin and the choroid fissure becoming obliterated by its lips coming together. The site of the fissure remains apparent for some time owing to the formation of pigment in the pigment-layer being delayed in its immediate neighbourhood. As the eye increases in size the Retina for a time grows more actively than the rest so as to be thrown into wrinkles (Fig. 74, E). The lens which was a hollow vesicle becomes solid its cavity being filled up by a great thickening of its deep wall, the cells of which grow out into a tall columnar form (Fig. 74, D and E). The essential parts of the eye as an optical instrument have now been laid down—the lens for the production of an image, the retinal wall of the optic cup for the reception of that image and the conversion of its light waves into nerve impulses, and the optic stalk which will become the optic nerve for the transmission of these impulses to the brain. To these essential parts there are added various accessory structures developed from mesenchymatous cells which accumulate round the parts of the eye already formed. In particular there is formed a protective capsule of tough con- nective tissue—the sclerotic with its transparent portion the cornea, covered externally by the ectoderm forming the corneal 1 The term “choroid” fissure is in reality misleading, having been adopted when the fissure was interpreted as a cleft in the choroid, in the days before the existence of the pigment-layer of the retina was recognized. 136 EMBRYOLOGY OF THE LOWER VERTEBRATES — CH. epithelium—and a highly vascular and deeply pigmented choroid between the sclerotic and the pigment-layer of the retina. As development proceeds watery fluid collects in the space between the lens and the cornea. Round its periphery this space is bounded internally by the inturned edge of the optic cup which overlaps the lens. For a considerable distance out from the lens, as far as the ora serrata, this portion of the wall of the optic cup does not undergo differentiation into functional retinal tissue. Its marginal portion covered on its outer surface by a prolongation of choroid becomes the iris. Here the inner “retinal” layer develops pigment and becomes in other respects like the outer layer. The outer layer itself gives rise to the radial and circular muscle fibres of the iris—a fact of morphological interest to be grouped along with the development of muscle fibres by the A Fig. 75.—Transverse sections through the brain of Lepidosiren, showing the solid optic rudiments (0.7). A, stage 21; B, stage 23. ectoderm of flask-glands referred to on a previous page. Outside the periphery of the iris the retinal layer remains comparatively undifferentiated as an unpigmented layer of columnar epithelium. It lies internal to a special development of choroid which forms the ciliary body and in which are developed the muscles of accommodation. The mode of origin of the eye as a whole and of its component parts as seen in the Fowl embryo having been sketched in outline, it will now be convenient to take the various parts in turn and note further features of interest without restricting consideration merely to the Bird. In the first place it is to be noted that the rudiment of the eye is at first solid in those Vertebrates in which the central nervous system is solid at the time of its appearance, the cavity of the optic rudiment developing secondarily (Fig. 75, A, B, or). As this secondary cavity increases in size both in the optic rudiment and in the brain a condition is gradually reached like that already described and the further development is on normal lines. I RETINA 1387 Retina.—The fully-developed retina—-which it will be re- membered is morphologically a specialized portion of the brain-wall —is an organ of extreme complexity. Its structure even in the adult is by no means completely worked out, and our knowledge of the details of its histogenesis is most imperfect. The most conspicuous feature is the great increase in thickness, the retinal cells becoming slender and columnar in form. Later on the nuclei are seen to become arranged in layers, this being an expression of the fact that the cells are also becoming specialized into layers of different structure and function. The details of development of these are almost completely unknown and there is here an interesting field for investigation. The layer of visual or percipient cells lies on the proximal side of the retina, and their rods and cones—the special parts of the visual cells which are believed to have the function of converting light Fic. 76.— Illustrating the development of the rods in Lepidosiren. The upper side of the figures represent the side turned away from the lens. A, B, C, D from stage 35; E, fully developed visual cell at stage 38, fixed during exposure to light ; E*, similar element killed in the dark. a.v, annular vacuole; fg, fatty globules stained black by osmic acid; m.1, external limiting membrane; 7, nucleus of visual cell; 7, rod. waves into nerve impulses—are at the ends of the cells which point away from the lens. To reach these rods and cones the light rays have therefore to traverse the whole thickness of the retina. This remarkable arrangement of the retina, precisely the opposite of what we should expect, is one of the characteristic features of Vertebrates. Its morphological significance is however at once made clear by a consideration of the main facts of development of the eye as already outlined. These, in fact, show that what becomes the proximal surface of the retina, ze. the surface which faces away from the lens, was originally part of the inner surface of the brain rudiment and therefore of the outer surface of the ectoderm before it became involuted to form the brain. The visual cells develop therefore on what was originally a part of the outer surface of the body and their rods point in a direction which was originally outwards. The mode in which the rods develop is illustrated by Fig. 76 which is taken from Lepidosiren (Graham 138 EMBRYOLOGY OF THE LOWER VERTEBRATES — CH. Kerr, 1902). Similar observations have been subsequently made in the case of Amphibians. ? ; The first obvious step in the specialization of the visual cell is the appearance of a fatty globule in its protoplasm. The end of the cell turned away from the lens now grows out into a projection and pushes back the tine cuticular limiting membrane (external limiting membrane) which has developed over this surface of the retina, into a little pocket. The oil globule which gradually increases in size passes into this pocket (Fig. 76, A, fy) and lies in it ensheathed in protoplasm. The protoplasm now becomes heaped up into a little conical protuberance (Fig. 76, B, 7) which is the rudiment of the rod. At first the limiting membrane is distinct over the surface of the rod but gradually, as the latter assumes a cylindrical shape, its protoplasm takes on a clear structureless appearance throughout: it apparently becomes in fact converted into cuticular material. This cylinder of cuticular material increases in size, assumes a character- istic appearance with alternating discs of dimmer and more traus- parent material as seen in the fixed specimen and the rod is complete (Fig. 76, E and E*). The rods complete their development sooner or later according as they are nearer or farther away from the optical axis of the eye and their time of development shows great variation in different individuals. The cones in those Vertebrates in which cones are present are merely specialized rods. Lrens.—The lens shows in its early stages, in various groups of Vertebrates, departures from the normal condition as described for the Bird, of exactly similar kinds to those seen in the development of the otocyst. In particular, the lens tends to develop out of a solid downgrowth of the deep layer of the ectoderm. This is well seen in Elasmobranchs (Fig. 77, A-E) where a rounded solid lens-rudi- ment is formed by proliferation of the ectoderm, this rudiment becoming isolated and developing a cavity secondarily. It is of interest to notice that even here a slight dipping down of the external surface into the lens rudiment is apparent for a time (Fig. 77, B). In Amphibians, Lung-fishes and Teleostomatous fishes the lens arises in a manner intermediate between what occurs in Elasmo- branchs and what occurs in Sauropsida. In the forms mentioned the lens arises as a downgrowth of the deep layer of the ectoderm (Fig. 77, F-I) and in some cases this downgrowth is simply an invagination of this layer, the only difference from the Sauropsidan condition being that here the opening of the invagination is closed by the superficial layer being continued across it (Fig. 77, J, K). As regards the later stazes in the development of the lens all that need be said is that it undergoes an enormous increase in size—by absorbing nourishment from its surroundings, for it has no blood- vessels—the cells of the deep wall becoming greatly elongated and taking on a clear glassy appearance, while the superticial wall remains as a layer of cubical epithelial cells over the outer surface of the lens. u EYE 139 SCLEROTIC, CORNEA, CHOROID.—These portions of the eyeball are gradually differentiated out of mesenchyme which becomes concentrated round the primary parts of the eye. In the case of the cornea the first stage in the developmental process consists in the accumulation between lens and ectoderm of a clear jelly-like secretion Fic. 77.—Variations in the early stages of the development of the Lens. A-E, Pristiurus; F-1, Siredon (after Rabl, 1898); J, K Phyllomedusa. (after Budgett, 1899). continuous, and identical in character, with that which fills the optic cup. As development goes on (Knape, 1909), a thin layer of this jelly-like material, about midway between the lens and the ectoderm, becomes condensed to form the rudiment of Descemet’s membrane. Amoebvid cells from the mesenchyme round the optic cup creep 140 EMBRYOLOGY OF THE LOWER VERTEBRATES — cu. along the deep surface of Descemet’s membrane and there settle down to form a single layer of flattened cells. On the deep side of this corneal endothelium a split gradually develops in the jelly-lke matrix: this contains a watery secretion (aqueous humor) and becomes the anterior chamber of the eye. The portion of matrix lying superficial to Descemet’s membrane becomes colonized by cells from all round its margin. It forms the main portion of the cornea, © while a thin layer lying next the ectoderm remains uncolonized and gives rise to Bowman’s membrane. Su Vitreous Bopy.—The cavity of the optic cup is from the beginning filled with clear fluid which keeps it distended and there is no ap- parent reason to assume that this arises otherwise than by the same method as holds with the eyes of many invertebrates @.e. as a secre- tion of the surrounding retinal cells. The fluid gradually acquires the jelly-like consistency characteristic of the fully-formed vitreous body. Amoebocytes wander at a com- paratively early period into the vitreous rudiment —in the Fowl embryo about the third day—and at a later period a continuous mass of mesenchyme tissue projects into it through the choroid fissure. This mass of mesenchyme develops a Fic, 78.—Semidiagrammatic figure of the network of blood-vessels continuous bisected eye of Vertebrate embryo with those of the surrounding tissue. (Rana 8 mm.) to show the course of says Pexcale the optic nerve-fibres (after Assheton, In the more primitive Vertebrates 1892). this mesenchymatous mass reaches c.f, Wall of choroid fissure; g, ganglion-cell; TO great development but in the 4, indifferent, supporting, cell; 1, lens ; nf, Teleostel and the Sauropsida, the ee haat tae Pigment layer of retinas % most highly specialized groups amongst the non-mammalian Verte- brates, it does so and persists throughout life, as the falciform process with its muscle-fibres for the purpose of accommodation in the one case (Teleostei), and the highly vascular, and probably mainly nutritive, pecten in the other (Sauropsida). Optic NeRvE.—As already indicated the optic nerve is not strictly speaking a peripheral nerve at all. It is a slender drawn-out portion of the brain analogous with the olfactory tract of a teleostean fish, con- necting the main portion of the brain with the small highly special- ized portion which has become converted into the optic cup. Its function being a conducting one the main mass of this stalk-like portion of the brain is composed of white substance or nerve-fibres. These fibres instead of passing outwards over the rim of the cup lI EYE 141 all round, as they possibly did originally, have become crowded together during the course of evolution into a single large bundle on _ the ventral side of the cup. In accordance with the general principle of economy of tissue this bundle of nerve-fibres has become sunk into a deep notch in the wall of the cup—the choroid fissure—so that it passes directly to the optic stalk. While this is probably a correct statement as regards phylogenetic evolution, matters are somewhat simplified in the development of the individual, inasmuch as the choroid fissure is brought about not by the notching of the already formed cup rim but by the rim ceasing to develop at the site of the choroid fissure while it grows actively everywhere else. As regards the development of the actual nerve-fibres, all that need be said is that they first make their appearance in the wall of the optic stalk ventrally and that they increase rapidly in number, passing between the epithelial cells of the stalk, loosening them out,and causing them in great part, if not entirely, to degener- ate. The individual fibres certainly for the most part become differentiated in a Fis. 79.—Transverse section through the still open neural : . . plate of Rana palustris near its anterior end, showing the centripetal direction position of the optic rudiments (Z) already marked out by 2.¢. from the retina to- the formation of pigment (after Eycleshymer, 1895). wards the brain, but. whether this means that they are actually sprouting out from ganglion- cells of the retina as is generally believed, or on the other hand that their fibrils are simply becoming differentiated centripetally within a continuous pre-existing protoplasmic connexion, remains to be demonstrated. EMBRYOLOGY AND THE EVOLUTIONARY ORIGIN OF THE Eyr.—The peculiar reversed position of the Vertebrate retina may perhaps be taken as an indication that that organ had already come into exist- ence, though no doubt in a very simple form, at a stage in Vertebrate phylogeny when the central nervous system had not yet sunk down below the surface. It is therefore of interest that in certain Verte- brates the rudiment of the retina does actually become apparent during embryonic development at a period when the medullary plate is not yet closed in. Thus Eycleshymer (1895) has described in Rana palustris and in Amblystoma how a patch of pigment appears for a time on the surface of the medullary plate (see Fig. 79) in the position which will later on form the optic outyrowth. Although we are perhaps justified in believing that the eye of existing Vertebrates was already present as a patch of epithelium sensitive to light in the far back evolutionary period when the fore- 142 EMBRYOLOGY OF THE LOWER VERTEBRATES | cu. runner of the central nervous system was still a portion of the outer surface of the body, we do not, in the preseut writer’s opinion, appear to be justified in connecting up the eye of typical Verte- brates with the “eyes” of Amphiowus or of Tunicates. It seems more probable that the eyes of these highly specialized creatures are organs which have developed independently within their own groups. Pirurrary Bopy.—To be included amongst the derivatives of the ectoderm is that enigmatical organ the pituitary body (“ Hypophysis cerebri”). This—the “anterior lobe” of the pituitary body in mammalian anatomy—arises normally as an ingrowing pocket of ectoderm on the ventral side of the head, situated as a rule close to the hinder limit of the stomodaeum but in the case of the Cyclo- stomes just outside its anterior boundary. This pituitary involution extends inwards beneath the infundibulum. In the Cyclostomata it retains its original form of a tube communicating with the exterior, but in the gnathostomatous Vertebrates its outer end becomes gradu- ally constricted into a narrow duct which in the great majority of cases becomes finally obliterated, so that the organ now forms a closed sac lying immediately underneath the infundibulum. The wall of the sac sprouts out into numerous tubular projections between which develops highly vascular mesenchyme, providing the rich blood supply necessary to the definitive function of the organ asa ductless vland. As regards variations in the development of the pituitary involu- tion: it may arise as a solid ingrowth of ectoderm (Lung-fishes, Am- phibians); it may be two-lobed (Teleosts) or three-lobed (Lacertilia) in its early stages ; its external opening may become secondarily dis- placed up on to the dorsal side of the head (Lampreys); its inner end may come to open secondarily into the pharynx (Myxinoids). As already indicated the wall cf the infundibulum in the Gnatho- stomata comes into intimate relation with the pituitary body in the restricted sense, forming the so-called posterior lobe, or cerebral portion, or nervous portion, of the pituitary body. LITERATURE Agar. Anat. Anzeiger, xxvili, 1906. Agar. Anat. Anzeiger, xxxv, 1910. Assheton. Quart. Journ. Micr. Sci., xxxiv, 1892. Assheton. Quart. Journ. Mier. Sci., xxxviii, 1896. Balfour. Phil. Trans. Roy. Soc., elxvi, 1876. Balfour. Monograph on the development of Elasmobranch Fishes. London, 1878. Bidder und Kupffer. Untersuchungen iiber die Textur des Riickenmarks und die Entwickelung seiner Formelemente. Leipzig, 1857. Boulenger. Tailless Batrachians of Europe. Ray Society, 1897. Braus. Anat. Anzeiger, xxvi, 1905. Budgett. Quart. Journ. Micr. Sci., xlii, 1899. Burrows. Journ. Experim. Zool., x, 1911. Cameron. Proc. Roy. Soc. Edinb., xxiv, 1903. Cameron. Proc. Roy. Soc. Edinb., xxv, 1904. Coggi. Atti Accad.Lincei, Anno 286, Ser. iv, Mem. Class. Sci. fis., vi, 1889, I NERVOUS SYSTEM 143 Cunningham and MacMunn. Phil. Trans. Roy. Soc., B, clxxxiv, 1893. Davies. Morph. Jahrb., xv, 1889. Dendy. Quart. Journ. Micr. Sci., xlii, 1899. Dendy. Quart. Journ. Micr. Sci., li, 1907. Dendy. Phil. Trans. Roy. Soc., B, cei, 1910. Dohrn. Mitt. zool. Sta. Neapel, viii, 1888. Dohrn. Mitt. zool. Sta. Neapel, xv, 1902. Eycleshymer. Journ. Morph., x, 1895. Goeldi. Zoolog. Anzeiger, xxiii, 1900. Gutzeit. Zeitschr. wiss. Zool., xlix, 1889. Harrison. Anat. Record, ii, 1908. Harrison. Journ. Experim. Zool., ix, 1910. i - Havet. La Cellule, xviii, 1901. Held. Die Entwicklung des Nervengewebes bei den Wirbeltieren. Leipzig, 1909. Hensen. Virchows Archiv, xxxi, 1864. Hensen. Arch. mikr. Anat., iv, 1868. Hensen. Die Entwicklunysmechanik der Nervenbahnen der Siugetiere. Kiel and Leipzig, 1903. : Hertwig, O. and R. Das Nervensystem und die Sinnesorgane der Medusen. Leipzig, 1878. Hill. Journ. Morph., ix, 1894. His. Untersuchungen iiber die erste Anlage des Wirbelthierleibes. Die erste Entwickelung des Hiihnchens im Ei. Leipzig, 1868. Kappers, Ariéns. Report XVIIth International Congress of Medicine. London, 1913. Keiffer. Arch. Biol., ix, 1889. - Kerr, Graham. Quart. Journ. Micr. Sci., xlvi, 1902. Kerr, Graham. Trans. Roy. Soc. Edinb., xli, 1904. Kerr, Graham. The Work of John Samuel Budgett. Cambridge, 1907. Klinckowstrom. Zool. Jahrb. (Anat.), vii, 1894. Knape. Anat. Anzeiyer, xxxiv, 1909. Kupffer. Hertwigs Handbuch der Entwicklungslehre. 1906. Leydig. Die in Deutschland lebenden Arten der Saurier. Tubingen, 1872. Neal. Journ. Morph., xxv, 1914. Neumayr. Hertwigs Handbuch der Entwicklungslehre. 1906. Novikoff. Zeitschr. wiss. Zool., xevi, 1910. Oppel. Lehrbuch der vergleichenden mikroskopischen Anatomie der Wirbeltiere. Jena, 1905. Paton. Mitt. zool. Sta. Neapel, xviii, 1907. Rabl, C. Zeitschr. wiss. Zool., lxiil, 189¢ Rathke. Untersuchungen iiber die Entwickelung und den Korperbau der Krokodile. Braunschweig, 1866. Ramon y Cajal. Histologie du systtme nerveux de l'homme et des Vertébres. Paris, 1909. Retzius. Proc. Roy. Soc., B, 1xxx, 1908. Roéthig and Brugsch. Arch. mikr. Anat., lix, 1902. Scammon. Keibels Normentafeln zur Entwicklungsgeschichte, xii, 1911. Schwann. Mikroskopische Untersuchungen iiber die Ubereinstimmung in der Struktur und dem Wachsthum der Thiere und Pflanzen. Berlin, 1839. Smith, Elliot. Anat. Anzeiger, xxxili, 1908. Spencer, Baldwin. (Quart. Journ. Micr. Sci., XXvii, 1886. Strasburger. Jahrb. wiss. Botanik, xxxvi, 1901. Studnicka. S.-B. Béhmisch. Gesellsch., xv, 1896. Thiele. Zeitschr. wiss. Zool., xlvi, 1887. Voeltzkow. Abh. Senckenberg. naturf. Gesellsch., xxvi, 1899. Warren. Quart. Journ. Micr. Sci., xlv, 1902. Wenig. Anat. Anzeiger, xlili, 1913. Winkler. Arch. Entw. Mech., xxix, 1910. CHAPTER III THE ALIMENTARY CANAL THE alimentary canal or enteron of the Vertebrate consists of a tube passing froma the mouth to the anus. The wall of this tube is known technically as the splanchnopleure in contradistinction to the somatopleure or body-wall (Balfour). It consists of an inner lining epithelium, the endoderm, ensheathed in a complex coating of mesoderm—the splanchnic mesoderm—consisting of connective tissue, blood-vessels, lymphatics, nerves, and coelomic or peritoneal epithelium. As is commonly the case in other metazoa the endodermal lining is in the Vertebrate more or less encroached upon at the oral and anal ends of the tube by the spreading inwards of ectoderm. The parts of the tube which come thus to be lined with ectoderm are known as stomodaeum and proctodaeum (Lankester, 1876) while the intervening region lined by endoderm is known as the mesenteron. In the Vertebrata there is very slight develop- ment of proctodaeum but an important section of the buccal cavity is, as will be seen later, stomodaeal in its nature. It is also customary in.embryological writings to use the some- what loose expression foregut for the anterior portion of the alimentary canal (reaching back to the pylorus or to the opening of the bile-duct), which in the meroblastic vertebrates becomes differentiated off from the yolk-sac comparatively early in de- velopment. A good idea of the blocking out. of the main regions of the alimentary canal in one of the lower vertebrates is got by inspecting sagittal sections of embryos and larvae at different stages of develop- ment such as those shown in Fig. 80. From the gastrula stage (A) on to the stage illustrated in Fig. 80, C, the endoderm forms a simple sac with its opening posterior (anus) and with its ventral wall greatly thickened owing to the fact that its cells contain the main store of yolk. From the stage of Fig. 80, D onwards the foregut (fg) becomes gradually constricted off in a tailward direction from the mass of yolk, while at the opposite end of the body, correlated with the outgrowth of the posterior trunk region of the embryo and the backward shifting of the anus, the yolky mass 144 CH. III THE ALIMENTARY CANAL 145 becomes extended into the form of a thick-walled tube—the rudiment of the intestine (en¢). From the stage of Fig. 80, E, onwards active growth of the true tail or postanal region is taking place, and it is noteworthy that, during this process, the endoderm retains for a considerable time its continuity with the mass of actively growing undifferentiated tissue at the tip of the tail and becomes drawn out into a cylindrical postanal gut (pa.g). This remains conspicuous for a time but eventually disintegrates and disappears completely. The main mass of yolk-cells, forming the ventral wall of the middle part of the enteron, gradually shrinks in volume, as the yolk is absorbed and carried off by the circulating blood for distribu- tion to the growing and developing tissues of the body, and eventually the gut wall is no thicker in this region than it is elsewhere. Buccat Caviry.—tThe alimentary canal of the adult Vertebrate commences with the buccal cavity which is in part at least—as shown by the presence within it of placoid and glandular elements corresponding with those of the skin—stomodaeal in its nature. The stomodaeum however is not as a rule developed, as is so usually the case in the Invertebrates, from a simple involution of the ectoderm forming a depression of the surface below the general level. It arises rather by the walling in of-an area on the ventral side of the head through the development of ridge-shaped outgrowths. These ridges may be termed respectively the maxillary ridge and the mandibular ridge accordingly as they give rise later on to the upper or to the lower jaw. The roof of the buccal cavity, or at least its anterior portion, is simply to be looked on as part of the primitive ventral surface of the head, delimited by the maxillary ridge on either side. The floor, on the other hand, represents the mandibular ridge (Fig. 80, H, m.r) which has grown forwards in a direction parallel to the roof. The inner wall of the buccal cavity is in close contact with the anterior extremity of the endodermal alimentary tube but for a time the two cavities remain separated by a thin membranous diaphragm made up of the apposed layers of ectoderm and endo- derm. This may conveniently be termed the velar membrane as the organ known as the velum in Amphiowus or Petromyzon consists simply of the remains of this membrane. The formation of the anterior, stomodaeal, portion of the buccal cavity is seen in its simplest form in some of the lower holoblastic Vertebrates such as Crossopterygians, Lung- fishes or Urodele amphibians. In Fig. 69, D (p. 126) in the case of Protopterus, or in Fig. 100 (p. 178) in the case of Polypterus (see also Fig. 80, H), what will become later the anterior part of the buccal roof is seen to be simply a portion of the ventral surface of the head, bounded behind by the transverse mandibular ridge—the rudiment of the lower jaw—and on each side by the longitudinal maxillary ridge. As is well shown in the figure of Polypterus, and as is also the VOL. II L 146 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. case with Protoplerus, the buccal roof in front, that is to say in the neighbourhood of the mesial plane, passes without interruption into the external skin: in other words the maxillary ridge is not con- tinued to the mesial plane so as to meet its fellow. In later stages pin. Fic. 80.—Sagittal sections through Polypterus. A, stage 14; B, stage 17; ©, stage 20; D, stage 23; E, stage 24 +4. endoderm ; ent, enteric cavity ; fb, cavity of forebrain; fg, of brain floor ; pin, pineal rudiment ; s.¢, cavity of spinal cor a, anus ; av, archenteron ; end, foregut; N, notochord 3 p.f, primary fold d; y, yolk. the roof of the mouth would be hidden in a view from the ventral side owing to the forward growth of the lower jaw. The anterior portion of the buccal cavity in Urodele Amphibians arises in a manner essentially similar to that described above. In the Gymnophiona and the Amniota a characteristic modifica- tion of the mouth margin is brought about by the fact that, as ul BUCCAL CAVITY 147 already mentioned, the maxillary ridge is cut across by the olfactory groove and so divided into the outer maxillary process and the inner median nasal process, the latter of which is continuous with its fellow across the mesial plane, forming with it the so-called fronto-nasal process (see Chap. X.). pin... i. SoORA aac SOURS N SST SEN AST ac. ‘, \ Fic. 80a.—Sagittal sections through Polypterus. F, stage 26; G, stage 29; H, stage 32. a, anus; a.c, anterior commissure; ch, optic chiasma ; l, cloaca; ent, enteric cavity; H, heart; h.c, habenular commissure ; li, liver ; m.r, mandibular ridge ; /p, pituitary involution ; p.c, posterior cominissure ; pa.g, postanal gut; pan, pancreatic rudiment ; pin, pineal rudiment; st, stomach; y, yolk. ’ It is of interest to notice that in various Vertebrates the buccal opening is at first elongated in an antero-posterior direction instead of from side to side. Such is the case with Seylliwm (Sedgwick, see Fig. 81) and Yorpedo amongst Elasmobranchs. In these cases the slit-like mouth is bounded on each side by a longitudinal ridge. 148 EMBRYOLOGY OF THE LOWER VERTEBRATES CB. Later on each ridge becomes sharply bent, about the middle of its length, in such a way as to give the buccal opening a rhomboidal shape and at the same time to mark off the ridge into a maxillary portion in front and a mandibular portion behind. In Anura a somewhat similar arrangement is found. “ENDODERMAL” SECTION oF Buccan Cavity.—The fully developed buccal cavity has incorporated in it a posterior portion— varying in relative extent in different Vertebrates—which is de- rived not from the ectoderm but from the anterior portion of the “endodermal” enteric rudiment. The simplest way in which this portion becomes added to the anterior portion is seen in those Vertebrates in which the anterior part of the enteric cavity is patent throughout develop- ment. In this case the velar membrane simply ruptures—its remnants soon be- coming absorbed — and the stomodaeal cavity is thrown into open communication with the enteric cavity. This is the case in certain Fic. 81.—Ventral view of head region of embryos of Scydlinm uw canicula, (After Sedgwick, 1892.) Anura (Rana) and ors Rat se oarenecan eee in Amniota. , 7-S mm. ; B, slightly more advance han A; ©, 11-12 mm. ; T D, 16 mm. In many Verte- brates no velar mem- brane is present, owing to the fact that the foregut either becomes solid for a time (Polypterus, Fig. 80, D-G) or is so at the beginning (Teleostei, Urodela, Lepidosiren and Protopterus). In such cases the peripheral layer of the yolky foregut rudiment gradually assumes an epithelial character and the yolk along its middle breaks down, so that a cavity arises—continuous with the stomodaeal cavity and forming the hinder section of the definitive buccal cavity. The pro- portion which this posterior portion bears to the anterior section derived directly from the outer surface is very different in different groups. It apparently attains its maximum in Teleosts where it forms practically the whole of the buccal cavity. Points of critical importance to the germ-layer theory are raised in this connexion by the fact that teeth, organs belonging originally to the outer surface, are developed in this posterior region of the buccal cavity from yolky “endoderm.” This is well seen in a Urodele, or a lung-fish such as Lepidosiren or Protopterus (Fig. 82). Theattempt is made to get round this difficulty by assuming that the layer of epithelium which makes its appearance over the surface of the buccal rudiment, and in relation with which the teeth develop, is really an ingrowth from the ectoderm. IL BUCCAL CAVITY 149 It is, as a matter of fact, quite continuous with the ectoderm, bt Th, Fic, 82.—Sagittal section through head region of a Protopterus larva (Stage 33). be, buccal cavity; b.t, anterior boundary of tongue; N, notochord; pin, pineal body; par, paraphysis ; Pit, pituitary body; Th, thyroid rudiment; t.o, tectum opticum. The position of dental rudiments is indicated by the two upward projections of the dorsal wall of the buccal cavity. but examination of carefully prepared celloidin sections (Fig. 83) shows that at its inner end the epithelium passes by imperceptible A B Fic. 83.—Sagittal sections through the region of the buccal cavity of (A) Lepidosiren, stage 80, and (B) Amblystoma, 7°5 mm. in length. b.e, buccal epithelium ; ect, ectoderm ; y, solid mass of yolk-cells in position of buccal cavity. gradations into the ordinary yolky endoderm, with no trace of the sharply defined edge which it would possess were it a layer of 150 EMBRYOLOGY OF THE LOWER VERTEBRATES — CH. ectoderm pushing its way inwards. It extends inwards simply by a process of delamination from the yolky “endoderm.” The real lesson to be learnt from these cases is that the characters of one germ-layer are liable to spread over its boundary into territory belonging to another layer or, in other words, that the territories of the various layers are liable to be separated by an indefinite debatable zone rather than by a mathematically sharp line. It follows that the apparent position of an organ-rudiment in relation to such a boundary is not necessarily to be taken as Fic, 84,---Sagittal sections illustrating the development of the tongue in Urodeles. Aand B, Triton; C, Salamandra (after Kallius, 1901); g.f, gland field ; A/, mandibular arch ; p.t, primary tongue. giving any definitive proof as to which of the two cell-layers that organ belongs to. THE ToncuE.—The tongue is a portion of the buccal floor which becomes demarcated off from the rest by a split formed by a down- growth of the lining epithelium of the mouth. Its mode of develop- ment is well illustrated by what happens in Urodele Amphibians as described by Kallius. Here there develops first a primary tongue, ensheathing the anterior and ventral portion of the hyoid arch (Fig. 84, p.t), which becomes marked off, except at its hinder end, by a deep groove in the floor of the mouth. A horseshoe-shaped thickening of the buccal epithelium now Ill BUCCAL CAVITY 151 develops external to, and parallel with, the groove bounding the primary tongue, and consequently lying on the floor of the mouth between the primary tongue and the lower jaw. The thyroid involution is situated hetween this thickening and the tip of the tongue. The ectodermal thickening develops numerous glands, each originating as a solid ectodermal down-growth, and is known as the gland-field. Externally it is bounded by a shallow groove. Later on the cleft or groove separating the gland-tield from the primary tongue becomes obliterated by fusion of its walls, and the gland-field becomes raised up in a dorsal direction (Fig. 84, B) the tongue-tip shrinking backwards so that eventually the demarcation between primary tongue and gland-field disappears (Fig. 84, C). Meanwhile the groove bounding the gland-field externally becomes deepened. It forms the outer limit of the definitive tongue which is thus a compound structure, its tip and edges developed from the original gland-field, its postero-median part from the primary tongue. In the fishes the tongue remains non-muscular and non-glandular: it is simply the primary tongue. In the Axolotl the tongue appears also to be a primary tongue, the gland-field making a transient appearance as a rudiment but eventually undergoing atrophy (Kallius). In the Amniota the tongue is, as in the terrestrial Urodeles, a compound structure, the primary tongue rudiment becoming fused with an elevation of the floor of the mouth lying in front of the Thyroid rudiment. This elevation, called by His the tuberculum wmpar, represents morphologically the gland-field of the Urodeles. The tongue of Cyclostomes is remarkable for its complexity: it has complex muscular and skeletal arrangements and on its surface it develops the horny spines which function as teeth and simulate teeth in their appearance. In Bdellostoma the tongue develops as a cushion-like swelling of the floor of the mouth at an early period while the velar membrane is still intact. In Petromyzon, on the other hand, it does not develop until the time of metamorphosis. It has already been shown how the olfactory organs come to communicate with the buccal cavity by the posterior nares. In the Amniota these become sunk into a recess in the roof of the mouth ‘and in the higher Reptiles, as in the Mammals, this recess becomes shut off from the buccal cavity. by a horizontal shelf which grows in from the side and meets its fellow to form the palate. How this has come about in evolution is illustrated by the three Lizards shown in Fig. 85. In ontogeny the mode of origin may be similar, the palatine out- growths meeting and fusing with one another in the middle line (Crocodi'es) or, as happens more usually, a median ridge or-septum extends backwards ‘from between the primitive posterior nares, and the palatine processes meet and fuse with its ventral edge. In the two cases the physiological result is the same—the shunting back- 152 EMBRYOLOGY OF THE LOWER VERTEBRATES CH. wards of the communication between olfactory and buccal cavities, a process which reaches its extreme in the Crocodilia where the palate extends back to about the level of the glottis. STOMODAEAL GLANDS.—Whereas in the majority of Fishes the stomodaeal lining possesses only isolated gland-cells, in the air- breathers on the other hand there are developed definite multicellular glands. These originate as a rule from solid down-growths of the Fic. 85.—View of the roof of the mouth in three species of Lizard (A, Lgernia Kingit ; B, Mabuia quinquetueniata ; C, Lygosoma rufescens), illustrating the shifting back of the communication between nose and mouth. (After Voeltzkow, 1899.) ech, recess into which primitive posterior nares open; pal, palatine; »t, pterygoid ; tr, transverse bone; vo, vomer. lining epithelium which develop a cavity secondarily. In Urodeles there is, as already mentioned, a special aggregation of these glands forming the gland-field in front of the tongue, while a single gland of considerable size develops from the roof of the mouth in the region between the olfactory sacs (Intermaxillary or internasal gland). In terrestrial Reptiles glands are present in numbers on the roof of the mouth (Palatine), beneath the tongue on each side of the middle line (Sublingual) and along the edge of the mouth just external to the row of teeth (Labial). The poison glands are specialized and enlarged labial glands of the upper jaw except in Ill PHARYNX 153 oo where they are the enormously enlarged sublingual glands. Similar localized developments of the buccal glands occur in Birds and some of them may reach a great size as, for exainple, the enormous sublingual glands of the Woodpeckers. PHARYNX.—The part of the alimentary canal which follows immediately behind the buccal cavity is highly characteristic from the fact that in Vertebrates it is concerned with the function of breathing. The special organs which are developed to carry out this respiratory function fall into two groups one represented by the Lung—adapted for respiratory exchange with the atmosphere, the other by the Gills—adapted for respiratory exchange with gases in solution in the water. As the balance of probability is in favour of the latter being the more archaic they will here be considered first. The gills are seen in their most typical and familiar form in the various groups of Fishes where there is present upon each side of the pharyngeal region a series of visceral clefts—slit-like openings lead- ing from the pharyngeal cavity to the exterior—separated from one another by masses of solid tissue knawn as the visceral arches or gill septa. The walls of the clefts are highly vascular and their surface is commonly raised into conspicuous plate-like projections— the respiratory lamellae — which serve to increase the area of respiratory tissue. In the most archaic arrangement, seen in Elasmobranch fishes, the front lip of each cleft, except the first, is prolonged backwards to form a small valvular flap overlapping the external opening. In the Holocephali, Teleostomi, and Dipnoi the anterior one of these flaps, that projecting back from the hyoid arch, becomes greatly enlarged to form the operculum which overlaps not merely one but the whole series of clefts lying behind it. Correlated with this the outer portion of each succeeding septum, which in the Elasmobranch gave origin to its valvular flap, has disappeared, leaving only the portion lying next the pharyngeal cavity. The cleft lying in front of the hyoid arech—the spiracle—is usually modified, its respiratory tissue having been reduced and even its opening being diminished in size or completely absent, but its general relations in the adult are such as to permit of no doubt as to its serial homology with the clefts behind it. Most usually there are on each side six clefts—a spiracle and five branchial clefts—but there is reason to believe that there was a greater number present in primitive Vertebrates—seeing that the number of persistent clefts becomes on the whole less as one ascends the vertebrate scale and that here and there among the more archaic forms a greater number than the usual is found (dellostoma up to 14, Notidanus cinereus 7, N. griseus 6). In a few of the more archaic Vertebrates there develop during larval life, in addition to the visceral clefts with their respiratory . 154 EMBRYOLOGY OF THE LOWER VERTEBRATES — cu. lamellae or internal gills, respiratory organs of another type—the external gills. As there is some reason to believe that these are more ancient organs than the gill-clefts they will here be considered first although they are much less familiar than the clefts with their internal gills. The branchial organs will therefore be considered in the follow- ing order : (I.) External gills, (II.) Visceral clefts, (II) Internal gills. (I.) ExreRNaL Gints.—The true external gills are organs which are commonly confounded with the ordinary or internal gills developed in the walls of the gill-clefts. They appear however to be quite in- dependent of these in their origin and they would probably have attracted more attention and interest than they have done had it not been for the fact that they occur in their typical form in only three subdivisions of the Vertebrates (Crossopterygii, Dipnoi, Amphibia) and that two out of these three groups comprise animals of extreme rarity,the developmental stages of which have not been generally accessible to embryologists. — The typical External gill is a projection from the surface of the body on the outer side of a visceral arch. It con- sists of a core of mesen- chyme with a covering Fic. 86.—Diagrammatic longitudinal section through the of ectoderm ; it is tra- early rudiments of the external gills of Lepidosiren (Stage 25). PM@osw’’ Sersed by a vascular loop e.g, external gill; end, endoderm ; v.a, visceral arch ; consisting of the poeae v.c, visceral cleft rudiment. aortic arch which passes out to its tip and then doubles back ; and it commonly has a pinnate form, paired projections growing out so as to increase its respiratory surface. It is provided with muscles by means of which the possessor is able to flick it sharply backwards so as to renew the water in contact with it. The external gill as a rule is without any special skeletal support but in the larval Polypterus a short rod of cartilage projects into its base, and in the extinct Dolichosoma of the Gas coal of Bohemia there was apparently present a well-developed segmented skeleton within the substance of the external gills. The external gill develops as an outgrowth from the tissue of the visceral arch at a period at which the clefts are not yet perforated. It arises as a bulging of the surface (Fig. 86) and in the author's opinion the endoderm of the cleft rudiments takes no part in its formation. At the same time it is only right to state that the pre- valent opinion in the past has been different. The outer surface of the visceral arch in the region where the external gill will develop is covered by a layer of cells thicker than the neighbouring ectoderm, and in some cases this thickened portion of the ectoderm shows in its Ill EXTERNAL GILLS 155 deeper portions a rich deposit of yolk, so as to look exactly like the yolk-laden endoderm. Greil explains this appearance by supposing that true endoderm cells actually spread outwards and replace the deep layer of the ectoderm, so that the external gill-rudiment would be partly endodermal in its nature. There is however no definite evi- dence of any such process taking place and the present writer would interpret the appearances as mean- ing simply that the ectoderm cover- ing the external gill - rudiment becomes thickened, and stores up a supply of yolk in its deeper layers, as a physiological preparation for the active processes of growth which are about to take place as the external gill rapidly increases in length. In this he agrees with Marcus (1908). The general appearance of the developing external gills is well seen in Aypogcophis (Fig. 87) or in Lepidosiven (Fig. 200). In Lepidosiren there are present four upon each side of the body. At first the four are quite independent of one another but as development goes on they become raised upon a common base so as to give the appearance of a single organ with four branches (Fig. 200, B-E). Fic. 87.—Hypogeophis embryos showing development of the external (After Brauer, 1899.) ey, external gill; olfactory organ, Hf, hyoid The rounded knobs seen gills. arch; olf, projecting in B from the hyoid arch, and also from the mandibular arch in front of it, are possibly external gill rudiments which do not yo on with their development. The distribution of true external gills amongst the main groups of Vertebrates is shown in the following table: ' i I ; IT. Il. IV. - ne a Visceral Arch. 1 . j First Secon Phir ‘ Mandibular. Hyoid. Branchial. | Branchial. | Branchial. \ cs Elasmobranchii. Crossoptery sii xl « Dipnoi. | x x x Amphibia | u,* v. x x x Amniota | | VI Fo Branchial. arth * y= Vestigial. In those animals in which they are well developed the external gills are for a time the main functional breathing organs. They are 156 EMBRYOLOGY OF THE LOWER VERTEBRATES Cu. richly vascular and the renewal of the water in contact with their surface is provided for by a well-developed muscular mechanism by which they are sharply flicked from time to time, or, in early stages, by rich ciliation of their surface as in the Frog (Assheton, 1896) or Cryptobranchus (Smith, 1912). They are as a rule merely temporary organs. As the respiratory function comes to be sufficiently per- formed by other organs their circulation becomes sluggish, their tissues moribund. ‘They become invaded by leucocytes and eventually undergo complete atrophy. In Protopterus distinct vestiges persist * for a prolonged period while in various Urodeles they remain func- tional throughout life. The external gills, highly vascular and projecting freely into the surrounding medium, present tempting objects for attack by other organisms. They are therefore extremely liable to injury, and cor- related with this they present a high power of regeneration. In correlation also with the same fact we find that they tend to be eliminated from development in certain members of groups which are as a whole characterized by their presence. Such is the case in the Amphibia where they are characteristic of the group in general but where in particular cases they are reduced (Hyla arborea) or com- pletely absent (Bombinator) although we must believe they were present in the ancestors of these forms. This tendency for the external gills to become eliminated from development in the process of evolution raises the interesting morpho- logical question: were External Gills at any period more widely dis- tributed amongst Vertebrates than they are at present? And, if so, are their vestigial representatives still to be found in any cases where they no longer develop as functional respiratory organs ? This interesting problem, which offers an inviting field for research, has not yet had sufficient attention devoted to it. Even if it were the case that external gills once existed in the ancestors of forms in which they are no longer present as functional organs there is always the possibility if not probability that their disappearance has been so complete as to leave no observable trace. Nevertheless such vestiges might persist and are worth looking for. Under these circumstances it is of interest to note that already certain structures are known which are interpretable as vestiges of once-present external gills. Thus in Gymnophiona what appear to be transient rudiments of mandibular and hyoidean external gills make their appearauce during development (Fig. 87, B). Again in the case of the Mandibular and Hyoid arches of Urodeles, on which no functional external gills develop, Driiner (1901) has found what appear to be vestiges of the muscles of external gills. Again in the larvae of various Urodeles there occurs in connexion with each mandibular arch a curious styliform projection known as the balancer, from the fact that the larva balances itself upon them as upon a pair of limbs (Fig. 88, 6). Each of these has a vascular loop within it and it in fact appears to be the modified external gill of UI EXTERNAL GILLS 157 the mandibular arch which has lost its respiratory and taken on a supporting function. While external gills occur within three main subdivisions of the Vertebrates, namely Teleostomatous fishes (Crossopterygians—the most archaic of existing Teleostomes), Lung-fishes, and Amphibians, there are two main groups—Elasmobranchs and Amniotes—in which they are conspicuous by their absence. Having regard to the tendency of the organs in question to disappear (as in the cases already alluded to amongst the Amphibia) their absence in a special group would not in any case constitute strong evidencethat they were never present in the ancestors of that group. As it happens however there is in the two groups mentioned a definite cause which seems quite competent to account for the dis- appearance of external gills, namely the de- velopment of a new organ —the yolk-sac with its highly developed vitelline network of blood-vessels—which in addition to its primitive function must neces- Vic. 88.