I ^Oo TEXT- BOOK OF THE EMBRYOLOGY OF INVERTEBRATES In Preparation. Parts 11. and III. of Drs. KORSCHELT and HEIDER'S "TEXT-BOOK OF THE EMBRYOLOGY OF INVERTE- BRATES," Translated and Edited by H. J. CAMPBELL, M.D., Senior Demonstrator of Biology and Demonstrator of Physiology in the Medical School of Guy's Hospital. Swan Sonnenschein & Co., Ld., London ; Macmillan & Co., New York. TEXT-BOOK OF THE EMBRYOLOGY OF INVERTEBRATES BY Dr. E. KORSCHELT Professor of Zoology (Sr» Comparative Anatomy in the University of Marburg Dr. K. HEIDER Professor of Zoology in the University of Berlin Translated from the German BY EDWARD L. MARK Ph.D W. Mc M. WOODWORTH Ph.D Instructor in Microscopical Anatomy in Harvard University With Additions by the Authors and Translators Hersey Professor of Anatomy in Harvard University PART I PORIFERA, CNWAA'JA, CTENOPHORA, VERMES, ENTEROPNE USTA, ECHINODERMA TA XouDon SWAN SONNENSCHEIN & CO., Limp NEW YORK: MACMILLAN & CO 1895 640078 9, ^.5Xy L TRANSLATORS' PREFACE The value of the Lehrhuch der vergleichsnden Entwicklungs- geschichte der ivirhellosen Thiere for students of animal morphology is too well understood by those who are familiar with its scope and execution to require any statement of our aims in undertaking an English translation of it. In presenting to zoologists the First Part of this work we consider ourselves fortunate in having had the valuable aid of the authors in supplementing the original text by numerous additions, made desirable by the rapid advance of the science since the date of first publication. Although the scope of the work has permitted the addition of only the most succinct statement of the results reached by embryo- logists in the last five years, these additions must prove to be of assistance to all students, and will, we believe, be especially acceptable to those who are already familiar with the original edition. In order to spare the reader the labor of comparing original and translation for the purpose of ascertaining what is new, the plan has been adopted of enclosing in brackets [ ] all new matter, which, so far as practicable, has been put in the form of footnotes. Each of these addi- tions is followed by the initial of the author, or by the word " Translators," to indicate the persons responsible for the new matter. Owing to an oversight, the initial has been omitted from several of the additions in the earlier chapters. It should be stated, therefore, that, unless otherwise indicated, the additions to Chapters I. — III. were made by Professor Heider, those to Chapters IV. — XIV. by Professor Korschelt. Brackets have also been freely used in the text to enclose such words or brief explanations as the translators deemed VI TRANSLATORS PREFACE useful supplements to the more literal translations of the original. In such cases an indication of the authority has been omitted, since no uncertainty is likely to result from the omission. To avoid confusion in citation and to indicate at a glance the additions to the Literature of the several chapters, the references not included in the original have been put in the form of Appendices and numbered with Roman numerals. It has been the aim to make these additions include all the important papers which have appeared since this Part was first issued. In translating Anlage we have employed the word funda- ment— a use which one of us has suggested and defended in the Translatoi's' Preface to Text-hook of the Emhryology of Man and Mammals, by Dr. Oscar Hertwig, etc. (Swan Sonnenschein & Co. : London, 1892). We are under deep obligation to our colleagues Doctors C. B. Davenport and G. H. Parker for their friendly and self-sacrificing assistance, and we desire to thank both of them for their aid — Dr. Davenport for having rendered us valuable service in revising the whole of the manuscript ; Dr. Parker for assistance in revising parts of the manuscript and reading the whole of the proof. It is with reluctance that we have felt compelled by the pressure of other duties to relinquish to others the task of completing the translation of this admirable work. We trust that one of the advantages of this change will be the more rapid publication of the translation of the remaining parts than could possibly have been hoped for from us. THE TRANSLATORS. Cambridge, Mass., U.S.A. AUTHORS' PREFACE The facts that the comparative embryology of Invertebrates has not had a broad and comprehensive presentation since the appearance of Balfour's Treatise on Comparative Embryology, and that the special literature of this subject has undergone an enormous increase since that time, have forced upon every one who has been concerned with questions of com- parative embryology the pressing need of a more modern treatment of the subject. Inasmuch as we had occasion to go over a considerable part of the literature of this subject during the last few years — partly for the purpose of courses of lectures to be given, partly from the requirements of special investigations — it was natural that the idea should have occarred to us to utilize this preliminary work, and by arranging the material acquired and further elaborating it to issue the whole in book form — a venture which was undertaken, and the first results of which have assumed the form of the present part. Since it has been our plan in writing the present work to proceed from the special to the general, and since naturally some time will elapse before the completion of the whole, we have thought that we should secure the gratitude of the reader if we published the first half of the special part at once. The second half of this part, embracing the Arthro- pods, Molluscs, MoUuscoidea, Tunicates, and Amphioxus, will appear shortly, while we hope to finish the general part, and therewith the whole book, in the course of the year 1890. Not to begin the special part of this work too abruptly, we have prefaced it with a short general introduction. Our decision to limit the subject matter to the so-called VIU AUTHORS PREFACE invertebrate animals may require an explanation, and perhaps also an excuse. We have been guided exclusively by prac- tical considerations, especially the fact that the comparative embryology of Vertebrates has very recently been compre- hensively treated in an excellent manner, and further the reflection that, with the limitation of our field to the In- vertebrates, the treatment of them might be so much the more thorough. We take the liberty of expressing here our best thanks to Herrn Geheimrath F. E. Schulze, who has aided us in the most amiable manner, both by his advice and by his assist- ance in procuring literature; and likewise to our publisher, who has made it possible to issue the book in its present form. THE AUTHORS. CONTENTS OF PART FIRST PAGB Introduction 1 Types of cleavage and gastralation 3 Formation of the mesoderm 10 Body cavity 11 Chapter I. Porifera. By K. Heider 13 I. Amphiblastula type 14 II. Coeloblastula type 20 III. Parenchymula type 22 Development of Spongilla 25 General considerations on the development and sys- tematic position of the Porifera 28 Origin of the different types of canal system . . 30 Non-sexual reproduction of sponges .... 32 Development of the gemmula 34 Literature 36 Chapter II. Cnidaria. By K. Heider 39 Systematic 39 I. Hydrozoa 39 1. Hydroidea 39 Development of metagenetic Medusae ... 