—Three stages in larval development of a newt sarily also function as a ae taeniatus) as seen from above. (After Egert, very efficient organ of " = h nd b, balancer ; e.g, external gill of first branchial arch. In Fig. A respiratory exCnaneera what looks like a posterior external gill is the pectoral limb. so render any pre-exist- In Figs. B and C the external gills have been cut away leaving ing respiratory organ no. only their basal stumps. longer necessary. Taking into consideration the presence of external gills in three archaic groups of Vertebrates it seems to the present writer to be clearly indicated that these organs are a very ancient characteristic of the Vertebrate phylum. The only alternative indeed is to regard them as having become evolved independently in the three groups in which they occur. It is difficult to accept this as in any way probable having regard to the similar morphological relations of the organs in question. It might be suggested that somewhere on the course of a large blood-vessel, such as an aortic arch, would be a most natural place for the development of a new respiratory organ. Such a suggestion however is entirely fallacious for simple physical reasons: for new breathing organs will tend to become evolved not on the course of a 158 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. large vessel where the quantitative relation of surface to volume in the blood-vessel is at its minimum but rather where there is present arich superficial network of capillaries, in which the ratio in question is at its maximum. (II.) VisceraL CLerrs.—The visceral clefts develop in what appears to be the most archaic method in Lampreys and Elasmo- branchs where each arises as a lateral pocket (visceral pouch) of the pharyngeal wall which meets and fuses with a, much shallower, ingrowth of the ectoderm, the apposed portion of endoderm and ectoderm breaking down so as to bring about a free communication between pharynx and exterior. Each cleft thus consists of a, usually much larger, inner portion lined with endoderm and an outer portion lined with ectoderm. ‘he most frequent type of modification of this probably primitive mode of cleft development is that so usually met with in the develop- ment of hollow organs, namely that the cleft-rudiment, instead of being a hollow pouch from the beginning, is for a time in the form of a solid lamina of endoderm, which only at a later period develops a cavity in its interior and becomes an open cleft. ‘This modification is found in Teleostomatous fishes, Lung-fishes and Amphibians. In the young Elasmobranch the gill-clefts are at first long slits traversing the whole dorsi-ventral extent of the lateral wall of the pharynx. Hach septum or arch grows back at its outer edge to form a valvular flap overlapping the cleft next behind it. In most cases this backgrowth fuses with the next septum at its dorsal and ventral ends so as to reduce the external opening of the cleft to a compara- tively small dorsi-ventral extent. In all Gnathostomes, excepting the typical Elasmobranchs but including the Holocephali, the hyoidean backgrowth becomes greatly enlarged to form the operculum which overlaps the whole series of clefts behind it. Correlated with this the outer portions of the sub- sequent septa with their backgrowths become reduced. In these cases we frequently find a marked tendency for the edge of the opercular backgrowth to become fused with the body so as to restrict the size of the opening behind it. Thus in the Eel the opercular opening becomes reduced to a small persistent ventral portion, while in Symbranchus the same holds but in this case the two openings have fused together to form a small ventrally placed median pore. A similar condition to this occurs in the tadpole of Discoglossus while in other Anura the persisting opening is displaced to the left side. Finally in Amniotes the fusion of opercular margin with body- wall takes place along its whole extent so that the branchial recion becomes completely enclosed (see Chap. X.). = Sprracte.—The spiracle or hyomandibular cleft always shows a considerable amount of modification. In Elasmobranchs its dorsal portion alone becomes perforate, although fusion of the pouch with the ectoderm takes place throughout its whole dorsi-ventral extent. IIL GILLS 159 Respiratory lamellae develop only on its anterior wall and these, as development proceeds, become vestigial forming the pseudobranch. In Teleostean tishes the spiracular pouch (Fig. 89 A, ve. 1) flattens out and disappears (Goette) so that the pseudobranch (ps.) on its anterior wall comes to he on the inner face of the base of the operculum and appears to belong to the second cleft (Fig. 89, B). In Lung-fishes the solid endodermal rudiment never becomes perforate. It becomes gradually reduced during development while its outer ectodermal portion becomes, as already indicated, converted into a special sense-organ. In Anurous Amphibians and in the Amniota the distal portion of the cleft rudiment becomes greatly dilated to form the tympanic cavity, while the proximal part forms the relatively narrow Eustachian tube. Just as the varying condition of the spiracle indicates a tendency for this cleft to undergo reduction so a similar but still more marked tendency exists for the gill clefts to become reduced at the other (posterior) end of the series. This is illustrated in the first place by the reduction in the number of functional clefts seen in passing from the lower Vertebrates to the higher. It is also frequently manifested in developmental stages. A Thus amongst the Amphibia we find that in the Gymnophiona (Hypoge- ophis, Marcus) a rudimentary 7th cleft makes its appearance though it never reaches the ectoderm, while the 6th is open for a time. In Urodeles a 6th rudiment appears and is for a time Fiu. 89,—Horizontal sections through Salmon embryos explaining position of pseudobranch on inner surface of operculum. (After Goette, 1901.) A.r, aortic root; a, aortic arch; an, anastomotic vessel counecting aortic arches T and IL; Hy, hyoid arch; op, operculum ; Ph, cavity of pharynx; ps, psendobranch ; ve, visceral cleft. connected with the ectoderm but does not become perforate, while in Anura this cleft appears only as a small and transient rudiment which never reaches the ectoderm. (III.) Inrernat Gitis.—The internal gills or respiratory lamellae arise as ridge-like or, at first, finger-lke projections of the cleft lining. The chief matter of dispute regarding their development has been the question whether they belong to the endodermal or the ectodermal portion of the cleft lining. In cases where, as frequently happens, the lamellae begin to develop after the cleft is completely formed, the appearances are sometimes in favour of the one sometimes in favour of the other interpretation. Goette (1901) in fact goes the length of regarding the lamellae as being of endo- dermal origin in the case of the spiracle and ectodermal in the case 160 EMBRYOLOGY OF THE LOWER VERTEBRATES CH. of the succeeding clefts, so that the spiracular pseudobranch would on a strict interpretation of the germ-layer theory not he serially homologous with the other gills. a In the present writer’s opinion, as already indicated, such obser- vations upon the first origin of organs which develop in the region of the blurred boundary between two layers are not to be taken as affording evidence of any serious importance 1n regard to the morphological nature of such organs. Greater weight however seems due to evidence obtained from cases where the first traces of gill lamellae are visible at a period before the bounding membrane of the cleft is ruptured, when the cleft consists still of two distinct pouches—one ectodermal, the other endodermal—separated by a still complete partition. Such is the case in Acipenser and Goette shows that in this case the lamella-rudiments arise outside the partition from what is undoubtedly an ectodermal surface (see Fig. 90, 9.2). The same discussion extends to the general lining of the cleft—as to how much of the lining of the adult cleft is ectodermal and how much endodermal. Goette and Moroff (1902) hold that only ; the portion of the cleft in the imme- YD diate neighbourhood of its pharyngeal penser showing the ectodermal Opening is to be regarded as endodermal, origin of the gilllamellae. (Atter all the rest being ectodermal. But here Goptbe 190i) again in view of the blurred character ies aortic arch ; g.l, rudiment of sill of the boundary between the two layers amella; Hy, hyoid arch ; op, operculum; . 2 Ph, cavity of pharynx. it seems hardly profitable to speculate on the matter. In certain fishes the gill-lamellae are for a time prolonged out- wards into long threads which project through the cleft opening into the surrounding fluid. Such is the case in the embryos of Elasmo- branchs, in which it is only the lamellae upon the posterior face of each arch that become prolonged, those on the anterior face not projecting beyond the edge of the septum. Eventually the pro- jJecting part of the filament disappears while its attached basal portion becomes the definitive lamella. In a few Teleosts a similar temporary modification of the lamellae takes place — perhaps the best example being Gymnarchus (Budgett, 1901; Assheton, 1907. See Fig. 199). Evo.urionary History oF THE BRANCHIAL RESPIRATORY OrGANS.—As regards the early evolutionary history of these branchial. respiratory organs one very generally accepted view looks upon the visceral clefts as being the most primitive, the internal gills as having developed next, and the external gills as being due to secondary extension of respiratory tissue outwards from the clefts. It seems however, bearing in mind what we now know regarding the development and distribution of external gills, at least equally if not \ III LUNG 161 more probable that the evolution of these organs has been in the opposite direction. On this latter hypothesis the external gills would be regarded as the primitive respiratory organs, inherited probably from pre- vertebrate ancestral forms. The evolution of clefts between their bases would be explicable as an arrangement for pumping water over the surface of the external gills, while it could be readily understood that the respiratory tissue would then tend to spread inwards along the lining of the clefts, where it would be both advantageously situated for carrying out its breathing function and, at the same time, protected from the dangers to which external gills are exposed. The development of respiratory lamellae to increase the area of this respiratory tissue on the wall of the cleft would be a further and natural development. The chief difficulty in the way of accepting this as a working hypothesis lies in the existence of animals admittedly near the base of the Vertebrate scale—such as Amphiowus and the Cyclosto- mata—in which there are no external gills and no vascular yolk-sac to account for their disappearance. This difficulty is undoubtedly a serious one but on the whole the present writer is inclined to think the difficulty is not so great as to justify the immediate rejection of the hypothesis: it becomes less formidable when it is borne in mind that the forms mentioned although evidently archaic in some of their characteristics bear in others equally convincing evidence of high specialization. Lunc.—In all the groups of Gnathostomata excepting the Elasmo- branch fishes the pharyngeal wall develops a great outgrowth which, as will become apparent later, is to be looked upon as’ homo- logous throughout the series and as primarily respiratory in its function—the lung. The lung appears in its most familiar and typical form in the tetrapod Vertebrates and its development in these will accordingly be considered first. Here in an early stage of its development the lung is in the form of a pocket of the pharyngeal floor projecting downwards in the mid-ventral line. This pocket commonly makes its first appearance as a longitudinal groove or gutter in the floor of the pharynx at about the level of the last visceral cleft. The groove becomes constricted off from behind forwards, so as to form a blindly ending pocket communicating in front with the pharyngeal cavity by a narrow opening—the glottis—and extending back immediately ventral to the pharynx. The blind end of the pocket grows actively tailwards and becomes deeply bilobed—the two lobes becoming respectively the right and left lung, while the unpaired portion connecting them with the glottis becomes the trachea or pneumatic duct. While the lung passes in its early history through stages corre- sponding on the whole with those described there are differences in detail in ditferent groups--the most conspicuous of these variations VOL. II M 162 EMBRYOLOGY OF THE LOWER VERTEBRATES CH. being, as is so often the case in the development of hollow organs, that the rudiment is at first solid and the cavity appears secondarily in its interior. This is the case in various anurous amphibians and in Lepidosiren and Protopterus. It has been indicated that the lung is primarily a ventrally placed pocket of the pharyngeal wall, that is to say its wall is a portion of splanchnopleure. It follows that the cavity of the lung is lined by endoderm while its outer layers (connective tissue, blood- vessels, muscles, ete.) are composed of splanchnic mesoderm. As regards the further development of the lung, the main steps are concerned with its respiratory function and have to do with the increase of the respiratory surface. In such an animal as the Newt, where the lung retains a relatively primitive condition, the endo- dermal lining grows equally as the organ increases in size, so that even in the adult the lung has the form of a simple sac with smooth endodermal lining. Ina Frog or a Lizard, however, growth activity is specially marked at particular spots so that at these spots the endoderm forms outward bulgings into the covering of splanchnic mesoderm. In these Sauropsida in which the pulmonary apparatus reaches its highest degree of evolution (Tortoises, Turtles, Crocodiles and Birds, in an ascending series) these pockets of the endodermal lining become more and more extensive, and more and more com- plicated, so as to give rise to a thick spongy mass, which forms the bulk of the lung, surrounding the now relatively small clear central space. The latter, forming as it does an apparent continuation of the bronchus or paired portion of the trachea, is spoken of as the intrapulmonary bronchus. Further the respiratory function becomes concentrated towards, the terminal portions of the pockets, their proximal portions forming simply conducting channels— branches of the intrapulmonary bronchus. In the Chameleons, towards the end of development, a number of the endoderm outgrowths bulge out beyond the general level of the surface of the lung upon its ventral side. These persist in the adult as large diverticula which when the animal blows itself out are inflated with air. In the embryos of Birds similar outgrowths make their appearance, four from each lung, but in this case as development goes on the outgrowths continue to increase in size and form the characteristic air-sacs of the adult bird. THE Lune or Birps.—As the Birds, in correlation with the intensely active metabolism as indicated e.g. by their high body temperature, stand pre-eminent amongst Vertebrates in the high stage of evolution which has been reached by their lung, the onto- genetic development of this organ will be followed out in a little more detail (Moser, 1902; Juillet, 1912). In the Fowl the pulmonary diverticulum of the pharyngeal floor makes its appearance about the beginning of the third day. By the end of this day the rudiment is bifurcated at its hind end, each lobe III LUNG 163 being the rudiment of a lung in the restricted sense and containing a prolongation of the enteric cavity lined by tall columnar endoderm cells. Outside the endoderm is a thick layer of mesenchyme and this in turn is covered by columnar coelomic epithelium. The endoderm-lined cavity is destined to become the main intrapulmonary bronchus—the mesobronchus. This remains un- branched until the fifth day when its endoderm begins to bulge out, near the point where it enters the lung, to form the first ento- bronchus. During further development a series of three other entobronchial outgrowths sprout out from the external surface of the mesobronchus close behind the first outgrowth. The four ento- bronchi so arising are closely contiguous and form a longitudinal row (Fig. 92, E1-E4). A set of similar outgrowths make their appearance spaced out along the mesial side of the mesobronchus posterior to the ento- bronchi: these are the rudiments of the ectobronchi. A third set of outgrowths on the lateral side of the mesobronchus ect are the rudiments of the small secondary lateral bronchi (Campana). Of these sets of outgrowths the first and second are the most important and they are arranged in a soae ontie y)) “mes. slightly spiral row along } we par the wall of the meso- De ent. bronchus. The mesobronchus, as Fic. 91.—Diagram illustrating the arrangement of the it - l ath main air-passages in the lung of the Fowl as seen 1b growS In length, from the mesial plane. (After Juillet, 1912.) assumes a somewhat br, bronchus ; ect, ectobronchi ; ent, entobronchi ; u-shaped curvature, by mes, mesobronchus; par, parabronchi. which the group of ecto- bronchi are carried towards the dorsal face of the lung while the entobronchi are nearer the ventral surface (cf. Fig. 91). Both ecto- bronchi and entobronchi grow rapidly parallel with and close to the surface of the lung-rudiment. They soon produce secondary branches as projections of their walls and these secondary branches increase greatly in length traversing the substance of the lung at first close to its median surface and, later, deep down in its sub- stance as well—the entobronchial branches growing in a dorsal and the ectobranchial in a ventral direction. The two sets of branches as their tips approach one another are seen to alternate in position (Fig. 91). When they have approached closely each branch bifurcates and its two tips become closely apposed to the two tips belonging to the other series which lie closest to them. About the thirteenth day these apposed tips become completely fused and their cavities continuous so that there is now 164 EMBRYOLOGY OF THE LOWER VERTEBRATES Cu. established a series of channels. running in a dorsiventral direction through the substance of the lung and communicating dorsally with the ectobronchi and ventrally with the entobronchi. The channels in question are termed parabronchi (Fig. 91, par). These are embedded in an abundant matrix of mesenchyme which from about the tenth day becomes divided up into more or less prismatic masses each having in itS axis an individual parabronchus—the prisms being delimited from one another by the development of intervening blood-vessels. The mesenchyme which constitutes the inner portion of this sheath round each parabronchus becomes later | replaced by a layer of smooth muscle fibres. At about the same period as the fusion of the parabronchial tips takes place, the wall of the parabronchus begins to grow out into numerous little pockets arranged in radiating fashion. These extend outwards, perforating the muscular sheath, and at a short distance from the parabronchus divide into branches which in turn elongate and become the air-capillaries of the fully developed lung. Judging from adult structure it would appear that the tips of these fuse with others to form the continuous air-capillaries so. that the latter would be formed much in the same way as the parabronchi but it has not been possible, so far, to demonstrate this by actual observation. The essential features of the development of the Bird’s lung as above outlined may be summed up in the statement that in this type of lung the diverticula of the intrapulmonary bronchus, which in other Vertebrates end blindly, become here joined together tip to tip to form continuous tubular channels. To allow this arrange- ment to function efficiently an apparatus is needed to force the air through the system of respiratory tubes: such an apparatus is provided by the air-sacs. ArR-SAcs.—The ventral part of the lung-rudiment is for a time formed of a thick mass of mesenchymatous tissue which has been termed by Bertelli the primary diaphragm, from the fact that it becomes continuous along its lateral margin with the side wall of the splanchnocoele, so as to form a kind of floor separating off the lung from the splanchnocoele which lies ventral to it. The air-sacs arise as outgrowths of the bronchial cavities and are on each side four in number: the first or most anterior giving rise to the cervical sac, the second by bifurcation to interclavicular and anterior thoracic sacs, the third to the posterior thoracic and the fourth to the abdominal sac. The rudiments sprout out into the substance of the primary diaphragm and become greatly distended within it, bulging out ventrally amongst the viscera so that the ventral layer of the diaphragm becomes stretched ‘out into a thin membranous wall delimiting the cavity of the air-sac on its ventral side. The dorsal part of the primary diaphragm, lying above the air-sacs, persists as the floor of the lung or secondary diaphragm (ornithic diaphragm of Bertelli, pulmonary aponeurosis of Huxley). Ill AIR-SACS 165 The air-sac rudiments sprout out (Fig. 92) from the main pulmonary cavities—the cervical from the first entobronchus, the interclavicular and anterior thoracic jointly from the third ento- bronchus, the posterior thoracic and the abdominal from the meso- bronchus. Later on additional secondary communications between the air-sac cavity and the pulmonary cavities are established (except in the case of the cervical air-sac) by means of the recurrent bronchi of Juillet. These arise in the ordinary fowl about the tenth day of incubation in the form of ‘outgrowths of the wall of the air-sac either near its tip (interclavicular and anterior thoracic) or just before it emerges through the general surface of the lung (posterior thoracic and abdo- minal) as shown in Fig. 92. These outgrowths burrow into the superficial layer of 7-~~ the lung, branch and become joined up, in a manner the ane details of which have not yet been worked out, with the a system of parabronchi. The © communications are visible in suitable preparations of the adult lung as groups of openings, each group leading ‘A into the lung from the appro- a priate air-sac—those of the interclavicular and anterior thoracic lying towards the Fic, 92.—Diagrammatic view of the right lung of lateral edge of the ventral a Fowl embryo pee cies Rd as Se ae surface of the Tung, about the he renal soe, Marat he ie attachment of entobronchi are shaded. the bronchus, and those of ab, abdominal air-sac; at, anterior thoracic air-sac ; the posterior thoracic and cer, cervical air-sac; E1 and £4, first and fourth ento- sbdominall sacs lying near ‘esl; to nbcivienss nate; na motecatts the hind end of the lung, ">" Staccato hi close to the direct opening Riwest it and the corresponding air-sac. It would appear that the function of these recurrent channels is to conduct the air forced out of the air-sacs in the expiratory effort through the system of air-capillaries, the muscular coat of the parabronchi doubtless playing an important part in directing the passage of the air through the system of air-capillaries rather than through the parabronchi themselves. The formation of the air-sacs does not exhaust the remarkable proliferative powers of the wall of the lung in Birds. Further out- cer _éL lc 166 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. growths arise from the walls of the air-sacs which burrow through the neighbouring tissue, even through bone, taking the place of the marrow and rendering the bones pneumatic. Such outgrowths may extend even into the terminal phalanges of the digits. They may also extend in amongst the connective tissue of the skin or between the muscles.? THE LUNG IN FISHES In the typical fishes or Teleostei, which of all Vertebrates are the most highly specialized in adaptation to a purely swimming habit, one of the most characteristic organs is the swim-bladder or air-bladder. In its most highly developed form (in the physoclistic Teleosts) this consists of a closed sac, lying above—dorsal to—the splanchnocoele and filled with gas containing a large proportion of oxygen. Special developments of the lining epithelium provide a mechanism whereby the amount of gas in the organ can be increased by a process of secretion or diminished by a process of absorption. This mechanism, which is under the control of the nervous system, has for its main function the keeping the body of the fish at the same constant specific gravity as the water in which it is swimming — counteracting changes in its specific gravity which would otherwise result from variations of pressure due to change of depth, or from variations in volume of gas produced by fermentative processes in the alimentary canal. The air-bladder with its com- pensating mechanism keeps the fish precisely at the specific gravity of the surrounding medium so as to obviate the expenditure of muscular effort in order to keep at one depth such as is necessary in the case of a shark or other fish unprovided with an air-bladder. The air-bladder arises in development as an outgrowth of the wall of the alimentary canal behind the region of the gill-clefts. This burrows its way backwards dorsal to the splanchnocoele and eventually attains to a large size. Its tubular connexion with the alimentary canal (pneumatic duct) becomes constricted across and severed so that the organ is completely isolated from the alimentary canal. In a good many cases however—namely, in the physosto- matous Teleosts—the duct persists and remains patent throughout life. In many fishes the dorsal wall of the air-bladder bulges out in a headward direction (Fig. 93) forming a diverticulum which may reach a great size so that in the adult the organ has the appearance of being composed of two segments marked off from one another by a constriction (Fig. 93, B), the pneumatic duct communicating with the hinder one of the two. The constriction may become accentuated ? Since the above account was written a full and well-illustrated description of the development of the Fowl’s lung has been published by Locy and Larsell (Amer. Journ. Anat., xix and xx, 1916). These authors’ results amplify and in the main confirm those of Juillet. ; Tir ATIR-BLADDER Souk Sars: \ Q Cc Fig. 93.—Development of the air-bladder of a Teleost. (After Moser, 1904.) A, Rhodeus, 5 mm., longitudinal section;, B, Rhodeus, 6 mm., longitudinal section; C, Rhodeus, 7 mm., transverse section, showing small pouch-like outgrowth of pneumatic duct; end, endoderm ; ent, enteric cavity; J, air-bladder; li, liver; N, notochord; xc, pronephric chamber; p.d, pneumatic duct ‘y, yolk. } 167 168 EMBRYOLOGY OF THE LOWER VERTEBRATES — CH. to form a kind of diaphragm perforated in its centre and capable of being thrown into vibration by air being forced from one chamber into the other so as to function as a sound-producing organ (¢.g. Gurnards). Other outgrowths may develop: thus for example in many Siluroids numerous branched projections are formed along each side of the air-bladder. The air-bladder rudiment is at its first appearance in some cases approximately dorsal in position (Salmo). In Rhodeus Moser (1904) has shown that the diverticulum is at first on the right side of the alimentary canal. The same observer found that during the early stages of development of the air-bladder the portion of alimentary canal from which it springs undergoes a process of rotation about its long axis in such a direction that a point on its dorsal surface: is carried towards the left side. Although the actual development has been worked out only in a few cases, we may infer safely from the adult relations (Rowntree, 1903) that the amount of this rotation differs greatly in different members of the group Teleostei. Thus in Siluroids and Cyprinodonts the glottis or pharyngeal opening is in the adult still to the right of the mesial plane; in others such as the genera Osmerus, Clupea, Chirocentrus it is practically median; in still others such as Mormyrids, Characinids, Gymnotids and Cyprinids it has passed the mesial plane so as to lie upon its left side, while in the case of the Characinids Macrodon, Erythrinus and Lebiasina the glottis has come to be completely lateral on the left side. This rotation of the gut in the region of the glottis is of much morphological importance as will be shown later. In the young Rhodeus, 7 mm. in length, Moser finds that a well- marked diverticulum from the pneumatic duct is present (Fig. 93, C). Later on it gradually disappears. A similar diverticulum occurs in Salmo and in the Carp, and in all probability in numerous other Teleosts : its morphological significance will be discussed later. ACTINOPTERYGIAN GaNnorps.—In these fishes the development of the air-bladder takes place on similar lines to that described for Teleosts. In Ama the additional detail has been made out that the rudiment is at first in the form of a longitudinally placed groove which becomes constricted off from the alimentary canal from behind forwards just as frequently happens in the case of the typical lung- rudiment of air-breathing Vertebrates (Bashford Dean, 1896; Piper, 1902). A rotation of the section of alimentary canal in the region of the glottis takes place similar to that which occurs in the Teleost. LuNG-FISHES.—In the adult Ceratodus an organ occurs which is equally lung and air-bladder. It forms an unpaired sac lying dorsal to the splanchnocoele just like a typical air-bladder, but the pneu- matic duct, instead of opening directly into the alimentary canal dorsally, passes round the right side and opens by a ventrally placed glottis. In Lepidosiren and Protopterus the general arrangement is Ir LUNG 169 the same except that here the organ is deeply bilobed: a right and a left lung or air-bladder occupying the place of the single organ of Ceratodus. The meaning of the ventral position of the glottis in these Lung- fishes, and, in fact, the morphological nature of the whole organ, is Fic. 94,—Transverse sections through the endoderm of the pharynx showing an early stage in the development of the lung. A, Polypterus, B, Ceratodus, and C, Bombinator (C after Goette, 1875). 7, lung-rudiment ; ph, pharynx. demonstrated by the examination of early stages in development. In these the organ is found to be a perfectly typical Jwng-rudiment (Fig. 94, B)—a mid-ventral projection from the pharyngeal floor of precisely the same kind as that found in tetrapodous vertebrates (C).! A Cc Fig. 95.—Views showing early stages of the lung-rudiment of Protopterus as seen from the ventral side (stages xxxii, xxxiv, xxxv). e.g, external gill; 2, lung; oes, oesophagus ; pan, dorsal pancreas ; p.f, pectoral limb; 7h, thyroid ; vc, visceral cleft rudiment. (Cut surfaces are indicated by uniform light tone.) Subsequent stages are illustrated by Figs. 95 and 96. The lung rudiment at first a rounded knob (Fig. 95, A) grows backwards and soon becomes bilobed (B). The figure does not bring out one im- portant fact namely that the lung-rudiment as it grows backwards 1 The projection is at first solid in the case of Lepidosiren and Protopterus. 170 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. twists upon itself, in such a way that points upon its ventral surface would move towards the embryo’s right side. (In other words the: lung-rudiment rotates about its long axis in a counter-clockwise direction as seen from behind, its front end remaining fixed.) The two lobes are the right and the left lung-rudiment but on account of the rotation just mentioned which extends through more than 180° the left lobe at this stage represents what was originally the right side of the rudiment. The two lungs of Lepidosiren or Protopterus are thus reversed in fan = A B Cc D Fic. 96.—Dissections of mid-gut of Lepidosiren at stages 32 (A), 35 (B), 36 (C), and 37 (D), showing the modelling of the intestine and also the later stages in the development of the lungs. Seen from the dorsal side. c.c, cloacal caecum ; int, intestine ; 1./, left lung; li, liver; mn.d, Wolffian duct ; pan, pancreas ; ph, pharynx ; 7.1, right lung; sp, spleen. : position—the right lung of these forms being homologous with the left of other Vertebrates. An important detail is that in early stages the original right lung, 7.e. the definitive left, is decidedly larger than its fellow (Fig. 95, B). In later stages this inequality disappears, the smaller lung overtaking the other in its growth (Fig. 96). In the case of most individuals the lungs assume their dorsal position simply by growing directly tailwards, the oesophagus being pushed out of the way towards the left side (Graham Kerr, 1910). In certain specimens however, which doubtless in this respect retain ur LUNG nlvae the archaic mode of development, the lung-rudiment (Fig. 97, /) describes a spiral curve round the oesophagus so that the bifurcated Hintunl Fie, 97.—Portions of-transverse sections through a Lepidosiren larva (stage 34) to illustrate the changing relations of lung to gut from a short distance behind the glottis tailwards. In A the lung is ventral to the alimentary canal ; in B it is directly to the right ; in C it has become displaced dorsally ; while in D (where it is commencing to bifurcate) it has come to be mid-dorsal in position. A, aort1; gl, glomerulus of pronephros ; 7, lung; N, notochord ; oes, oesophagus. hinder end of the rudiment, which will give rise to the lungs in the restricted sense, comes to lie dorsal to the alimentary canal. The lungs continue their tailward growth in the substance of the 172 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. dorsal mesentery (Fig. 97, D) but eventually the portion of this mesentery containing the lung and dorsal to it becomes greatly thickened from side to side and finally merges completely in the roof of the splanchnocoele, so that in the adult condition the lungs he completely outside the body-cavity—between it and the vertebral column. In Ceratodus (Gregg Wilson, 1901; Neumayr, 1904) the lung is at first, as in the other two lung-fishes, ventral in position (Fig. 94, B) but in this case the originally left lung, which in Lepidosiren and Protopterus is for a time during development reduced in size, seems to have disappeared almost entirely, being represented only by a small and transient rudiment. Further detailed studies of the early stages in the development of the lung of Ceratodus are much needed to make clear the origin and fate of this vestigial left lung. But it seems clear from what is already known that the monopneumatic condition of Ceratodus has come about in evolution through the suppression of the originally left lung. As the lung completes its development, its cavity becomes en- eroached upon by two median longitudinal ridge-like ingrowths, one dorsal and the other ventral. It used to be supposed that these marked an incipient division of the lung into a right and a left half so as to bring about the condition seen in Lepidostren or Protopterus —the monopneumatic condition being supposed to be the more nearly primitive. It will have been gathered from what has been said that this point of view is no longer tenable and that the mono- pneumatic condition of Ceratodus is to be looked on as secondary and not primary. CROSSOPTERYGIANS. — Of the two surviving examples of the Crossopterygian ganoids—the most archaic existing members of the Ganoid-Teleostean stem—a few stages in the development of the lung have been investigated in Polypterus (Graham Kerr, 1907). In the earliest stage observed the lung-rudiment -was in the form of a mid- ventral groove formed by an outgrowth of the pharyngeal lining (Fig. 94, A,/). This groove becomes deeper and towards its posterior end widens out ventrally soas to have a L-shape in transverse section. Posteriorly the lung-rudiment grows back into a pair of horn-like projections—the rudiments of the right and left lung. These extend backwards in the connective tissue of the splanchnopleure and they very soon show a marked inequality in their rate of growth the left lagging behind the right. As growth goes on this inequality becomes more and more marked, so that in a larva of about 30 mm. in length the right lung extended right back to the cloaca while the left pro- jected back only about 3 mm. behind the glottis. In these later stages another important feature is to be noticed, one which is correlated with the fact that the air-filled lung neces- sarily acts as a float in an aquatic animal. This feature is that the lung tends to assume a position symmetrical about the median plane. Thus in the anterior region where both lungs are present they are Ill ATR-BLADDER 173 situated laterally, balancing one. another, while farther back where only the right lung is present this shifts towards the mesial plane until it is symmetrical about that plane, lying in the dorsal mesentery (Fig. 98, A and B). EVOLUTION OF THE AIR-BLADDER.—The facts that have been enunciated above, with regard to the development of the lung in Dipnoan and Crossopterygian fishes, are of much morphological interest. When pieced together with what has been said regarding the development of the air-bladder of Teleostean fishes they afford data from which the evolutionary history of the Teleostean air- bladder can be traced out with a high degree of probability. That history may be stated in a few words to have probably been as follows: 1. The primitive condition was that of a lung, communi- cating with the pharynx by a ven- trally placed glottis —for we have seen that the embryonic rudiment of the organ in the most archaic forms pos- sessing it is a typical lung-rudiment. 2. The organ became bilobed, growing back into a right lung and a lett Fig. 98.—Sections through the lungs of a larva of lun Polypterus 30 mm. in length. B. Aunt 3 In the forms A, more anterior; B, more posterior; A, aorta; ent, enteron; 1.1, . left lung; N, notochord ; opn, opisthonephros ; p.v, pulmonary veins ; which took to a r.l, right lung; v, interrenal vein. purely swimming existence, and became specialized in the direction of adaptation to this, there came about an asymmetry of the lungs, the right lung increas- ing and the left lung diminislting. Why this should have happened is not yet absolutely certain : it may probably have been in adapta- tion to active movements of lateral flexure, for we see the same thing taking place in Gymnophiona, Snakes and Snake-like Lizards. That it has been the right rather than the left lung which has increased in size, is probably correlated with the rotation of this region of the alimentary canal in a counter-clockwise direction as seen from behind (see p. 168) which would tend to interfere more with the circulation through the left lung than with that through the right, by lengthening the course of the left pulmonary artery. Steps 174 EMBRYOLOGY OF THE LOWER VERTEBRATES cn. in the development of this asymmetry are seen in Polypterus and in the Lung-fishes. 4. In purely aquatic creatures the dictates of adaptation would naturally cause the air-filled lung to assume a dorsal position. An. initial phase of this is repeated in Polypterus where the right lung has become dorsal and median in its hinder portion. In the Lung- fishes a further step is taken—the whole of the lung becoming dorsal except the pneumatic duet which still remains to mark out the path by which the lung moved dorsalwards round the right side of the alimentary canal. That the movement dorsalwards was round the right side was no doubt due to the right lung being predominant and the left reduced in size. In the case of Ceratodus the predominance of the original right lung has been retained, the other being completely obsolete except for a short period during development. In Lepidosiren and Protopterus, on the other hand, the lopsidedness disappears, the original left lung regaining during ontogeny its primitive equality in size with its fellow. 5. In the Actinopterygians—those fishes which show the highest degree of evolution in adaptation to a swimming mode of life—the lung has in the course of its evolution passed through similar stages to those exemplified by Polypterus and Ceratodus. Here again only the original right lung persists as the air-bladder, the vestige of the left lung being possibly represented by the little diverticulum found by Moser upon the pneumatic duct in early stages of development.1 In the Actinopterygians a further step onwards has been made in that the glottis has assumed a dorsal position. This is fully ex- plicable by the rotation which this part of the gut has undergone, aided no doubt by the principle of economy of tissue which would tend to bring about a shortening of the unnecessarily long pneumatic duct. In some cases there still persist vestiges of the ancient cellular respiratory lining of the swim-bladder (e.g. Lebiasina, Erythrinus). 6. Finally in the Physoclistic forms—the most highly specialized of all—the swim-bladder has become completely isolated from the gut, its respiratory function has gone and it subserves a mainly hydro- static function. The outline given above represents a scheme of evolution which in the light of modern research has a high degree of probability. Of course as in all such evolutionary speculations there exist details which are still difficult to explain. While most of the facts of com- parative anatomy fit in well with it, some do not—such as, for example, the nerve-supply and the blood-supply of the air-bladder of Amta—but it may be anticipated with considerable confidence that these difficulties will be lessened or disappear with the progress of research. 1 See p. 168. This matter affords an interesting subject for further research. III THYROID 175 DERIVATIVES OF PHARYNGEAL WALL OTHER THAN THE RESPIRATORY ORGANS THYROID.—The Thyroid gland arises as a mid-ventral outgrowth of the pharyngeal or buccal floor about the level of the Hyoid hull fiviiliuil Fic. 99.—Sagittal sections through anterior portion of alimentary canal of Lepidosiren illustrating the development of the Thyroid. A, B, C from specimens of stage 30; D, stage 31; Th, thyroid; t, tongue. arch. In those Vertebrates in which the pharyngeal rudiment is solid at this stage the thyroid outgrowth is also solid at its first appearance (Fig. 99, A, 7h) and develops its cavity secondarily by cytolysis. 176 EMBRYOLOGY OF THE LOWER VERTEBRATES cn. The Thyroid becomes gradually constricted off from the pharynx (Fig. 99, B and C) remaining for a time connected by a narrow stalk or duct with the pharyngeal or rather buccal floor just in front of the primary tongue (see Fig. 82, p.. 149). This stalk of attachment becomes nipped across and the thyroid forms a mass (Fig. 99, D) or vesicle rounded in form or somewhat elongated in an antero-posterior direction lying in the mid-ventral line beneath the pharynx and just in front of the ventral aorta. The originally simple vesicle undergoes a process of sprouting and division by which it becomes converted into a mass of rounded vesicles, each possessing a wall composed of a single layer of cubical epithelial cells and separated from its neighbours by highly vascular mesenchyme which penetrates in between the vesicles to form the stroma of the organ. During later development the Thyroid undergoes characteristic changes of form in different subdivisions of the Vertebrata. Thus in Teleosts it frequently assumes a more or less diffuse character, the follicles being distributed in the neighbourhood of the ventral aorta and roots of the afferent branchial vessels. In the Amphibia and Amniota the organ becomes deeply constricted into two laterally placed lobes which may remain connected or may become separated, so that it assumes a paired character as happens in Amphibians and Birds. With the processes of differential growth involved in the develop- ment of the neck, the thyroid may undergo considerable displace- ment from its point of origin. Thus in adult Lizards it lies across the trachea well forwards from its hind end while in other reptiles and in birds it lies farther back close to the roots of the great arteries. It is now generally accepted that the clue to the phylogenetic history of the Thyroid is afforded by its development in Petromyzon (W. Miiller, 1871). Here there develops a mid-ventral outgrowth of the pharyngeal floor, forming a short gutter in the branchial region, the lining of which is composed partly of glandular cells which secrete a sticky mucus and partly of cells which bear powerful flagella. Morphologically this gutter is the same as the endostyle of Amphiovus and during larval life its function is also similar: it appears to be in fact simply a shortened up endostyle. The slit-like pharyngeal opening becomes gradually reduced in length till it forms merely. a small pore. At the time of metamorphosis the pore becomes obliterated so that the organ becomes a closed vesicle underlying the pharynx. This vesicle divides up into a number of small vesicles and its mucous secretion accumulates in their interior as a colloid substance like that of the Thyroid vesicles of the Gnathostomata. Ina word, the endostyle of the Ammocoetes stage becomes the Thyroid of the adult, and there seems no reason to doubt that the same has happened in phylogeny and that the thyroid of the Vertebrate is UI BRANCHIAL BUDS 177 simply the modern representative of the endostyle of the proto- chordate ancestor. An interesting feature is that while the physiological importance of the thyroid in the modern Vertebrate is that of a ductless gland for the production of internal secretion to be absorbed by the blood, it still goes on producing the mucous material used by the far back protochordate ancestor for entangling food particles, though that substance is no longer, owing to the disappearance of the duct, discharged into the pharyngeal cavity. BraNncuiaL Bups.—There make their appearance in the develop- ing Vertebrate a series of bud-like proliferations of the endodermal epithelium of the branchial clefts which may be known as branchial buds. They appear at the upper and lower angles of the clefts and the series shows its fullest development in the Lampreys, where buds develop at the dorsal and ventral angles of all the clefts. In the majority of fishes investigated they have been found to appear at the dorsal angles of all the clefts except the first; in Urodele Amphibians at the dorsal angle of all clefts and at the ventral angle of IL, III. and IV.; in Anura at the dorsal ends of I. and II. and at the ventral ends of II.-V.; in Lacerta at the dorsal ends of I-III. and the ventral ends of III. and IV.; in Gallus at dorsal and ventral ends of ITI. and IV. The morphological significance of these organs is still completely obscure. Physiologically some of thera appear to be of importance during the later stages of development preceding sexual maturity inasmuch as they give rise to that often bulky organ the Thymus. This arises by the fusion together of more or fewer of the dorsal buds, the others undergoing no further development. Thus in Lepidosiren (Bryce, 1906) dorsal buds III. and IV. develop into thymus while IJ. and V. undergo no further development: in Ceratodus (Greil, 1913) IL., III. and IV. give rise to Thymus while V. and VI. do not develop further: in Hypogeophis II., II., IV. and V. give rise to Thymus while rudiments on I. and VI. atrophy. In regard to the much discussed histogenesis of the thymus all that need be said here is that the originally solid epithelial rudiment becomes in the course of development loosened out into a sparse reticulum interpenetrated by mesenchyme richly traversed by blood- vessels and crowded with leucocytes. The ventral buds, where they occur, become constricted off from the branchial epithelium forming simple rounded masses of epithelial cells (Amphibians) or they may be subdivided up by intrusive con- nective tissue into solid portions (Reptiles) or hollow vesicles (Birds). The small organs so formed are termed by their discoverer Maurer epithelial bodies: their physiological significance is quite unknown. There normally develops in the Vertebrate either on both sides or only on the left side a small pouch-like diverticulum of the pharyngeal wall close to the ventral edge of the last gill cleft, what- ever the number of this be in the morphological series. The VOL. II : N 178 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. diverticulum becomes separated from the pharynx and commonly gives rise to numerous rounded vesicles somewhat resembling those of the thyroid in appearance. The organ thus formed was named by van Bemmelen who discovered it in Elasmobranchs-—suprapericardial body—while Maurer has termed it the postbranchial body. Nothing is definitely known regarding either its function or its evolutionary history, though it is sometimes regarded as representing a vestigial last gill-pouch. A curious point is the tendency of the organ to unilateral development as it makes its appearance only upon the left side in a large number of cases (Acanthias, Lepidosiren and Protopterus—see Fig. 109, B—most Urodeles, some Lizards). CEMENT ORGANS OF TELEOSTOMATOUS FisHES.—It has long been Fic. 100.—Ventral views of Polypterus larva to show the cemeut-organs. A, Stage 30; B, Stage 33; c.o, cement-organ ; ¢.n, olfactory organ; m, mouth ; V, ventricle of heart. known that the larvae of Actinopterygian ganoids possess cement- organs on the head in front of the mouth. Balfour (1881) wrote of this as “a very primitive Vertebrate organ, which has disappeared in the adult state of almost all the Vertebrata; but it is probable that further investigations will show that the Teleostei, and especially the Siluroids, are not without traces of a similar structure.” The organs in question were generally regarded as being developed from a thickening of the ectoderm. Miss Phelps (1899) first stated that they originated from endoderm (Amia) and the present writer, at the time ignorant of her work, was greatly surprised to find himself forced to this same conclusion by the examination of Budgett’s material of Polypterus. The cement-organ of Polypterus (Graham Kerr, 1906 and 1907), when at the height of its development, forms a stout cylindrical structure with a deep hollow at its free end, projecting from the Ill CEMENT-ORGANS 179 head on each side as shown in Figs. 100, A, and 197, C, ¢.0. A longi- tudinal section through the centre of the organ at about this stage (stage 26, Fig. 101, E) shows that the organ is covered by the ordinary 2-layered ectoderm. Round the lip of the opening at its Vt eet Fic. 101.—Tllustrating the development of the cement-organ of Polypterus. B represents part of a transverse section, the other figures portions of horizontal sections. A and B, stage 20; C, stage 23; D, stage 24; B, stage 26. ¢.0, cement-organ. The darker tone indicates ectoderm. tree end, the superficial layer of ectoderm stops, while the deep layer seems to dip down asa deep involution toform the secretory epithelium (c.o) which lines the cavity. All the appearances seem to point to the secretory epithelium being ectodermal in its nature. How deceptive these appearances are will be gathered trom an inspection of Fig. 101, A-E. 180 EMBRYOLOGY OF THE LOWER VERTEBRATES — cu. The first rudiment of the organ is seen to be a simple pocket-like outgrowth of the gut-wall (A, ¢.0): this becomes more and more prominent (B, ©): it becomes gradually constricted off at its base from the gut-wall, its cavity becoming isolated first (D). Finally it separates completely from the main endoderm and its outer end undergoes fusion with the deep layer of the ectoderm. Its cavity then opens to the exterior and the fully functional condition is reached—the endodermal origin of the secretory lining being for a time betrayed by the conspicuous persistent yolk granules in its cells. It will be noted that the exposed side of the secretory epithelium, that on which the secretion is extruded, is that which originally faced inwards towards the lumen of the alimentary canal. In other words the direction in which the extrusion takes place is morpho- logically the same as that of any other part of the glandular lining of the gut-wall. As is the case in other forms the cement-organ is a transient, purely larval, structure. About stage 31 (Fig. 197, D) degeneration commences: the gland shrivels up, the gland-cells becoming more slender and dark pigment making its appearance in their interior, the epithelium becomes penetrated by ingrowing blood-vessels, its cell-boundaries become indistinct. The process of atrophy goes on rapidly and by stage 36 (Fig. 197, F) the organ has completely disappeared. An interesting variation from the normal course of development is found in specimens in which the cement-organ rudiments are more or less approximated to one another. This variation reaches its maximum in occasional individuals in which they are completely fused and form an unpaired structure, continuous across the mesial plane. In the actinopterygian Ganoids the cement-organ develops along the same general lines as those just indicated. In the Sturgeons the development has been worked out recently by Sawadsky (1911) in Acipenser ruthenus. Here the organ forms a rounded projection, very much in the same position as that of Polypterus, but in this case each becomes divided by a groove so as to form two rounded knobs. These knobs eventually grow out to form the tactile barbels of the adult, the secretory epithelium being carried out on the surface of the barbel as it grows. The secretory epithelium is here also endodermal, its rudiment being the gut-wall immediately dorsal to the position in which the mouth will develop later and being continuous across the mesial plane. The unpaired condition which occurs in Polypterus as a variation is thus normal in the case of the sturgeon. As the head increases in length the secretory epithelium becomes carried out on its ventral surface, looking just as if it were the thickened ectoderm of this surface. Finally the paired condition comes about, the lateral parts of the secretory epithelium coming to be supported by the knob-like projections already mentioned. Il CEMENT-ORGANS 181 Amia is of special interest in regard to its cement-organs as it was in this form that their endodermal origin was first announced. The organs are for a time in the form of a pair of rounded knobs, one on each side, but these take on a crescentic shape so that together they form a circular wall, interrupted anteriorly and posteriorly. Each organ contains a pocket-like projection of the gut-wall which takes on a somewhat sausage-like form in correlation with the curved shape of the organ as a whole. This endodermal sac separates from the main endoderm and becomes constricted across, so as to form a curved row of closed vesicles from six to ten in number. Each vesicle fuses with the ectoderm and develops an opening to the exterior so that it takes on the appearance of a cup at first deep and narrow, later shallow and wider, its lining continuous with the deep layer of the ectoderm. When the larva reaches a length of 13-14 mm. it makes less use of its cement-organ and the latter commences to degenerate, sinking beneath the surface with which, however, it remains connected by a narrow tubular channel. By about the 20 mm. stage this has disappeared and soon there is no trace of the organ to be found even in sections. In Lepidosteus the organ appears to be similar while in the other ganoids its development still remains to be worked out. These cement-organs are of special interest and importance for more than one reason. In the first place they are of importance in revealing a quite unexpected pitfall in the way of the investigator trained to have implicit faith in the germ-layer theory, for they show how a particular organ may become transferred from one germ-layer to another even though not belonging to the transitional zone where the two layers are continuous. A very common modification of ontogenetic development consists in the slurring over or even omission of particular stages in early development. Were this to happen in the case of the early stages in the development of the cement-organ say of Polypterus, it is easy to see that the organ might have every appearance of being purely ectodermal in its nature, although it is, as a matter of fact, endodermal. It appears to the present writer quite possible, if not probable, that this modification has actually come about in the Dipnoi and Amphibians, and that the cement-organs of these groups, although they develop from the ectoderm in those forms which have been investigated (p. 79), are really homologous with the cement-organs of the Teleostomi, their endodermal stage having been eliminated from ontogenetic development. Further investigations are needed in the Amphibia—to see whether no trace exists, in any member of the group, of an original connexion with the endoderm. ; As regards the original nature of these organs it is impossible to arrive at any certain conclusion. Arising as they do in the form 1 Phelps (1899). The actual discovery seems to have been made by, Reighard. Cf. Reighard and Phelps (1908). 182 EMBRYOLOGY OF THE LOWER VERTEBRATES cH. of endodermal pockets, they obviously recall gill pouches on the one hand and coelenteric pouches on the other. Their position suggests a pair of premandibular gill pouches: their function, that of forming an excretion (cement), perhaps indicates rather coelomic affinities and the present writer suggested (1906) their possible correspondence with premandibular head cavities of other Vertebrates. Reighard and Phelps (1908) homologize them with the anterior pair of head-cavities of Elasmobranchs while ‘ van Wijhe (1914) sup- Fre. 102,—Larva of Sarcodaces odie. (After Budgett, 1901.) ports a homology with the ciliated organ of Amphioxus. Altogether these cement-organs are very interesting and puzzling structures and would well repay further investigation. A thorough comparative study should be made of their development in the archaic Crossopterygians and of their possible homologues in Elasmobranchs. Little is known regarding cement-organs in Teleosts, though it is probable they will be found to occur in, various tropical fresh- water fishes. Budgett Lf c.o, cement-organ, (1901) found a_ large cement-organ on the head of the larva of the Characinid Sarcodaces odée (Fig. 102,¢.0). Ina larva believed to be that of the Mormyrid Aypero- prsus bebe he found six well-marked cement glands on the head which in this case secrete fine threads by which the larva hangs suspended in the water until the Fic. 103.—Teleostean larvae, supposed to be those of . Hyperopisus bebe, suspended from the rootlets in the yolk is used up (Fig. ; nest. (From Budgett, 1901.) 103). Heterotis and Gymnarchus also possess similar organs—very small in the latter case (Assheton, 1907). The organs in these various fishes present the appearance , of being ectodermal thickenings: we have as yet no information as to whether, as may be suspected, they really originate from the endoderm. DicestivE Tract.—- The respiratory region of the alimentary canal is succeeded by the true digestive tract and this shows more or less pronounced differentiation into successive portions— é III THE ALIMENTARY CANAL 183 oesophagus, stomach, intestine and its subdivisions, cloaca. In correlation with the digestive and assimilative function of the intestinal endoderm this serves during early stages as the favourite storehouse of food-yolk, and the concentration of yolk in the ‘abapical portion of the unsegmented egg is to be looked on as a foreshadowing of the fact that this portion of the egg will later become the endoderm. In the holoblastic Vertebrates the mass of heavily yolked endo- derm cells becomes, as it were, modelled into a tubular shape by the l= Fic. 104.—Illustrating the modelling of the yolk in Lchthyophis. (After Sarasins, 1889.) A and B illustrate the same stage, B representing a view from the dorsal side. The small-celled epithelial portion of the gut-wall is seen passing down the centre of Fig. B. C, D, and E represent later stages drawn from the ventral side; F (7 cm. embryo) ventro-lateral view from the right side. reciprocal activity of endoderm and splanchnic mesoderm; the rudiment so formed undergoing active growth in length and differentiation of structure while the yolk is being assimilated. In the two most archaic groups of holoblastic gnathostomes, the Crossopterygians and the Lung-fishes, a feature of special interest is the development of the spiral valve. In Lepidosiren, as is indicated by Figs. 105 and 106, this takes its origin by the solid mass of yolk-laden endoderm becoming modelled into a right-handed spiral coil—the deep incision which separates successive turns of the spiral being filled up by ingrowing mesenchyme belonging to the splanchnic mesoderm. There can be little doubt that this is a secondarily modified mode of development, but nevertheless it is probable that the spiral coiling of the endodermal rudiment is to be 1 184. EMBRYOLOGY OF THE LOWER VERTEBRATES CH. explained as a repetition of an ancestral condition in which the intestine as a whole was long and spirally coiled. ; An important feature of such a spiral coiling of the gut rudiment is that it would necessarily tend to bring about a twisting of the Fic, 105.—Dissections of young Lepidosirens of stages 32 (A), 35 (B), 36 (C), and 87 (D), from the ventral side to show the modelling of the intestine. g.b, gall-bladder ; li, liver; V, ventricle. alimentary canal just in front of the spirally coiled portion in a counter-clockwise direction as seen from behind, ¢.e. a movement in which points on the ventral side of the alimentary canal would become shifted towards the right side. As already indicated such a twisting of this region of the alimentary canal actually does take III THE ALIMENTARY CANAL place in development causing the lung rudiment to shift dorsally round the right side of the alimentary canal. In the more richly yolked Vertebrates the ventral portions of the gut-wall are more and more clogged up with yolk and this results in a greater and greater concentration of de- velopmental activity in the dorsal wall. This is clearly in- dicated by transverse sections through the developing gut of Vertebrates which though rich in yolk are still holoblastic. Such sections (Fig. 107) show the dorsal wall of the gut to consist of small active cells ar- ranged as a columnar epithelium, while the side walls and floor consist of large comparatively inert yolk-laden elements. It is only as development goes on, and as the yolk is consumed, that the epithelial small-celled character gradually spreads ventrally. In the actually meroblastic Vertebrates, the heavily yolked portions of the primitive gut- wall never undergo segmenta- tion at all, unless possibly as regards a thin superficial layer. They remain as a continuous mass of yolk, round which the epithelium gradually spreads. In this case the formation of all the important organs of the ali- mentary canal is concentrated in the dorsal portion which be- comes gradually folded off from the main mass of the yolk. This folding-off process takes place most actively in the anterior region, so as to form the tubular fore-gut, and also posteriorly, the intermediate portion re- 185 Dissection of Lepidosiren larva of stage 35. 106. Fic. h, hemisphere; ht, heart; 1, lung; li, liver; o.c, auditory capsule ; ocr, occipital rib; p.n, pronephros ; f.0, tectum opticum. 186 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. maining for a time as a longitudinal groove opening ventrally towards the yolk. As the lips of this groove gradually coalesce at each end the communication between the gut cavity and the yolk becomes gradually narrowed down to the tubular cavity of the yolk- stalk situated at first behind the liver but later becoming’ shifted forwards by differential growth. Eventually this becomes obliterated and the definitive alimentary canal becomes completely isolated from what remains of the yolk. In many Teleostean fishes this isolation takes place at a very early stage in development. The alimentary canal is, in correlation with its digestive function, necessarily a highly glandular organ. Primitively the secretory functions are carried out by unicellular glands, scattered about amongst the other epithelial cells of the endoderm, but in the Verte- brates, as in all the more complex Metazoa, special concentrations of gland cells and of secretory activity take place in localized portions of the enteric wall. Each of these specially glandular patches under- goes a great increase in its area, which causes it to bulge outwards as a simple or much subdivided Fic. 107.—Transverse section through hind and complicated pocket, forming portion of intestine of a larva of Ich- & distinct glandular appendage of te, (Aber era, 08) THe the alimentary canal Te 104, ¥ The sheath of splanchnic Liver. UE these glandular mesoderm is omitted. appendages, in the case of Verte- brates, the most ancient appears to be the liver, which is already present in Amphiozus. In this animal the liver originates in ontogeny (Hammar, 1893) as a pocket-like outgrowth of the alimentary canal wall on its ventral side and slightly posterior to the hind end of the pharynx. Apart from increase in size and relative narrowing of its base of attachment the liver in Amphioxus undergoes no further complication but retains its extraordinarily primitive pouch-like condition throughout life. In the holoblastic Craniates the liver arises similarly as a ventral projection of the alimentary canal wall. This shows the customary modifications in correlation with the presence of yolk, arising in some cases in the more primitive fashion as a hollow pocket (Lampreys, many Amphibians, Ceratodus), in others (many Amphibians, Lepido- siren and Protopterus) as a solid knob of yolk-laden cells (Fig. 105, ii). This grows rapidly in size, as it uses up its food-yolk, and becomes constricted off from the main mass of yolk by ingrowing mesenchyme, until its attachment becomes narrowed down to a slender stalk—the rudiment of the bile-duct. The pouch-like rudiment of the liver undergoes an active process of sprouting into numerous secondary pockets, each of which becomes III LIVER 187 greatly elongated and branched, and gives the gland a tubular character. This character may be retained throughout life (Lampreys) but normally the tubules undergo anastomosis so as to form a net- work of trabeculae. While this is to be regarded as the primitive mode of development of the tubules it is to be noted that they more usually in actual fact show the modification of development which we have learned to associate with the presence of yolk, being at first solid and taking their origin not by a process of outgrowth but rather by a process of modelling by ingrowing mesenchyme. In the meroblastic Vertebrates also the liver may be described as originating from a mid-ventral outpushing of the enteric wall. Variations occur in detail, in correlation with the varying relations of the hepatic portion of enteric wall to the fore-gut and yolk-sac. If this part of the gut-wall has already been folded off from the yolk-sac and incorporated in the fore-gut, then the early stages of development of the liver diverticulum pursue their normal course. If, on the other hand, it still forms part of the yolk-sac wall, the hepatic rudiment makes its appearance as a projection from this, and it may be in its first beginnings paired, its two halves separated ‘by the longitudinal slit by which the cavities of the definitive gut and the yolk-sac are still continuous. ELASMOBRANCHIL—The hepatic diverticulum at an early stage bulges out to form a conspicuous outgrowth on each side anteriorly —the rudiments of the right and left lobes of the liver. The median portion between these becomes in its anterior region converted into secretory tissue while its posterior part becomes the bile-duct, with its dilatation the gall-bladder. In Acanthias (Scammon, 1913) the first rudiment of the liver, which makes its appearance at a time when this region of the enteron is not yet floored in but opens freely into the subjacent yolk-sac, is distinctly paired. In view of the unpaired condition in Amphiozxus and the holoblastic Craniates there can be little doubt that this condition in Acanthias is a secondary modification as indicated above. Secondary pockets soon make their appearance on the wall of the secretory portion of the rudiment, and grow actively into elongated and much-branched tubules. These fuse together secondarily to form the network characteristic of the fully developed liver. This network is bathed by the blood of the vitelline veins (see Chap. VI.). : After the embryo (Acanthias) has reached a length of 25-28 mm. the walls of the tubules, or trabeculae of the network, increase greatly in thickness so that both their own cavities and the intervening blood-spaces become relatively reduced and the organ assumes the compact definitive condition. ; Whereas the tubules become throughout the greater part of their extent secretory in function the proximal portions, each common to a group of tubular branches, functionmerelyas ducts. These communt- cate with the main bile-duct formed from the posterior and median 188 EMBRYOLOGY OF THE LOWER VERTEBRATES CH. fi. Fic. 108,.—TIlustrating early development of the liver in Birds. A, 47-hour chick; B, 52-hour chick ; C, 50-hour chick (after Brouha, 1898); D, fourth-day chick; E, 7 mm. embryo of the Roseate Tern—Sterna paradisiaca—(after Hammar, 1897). 6d 1, rudiment of anterior (‘left’) bile- duct ; bd2, posterior (“right”) bile-duct; ent, cavity of fore-gut; gb, rudiment of gall-bladder ; 17. 1 and 2, anterior and posterior liver -rudiments; pan,\ dorsal rudiment of pancreas, portion of the rudiment. The gall bladder originates as a bulging of the floor of the bile-duct towards its anterior end. The formation of the pos- terior and longer section of the bile-duct, which will be extrahepatic in the adult, lags in its development behind the anterior portions of the rudiment. Such differences in the time of appearance of different parts of the hepatic apparatus—liver, gall-bladder, bile-duct—are to be looked on as mere secondary modifica- tions of development, — the primitive condition being that of a simple pocket of the gut- wall such as persists in Am- phioxus. Sauropsipa.—The hepatic apparatus here again makes its appearance as a longi- tudinally situated pocket of the morphologically ventral wall of the gut. In birds this is situated at first on the anterior wall of the yolk-stalk (Fig. 108, A). The diverti- culum grows actively into an anterior (dorsal) and a pos- terior (ventral) pocket (Fig. 108, C, . 1 and 1. 2) while the intervening portion be- comes flattened out and incor- porated in the eut-wall. There thus come to be two distinct liver-rudiments an anterior and a posterior. Of these each sprouts out at its end into irregular projections which eventually fuse and form a spongy mass, surround- ing the cavity of the ductus venosus, and having in its meshes blood-spaces which 1Ul PANCREAS 189 communicate with the just - mentioned vessel. This spongy mass, the trabeculae of which are at first solid and only secondarily develop a lumen, forms the secretory portion of the liver, while the proximal portions of the outgrowths persist as the two con- spicuous bile-ducts of the adult bird (Fig. 108, D, E, bd. 1 and bd. 2). In such birds as possess a gall-bladder this is formed by a dilatation close to the point of junction of the posterior bile-duct with the. gut-wall (Fig. 108, D, E, 96). PancrEAS.—The pancreas, though in the adult a single structure, arises typically from three distinct rudiments, each of which is at first a simple pocket-like outgrowth of the splanchnopleure. One of the rudiments (cf. Fig. 80, H) is situated dorsally a little posterior to the stomach, the other two, which appear somewhat later, are ventral and arise as outpushings of the hepatic diverticulum in the region of the bile-duct. The ventral pancreatic rudiments are commonly paired, arising one on the right and one on the left of the bile-duct. The three rudiments increase in size, secretory tubules sprout out from them and the two ventral rudiments become carried in a dorsalward direction, up the right side, by the rotation which the gut undergoes in this region (see p. 168). The right ventral rudiment comes in contact with the dorsal rudiment and fusion takes place— all three rudiments forming a single organ the three-fold origin of which is indicated by its three communications with the alimentary canal. Such may be considered the typical mode of development of the pancreas, but important variations in detail occur in the different groups. In Cyclostomes and Elasmobranchs only the dorsal pancreas is known to occur. Its development in the former group requires further investigation. In Elasmobranchs it arises as a longitudinal groove of the enteric wall dorsally and a little posterior to the open- ing of the bile-duct. It becomes constricted off from before backwards and in accordance with the rotation of the alimentary canal it becomes shifted to the left side and ends up by being ventral. In Crossopterygians the three typical rudiments appear (Fig. 80, H) but their development has not been followed in detail. Eventually the pancreatic complex extends forwards beneath the liver and com- pletely fuses with it forming a thick layer over its ventral surface in the region near the opening of the bile-duct. In Actinopterygian Ganoids also (Piper, 1902; Nicolas, 1904), the pancreatic complex derived from the original three rudiments becomes fused with the substance of the liver, only its posterior dorsal portion remaining extrahepatic. The main duct of the pancreas is the persistent stalk of the right ventral rudiment which opens into the gall-bladder formed by the dilated terminal part of the bile-duct. Of the two other pancreatic ducts the left ventral apparently atrophies entirely, while the dorsal is said in the case of Amia to disappear but in the Sterlet (Acipenser ruthenus) to persist. 190 EMBRYOLOGY OF THE LOWER VERTEBRATES _ cH. In Teleosts the early development agrees closely with that of the ganoids, only a doubt exists whether the definitive pancreatic duct (Duct of Wirsung) may not be formed by a fusion of the two ventral ducts rather than by the persistent right duct alone. During later stages great differences arise between different members of the group. In : some (Stlwrus, Hsox) the complex forms a single compact gland, in others (Scomber, Cyprinus) it becomes divided into a number of independent lobes, in others, including the majority of the more familiar Teleosts, it becomes greatly branched and is diffused in the substance of the dorsal mesentery while in still others (Labridae, Syngnathus) the condition resembles that of the ganoids a large part of the organ being intrahepatic (Laguesse, 1894). ~|-—-— , i mm, ‘ r i a mit Fic. 109.—Dorsal view showing 1udiments of dorsal pancreas and lung in larvae of Protopterus (stages 32 and 34). 1, lung; op, operculum ; pu.d, dorsal pancreas ; p.b, postbranchial body ; p.f, pectoral limb ; v.e, visceral cleft rudiment. In Lung-fishes the three typical rudiments make their appearance. In Protopterus the dorsal rudiment (see Fig. 109, A, pa.d) is a solid outgrowth (hollow in Lepidosiren) from the gut-wall, usually rounded in form but occasionally elongated in an antero-posterior direction as in the specimen figured (Hig. 109, A). The attachment to the gut becomes rapidly constricted to a narrow stalk and a cavity develops in the interior of the rudiment. The ventral rudiments appear a little later, as solid projections one on each side of the attachment of the bile-duct to the gut. The two ventral rudiments, as they increase in size, meet and fuse dorsal to the bile-duct, and later on the dorsal surface of the right ventral rudiment comes in contact and fuses with the ventral surface of the dorsal rudiment. The stalks of the three rudiments remain as three ducts, the two ventral opening just posterior (original right rudiment) and anterior Ill PANCREAS 191 (original left) respectively to the opening of the bile-duct, while the dorsal opening is situated at the extremity of the spout-like pyloric . valve. The general course of development in Zepidosiren is similar and in both it is characteristic that the pancreas never bulges beyond the mesodermal coating of the splanchnopleure. It remains embedded throughout life in the gut-wall and is consequently not noticeable in an ordinary dissection. In Ceratodus (Neumayr, 1904) the development of the pancreas is similar though here the left ventral rudiment, which in Protopterus is smaller in size than the right, remains rudimentary. The Amphibia are of special interest from the fact that it was a member of this group (Bombinator) in which Goette (1875) first observed the origin of the pancreas from three separate rudiments. Goeppert (1891) was able to extend the observation to various other Amphibians, both Urodele and Anuran, and to show that in Urodeles the dorsal rudiment retains its duct, opening just behind the pylorus, while in the Anura this duct disappears. In both cases the ducts of the two ventral rudiments undergo fusion to form a duct of Wirsung which opens into the bile-duct. In Reptiles (Lacerta—Brachet, 1896) the right ventral and the dorsal rudiments fuse to form the definitive pancreas, the left ventral atrophying (cf. Lung-fishes). According to Brachet the duct of the dorsal rudiment does not disappear but fuses with that of the right ventral to form the definitive pancreatic duct. Birds show three rudiments which undergo fusion into a complex in the normal fashion, all three ducts remaining functional and conspicuous in the adult. Suppression of the left ventral rudiment occurs as an occasional variation. The observed facts of development of the Pancreas clearly justify the conclusion that this organ of the modern Vertebrate has arisen in the course of evolution from three originally separate diverticula of the glandular enteric wall—a pair arising from the hepatic pouch and the third from the dorsal wall. The precise localization of the rudiments at comparatively distant points of the enteric wall point to the probability that the nature of the secretion was originally different in the case of the ventral pancreas from that of the dorsal. Pytoric Cazca.—The caeca which are present in the pyloric region in many actinopterygian fishes arise as simple outgrowths of the gut-wall. The interesting suggestion has been made (Taylor, 1913) that the simple circle of these caeca, which is apparently their most primitive arrangement, corresponds morphologically with the curious valve found in various fishes (Ama, Lung-fishes, Symbranchus, Anguilla, etc.) in which the pyloric end of the stomach is prolonged back into a kind of spout which is ensheathed by the anterior end of the intestine. The circular prolongation forward of the intestinal cavity round the gastric spout might clearly give rise to a circle of pyloric caeca simply by subdivision into a number of separate 192 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. portions each of which continued to open into the gut cavity at its hinder end. Recta, Gianp.—This organ, which occurs in Elasmobranchs, arises as a simple pocket-like outgrowth of the gut-wall. The super- ficially similar caecum of Lung-fishes will be dealt with in con- nexion with the renal organs. CLoaca.—-In the more archaic Vertebrates the ducts of the excretory organs open into the terminal part of the intestine which is thus a cloaca. It is believed by many that the excretory ducts originally opened at the hind end of the trunk independently of the alimentary canal and it is natural to suppose that the openings of the ducts have become gradually shifted first into close proximity to the anus and finally on to the lining wall of the ali- mentary canal. This again suggests that the cloaca may really be a proctodaeum—that the skin has been involuted to form its lining and that with this involution the renal openings have also been carried inwards. Unfortunately the facts of ontogenetic development do not so far as can be seen at present fit this simple and attractive hypothesis. The cloaca is, except for a small portion close to its opening, of purely endodermal origin—the renal ducts open on what is part of the primary enteric wall. A suggested explanation of this fact differing from that mentioned above will be found in the chapter dealing with the renal organs. A cloaca seems always to be developed though in some cases (e.g. Teleostean fishes) it flattens out and disappears later so that the renal organs and the gut come to have independent external openings. The bursa Fabricii, a conspicuous glandular appendage of the dorsal wall of the cloaca in young birds, has usually been regarded as proctodaeal in its origin but it is now known to arise in ontogeny from vacuolar spaces in a solid projection from the cloacal rudiment, dorsal to the stalk of the allantois (Wenckebach, 1888) and would therefore appear to belong to the mesenteron rather than to the proctodaeum. The anal opening of the Vertebrate, as may have been gathered from Chap. Il., is to be regarded as representing morphologically a portion of the gastrular mouth or protostoma. In a large number of Vertebrates however the opening arises in ontogeny not in this way but rather as a secondary perforation, although even in such cases the perforation arises in the line of the closed protostoma. TEMPORARY OCCLUSION OF THE ALIMENTARY CAaNAL.—The ali- mentary canal is, in correlation with its function, a hollow tube. In a large number of Vertebrates, however, there are more or less extended periods of development during which the cavity is com- pletely absent, either throughout the length of the canal or in certain portions. In its simplest condition this occurs as a special case of the HL THE ALIMENTARY CANAL 193 temporary absence of lumen so frequently found in the idevelopment of eventually hollow organs from a richly yolk-laden rudiment. An idea of how it has come about will be got from an inspection of the various stages of the development of the alimentary canal of Polypterus as shown in Fig. 80 on p. 146. During early stages the archenteric cavity is seen to be widely patent throughout, except that there is no mouth opening. During the later stages of develop- ment, immediately prior to the canal becoming functional, its walls throughout the region between the fore-gut and the cloaca become closely apposed, so as almost entirely to obliterate the cavity. Later on the walls recede from one another and the lumen becomes again patent. It would obviously be merely a slight accentuation of this modification of development for the cavity to be completely obliter- ated for a time. A still further modification would be brought about by the omission altogether of the original hollow stage from the ontogenetic record. This actually occurs in the case of the fore-gut in those Vertebrates in which this region of the enteric rudiment is yolk-laden: where, on the other hand, the yolk is practically completely concentrated in the mid-gut region as in meroblastic Vertebrates it does not occur as a rule. The most striking temporary occlusions of the alimentary canal during development have to do with its terminal apertures. Thus there is not a single existing Vertebrate, so far as is known, in which the mouth opening persists from the gastrular stage, or in which even any connexion has so far been traced between the definitive mouth opening and the protostoma. In every case, even in Amphioxus, the mouth opening develops comparatively late as a secondary perforation. This modification of development is in the present writer's opinion to be attributed to the entire dependence of members of the Vertebrate phylum upon food-yolk during early stages of their development, the need for a functional mouth having thus disappeared. The anteroposterior extent of this occlusion of the alimentary canal in the region of the oral opening differs in different sub- divisions of the phylum. It may include a large part of the stomodaeal as well as the endodermal portion of the buccal cavity as in the Lung-fishes (p. 148) but more usually it is confined to the boundary between the two, i.e. to the site of the original mouth opening the closely apposed ectoderm and endoderm being at this level continuous across the site of the future opening as the velar membrane (p. 145). The secondary perforation by which the alimentary canal comes to communicate with the exterior at its front end is in the case of some larval Vertebrates (e.g. Lepi- dosiren) closely correlated with the commencement of pharyngeal respiration but where the development is embryonic it commonly still takes place long before the existence of any obvious functional need (¢.g. Chick, fourth day). At its hinder end the archenteron is, VOL. II ) 194 EMBRYOLOGY OF THE LOWER VERTEBRATES CH. as has been shown in Chap. I., widely open to the exterior in all the lower Vertebrates during early stages and in various cases this opening can be traced either into direct continuity, or into less direct but still clear relationship, with the anal opening. The explanation of this lesser degree of modification of the development of the anal opening as compared with the mouth may probably be associated with the less accentuated delay in the functional need for this opening. At stages long before ingestion or inspiration takes place by the mouth, the formation of waste products during the digestion of the yolk necessitates an outlet from the enteric canal at its hinder end. Where obliteration does take place during still earlier stages this is probably correlated with the fact that the need of the opening is still non-existent. It is of interest to notice that obliteration of the anal opening which is of a directly adaptive significance may take place at a later stage. Thus in Lepidosiren during about the first two weeks of larval life, when large numbers of practically motionless larvae are lying crowded together in the nest, the anal opening, which had been continuously patent in earlier stages, is closed, so as to prevent the poisonous excretory products from finding their way out. So also in the case of the Elasmobranch embryo enclosed within its egg-shell. In the Amniota the perforation of the anus is delayed to a relatively late period doubtless for a similar reason. It is characteristic of the phylum Vertebrata that the anal opening no longer occupies its primitive position at the extreme end of the body but has become shifted forwards along the ventral side. This shifting has probably come about with increased specialization for swimming by lateral flexure of the body, the withdrawal of the alimentary canal with its surrounding splanchnocoelic cavity from the hinder portion of the body, leaving the space they occupied free for increased development of the lateral muscles. This shifting forwards of the anus, leading to the differentiation of a distinct postanal or tail region, has occurred in all Vertebrates, least markedly in the more archaic groups. It reaches its maximum in some members of that group of Vertebrates which is above all others highly specialized for active swimming, the Teleostei, in some families of which the anus has actually assumed a jugular position. During the actual ontogeny of the Vertebrate the process by which the anus comes to occupy a position more or less distant from the tip of the tail region is somewhat modified from that which probably occurred during phyletic evolution. We do not find that the anus remains at the tip of the tail during the growth in length and that it then gradually shifts forwards along the ventral side. What happens is that the opening at an early stage assumes a ventral position and that the tail region proceeds to sprout out dorsal to it. The process will be understood from an inspection of Fig. 80 (p. 146). In B the anus is at the hinder end, in C it has 1 Ill THE ALIMENTARY CANAL 195 assumed a ventral position being overhung by the bulging tail rudiment, in D, E, F, G the tail rudiment is seen to be extending actively past the position of the anus, the specially actively growing tissues being indicated by the darker shading. In Fig. 80, G, a feature is well shown which occurs in the embryos of most Vertebrates —the postanal gut (pag). It was shown in Chap. I. how a connexion—the neurenteric canal—existed in some Vertebrates between the cavity of the enteron and that of the neural rudiment at their posterior ends. Here, in the postanal gut, we have such a connexion still persisting in a drawn-out form though, as in the present case, it may be a solid strand of yolky cells and not a hollow tube. The postanal gut is a purely transitory structure which at a relatively early period of development dis- integrates completely. In endeavouring to determine the morphological significance of the postanal gut it is necessary to bear in mind that the Vertebrate in early stages develops from before backwards and that the growth in length by the addition of new segments takes place at its hinder end where there is a mass of actively growing embryonic tissue forming a kind of “growing point.” The tissue of this, although to the eye quite undifferentiated, contains the elements which form all the various tissues such as nerve cord, notochord, myotomes, alimentary canal, etc. As growth goes on these gradually become differentiated out, the differentiation always proceeding from before backwards. If we now look at such a young Vertebrate as that shown in Fig. 80, G, we see the typical Vertebrate structure, includ- ing alimentary canal (pa.g) extending right back practically to the tip of the tail: it is only at the extreme tip that the various organs merge together into undifferentiated embryonic tissue. The only striking peculiarity is that the communication of the alimentary canal with the exterior, the anus, is not in the midst of the growing tissue of the tip, as it would be, for example, in a young Chaetopod worm, but well forwards on the ventral side. This peculiarity, in the writer’s opinion, finds its explanation in the development from before backwards already alluded to. The appearance of the anus at a point relatively far forwards means that it and the organs related to it such as the excretory ducts complete their development at an earlier period of time. As it is of functional importance that the organs in question should do so, in contradistinction to the purely motor arrangements farther back, we see a physiological reason why evolution should have brought about a development of the anal opening in its anterior position from the beginning, and the elimination of those stages in which it was situated farther back. As regards the phyletic evolution of this part of the enteron, we may sum up probabilities as follows: that the alimentary canal with its surrounding splanchnocoele originally extended to the hind end of the body: that the anal opening came to be shifted on to the 196 EMBRYOLOGY OF THE LOWER VERTEBRATES _ Il ventral wall of the canal: that it then underwent a gradual shifting forwards along the ventral side: that as it did so the now postanal portion with its splanchnocoele gradually atrophied the position they occupied becoming filled mainly with muscle. LITERATURE Assheton. Quart. Journ. Micr. Sci., xxxviii, 1896. Assheton. The Work of J. S. Budgett. Cambridge, 1907. Balfour. Comparative Embryology, ii, 1881. Brachet. Journ. de ]’Anat. et de la Physiologie, xxxii, 1896. Brauer. Zool. Jahrb. (Anat.), xii, 1899. Brouha. Journ. de l’Anat. et de la Phys., xxxiv, 1898. Bryce. Journ. Anat. and Phys., xl, 1906. Budgett. Trans. Zool. Soc. Lond., xvi, 1901. Dean, Bashford. Zool. Jahrb. (Syst.), ix, 1896. Driner. Zool. Jahrb. (Anat.), xv, 1901. Egert. Zool. Anzeiger, xlii, 1913. Goeppert. Morph. Jahrb., xvii, 1891. Goeppert. Morph. Jalirb., xx, 1893. Goette. Entwicklungsgeschichte der Unke. Leipzig, 1875. Goette. Zeitschr. wiss. Zool., lxix, 1901. Greil. Semons Forschungsreisen in Australien, i. Jena, 1913. Hammar. Arch. f. Anat. u. Entwicklungsgesch., 1893. Hammar. Anat. Anzeiger, xiii, 1897. Juillet. Arch. zoo]. expér. [5], ix, 1912. Kallius. Anat. Hefte (Arb.), xvi, 1901. Kerr, Graham. Proc. Roy. Phys. Soc. Edin., xvi, 1906. Kerr, Graham. The Work of J.S. Budgett. Cambridge, 1907. Kerr, Graham. Quart. Journ. Micr. Sci., liv, 1910. Laguesse. Journ. de l’Anat. et de la Phys., xxx, 1894. Lankester, E. Ray. Quart. Journ. Mier. Sci., xvi, 1876. Marcus. Arch. f. mikr. Anat., 1xxi, 1908. Moroff. Arch. f. mikr. Anat., lx, 1902. Moser. Arch. f. mikr. Anat., lx, 1902. Moser. Arch. f. mikr, Anat., lxili, 1904. Miller, W. Jenaische Zeitschrift, vi, 1871. Neumayr. Semons Forschungsreisen in Australien, i, 1904. Nicolas. Arch. Biol., xx, 1904. Phelps. Science, N.S. ix, 1899. Piper. Arch. f. Anat. und Entwicklungsgesch., Suppl. Bd., 1902. Piper. Verh. Anat. Ges., Halle, 1902. Reighard and Phelps. Journ. Morph., xix, 1908. Rowntree. Trans. Linn. Soc, Lond., (2) ix. 1903. Sarasin, P. and F. Ergebnisse naturwiss. Forschungen auf Ceylon, ii, 3. Wiesbaden, 1889. Sawadsky. Anat. Anzeiger, xl], 1911. Scammon. Amer. Journ. Anat., xiv, 1913. Sedgwick. (Quart. Journ. Micr. Sci., xxxiii, 1892. Smith. Journ. Morph., xxiii, 1912. Taylor. Quart. Journ. Micr. Sci., lix, 1913. Voeltzkow. Abh. Senck. Ges., xxvi, 1899. Wenckebach. Ontwikkeling en de bouw der bursa Fabricii. Proefschrift. Leiden, 1888. Wijhe, van. Verhand. Konink. Akad. Wetensch. Amsterdam, Tweede Sectie, * Deel xviii, 1914. Wilson, Gregg. Proc. Roy. Phys. Soc. Edin., xiv, 1901. CHAPTER IV THE COELOMIC ORGANS Iytropuction.