40 Budding of Hydroid polyps 43 Development of the Medusa 43 Comparison of polyp and Medusa. ... 45 Budding and division in Medusae. ... 48 Frustulation 49 Development of polyps with sessile gonophores . 50 Development of hypogenetic Medusae ... 53 Developmental cycle of the Cuninas ... 58 2. Siphonophora 59 Systematic • . .59 (a) Physophoridse 60 Disconula larva of the Velellidse . . 65 Laws of budding in the Siphonophore stock 66 (h) Calycophoridse 68 General considerations on the origin of the Siphonophora 72 ix X CONTENTS PAGB II. Anthozoa 75 (a) Alcyonaria 75 Cleavage, and the development of the planula 76 Attachment and development of the polyp . 77 Budding 82 (b) Zoantharia 88 Cleavage and development of the planula . . 88 Development of the septa 90 Formation of the calcareous skeleton . . 98 Division and budding 100 III. Scyphomedusae 102 («) Lucernaridse 102 (6) Charybdeidse 103 (c) Discophora 103 Alternation of generations in the Discophora . 105 Development of the Scyphistoma . . . 105 Strobilation 113 Hypogenetic development of Pelagia . .117 Metamorphosis of the Ephj'-ra .... 119 General considerations regarding the Scypho- medusae 122 General considerations regarding the Cnidaria 125 Literature 127 Chapter III. Ctenophora. By K. Heider .... 136 Tectonic 136 Embryonic development 138 Cleavage 139 Epiboly 142 Formation of the gastro vascular system 146 Development of the mesoderm 148 Metamorphosis 151 Heterogeny 153 General considerations 154 Literature 158 Chapter IV. Platyhelminthes. By E. Korschelt . . . 159 L Turbellaria 159 Systematic survey ; different forms of development . 159 1. Polycladida 160 A. Direct development. Oviposition; cleavage 160 Formation of the germ-layers . . . 161 External and internal development of the body 163 B. Indirect development 165 Miiller's larva ; transition to the adult animal 165 Aberrant larval forms (Oligocladus, Sty- lochus pilidium) 167 CONTENTS XI PAGB 2. Tricladida. Oviposition ; yolk-cells ; cleavage . 169 Formation of the embryo and its organs . . 169 3. Rhabdocoelidae 173 [Acoela] . 175 General considerations : relation of the Tur- bellaria to the Ctenophora (Cteno plana and Coeloplana) 176 11. Trematoda 178 Composition of the egg from egg-cells and yolk-cells . 178 1. Distomidse 178 Embryonic development 179 Further course of development; Distomum hepaticum 180 Sporocyst ; E,edia ; Cercaria .... 180 Differences in the development of various Distomidae 185 Cercaria setifera, Bucephalus, Leucochlori- dium, etc. 186 2. Polystomidse 188 Diplozoon, Gyrodactilus, etc 189 III. Cestoda 190 Character of the eggs ; resemblance to those of Tre- matodes 190 Embryonic development of the Bothriocephalidse . 191 Embi'yonic development of the Tseniadse . . . 193 Further development ; transference of the eggs ; cysticercus 194 Formation of the scolex ; transference of the cysti- cercus ; development of the tapeworm . . . 197 General considerations : significance of the develop- ment (Archigetes, Caryophyllseus, Amphiptyches, etc.) 198 Taenia coenurus and T. echinococcus ; relation to the Trematodes (Amphilina) and Turbellaria . . 200 Literature 201 Chapter V. Orthonectid^ and Dicyemid^. By E. Korschelt 206 I. Orthonectidae 206 Systematic survey ; occurrence ; shape of the body ; organization; eggs 207 Development of the male 208 Development of the female 209 II. Dicyemidae 209 Systematic occurrence ; organization .... 209 Development of the vermiform embryos . . . 210 Structure and development of the inf usoriform embryos 211 Xll CONTENTS PAGE General considerations: interpretation of the Ortho- nectidse, etc 215 Literature 215 Chapter VI. Nkmertini. By E. Korschelt .... 217 Oviposition ; different forms of development .... 217 1. Development through the Pilidium larva . . . 217 Blastula ; gastrula ; Pilidium 218 Origin of the worm in the Pilidium .... 221 2. Development after the type of Desor .... 224 3. Direct development 227 General considerations : interpretation of the various forms of development 229 Belationships of the larvae and adult animals to other groups 230 Literature 232 Chapter VIL Nemathelminthes. By E. Korschelt . . . 234 L Nematoda . 234 Embryonic development 234 Cleavage ; formation of the germ-layers .... 234 Formation of the embryo 237 The post-embryonic development ; Trichocephalus ; Heterakis 239 Dochmius; Mermis 239 Sphserularia ; Atractonema 240 Heterodera ; Allantonema 241 E/habditis nigrovenosa ; Rhabdonema ; Cucullanus . 242 Dracunculus ; Spiroptera ; Trichina .... 243 IL Gordiidse 244 Embryonic and subsequent development . . . 244 General considerations : Gordiidae, Nematoda, and Acanthocephali 246 Literature 247 Chapter VIIL Acanthocephali. By E. Korschelt . . . 249 Cleavage ; embryonal membrane 249 Embryo and larva : their migration and further development 250 Literature 255 Chapter IX. Rotatoria. By E. Korschelt .... 256 Heproduction by means of summer, winter, and male eggs ; cleavage 256 Formation of the germ-layers and further development . . 257 General considerations : Trochospheera sequatorialis, com- parison with the Trochophore larva of the Annelida, etc. . 259 Relationships to the Arthropod a (?) 260 Literature 261 Chapter X. Annelida. By E. Korschelt 262 I. Chaetopoda and Archiannelida . ^ . . . 262 CONTENTS xiil PAGE 1. Daveloptnent through free-swimming larvae (Poly- chsbta and Archiannelida) 262 Oviposition ; brooding 262 Embryonic development (cleavage, blastula) . 263 Formation of the germ-layers ; mesoderm . . 265 Trochophore [larva] 266 Metamorphosis of the Trochophore larva into the worm 268 The various larval forms (atrocha, monotrocha, telotrocha, etc.), with remarks on their meta- morphoses 270 Suppression of the larval form through brooding 280 2. Development without free-swimming larvae (Oli- gochseta) 281 Oviposition (cocoons) ; cleavage .... 281 Formation of the germ-layers (mesodermal bands) 282 The embryos in the larval condition (head kid- ney) 283 Formation of the worm 284 Primitive segments ; formation of the mesoderm 285 3. Formation of the organs 286 Ectodermal structures : epidermis ; setigerous sacs ; nervous system and sensory organs . 287 Mesodermal structures : body cavity ; muscula- ture ; blood-vessels 289 Head kidney and segmental organs . . . 295 Genital organs 297 Entodermal structures : intestinal canal . . 299 Non-sexual reproduction ; regeneration ; divi- sion (Lumbriculus, Ctenodrilus) . . . 301 " Budding " of the Naidides, Autolytus, Syllidse, Nereis 302 Stock-formation of Syllis ramosa .... 304 II. Echiuridae 304 1. Oviposition; cleavage and formation of the germ- layers 304 2. Larval form and metamorphosis of Echiurus and Thalassema 306 3. Larval form and metamorphosis of Bonellia . . 310 Development of the male of Bonellia; general considerations 312 in. Dinophilus . . . 313 Resemblance to polytrochal Annelid larvae . . 313 Oviposition ; development ; sexual dimorphism . 314 IV, Myzostoma 315 Embryonic development and larval form . . , 315 Resemblance to Annelid larvae; metamorphosis . . 316 XIV CONTENTS PAOB V. Hirudinea 318 Oviposition (cocoons) 318 1. Cleavage ; formation of the germ-layers and estab- lishment of the external form of the body (Rhyncobdellidae) 319 Formation of the germ band 321 Development of the embryo 323 Gnathobdellidse : cleavage ; germ bands . . 