—The mesoderm of Amphiozus consists in an early stage, as already indicated (p. 57), of a row of closed sacs arranged serially one behind the other upon each side of the body. At this time the coelome of Amphiozxus is in the extremely archaic condition of a series of metamerically arranged paired compartments —a condition resembling that of the less modified forms of Annelids. The coelomic sacs gradually spread in a ventral direction until they meet. For a time after this happens the sacs of opposite sides of the body remain separated by a longitudinal partition the ventral mesentery. Similarly the apposed posterior and anterior walls of neighbouring sacs belonging to the same side of the body, form thin membranous septa like those of Annelids. A highly characteristic difference from the Annelid arrangement begins to show itself a little before hatching in the ventral portion of the body, in as much as the transverse septa break down and disappear thus converting what was hitherto a chambered coelome in this region into a continuous space. There is no obvious reason why this loss of segmentation of the ventral portion of the meso- derm has come about in evolution. A general characteristic, however, of the phylum Vertebrata is the loading up of the ventral part of the endoderm with yolk and it may well have been that the loss of the mesoderm septa ventrally arose in correlation with the presence of a greater amount of yolk in the ancestral condition than exists in the present-day Amphioxus. A further striking difference between the Vertebrate and the Annelid is expressed in the extent to which the coelomic wall gives rise to muscular tissue. In the Annelid practically the whole extent both of the somatic layer lining the body-wall and the splanchnic layer covering the gut gives rise to muscular tissue. In Amphioxus however, and the same holds for Vertebrates in general, the ventral portion of the somatic mesoderm, the portion which loses its segmental character—loses also its capacity for producing muscle. : On the other hand the dorsal portion of the mesoderm, which retains its segmentation, retains also, and to an accentuated degree, its muscle-forming capacity. It separates off from the ventral or 197 198 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. splanchnocoelic portion of the mesoderm in the form of a series of segmentally arranged sacs—the myotomes—and the wall of these gives rise to almost the whole of the muscular system. The myotomes are at first, from their mode of origin, restricted to the dorsal side of the body, but as development goes on active growth of their ventral portions takes place and they extend downwards, overlapping and covering in the splanchnocoelic mesoderm right down to the mid- ventral line. In this way a muscular body-wall is provided for the ventral region of the body in which the original muscle-producing capacity of the somatic mesoderm had disappeared. The evolutionary origin of this curious secondary muscularization of the ventral body-wall of the Vertebrate is unexplained but the suggestion may be hazarded that it was associated with the loss of segmentation of the ventral splanchnocoelic mesoderm, the primitive mode of movement of the Vertebrate—by waves of lateral flexure— being only able to utilize longitudinal muscles divided into segments. We may take it that the splanchnocoelic muscular layer, as it lost its segmentation, would become less efficient for purposes of move- ment, and that, correlated with this, its territory would then tend to be encroached on by the still segmented, and therefore more efficient, dorsal portion of the muscular layer until eventually it came to be replaced completely by it. As a result of the developmental processes which have just been indicated the mesoderm of Amphioxus, which for a time consisted of a metameric series of paired sacs, is now represented by (1) the segmentally arranged myotomes and (2) the unsegmented splanchno- coelic lining. To these a third element becomes added in the form of a pocket-like outgrowth from the myotome wall close to its lower end (Fig. 144, A, sel, p. 285). This grows first towards the mesial plane and then dorsally, insinuating itself into the space between myotome on the one hand and notochord and spinal cord on the other, until it occupies practically the whole of that space right up to . mid-dorsal line. This pocket-like diverticulum is the sclerotome p. 286). In the typical Vertebrate a fourth derivative of the mesoderm segment is of importance: it takes the form of a connexion which persists for some time between the myotome and the splanchno- coelic mesoderm as a narrow stalk or isthmus. This—the proto- vertebral stalk or nephrotome (Riickert, 1888) with its cavity the nephrocoele—is of great importance from its relation to the nephridial organs but its existence has not up to the present been demonstrated in Amphiowus. We will now proceed to trace out the subsequent fate of these - various derivatives of the primitive mesoderm segments. CoELOMIC CAVITIES.—The only portions of the coelomic cavities which remain patent are the nephrocoeles (which will be dealt with later on) and the splanchnocoele or peritoneal cavity. It may be taken as probable that the body-cavity of the Iv COELOME 199 ancestral Vertebrate was divided up into segmentally arranged compartments by transverse septa, and into a right and left half by a sagittally placed partition supporting the alimentary canal and forming the dorsal and ventral mesentery; in other words that the general arrangement was like that of a primitive Annelid worm. This seems to be indicated by the mode of development of the mesoderm in Amphioxus. In Vertebrates above Amphioxus the segmented condition of the splanchnocoele has disappeared even from development! The sagittally placed mesentery on the other hand still appears in ontogeny in the form of the partition remaining between the edges of the lateral mesoderm as they approach one another on the ventral and on the dorsal sides of the alimentary canal respectively. In correlation with the great increase in length, and consequent coiling, of the alimentary canal of the Vertebrates—a condition which probably existed even in the ancestors of those gnathostomes in which the alimentary canal is now short (p. 184)—the ventral mesentery disappears at an early stage of development throughout that portion of its extent which lies on the tailward side of the liver. The dorsal mesentery on the other hand persists throughout life, serving as a bridge to carry the complicated connexions of the gut wall with the vascular and nervous systems, although perforations may appear in it, more or less extensive in different groups of Vertebrates. The complicated foldings and frillings which the dorsal mesentery undergoes, owing to its enteric edge having to keep pace with the increase in length of the gut, are of interest mainly to specialists in the anatomy of particular groups and need not be dealt with here. In the fishes, in which the lung performs an important hydro- static function, that organ grows back in the substance of the dorsal mesentery, and in accordance with its tendency to assume a more and more dorsal position, the portion of mesentery lying above it may become incorporated in the dorsal wall of the splanchnocoele, the result being that the lung in the adult now lies entirely dorsal to and beyond the limits of the body-cavity (Dipnoi,? Actinopterygii). Apart from its primary segmentation, the splanchnocoele shows a tendency for special portions to become secondarily separated off from the main cavity. The most important case of this occurs at 1 While it has to be granted that the splanchnocoele of the Vertebrates represents the ventral portion of the coelome which has lost its segmentation, care must be taken not to assume that this loss of segmentation has necessarily extended dorsal- wards to precisely the same level in all Vertebrates. Like other anatomical boundaries the dorsal limit of the splanchnocoele is doubtless fluctuating and vague. It is therefore wise not to attach too great importance to the exact position of the first rudiment of an organ which develops in one case on the dorsal and in another on the ventral side of the boundary between segmented and unsegmented mesoderm such as for example the gonad (p. 270). ; 2 Cf. Graham Kerr, 1910. ‘ 200 EMBRYOLOGY OF THE LOWER VERTEBRATES cn. the hinder end of the heart where there exists on each side a broad bridge by which the duct of Cuvier passes from the somatopleure to the sinus venosus. This bridge becomes extended headwards and dorsally on each side of the oesophagus until it meets the dorsal wall of the splanchnocoele thus forming with the oesophagus a floor separating the anterior portion of the splanchnocoele into two cavities, one dorsal and one ventral, each opening posteriorly into the main splanchnocoele. Of these two cavities the dorsal becomes completely obliterated by fusion of its splanchnic (oesophageal) and somatic walls from before tailwards. The ventral one on the other hand roofed in by the oesophagus persists as the pericardiac cavity. The communication of this posteriorly with the main splanchno- coele is obstructed in the middle by the flattened headward surface of the liver which is embedded in the distended ventral mesentery, while laterally the communication is for a time open. As develop- ment goes on however the opening on each side becomes obliterated by an ingrowth from the somatopleure which spreads downwards from the bridge of tissue containing the duct of Cuvier and the free edge of which meets and fuses with the mesoderm covering the headward surface of the liver. The pericardiac cavity comes in this way to be bounded posteriorly by a complete wall of tissue a large part of which consists simply of the mesodermal sheath of the liver. As the body of the embryo increases in diameter this wall of tissue keeps pace with it as does also the liver. The latter organ however in subsequent growth of its anterior or headward surface does not keep growing in continuity with the substance of the septum but becomes separated from it by a deep cleft, the region of continuity between liver and septum becoming thus restricted to a small area dorsal and close to the mesial plane. Similarly the region of continuity between the headward face of the septum and the wall of the sinus venosus which is at first of relatively considerable dorsiventral extent becomes reduced to a narrow bridge of tissue. In the Elasmobranchs the isolation of pericardiac cavity from the main splanchnocoele is only temporary. A median pocket-like extension of the pericardiac cavity spreads tailwards immediately dorsal to the sinus venosus in the substance of the mesodermal sheath covering the ventral surface of the oesophagus. This develops on each side a communication with the main cavity of the splanchno- coele which persists throughout life as a crescentic slit on the ventral surface of the oesophagus (Hochstetter, 1900). This secondary com- munication between pericardiac coelome and splanchnocoele is known as the pericardioperitoneal canal. In Myxinoids, throughout life, and in Petromyzon, during the larval period, the rudiment of the wall separating pericardiac from splanchnocoelic cavity remains in the form of a simple bridge enclosing the duct of Cuvier so that the two cavities are in wide communication with one another. Iv In the Am- phibia and Aimni- ota the pericardiac cavity becomes telescoped — back into the general peritoneal cavity, its hinder wall be- coming extended so as to form a thin membranous bag enclosing the heart and separat- ing it from the other viscera. Apart from the walling off of the pericardiac from the main periton- eal cavity there is found in the case of the Amniota a well-marked tend- ency for the latter cavity to undergo further subdivi- sion, special por- tions becoming more or less com- pletely walled in by secondary fu- sions taking place between apposed portions of the peritoneal lining. Fic. 110.— Differentia- tion of the myotome as seen in transverse sections of Lepidosiren larvae. A, stage 30; B, stage 31+; C, stage 32; D, stage 35-; E, dividing myoblasts of inner wall from stage 36. mb’, myoblasts of inner wall; mb”, myoblasts of outer wall; mf, contractile fibrils; vac, vacuole. The contractile fibrils cut across are shown as distinct black dots. COELOME 201 202 EMBRYOLOGY OF THE LOWER VERTEBRATES — cH. For example in Birds! the mesodermal coating of the lung upon its ventral side becomes continuous (1) with that lining the body-wall laterally so as to enclose the portion of splanchnocoele dorsal to the lung as a pleural cavity, and (2) with that covering the surface of the liver, forming a ventral pulmonary ligament which serves to wall in a pulmo-hepatic recess lying between it and the mesen- tery. A third connexion, the origin of which is associated with the development of the abdominal air-sacs, forms the thin, post-hepatic septum which stretches from the ventral surface of the lungs obliquely downwards and backwards to the ventral body-wall. Amongst Reptiles somewhat similar arrangements exist, differing in detail in the different groups. Tue Myoromes.—The developmental changes by which, in a gnathostomatous Vertebrate, the myotomes become converted into masses of muscle-fibres are excellently shown by Lepidosiren in which the cellular elements are particularly large and distinct. In this animal the myotome is at first solid, but later on develops a small cavity or myocoele by the breaking down of its central cells. This myocoele soon becomes obliterated by its inner and outer walls coming together. The cells of the inner wall assume a more regular shape, taking the form of large parallelepipedal cells (Fig. 110, A, mb’), flattened dorsiventrally and stretching in an anteroposterior direction throughout the whole length of the myotome. The nuclei of these large cells—myoblasts or myoepithelial cells— divide, mitotically, so that they assume a syncytial character. Their protoplasm develops a longitudinally fibrillated appearance and presently distinct cross-striped contractile fibrils (m/f) make their appearance in the protoplasm—each fibril running through the whole length of the myoblast or in other words from end to end of the myotome. The contractile fibrils, which as seen in a trans- verse section are arranged in a > -shaped pattern (Fig. 110, A, m/f), become more and more numerous and soon fill up the inner two- thirds of the myoblast almost entirely, there remaining only a relatively small amount of perifibrillar protoplasm between them (Figs. 110, B, and 111). The outer end of the myoblast does not for some time develop any contractile fibrils but there appear in its protoplasm large vacuoles (vac) which form a broad clear band in horizontal sections— of much use as a landmark ‘to indicate the outer limit of the inner wall of the inyotome. The cells of the outer wall of the myotome take the form of elongated cylinders stretching throughout the length of the myotome and in their protoplasm longitudinal fibrils make their appearance as in the case of the -inner wall myoblasts (Figs. 110, GC; 111, B). The longitudinal fibrils become fused at their ends with connective-tissue septa formed by mesen- chyme cells which wander in between consecutive myotomes. 1 For a well-illustrated account of the complicated arrangements in detail see Poole (1909). Iv ' MYOTOMES 203 Some such mesenchyme cells also penetrate into the substance of the myotome and settle down there to form connective tissue. The cylindrical myoblasts of the outer wall undergo active multiplication (Fig. 110, C) so that it comes to be greatly thickened, composed of many layers of muscle-cylinders—those towards the outer surface going on dividing actively while those further in towards the mesial Fic, 111.—Differentiation of the myotome as seen in horizontal sections of Lepidosiren larvae, A, stage 31; B, stage 31+. mb’, myoblast of inner wall; mb”, myoblasts of outer wall ; mf, contractile fibrils ; vac, vacuoles ; y, yolk. plane increase much in size as they develop more and more fibrils in their interior. As the outer wall of the myotome continues to increase in thickness the myoblasts of the inner wall become relatively more and more insignificant. Eventually they divide up into muscle- cylinders like those of the outer wall so that it is no longer possible to distinguish the inner wall portion of the myotome from the outer wall part. The muscle-cylinders become the muscle-fibres of the adult, the undifferentiated protoplasm between the fibrils persisting as the sarcoplasm the superficial layer of which may be somewhat condensed to form the sarcolemma. 204 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. A point to be noticed, of much morphological interest, is that the inner wall myoblasts of Lepidosiren are for a time (Fig. 110, A) in the form of typical myoepithelial cells such as are familiar in some of the lowest invertebrates. They are, as indicated in Chap. II, in continuity with the central nervous system by a protoplasmic tail-like extension of the cell-body closely resembling that which occurs in Nematode worms (Fig. 112). The peripheral portion of this remains as a mass of granular protoplasm on the surface of the muscle-fibre—the motor end-plate. The latter is therefore to be regarded as a portion of the muscle-cell which retains its proto- plasmic condition rather than as a portion of the nerve-fibre. The mode of conversion of the embryonic myotome into the muscle-segment has been described as it occurs in Lepidosiren because of the two special safeguards against error which exist in that animal, (1) the large size of the histological units and (2) the fact that the boundary between outer and inner walls of the myotomes is marked by a clear and unmistakable land- mark in the form of the vacuolar mm, zone constituted by the outer portions of the inner wall myo- blasts. It now remains to indicate shortly the more important differ- Fic. 112.—Diagram of a motor ganglion- aan an detail which are to be cell in the spinal cord continuous found in descriptions of the process through the substance of a nerve-fibre aS observed in others of the lower with a muscle-cell in the myotome. Vertebrates. c.f, contractile fibrils in myoepithelial cell; The chief of these concerns Oe ee the fate of the outer wall of ‘the embryonic myotome. In Lepido- siren as has been stated the outer wall gives rise to muscle. In the case of Elasmobranchs and Ganoids, Balfour stated explicitly that the outer wall of the myotome similarly takes part in the development of muscle. Many authorities (Hertwig, Rabl, Maurer), however, deny that this is the case: according to them the outer wall plays no part in muscle-formation: it simply breaks up into amoeboid cells which contribute to the dermal mesenchyme. Hence ‘these investigators term the outer wall of the myotome the “Cutis-layer.” In the case of the Sturgeon, Maurer corroborates - Balfour’s statement that the myotome is composed of two layers of muscle-elements but according to him the outer layer is simply budded off from the inner and does not represent the original outer wall of the myotome as Balfour supposed. In the Amniota the myotome in early stages is almost square as seen in a transverse section practically the whole of the wall Iv MYOTOMES 205 next the endoderm representing the sclerotome. Cells proliferat- ing from this invade the myocoele and completely fill it up. It is only in later stages that the myotome becomes extended into the normal plate-like form by active growth at its inner (dorsal) and outer edges. Of the two walls of this stage the inner admittedly becomes converted into muscle-cylinders. The outer becomes loosened out into a mass of irregularly shaped cells and these are commonly believed to give rise to dermis. In view of what happens in Lepidosiren, where accuracy of observation is so much more easily attained to, it seems advisable not to accept this as absolutely certain. At the same time it may be allowed that there is no a priori difficulty in the way of admitting that portions of myotome which in one type of Vertebrate give rise to muscle, may in another have ceased to do so, for, as already indicated, a quite similar process of concentration of muscle-development in a localized portion of somatic mesoderm is a fundamental characteristic of the whole Vertebrate phylum. The series of paired myotomes, each composed of a mass of longitudinal muscle-fibres traversing it from end to end, forms the material out of which is formed the, often extremely complicated, system of voluntary muscles of the adult Vertebrate. The various myotomes as they increase in size become divided up into it may be numerous pieces and these are pushed hither and thither by processes of differential growth until the arrangement of the numerous adult muscles contrasts greatly with the simple longitudinal arrangement of the original myotomes. During the various displacements which it undergoes the individual muscle or fragment of myotome remains in organic connexion with its nerve-centre by means of its motor nerve and the course of these nerves in the adult frequently gives an important clue to the developmental migrations of the particular muscles. No attempt will be made here to follow out the evolution of the complicated muscular arrangements of the adult beyond a short sketch of the method in which the muscles of the fins or limbs originate. The median fin is simply the extension of the body in the median plane and we should therefore naturally expect it to be muscularized by prolongations of the myotomes growing into it. The actual process is clearly illustrated in Fig. 113. In A a muscle-bud is seen to be projecting from the end of each myotome where a median fin is developing—the upper group of buds belonging to the dorsal fin, the lower to the anal. The buds diminish in size towards each end of the series and in the case of the dorsal fin, towards its _anterior end, there are a considerable number of abortive buds which never come to anything. The muscle-buds grow into the fin fold and then become cut off from the main part of the myotome to form the muscles of the fin as is shown in B. 206 EMBRYOLOGY OF THE LOWER VERTEBRATES _ cu. The paired fins or limbs become muscularized by very similar, segmentally arranged, buds and it is necessary from the outset to bear in mind that this similarity need have no deeper signifi- cance than that the paired fins also necessarily obtain their muscular- ization from the segmentally arranged myotomes. The process as it occurs in the pelvic fin of the shark Spinaa is illustrated by Fig. 114. In the 20 mm. embryo (A) the fin rudiment is seen as a longitudinal ridge and a series of myotomes in the neighbour- hood of this ridge are seen each to be forming at its lower edge two projecting muscle - buds. These sprout out into the limb rudiment, assume an elongated form (B) and then become separated off from the myotome (C). Each bud now splits into two layers a dorsal and a ventral and each of these undergoes histological differentiation and be- : Seavone of the aN Se myotome, a dorsal and a Fic, 1138.—Muscularization of median fin in Lepidosteus. y tral fi He ot ih (After Schmalhausen, 1912.) ventral trom each of the two original buds. Such is the process in its main outlines. The existence of a disturbing complication of this simple scheme is indicated by the adult arrangements,in as much as it can be shown that a single motor spinal nerve (7.e. the nerve belonging to a single myo- tome) is related to more than the four radial muscles to which alone we should expect it to be related were the account which has just been given complete. This discrepancy is brought out particularly clearly by physiological experiments. Careful stimulation of a single spinal nerve very commonly causes three consecutive (dorsal or ventral) radial muscles to contract instead of only two, and in some cases apparently a still greater number. This seems clearly to indicate that the end- organs, in other words the muscle-fibres, belonging to a particular motor nerve or myotome are in the adult not strictly confined within the limits of the two pairs of radial muscles corresponding to that motor nerve or myotome. To those who believe in the organic continuity of muscle-cell and nerve-fibre from an extremely early stage of development the idea obviously suggests itself that a shifting of some of the constituents from one muscle-bud into its neighbours takes place A, 13mm.; B, 21mm. The muscle-buds, and, in the lower figure, the nerves connected with them are shown in black. Iv MUSCLE-BUDS 207 during development. According to Mollier (1893) and Braus (1899) such a process actually occurs. Broad anastomoses or bridges make their appearance connecting the various radial muscle rudiments NNECy Fic. 114.—Illustrating the muscularization of the pelvic fin in Spinax. (After Braus, 1899.) A, 20 mm. (70 mesoderm segts.); B, 25 mm. (74 m.s.); C. 26mm.; D, 32mm. The myotomes are indicated by Arabic numerals. The muscle-buds are shown in black, those within the fin rudiment being numbered with Roman numerals. The nerve-trunks are shown with double contour. with their neighbours near their proximal ends. These connecting bridges persist for a short time and then disappear. According to the authors mentioned they are the expression of a cellular interchange taking place between neighbouring muscle rudiments. 208 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. E. Miiller (1911) believes the connecting bridges in the case of Acanthias to be special developments of a syncytial network which lies between the buds from the commencement: he fails to find in this animal any evidence of shifting of muscle-cells along the bridges. The matter appears to stand in need of further investigation. Already within the group of Elasmobranchs we find modification of the typical mode of muscularization of the fins outlined above. In the case of the most anteriorly placed muscle-bud of the pectoral fin of Spinaa the bud resolves itself into its constituent cells which separate before giving rise to muscle-cells. Again at the anterior and posterior limits of the pectoral limb musculature in Pristiwrus and Yorpedo the compact stage of the muscle-bud is eliminated entirely and the cells which muscularize the fin are budded off * ot. z u. u Fic. 115.—Diagram to illustrate the arrangement of mesoderm segments in the head- region of a young Elasmobranch embryo. (From a drawing by Agar. ) ot, otocyst ; sple, splanchnocoele ; t, u, v, w, occipital myotomes ; 1, 2, 3, anterior myotomes ; I, HL, ete., visceral clefts; *, ‘‘ Fourth” myotome of van Wijhe. separately from the myotome, wandering from their place of origin into the limb rudiment and there settling down (Braus). Amongst Vertebrates outside the group of Elasmobranchs such modification appears to be the rule. Thus in Acipenser and apparently in Lacerta typical .muscle-buds arise singly from the myotomes concerned. In Teleosts Harrison finds muscle-buds in the pelvic fin but a diffused origin in the pectoral. In Lung-fishes and Amphibians the origin seems to be again diffuse and the same appears to be the case in Birds. MESODERM OF THE Heap-REGION. — There are two important characteristics of the head-region of the Vertebrate ultimately connected with the muscular system: (1) loss of flexibility, associated with the evolution of brain and skull and (2) special muscularity of the wall of the alimentary canal, associated with the presence of important movable skeletal structures enclosed in the substance of the visceral arches. These peculiarities find their expression (1) in the tendency to suppression of the myotomes of the head- Iv MESODERM OF HEAD 209 region and (2) in the retention, to a greater extent than in the trunk, of the muscle-forming capacity of that part of the mesoderm which lies ventral to the myotomes. The mesoderm of the head-region shows the least amount of moditication posteriorly where its relation to the mesoderm of the trunk is still clear. In the occipital region—the region between the otocyst and the occipital arch, which may be taken as the hinder limit of the skull—we find a series of typical (“ occipital ”— Fiirbringer) myotomes, the mesoderm ventral to which takes part in the lining of the splanchnocoele just as in the trunk-region. This series of occipital myotomes seems clearly to be undergoing a process of reduction. It is largest in such, comparatively primitive forms as Elasmobranchs. Again during ontogenetic development the series commonly shows progressive reduction. In Spinas: for example seven occipital myotomes make their appearance, but as development” goes on the anterior three (¢, u, v)! break up and dis- appear, the fourth (wz) does so incompletely, while the last three (a, y, 2) develop into definite muscle-segments though of small size. As each anterior myotome disappears those behind it become shifted forwards so that its place becomes occupied by the myotome originally behind it in the series. It will be realized that there is thus introduced a serious source of possible error which has to be carefully borne in mind in observations on the development of the occipital region where the identification and correct reference of individual myotomes to their place in the series is of importance. Auteriorly the series of occipital myotomes is prolonged forwards past the otocyst by a mass of mesoderm (* in Fig. 115) which was regarded by van Wijhe (1883)—who may be said to have laid the foundations of modern work upon the segmentation of the mesoderm of the head—as the equivalent of a single (“fourth”) myotome. It has already been indicated that the series of occipital myotomes is undergoing reduction from its front end backwards and it seems on the whole more probable that van Wijhe’s “ fourth” myotome in the Gnathostomata is to be interpreted not as a single myotome but rather as the degenerate remnant of a series of myotomes. The number of myotomes originally present in this series does not appear to be capable of decision with any degree of certainty. Possibly it was very considerable and Froriep finds even in ontogeny (Zorpedo) that during early stages (Stage “D” of Balfour) as many as six distinct segments are recognizable in the region in question—in other words that the series of myotomes commences not with ¢ but with m, the anterior members of the series disappearing in turn as development proceeds. A point of 1 The hind end of the series—the occipital arch—being taken as a fixed point while the front end varies, Fiirbringer has introduced the convenient method of designating the individual occipital myotomes (or their nerves) by the terminal letters of the alphabet—the last being z, the one next in front y, and so on. The myotomes behind the occipital arch are counted as belonging to the trunk and are designated by numerals 1, 2 and so on (cf. Fig. 220). VOL. IT P Sets: 210 EMBRYOLOGY OF THE LOWER VERTEBRATES — CH. ‘interest is that the anterior limit of this series of recognizable ‘segments agrees approximately with the anterior end of the definitive notochord. In front of the “fourth” myotome of van Wijhe we find what appear to be fairly typical third and second myotomes, each con- ‘tinuous ventrally with the wall of the pericardiac portion of the \splanchnocoele. Of these myotome III gives rise to the External Rectus muscle and IJ to the Superior Oblique. At the front end of the series we have the first or premandibular or oculomotor myotome, - peculiar in that it is fused with its fellow across the mesial plane and that it no longer shows any connexion with the splanchnocoelic mesoderm. It gives rise to the four eye-muscles supplied by the Third cranial nerve—the Superior, Internal, and Inferior Rectus, and the Inferior Oblique. We have so far dealt only with the myotomes but the lateral or splanchnocoelic mesoderm is also continued well forwards into the head-region. Its more ventral portion forms the lining of the pericardiac cavity, while its more dorsal portion becomes traversed by the visceral pouches or clefts. The splanchnocoelic mesoderm ventral to myotomes II and III comes to form a stalk - like connexion between the myotome and the pericardiac wall (Fig. 115). This stalk is hollow in the case of myotome II and hes in the mandibular arch: in the case of myotome III it is solid and lies in the hyoid arch. In both cases the wall of the stalk gives rise to the muscular apparatus of the particular arch—in the one case the masticatory muscles and in the other the hyoidean musculature which is destined to attain to such a development in the mammals as the musculature of the face. The splanchnocoelic mesoderm corresponding to the myotomic mass behind myotome III (* in Fig. 115) is said to give rise to the musculature of the branchial arches. As the myotomic mass in ' question shrivels up during development, and the occipital myotomes pA move forwards to take its place, these myotomes come to overlie the splanchnocoelic mesodefm which gives rise to the branchial muscles. Consequently as will be realized the position of myotomes ¢, uv, and v in relation to clefts III, IV, and V as shown in Fig. 115 is secondary, the myotomes having moved forwards before the formation of these clefts. The above sketch has dealt with the cephalic mesoderm of Elasmobranchs but a similar scheme of development with minor variations in detail holds for other Vertebrates. Upon the whole it may be said that with upward progress in the evolution of the Vertebrata the segmentation of the mesoderm in the hinder part of the head becomes more and more obscured. Right up to the highest forms however traces of it persist. In Fowl embryos of about the third day of incubation the series of obvious myotomes may often be seen to be prolonged forward (see Fig. 236) by faintly visible ‘blocks agreeing in size and exactly in series with the myotomes. IV MYOTOMES OF HEAD 211 These blocks may be indistinguishable in ordinary thin sections but quite distinct in stained preparations of the whole embryo. It will > ' require strong evidence to justify the refusal to give them the inter- pretation that at once suggests itself—that these slight condensations | of the mesenchyme are as it were the ghostly remnants of once existing myotomes which in Birds have ceased to become functional. An important side issue of their presence to be borne in mind is that the slightly greater resistance of the more condensed portions of mesenchyme must necessarily exercise pressure upon the soft surface of the rapidly growing brain and produce a modelling of its surface which may be adequate to explain at least some of these appearances of segmentation of the brain-region which are included under the term neuromery. The blocks in question extend well forwards—in the specimen figured (Fig. 236) there are four distinguishable anterior to the middle of the otocyst and they may be taken as additional evidence in favour of there being not one but a number of myotomes repre- sented in the region of van Wijhe’s “fourth ” myotome. It is of interest to note that in the Lampreys the blurring of the segments immediately posterior to the third of van Wijhe’s series seems not yet to have come about and there is an undoubted simple “fourth” myotome (Koltzoff, 1901). We may justifiably associate this with the low degree of cephalization in these creatures which has involved a persistence of, or more probably a reversion to, an apparently archaic condition of this myotome and its immediate successors in the series. The relations of segments I, II and III to the eye-muscles have been worked out in a number of Elasmobranchs and similar conditions have been described in Reptiles and Birds. Our knowledge of the holoblastic Vertebrates in this respect is still fragmentary. In the case of Lepidosiren and Protopterus the eye-muscles develop out of compact masses of mesenchyme in which it is impossible to recognize definite segments (Agar, 1907) while on the other hand in Cerutodus (Gregory, 1905) these segments make their appearance much as in Elasmobranchs. Before leaving this part of the subject it should be pointed out that not all morphologists are convinced that segments I, II and III are actually serially homologous with the undoubted mesoderm segments or myotomes of the trunk-region: the blurring of the mesoderm arrangements between them and the admitted myotomes, and more especially their late appearance in ontogeny, at a time when the anterior members of the occipital series have already degenerated, are brought as evidence against the more generally accepted view. The present writer does not feel inclined to attach great weight to these objections. (1) The break or blurring of the series immediately behind III seems adequately explained by the disappearance of functional muscles in this region and (2) the relatively late appear- ance of myotomes I to III is explicable by the fact that the 212 EMBRYOLOGY OF THE LOWER VERTEBRATES CH. functional muscles derived from them are connected with the eye- ball an organ which becomes complete and functional only at a ‘relatively late period of development. + HypoprANCHIAL on HypocLossan Muscu.ature.—In addition to the musculature already indicated the Vertebrate head possesses on its ventral side a system of hypobranchial muscles which have the appearance of a prolongation forwards of the longitudinal muscles of the ventral body-wall. This hypobranchial musculature as a matter of fact does arise in ontogeny as a prolongation forwards of the anterior trunk and occipital myotomes, as is well shown by | Lepidosiren or Protopterus (Agar, 1907). About stage 29 the ventral ends of myotomes y, 2 and 1 are seen to be growing out at their ventral ends into a long slender prolonga- tion (Fig. 116). These processes grow outwards in front of the pronephros and undergo complete fusion at their tips. The fused apical portion ch soon separates » from the parent myotomes and grows forwards, on each side of the pericardiac cavity, until it reaches the hyoid arch. It now spreads ventrally until it meets its fellow below the pericardiac cavity. The common mass so formed be- comes converted into a sheet of longitudinal muscle-fibres, attached posteriorly to the shoulder girdle _ and anteriorly for the most part to Fic. 116.—Dorsal view of anterior myo- the hyoid arch (coracohyoid muscle, \ pee re eg of stage 29. Fig. 117, cor. hy), the branchial naa arches being reduced in the fishes c.h, coracohyoid muscle; N, notochord ; p.f, : : muscle-bud to pectoral limb ; pn, pronephros. in question. As the muscle goes on with its development the an- terior boundary of the portion belonging to myotome 1 becomes marked by a connective-tissue intersection, while in some specimens a similar intersection appears to demarcate y from z. In other Vertebrates the hypobranchial or hypoglossal musculature appears to originate in the same way—difference occurring only in the number of myotomes which take part. Five appears to be the most usual number (Seylliwm, Corning t Teleosts, Harrison). ELECTRICAL OrRGANS.—The conspicuous sign of a muscle becom- ing active is that it changes its shape: an inconspicuous accompani- ment of this change of shape is the production of a slight electrical disturbance. In the case of most electrical organs we have to do with portions of the muscular system in which the function of contraction has been reduced to a subsidiary réle or abolished 1Vv ELECTRICAL ORGANS 213 entirely, while the production of electrical disturbance has become predominant. We have here an excellent example of the principle of “substitution of functions,” which is constantly at work during evolution, the previously predominant functions of organs becoming subsidiary or falling into abeyance and being replaced by functions which were previously subsidiary. The development of the electrical organ can be conveniently studied in the case of the Skate (Raia) of which the most complete description has been given by Ewart (1888, 1889, 1892). In this animal the electric organ forms an irregularly spindle-shaped body which lies embedded in the lateral muscles on each side of the tail region. It varies in size in different species and is distinguishable to the naked eye from the muscle by its more gelatinous appearance. Fic. 117.—Side view of skull and myotomes of Lepidosiren, stage 38. (From Agar, 1907.) Cartilage dotted, myotomes indicated by outlines, nerves black. a.o.p, antorbital process ; aud. cups, auditory capsule ; br.n, brachial nerve ; ¢.ph.n, nerve to dorsal portion of constrictor of pharynx ; cor. hy, coracohyoid muscles; hy, hyoid arch; hypog.n, hypoglossal nerve; M.y, M.z, M.1, M.2, etc., myotomes; mand, Meckel’s cartilage; nas. caps, nasal capsule; occ. arch, occipital arch; occ. rib, occipital rib; pg, pectoral girdle; quad, quadrate; y, z, nerves; 1, 2, 3, etc., spinal nerves; 3 br, branch from 8 to brachial nerve. On examining transverse sections through the tail it is seen that the electric organ occupies the place of the middle one of five super- imposed portions into which the muscle is divided. And this clearly suggests, as Babuchin first pointed out, that the electric organ is morphologically part of the muscular system. That this is actually so is placed beyond dispute by the facts of development. In an embryo of &. batis about 7 cm. in length the position of the future electric organ is indicated by a slight moditication of the muscle fibres, inasmuch as some of these (Fig. 118, B) show a tendency to assume the shape of a club, the anterior end of the fibre being slightly thickened. In contact with this thickened end is a mass of protoplasm crowded with nuclei. This represents the motor end- plate which has assumed a terminal position. In a slightly older embryo (Fig. 118, C) the club-shaped fibre of the preceding stage has become further modified, the anterior end 214 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. Bia ln) —_ ee bse nes being now still thicker and = the whole fibre having assumed the shape of a mace. In the expanded head portion of the mace the cross striation is becoming closer, while in the slender handle the striation is becoming blurred and in the portion next the head is almost disappearing. The end-plate forms a very definite layer of uniform thick- ness covering the truncated an- terior end of the mace. It is crowded with large nuclei and to it pass nerve-fibres which show a regular dichotomous branching. In the fibre shown in Fig. 118, D, taken from the same 10 cm. embryo, the head of the mace is still more ex- panded as compared with the stem. The main portion of the head, in which the muscle striation had become closer, now forms a thick plate bent into a Fic. 118.—Development of the electric organ of Raia batis. (After Ewart, 1888.) A and B from an embryo slightly over 7 cm. in length; C and D from an embryo about 10 em. in length; E, from a specimen about 67 cm. in length. al, alveolar layer; ¢.1, electric layer; el, electrolemma ; 7./, nerve-fibres ; s./, striated layer; ¢, vestigial remains of muscle-fibre. Iv ELECTRICAL ORGANS 215 saucer shape with its concavity posterior and composed of numerous closely packed lamellae. It forms what is termed in the fully de- veloped organ the striated layer. On its anterior face the striated layer is covered by the end-plate, now known as the electric layer, while its posterior face is covered by a thick layer of richly nucleated protoplasm which, from the deeply pitted character of its posterior surface, is known as the alveolar layer. From this passes back- wards the main part of the muscle-fibre which shows symptoms of degeneration especially in the portion next the alveolar layer where it becomes vacuolated. Whether the alveolar layer represents, as seems probable, a localized thickening of the sarcolemma is not clear from the descriptions. In the fully developed condition (Fig. 118, E) the muscle-fibre has become converted into the functional electroplax (Dahlgren, 1908) or electrical unit. What was the head of the mace in earlier stages is now expanded to form a broad thin circular disc, lying perpendicular to the long axis of the body—the stem of the mace having degenerated into an apparently insignificant and functionless vestige (Fig. 118, E) or having disappeared entirely. The electro- plax is formed of the striated layer which is almost flat except round its edge where it is bent in a tailward direction. It is com- pletely ensheathed in syncytial protoplasm, that on its posterior face forming the alveolar layer, probably nutritive in function, that on its anterior face forming the electric layer. Into the latter there pass the numerous end-twigs of the nerve-fibres, the superficial (we. headward) layer showing a characteristic fibrillation of the protoplasm in a direction perpendicular to the surface (nervous layer—Ewart) in contrast to the deeper portion in which the proto- plasm is granular and nucleated (nuclear layer—Ewart). The tail- , like vestige of the posterior portion of the muscle-fibre is directly continuous with the striated layer. With the latter it represents the contractile portion of the original muscle-fibre, while the ensheathing protoplasm whether electric layer, or alveolar layer, or sheath of the tail-like vestige, is probably to be regarded as repre- senting the superficial portion of the sarcoplasm. As the muscle-fibres pass through the above-described modifica- tions, the connective tissue between them increases in quantity and becomes condensed between the electroplaxes in such a way that each electroplax becomes enclosed in a disc-shaped compartment. The walls of this fit close to the electroplax round its edge while, on the other hand, the anterior and posterior walls are separated, especially the latter, by a wide space from the face of the electro- plax. This space is occupied by connective tissue with sparsely scattered cells and a jelly-like appearance. That on the anterior side is traversed by the very numerous nerve-fibres which branching dichotomously pass towards the electric layer, while that on the posterior side is traversed by blood-vessels. ; During the earlier stages of development the electric organ 216 EMBRYOLOGY OF THE LOWER VERTEBRATES — cu. increases in size, partly bythe adding on to it of new electroplaxes formed at its periphery, but the marked growth which takes place in the organ later on is due to actual growth of the individual units which form it. Thus comparing a skate of 180 cm. length with one of 45 cm. the individual electroplaxes are found to have increased in size practically in the same proportion as has the body as a whole. © The above description deals with the development of the electric organ as it takes place in Raia batis. In other species of skate the process appears to be similar as regards its main features, but it is interesting to notice that the relative expansion of the front end of the muscle-fibre to form the electroplax is much less pronounced in cer- tain species than is the case in &. batis. Of the species so far investigated A. radi- ata shows the least ad- vanced stage of evolu- tion. In this species (Fig. 119, A) the elec- troplax is, a8 in vari- ous other species (e.g. RL. cwrewularis and £. fullonica, Fig. 119), in the form of a cup rather than a disc. In R. radiata the wall of the cup is very thick and retains throughout ee life only slightly modi- Fe ee a ea Vy ed cade, cence, fullonica, and (D) R. batis. (After Ewart, 1892.) The . electric layer ; is relatively feebly devel- oped, the thick alveolar layer is represented by hardly modified sarcolemma and the tail is only comparatively slightly degenerate. The skate has been taken as the basis for the description of the development of the electric organ since the phenomena concerned have been particularly clearly worked out in this fish. In the Torpedoes the electric organ develops from muscles in the region of the visceral arches by very similar stages. As regards the electric organs of Teleosts our knowledge is still very insufficient. ° In Mormyrids and in Gymnotus they are clearly modified portions of the lateral muscles as in the skate ; in Astroscopus (Dahlgren) they are believed to be derived from eye- muscles; while in Malopterurus though generally believed to be modified skin glands they are believed by Dahlgren and Kepner (1908) to be more probably of muscular origin. IV LATERAL MESODERM 217 LateRaAL MrsopErM.—The lateral mesoderm forms the lining of the splanchnocoele. Its superficial layer persists throughout life as the coelomic or peritoneal epithelium, while its deep surface produces by proliferation abundant mesenchyme cells which forma con- nective-tissue backing to the epithelium. The development of muscle- fibres, which is so characteristic a feature of the coelomic lining in the dorsal or myotomic region, is here to a great extent suppressed, this portion of the mesoderm no longer playing any part in the muscularization of the body-wall. It still however takes place in restricted areas, smooth or striped muscle-fibres being developed in those portions of the mesoderm which invest particular organs such as heart and blood-vessels, alimentary canal with its appendages, oviduct. The development of the musculature of the heart will more suitably be treated in the chapter dealing with the vascular system. As regards the muscles of the gut-wall we have little detailed knowledge, what there is being related mainly to the musculature of the skeletal elements contained in the visceral arches. In the larva of Lepidosiren the important point has been estab- lished by Agar (1907) that the sheath of muscle which forms the con- strictor of the pharynx is of double origin, its ventral and larger portion being a development of the splanchnic mesoderm covering the pharynx, while its dorsal portion arises as an outgrowth from one (y) or more of the occipital myotomes. The fact that muscular tissue derived from myotomes may join the splanchnic muscle to form part of the muscular sheath of the alimentary canal is of importance (1) by impressing upon us that an apparently homo- geneous muscular apparatus may really be heterogeneous —muscular- ized from two quite distinct sources, and (2) by indicating the possibility of splanchnic musculature being replaced by myotomic or conversely. Obviously the muscular sheath just mentioned might, by reduction of one or other of its component parts, become purely myotomic or purely splanchnic. As will be gathered later (Chap. VII.) the point is an important one from its bearing upon the discussion of certain problems of morphology. Rena Orcans.—In triploblastic Metazoa the function of excreting nitrogenous waste products is commonly carried out by tubular organs to which Lankester (1877) gave the name nephridia. Under this term were included the excretory tubes of Chaetopods, Molluscs, Rotifers, Trematodes, Turbellarians and Vertebrates. Subsequent research soon brought to light an important structural difference as regards the inner ends of these nephridial tubes in different groups of animals. In certain groups the tube possesses at its inner end an open funnel or nephrostome! which leads from the coelome into the cavity of the tubule, while in other groups the inner end of the tubule is without any coelomic funnel but is on the other hand provided with an arrangement of flame- 1 Goodrich terms such funnels ‘“ coelomostomes”’ and uses the word nephrostome in a special restricted sense. 218 EMBRYOLOGY OF THE LOWER VERTEBRATES cH. cells. The appreciation of this difference gave rise, not un- naturally, to a suspicion —which would now appear to be unfounded—that under Lankester’s name nephridium were included excretory tubes of two morphologically distinct types and the use of the word nephridium was often restricted to the one of these types in which the coelomic funnel was present. ; Later researches brought out the fact that in some cases—certain Polychaete worms—the excretory tube may possess both flame-cells and coelomic funnel. And finally the hypothesis was developed—by Meyer and especially Goodrich (1895)— that the nephridial tube and the coelomic funnel were originally quite distinct organs with separate openings to the exterior. On this view the primitive excretory tube or protonephridium (Goodrich) was provided with flame-cells at its inner end, while apart altogether from it and opening independently to the exterior was the coelomic funnel which formed the primitive exit for the reproductive cells. In the course of evolution there came about a fusion of the two structures, the coelomic funnel becoming as it were grafted on to the nephridium and in many cases shifted up the wall of the tubule right to its inner end. Such a compound organ (Nephromixium — Goodrich) might retain for a time both flame-cells and coelomic funnel—as in the Polychaetes alluded to above—or the flame -cells might, as is more usual, disappear leaving an excretory tube possessing at its inner end a coelomic funnel which shows no trace of its morpho- logically independent origin. To support the hypothesis which has just been outlined there is brought in the evidence of embryology which testifies (see Vol. I. p. 158) that the main part of the excretory tube is developed as an ingrowth of the ectoderm, while the coelomic funnel arises as an outgrowth of the mesoderm. This hypothesis has met with very general acceptance not merely with regard to the excretory organs of Annelids alone but also as a theory of the morphology of excretory tubes in general. As, however, the writer of this volume takes up a somewhat different standpoint it will now be necessary to state shortly what that standpoint is. The word nephridium will be used in the original sense as meaning an excretory tube whether possessing flame-cells or a coelomic funnel at its inner end. Physiologically the open funnel and the flame-cell appear to be associated primarily with two different sets of spaces. The funnel is associated with coelomic spaces and it serves to transmit to the exterior the products of the lining of such spaces—fluid, excretory, or reproductive. The flame-cell is associated rather with the meshes of the mesenchymatous spongework: it serves to filter off from these spaces watery fluid containing excretory salts in solution. The activity of the “flame” is in direct relation to the pressure of fluid within these spaces: if the pressure is lowered by making a minute puncture in the body-wall the movement at once ceases—to IV RENAL ORGANS 219 commence again when pressure is restored. This association of flame-cells with the spaces of the mesenchyme is seen in the more lowly forms in which they occur and it is therefore justifiable to regard it as primitive in spite of the exceptional cases in which flame-cells occur in coelomic cavities. The question of the relative antiquity in evolution of coelomic funnel and flame-cell is one which cannot be decided with certainty, depending as it does in turn on the unsolved question as to whether coelome or mesenchyme was evolved first. Looking to the occurrence of mesenchyme in Coelenterates (e.g. the Alcyonarians), in animals in which there is not as yet any closed-off coelome, the balance of probability seems to be on the whole in favour of the flame-cell having originated first, in other words in favour of the original nephridium being of the type called by Goodrich protonephridium. The fact that existing excretory tubes of this type arise from the ectoderm is also an argument for its antiquity, as it seems natural to suppose that primitively excretory products were got rid of at the outer surface of the body. In the primitive ancestral form the genital cells, formed by the lining of the enterocoelic pouches, would reach the exterior through the protostoma or primitive mouth but as evolution proceeded and the coelenteric pouches became separated from the enteron to form a closed coelome+ another mode of exit would have to be evolved. The natural mode of such exit would be by rupture of the coelomic wall at its weakest spot. Such weak spots would be provided at points where the cavity of the nephridial tube came into proximity with that of the coelome. At such points rupture would take place and the tendency would be for such a temporary rupture, at the time of maturity of the genital cells, to be replaced by a permanent ? opening from coelome into nephridium. This permanent opening would be the coelomic funnel (coelomostome, nephrostome in the original sense). The coelomic funnel, though originally developed to transmit the genital cells, would necessarily also serve as an exit for super- fluous coelomic fluid, and the fluid so transmitted would necessarily serve incidentally to flush out excretory matters passed into the lumen of the tube by the activity of its walls and would thus fulfil the function originally fulfilled by the fluid drawn in by the flame- cell. The function of the flame-cells being in this way otherwise provided for they would tend to disappear. The nephridial tube thus came to transmit (1) the reproductive cells and (2) the highly poisonous excretory products. There is, as it appears to the writer, ample evidence that under such circum- 1 The argument involves as will be seen the assumption that the coelome was in its evolutionary origin enterocoelic. This assumption appears to be justified by the numerous cases in which the coelome so arises in ontogeny. 2 An actual case where such a temporary rupture, brought about at parturition, has come to he a fixed character of the species and develops independently of mechanical rupture, is seen in the ‘‘median vagina” of certain Marsupials. 220 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. stances the tendency of subsequent evolution would be to separate from one another the paths to the exterior of the genital cells and of the poisonous excretory products respectively. It might fairly be anticipated for physiological reasons that there would be such a tendency but that it actually exists is demonstrated by the facts of comparative anatomy and embryology. Over and over again we find cases where such separation has undoubtedly come about. For example in the evolution of the Gasteropoda the right nephridium has lost its excretory function and come to be merely a genital duct. In Vertebrates there are several familiar examples of parts of the renal system which have to do with transmitting genital cells becoming separated off from those which retain a renal function.’ The present writer then believes the balance of probability to be in favour of the evolutionary origin of the type of nephridial tube commonly met with in coelomate animals, possessing a coelomic funnel or nephrostome at its inner end, having come about in the manner outlined above. The essential difference between the view here outlined and that developed by Goodrich is that it rejects the idea that evolution has brought about a more and more intimate connexion between originally independent genital funnel and nephridial tube as opposed to physiological probability. On the contrary it regards the funnel as having opened into the tube at the time of its first appearance, the progress of subsequent evolution having been in the direction of separating genital funnel and nephridial tube and not of uniting them. Even in the case of Polychaete worms the arguments against interpreting the anatomical arrangements in different genera as illustrating evolutionary sequence in the reverse order to that believed in by Goodrich seem unconvincing and insufficient to counterbalance the weight of physiological probability. In the case of a tube leading from the coelome to the exterior the two ends are almost of necessity mesodermal and ectodermal in their nature respectively. Consequently; the fact that the “nephromixium” has such a twofold origin in ontogeny does not appear to the present writer to constitute evidence of any particular weight that it actually arose in phylogeny by the fusion of two pre- existing independent organs. As regards the proportion derived from the two layers the probability would be that the specially excretory portion was originally ectodermal—excretory products being naturally got rid of by the outer surface—and that the portion specially concerned with the getting rid of coelomic products would be mesodermal—arising as a bulging of the coelomic lining. Accepting as a working hypothesis that the nephridial system of tubes with their nephrostomes aroge in the manner outlined above, it is important to bear in mind how greatly the system would be influenced in its subsequent evolution by the establishment of 1 fg. the separation of the Miillerian duct from the kidney system or the separation of the renal collecting tubes from the Wolffian duct. IV NEPHRIDIAL ORGANS 221 circulating mesenchyme or blood. This would render possible the shortening up of the nephridial tubes and the more definite localiza- tion of the excretory tissue. Whereas the original flame-cell type of excretory apparatus was diffuse—the flame-cells being scattered ' throughout the mesenchyme sponge-work—as it is still to be seen in the more lowly organized forms—it would now become compact, the waste-products being brought to it by the movements of the circulating blood. NEPHRIDIAL ORGANS OF VERTEBRATES.—Before passing on to the details of development of the renal organs in Vertebrates it is necessary to notice one or two points of general importance regard- ing the morphology of these organs within this particular phylum, and also to define precisely the sense in which certain technical terms will be used. In the first place the kidneys or renal organs of Vertebrates are built of tubules each of which is a nephridium according to the original definition of the term. The conclusion, already arrived at, that the ancestral Vertebrate possessed a completely segmented coelome, carries with it the further conclusion that in all probability a pair of nephridial tubes originally opened to the exterior from each segment. A characteristic feature however of the Vertebrates (with the exception of Amphioxus) is that the nephridia open not directly to the exterior on the surface of ‘each segment as in a typical Annelid but into a longitudinal duct which passes back along each side of the body and communicates at its hind end with the cloaca. The whole series of nephridial tubes on each side of the body is known as the archinephros! and the duct as the archinephric duct. In the embryos of Vertebrates development takes place from the head end backwards. We should therefore expect the nephridial tubules to appear in regular sequence from before backwards. It is however highly characteristic of the Vertebrate that the tubules, instead of developing in this regular sequence, develop in three batches one behind the other—an anterior, a middle, and a posterior. These constitute respectively the pronephros, mesonephros, and metanephros (Lankester, 1877). In many of the lower Vertebrates there is no separation between mesonephros and metanephros, the two forming a continuous structure which acts as the functional kidney. Such a type of renal organ consisting of the series of tubules corresponding to mesonephros together with metanephros may conveniently be termed the opisthonephros.” Of the four types of kidney just mentioned the first—the pro- nephros—forms the functional kidney in /arval Vertebrates. It is well seen in the larvae of Crossopterygians, Actinopterygians, Lung- fishes, and Amphibians, while, as Sedgwick first pointed out, it is 1 Archinephron, Lankester (1877). Price's term Holonephros is also frequently used in the same sense. ; : 2 In analogy with the use of the word opisthosoma in the group Arachnida, 222 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. reduced in forms with richly yolked eggs where the development is not larval. Its reduction or disappearance in the last-mentioned forms may probably be due to the facilities afforded for getting rid of excretory products by simple diffusion from the blood circulating on the surface of the yolk-sac into the surrounding medium. The opisthonephros forms the functional kidney in the adults of most if not all anamniotic Vertebrates. Distinct mesonephros and metanephros are found in the Amniota —the mesonephros being functional during the later embryonic ‘period and in the Reptiles during the first few months after hatching, Fie. 120.—Renal organs of the Frog (Rana temporaria) as seen from the ventral side after the ventral wall of the splanchnocoele and the portion of the alimentary canal contained within it have-been removed. (After Marshall, 1893.) A, 12mm. tadpole; B, 40 mm. tadpole; C, frog at time of metamorphosis. A, dorsal aorta; 4.7, aortic root; I’, fatty body ; gl, glomerulus ; op, opisthonephros ; pn, pronephros. while the metanephros forms the definitive kidney of the adult, its excretory activities being reinforced during the first few months in the case of Reptiles by the still functional mesonephros. THE PRONEPHROS.—A typical functional pronephros is well seen in a frog tadpole of about half an inch in length (Fig. 120, A, pn). It consists of a massive organ lying dorsal to the anterior portion of the splanchnocoele on each side. It consists mainly of a much con- voluted tube the anterior portion of the archinephric duct, and into. this there open three segmentally arranged pronephric tubules also, except the anterior one, much coiled and twisted. While the organ as a whole is retroperitoneal, 7.e. outside the coelomic lining, there exists an opening leading from the splanchnocoele into each tubule the nephrostome (Fig. 121, ms). Duct and tubules are lined with cubical, or almost columnar, epithelium and in the neighbourhood ‘IV PRONEPHROS 223 of the nephrostome the cells become pigmented and carry powerful flagella. At the lip of the nephrostome the lining epithelium of the tubule is continued into the flattened epithelium lining the splanch- nocoele which is richly ciliated in its immediate neighbourhood. The archinephric duct is continued back from the pronephros along the coelomic roof to open at its hinder end into the cloaca. The pronephros has very characteristic relations to the blood- vascular system. The tubules serve to transmit to the exterior the fluid secreted by the coelomic epithelium and a patch of this epi- thelium, lying on the roof of the splanchnocoele at its mesial side ° and facing the nephrostomes, has its secretory activity much exag- gerated. This specially secretory epithelium has its area increased by bulging into the splanchnocoele, the bulging portion enclos- ing a vascular skein connected with the aortic root. This bulg- ing structure is known as the glomerulus (Figs. 120 and 121, gl). The anterior convoluted part of the archinephric duct and the tubules opening into it have other relations to the vascularsystem, for their surface is bathed by the blood of the posterior cardinal sinus which forms a system of ir- ead oe ' regular spaces between gl, glomerulns; int, intestine; 1, lung; be livers M, medulla a oblongata; N, notochord; ns, nephrostome; oes, oesophagus ; them. This double op, operculum ; pn, pronephros. relation to the blood- system is doubtless correlated with the double function of the organ. It serves in the first place to get rid of watery fluid secreted by the coelomic epithelium, and with this function the glomerulus with its aortic blood supply is concerned; secondly it has to extract poisonous waste products from the circulating blood, and this is done by the wall of the tubule acting on the venous blood which bathes its surface. DEVELOPMENT OF THE PRONEPHROS IN HyYPOGEOPHIS.—Hypo- geophis, a member of the Gymnophiona, will be taken as an example of the mode of development of the pronephros for the following reasons : (1) In this as in other Amphibians the pronephros still becomes an actively functional organ. Consequently the probabilities are in favour of its developmental processes having departed less fronr the Fic. 121.—Transverse section through a 12 mm. Tadpole at the level of the pronephros. (After Marshall, 1893.) 224 EMBRYOLOGY OF THE LOWER VERTEBRATES CH. ancestral methods than in the case of those Vertebrates (¢.g. Elasmo- branchii, Amniota) in which the organ is modified to the extent of being reduced to a functionless rudiment. . (2) Its histological texture is comparatively coarse and the general structural arrangements in the embryo are so distinct as to eliminate to a great extent risk of observational errors. (3) Its development has formed the subject of a particularly careful and complete investigation (Brauer, 1902). Fic. 122.—Early stages in the development of the pronephros of Hypogeophis. Hach figure represents a' longitudinal section, so arranged as to pass outwards through the nephrotomes, cutting them across, and viewed from the dorsal side. (After Brauer, 1902, slightly simplified. ) A, from an embryo with 15 mesoderm segments; B, 12 segments; C, 16 segments; D, 27 segments. a.n.d, archinephric.duct ; pr, pronephric tubule. The Roman figures are placed in the nephrocoeles. The first signs of the pronephros make their appearance—in embryos with about 9 or 10 mesoderm segments—in the form of bulgings outwards of the outer or somatic wall of the nephrotome of segments IV and V. These outward bulgings are the rudiments of the pronephric tubules. A third soon appears in segment VI (Fig. 122, A, cf. also Fig. 123, A, pn). The three rudiments grow actively in length pushing their way tailwards along the body just external to the nephrotomes. They come to be in close contact and presently fuse to form a rod-like structure (Fig. 122, B) which continues to extend backwards towards the tail and becomes tubular through developing a cavity secondarily in its interior. This, at first solid, rod-like structure (Fig. 122, C, a.n.d) is the rudiment cf IV PRONEPHROS 225 the archinephric duct, which thus owes its origin to the fusion nC. gf pe C Fic. 123.—Development of pronephros of Hypogeophis as seen in transverse sections. (After Brauer, 1902.) A, embryo with 22 segments; B, with 29 segments; C, with 44 segments. .4, dorsal aorta; end, endoderm; gil, glomerulus; U.m, lateral mesoderm ; mc, myocoele ; my, myotome; N, notochord ; ne, nephrocoele; ns, nephrostome; p.c, peritoneal canal; pn, pronephric tubule; s.c, spinal cord; scl, sclerotome ; splc, splanchnocoele. together of the outer ends of the tubule rudiments belonging to segments IV, V, and VI. VOL. IT e 226 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. Additional tubule rudiments to the number of about 8 arise in order behind those first laid down. They arise in exactly the same way as the first (Fig. 122, B, VII) but as the archinephric duct rudiment has already grown past their point of origin they become joined on to the duct by their outer ends undergoing secondary fusion with it. Each tubule rudiment grows actively in length so | that it eventually forms a much-coiled tube connecting the nephro- coele or coelomic cavity of the nephrotome with that of the duct. In the meantime the nephrotome is undergoing important changes apart from the ‘tubule rudiment. Its cavity, the nephro- coele, from being a mere slit with its floor and roof in contact, becomes widely dilated and it becomes cut off from the dorsal part of the segment which forms the myotome and sclerotome (Fig. 123, C). The nephrotome also becomes gradually constricted off from the lateral mesoderm but in this case the separation either never becomes completed (Fig. 123, C, p.c) or if it does so, is merely temporary—communication being soon re-established at the point where the constriction took place. The nephrocoele is thus, even in the fully developed pronephros, in open communication with the splanchnocoele by a more or less narrow channel the peritoneal canal (Fig. 123, C, p.c), the splanchnocoelic end of which forms the peritoneal funnel. As the nephrotome and tubule go on with their development there arise characteristic relations with the blood-vascular system. An intersegmental branch from the dorsal aorta passes to each nephrotome, causing its floor to bulge into the nephrocoele (Fig. 123, C, gf) and form a conspicuous projection—the glomerulus— , which later on fills up most of the nephrocoelic space. From the glomerulus the vessel passes (as the vas efferens) into a network of blood-spaces lying between the coils of the tubule and belonging to the posterior cardinal vein. The fully formed pronephros of Hypogeophis is composed of about a dozen segmentally arranged units each developed in the way described. It is to be noted however, that the last three of these units never become fully developed and further that behind the last as well as in front of the first unit of the functional pronephros each segment has its typical nephrotome though this never proceeds with its development. In other words the pro- nephros of Hypogeophis possesses at its anterior and posterior ends a number of units more or less reduced or vestigial. It is also of interest to notice certain variations which occur in connexion with the relations of the tubule to the nephrotome. In what may be termed the typical arrangement the nephrostome opens from the dilated part of the nephrocoele (Fig. 124, A). Frequently however it has become shifted on to the constricted peri- toneal canal (Fig. 124, B). When it does this there are apt to arise very misleading appearances as illustrated by the accompanying figure, whereby on the one hand the tubule appears to lead directly from Iv PRONEPHROS 227 the splanchnocoele, the chamber containing the glomerulus appearing to be a side branch (Fig. 124, C), or on the other hand the pro- nephric chamber appears to form the dilated end of the tubule while the peritoneal canal appears to form a side branch (Fig. 124, D). In connexion with what has been said it is important that the student should get clear in his mind from the beginning (1) that the cavity into which the glomerulus projects (known as the cavity of the Malpighian body in the more highly evolved types of kidney) is simply a more or less completely separated off portion of the ns na Zz and and . cs ze ph RC. mc it 2 and. B / ar. dad Cc oe pr of Fic. 124,—TIllustrating variations in the relations of nephrocoele, tubule and peritoneal canal in the pronephros of Hypogeophis. a.n.d, archinephric duct; ne, nephrocoele ; ns, nephrostome ; p./, peritoneal funnel ; t, tubule. coelome (nephrocoele) and that neither it nor the peritoneal canal is to be regarded as a portion of the tubule, and (2) that the actual tubule commences at the nephrostome or opening leading into it from the nephrocoele. The word nephrostome throughout morpho- logy means an opening leading from coelome into nephridium. It is necessary to accentuate this because in many embryological writings the term nephrostome, or nephrostomal canal, is applied to the peritoneal canal which is not an opening leading from coelome into nephridium but simply a communication between the splanch- nocoelic and the nephrocoelic portions of the coelome. A clear appreciation of these points is of help in facilitating the comprehen- sion of a difficult chapter in Vertebrate morphology. 298 EMBRYOLOGY OF THE LOWER VERTEBRATES Cu. Fic. 125.—Transverse sections of Polypterus—stages 20, 23, and 28—passing through the rudiment of nephrostome B, which is seen projecting outwards from the wall of the nephrocoele. A, dorsal aorta ; (vel, coelome ; ent, enteric cavity ; me, myocoele ; N, notochord ; xe. uephrocoele ; p.c.v, posterior cardinal yein 5 splc, split representing splanchnocoele. In other Verte- brates possessing a functional pronephros the appearances seen in early stages are readily reconcilable with those described above for Hypogeo- phis, and we may take it that, apart from variations in detail, this represents the normal mode of development of the organ. CROSSOPTERYGII —In Crossoptery- gians, so far at least as Polypterus — the only member of the group investigated— is concerned, the first rudiments of the pro- nephric tubules are in the form of pro- jections which pass outwards and_ back- wards from the exter- nal side of each of the anterior nephrotomes (Fig. 125, A). The number of these tubule rudiments pre- sumably varies, seven being seen in one specimen and nine in another. Apparently the tubule rudiments become fused at their outer ends to form a solid mass—the rudi- ment of the archi- nephric duct. In the stages shown in Fig. 126, A and B, five tubule rudiments are seen passing at their outer ends into the IV PRONEPHROS 229 duct rudiment. As development goes on however the tubules belonging to nephrotomes I, III, and IV, those labelled A, CG and D in the figure, become reduced in size and finally A B Cc D E Fic, 126.—Dorsal view of pronephros of Polyplerus at stages 20, 28, 24+, 25 and about 28. a.n.d, archinephric duct. The tubule rudiments are indicated by letters, the nephrotomes by Roman numerals. disappear, while B and E on the other hand increase in length and become the functional tubules. The anterior end of the archinephric duct becomes gradually modelled out of the solid rudiment already referred to in the way indicated in the figure. HILT end Fic. 127.—Part of a longitudinal vertical section through the series of nephrocoeles in the pronephros of Polypterus—stage 24+. a.nd, archinephric duct; ect, ectoderm ; end, endoderm ; nc. B, nephrocoele “B” ; ne.F, nephrocoele ‘‘F.” After it has assumed its definitive tubular form this front part of the archinephric duct commences to grow actively in length, it becomes thrown into complicated coils and forms a large fraction of the entire bulk of the pronephros in its early functional stages. 230 EMBRYOLOGY OF THE LOWER VERTEBRATES CI. Ul op oo, and 10 15 20 29 30 35 39 Fig. 128.—Renal organs of the right side of a Protopterus larva of stage 34, (From a reconstruction by M. Robertson.) a.n.d, archinephric duct; op, opistho- nephric tubules; pn, pronephros. The capital letters indicate nephrostomes and the figures metotic * mesoderm segments. * “ Metotic ”= posterior to the otocyst. In its later functional stages the pro- nephros reaches relatively enormous bulk, occupying the whole thickness of the body-wall, but in these later stages the two tubules become much elongated and coiled as well as the duet itself. The nephrocoeles belonging to the various nephrotomes which develop ‘tubules form a series of closed cavities lying in a row one behind the other (Fig. 127, nc.B, ne.F). They are for a long time, in Polypterus, the only coelomic spaces which are widely open (Fig. 125, B). As development goes on the nephrocoeles connected with the func- tional tubules (B and E) become more and more dilated, their wall becoming thinner as they do so, and the floor bulging into the cavity to form the glomerulus. Eventually the cavity of the nephrocoele becomes continued ven- trally, a split spreading downwards to form the splanchnocoele, into which the nephrocoele opens freely. The portions of splanchnic mesoderm to which the glomeruli are attached, 2. the floors of the original nephrocoeles, become folded in towards one another, as the splanchnocoelic cavity dilates, to form the dorsal mesentery so that the glo- meruli are eventually borne by the mesentery one on each side. Meanwhile the nephrocoeles belong- ing to the tubules which atrophy gradually shrink up and disappear, and as they do so the two large functional nephrocoeles increasing still more in size meet and their cavities as well as their glomeruli become continuous. No definite constrictions (peritoneal canals) are formed between nephrocoeles and splanchnocoele, unless possibly during late stages, but the dorsal portion of the more posterior nephrocoele becomes cut off from the splanchnocoele by another method—the free edge of the glomerulus coming to fuse with the somatopleure so as to form a floor to the nephrocoele. Iv PRONEPHROS 231 This posterior nephrocoele is still in wide communication with the splanchnocoele indirectly by way of the anterior nephrocoele. Dipnor.—In Lepidosiren and Protopterus! the fully functional pronephros of the larva possesses usually two tubules (Fig. 128, B and D). These are the surviving members of a series of tubule rudiments extending through at least the anterior 4-7 segments but probably extending much further back. The tubules which become fully developed are normally “B” and “D” i.e. those corresponding to the second and fourth mesoderm segments. Thus the second tubule does not correspond with the second tubule in the fully developed pronephros of Polypterus. The tubules appear to originate (cf. Fig. 129, A, B) as in Hypogeophis except that the outgrowths from the nephrotomes are solid as in Polypterus and such is the case also with the archinephric duct rudiment. The nephrocoeles of the two main pronephric tubules undergo fusion as in Polypterus so as to form a large pronephric chamber on each side. This is continuous with the pericardiac portion of the splanchnocoele and the two glomeruli as usual become fused together to form a compound glomerulus.? In Lepidosiren the fusion of the pro- nephric chambers takes place before the appearance of the glomerular rudiments. These appear first on the floor of the continuous cavity (Fig. 129, C, gl) and very soon undergo fusion themselves. By differential growth the root of attachment of the glomerulus becomes gradually shifted towards the mesial plane and dorsally (Fig. 129, D and E) so that it comes to hang down into the pronephric chamber or nephrocoele from a point in close proximity to the dorsal aorta. The pronephric chambers are at first perfectly continuous with the splanchnocoele which spreads outwards from them. Later on the pronephros becomes greatly enlarged and bulges across the splanchno- coele until it comes in contact with the mesodermal sheath of the oesophagus. Fusion then takes place (at the point marked with * * in Fig. 129, E) between the surfaces in contact so that the glomerulus comes to be enclosed in a secondary pronephric chamber, which however remains freely open to the splanchnocoele at its hinder end. The glomerulus becomes firmly slung diagonally across this chamber by its tip undergoing fusion on the ventrolateral side of the chamber with the mesoderm investing the pronephros. In Ceratodus (Semon, 1901) the pronephros probably develops in a manner similar to that described in the case of the other two Lung- fishes. The organ in its first stage is a solid projection of the meso- derm the appearance in section being similar to that figured for Lepidosiren. The portions of the rudiment corresponding to the individual tubules are in such close apposition as to be at first indistinguishable (as is often the case in the other two Lung-fishes) : 1 A large part of the investigations upon which this account is based were carried out by Miss Muriel Robertson in the University of Glasgow during 1904. 2 The word glomus is often used for such a compound glomerulus, 932 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. it is only when nephrocoeles begin to appear (in the regions of the jill B fila ie bul Fic. 129.—Development of the pronephros in Lepidosiren as shown in transverse sections. A, stage 21; B, stage 21; C, stage 244. a.n.d, archinephric duct; end, endoderm; ent, enteric cavity ; gl, glomerulus ; l.m, lateral mesoderm; my; myotome; N, notochord; ne, nephrocoele; pn pronephric tubule ; s.c, spinal cord ; scl, sclerotome. fifth and sixth segments) that the segmented nature of the rudiment IV PRONEPHROS 233 becomes apparent. The fully functional pronephros has two tubules on each side, corresponding to the segments above mentioned: it may be presumed that these are the survivors of ,a once greater number, though there is no record of other rudiments having been actually observed. Fic. 1294.—Developinent of the pronephros in Lepidosiren as shown in transverse sections. D, stage 30; E, stage 31+. A, dorsal aorta; a.n.d, archinephric duct ; end, endoderm ; gl, glome- rulus; /i, liver; my, myotome; N, notochord; ne, nephrocoele; ces, oesophagus; pn, pronephric tubule ; p.v.c, posterior vena cava; splc, splanchnocoele. ACTINOPTERYGII—The acquirement of a thorough knowledge of the development in the more primitive members of the group—the ganoids—is an essential preliminary to the proper comprehension of the development of the more highly evolved Teleosts but un- fortunately our knowledge of renal development in the ganoids is still far from complete. 234 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. The tubule rudiments appear to arise in normal fashion, as out- growths of the lateral wall of the nephrotome. These outgrowths show the familiar variation of being sometimes hollow sometimes solid. Thus in Amia according to Felix (1904) the anterior three rudiments are hollow pockets while those farther back are at first solid. Tubule rudiments make their appearance from segment III to segment XIII but here as elsewhere only relatively few of these complete their development and are to be found in the pronephros at the height of its functional activity. Thus in a six-day Acipenser larva Jungersen found six functional tubules while in Amia Felix finds only a single tubule functional. In the latter case the tubule opens’ from a large pronephric chamber apparently formed by the fusion of at least three nephrocoeles. The tubule belongs originally to the most anterior of these and corresponding to it there is present a single open peritoneal canal. Later on this becomes replaced functionally by another peritoneal canal situated farther back. In Lepidosteus the functional pronephros has at least three tubules each with its nephrocoele (Felix, 1904). As in the case of the Lung-fishes the dorsal part of the splanchno- coele in the pronephric region becomes floored in by the approxi- mation of the mesial surface of the pronephros to the lateral surface of the oesophagus (cf. Fig. 1294, E) so as to form a secondary pro- nephric chamber. In Lepidosteus this forms a widely patent cavity with which the first nephrocoele becomes completely merged and which remains ventrally in continuity with the main splanchno- coele by a narrow richly ciliated tubular channel. In Acipenser the first nephrocoele undergoes a similar modification while the remain- ing five are fused with one another but isolated from the splanchno- coele. TELEOSTEIL—The development of the renal organs has been worked out in detail in the case of the genus Salmo by Felix (1897). In this case the myotomes are already separate from the more ventrally situated portions of the mesoderm at a very early stage. The first rudiments of the pronephros are in the form of a series of somewhat conical, segmentally arranged, solid projections from the median edge of the lateral mesoderm towards the mesial plane. These projections—five in number (segments 3-7) in a 26-day Trout—are probably to be regarded as nephrotomes which have been precociously separated from the myotomes, if indeed they ever were continuous. These five nephrotomes soon come into intimate contact so as to be no longer distinguishable. They now together form a continuous mass of mesoderm the so-called pro- nephric fold. The dorsal and outer portion of this mass becomes nipped off to form the anterior portion of the archinephric duct except at one point where a connecting isthmus remains to form a tubule. The mesial portion of the mass becomes the wall of the single pronephric chamber. IV PRONEPHROS 235 The whole mass is at first solid, the cavity of duct, tubule, and pronephric chamber, developing secondarily. The cavity of the pronephric chamber is for a time continuous with the split-like splanchnoccele, but it soon becomes constricted off from it and forms a completely closed cavity. Bearing in mind the segmented condition of the pronephric rudiment in its first stage of development and the process of fusion of successive nephro- coeles which takes place in Ganoids, we may conclude that the pronephric chamber of the Teleost probably represents a number of nephrocoeles fused together. The single pronephric tubule is very possibly the same member of the series as that which occurs in Amia although this has not yet been actually determined. A remarkable peculiarity found within the group Teleostei is that in a few genera (eg. Fierasfer, Zoarces, Lepadogaster) the pronephros retains its renal function throughout life (cf. Guitel, 1901, 1902). AMPHIBIA.—In Amphibians other than Gymnophiona the pro- nephric rudiment first becomes visible as a solid swelling of the somatic mesoderm at the level of the anterior mesoderm segments (Rana segments 2-9, Triton alpestris 1-6, Mollier). Though at first no segmentation is to be detected by the ordinary methods of observation in this swelling it is reasonable to interpret it as representing. morphologically a series of closely apposed or fused nephrotomal projections like those of Hypogeophis. This pronephric rudiment gradually becomes demarcated off from the rest of the mesoderm by a split which becomes apparent on its ventral side at first laterally and then spreads inwards. * The rudiment now forms a thick flap (cf. Lung-fishes, Fig. 129, A and B) hanging down on the outer side of the mesoderm, and continuous with the somatic mesoderm along its dorsal and median edge. Segmentally arranged coelomic splits make their appearance along the line of attachment of the pronephric flap and these we may interpret as incipient nephrocoeles. The split already mentioned as demarcating the pronephric rudiment ventrally spreads round its median edge, so as to detach it completely from the (nephrotomic) mesoderm except at segmentally arranged points where a connect- ing isthmus remains as the nephrostomal end of a tubule. The pronephric rudiment now undergoes a kind of modelling process similar to that occurring in Crossopterygians and Lung-fishes, its outer portion being gradually cut off from behind forwards so as to form the archinephric duct, while the part nearer the mesial plane forms the recurrent portion of the duct with the tubules connected with it. The whole rudiment is at first solid. The earliest coelomic spaces to appear are the nephrocoeles and from these split-like extensions spread outwards in each tubule rudiment, while towards the outer margin of the rudiment the continuous longitudinal cavity of the archinephric duct develops. Of the tubule rudiments, as usual, only a few become functional— 236 EMBRYOLOGY OF THE LOWER VERTEBRATES — cH. in Anura commonly 3, in Urodeles commonly 2 (in Amphiuma 3 according to Field). Probably here as elsewhere the number is really a variable one. As the tubules develop they show active increase in length so that they become much coiled and the same applies to the part of the archinephric duct lying in the pronephric region. “Tt is only when they first appear that the nephrocoeles show a segmental arrangement: later on they become merged in the general splanchnocoele. Along the inner wall of the dorsal portion of this cavity, i.e. the portion which represents the fused nephrocoeles, the glomerulus develops as a continuous laterally projecting fold of splanchnic mesoderm. Usually the portion of the body cavity con- taining the glomerulus becomes for a time incompletely shut off from the rest to form a secondary pronephric chamber as in Lung- fishes, the mesoderm covering the lungs undergoing fusion with that covering the bulging surface of the pronephros. The secondary pronephric chamber may in turn be subdivided by the edge of the glomerulus fusing with the mesoderm covering the pronephros. MEROBLASTIC VERTEBRATES. — As a rule, in the Meroblastic Vertebrates the pronephros never becomes a functional organ, and correlated with this it shows a reduction in its structure. Possibly, as already indicated, this may be due to the presence of the large yolk-sac with highly vascular surface in contact with the external medium, which will facilitate the getting rid of excretory material by diffusion outwards. ELASMOBRANCHII.—In Elasmobranchs the ventral ends of certain of the anterior mesoderm segments, usually commencing with segment VII, become dilated to form vesicular cavities (van Wijhe, 1889) which are probably to be interpreted as nephrocoeles. The tubule rudiments appear as thickenings of the somatic wall of these nephrocoeles which grow outwards and being in close apposition form at their outer ends, apparently by fusion, a solid continuous pronephric swelling. The tubule rudiments make their appearance in sequence from before backwards. Different workers vary in their. statements as to the number of rudiments in different forms [Scyl/iwm, 5—Riickert, 3—van Wijhe; Pristiurus, 5—Riickert, 4—Rabl, 3—van Wijhe; Raia clavata, 5—van Wijhe; &. alba, 8—Rabl; Torpedo, 7—Riickert (Fig. 130)] from which we may conclude safely that the number of tubule rudiments is very liable to variation both as between different species and different individuals. This variability may be taken in correlation with the fact, observed by van Wijhe, that in Pristiwrus dilated nephrocoeles made their appearance from segment I to segment XIV, gradually diminishing in size towards the end of the series, although tubule rudiments appeared only in 3 segments. Both phenomena indicate that the pronephros in Elasmobranchs as in other groups has under- gone reduction from a once much greater anteroposterior extension. In a comparatively late stage the tubule rudiments develop their IV PRONEPHROS 237 lumen. The pronephric swelling extends backwards into the archi- nephric duct. No glomerulus develops but segmentally arranged branches of the dorsal aorta appear on the right side corresponding in number and degree of development with the pronephric tubules. These give rise either one of them (Riickert, van Wijhe) or by fusion together (Rabl) to the root of the vitelline artery but are termed by Rabl pronephric arteries. The pronephros undergoes rapid degeneration and eventually nothing is left of it but the coelomic funnel of the Miillerian duct (see below). Sauropsipa.—In the fowl pronephric tubule rudiments develop to the number of about 12, in the form of solid outgrowths of the somatic mesoderm at the level of the nephro- tomes although, except in the case of the most anterior, the mesoderm is not yet segmented at this level at the time when the rudiments appear. In at least some cases segmental dilatations of the otherwise split-like coelome occur op- posite the tubule rudiments and are no doubt to be interpreted as the nephro- coeles of the corresponding tubules. The first tubule rudiment makes its appearance in embryos with 8 or 9 segments, its position showing consider- able variation (usually segment 4, 5 — or 6). , The successive tubules appear in : rapid succession—almost synchronously. "130 H aa ne At about the 19-segment stage the Torpedo. (After Riickert, 1888.) myotomes become separated from the a.nd, archinephrie duct; b.v, blood- nephrotomes, the latter remaining in vessel; pn.1, ete., pronephric tubules. continuity with the lateral mesoderm 77° SPvens ate numbered with and their cavities (nephrocoeles) with the splanchnocoele. About the same stage the hackwardly projecting tip of each tubule rudiment undergoes fusion with its successor in the series and thus gives rise to a continuous longitudinal rod-like structure —the rudiment of the archinephric duct (Felix, 1904). The archinephric duct in its anterior portion thus would appear to develop in a manner essentially the same as that found in the Gymnophiona and so many other of the lower vertebrates. The hinder end of each tubule rudiment, as well as the archi- nephric duct itself, is at first solid. The definitive lumen makes its appearance (about 20-segment stage) secondarily in the form of dis- continuous chinks which gradually become continuous and spread backwards. Pronephric glomeruli develop in the Bird though at a late stage 238 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. when the pronephros is already degenerating. They were discovered first by Balfour and Sedgwick (1878) in the Fowl where they vary in number from about 3 to about 7. They may, as so commonly occurs in pronephric glomeruli, undergo a less or greater amount of fusion with one another and also with the anterior glomeruli of the opistho- nephros. The whole pronephros in the Bird undergoes rapid atrophy and by the sixth day of incubation has usually in the Fowl com- pletely disappeared except the glomeruli which may still be detected for a day or two longer. In the Reptiles also a rudimentary pronephros makes its appear- ance but degenerates without becoming functional. The nephro- tomes or protovertebral stalks, at first solid, develop a patent cavity or nephrocoele. In a varying number of segments (in Lizards 6-8, commencing with segment V) pronephric tubule rudiments develop as outgrowths of the somatic wall of the nephrotome after the ordinary fashion and fuse together at their outer ends to form the archinephric duct. THE ARCHINEPHRIC Duct.—As has already been indicated, it is characteristic of the Vertebrate that its nephridial tubes no longer open directly to the exterior, but that, on the contrary, they open into a longitudinal duct on each side—the archinephric duct—which in turn opens into the alimentary canal towards its hinder end. The first steps in the evolution of the archinephric duct have passed beyond our ken and to decide as to how they came about we have to balance probabilities on a basis of somewhat scanty embryological and anatomical data. Two obvious possibilities present themselves —(1) that the row of segmentally arranged nephridial openings came to be sunk beneath the general surface in a longitudinal groove and that this groove became covered in to form a longitudinal duct, and (2) that the external opening of each tubule became shifted back- wards so as to open into its successor in the series and so give rise first to a common opening with it and later to a common longi- tudinal duct (Fig. 131) in the way exemplified by the posterior kidney collecting tubes of male Elasmobranchs. On which side the balance of probability lies will be apparent on considering the developmental facts so far as they are known to us at present. It will be recalled that in Hypogeophis, according to Brauer, the anterior portion of the archinephric duct arises by a number of pronephric tubule rudiments bending tailwards at their outer ends and undergoing fusion together. The fused portion forms the duct rudiment and it proceeds to extend backwards by independent growth until eventually it reaches and fuses with the wall of the cloaca. It is only a small portion of the duct close to its anterior end which is formed by the direct fusion of tubule rudiments—the tubules farther back growing out and fusing secondarily with the already formed duct. If we turn to other Vertebrates we find considerable evidence for believing that Hypogeophis presents to us a mode of development, IV ARCHINEPHRIC DUCT 939 of the archinephric duct which is relatively primitive. In a number of Vertebrates there appear to be distinct traces of the formation of the front end of the archinephric duct by fusion of the outer ends of tubule rudiments in a manner essentially the same as that which holds for Hypogeophis. As will have been gathered from the pre- ceding pages this is the case with such different groups of Vertebrates as Elasmobranchs, Crossopterygians, Lung-fishes, Reptiles and Birds. ; L—-frc. end. Ln. | _—eci. coel. ™YJ A D Te B c E Fic, 131.