325 2. Larvae of the Gnathobdellidse (ciliation and primi- tive kidneys) 327 3. The further development of the body ; establish- ment of head and trunk 329 4. Formation of the organs 331 General considerations on the development and interpretation of the Hirudinea . . . 336 VI. Branchiobdella 336 Systematic position 336 Oviposition ; cleavage ; formation of the germ- layers, etc 337 General considerations regarding Annelida : larval forms (Trochophore) 341 Trochosphaera, Pilidium, and other larvae : descent . 342 Head and trunk ; segmentation 346 Literature 350 Chapter XL Sipunculid^. By E. Korschelt .... 357 1. The development of Sipunculus (blastula, formation of the germ-layers) 357 Development of the embryo 358 The larva of Sipunculus 361 Metamorphosis into the adult animal .... 363 2. The development of Phascolosoma 364 General considerations : Sipunculidae and Echiuridse as Gephyreans ; relationships to the Annelida . . 365 Literature 366 Chapter XII. Ch^tognatha. By K. Heider .... 367 Systematic position ; oviposition ; cleavage ; formation of the germ-layers 367 Further development 370 General considerations : relationships to the Annelida ; absence of the larva ; coelomic sacs 371 Literature 371 Chapter XIIL Enteropneusta. By E. Korschelt . . . 373 Anatomy 373 Development without Tomaria larva: cleavage ; formation of the germ-layers ; development of the embryo ; larva . . 379 Development by means of the Tornaria 381 CONTENTS XV PAGE Further developmental processes of both types : coelomic sacs, gills, genital organs, nervous system, etc. . . . 384 General considerations : relationships to the Echinodermata ; segmentation; relationships to the Chorda ta . . . 388 Literature 390 Chapter XIV. Echinodermata. By E. Korschelt . . . 392 1. Formation of the primary germ-layers, the mesenchyma, mouth, and anus 392 Holothurioidea 893 Echinoidea 398 Asteroidea 401 Ophiuroidea ; Crinoidea 403 2. The origin of the enterocoele and the hydrocoele. . . 405 Asteroidea 405 Ophiuroidea 406 Asteroidea (Asterina) . 407 Echinoidea ; Holothurioidea 409 Crinoidea 411 Divergent statements in regard to the formation of the entero-hydrocoele : Ophiuroidea ; Echinoidea (two pairs of enterocoeles, internal segmentation) . . 413 3. Development of the typical larval forms ; simple funda- mental form 415 Crinoidea 415 Holothurioidea : Auricularia, and direct development . 417 Asteroidea : Bipinnaria, Brachiolaria .... 419 Aberrant larval forms of the Asteroidea . . . 421 Ophiuroidea : Pluteus, direct development . . . 422 Echinoidea : Pluteus, direct development . . . 424 4. Metamorphosis of the larva into the echinoderm . . 426 Holothurioidea 426 Asteroidea 432 Ophiuroidea 437 Echinoidea 438 Crinoidea 443 Regeneration and division 455 General considerations : common features of the develop- ment 456 E-adial structure of the adult forms : its derivation , . 458 Relationships to other divisions (value of the larval forms, ambulacra 1 system) 460 Literature 461 CORRIGENDA p. 30, 1. 6 Schulze, No. 12 change tc Schulze, No. 26. p. 33, 1. 5 (No. 4) ... ... (Ganin, No. 4). p. 43, footnote ... Literature on Cnidaria ,, Literature on Hydroidea. p. 49, 1. 12 fr. bot. No. 39 No. 38. „ 1. 13 „ No. 37 No. 35. „ 1. 14 „ Stromobrachium ,, Stomobrachium. p. 51, 1. 2 (No. 42); (No 21) ... (No. 43); (No. 22). „ 1. 14 fr. bot. (No. 33) (No. 32). „ 1. 15 „ (No. 32); (No. 35) ... (No. 33); (No. 87). p. 52, 1. 14 „ (No. 28) „ (No. 26). „ 1. 16 „ Mekeschowsky ,, Mereschkowsky. p. 53, 1. 1 Hydrocorallia ,, the Hydrocorallia. „ 1.11 ... (No. 30) (No. 29). p. 58, last 1. (No. VII.) „ (No. v.). p. 120, 1. 3 Cyanidse ,, Cyaneidse. p. 205, 1. 4 prefix " V." to the line. INTRODUCTION jOOLogical research of the last decade has led to a sharp separation of two chief divisions of the animal kingdom: the Protozoa and the Metazoa. In the group Protozoa the individual can, from its structure, be referred to the funda- mental type of a cell. These unicellular individuals exist either separately or united in great numbers to form colonies or corms. In the latter case, however, the different indi- viduals remain equivalent to one another in structure and function. In the group Metazoa, or Germ-layer animals, on the contrary, there always results a multicellular or- ganism (cell-community or cell-corm), in which the single cells give up their independence for the good of the com- munity, and accommodate themselves to a division of labour, in consequence of which there is brought about a diversity in the structure and function of the cells of the Metazoan organism. While the development and differentiation of distinct tissues with specific functions result from this poly- morphism of the cells, the entire colony gains a higher functional capacity and a more complete unity. In this way there arises an individual of higher rank or second degree, w^hich we designate as person. These Metazoan in- dividuals also may, through incomplete separation after budding, remain united in colonies, and then there results an individual of the third degree, the stock or corm. By- adaptation of the stock-forming persons to various functions, accompanied with their polymorphous development, a higher functional unity may be reached in this case also. As a result of the division of labour which is effected among the cells of the Metazoan organism, it comes about K. H. E. 1 B Z INTRODUCTION that the ability of reproducing the entire organism does not belong alike to all the cells. It is confined rather to very special cells, which are known as reproductive cells (egg-cells and sperm-cells) ; these are cells which for the most part are developed only in definite regions of the organism {genital organs, gonads). The development of the Metazoan begins with the fusion of two morphologically different reproductive cells derived, as a rule, from two different individuals {fertilization). This kind of reproduction, known as sexual reproduction, is typical for all Metazoa. In many forms, however, nOn-sexual modes of reproduction (by division or budding) are interpolated in the life-history. If such an interpolation is the rule, so that two morphologically differ- ent generations, one of which multiplies by sexual and the other by non-sexual reproduction, regularly alternate with each other, then this condition is known as metagenesis or alternation of generations. It may also happen, however, that there is a regular alternation of sexual generations, in one of which reproduction is hermaphroditic or parthenogenetic, while in the other it is by means of separate sexes. Here also there occurs a heteromorphous development of the two generations. We call this condition heterogeny. Inasmuch as the individual Protozoan has the morpho- logical value of a single cell, the embryology of the Protozoa belongs to the province of cell morphology. For this reason it is usually excluded from the domain of comparative em- bryology of animals in the stricter sense ; in this book, too, it will receive no consideration. Comparative embryology accordingly has to do with the development of the Metazoa, and, above all, with their development from the fertilized e^g. Its chief problems consist in the investigation of the formation of the germ-layers, the origin of organs, and the development of the general form of the body. Its purpose is the recognition of the laws of development, the determina- tion of the homologies of organs, and the