—Diagram illustrating a possible mode of evolution of the archinephric duct. A, the coelomic compartments are bulging towards the nephridial tubes; B, the compartments have come to open into the nephridial tubes and the flame-cells have disappeared ; C, D, the external openings of the nephridia are becoming shifted backwards so as to give rise to the archinephric duct ; E, the archinephric duct is completely formed and communicates with the enteron through one of the segments retaining, or reverting to, its primitive enterocoelic connexion. a.n.d, archinephric duct; ef, coelomic funnel; coel, coelomic cavity; cl.o, cloacal opening of archinephric duct; ect, ectoderm ; end, endoderm; f.c, flame-cell ; n, nephridial tube. If we are justified in looking upon this mode of formation of the duct in ontogeny as relatively primitive, it obviously affords strong support to the second of the two above-mentioned hypotheses as to the evolutionary origin of the archinephric duct: the bending back of the tubule rudiments would then be interpretable as a develop- mental reminiscence of the backward shifting of their external open- ings which took place during phylogeny. 240 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. The independent backgrowth of the remainder of the duct in Hypogeophis is probably to be regarded as a case of accelerated or precocious development to allow the anterior tubules to become functional at an early stage of development before those farther back have developed. As regards the ontogenetic development of the main part of the duct in other Vertebrates we find the most divergent statements and it seems clear that this divergence can only be explained by the actual facts not being always the same. In the Sauropsida it is admitted that the main part of the duct is formed as in Hypogeophis by independent backgrowth. Amongst the Anamnia the same is said to be the case in Elasmobranchs by Balfour and by Rabl, and in Alytes according to Gasser, but other authors describe two other methods of formation as occurring. The first of these is found in Elasmobranchs according to van Wijhe, Beard, Riickert and others. According to these investigators, the archinephric duct makes its first appearance as a longitudinal ridge -like thickening projecting inwards from the ectoderm. This becomes split off as a solid ectodermal rod which develops a cavity secondarily and forms the archinephric duct. Such a mode of development would be of great morphological interest as it would lend decided support to the view that the archinephric duct originated in evolution as an ectodermal groove—it being a common ontogenetic modification that what is morphologically a groove develops ontogenetically in the form of a solid ridge-like ingrowth. It has however to be borne in mind that there exists a serious source of possible error in making observations upon the archi- nephric duct in early stages. The duct lies between ectoderm and somatic mesoderm—the two cell-layers mentioned fitting close round it. During the various processes to which the embryo is subjected preparatory to being cut into sections the ectoderm usually separates slightly from the mesoderm, and the archinephric duct tends to adhere firmly to one or other of these layers. This is the case more particularly at its tip, where it is pushing the ectoderm and meso- derm apart as it grows back and is therefore in par‘icularly intimate contact with them. It is exceedingly difficult in studying sections to distinguish with certainty between such intimate contact and actual organic continuity. In cases where the hinder part of the duct is adherent to the ectoderm an appearance is produced which simulates closely a development by splitting off from the ectoderm. As a matter of fact C. Rabl’s very careful investigations (1896) fail to confirm the ectodermal origin of the duct in Elasmobranchs and upon the whole in the writer’s opinion there does not appear to be any longer justification for accepting it as actually occurring. The other mode, by which the extension of the archinephric duct backwards has been described as taking place in the Anamnia, is that the duct becomes split off from the underlying somatic mego- derm. It is necessary again to bear in mind the caution expressed IV ARCHINEPHRIC DUCT 241 above but making full allowance for this it seems impossible to escape the admission that in many forms (Petromyzon, Lung-fishes, most Amphibians, Teleosts and probably actinopterygian Ganoids) the duct is prolonged backwards by a process of this kind. It being accepted that in a number of Anamnia a large part of the archinephric duct arises in development by being split off from the mesoderm, we are faced by the problem how this mode of develop- ment is to be correlated with the mode of development. by fusion of the outer ends of tubule rudi- ments. It may be suggested sae that what has happened is that 2% the development has been accelerated—as often happens —by skipping over the early stages. The mode of develop- ment in question may have been derived from the more primitive mode by the omission of the separate tubule stage and the passage at once to the stage in which the tubule ends are fused into. a continu- ous structure. In some cases however the primitive mode of development 11111) has undergone a further modi- wT war -. fication. This is exemplified el. and by Polypterus (Graham Kerr, pyc, 132.—Transverse section through Polypterus 190'7) where the hinder portion of stage 23 at level of cloacal opening. of the duct appears to be a.nd, opening of archinephric duct into cloaca; cl, formed by bodily conversion opening of cloaca to exterior; end, alimentary canal of the series of nephrotomes. pra my, myotome; N, notochord; s.c, spinal These are not segmented but : form a continuous structure which becomes converted directly into the archinephric duct. In whichever way the archinephric duct completes its extension backwards, it eventually comes to open into the cloaca. This is, in the great majority of Vertebrates, described as coming about by fusion of the previously freely-ending tip of the duct with the cloacal wall. It is obvious that such a process cannot correspond with what happened during evolution as the duct must have had its posterior aperture throughout in order to perform its function. It is possible that a clue to the evolutionary origin of the com- munication between archinephric duct and alimentary canal is given by Polypterus. It has already been mentioned that in this animal the hinder part of the archinephric duct arises by bodily conversion of the row of fused nephrotomes. Fig. 132 shows that the opening of archinephric duct into the alimentary canal presents a striking VOL. II R 242 EMBRYOLOGY OF THE LOWER VERTEBRATES = Cu. resemblance to the primitive communication of mesoderm segment with enteron, and it is suggested that it actually is this primitive communication which has remained patent while in all the other segments it has disappeared. The two archinephric ducts open at first separately into the cloaca, one on each side. In some groups of Vertebrates however their terminal portions become gradually approximated and eventu- ally fused together into an unpaired dorsal vesicle which may undergo various modifications. In Elasmobranchs it forms the urinogenital sinus which bulges forwards and on each side becomes prolonged into the sperm-sac. In Lung-fishes it forms the cloacal caecum: in Teleostei the urinary bladder. It is noteworthy that in the adult Lung-fish the communication of the kidney ducts with the caecum is close to the posterior opening of the latter, so that a small amount of shifting would cause these ducts to open into the cloaca independently of the caecum. This suggests a possible evolutionary origin of the allantois. It is conceivable that a caecum similar to that of Lung-fishes arose by a fusion of the terminal portions of the kidney ducts ventral, instead of dorsal, to the alimentary canal and that the ducts then came to be emancipated from the caecum which remained as a ventral diverticulum of the cloaca to form the allantois. We have no detinite evidence as to the evolutionary origin of the allantois and it is well to bear in mind the possibility here indicated in addition to the simpler and perhaps more probable hypothesis that the allantois was from the beginning simply a bulging outwards of the ventral cloacal wall as it is actually in ontogenetic development. DEGENERATION OF THE PRONEPHROS.—The role of the pronephros as the functional renal organ is usually confined to comparatively early stages in development and at the end of this period, when its function is being taken over by the opisthonephros, the pronephros commences to undergo characteristic degenerative processes which normally culminate in its almost complete disappearance. In the frog (Marshall and Bles, 1890) these processes become apparent in the tadpole of about 20 mm. in length. The archi- nephric duct becomes more or less obstructed behind the pronephros and as fluid continues for a time to pass into the tubules the latter become greatly distended in places, their lining cells assuming a cloudy appearance, the cell boundaries becoming indistinct and their inner surfaces losing their smooth outline and becoming ragged. The whole organ shrinks in size, becomes invaded by leucocytes, the nephrostomes close, one after the other, and by the end of the first year the whole organ with the adjacent portion of the archinephric duct has practically disappeared. Mé.uertan Duct.—Throughout the series of gnathostomatous Vertebrates, with the exception of the teleostomatous fishes, the oviducts are admittedly homologous. They—the Miillerian ducts— are above all characterized by the fact that they open freely into the IV MULLERIAN DUCT 243 splanchnocoele at their anterior end by an open funnel (ostium tubae). There exists in some of the more archaic fishes what appears to be distinct evidence that the Miillerian duct has been evolved out of the tubules and duct of the pronephros and it will therefore be convenient to consider this evidence now. The Elasmobranchs are the fishes in question. In Torpedo (Riickert, 1888) as thé pronephros degenerates its tubules become reduced to the three hindermost. Of these three the two posterior degenerate while the other—tubule E—persists and its enlarged nephrostome becomes the coelomic funnel of the Miillerian duct. Other workers (e.g. van Wijhe and Rabl), working on other Elasmo- branchs (Pristiwrus), trace back the coelomic funnel of the Miillerian duct also to an opening derived from the pronephros and nephro- stomal in its nature, but they believe the opening to be formed not by the persistence of a single enlarged nephrostome but rather by the fusion of three or four nephrostomes together. That it is morpho- logically a single nephrostome is however rendered more probable by what we now know regarding the development of the pronephros in those of the more archaic fishes in which it develops as a functional organ. It will be recalled, for example, how in Polypterus tubule E (ike B) becomes enlarged as compared with A, C,and D. A pro- nephric tubule enlarged in this manner in correlation with purely excretory needs would provide an obviously adequate beginning for the evolution of a funnel for the transmission of the eggs like that at the front end of the Miillerian duct. While the funnel of the Miillerian duct is nephrostomal in origin the main part of the duct is developed in the Elasmobranchs (Semper, 1875; Balfour, 1878) from the archinephric duct. The latter undergoes a process of splitting from before backwards into a ‘dorsal and a ventral tube, the latter being at first a solid thickening of the ventral wall of the archinephric duct. Of the two tubes so formed the ventral is continuous with the pronephric funnel, while the dorsal carries the openings of the kidney tubules farther back in the series: the former becomes the Miillerian duct, the latter persists as the functional duct of the opisthonephros (Fig. 133, C, W.d). This mode of development is satisfactorily explained by the assumption that the relatively archaic fishes in which it occurs are repeating the process by which the Miillerian duct arose in evolu- tion. Such a splitting of an originally common duct into two, so as to separate the routes by which two different products reach the exterior, is probably of frequent occurrence in evolution. Good examples are seen in the splitting of the common genital duct of hermaphrodite gasteropods (eg. the ordinary snails) to form a separate oviduct and vas deferens. It appears then justifiable to accept as a working hypothesis that the Miillerian duct arose in evolution by being split off from the archinephric duct and that its coelomic funnel is a persistent pronephric funnel. Turning to Vertebrates other than Elasmobranchs, well-marked 244. EMBRYOLOGY OF THE LOWER VERTEBRATES CH. differences are found to exist between the phenomena as described for different groups and even for members of the same group by Pp? -, 5 ~) SRoouca 4 RQ and... Fo > 29 L) SS a) WY? ; |e cs &.. pe lo Si Fic. 133.—Arrangement of archi- nephric duct, etc., in embryos of Pristiurus. (Based on Rabl’s figures. ) oO A, male 17 mm.; B, female 19 mm.; C, female 27 mm. a.n.d, archinephric duct ; cl, cloaca; M.d, Miillerian duct; os, coe- lomic opening of Mullerian duct; pn, pronephric nephrostome; W.d, duct of opisthonephros. different observers. While some of these may be due to observations being pushed to within the limits of prob- able error it is impossible to avoid the conclusion that great differences do actually exist in the details of develop- ment of the Miillerian duct. It is possible on general embryo- logical principles to arrive at an idea of the kind of variations which might be expected to show themselves from the supposedly primitive mode of de- velopment. I. The Miillerian duct might con- tinue to arise in an unmodified manner by splitting from the archinephric duct, its funnel being a persisting nephro- stome. II. In correlation with the fact that the one derivative of the archi- nephric duct (duct of the opistho- nephros) is required to be functional at a very early period, while the other (Miillerian duct) does not function until adult lite, there would be a tendency for the two ducts no longer to keep exactly abreast in their development but to become separated, the Wolffian duct developing relatively earlier, the Miillerian relatively later. To enable this to take place, the primitive stage in which the two ducts were still one would tend to be more and more curtailed until it was eventually elimin- ated and the two ducts were independent from the beginning. Ill. The independently arising Miillerian duct might retain the mode of extension backwards by intrinsic growth, eventually reaching and fusing with the wall of the cloaca. IV. Its separation from the somatic mesoderm might take place relatively later than its extension backwards so that it would arise in develop- ment comparatively late, as a longitudinal ridge or fold gradually separating off from the mesoderm from before backwards. Iv MULLERIAN DUCT 245 A survey of the phenomena as described for the various sub- divisions of the Vertebrata shows that as a rule they may be fitted without difficulty into one or other of these types of moditication. Thus in the Amphibia some, especially of the older observers, described the extension of the Miillerian duct as taking place by splitting off from the archinephric duct as in Elasmobranchs, others as being due to independent intrinsic growth, still others as a process of folding or splitting off from the splanchnocoelic epithelium. One of the most careful modern accounts (H. Rabl, 1904) based upon the phenomena observed in the relatively primitive Urodela (Salamandra) states that the funnel is the persisting “second” nephrostome of the pronephros and that the portion of duct behind this arises as a thickening of the coelomic epithelium—the cells first assuming a columnar shape, then becoming arranged in several layers to form a ridge projecting into the subjacent connective tissue, and finally becoming split off as a solid rod. Only the anterior portion of the Miillerian duct is formed in this way, the rod-like rudiment so formed proceeding to grow back independently to form the hinder part of the duct. In Reptiles and Birds the ostium is described as originating as a pit in the coelomic epithelium, which we may look on as a delayed and modified nephrostome, and the extension backwards as taking place by independent growth. The evagination of the epithelium to form the pit is, as is usual in such cases, preceded by the epithelium becoming somewhat thickened. The mode of origin of the Miillerian duct has not yet been worked out in detail in the Ganoids and Lung-fishes. In ordinary fishes (Teleostei) the conditions are peculiar and will be dealt with along with the development of the ovary. As regards the further development of the Miillerian duct, it should be noted that its completion and opening into the cloaca is commonly delayed till a comparatively late stage—often till a period but shortly before sexual maturity. Though primarily retroperitoneal the Miillerian duct comes, with increasing growth, to bulge into the splanchnocoele, pushing inwards the peritoneal lining which comes to surround it as a sheath con- taining muscles, blood-vessels, etc. Its lining epithelium becomes glandular and specialized to minister to the nutritive and protective needs of the egg in ways which differ in the different groups. Various modifications make their appearance in later stages. Very frequently the coelomic opening becomes shifted by the addition to the tube of a secondary extension formed from the peritoneal lining. In Elasmobranchs this shifting is towards the mesial plane and except in a few species leads to complete fusion so as to form a single median opening for the two oviducts. Again the hinder ends of the Miillerian ducts are in many cases approximated and they too may fuse to form a terminal unpaired portion. : In the case of the Birds the right oviduct lags behind in 246 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. development from about the eighth day of incubation; it never opens into the cloaca, and it persists in the adult as a functionless vestige. There is considerable probability that the genital pores—paired openings leading from the hinder end of the splanchnocoele directly into the urinogenital sinus (Cyclostomata) and through which the gametes pass out—are to be looked on as Miillerian ducts in the last stage of reduction, the whole duct having disappeared except its hinder opening. Whether there is any evidence bearing on this in their ontogeny is not yet known. The Miillerian duct goes through the early stages of development in the male as well as in the female. It usually however never opens into the cloaca and it soon becomes reduced to a vestige. This may persist to a greater or less extent as an individual variation or as a normal characteristic, eg. in the male Elasmobranch or Lung-fish well-marked vestiges remain in the adult, and so, still more markedly, in some of the Amphibia such as the Bufonidae and some of the Gymnophiona. OPISTHONEPHROS.—Here again Brauer’s excellent account of the development in Hypogeophis (1902) may be taken as a basis of our description. The opisthonephros in this amphibian is composed of segmentally arranged units extending from segment 24 to segment 100. Each unit is identical in composition with those of the pronephros, consisting of a tubule and a chamber (Malpighian body) containing a glomerulus and communicating with the splanchnocoele by a peritoneal canal. As in the case of the pronephros, each unit arises in development from the nephrotome or protovertebral stalk, the tubule rudiment being in the form of a diverticulum of the lateral or somatic wall of the nephrotome, the blind end of which comes in contact and fuses with the wall of the duct secondarily. Again as in the case of the pronephros, the nephrotome becomes completely separated from the myotome. It also becomes constricted off from the splanchnocoelic mesoderm, incompletely in some cases, a narrow communication — the peritoneal canal— remaining open between the nephrocoele and the splanchnocoele, but more usually completely. In this latter event a new peritoneal canal is developed secondarily in place of that which has been obliterated, a diverticulum growing out from the wall of the nephrotome which meets and fuses with the splanchnocoelic epithelium. There are differences in detail between the development of pronephros and opisthonephros, eg. the tubule rudiment makes its appearance relatively later in the case of the latter—at a period after the nephrotome has become constricted off from the splanchnocoelic mesoderm. viduality—forming a group of distinct PIO tubes, varying in number in different Bees Ss = nseating teer lates forms from about 4 (Spinaa) to about development of a segmental unit 15 (Acanthias), and converging SO as of the opisthonephros in male to open close together into the urino- Pristiurus. The figure is ineach genital sinus. In some cases the split- case a view from the mesial side. ?. . (After C. Rabl, 1896.) ting apart is not complete and more or A, 15th unit of 17 mm. embryo; B, 15th fewer of the tubes may be united to- unit of 225 mm.; C, 15th unit of 253 gether to form a longitudinal “ ureter.” mm.; D, 25th unit from eam embryo as The mode of origin of the Mal- canal; pf peritoneal fumel; 4 tubule. Pighian body — the definitive condi- tion of the nephrotome — from which each opisthonephric tubule leads has already been indicated. It is for a time rounded in form (Fig. 135) but eventually one portion of its wall—varying greatly in position—comes to bulge inwards to form the glomerulus containing a loop of blood-vessel. The peritoneal canal during development lengthens out consider- ably (Fig. 135, D) and becomes narrower. This narrowing is most marked in the posterior third of the opisthonephros and in this we IV OPISTHONEPHROS 251 see what is probably the expression of a general tendency for the portion of coelome containing the glomerulus to become more and more completely isolated from the main splanchnocoele as the renal unit becomes more and more highly evolved. Eventually, in the adult of the majority of Elasmobranchs, the peritoneal canal becomes completely obliterated, but in a considerable number of others! this happens, if at all, only towards the anterior and posterior ends of the opisthonephros so that the greater part of the organ retains open - peritoneal funnels throughout life. Bles (1897) has made the interesting suggestion that there is a physiological correlation between the persistence of open peritoneal funnels and the absence of abdominal pores — secondary perforations of the wall of the splanchnocoele in the neighbourhood of the anus which make their appearance, at a late period of development, in various Elasmobranchs and other Vertebrates. UropELa.—The third type of development of the opisthonephros amongst the more primitive Vertebrates is found in the Amphibians, especially in the Urodeles. The excellent account given by Fiir- bringer (1877) still forms a thoroughly adequate basis for the description. The Amphibians possess, as has already been shown, a large and highly developed pronephros amply sufficient for their excretory needs during early periods of development. In correlation with this there is marked delay in the development of the opisthonephros, the myotomes having already become separated and their stalks or nephrotomes breaking up into mesenchyme before the opisthonephric units make their appearance. The rudiments of these units—the nephrotomes—become reconstituted in the midst of the mesenchyme as solid cellular strands which may retain their metameric arrange- ment (Amphiuma— Field, 1891; anterior segments in Triton, Amblystoma, etc.) but usually have completely lost it. Each of these nephrotome rudiments is a solid strand of cells which curves out- wards dorsal to the duct. In the anterior region where, as is specially clear in Yriton, the inner end of the strand is for a time continuous with the lining of the splanchnocoele, the general arrangement is clearly the same as that of the Elasmobranch (cf. Fig. 136, A, with Fig. 134). The splanchnocoelic end of the nephrotome disappears for a time while the main portion develops a cavity in its interior and becomes converted into a vesicle with epithelial wall lying immediately dorsal to the duct (Fig. 136, B). This vesicle becomes elongated in a mediolateral direction (? by active growth of its outer wall) and then assumes a characteristic curvature first O- and then u)-like in shape (Fig. 136, C). The mesial end of the w gives rise to the Malpighian body, the remainder to the actual tubule, its outer end undergoing fusion with the wall of the duct (Fig. 136, D). The tubule grows rapidly in length and is forced into complicated 1 Eig. Cestracion philippt, Rhina squatina, Seyllium canicula,S. stellare, Pristiurus melanostomus, Spinax niger, Acanthias vulgaris, Scymnus lichia—(Bles, 1897). 252 Fic, 186.—Transverse sections showing various stages in the development of the opisthonephros. (After Fiirbringer, - 1877.) A, Triton alpestris; B, Salamandra maculata, 14 mm.; C, D, Salamandra maculata, 17 mm. ; £, Salamandra maculata, 21 mm.; F, Salamandra maculata, 25 mm. A, dorsal aorta; a.n.d, archinephric duct; g, gonad; gi, glomerulus ; n, nephrotome ; nc, nephrocoele ; p.c, rudiment of peritoneal canal; sple, splanchnocoele; t, tubule; tl, #2, ¢, primary, secondary, and tertiary tubule rudiments, EMBRYOLOGY OF THE LOWER VERTEBRATES CH. coils and windings as it does 80 (Fig. 136, E) while the Malpighian body dilates and its dorsal wall becomes invaginated to form the glomerulus. As a rule the primitive con- tinuity of nephrotome with the splanchnocoelic lining disappears in the Amphibian as already indi- cated, but it becomes re-established by a peritoneal canal developing secondarily (Fig. 136, D) as an out- growth, arising in Urodeles usually from the neck of the Malpighian body and in Anura from a point farther down the apparent tubule, which grows towards and fuses with a thickening of the coelomic epithelium. Such displacements of the communication between Mal- pighian coelome and splanchnocoele are probably of a similar nature to those mentioned in the case of the pronephros of Hypogeophis (see p. 226). In those parts of the opistho- nephros which are actively renal in function, 7.¢. the hinder portion in Urodeles and the greater part of the whole length in Anura, there takes place great increase in bulk, associated with the development of generations of subsequent tubules. Such secondary, tertiary, ete. tubules make their appearance amongst the mesenchyme in the form of cellular strands which resemble closely — both in their original appearance and in the series of changes which they pass through —those from which the primary elements arise (Fig. 136, F, ?°). Eventually the secondary tubule comes to open into the primary tubule, the terminal section of which thus forms a collecting-tube com- mon to both, while the tertiary tubule similarly comes to open into IV OPISTHONEPHROS 253 the secondary. As the various generations of tubules go on with their development, undergoing the same histological differentiation and increasing enormously in length, they become inextricably mixed up together to form the compact fully developed opisthonephros of the adult. Eventually, in the Urodele, the duct is slightly displaced out- wards so as to leave a distinct gap between it and the opisthonephros across which pass the terminal parts of the collecting-tubes. In the male Urodele the openings of these become, as a rule, shifted back- wards to the hind end of the duct as in Elasmobranchs. The Amphibia alone among tetrapod Vertebrates retain the relatively primitive feature of possessing open peritoneal funnels in the adult, and they can be excellently demonstrated with their actively moving flagella by examining the slender anterior portion of the excised and still living kidney of a female Urodele in normal salt solution under the microscope. In the anterior genital portion of the opisthonephros of the male they as a general rule (not in Spelerpes, Spengel) remain however obliterated. In the Anura (Nussbaum, 1886) the peritoneal canals at an early stage of larval life lose their connexion with the Malpighian body or tubule and establish a secondary connexion with the blood spaces between the tubules, thus affording a route by which the fluid in the splanchnocoele is returned to the blood, analogous to that provided by the lymphatic system in higher Vertebrates. Amniora.—tIn the Amniota the opisthonephros of the Fishes and Amphibians is represented by the mesonephros and metanephros— and it will be convenient to consider the mesonephros first. MESONEPHROS OF Brrps.—As has already been pointed out one of the marked differences between Amphiovus and the Craniata is that in the latter segmentation is no longer apparent at any stage of development in the ventral or splanchnocoelic region of the meso- derm. The Amniota show a further accentuation of this difference inasmuch as the loss of mesodermal segmentation has extended so far towards the dorsal side as to involve the region of the nephro- tomes. In the early embryo of the bird the nephrotomic part of the mesoderm has the form of an unsegmented mass—the intermediate cell-mass—showing more or less distinct traces of being composed of a somatic and a splanchnic layer continuous with the correspond- ing layers of the splanchnocoelic mesoderm and of the myotome. Although the intermediate cell-mass no longer consists of discrete nephrotomes, traces of its primitive segmental nature persist in its connexions with the segmentally arranged myotomes and in the fact that its connexion with the lateral mesoderm is not continuous in a longitudinal direction. ; ‘As regards the mode of origin of the actual mesonephric units differences exist, as was shown long ago by Sedgwick (1880), which are of much interest owing to the fact that the less modified mode of development found at the front end of the series is readily correlated ‘ 254 EMBRYOLOGY OF THE LOWER VERTEBRATES | cH. with that which is found in the Anamnia, while the more highly modified mode of development occurring posteriorly is equally readily correlated with what happens in the metanephros of the Amniota. In the anterior region (approximately segments 12-15) the inter- mediate cell-mass is compact, recognizably two layered, and the split which separates the two layers may be obviously continuous with the splanchnocoele (Fig. 137, A). It separates at an early stage from the myotome, but it remains continuous at intervals with the lateral 720. SJR Sun C. Fre. 137.—Sections illustrating the development of the mesonephros in Birds, (A and B, after Sedgwick, 1881; C, D, E, after Schreiner, 1902). A, 22-segment chick at level of the 15th segment; B, 34-segment chick at level of 13th or 14th segment (combined from two sections); C, 45-segment duck at level of 29th segment; D, 45-segment duck at level of 25th segment; B, 45-segment duck at level of 24th segment. A, dorsal aorta; a.n.d, archinephric duct; glom, glomerulus; ne, nephrocoele; nl, nephrotome ; p.c, peritoneal canal; p.c.v, posterior cardinal vein ; sple, splanchnocoele ; t, tubule rudiment. mesoderm. The intermediate cell-mass becomes closely apposed to and very soon directly continuous with the duct by a narrow isthmus in each segment—the tubule rudiment (Fig. 137, A). Ventrally, 2.e. near its junction with the splanchnocoele, the split between the two layers of the nephrotome dilates and forms a definite nephrocoele which opens into the splanchnocoele by a wide peritoneal canal (Fig. 137, B, pc). The tubule rudiment develops a lumen leading from nephrocoele into duct,! and the dorsal wall of the nephrocoele becomes 1 The opisthonephric duct in the Amniota is known as the mesonephric or Wolffian duct as its function is restricted to draining the mesonephros or ‘‘ Wolffian body.” \ IV MESONEPHROS 255 invaginated into the cavity to form a glomerulus (glom) which may become much enlarged so as to extend right out into the splanch- nocoele. As the process of development is traced back into the second region of the mesonephros (stretching approximately from segment 16 to 19 or 20) a distinct modification becomes apparent. The inter- mediate cell-mass in this region becomes loosened out into mesen- chyme, and amongst this loose tissue what may be termed the definitive nephrotomes make their appearance secondarily in the form of roundish condensations of cell elements. Each of these becomes more and more sharply marked off from the surrounding mesenchyme, its cells assume a radial arrangement, and presently a small rounded cavity appears in the centre. This cavity dilates and the result is a hollow vesicle with a wall composed of a single layer of cells—the definitive nephrotome. In the third, hinder, region of the mesonephros, extending from about segment 20 or 21 backwards, the process in the Fowl, though not in the Duck, has undergone the further modification that the intermediate cell-mass is from an early period completely isolated from the peritoneal epithelium. The peritoneal canals have here completely disappeared, except for faint vestiges, the cells of the peritoneal epithelium still showing here and there traces of the same arrangement as they have further forwards where they are passing into a peritoneal canal. Apart from this separation from the peritoneal lining, the process is similar to that already described. Here also the intermediate cell-mass becornes separated out into loose mesenchyme in which the definitive nephrotomes make their appearance secondarily. An important feature in the above described processes of develop- ment is the obliteration of the primitive segmentation of the nephro- tome region. When the definitive nephrotomes become visible, and so bring into view the metameric segmentation of the mesonephros, a further modification becomes apparent in that the mesonephric segments, except towards the front end of the series, are more crowded together than are the primitive mesoderm segments as represented by the myotomes (Sedgwick, 1880). Thus in the Duck Schreiner (1902) found in the region of myotome XX, 4 or 5 meso- nephric rudiments, in that of myotome XXV—7, in that of XXVI —9, in that of XX VII as many as 13. As development proceeds, the mesonephric elements become still more crowded together inasmuch as from segment 21 or 22 back- wards “subsequent” nephrotomes make their appearance in the mesenchyme. ‘These closely resemble in appearance the primary nephrotomes, with which they are at first in close proximity if not in actual continuity, and they develop in succession one over the other, each series forming a vertical row over its primary nephrotome. The number of subsequent tubules is greatest posteriorly where there are commonly four to a segment. 256 EMBRYOLOGY OF THE LOWER VERTEBRATES CH. The later development of the individual nephrotome of the meso- nephros takes place in the Birds along lines exactly similar to what takes place in lower forms such as the Amphibia. The tubule rudiment originates as an, at first solid but later pocket-like, outgrowth of the lateral wall of the nephrotome (Fig. 137, C, t). The tip of this presses against the mesial wall of the duct and, as the tubule grows in length, fusion takes place and the lumina of duct, tubule rudiment and nephrotome—which together form a characteristic cu-shaped structure as seen in a transverse section—become continuous (Fig. 137, D and E). The portion of the ~ nearest the mesial plane represents the nephrotome in the strict sense, 7.e. the forerunner of the Malpighian body, and has assumed a watch-glass shape, its dorsal wall being involuted into the cavity as the rudiment of the glomerulus (glom). The further development of the mesonephric unit, which need not be followed out in detail in this book, consists in (1) the immense growth in length of the tubule, which leads to its becoming in- extricably intertwined with its neighbours, (2) the histological ditferentiation of its wall, and (3) the differentiation of the Malpighian body. It should be mentioned that where the tubules are much crowded together they do not all establish a communication with the duct in the typical manner above described. Some, even of the primary tubules, come to open into neighbouring tubules. In the case of the subsequent tubules, some open into the duct in the typical fashion, others open into neighbouring tubules, while the majority become con- nected with pocket-like outgrowths from the duct. These outgrowths are greatly developed in some birds (cf. Duck, Fig. 138, B), becoming much elongated and taking the form of branched collecting-tubes into each of which open a series of subsequent tubules (ef. Hypogeophis), the whole condition distinctly foreshadowing arrangements presently to be mentioned in the metanephros. The mesonephros acts as the renal organ only for a short period during the early stages of development. In the Fowl it begins to develop about the end of the second day of incubation, it reaches its maximum about the 7th or 8th day, and almost immediately thereafter begins to show signs of degeneration as the renal fanc- tion becomes concentrated in the metanephros. The mesonephros never completely disappears though it ceases to be of any importance as a renal organ: its persistence is correlated with the fact that this portion of the opisthonephros has already in the forerunners of the Amniota important functions connected with reproduction. Its modification in relation to these functions will be gone into later. METANEPHROS OF Brrps.— The continuous mass of mesen- chymatous tissue representing the nephrotomes or protovertebral stalks does not cease at the hinder limit of the mesonephros at segment XXX: itis continued on through segments XXXI, XXXII, IV METANEPHROS 257 XXXIII and XXXIV to the level of the opening of the duct into the cloaca. The nephrotomal tissue in the segments mentioned remains for a time passive but eventually it gives rise to the definitive nephrotomes of the metanephros. The metanephros is therefore ontogenetically as was indicated long ago by Sedgwick (1880) in its origin simply a tailward continuation of the mesonephros. In the terminology used in this book it consists of the greatly enlarged posterior segment or segments of the opisthonephros. The develop- ment of the metanephros is inaugurated by the appearance of the rudiment of the ureter or meta- \ at 32 391347 nephric duct. This arises aS an out- A Baa growth (Fig. 138, B, ur) from the dorsal myn. Vee Tae apy wall of the meso- - nephric duct near its posterior end. The outgrowth ex- tends in a dorsal direction and then spreads out at its tip, projecting very slightly tailwards but growing much more actively in a headward direction along the outer side of the hinder or metanephric POLr- ye, 138.—Reconstructed outlines of hind end of mesonephric _ tion of the nephro- duct and ureter in Bird embryos as seen from the left side. tomal mesenchyme. (After Schreiner, 1902.) This latter becomes A, duck embryo with 48 segments ; B, duck embryo with 50 segments ; | C, duck embryo, 10°75 mm.; D, fowl embryo, 13°5 mm. mn, meso- secondarily (about nephros; wr, ureter. The Arabic numerals indicate the position of the end of the fifth the mesoderm segments. day) marked off by a distinct break from the mesonephric portion. About the same time the dorsal wall of the actively growing ureter begins to develop pocket-like outgrowths (Fig. 138, D). These increase in length, branch repeatedly, especially the hinder ones, and ‘become collecting-tubes. As this takes place the nephrotomal mesenchyme becomes condensed into small portions, one of which ensheaths the growing tip of each branch of the collecting - tubes. In these terminal caps of mesenchyme definitive nephrotomes gradually come into view, similar to those of the mesonephros. In other words the definitive nephrotome is at first a mere rounded cellular mass. This develops a lumen and, as the latter dilates, assumes a vesicular form, and finally the actual tubule makes its VOL. II s o | 2 | a | oe 1) 99 | a 1 oo | ae | ga | 258 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. appearance as an outgrowth which fuses secondarily with the tip of the collecting tube. The Malpighian bodies begin about the ninth day to develop their special characteristics in a manner similar to those of the mesonephros. An important point to notice is that the metanephros differentiates from behind forwards instead of in the opposite direction as does the mesonephros. About 24 hours (the exact time varies greatly) after the first appearance of the ureter the part of the mesonephric duct between it and the cloaca becomes incorporated in the cloaca so that meso- nephric duct and ureter come to have independent openings into the cloacal cavity. As the metanephros goes on with its development it comes to be situated in great part dorsal to the mesonephros but it will be under- stood that this topographical relationship is secondary. At first it is completely posterior to the mesonephros. Even the ureter is in its first stage localized about segment XXXIV (Fig. 138, B) and its extension forward as far as segment XXV or even farther is purely secondary. It will be noticed in Fig. 138 how exactly the ureter in its first beginnings resembles one of the pocket-like outgrowths of the duct which in the mesonephric region develop into collecting-tubes, and it seems scarcely possible to avoid the conclusion that the meta- nephros of the Fowl is simply the enormously hypertrophied nephridial apparatus of a single segment, the ureter being a greatly elongated collecting-tube with an immense number of subsequent tubules opening into it. OPISTHONEPHROS IN OTHER GROUPS OF VERTEBRATES CROSSOPTERYGII.—Our knowledge of the early stages of develop- ment is still fragmentary being based upon three specimens of Polypterus (stages 32, 33 and 36) obtained by Budgett (Graham Kerr, 1907). In the youngest of these stages a number of the opisthonephric units have made their appearance in the form of rounded cell masses arranged segmentally in the mesenchyme ventral to the myotomes, those which are best developed possessing a distinct lumen. In the specimen of stage 33 these rudiments have become elongated forming thick curved masses, one end of which is closely applied to, or even fused with, the dorsal wall of the duct. The lumen is restricted to the end farthest from the duct, which represents the definitive nephrotome, while the part which extends towards the duct—the tubule rudiment—is solid. In the 30-mm. larva described by Budgett (1902) the opistho- nephros commences about 4 segments behind the pronephros and stretches through about 39 segments with from two to five Mal- pighian bodies and tubules in each segment. On the right side of the body 18 of the Malpighian bodies—roughly one to each segment —communicated with the splanchnocoele by a nearly straight peri- lv OPISTHONEPHROS 259 toneal canal. The earlier developmental material does not suffice to show definitely whether or not, as is probable, this peritoneal canal is a secondary connexion with the peritoneal epithelium. The peritoneal funnels exist only for a time during larval life: in specimens 90 mm. in length they had disappeared. In Calamichthys (Lebedin- sky, 1895) the peritoneal canals have been found still persisting in a larva of 15 em. ACTINOPTERYGIAN GANOIDS.—Here again the definitive nephro- tomes appear as solid masses of cells arranged segmentally. The gap separating them from the pronephros is in the more primitive Sturgeons about 3 or 4 segments, in the more highly evolved Amia 16 or 17 (Jungersen, 1893-4). © The rudiment grows in length, develops a lumen secondarily, joins on to the duct by its lateral end while its mesial end dilates to form the Malpighian body—all in the usual fashion. Ata late period—after the Malpighian bodies have already assumed their characteristic features—they develop peritoneal canals as outgrowths from their walls which meet and fuse with the peritoneal epithelium secondarily. Later on the peritoneal canals become again obliterated and appear to be absent in the adult except in the case of Ama. TELEOSTEL—JIn the Teleostean fishes, as is indeed the case to a certain extent in all the members of the Teleostomi, the opistho- nephros is delayed in development in correlation with the prolonged functioning of the pronephros. According to Felix (1897) in the Trout the first opisthonephric units or definitive nephrotomes begin to make their appearance about 70 days after fertilization as rounded clumps of cells, in the centre of which a small lumen appears. These lie immediately dorsal to the duct, in the connective tissue trabeculae which at this stage of development traverse the’ cavity of the interrenal vein. These rudiments appear first about the middle third of the duct and gradually spread backwards, those in front being segmental in position while those farther back are no longer segmental and fuse together into irregular masses. Each rudiment grows actively in length and goes through the usual series of changes before joining up to the duct. To the primary units just described are added secondary and tertiary units. These develop exactly as do the primary except that in the case of the tertiary set the tubule may fuse either with the duct directly or with an already developed tubule. As the tubules increase enormously in length they become inex- tricably entangled together extending even across the median plane so that the substance of the two kidneys becomes continuous through the interrenal trabeculae. It is further characteristic of the tele- ostean kidney that there takes place in it a great development of round-celled pseudolymphoid (Felix) tissue. This forms a packing tissue between the tubules and appears to be formed by proliferation from the walls of the interrenal venous spaces. ; The opisthonephros extends back for a short distance behind the \ 260 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. cloaca. This postcloacal portion drains into a pocket-like prolonga- tion which grows back from the duct, usually on the right side only. A remarkable peculiarity has been observed in certain Teleosts (Lepadogaster, Guitel, 1901) in which, correlated with the persist- ence of the enlarged glomerulus of the pronephros, the Malpighian bodies of the opisthonephros have, at least in the adult, completely disappeared. Dipno1—In the Lung-fishes the development of the opistho- nephros closely resembles that in Amphibia. In Lepidosiren and Protopterus the units appear as rounded, at first solid, masses inde- pendent alike of myotome and of peritoneal epithelium. In Protop- terus they commence about segment 14-18 but in some specimens they appear to be represented by slight condensations of mesenchyme right forwards as far as the hind limit of the pronephros. The rudi- ments extend back to about segment 36 «e. to about one segment in front of the cloaca. They are roughly segmental in position and remain so during the greater part of larval life. Each rudiment gives rise to a typical Malpighian body and a tubule which joins on to the duct secondarily. There does not appear to be any trace of peritoneal canals developed although they are for a time present in Ceratodus. The development of the primary units is followed by the develop- ment of subsequent ones but the origin of these has not so far been worked out. In Protopterus, though not in Lepidosiren, the hinder ends of the kidneys become continuous across the mesial plane and this fused portion becomes gradually marked off conspicuously by its pale colour the cortical region of the paired kidneys becoming crowded with amoebocytes laden with melanin which settle down there and give it a coal-black appearance. Reptitra.