deduction of the ancestral history of the Metazoa. The Metazoa constitute a single stem of the animal king- dom. It is very probable that all Metazoa can be referred to a common ancestral form, and that certain correspond- INTRODUCTIOJi 3 features in the mode of development are the result of common descent. The earliest stages of development in the Metazoa can readily be reduced to a uniform plan characterized by the appearance of the blastula- and gastrula' stages at the end of cleavage. One is justified in the as- sumption that in these two stages there exists a repetition of ancestral forms which are common to all the Metazoa. In the first stages of development of the Metazoa the existence of a chief or 'primary axis can be recognized, the ends of which are distinguished as the animal pole and the vegetative pole, because in the differentiation of the two primary germ-layers, which soon follows, the layer arising in the vicinity of the animal pole (ectoderm) presides over the animal functions (sense perception, locomotion), while the germ-layer at the opposite pole (entoderm) is mainly de- voted to the functions of vegetation (e.g., nutrition). The Metozoa accordingly at first show a monaxial, heteropolar structure. Frequently the chief axis can be recognized in the egg-cell of the Metazoa before the beginning of development, since the germinative vesicle (nucleus of the egg-cell) and a dense accumulation of protoplasm are situated near the animal pole, whereas in the region of the vegetative half of the egg a great accumulation of yolk particles can be recognized. The animal pole, furthermore, is characterized by being the place at which the expulsion of the polar globules takes place before fertilization. The process of the cleavage of the egg, by which, after fertilization has taken place, the embryonic development is initiated, is essentially an ever-progressing division of the egg, which takes place according to fixed laws, and by which the egg is divided into a number of cells (cleavage spheres, blastomeres), whicli at first are still undifferentiated. Ac- cording to the direction which the planes of cleavage occupy in this process, we distinguish meridional and equatorial furrows, the former coinciding with the chief axis, the latter being perpendicular to it. In this manner there arise blastomeres that are at first spherical, but in later stages more or less pyramidal in form, and which are arranged radially about a point occupying the centre of the egg. By 4 INTRODUCTIOX separation of the cells there soon arises a central cavity, the cleavage cavity or Von Baers cavity (hlastocoele), which con- tinually increases during the succeeding cell divisions, while the blastomeres arrange themselves about this cavity in a single-layer epithelium {blastoderm). The stage of develop- ment thus reached is known as the hlastula or hlastosphere. In the one-layer blastula an arrangement of the parts of the e^^ about the chief axis is also clearly recognizable. The cells in the vicinity of the animal pole are, as a rule, smaller and not so rich in nutritive yolk particles ; whereas the cells of the vegetative portion are larger and richer in yolk, and, in consequence of the impeding influence offered by the nutritive yolk, divide more slowly. ^ The wall of the single- layer blastosphere represents the first of the primitive organs of the Metazoan body. In the simplest cases a gastrula- stage is developed out of the blastula-stage by the cell-layer of the vegetative half becoming flattened and gradually depressed, so that there arises an ever-deepening invagination at the vegetative pole. In this way the cleavage cavity (primitive body cavity) becomes gradually reduced, and often is preserved only as a narrow cleft between the two layers of the body-wall produced by the metamorphosis already described. The gastrula- stage has substantially the form of a sac. It encloses a cavity which has arisen by invagination, called the archenteric cavity, and opens to the exterior in the region of the vegetative pole by means of the primitive mouth or prostoma (blastopore). The wall at this stage consists of two cell-layers : an outer, the ectoderm, which is derived from the cells of the animal portion of the blastosphere, and an inner, the entoderm,, which consists of the cells of the former vegetative half, and which has reached the inside of the embryo by the process of invagination. In the region of the blastopore the ectodermal and entodermal layers be- come continuous with each other. Ectoderm and entoderm ^ There are reasons for thinking that the rate of cleavage is not wholly dependent on the proportion of nutritive yolk in the blastomere. (See KoFOiD, C. A., " On Some Laws of Cleavage in Limax," Proc. Amer. Acad. Arts and Sciences, vol. xxix., p. 180, 1894) [Translators]. i INTRODUCTION 5 represent the two primitive organs — or, as they are called, the two primary germ-lai/ers — resulting from the differentia- tion of the simple blastosphere. The gastrula-stage, which recurs under various modifications in all the Metazoa, ap- pears as the recapitulation of a hypothetical ancestral form {Gastrea), which was characterized by the development of the archenteron. Among the Metazoa now living many of the Cnidaria have retained essentially the structure of this hypothetical ancestor. In the more highly developed forms the two primary germ-layers undergo various modifications, whereby additional organs are differentiated. A third layer, the mesoderm or middle germ-layer^ also grows in between the two cell-layers. Concerning the origin of this we shall speak farther on. Of the primary germ-layers, the entoderm retains, even in the higher Metazoa, the function originally belonging to it : that of receiving and digesting food. In general it constitutes the epithelium of the mid-gut. From the ectoderm, on the other hand, arise usually the epidermis, and the nervous system, and sense organs, as well as the epithelial lining of the stomodeal and proctodeal invagi- nations. We have described above a method of origin of the blastula- and gastrula-stages as it occurs in some, but by no means all, Metazoa. It -was chosen as the type because, with due regard to the disturbing influences present, many of the aberrant modes of development can readily be reduced to the plan here presented. Frequently cleavage, the formation of the blastosphere, and the process of gastrn- lation are modified by the presence and definite arrangement of large masses of food-yolk. Certain eggs with little food-yolk approach most nearly to the plan presented above (e.gr., those of Amphioxus, Sagitta, and the Echinoderms). In such cases cleavage results in the production of blastomeres which are nearly uniform in size, so that in the completed blastosphere only a slight difference can be detected between the size of the blastomeres of the animal and vegetative poles. However, even here those of the vegetative pole are, as a rule, a little more voluminous. This kind of cleavage is called total and equal cleavage. It b INTRODUCTION is called total because the entire mass of the egg is separated by the division into blastomeres, and it is called equal on account of the approximately equal size of the resulting blastomeres. The blastula-stage which, with large central segmentation cavity, is produced by this cleavage, is