—In the Reptiles we find, as we should expect, that the process of development follows upon the whole the same lines as in Birds but at the same time shows various features in which the con- dition remains more primitive. Thus in Lacerta Schreiner (1902) finds that, except in the hinder portion of the opisthonephros, the units arise directly from typical nephrotomes or protovertebral stalks. These become isolated from the peritoneal mesoderm and then from the myotome. Each develops a lumen and_ becomes vesicular and its lateral wall gives rise to an outgrowth which becomes the tubule rudiment and fuses with the duct. No peri- toneal canals are developed, though vestiges of these may appear—a vestigial peritoneal funnel appearing as a slight projection from the splanchnocoele into the ventral end of the nephrotome (Lacerta), or the latter remaining for a time connected with the peritoneal lining by a solid stalk representing the peritoneal canal (Anguis). In the posterior segments the nephrotomes are no longer distinct: they form a continuous mass of mesenchyme stretching uninterruptedly from segment to segment. In this, cellular conden- IV RENAL ORGANS 261 sations occur which give rise to definitive nephrotomes and these also are no longer strictly segmental, there being about 2 to each segment from segment 25 to 30. The definitive nephrotomes pursue the normal course of develop- ment. The first to appear are towards the ventral edge of the nephrotomal tissue but later other subsequent units appear in succession more dorsally. GENERAL MorPHOLOGY oF THE RENAL ORGANS OF VERTE- BRATES.—The main problem connected with the morphology of the renal organs is that which deals with the serial homology of its con- stituent elements. Lankester (1877) clearly implied this homology when defining his terms archinephros etc. while, looking at the matter from a more strictly embryological standpoint, Sedgwick (1881) formulated the view that pronephros, mesonephros and meta- nephros are simply successive portions of a single elongated ancestral excretory organ possessing a duct and segmentally arranged, serially homologous, tubules. ‘In discussing this archinephros theory it is necessary to bear in mind the following points :— (1) The names pronephros, mesonephros and metanephros accord- ing to their original definitions signify three sets of renal structures forming a succession along the length of the body in a tailward direction :—(a) an anterior or headward set, (b) a middle set and (c) a posterior or tailward set respectively. It is inadmissible by the terms of the original definition to use them in any other sense and to do so is bound to lead to confusion. (2) In addition to the anteroposterior series of renal units there may develop a sequence of elements within the same body segment— i.e. the development of the primary unit may be followed by the production of a series of subsequent units, secondary, tertiary, quaternary and so on, probably derived originally from the primary nephrotome by a process of budding. ‘he extent to which such subsequent units may develop differs greatly in different animals and in different segments. In the pronephric region there are commonly none, in the opisthonephros of Hypogeophis there may be as many as twenty in a segment, while it is possible that the metanephros of the Bird is to be looked on as a gigantic mass of subsequent tubules belonging to a single segment. It is obvious that in comparing renal elements of different parts of the series care must be taken that the comparisons are made between elements of the same order, and it is further obvious that a danger to be guarded against is involved in the theoretical possibility of the suppression of the tubules of one.order—say the primary tubules—in some particular region. The comparison of mesonephros with pronephros involves then these two fundamental questions :— (1) Does the mesonephros contain a set of units of the same order as those of the pronephros ?—1.¢. in this case primary elements and 262 EMBRYOLOGY OF THE LOWER VERTEBRATES — cH. (2) Are these elements serially homologous throughout the length of pronephros and mesonephros ? From the facts of development as stated earlier in this chapter it is clear what the answer to these two questions must be. It has been shown that in Hypogeophis and other forms the first tubule to appear in each segment of the opisthonephros arises as a direct out- growth from the nephrotome exactly in the same way as the pro- nephric tubule: it is clearly then a primary tubule, and its Mal- pighian body, arising directly from the main part of the stalk, is also primary. The evidence then seems conclusive that in Aypo- geophis the pronephric and opisthonephric tubules form a homologous series, and naturally if this is true of Hypogeophis it.is, in all proba- bility, true of other Vertebrates. Yet the view has been strongly advocated and is still held by many morphologists that there is no precise homology between the units which build up pronephros and opisthonephros. Riickert, van Wijhe, Field, Semon, Boveri, Felix, have been among the more _ important protagonists of this view. They have brought forward such arguments as the following :— (1) While the. pronephric tubule arises as an outgrowth of somatic mesoderm, the mesonephric is derived partly from somatic and partly from splanchnic. (2) The pronephric tubules arise relatively early and in continuity with the archinephric duct, the mesonephric tubules arise much later and in discontinuity with the duct. (3) The glomerulus of the pronephros is unsegmented and lies in the general splanchnocoele: that of the mesonephros is segmental and lies in a special chamber the cavity of the Malpighian body. These arguments however do not appear any longer to have the weight which formerly attached to them. (1) The evidence of Hypogeophis that opisthonephric tubules arise as outgrowths of the somatic wall of the nephrotome just as do the pronephric tubules seems quite convincing. (2) In Hypogeophis all the pronephric tubules except the first three join up to the duct secondarily precisely as do the opistho- nephric tubules. Further the precocious completion of the archi- nephric duct is a physiological necessity, in view of the early function- ing of the pronephric tubules, and this in turn involves as a necessary consequence that the tubules behind those which first function become joined to it secondarily. (3) The glomerulus of the pronephros is segmental and the pronephric chambers are also segmental at first in some of the more archaic forms and the unsegmented condition is purely secondary. Another line of argument is directed not against the view that pronephros and mesonephros are built up of serially homologous units but rather against the strict homology of the functional parts of these units. Thus it is stated that in the region of the pronephros in addition to the main tubules there occur rudiments of other Iv RENAL ORGANS 263 tubules which resemble more closely those of the mesonephros and similarly that in the region of the mesonephros, in addition to the ordinary tubules, there occur vestiges of another set of tubules resembling more closely those of the pronephros. Consequently, of the set of potential tubules (primary, secondary etc.) which is repeated in each segment, it is not the corresponding member which becomes the functional or main tubule in the pronephric and opisthonephric regions respectively. To the present writer the various observations which have been brought to support this argument do not appear to be anything like so convincing as the very clear evidence afforded by Hypogeophis and he consequently holds that in the present state of our knowledge there is no adequate reason to refuse to accept the precise homology of the first-appearing (“primary”) tubules of the opisthonephros with those of the pronephros. The idea of the primitive continuity between mesonephros and metanephros is less open to attack than that between the pronephros and the anterior (mesonephric) portion of the opisthonephros. Apart from the evidence of embryology we find in various of the lower vertebrates (Elasmobranchs, Urodeles) an elongated opisthonephros in the adult which shows in the clearest possible manner an incipient stage in the differentiation of the organ, into an anterior genital region and a posterior renal region, of precisely the same kind as we believe to have taken place in the Amniota. Further we have seen that in actual ontogeny the tubules of mesonephros and metanephros arise from an at first perfectly con- tinuous mass of nephrotomal mesenchyme. As regards the minor problem whether one or more primary tubules still persist in the metanephros among its immense mass of subsequent tubules there is, as yet, no adequate evidence. Accepting then the idea of the archinephros as a sound theory of the primitive condition of the renal system of Vertebrates we may sketch out the probable course of the modifications which have come about in its development somewhat as follows. Primitively its tubules developed—in accordance with the develop- ment of the body-segments generally—in regular sequence from before backwards. The disappearance of segmentation in the ventral portion of the coelome enabled the early-formed tubules—those towards the head end—to drain the whole length of the splanchnocoele. Correlated with this these tubules became greatly enlarged and their efficiency greatly increased. This high development of the anterior tubules to drain the whole splanchnocoele enabled them to cope with the entire excretory needs of the developing animal for a prolonged period and_ the tubules bebind them in the series being unnecessary were either delayed in their appearance or ceased entirely to develop. Thus a gap arose separating off the precociously developed tubules as the pronephros. Within the pronephros itself there was 264 EMBRYOLOGY OF THE LOWER VERTEBRATES — cu. a tendency for functional activity to become specially marked in certain tubules these becoming enlarged in comparison with the others. The increase in size of pronephric tubules was accompanied by increase in the size of their glomeruli, which con- sequently came into contact and fused together. As the pronephric tubules drained the whole splanchnocoele the peritoneal canal leading to their nephrocoeles became wider and wider until at last they ceased to be marked off from the rest of the splanchnocoele. The opisthonephric tubules—the renal functions being still for a time undertaken. by the pronephros—developed in regular sequence from before backwards. With the acquisition of new outlets for fluid in the splanchnocoele, such as abdominal pores, or connexions with lymphatic or blood vessels, the peritoneal canals leading from it into the nephrocoeles (Malpighian bodies), in which the secretion of coelomic fluid was specially concentrated, became gradually reduced and finally disappeared, there being no longer any physio- logical need for them. Within the series of opisthonephric tubules, the excretory function became more and more concentrated in the segments nearest the cloacal’ opening. In these segments the opisthonephros increased in bulk owing to the specially active budding processes which gave rise to successive generations of subsequent (secondary, tertiary, quaternary and so on) tubules. The final stage in this process was reached in the Birds, where renal activity became concentrated in a single segment close to the cloacal opening. In this segment an immense hypertrophy of the opisthonephric elements took place, successive generations of tubules being added on in front. Thus the opisthonephric mass belonging to this segment came to extend headwards dorsal to the anterior portion of the opisthonephrog (mesonephros) and became the definitive kidney or metanephros. : ORIGIN oF THE NeEpHRIDIAL Ducrs.—As already pointed out the nephridial tubes in craniate Vertebrates open primitively into a longitudinal archinephric duct—the presence of this duct being the most conspicuous feature which differentiates the renal system in Vertebrates from the presumably ancestral condition as exemplified by Annelids, where the tubules open separately upon the external surface. Two possible ways in which this duct may have originated in evolution have already been indicated and it has also been indicated that on the whole the balance of probability seems to be in favour of the view that it came into being through the backward shifting of the external opening of each tubule till it became coincident with the next behind it. Those who take this view usually assume that the archinephric duct originally opened posteriorly upon the outer surface of the body and that its opening’ became secondarily shifted into the IV NEPHRIDIAL DUCTS 265 cloaca. But as already pointed out there is no embry ological support for this view. Everywhere the archinephric opening is at first within the endodermal part of the alimentary canal and this suggests that the communication of duct with cloaca has come about in some other way. The evidence of Polypterus suggests as already indicated that the opening into the cloaca represents the persistent primitive communication of a mesoderm segment with the enteron. It is quite conceivable that a secondary communication between archinephric duct and gut may have come about in this way, in correlation with the pronephric part of the archinephros reaching the actively functional condition at a period when the mesoderm segment at the level of the anus had not yet been completely separated from the endoderm. Once this secondary opening was estab- lished it would be a natural consequence for the post-anal portion of the nephridial system to atrophy and disappear. The hypothesis indicated in this descrip- tion derives the nephridial apparatus of the Vertebrata from an ancestral condition resembling that characteristic of Annelids —the main difference being that in the Vertebrates the nephridial tubes open into a longitudinal duct which at its hinder end \d accu a ne communicates with the alimentary canal. It is of great interest then to find even within the group of the Annelida clear igs Sue a a4 expressions of the tendency for the nephri- peee a ame ai oe a dial tubes to open into such a duct. The Allolobophora antipae as seen best marked case of this known up to the — from the dorsal side to show present appears to be that of the Earthworm Shea ae ay Allolobophora antipae described by Rosa according to Rosa (1906). (1906). Here (Fig. 139) in the posterior portion of the body the nephridial tubes lead into a longitudinal duct which fusing posteriorly with its fellow opens into the alimentary canal on its dorsal side and near its posterior end. In other words in this particular case an arrangement precisely like that of the vertebrate has been evolved out of an ancestral condition in which segmentally placed nephridial tubes opened independently upon the outer surface. In regard to the origin of the typical metanephric duct or ureter as seen for example in a Bird there are two obvious possibilities. If the metanephros represents a number of nephridial segments its special duct may have originated by such steps as are represented by the adult condition in male Urodeles and male Elasmobranchs 1.e. by the openings of the collecting-tubes into the original duct becoming displaced backwards. Or on the other hand if the metanephros 266 EMBRYOLOGY OF THE LOWER VERTEBRATES — cu. represents the greatly enlarged tubule system of a single segment the ureter would probably have arisen simply by the enlargement of the collecting-tube of that segment. When one studies the facts of development as now known (see p. 258 and especially Fig. 138) the balance of probability appears to be decidedly in favour of the second of these hypotheses representing the method by which the ureter has actually arisen in phylogeny. THE Gonap.—The great mass of the cells which constitute the body of a Vertebrate or any other of the higher Metazoa are specialized for the performance of particular functions in the ordinary life of the individual, and, correlated with this specialization, such cells have lost the power of giving rise to reproductive cells or gametes. The main mass of the body constituted of such specialized cells is known technically as the soma. At one or more points in the body there remain however patches of cells which have not undergone this specialization for ordinary vital functions and which retain the power of giving rise under favourable circumstances to gametes. The sum total of such cells con- stitute the gonad. The word gonad iscommonly used in a loose sense as an equivalent of ovary or testis but it should be borne in mind that each of these organs contains a large proportion of immigrant tissues —connective tissue, blood, nerves and so on—which are strictly speaking part of the soma. The problem of greatest general importance attaching to the development of the gonad of Vertebrates is that which concerns the origin of the cells (gonocytes) which constitute it. And the interest of this question rests especially on the fact that in certain invertebrates the germ-cells have been traced back to blastomeres specially set apart at early stages of segmentation. All the probabilities seem to indicate that such a process if it occurs in the animal kingdom at all, is of a fundamental character and that indications of the same process may be confidently looked for in other groups. ‘ The most however that we seem to be justified in asserting to be definitely established for Vertebrates is that genital cells are derived from the mesoderm of the coelomic wall. Apart from the actual facts of observation such development of gonocytes from coelomic lining fits in well with general morphological ideas. It is clear that we must believe that in the simplest diploblastic ancestor of the Vertebrates the gonocytes were derived from epithelial cells. Itis also clear that, on the view that the Coelomata passed through an Actino- zoan-like stage during their evolution, we must regard it as probable that during that stage the gonocytes were situated, as in existing Actinozoa, in the endodermal epithelium lining the pockets between the mesenteries—an epithelium which, on that view, is represented by the endoderm of the enterocoelic pouch of an Amphioxus embryo aud by its derivative the coelomic mesoderm of an adult Amphioxus or other Vertebrate. In Amphioxus the gonad of the adult shows special peculiarities W GONAD 267 which mark it off from all other Vertebrates. Bearing in mind however that the general arrangement of the mesoderm of the adult Amphioxus, which also shows striking peculiarities, is preceded by a condition in ontogeny which there is reason to regard as more nearly primitive than occurs in any other Vertebrate—the possibility at once suggests itself that this may also be the case with the gonad. Consequently it becomes important to enquire what are the early conditions of the gonad in Amphiowus and whether it is reasonable to interpret the conditions in the more typical vertebrates as being modifications of those illustrated by Amphioxus. The earliest so far recognized stage of the gonad (Boveri, 1892; Zarnik, 1904) consists of a thickened portion of coelomic epithelium at the ventral end of the mesoderm segment 1.e. in the region where at an earlier stage the segmented part of the mesoderm was continuous with the portion which loses its segmentation. The thickening lies close to the headward boundary of the segment and within its ventral angle. As the segment has already become nipped off from the lateral mesoderm it is not possible to say from actual observation that the thickening belongs to the splanchnic rather than the somatic wall though this is probable from the condition in the more typical vertebrates. The genital thickening is repeated over a number of segments (from about the 9th or 10th to about the 34th or 35th—Zarnik). There are then three important points to be gathered from the study of the origin of the gonad in Amphiowus :— (1) It arises as a thickening of coelomic epithelium 7e. it shows the mode of origin characteristic of coelomate animals in general, (2) It arises close to the boundary of seemented and unsegmented mesoderm, and (3) It arises on the dorsal side of that boundary. In the more typical Vertebrates the ovary or testis first becomes clearly recognizable as a rule in the form of a longitudinal ridge— the genital ridge—which runs along the dorsal wall of the splanchno- coele on each side, at a varying distance from the line of attachment of the dorsal mesentery, and projects into the splanchnocoelic cavity. The genital ridge commonly extends over a greater antero-posterior extent than does the functional gonad later on—e.g, in the Salmon of the 185th day it extends from about the level of the fourth trunk myotome back to behind the anus (Felix). The restricted portion of the ridge which is destined to develop into functional ovary or testis is termed by Felix the gonal portion to distinguish it from the portions in front (progonal) and behind (epigonal) which remain sterile. The relatively great anteroposterior extent of the gonad during early stages in its development is probably to be regarded, along with the greatly elongated condition in the adult of some of the more archaic Vertebrates, as evidence that at one period of evolution the gonad extended throughout the whole length of the splanchnocoele. 268 EMBRYOLOGY OF THE LOWER VERTEBRATES cH. As development goes on the genital ridge increases in depth and is now termed the genital fold. This is composed of peritoneal epithelium covering a supporting and, later on, vascular core of mesen- chymatous connective tissue. The rudiment of the actual gonad in the strict sense consists of a thickening of the peritoneal epithelium covering the genital fold— the germinal epithelium. This thickened germinal epithelium may extend over both mesial and lateral surfaces of the genital fold as in most Amphibians, Reptiles and Birds or it may be restricted to its lateral (Ichthyophis ?, most Teleosts) or median (Elasmobranchs except in very early stages, Ichthyophis 3) surface. Of the more primitive holoblastic Vertebrates the Amphibia are the only group on which detailed observations on the origin of the gonad have been recorded. We shall accordingly summarize the early stages in the development of ovary and testis in this group and where possible in its more primitive subdivision the Urodela. Fig. 140 illustrates the earliest stages of the gonad so far identified in Urodeles, as described by Schapitz (1912) for the Axolotl Fig, A is taken from an embryo in which the protovertebral stalk or nephro- tome is not yet completely restricted off from the myotome. On its outer side is seen the rudiment of the archinephric duct. The stalk is continuous ventrally with the lateral or splanchnocoelic mesoderm. In its inner portion certain of its cells (eg. the two adjoining cells in the figure in which the nucleus is shown in a darker tone) are begin- ning to show recognizable indications of nuclear and cytoplasmic features which are characteristic of the gonad later on. It will be borne in mind that the wall of the protovertebral stalk is morpho- logically part of the coelomic wall to which therefore these gonad cells also belong. In the sections shown in B and C the mass of cells showing these peculiarities has become more and more distinctly marked off from the lateral mesoderm (mes) and may now be spoken of definitely as the gonad. In the stage illustrated by D the lateral mesoderm is seen to be spreading inwards towards the mesial plane ventral to the gonad and it is beginning to show here and there a distinct split separating its somatic and splanchnic layers. In the later stages (E, F, G) this split becomes a patent cavity—the splanch- nocoele (sp/c)—and the gonad is seen to lie on the dorso-lateral side of this, separated from the actual cavity by the somatic layer of peritoneal epithelium. In the last stage figured (G) the gonad is causing a slight bulging of the peritoneal lining into the splanchno- coele: this bulging is the incipient genital ridge (9.7). During the earlier of the stages illustrated the gonocytes gradu- ally acquire the superficial histological characters of germ-cells. The cell-body is larger than that of the other cells, it remains full of yolk particles, and in the spaces between the latter are to be seen fine granules of dark pigment. The nucleus is elongated or lobed in shape, the chromatin distributed in fine particles so that the nucleus as a whole stains less deeply than do the nuclei of other cells, and IV GONAD 269 large round nucleoli are present, frequently corresponding in number with the lobes of the nucleus. The embryonic gonad during the stages which have been described Fic. 140.—Origin of the gonad in Amblystoma. (After Schapitz, 1912.) A, 7-8 days after fertilization; B, 10 days; C,12 days; D, 18 days; HE, 17-18 days; F, 19 days ; G, just after hatching. ‘ o.n.d, archinephric duct; g, gonad ; g.r, genital ridge ; mes, lateral mesoderm ; my, myotome ; sple, splanchnocoele. At * in Fig. D is seen one of the heavily yolked cells which are interpreted by some as accessory gonocytes. is not as a rule continuous from end to end. On the contrary it consists of isolated pieces and these in many cases show distinct traces of metameric arrangement, the pieces being directly opposite 270 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. the mesoderm segments. The discontinuity becomes less marked in _ the later stages but even in an 18-day embryo Schapitz found the gonad still consisting on one side of the body of metamerically arranged blocks while on the other it had become a continuous strand, except for a single small isolated piece posteriorly. From what has been said it seems clear that the gonad of the Urodela is a derivative of the coelomic wall lying close to the boundary between the segmented and the unsegmented (lateral) portions of the mesoderm. As in early stages it consists of blocks with a roughly segmental arrangement it would appear to lie on the dorsal or nephrotomal side of the boundary mentioned. There is no apparent reason for declining to interpret this early segmented stage of the gonad as a persistent trace of a primitive segmental arrange- ment like that of Amphioxus. The tendency for the segmented condition to disappear in the typical Vertebrates is adequately explained by the gradual dorsal- ward encroachment of the unsegmented splanchnocoele. The boundary between segmented and unsegmented (lateral) mesoderm has altered much in position during the course of evolution, and there is no adequate reason to suppose that this boundary is not still a fluctuating one and if it is so we may expect varying traces of the original segmented condition to present themselves during develop- ment. The gonad has been described as being paired throughout but it may be mentioned that various observers have noticed an unpaired condition at one or other period during the early stages of develop- ment. This appears to be adequately interpreted as a secondary fusion similar to that occurring between the right and left opistho- nephros in a Teleost or in Protopterws rather than as a primary condition. We have traced back the gonad to its first recognized beginnings in one of the relatively primitive holoblastic Vertebrates. Before passing on to its farther development it has to be noticed that there exists a considerable volume of evidence pointing to the existence of additional germ -cells which arise independently of the coelomic lining and some of which migrate into the germinal epithelium and may give rise eventually to functional gametes. It is not proposed to describe this evidence as it has not as yet, in the present writer’s opinion, reached the stage of being convincing. It does not appear to have been satisfactorily demonstrated that the supposed extra gonocytes are really gonocytes at all rather than somatic cells. What is needed to provide such a demonstration is a careful study by skilled cytologists of the nuclear features of these cells, to determine whether there are any definite nuclear characters (such as Boveri discovered to be present in the gonocytes of Ascaris megalo- cephala) showing them to be beyond doubt gonocytes and affording a means of tracking them down in their supposed migration. Mere shape and staining capacity of the nucleus as a whole, or presence of IV GONAD 271 nucleoli, do not seem sufficiently definite characters as these are probably directly related to volume and metabolic activity of the cell. Cytoplasmic features—of which much use is made in this con- nexion—such as richness in yolk or roundness in shape are also unreliable. As regards the first of these, the study of the develop- ment of embryos rich in yolk brings out clearly the fact that the cells in particular tissues do not, by any means, all keep pace with one another in their developmental processes. Individual cells lag behind, and one of the commonest characteristics of such cells is that the yolk stored up in their cytoplasm remains unaffected for some time after that in the neighbouring cells has been completely used up. Obviously in such a case richness in yolk, even when occurring along with greater size due to less active division, does not constitute evidence of any weight as regards difference in morphological nature between the heavily yolked cell and those round about it. Again there is reason to believe that yolk may be stored up secondarily in particular cells or portions of tissue of a developing embryo as a preparation for future needs quite apart from the actual germ-cells. As regards approximation to a spherical shape, it should he remembered that there is a usual tendency for irregularly shaped or branching cells, such, as those of ordinary mesenchyme, to assume temporarily a rounded form at the period during and about mitosis. Such cells are apt to assume an appearance misleadingly like that of young germ-cells. The various features above indicated occurring together are sufficient to give a characteristic appearance to the cells in the main gonad but they form hardly definite enough criteria to prove that cells elsewhere are germ-cells in face of the strong probability that the whole mass of germ-cells in the body are of a common origin. GenitaL RipGk AND GENITAL Fotp.—The genital ridge was left as a slight bulging inwards of the peritoneal epithelium covering in the gonocytes. As development goes on the ridge becomes con- verted into a prominent fold—the genital fold. The peritoneal epithelium at first passes continuously over the surface of the strand of gonocytes but soon a change comes about in their relative positions the gonocytes coming to be incorporated in the thickness of the epithelium which may now therefore be spoken of as germinal epithelium. The gonocytes are to be seen first along the tree edge of the fold (Fig. 141, A) and this during subsequent development swells out greatly and forms the functional ovary or testis, while the proximal portion acts merely for suspensory purposes. The gonocytes increase in number by mitotic division but are also reinforced from small apparently indifferent cells lying between them (Fig. 141, ©, ge’). We may take it that these small cells are in all probability to be interpreted as cells of the original gonad which have lagged behind in development, though it is naturally difficult from mere observation to make certain that they are not ordinary peritoneal cells. At a particular stage in development (between 26 and 33 mm. in 272 EMBRYOLOGY OF THE LOWER VERTEBRATES — cH. Rana temporaria—Bouin) the total number of gonocytes in the gonad undergoes a remarkable reduction, e.g. from between 200 and Fic. 141.—Development of the gonad in Amphibia as seen in transverse sections. (A, after Schapitz, 1912 ; B-E, after Bouin, 1901.) A, larva of Axolotl (Amblystoma mexicana) ; B, Rana temporaria, 20 mm. tadpole ; C, 24 mm. tadpole ; D, 33 mm. tadpole; E, 35 mm. tadpole with hind legs completely developed. ol, nucleus of follicle- cell; g.f, genital fold ; g.s, genital strand; gc, gonocyte; gel, transitional stages showing conversion of apparently indifferent cells into gonocytes; mes, mesenchyme; som, somatic mesoderm; spl, splanchnic mesoderm ; y, yolk. 250 in a frog tadpole of 26 mm. to between 37 and 46 in a tadpole of 33 mm. (Bouin). During this process individual gonocytes Iv GENITAL ORGANS 273 degenerate and in many cases they appear to be shed into the splanchnocoele leaving behind them the spaces or follicles in which they lay walled in by indifferent cells. The meaning of this pheno- menon is obscure but a suggestion is made regarding it below. A period of active mitotic division of the gonocytes now sets in which leads to the formation of solid masses of gonocytes projecting down into the interior of the young genital gland (Fig. 141, E). These gonocytes are the ancestral oogonia or spermatogonia as the case may be. Up till now the genital fold has been spoken of as if it were merely a fold of the coelomic epithelium. As a matter of fact the fold very soon becomes invaded along its attached edge by immigrant mesenchyme cells forming a solid connective tissue core or frame- work which serves both a supporting and later, when it develops blood-vessels, a nutritive function to the developing germ-cells. The penetration of nests of gonocytes into this in the form of solid down- growths of the germinal epithelium may be interpreted as represent- ing an ancestral increase in area of the fertile portions of the germinal epithelium — the increase being originally brought about by the formation of hollow downgrowths into the vascular stroma, and these downgrowths having secondarily lost their cavities. The otherwise mysterious degeneration and shedding of large numbers of gonocytes already referred to may probably be regarded as a means of providing room for the localized parts of germinal epithelium which undergo this active proliferation. URINOGENITAL NETworK.—A characteristic feature of the Verte- brate is the set of tubular channels—vasa efferentia—which in most of its subdivisions connects the testis with the opisthonephros and is frequently to be recognized in more or less vestigial form in the female sex as well. Apart from variations in detail, these channels may be said to pass from the cavity of the testis to the cavities of the Malpighian bodies of the opisthonephros. They are clearly recognizable during early stages of development as solid strands of cells lying within the genital fold (Fig. 141, D, gs), the cavity in their interior develop- ing secondarily. As regards their first origin the majority of observers state that they are first recognizable at their renal ends and have therefore interpreted them as outgrowths from the coelomic epithelium of the Malpighian body, ze. from the nephrotome. Other observers seeing them make their appearance gradually in the core of the genital fold and reaching the Malpighian body secondarily regard them as differentiating im situ from the mesen- chyme, while still others have adduced evidence in favour of the cells which give rise to the strands being budded off from the peritoneal epithelium close to the attached base of the genital fold. The disparity between the statements of different observers is most reasonably to be attributed to actual variation in the mode of development. It may be assumed that originally the connexions VOL. II T 274 EMBRYOLOGY OF THE LOWER VERTEBRATES ch. of the genital portion of the peritoneal epithelium with the peri- toneal funnels, or with the nephrostomes, were in the form of open ciliated grooves or gutters on the surface of the peritoneum, that later on these became closed in to form tubular channels, and that in actual ontogenetic development in the modern amphibia the development from the coelomic epithelium has become obscured except for traces now at one end now at the other. At their distal ends the cell-strands in the male can be traced gradually farther and farther into the genital fold until they come into immediate relationship with the cell-nests of gonocytes. In the female of Urodela and Anura the strands do not spread so far into the genital fold, nor are they, even in early stages, so well developed as in the male. ‘The fatty body is developmentally simply a portion of the genital fold which becomes specialized as a store-house of fat. In Anura it is the progonal portion which undergoes this differentiation while in Urodeles and Gymnophiona the rudiment of the fatty body is continued backwards as a ridge along the mesial face of the genital fold throughout its extent. The fat is stored in the connective tissue of the organ, the fat cells being usually interpreted as lmmigrant mesenchyme cells which have invaded the rudiment by its base of attachment. It has also been suggested that these fat cells are peritoneal in their origin (Abramowicz, 1913)—a suggestion of obvious interest in view of the general tendency in the animal kingdom for potential germ-cells to ~ undergo degeneration in order to provide nourishment for the germ- cells which become functional. Testis.—The development of the functional testis out of the genital fold is seen in peculiarly simple and diagrammatic form in the Gymnophiona. Here the strands of the urinogenital network, as they sprout into the interior of the testis, anastomose together along its axis so as to form a central canal—around which, embedded in the stroma of the organ, lie the rounded nests of gonocytes. Fusion takes place between each gonocyte-nest and the wall of the central canal and then each nest develops a cavity in its own interior and becomes a hollow ampulla opening into the canal at its inner end. Various modifications of this simple scheme are to be found. In Gymnophiona themselves ampulla-formation becomes suppressed except in localized regions between successive vasa efferentia, so that “intervening portions of the testis are sterile and form merely thin tubular connexions between the bead-like fertile portions. Again : the ampullae vary in shape: they may be elongated and tubular (Discoglossus) or, as in the majority of cases, flattened against one another by pressure. The “axial” canal again may lie close to the surface: it may become greatly branched, as in most Urodeles, or may form a complicated network as in most Anura. Ovary.—lIn the differentiation of the ovary (Bouin, 1901) the most important points to be noted are the following. As regards IV OVARY 275 the germinal epithelium the most conspicuous feature is of course the immense increase in size, accompanied by the storing up of yolk in the cytoplasm, exhibited by those gonocytes which are to become functional eggs. Synchronously the indifferent cells of the germinal epithelium in their neighbourhood become converted into follicular cells, having for their main function the ministering to the metabolic needs of the growing egg-cells. The intervening portions of germinal epithelium, which do not undergo this modification, retain their germinal character and provide successive batches of eggs in successive breeding seasons. The cellular strands of the urinogenital network assume, as in the male, a tubular form, their wall becoming a cubical epithelium. As in the male the ovarian ends of these channels come into close relation with one another and fuse to form a central canal. In the Anura the fusion together to form an axial cavity appears to be less complete than in the male, a number of isolated central cavities being formed one behind the other. Fusions which take place later lead merely to reduction in the number of these central spaces (in Lana from about 12-15 down to about 5-7—Bouin). With further development, and as the functional egg-cells increase in size, the epithelial walls of these spaces become thin and flattened. Eventually they become pressed together and the cavity is reduced to a mere slit. The portions of the tubes lying nearer to the attachment of the ovary become vestigial. The presence of these axial cavities in the ovary, homologous with the central canal into which the microgametes are shed in the male, is of great morphological interest. It suggests the possibility that at one time the eggs were shed into this central space and therefore that the condition now holding in the Vertebrata, where the eggs are shed into the open splanchnocoele, is to be interpreted as a reversion to, rather than a persistence of, the primitive method of shedding the eggs. Leaving on one side the elaboration of histological detail which is not dealt with in this volume, the development of ovary and testis shows in its main features great uniformity throughout the gnathostomatous Vertebrates, and may therefore be dismissed with a few general remarks. Everywhere we see the gonad consist- ing at an early stage of a localized patch of coelomic epithelium in contact with nutritive and supporting mesenchyme: everywhere we see this coming to project into the splanchnocoele as a more or less prominent ridge or fold. A conspicuous feature is the widespread tendency towards increased localization of the actual functional areas of the germinal epithelium. This is seen on a small scale in the development of cell- nests of gonocytes separated by indifferent or sterile portions: it is seen again in the restriction of fertility to a relatively small antero- posterior part of the genital fold, long progonal and epigonal portions becoming sterile: it is seen again even in the actual differentiated 276 EMBRYOLOGY OF THE LOWER VERTEBRATES CH. testis where a considerable length towards the posterior end may lose its fertility and assume a merely conducting function,’ or where such sterile portions may be repeated at regular intervals throughout the length of the testis (@ymnophiona). ; This concentration of the activity of the gonad may affect its bilateral symmetry. In Elasmobranchs both ovaries may be present and functional (Laemargus, Notidanus griseus), or one may be function- ally inactive (Centrophorus, Trygon), or as happens in the majority, one, usually the left, fails to complete its development and is reduced to a more or less insignificant vestige. A similar reduction of one ovary takes place in many Teleosts. In the Birds the right ovary ceases its development at an early period and soon disappears entirely in the majority of individuals, although exceptions are of com- paratively frequent occurrence. Ovary AND Ovipucr or TELEOstomaTous FisHEs.—In the most archaic of existing teleostomatous Fishes—the Crossopterygian ganoids—the ovary is in the form of a typical genital fold which sheds its eggs into the splanchnocoele, from which in turn they pass out by a Miillerian duct. Consequently we may take it, in the absence of convincing evidence to the contrary, that the ancestral condition of the ovary and oviduct in the teleostomatous Fishes did not differ from that in other primitive gnathostomes such as Elasmo- branchs, Dipnoans, or Urodeles. A peculiarity however of the Crossopterygian oviduct as compared with that of the other groups mentioned is seen in the reduction of its glandular activity, and this reduction—which finds its physiological expression in the reduction of tertiary egg envelopes—probably gives a clue to the subsequent evolutionary history of the oviduct within the group Teleostomi, in the majority of which the whole Miillerian duct has apparently been reduced to the verge of complete disappearance. As regards the ovary itself there has come about secondarily— perhaps in correlation with increase in number and diminution in size of the eggs—a condition in which the eggs are not set free in the general splanchnocoele but are shed into an ovarian cavity, the wall of which is in complete continuity with that of the oviduct. The ovarian cavity is formed by the walling in of the portion of splanchnocoele which lies along the fertile (usually lateral) face of the ovary. The precise method of enclosure differs in detail in different Teleosts. In some (¢.g. Perca, Gasterosteus, Acerina, Zoarces) the ovigerous surface of the genital fold becomes invaginated, or overgrown by flaps which eventually meet and undergo fusion (Fig. 142, A): in others (eg. Cyprinoids) the free edge of the genital fold meets and undergoes fusion with the wall of the splanchnocoele (Fig. 142, B). Which of these two types of development is the more nearly primitive cannot be stated with certainty but the balance of 1 Lepidosiven and Protopterus (Graham Kerr, 1901); Polypterus and probably Teleosts (see below). Iv TELEOSTEAN OVARY 277 probability seems on the whole in favour of the first: mentioned, for the formation of folds or grooves of the fertile surface of the genital fold, so as to give increased area, is a very usual phenomenon, and the formation of a single longitudinal groove would readily lead to the first-mentioned condition. On the other hand the replacement of this condition by the second is also readily understandable. The ovary passes without a break into the oviduct which is simply the posterior sterile portion of the genital ridge in which a cavity develops secondarily—not always continuously—from before backwards. The oviduct differs greatly in length in different Teleosts: in some (Zoarces, Cyclopterus) the ovary itself may stretch right back to the genital pore. Although the above description fits in with the normal con- ditions, there are various Teleosts in which the processes of fusion connected with the ovary do not take place and in which the ovary remains as a genital fold hanging down into the splanchnocoele, Fic. 142.—Diagram illustrating the conversion of the genital fold into a closed ovary in the Teleostean fishes. e.g. in the case of the Salmon fusion of the ovarian edge with the body wall takes place anteriorly for a short distance and again in the posterior sterile region, but the greater part of the fertile region of the ovary hangs free. In such cases the eggs are shed into the splanchnocoele and pass to the exterior by genital pores (compare Cyclostomata, p. 246). Unfortunately we are still in almost complete ignorance regard- ing the development of ovary and oviducts in the Ganoids. From the little we do know it would appear that in Lepidosteus (Balfour & Parker, 1882) the ovary becomes enclosed in the same manner as in Cyprinoids (Fig. 142, B). Posteriorly it is continuous with the oviduct as in Teleosts generally. In the other Ganoids the ovary retains the form of a genital fold hanging down into the splanchno- coele while the oviduct is provided anteriorly with a coelomic funnel. The position of this funnel, far removed from the front end of the splanchnocoele, is sometimes used as an argument against the homology of this opening with the ostium of a true Miilerian duct, but such an argument carries little weight as we know from 278 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. the higher vertebrates that the ostium of an undoubted Miillerian duct is liable to undergo secondary shifting into such a position. Again the fact that the opening lies on the mesial side of the ovary is adduced as an argument in the same sense but in this case we have definite embryological evidence from Polypterus (Budgett, 1902) that this position is secondary, the early rudiment of the duct lying external to the ovary and immediately ventral to the Wolffian duct as is the case with a typical Miillerian duct. Consequently there is no sufficient reason to doubt that these oviducts'with open ostia in ganoids are really Miillerian ducts. PHYLOGENY oF TELEOSTEAN Ovipuct.—The facts of development show clearly that the main part of the Teleostean oviduct is of the same morphological nature as the ovary with which it is continuous. It arises from the hinder part of the primitive ovary which has become sterile and assumed a merely conducting function. The main difficulty connected with the morphology of the organ is that of accounting for the joining up of the part of the oviduct of ovarian origin with the cloaca or exterior. Balfour suggested that this had come about by the oviduct becoming fused with the lips of the “abdominal pores.” As an objection to this was adduced the observation by Hyrtl that in Jformyrus abdominal pores exist along with oviducts. This objection disappears, however, if we remember that in Balfour’s time there were confused together under the same name two different types of aperture—true abdominal pores‘and genital pores. Substituting genital pores for abdominal Balfour’s view seems still the most feasible. The probability seems to be that the main steps in the evolution of the Teleostean oviduct were as follows :— (1) The primitive oviduct or Miillerian duct underwent gradual atrophy becoming gradually shorter? until eventually nothing was left but its external opening—the genital pore. This process would doubtless be correlated with the loss of its glandular function and this in turn may have been connected either with the adoption of pelagic spawning, in which special tertiary investments for the eggs were no longer required, or with a special development of primary envelopes within the group. A stage was thus reached which is represented by the condition in Salmo. Of course we do not know whether Salmo has retained this condition or has reverted to it: the latter is more probable. (2) The portion of splanchnocoele along the ovigerous surface became enclosed so as to form a cavity which served to conduct the shed ova back into the neighbourhood of the genital pore. Anteriorly the ovarian surface abutting on this cavity remained fertile, while posteriorly it became sterile, so that the posterior portion of the cavity performed merely a conducting function (oviduct). (3) The lips bounding the posterior end of the oviduct from 1 We may see early stages in this process illustrated by the Ganoids 4mia and Acipenser. IV URINOGENITAL CONNEXION 279 being merely in proximity to the genital pore came to be com- pletely fused with the edges of the latter opening which consequently became the oviducal aperture. URINOGENITAL CONNEXION.—We have already summarized for the Amphibia the course of development of the urinogenital network —the system of tubes or vasa efferentia which connect testis and kidney and which serve as the outlet for the sperm. It is now necessary to glance at some points in the general morphology of this system of tubes. It has already been indicated that at their genital end the tubes become merged together in the axial cavity of the testis. The latter we must regard as morphologically an isolated portion of splanchnocoele into which the spermatozoa are shed although itis no longer traceable to splanchnocoele in actual ontogeny. It has also been suggested that the tubular channels were probably originally open grooves of the peritoneal lining which became con- verted into closed tubes as the gonad became isolated from the main splanchnocoele. The vasa efferentia frequently show a tendency, more or less pronounced, to anastomose together into a network. In the Amphibia it is a very general though not invariable rule that anastomosis takes place close to the edge of the kidney, forming the longitudinal “marginal canal” which is conspicuous in most Amphibians. A similar marginal canal is formed in Elasmobranchs. In taking a general view of the system of vasa efferentia we find that one of its characteristics is, as in the case of the gonad itself, a tendency to increased localization of its functional portions. Thus during ontogeny in Amphibians the vasa efferentia towards the hinder end of the series become blocked and non-functional, or dis- appear entirely, leaving only those at the anterior end functional. This process reaches its limit in such forms as Alytes or Discoglossus where only two or a single member of the series persist. Similarly in Elasmobranchs the number of functional vasa efferentia becomes reduced to a few at the anterior end of the series (Centrophorus 9, Scylliwm 6, Acanthias 4-6, Pristiurus 3, Mustelus 2-3, Rava 1). The same happens in Amniota. In the Dipnoi on the other hand the localization takes place at the hind end of the series, the functional vasa efferentia being reduced to about half-a-dozen (Lepidosiren) or to a single one (Protopterus). Another phenomenon which makes its appearance is the simplifica- tion and shortening of the route by which the spermatozoa pass from the vasa efferentia towards the exterior. Primitively the vas efferens opens into an otherwise normal Malpighian body containing its glomerulus and continued into a functional renal tubule. This condition may persist (Rana esculenta, Bufo), or the glomerulus may disappear (R. temporaria), or finally the whole Malpighian body and its tubule may be shortened and widened and converted into a simple tubular continuation of the vas efferens towards the opisthonephric CH. 280 EMBRYOLOGY OF THE LOWER VERTEBRATES “41 ySnoryy Suluado sourojsorydau YPA Suuly [wauopasead ‘ad $ sorydou -orystdo JO youp ‘po £ (amo[eoo Jo Sully jo word) peuos f -soajay, Sunoé so sniaydhjog &q pagiauises) Sonus ary sey aie He Roe ea pue 7 jo pue punt at} Usengag UorwolUNUWOS sy} “| + (snsapfoporg Aq pagrduiaxa) 47 Jo pue puly 4e ouo e[SuIs v 07 paonpar eIyMeraye wsea ‘qt (u oprday Aq pegydwexe) 47 Jo uonsod tepuny UL MOF BOF paonpat BIQuUalaya vsvA pur suotqiod Gz) eylzeays pue (Lz) euoouny ojUL paplatp 's14s0} 9 Coq Srasuallor Aq. payriduexa) 81980} pur AOUPH] JO YQSUET 8Y} YUOYTNoIYY pesszqes svlpog ULsdpey_ urezzeo Jo SAIPAV GY} JIA SuQvorunumtos pux afaooouyourfds woay yo qnys 17893 ‘gq + sorydeuoyysido eyg Jo setnqny pue sauiojsoryden ayy Ysnoryy Spavaquo passed puv efeooouyourlds ayy 07 Ur pays area suitads of] YOM UL UOTpuoD “y \ ‘SOOO, STVU at} JO Joup [eps oY} Jo WoOLN[ors oy} Suyvaysnql weiseig— ‘EFL ‘PLT , [ : por NI ad {aa} UO Mp oa SeUSTIES: IV URINOGENITAL CONNEXION 281 t duct (Discoglossus, Alytes, anterior vasa efferentia of Bombinator). In Alytes the single vas efferens with its continuation becomes completely emancipated from the kidney tissue and lies in the adult some distance from the anterior end of the kidney. The same phenomenon is seen in Elasmobranchs and Amniotes where the opisthonephric tubules connected with the vasa efferentia never reach the length and complicated convolution of the normal tubule and the Malpighian bodies either degenerate (Seylliwm, Pris- tiurus, Birds) or are eliminated entirely from ontogeny (Skates). On the other hand the simplification of the route from testis to Wolffian duct may come about in a different fashion, as is seen in Amia, where the opening of the vas efferens has become shifted from the Malpighian body down the course of the tubule, in some cases till it has come to open directly into the duct. A careful study of the method by which these various modifica- tions come about during ontogeny is greatly needed. Application of the general principles outlined above seems to afford a probable explanation of the remarkable arrangement in Teleostean fishes, where the testis is continued back into a special sperm duct which opens to the exterior near the opening of the kidney duct (Fig. 143). The presence of a urinogenital network along the whole length of the testis in Ganoids (Acipenser, Lepidosteus) justifies the assump- tion that the ancestral Teleost possessed this primitive arrangement of the network. In the case of Polypterus the testis is continued back as a duct which opens into the urinogenital sinus formed by the hinder ends of the Wolffian ducts. The duct, however, is not, except near its termination, a simple tube but contains a network of cavities continuous with those of the testis. It is in fact not a simple duct but a portion of the testis which has become sterile. Similarly in the case of various typical Teleosts it has been shown (Jungersen, 1889) that the duct is formed by the hinder part of the genital ridge, that it contains for a time a network of cavities con- tinuous with those of the functional testis—that it is in other words the modified and sterile hinder portion of the testis—and, finally, that posteriorly it opens into the Wolffian duct. Now the method by which the condition met with in Polypterus or in the young Teleost has arisen is probably indicated by what has happened in the two Lung-fishes Lepidosiren and Protopterus. In the former the testicular network is reduced to the extent that only about half a dozen vasa efferentia persist at the hind end of the series. In Protopterus these are still further reduced to a single vas efferens which passes from the hinder end of the sterile portion of the testis—“ sperm duct” of the older descriptions—into the hind end of the kidney and communicates with the Wolffian duct through the hindermost kidney tubules. The only further step needed from the condition exemplified by Protopterus to that of Polypterus or the young Teleost is that 282 EMBRYOLOGY OF THE LOWER VERTEBRATES CH. the communication between the sterile or duct portion of the testis and the Wolffian duct should come to be by a direct tubular prolonga- tion of the vas efferens instead of by tortuous kidney tubules. That such a shortening up and simplification of the channel from testis to Wolffian duct does actually tend to come about in evolution is demonstrated by the precisely similar series of modifications which have occurred in the Anura at the front end of the urinogenital network. In these not only has the series of vasa efferentia become reduced to a single (anterior) member in such a form as Discoglossus but the kidney tubule interpolated between it and the Wolffian duct has become shortened and widened, so that there exists simply a single tube leading from the testis and continued at its other end into the Wolffian duct. Giving consideration to these various points it appears to be justifiable to relate the probable evolutionary history of the sperm duct of the Teleost in the following terms :— I. Primitively the elongated testis communicated with the Wolffian duct by way of (a) a series of vasa efferentia distributed along its length, and (d) the tubules of the opisthonephros. IL. The posterior portion of the testis became sterile and functioned merely as a reservoir and duct. III. The vasa efferentia became reduced to those connected with this sterile region and finally to the hindermost one of these. IV. The channel formed by this together with the kidney tubules into which the spermatozoa passed from it became shortened and widened until it reached the condition of a simple tube leading from the hind end of the testis into the hind end of the Wolffian duct. V. The final stage was reached by the opening of this tube into the Wolffian duct becoming shifted back until its opening to the exterior came to be independent. , SUPRARENAL OrGANS.—The organ familiar to students of the Amniota and especially of the Mammalia under the name Suprarenal or Adrenal is now generally recognized as being not a single organ but an organic complex formed by the union of two originally separate elements—the medullary substance and the cortical sub- stance. These two elements arise quite independently in ontogeny, the medullary substance being derived from the sympathetic ganglia, while the cortical substance arises in the form of a number of thick- enings of coelomic epithelium on the roof of the splanchnocoele between the two kidneys. That this independence in early stages of ontogeny is a repetition of a condition which occurred during phylogeny is indicated by the fact that in Fishes the two elements are still independent. The names medullary and cortical substance, referring as they do to a topographical relation which occurs only in mammals, are obviously unsuitable from the point of view of comparative morphology and it is becoming customary to apply other more characteristic names. The medullary substance in mammals and what corresponds with it in other Vertebrates has a very characteristic IV THE SUPRARENAL ORGANS 283 chemical or physical reaction, in that it takes on a deep yellow or brown colour when treated with salts of chromic acid. Hence it is convenient, and usual, to apply to it a name expressive of this reaction— Ca Chromophile (Stilling), Chromaffine (Kohn) or Phaeochrome oll). The cortical tissue has also characteristic features—in particular the fact that its cytoplasm contains numerous granules of lipoid or fat-like substance, soluble in Ether, Xylol, etc., staining deeply with various Aniline stains, and giving the characteristic black with Osmic Acid. For masses of this tissue the name Interrenal organ may be used (Balfour) which although a topographical term like cortical substance has the advantage of being correct for vertebrates in general during at least the early stages of their development. Of the more primitive groups of gnathostomatous Vertebrates only the Elasmobranchs and the Amphibians have been studied carefully in regard to the development of these organs and we shall consequently use them as illustrating the general mode of develop- ment which, with variations in detail, holds throughout the groups dealt with in this volume. ELASMOBRANCHII.1—The Interrenal organs are here interrenal in position through life, forming either one (Sharks) or a pair (Skates and Rays) of elongated bodies lying in the region of the mesial plane and extending for some distance opposite the hinder part of the opisthonephros. In Seyllium (Poll) the interrenal makes its first appearance (7 mm. embryo) in the form of a number of irregularly distributed thickenings of the splanchnic mesoderm in the region of the root of the mesentery, just ventral to the dorsal aorta. The possibility of metameric arrangement in the very earliest stages does not seem to be absolutely excluded but there is no evidence of this so far. The rudiments are most numerous in the genital region but they occur as far forwards as the hind end of the pronephros and back as far as the cloaca. The rudiments of the two sides, projecting towards the median plane, meet and become continuous, and as antero- posterior fusion also comes about, the rudiment takes the form (10 mm. embryo) of a cellular rod lying beneath the dorsal aorta and above the mesenteric root, and for a time still continuous with the splanchnic mesoderm which gave it origin. For a time there is close apposition, amounting to apparent continuity of tissue, between this rod and the opisthonephric nephrotomes lying on either side of it, but it is doubtful whether any special morphological significance is to be attached to this. In embryos of 16-28 mm. in length the interrenal organ gradually becomes separated in a tailward direction both from the coelomic epithelium and the nephrotomes, and assumes its definitive form. Only the tailward part of the series of original rudiments 1 The best general account of the development of the Suprarenal organs is that by Poll (1905). 284. EMBRYOLOGY OF THE LOWER VERTEBRATES cH. completes its development in the way described. The whole series extends through about 25 segments but of these only about the posterior half take part in the formation of the interrenal rod: the anterior ones either atrophy completely or develop into small accessory interrenals. The chromophile organs of the Elasmobranch (Swale Vincent, 1897) are small, rounded, segmentally arranged bodies lying ventral to the intercostal arteries—the anterior few on either side forming a continuous structure which was regarded by the earlier workers as an accessory heart (Duvernoy). These bodies are, as Balfour showed (1878), derivatives of the sympathetic ganglia. In a Scyllium of about 53 mm. the lateral part of the ganglion rudiment begins to show differentiation from the rest, its cells being relatively smaller than those which are destined to become ganglion-cells, and their protoplasm not only staining more deeply with ordinary stains but also developing the characteristic chromic acid reaction. In the Seyllium of 90 mm. the chromophile organ has assumed its definitive rounded form. Intrusive connective tissue forms a sparse stroma and capsule and in the former a capillary network is present. The series of segmentally arranged chromophile masses undergoes much modification in subsequent development—some, particularly at the ends of the series, aborting, others undergoing fusion. The details vary in different genera, the result being a striking variety in the adult arrangements in the various members of the group. AmpuHIpia.—Brauer (1902) in his work on the renal organs of FHypogeophis gives a clear account of the development of the supra- renals. The interrenals appear as in Elasmobranchs in the form of cellular proliferations of the coelomic epithelium, in this case a little external to the root of the mesentery. These proliferations are paired and segmental in their arrangement, and extend from the region of the pronephros-to that of the cloaca. The cellular buds become constricted off from the coelomic epithelium and lie above it as rounded masses embedded in the mesenchyme. As the two posterior cardinal veins approach and fuse the interrenal buds become displaced upwards so as to lie between the cardinal vein and the dorsal aorta. As development goes on processes of fusion take place between the rudiments more especially anteriorly where they come to form an unpaired elongated mass lying below the dorsal aorta and for the most part dorsal to the posterior vena cava (i.e. the fused posterior cardinals) but here and there passing laterally round the vein to its ventral surface or even piercing it— the fusion between the two cardinals having been obstructed at such points. In the posterior half of the organ the several rudiments retain their distinctness and lie on the ventral face of the opistho- nephros. The chromophile bodies develop as in the Elasmobranchs from split off portions of the sympathetic ganglion rudiments. These Iv THE SUPRARENAL ORGANS 285 become shifted in a ventral direction round the dorsal aorta and take up their position in intimate contact with the interrenal bodies, lying in the posterior paired region of the interrenal on its mesial face, elsewhere dorsal or lateral or even completely surrounding the ent Fic, 144,—Illustrating origin of the sclerotome as seen in transverse sections through young stages of A, Amphiovus (for the sake of clearness the spaces have been exaggerated, and the ventral portion of the diagram is taken from a rather younger stage than the dorsal); B, Lepidosiren, stage 24; C, Polypterus, stage 23. ent, enteric cavity; my, myotome; N, notochord; ne, nephrocoele; ym, pronephros ; s.c, spinal cord ; scl, sclerotome ; sn, subnotochordal rod ; sple, splanchnocoele. interrenal. The chromophile bodies stand out with great distinct- ness from the interrenal by their fine grained deeply-staining proto- plasm and their larger nuclei. _ a We thus see that in the Amphibia the originally separate inter- renal and chromophile bodies become during the course of develop- 286 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. ment associated tovether to form a suprarenal complex of the type seen in the higher Vertebrates. Incidentally the unsuitability of the terms medullary and cortical is accentuated, for here when one of the elements comes to surround the other it is the chromophile which does so—precisely the opposite to what happens in the Mammalia. THE ScLEROTOME—In Amphioxvus the sclerotome (Fig. 144 A, sel) arises as a pocket-like diverticulum of the splanchnic mesoderm just ventral to the myotome. It grows inwards and dorsalwards, pushing its way between the notochord and spinal cord on the one hand and the myotome on the other, until it reaches the mid-dorsal line where it meets its fellow of the. opposite side. The epithelial walls of the sclerotome finally break up into mesenchyme—amoeboid connective tissue cells. The cells derived in this way from the outer wall of the sclerotome apply themselves to the mesial face of the myotome, penetrating in between its muscle cells and forming septa of connective tissue between adjacent myotomes, while those derived from the inner wall go to form packing tissue in the inter- stices round spinal cord and notochord. Over the spinal cord this packing tissue forms a tough protective roof. During this resolution of the sclerotomes into mesenchyme all trace of the original segmental character of the sclerotomes disappears. It is customary — although the present writer regards it as questionable whether this is wholly justified—to regard the mode of origin of the sclerotome seen in the developing Amphioxus as representing the primitive mode of development. Upon this assump- tion we may describe what takes place in the typical Vertebrates as follows. The breaking up of the sclerotome into mesenchyme tends to take place at earlier and earlier periods of development—the diverticulum stage becoming more and more transient and eventually disappearing completely so that sclerotome formation comes to be represented merely by a very active proliferation of mesenchyme cells from the splanchnic surface of the mesoderm ventral to the myotome (cf. Fig. 144 C, se/). It must not be supposed that the whole of the connective tissue in the body is necessarily derived from the sclerotome. On the contrary it would appear that other regions of the mesoderm also give rise to mesenchyme cells. Thus the inner surface of the splanchnic mesoderm of the gut-wall would appear to give rise to the connective tissue of this region, and the whole of the splanchnocoelic mesoderm of the postanal region apparently becomes resolved into mesenchyme. On the whole perhaps the safest position to take up is that of regarding the power of forming mesenchyme as a general property of the mesoderm, and of regarding the sclerotome merely as expressing a localized concentration of this power, rather than as being the representative of some primitive pocket-like diverticulum of unknown function. IV THE COELOMIC ORGANS 287 LITERATURE Abramowicz. Morph. Jahrb., xlvii, 1913. Agar. Traus. Roy. Soc. Edin., xlv, 1907. ot Balfour. Monograph on the development of Elasmobranch Fishes. London, 1878. Balfour and Parker. Phil. Trans. Roy. Soc., clxxiii, 1882. Balfour and Sedgwick. Proc. Roy. Soc,, xxvii, 1878. Bles. Proc. Roy. Soc., lxii, 1897. Bouin. Arch. de Biol., xvii, 1901. Boveri. Anat. Anz., vii, 1892. Brauer. Zool. Jahrb. (Anat.), xvi, 1902. Braus. Morph. Jahrb., xxvii, 1899. Budgett. Trans. Zool. Soc. Lond., xvi, 1902. Dahlgren and Kepner. Animal Histology. New York, 1908. Ewart. Phil. Trans. Roy. Soc., 179 B, 1888. Ewart. Phil. Trans. Roy. Soc., 179 B, 1889. Ewart. Phil. Trans. Roy. Soc., 183 B, 1892. Felix. Anat. Hefte (Arb.), viii, 1897. Felix. Hertwigs Handbuch der Entwicklungslehre, iii, 1904. Field. Bull. Mus. Comp. Zool. Harvard, xxi, 1891. Furbringer. Entwick. der Amphibienniere. Heidelberg, 1877. (See also Morph. Jahrb., iv, 1878.) — Goodrich. Quart. Journ. Mier. Sci., xxxvii, 1895. Guitel. Bull. Soc. Sci. et Méd. de l'Ouest, x, 1901, and xi, 1902. Gregory. Semons Forschungsreisen. i, 1905. Hochstetter. Morph. Jahrb., xxix, 1900. Jungersen. Arb. zool.-zoot. Inst. Wiirzburg, ix, 1889. Jungersen. Zool. Anzeiger, xvi and xvii, 1893, 1894. Kerr, Graham. Proc. Zool. Soc. Lond., 1901. Kerr, Graham. The Work of John Samuel Budgett. Cambridge, 1907. Kerr, Graham. Quart. Journ. Micr. Sci., liv, 1910. Koltzoff. Bull. Soc. Imp. Nat. Moscow, xv, 1901. Lankester. Quart. Journ. Micr. Sci., xvii, 1877. Lebedinsky. Arch. mikr. Anat., xliv, 1895. Marshall. Vertebrate Embryology. London, 1893. Marshall and Bles. Studies Biol. Lab. Owens College, ii, 1890. Mollier. Anat. Hefte (Arb.), iii, 1893. Miller, E. Anat. Hefte (Arb.), xliii, 1911. Nussbaum. Arch. mikr. Anat., xxvii, 1886. ‘Poll. Hertwigs Handbuch der Entwicklungslehre, iii, 1905. Poole. Proc. Zool. Soc. Lond., 1909. Rabl, ©. Morph. Jahrb., xxiv, 1896. Rabl, H. Arch. mikr. Anat., lxiv, 1904. Rosa. Archivio zoologico, iii, 1906. Rickert. Arch. f. Anat. u. Entwick., 1888. Schapitz. Arch. mikr. Anat., xxix, 1912. Schmalhausen. Zeitschr. wiss. Zool., C, 1912. Schreiner. Zeitschr. wiss. Zool., 1xxi, 1902. Sedgwick. (uart. Journ. Micr. Sci., xx, 1880. Sedgwick. Quart. Journ. Micr. Sci., xxi, 1881. Semon. Keibels Normentafeln. iii, 1901. (Also in Zool. Anzeiger, xxiv, 1901.) Semper. Arb. zool.-zoot. Inst. Wiirzburg, il, 1875. Vincent, Swale. Trans. Zool. Soc. Lond., xiv, 1897. Wijhe, van. Uber die Mesodermsegmente und die Entwicklung der Nerven des Selachierkopfes. Verhand. Akad. Wet. Amsterdam, xxii, 1883. (Reprinted, Gron- ingen, 1915). Wijhe, van. Arch. mikr. Anat., xxxiii, 1889. Zarnik. Zool. Jahrb. (Anat.), xxi, 1904. CHAPTER V THE SKELETON Tue skeletal tissues of the animal body show a variety which is at first sight quite bewildering. Closer scrutiny however reveals certain general principles which are at work. In a very restricted set of cases we see that the supporting structure consists of a row or rod of cells which is rendered stiff through the individual cells being blown out or distended with fluid. Such turgor of cells is a far less con- spicuous feature in the animal kingdom than it is in the vegetable. It is well seen in the axial row of endoderm cells which supports the tentacles of the Hydrozoa. In the Vertebrate it is seen im the notochord. Far more usually the support is given by a definite supporting substance with such physical qualities as rigidity, tensile strength, elasticity, as may be required in the particular case. These supporting substances of the animal body again show the greatest variety in their morphological nature but they may all be classed between two extremes—in one of which the supporting substance consists clearly of modified cells or portions of cells and in the other of dead intercellular substance. Examples of the former are seen in the remarkable phagocytic organs of nematode worms where an enormous cell becomes developed into an immensely com- plicated branched structure of stiff horny consistency upon the terminal twigs of which are perched innumerable minute blobs of phagocytic protoplasm. A good example of the second type is seen in the skeleton of ordinary coral—a mass of hard calcareous material lying clearly outside the limits of the living cells. It is necessary to emphasize the fact, which is frequently lost sight of, that the differences between these two types are superficial rather than fundamental. They are merely the extremes of a series and are connected up by innumerable intermediate conditions. The skeleton of an Arthropod such as a Lobster is in the early stages of its development simply the stiffened and hardened outer layer of the cytoplasm of the ectoderm cells, while in its latest stage, immediately before it is shed, it has become a thick layer of dense chitinous and calcified non-living substance lying outside the limits of the living protoplasm. 288 OH. ¥ THE SKELETON 289 Non - living material “secreted” by cells consists no less of modified cytoplasm, although here the cytoplasm so modified does not form a continuous mass and retain its original position in regard to the rest of the cell body. It arises commonly as isolated droplets or particles which may secondarily run together within the body of the cell or, without this happening to any obvious extent, are extruded from it, passing on to a free cell surface or into the intercellular matrix. The process can be followed by observation, naturally, only in cases where the secretion runs together into discrete droplets or particles sufficiently large to be visible under high powers of the microscope, as commonly happens in gland-cells. Far more frequently the extruded particles are so small—possibly molecular— as completely to elude observation. Such is the case where the inter- cellular substance undergoes skeletal modification: all that can be observed is simply a gradual transformation in the physical and chemical characters of the matrix, due in some cases to a gradual change in the metabolic activities of the cells which inhabit it, in others to the immigration into it of cell-colonists of a new type. The supporting skeleton of the Vertebrate is an endoskeleton ;+ it is developed not on the outer surface of the body but within its substance. In this it contrasts with the skeleton of the Arthropod or Molluse which is exoskeletal—consisting of thickened and stiffened cuticle. In the case of the most ancient skeletal structure in the Vertebrate body—the notochord—the stiff supporting character is due to the individual cells being distended by fluid secreted in their interior but as a rule in other skeletal tissues the stiffness is given not by the cells but by the intercellular matrix. THE NOTOCHORD AND ITS SHEATHS.—A comparative study of the Vertebrate skeleton shows that it illustrates three phases of evolu- tionary progress (1) the notochordal phase, (2) the cartilaginous or chondral phase and (3) the bony or osseous phase. Of .these the primitive is indisputably the first. It is a phase which is passed through during ontogeny in all Vertebrates and it remains permanent throughout life in Amphiozus. The notochord is in its origin a rod of cells split off from the endoderm along its mid-dorsal line. This is seen in all the lower Vertebrates. In some of the more primitive members of the group the notochordal rudiment is for a time deeply grooved on its lower side, so as to form an inverted gutter along the middle of the enteric roof, and it may well be that this is to be regarded as the primitive mode of formation of the organ. The notochord becomes constricted off as a cylindrical rod extending along the dorsal side of the alimentary canal from a point just behind the tip of the infundibulum to the tip of the tail. The 1 It is regrettable that the term exoskeleton has crept into use by writers on Vertebrate anatomy for structures such as fish-scales. As will be seen later these are really endoskeletal, even the enamel being: developed on the inner surface of the epidermis. ; VOL. II U 290 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. individual cells develop in their cytoplasm fluid vacuoles which increase in size and become confluent until at last the cell takes the form of a comparatively thin layer of protoplasm surrounding an enormous vacuole and containing embedded in its substance at one point the nucleus. The turgescent condition of the cells inflated with fluid gives them the firmness which enables the notochord to carry out its function as a supporting structure. The inflation of the cells with fluid carries with it another result namely a great increase in size of the individual cells. This in turn causes a great increase in size of the notochord as a whole, showing itself particularly by increase in diameter but also by increase in length. The latter is not able to take place with perfect freedom and the result is that the individual cells tend to be compressed into the form of transverse discs. The notochordal rudiment at an early stage becomes covered by a thin, elastic, highly refracting, cuticular formation known as the primary sheath of the notochord (“elastica eaterna”). After the formation of this the superficial layer of the notochord soon resumes its cuticle-forming activity but now in a somewhat moditied form—a secondary or “fibrous” sheath, thicker and more jelly-like in appearance, being produced internal to the primary sheath. In Cyclostomes and Sturgeons this secondary sheath remains through- out life without conspicuous change beyond increase of thickness and the assumption of a tough fibrous character and is physiologically the most important part of the axial skeleton of the trunk region. The superficial layer of notochordal cells, lying in immediate contact with the inner surface of the secondary sheath, do not as a rule undergo the process of vacuolation which affects the inner cells. They remain as a layer of compact protoplasmic cells known as the notochordal epithelium. In some Vertebrates (Dipnoi, Agar, 1906) a short stretch of noto- chord, from the tip backwards, degenerates within its primary sheath at an early stage, breaking up into loose mesenchyme. As the notochord behind the degenerated portion grows in length its front end is pushed forwards so as to re-occupy the vacated portion of primary sheath. The extent to which this process takes place throughout the Vertebrata in general, and also its meaning, are deserving of further enquiry. Hyeocnorpd (Subnotochordal Rod).—In the anamniotic Verte- brates there is formed what is apparently an accessory notochord lying ventral to the true notochord and hence known as the Hypo- chord or Subnotochordal rod. This organ (see Gibson, 1910) arises after the notochord and in an entirely similar manner, i.e. as a longitudinal rod of cells split off from the endoderm in the mid- dorsal line and sometimes possessing a distinct groove along its lower surface facing the enteric cavity. On its surface it normally develops a primary sheath precisely like that of the notochord. We may be sure, from its wide distribution amongst the more Vv THE SKELETON 291 primitive Vertebrates (Lampreys, Elasmobranchs, Teleostomes, Dipnoi, Amphibia) and the early stage of development at which it appears, that the hypochord is an organ of great antiquity in the Vertebrate stem, but we have no definite knowledge of its ancestral significance. The fact that it does not occur in Amphiorus has rendered possible the suggestion that it represents the longi- tudinal groove which in this animal runs along the mid-dorsal line of the pharyngeal wall. But this idea is negatived by the fact that the hypochord- extends right back to the tail and is uot merely a pharyngeal organ, and the probability seems to be that it has come down from a period in evolution long before the appearance of Amphioxus. It is perhaps simplest to regard it merely as an accessory notochord. Whereas the true notochord plays an important physiological role—as the main part of the axial skeleton during early stages, and as the foundation for the vertebral column of later stages—the hypochord has no such justification for its persistence. It lasts only for a short time and eventually breaks up and completely disappears. In the Amniota there is no typical hypochord developed but it is possible that a thickening of the mid-dorsal endoderm which is frequently found in the pharyngeal region (¢g. in the Second day Fowl embryo) may represent a last vestige of it. SKELETAL DEVELOPMENTS OF THE CONNECTIVE TISSUE Whereas the notochord is derived directly from the endoderm, the cartilaginous and bony components of the skeleton on the other hand are modifications of the mesenchyme or connective tissue, which forms a considerable proportion of the entire bulk of a typical vertebrate. Connective tissue in its least specialized form may be seen in practically any late vertebrate embryo as a reticulum or spongework—a syncytial framework—of much-branched cells, the processes of which are continuous from one cell to another, while the meshes are occupied by a clear fluid or jelly-like matrix. Masses of this tissue form a kind of packing between and around the various epithelia of the body, while it also, in the form of discrete wandering cells, actually invades the epithelial tissues and colonizes them. Such immigrant elements are found for example between the muscle-fibres, in the substance of the central nervous system, and even frequently between the epithelial cells of the epidermis. The primitive or embryonic connective tissue undergoes gradual differentiation in accordance with the physiological rdle which it has to play in different localities. This differentiation finds expression in such superficial features as shape and arrangement of the individual cells and more fundamentally in the peculiarities in metabolism which lead to its storing up particular substances in its protoplasm 292 EMBRYOLOGY OF THE LOWER VERTEBRATES cu. —pigment of chromatophores, fat of the cells of adipose tissue—or again in the influence exerted by the metabolic activity of the cell upon the character of the matrix. This matrix is commonly described as intercellular, which is quite correct, but the important point is not the question whether it is inter- or intra-cellular but the fact that it is in immediate relation to, and under the influence of, the living protoplasm of the cell. The portion of matrix in con- tiguity with one of the irregularly shaped connective-tissue cells fis comparable with an intracellular vacuole the outer wall of which has thinned out and disappeared. The matrix has been formed by the breaking down of living substance and it seems merely a matter of phraseology whether we speak of it as modified protoplasm or as dead “formed” material. The most familiar differentiation of the matrix of connective tissue consists in the development within it of thin tough fibres, characterized by the physical property that they soften and dissolve, yielding gelatin, under the action of boiling water, and that they become further toughened by the action of tanning agents. These fibres run indiscriminately in all directions or, in the more specialized conditions, are definitely orientated, as in the case of tendon where they are parallel and arranged in longitudinal strands, or of aponeuroses where they are arranged in thin layers, those of one layer perpendicular to those of the next. Other portions of the matrix take the form of elastic fibres—characterized by their elasticity, by their connexion together to form a network, by their being much less easily affected by boiling water, and by their not yielding gelatin. The amount of matrix present differs greatly in different localities. It may he reduced to a very small amount—to a mere demarcating line—between closely fitting plate-like cells, as in the case of the endothelium covering the surface of a tendon, or it may be large in © amount and comparatively rigid as in the case of the two great skeletal tissues cartilage and bone. CARTILAGINOUS OR CHONDRAL SKELETON.—The cartilage is char- acterized by its cells taking on a rounded form and becoming separated by an abundant semitransparent, elastic, chondrin-containing matrix. The process of chondrification becomes apparent first of all in the somewhat dense packing tissue (“skeletogenous layer”)! immediately surrounding the notochord. This connective tissue becomes locally modified to form little blocks of cartilage known as the arch-elements (arcualia—Gadow, 1895), lying just outside the primary sheath and arranged in four longitudinal rows, two dorsal composed of the rudiments of the neural arches, two ventral—the rudiments of the haemal arches. These arch-elements are apparently in the primitive condition duplicated in each segment, te. within the limits of a myotome or sclerotome there are situated two pairs of neural and two pairs of haemal arch-elements. 1 The term prochondral is applied to the young cartilage in its early stages before the characteristic intercellular matrix makes its appearance. v THE SKELETON 293 There now takes place in two of the more primitive groups of Vertebrates—the Elasmobranchii (including the Holocephali) and the Dipnoi—a remarkable process whereby the secondary sheath of the notochord becomes converted into a sheath of cartilage. Certain of the cartilage cells in the arch rudiment take on an amoeboid character and burrowing their way through the primary sheath, apparently by the help of a digestive ferment, invade the secondary sheath (Fig. 145, m.c). Continuing their migration they become distributed equally throughout the whole substance of the secondary sheath, including those portions in the head re- gion which will later on form part of the cranium. The immigrant cells fin- ally settle down in the substance of the second- ary sheath and the latter becomes a cylinder of cartilage. It is important, with an eye to the evolution of the vertebral column in Vertebrates higher in the scale, to bear in mind that this invasion of the secondary sheath by im- Fic. 145. — Part of a transverse section through a migrant cartilage cells Lepidosiren of stage 38, traversing one of the neural takes place at four points rch rudiments, in the transverse plane, Ce, notochordal epithelium ; mc, migrating cartilogt: cell 5 corresponding to the bases See a neural arch; sl, primary sheath; s%, of the four arch rudi- ments, and that this arrangement is repeated twice within the limits of one segment owing to the arch rudiments being so repeated. Consequently if we suppose the colonization of the secondary sheath to be restricted to the neighbourhood of the trans- verse plane in which the arch rudiments are situated the result would be the formation of two rings of cartilage within the limits of a single segment. In the case of Lung-fishes and Holocephali the chondrified secondary sheath undergoes no further modification but in typical Elasmobranchs it becomes divided up into segments, which form the centra or bodies of the vertebrae, in the manner to be described later on. In this process the originally uniformly flexible notochord with its sheaths becomes replaced physiologically by a series of rigid masses, flexibility being given to the whole by the presence of the intervening joints. As this jointed condition of the vertebral column originated in evolution at a time when the longitudinal muscles of the body were already divided into myotomes, we may 294 EMBRYOLOGY OF THE LOWER VERTEBRATES CH. take it as probable, for obvious mechanical reasons, that the rigid skeletal masses arose in a position alternating with the muscle segments. The individual vertebral centra were in other words from the beginning intersegmental in position in relation to the general body metamerism. In sketching out in somewhat greater detail the further develop- ment of the vertebral column the assumption will be again made use of, as it was in dealing with the mesoderm segments, that the trunk region has in all probability departed least from the primitive con- dition, and the facts quoted will in the main be taken from this region of the body. ; ; ; The student who goes on to peruse original memoirs will notice that this rule is by no A UB. Saye ie A. B. meansalwaysaccepted. Some an a i § “ writers will be found to assume that the caudal re- gion is more nearly primitive, and, in accordance with this assumption, to interpret the phenomena observed in the trunk vertebrae by those observed in the caudal, in- Fic, 146.—Arrangement of dorsal arch-elements in stead of wee oes . hinder trunk region of a Petromyzon larva 95 In this connexion 1¢ must mm, in length. (After Schauinsland, 1906.) be borne in mind that the A, anterior, B, posterior neural arch-elements ; d.7, dorsal Vertebrate is above all rot of spina nerve suraceataskchenl mE“EE. things essentially a coelomate * animal. No one doubts that whatever the common ancestor of the Vertebrates was like it was at least coelomate. And most morphologists would admit further that the weight of evidence indicates that in this ancestor the splanchnocoele extended throughout the greater part of its length and that the existence of a considerable stretch of body towards the hind end devoid of splanchnocoele (@.e. a tail region) is secondary. But if the caudal region has in this way undergone profound secondary modification of its structure it is clear that it is not in this region of the body that we should expect to find persisting primitive modes of development of the axial skeleton. It is now necessary to follow out the fate of the arch-elements. As already mentioned the primitive arrangement of these appears to have been two pairs to each segment, above and below, so that corre- sponding with each myotome there were, on each side, two neural elements an anterior (A) and a posterior (B), and two haemal elements an anterior (@) and a posterior (0). NevurAL ArcuEs.'— Apparently the most nearly primitive arrange- 1JIn writing these sections on the vertebral column much use has been made of Schauinsland’s descriptions (1906) to which the student is referred for a more detailed account than is here possible. v NEURAL ARCHES 295 ment of the arches is that which occurs in the hinder trunk region of the Lamprey (Fig. 146). In this animal, as is well known, the dorsal (sensory) and ventral (motor) nerve-roots are still separate and are spaced out alternating with one another at approximately equal distances along the sides of the spinal cord. The dorsal arch elements alternate, in their turn, with the nerve-roots, so that there are, on each side, an anterior (A) and a posterior (B) neural arch- element within the limits of a single myotome.' It should be noticed particularly that of these the anterior is situated between the sensory and the motor nerve-root belonging to the segment. This suggests a possible explanation of the later evolutionary history of these cartilages (A) which in the typical Fishes tend very usually Tax ab : bob & Fic. 147.—A, arrangement of arch-elements in mid-trunk region of a Carcharias embryo 85° mm. in length ; B, do. in anterior region of a Sturgeon (Acipenser huso) 36 mm. in length. (After Schauinsland, 1906.) A, anterior neural arch-element ; B, posterior do. ; a, anterior haemal arch-element ; b, posterior do. ; d.v, sensory nerve-root ; v, blood-vessel ; v.r, motor nerve-root. to become reduced in size, even to the point of disappearance. It may be that this reduction in size is connected with the fact that in the Fishes, as indeed in all gnathostomatous vertebrates, the two: nerve-roots have become approximated together to form a common sensori-motor spinal nerve. On the other hand this explanation would leave untouched the fact that a similar reduction in size may occur in the corresponding ventral or haemal arches. The reduction in size of the “A” elements, which is of so frequent occurrence amongst the typical fishes, is well shown in Figs. 147, A and B, which are based upon Schauinsland’s recon- structions. This marked reduction is by no means of universal occurrence. The two common Dog-fish—Seyllium and Acanthias— are familiar examples of fishes in which the “A” elements 1 In the anterior trunk region the arrangement is apt to be modified—the inter- - segmental vessel, which forms the anterior limit of the segment, coming to lie on the tailward side of the A cartilage of that segment (Schauinsland). 296 EMBRYOLOGY OF THE LOWER VERTEBRATES — cn. (“intercalary pieces,” “ interdorsals”) remain nearly as well developed in the adult as the “B” elements." _ In Lung-fishes (here and there) and in Urodele amphibians the “A” pieces can still be recognized (cf. Fig. 148); they have also been observed in the embryos of various Reptiles. In this case they usually lose their individuality at an early period, becoming completely merged in the definitive neural arch formed by the “B elements lying next to them on their headward side, but in some cases, ¢.g. in the tail region of Lacerta, they have been found to persist as discrete structures even in the adult, forming a vestigial second neural arch behind the main arch. ; The neural elements become prolonged dorsally and meet so as to form a complete neural arch and the apex of this becomes pro- longed as an unpaired piece in the mesial plane to form the neural spine. The complete neural arch formed in this way frequently becomes segmented up into separ- ate pieces of cartilage. The arcualia in such cases become each divided into a larger basal (basi- dorsal—B, interdorsal—A,Gadow ) and a smaller apical (supradorsal) portion. The spine may segment into three superimposed rod-like portions. HarmMaL ArcHzes. —In_ the Fic. 148.—Arrangement of arch-elements lostomes typical haemal arches in anterior caudal region of a Siredon 50 Cye ce yP mm. in length. (After Schauinsland, ieee absent, although possibly 1906.) vestiges of them are represented Reference letters as in Fig. 147. by a continuous ridge of cartilage occurring in the tail region of Petromyzon where the neural arches have also been reduced to a similar continuous ridge (Schneider). Of haemal arch elements there were apparently primitively two pairs to a segment just as in the case of the neural arches. This seems to be clearly indicated by Callorhynchus (Fig. 149). It is also well shown in the young Sturgeon (Fig. 147, B) where the anterior element (a) in each segment has undergone reduction in size exactly as was the case with the corresponding neural element (A).