called a coelohlastula or archiblastula, whereas the gastrula arising from the latter by a process of enfolding is called an inva- gination gastrula or embolic gastrula. In the eggs of some Cnidaria, especially Hydroids, whose earliest development is accomplished exactly in the mariner described above — that is, by total and equal segmentation and subsequent development of a coelohlastula — there exists a method of entoderm formation (gastrulation) which differs somewhat from the method by invagination just described, although it can be reduced to the same. This is the forma- tion of the entoderm by polar ingression. In this case the entoderm does not arise by an invagination of the cells of the vegetative pole, but the latter detach themselves from the blastoderm and migrate into the blastocoele, which in this way gradually becomes filled with a closely packed mass of entoderm cells. It is only secondarily that the archenteric cavity arises in this mass as a fissure, and that a mouth is formed by dehiscence of the wall. One sees that this kind of entoderm formation can readily be derived from that by invagination, inasmuch as the essential difference from that method of formation consists in the fact that the entoderm cells give up their epithelial continuity at the beginning of the ingrowth. Closely allied to the above-described type of total and equal cleavage are forms in which a more or less con- siderable amount of food-yolk is deposited in the vegetative half of the egg. On account of this accumulation the vege- tative portion of the egg considerably exceeds the animal portion in mass. It follows from this that in the course of cleavage, which here also is total, the cleavage cavity appears relatively small, and occupies a very eccentric position near the animal pole. The wall of the blastosphere, which can still be called a coelohlastula, in this case presents a con- siderable difference in thickness at the animal and vegetative INTRODUCTION 7 poles. We call this kind of cleavage total unequal cleavage, and group together as holohlastic eggs the forms belonging to this type and those previously mentioned. The unequal- walled blastula that has arisen through total unequal seg- lentation can in its further progress lead to the formation a gastrula by invagination, but in this case the gastrula "avity will be relatively shallow, corresponding to the small size of the segmentation cavity. In some other cases, on the contrary (e g., in some Annelids), after the conclusion of the process of total unequal cleavage, there is formed a blastula in which the cleavage cavity is reduced to a minimum. Accordingly there results from segmentation a more or less solid cell-mass (sterroblas- tula), in which we can distinguish a portion composed of large entodermic elements rich in food-yolk from an ectodermic portion consisting of small cells. The latter is placed upon the former like a small cap in the region of the animal pole. Here gastrulation by invagination cannot take place ; but the gastrula-stage is formed by the growth and consequent increase in size of the cap-shaped ectodermic part, whereby its edges push themselves more and more over the entodermic mass, so that finally the latter is entirely included within the ectodermic sac. We designate the solid gastrula-stage arising in this manner as a circumcrescent or epiboUc gastrula (sterrogastrula) . By this means a gastrula cavity is not formed primarily, but arises only secondarily as a fissure in the entodermic cell-mass. The edges of the spreading ecto- dermic layer must be regarded as the blastopore, which accordingly is filled by a so-called yolk-plug. The presence of large quantities of yolk matter in the region of the vegetative half of the egg presents obstacles to the progress of cleavage in that region. The accumula- tion of large masses of yolk can go so far that this portion of the egg does not at first take part in the segmentation ; but only a small portion, situated in the vicinity of the animal pole and consisting principally of formative yolk, is divided into blastomeres. Such, which undergo only a partial cleavage, are known as meroblastic eggs, in contra- distinction to holoblastic ones. There is developed in this 8 INTRODDCTION way a disc-shaped embryonal part which is situated on the un segmented yolk-mass at the animal pole. We call this type of segmentation, which represents the most extreme case of unequal cleavage, dlscoidal cleavage. It occurs, for example, in the Cephalopods. A particular type of cleavage, which does not fit into the above series, occurs in the class of Arthropods. Whereas all eggs thus far considered were characterized by a more or less considerable accumulation of yolk in the region of the vegetative half (telolecithal eggs), the distribution of the yolk being accordingly eccentric, the eggs of Arthro- pods exhibit a regular distribution of the yolk masses of such nature that their centre coincides with the middle point of the egg (centrolecithal eggs). The first cleavage nucleus here lies in the centre of the egg, where by divi- sion it separates into a large number of nuclei, which are distributed uniformly at the periphery of the egg, and thus give rise to the formation there of a layer of small uniform blastomeres. This cell-layer represents the blastoderm, while the cleavage cavity of the blastula-stage produced in this way is filled with the unsegmented yolk-mass. This kind of cleavage is known as superficial cleavage. The modifications of development hitherto considered ap- pear to be dependent principally upon the amount and mode of distribution of the yolk matter. We have still to consider some forms which in the mode of distribution of the yolk recall the centrolecithal eggs, but which by their peculiar mode of entoderm formation prove to be aberrant forms. In the first place, there should be mentioned in this connection the kind of entoderm formation (by delamination) occurring in the Cnidaria (Hydroids). The typical case of this kind exists in the development of the Geryonidae. A cceloblastnla is here formed by total and equal cleavage, and there follows a division of the cells in such a manner that an inner portion rich in yolk becomes separated from a superficial part with little yolk matter. In this way there arises out of the one- layer sphere an arrangement of the cells into two concentric hollow spheres, of which the inner contains the elements of the entoderm, and the outer those of the ectoderm. One INTRODUCTION ft sees that in this mode of formation, which cannot be com- pared to the plan of gastrulation by invagination, the gastrula cavity arises from the cleavage cavity. Apparently a transition between the formation of the entoderm by delamination and by polar ingression is effected by a kind of entoderm formation which has been observed by Metschnikoff in different Hydroids, and which is designated as multipolar ingression {i.e., ingression from all sides), in which single cells of the blastoderm migrate into the blastoccele from different points of the surface and here form an entodermic cell- mass. Nevertheless the process of forming entoderm by delamination remains, in contrast with the other types of entoderm formation, some- what isolated and unexplained. Closely related to delamination is a kind of entoderm formation which was formerly held to be of frequent occurrence, but whose range of dis- tribution has become more and more restricted by careful investigation of the individual cases. They are the cases in which the blastomeres present no radial arrangement about a point within and no definite relation to a cleavage cavity. Such a stage, which is an apparently irregular solid mass of cells, without cleavage cavity, has been designated as morula ; and it is assumed that by a rapid division of the cells at the surface an outer cell-layer is differentiated from the inner cell-mass, so that here also the separation of ectoderm from entoderm would be brought about by a splitting off which takes place uniformly over the entire circumference. We shall see that examples of such a mode of origin of the two primary germ-layers are still ascribed to many Hydroids and Anthozoa, though probably the greater part of the cases referred to this method can be reduced to epibolic gastrulation, in which event the morula- stage, as being a Schema founded on erroneous assumptions, would have to be omitted. Even though the last-mentioned modes of entoderm formation, re- stricted as they are to a few kinds of Metazoa, place many difficulties in the way of the conception that there is uniformity in this process, it is probable that more careful investigation may succeed in bringing them into accord with the less aberrant types already mentioned. We liave seen that the chief axis of the gastrula- stage unites the anterior or apical (animal) and the posterior or prostomial poles with each other. In the lowest types of the Metazoa — the Porifera, Cnidaria and Ctenophora — this primitive axis becomes the permanent chief axis of the body ; therefore these groups have been contrasted by Hatschek ^ as Frotaxonia with the rest of the Metazoa, which he terms * Compare Hatschek's Lehrbuch der Zoologie. Jena, 1888, p. 40, as well as p. 69 et seq. 10 INTRODUCTION Seteraxonia or Bilateria. In the latter the blastopore undergoes a secondary shifting, so that the later chief axis can no longer be identified with the primitive axis. The layered structure of the Metazoa becomes more com- plicated by the appearance of a cell-layer introduced between the ectoderm and entoderm, which takes a position in the primitive body cavity, the remnant of the cleavage cavity, and is designated as mesoderm or middle germ-layer. This name is applied to any cell-layer introduced between ecto- derm and entoderm and separated from both by a sharp boundary, but it is not intended thereby to maintain the homology of this layer for all the Metazoa. On the contrary, it appears that in the Protaxonia mesodermic layers w^ere independently acquired in various ways. Even in the Bilateria the homology of the mesoderm in all groups is not absolutely established, although it may be assumed as probable. The mesoderm of the Bilateria arises as a rale out of the primary entoderm, which in such cases is divided into two parts : mesoderm and secondary entoderm. In regard to the mode of origin, we can distinguish two sharply separated types : the formation out of two priTnitive mesoderm cells and the formation by the production of diverticula of the archenterooi} The first type is widely distributed among the Bilateria. At an early period there become conspicuous at the prostoma of the gastrula-stage two peculiar cells, by whose position the median plane, which passes between the two, is determined. These cells are known as the primitive mesoderm cells. They move into the space betw^een the ectoderm and entoderm (therefore into the primitive body cavity), and by proliferation give rise to two paired cell- bands, which are called the mesoderm hands, and out of which the organs of the mesoderm are constructed. The formation of the mesoderm by the production of diverticula * As a third type of mesoderm formation one might perhaps cite the formation of a mesenchyma (compare p. 11) in those cases in which, as in the Nemerteans and Echinoderms, numerous wandering cells migrate into the blastocoele at an early period. Yet this type could perhaps be reduced to one of those mentioned above. of the i INTRODUCTION 11 of the arclienteron, as it occurs in the Chsetognatha, Brachio- poda, and Cliordonia, consists in the development of paired sac-like diverticula of the arclienteron, which become constricted off, and then as independent coelomic sacs give rise to the systems of organs of the mesoderm. Different as these two kinds of mesoderm formation appear to be, they nevertheless can be reduced, like the processes of gastrulation by invagination and by polar ingression already described, to a uniform plan, if we assume that in the first case the mesodermic elements at an early period abandon (as primitive mesoderm cells) epithelial continuity with the entoderm, whereas in the second case the mesodermic cell- mass retains provisionally its epithelial continuity, and only later becomes separated from the entoderm by the formation of the diverticula. As regards the subsequent fate of the mesoderm, we can, if we disregard the formation of the individual organs, distinguish two types. In the one case the union of the mesodermic elements is loosened, and these distribute them- selves in the manner of amoeboid wandering cells in the space of the primitive body cavity, which eventually they completely fill with a tissue consisting of stellate migratory cells embedded in a gelatinous stroma. This tissue is known as inesenchijma (0. and R. Hertwig). By separation of the cells of the mesenchymatous tissue, spaces (lacunae) may be formed in it, which may coalesce to form larger spaces, and so apparently represent a kind of body cavity. To such spaces the name of pseudocoele is given. In other cases the largest part of the mesoderm is employed in the formation of paired sacs, the coelomic sacs, the walls of which consist of a continuous epithelium of mesoderm cells. The cavities contained in them represent the true body cavity or coelom. The paired coelomic sacs entirely surround the intestinal canal, so that the walls of the sacs come together in the middle line above and below the intes- tine to form the so-called mesenteries. The body cavity divides the mesoderm into two layers. The outer layer, the one applied to the ectoderm, is known as the somatic layer, the inner one, applied to the entoderm, as the splanchnic layer. 12 INTRODUCTION There are a number of animals in which the mesoderm produces, in addition to the specific organs that have arisen from it (genital organs, excretory organs), only mesenchyma. Such is the case in the Platyhelminthes. In the great majority of the Bilateria, on the contrary, the formation of mesenchyma and coelom occurs together, and there is there- fore a sort of competition between the two methods of mesoderm development, so that in one case (Annelids, Sagitta, Phoronis) the formation of a ccelom predominates, in the other (Mollusca, Arthropoda) that of a pseudocoele (mesen- chyma). In the Bilateria there arise from mesoderm the muscula- ture, the genital organs,^ the excretory organs known as nephridia, the connective tissue, and fatty tissue. 1 [The statement that the fundament of the genital organs of the Bilateria comes from the mesoderm ought to be considerably restricted. In recent times the observations have been increasing which appear to support the Weismannian doctrine of the continuity of the germ cells. Grobben some time ago observed the early differentiation of the sexual cells in Moina. The same has been known for a long time to be true of Diptera and Aphidse. Recently Heymons {Sitz.-Ber. Gesell. Nat. Frennde, Berlin Jahrg., 1893, p. 263) has found similar conditions in various Orthoptera. The investigations of Faussek on Phalangidae [Biol. Centralbl., Bd. xii., p. 1, 1892) and the very recent observations of A. Bkauer {Zeitschr. f. wiss. Zool, Bd. Ivii. 1894) on the scorpion have shown the early independence of the genital fundaments in these forms. Special importance in the present question' is to be attached to the observations of Boveei on Ascaris (Sitz.-Ber. Gesell. f. Morph. u. Physiol. Milnchen, Bd. viii., 1892), according to which the sexual cells are dis- tinguished from the somatic cells even in the first stages of cleavage owing to the special structure of the chromatic elements of their miclei.] ■PTER I. PORIFEEA. ipoxGES reproduce by sexual and non-sexual means. To the non-sexual kinds of reproduction belong — (1) sprouting or budding, which may lead to the formation of complicated stocks or colonies ; (2) the formation of small buds which separate from the parent body and grow np independently into new individuals ; (3) reproduction by means of gemmulfE. The investigations on the development of sponges from the fertilized egg have not up to the present time yielded a uniform plan for the embryology of this group, and they frequently contradict one another. The following may be mentioned as features common to the development of all sponges. (1) The sexual products arise in the connective tissue of the so-called mesoderm out of cells which at first are not to be distinguished from the connective tissue cells of this layer. (2) The eggs are not surrounded by any cuti- cular envelope (chorion) or vitelline membrane. They lie naked in a cavity lined with endothelium (Fig. 1 e) in the mesoderm of the parent body. Here the expulsion of the polar globules, fertilization, and early de- velopment take place. 13 Fig. 1.— Egg of Placina trilopha in the parent body (after Misdeburg). r, polar globules ; e, endothelial lining. 14 EMBRYOLOGY The polar globules of the sponges have been overlooked up to the pre- sent time. According to recent observations by Magdeburg, not yet published, they present in Placina the appearance typical for most of the other Metazoa (Fig. 1 i'). Also the processes of their formation may well find place in the general plan, while, according to Fiedler's com- munication {Zeitschr. f. loiss. Zool., Bd. 47) on Spongilla, it would almost seem as if there existed here a peculiar type of formation. (3) The eggs undergo total cleavage, and develop in the parent body into spheroidal or ovate embryos covered on the surface witli flagella. (4) When the embryos have readied the stage of the oval, flagellate, so-called planula larva, they emerge and pass through a swarming stage, during which development makes but little progress. (5) After attachment to a fixed support is effected there follows a rapid transformation into a young sponge, resem- bling substantially the parent. We may best arrange the types of sponge development hitherto known in accordance with the characteristic con- dition of the swarming stage. I. — Type of development through a so-called Amphiblastu la-stage. The development of Sycandra raphanus, which has been described by Metschnikoff (Nos. 12 and 13) and F. E. ScHULZE (Nos. 19 and 22), serves as an example of this type. The egg of this calcareous sponge undergoes a total and nearly equal cleavage, but the course of cleavage is somewhat modified by the relation which the embryo acquires to the wall of one of the radial tubes of the parent (Fig. 2). The egg is a naked cell, and lies in the parenchyma close to the wall of a radial tube. It is first divided into two blastomeres of equal size (Fig. 2 A) by means of a furrow which is perpendicular to the radial tube, and in relation to the orientation of the developing embryo must be con- sidered as meridional. By means of another meridional furrow perpendicular to the first one, the two cleavage spheres separate into four blastomeres, now arranged in the form of a cross (Fig. 2 B), which are applied to the radial th a flattened basal surface ; and since thej do not come into close contact with one another at the centre, thej iclose between them a cavity (cleavagt> cavity) open above id below. With the next act of cleavat^'e each of these four Is is divided into two equal parts by a new meridional irrow (Fig. 1 G and D) . The embryo now consists of a circle C / Fig, 2.— Cleavage stages of Sycandra raphanus (after F. E. Schulze). A, two- 3II stage ; B, four-cell stage ; C eight-cell stage ; B, the same in vertical section in Its relation to the collared epithelium of the maternal radial tube (diagram) ; E, ixteen-cell stage ; F, the same in vertical section (diagram) ; G later stage of leavage with eight granular (ectodermal) cells at the lower pole ; H, blastosphere Stage in side view. In the interior the cleavage cavity j granular cells below, lerwise an epithelium of tall columnar cells. 16 EMBRYOLOGY of eight cells, which enclose the cleavage cavity. Since the cells are applied to the wall of the tube with their broad bases, and taper conically in the opposite direction, the embryo has nearly the form of a cup-cake (Fig. 2D). By means of a subsequently appearing equatorial furrow each of these 8 cells is separated into an upper smaller [entodermal] and a lower larger [ectodermal] segment ; and at the same time the whole shape of the embryo changes in this 16-cell stage, taking on the form of a biconvex lens by the bulging out of its basal surface (Fig. 2 E and F). The cleavage cavity is still open at both poles, although the opening of the upper Fig. 3.— Swarming larval stage of Sycandra raplianus (after F. E. Schulze, from Balfouk's Comparative Emhryology). A, amphiblastula-stage ; B, stage at the beginning of the gastrula invagination; cs, cleavage cavity; ec, future ectoderm cells ; en, future entoderm cells. [The orientation of the larva in this figure is the reverse of that in Fig, 2 D, F and H, the entodermal pole being there the upper one.]— Tbanslatobs. side is already considerably narrower than that of the lower one. By means of new meridional and equatorial furrows the embryo gradually passes into a multicellular stage, which has an almost spherical shape, corresponding to which there is an extensive cleavage cavity within. The opening at the upper pole has disappeared by the apposition of the cells, while the one corresponding to the former basal sur- face is still retained (Fig. 2 G). It is surrounded by eight [ectodermal] cells, which are soon distinguished by increasing PORIFERA 17 size and by their granular plasma. With the closure of this lower opening the embryo becomes a spherical bjastosphere. The granular cells now enlarge and multiply to the num- ber of about thirty-two ; the other cells meanwhile increase in numbers, and lengthen out into tall columnar prisms (Fig. 2 H), each of which develops a flagellum at the surface. The large, richly granular cells now fold into the segmentation cavity, and the last stage to be passed in the body of the parent, the so-called -pseudo- gastrula stage, is thus reached. It has nothing to do with the true process of gastrulation, but re- presents a transient condition, which was perhaps acquired in connection with the mechanism of hatching. , ., m / \ { I En ^• Fig. 4.— Attached gastrnla-stasre of Sycandra rapTianits (after F. E. Schui-ze). Ec, ectoderm; En, entoderm; m, gelatinous secretion between the two layers (remnant of the cleavage cavity). When the embryo is hatched the invaginated part (ecto- derm) returns to its former position, and an elongation in the direction of the chief axis follows. The ovate swarm- ing stage now reached is known as the amphiblnstula (Fig. 3 A). It consists of histologically differentiated halves. The half of the body directed forwards in swimming is composed of tall columnar flagellate cells, whereas the large granular cells of the posterior half of the body bear no flagella. Within is seen the considerably reduced cleavage cavity (csy. * [According to recent investigations of Dexdy (Appendix to Literature on Porifera, No. II,), small cells, which perhaps become mesoderm cells, make their appearance in this cavity at an early period.] K. H. E. C 18 EMBRYOLOGY After the completion of the swarming stage, and shortly before the attachment of the larva, a shortening in the direction of the chief axis takes place, which is accomplished principally by a flattening of the flagellate, formerly bulging cell-layer ; and an invagination of this layer quickly follows the flattening, the result being that the cleavage cavity is entirely obliterated. In this way a cap- shaped gastrula-stage (Fig. 3 B) is reached. The outer granular layer of cells can henceforth be considered as the ectoderm, and a circle of Fig. 5a.— Young, mortar-shaped Olynthus stage of Sycandra raphanus (after F. E. Schulze). Os, osculum ; po, lateral inhalent pores of the body-wall. about sixteen of these cells, which are particularly prominent and are known as marginal cells {Randzellen)^ surrounds the wide gastrula mouth or blastopore, while the invaginated flagellate layer represents the entoderm. The attachment of the larva now takes place by the fixation of the edge of the gastrula mouth by means of pseudopodia-like processes of the marginal cells to some 1 support (Fig. 4). The entire process of gastralation and attachment proceeds with uncommon rapidity. ^^bn the gastrula-stage ectoderm and entoderm are not ^^rosely applied to each other ; but one notices between them a space which must be interpreted as the remains of the segmentation cavity (Fig. 4 m), and which is filled with a gelatinous hyaline mass. According to Metschnikoff, in- dividual cells of the granular ectodermal layer migrate into this mass, and lead to the formation of the mesenchyma, the so-called mesoderm, between the two primary layers. The first skeletal structures arise in these cells in the form of small rod-like needles ; triradiate ones are formed later, and finally quadriradiate ones. After the gastrula mouth has be- come narrowed and finally closed, the hollow body of the larva, which has no external opening, elongates in the direction of the chief axis, and grows out into a cask-like or cylindrical form (Fig. 5a), the upper surface of which consists of a thin membrane, which acquires at its centre a circular opening, the beginning of the exhalent orifice (o-fculum, Os), which soon enlarges. At the same time the inhalent openings or pores (po) appear as perforations in the lateral walls. Since, moreover, the epithelial layer of the ento- derm acquires the character of flagellate collar-epithelium, the characters typical of the Porifera are completed in this ascon-like stage (Fig. 6a, Olynthus). The development into the Sycon takes place by the radial tubes becoming established as simple evaginations of the body- wall (Fig. 56) ; at first a circle of radial tubes makes its appearance at about the middle of the body ; to this a second is soon added, and so on. Fig. 5b. — Older, attached stage of Sycandra raphanus with the fundaments of the first radial tubes, r. Po, inhalent pores. 20 EMBRYOLOGY The amphiblastula-stage seems to occur in the Hfe-history of many Calcarea. It has also been found in Ascandra contorta (Barrois), in Ascandra Lieherkuhnii (Keller), and Leucandra aspera (Keller, Metschnikoff). The genus Ascetta develops according to another type. il. — Type of Development through a Swarming Coeloblastu la-stage. The egg of Oscarella lohularis (Halisarca lohularis) de- velops in the trabeculse of the tissue in the internal parts of the parent, which are without flagellate ampullae, and under- goes total and equal cleavage, by which there are formed at first two, then four, eight, sixteen, etc., blastomeres of equal size. At the sixteen-cell stage a distinct cleavage cavity can be recognized within. By further cell proliferation there is formed a hollow sphere (coelohlastula or archiblastula), the wall of which is composed exclusively of cubical cells of equal size arranged in a single layer (Carter, No. 3 ; Barrois, No. 2 ; F. E. ScHULZE, No. 20). Shortly before swarming, the elements of the body-wall lengthen out into columnar epithelial cells, each of which acquires a flagellum at its outer end. The swarming larva (Fig. 6 A) possesses an approximately ovate form, and ex- hibits a blunt yellowish pole, which is directed forwards in swimming, and a posterior, more pointed brownish-red pole. The wall of the blastula consists of a single layer of cylindrical flagellate cells. The internal cavity contains no cells, and is filled with an albuminous fluid (F. F. Schulze). By the invagination of one pole of the larva this stage passes into a hemispherical gastrula-stage, which, like that of Sycandra, now attaches itself by its gastrula mouth to a support (Fig. QB). Thus there arises a shallow, cap-shaped larva, the wall of which consists of two layers (ectoderm and entoderm) and the inner cavity of which must be con- sidered as the archenteron. The gradual closure of the wide gastrula mouth now follows ; and, at the same time, by a complicated process of folding, the first flagellate ampullae arise as diverticula3 of the archenteron (Fig. 6 0). The mesodermal connective-tissue layer originates by the mi- gration of cells into the space embraced between the PORIFERA 21 ectoderm and entoderm. Lastly, there occurs at the apical ile an evagination of the bodj-wall like a chimney-pot, at Fig. 6.— Development of Osearella, diagrammatic (after Hkideb). A, swarming blastula larva ; B, attached gastrula-stage ; C, stage initiating the closure of the gastrula mouth (Gm) and the folding of the entodermic sac; D, young sponge ; Os, oeculum po, inhalant pores ; Ec, ectoderm ; En, entoderm. 22 EMBRYOLOGY the apex of which the osculum (Os) breaks through (Fig. 6 D) The inhalent pores (po) are formed as perforations at places where the ampullae lie close to the ectoderm. The system of inhalent canals, which arises later, must be referred to invaginations of the ectoderm, that of the exhalent canals to evaginations of the entoderm. The larva is not attached by its entire base, but rests upon a few foot-like supports (K. Heider, No. 8). The development of the Plakinid® appears to be closely related to that of Oscarella. The swarming larva can on the whole be included in the type just described. The metamorphosis into the attached stage has not been accurately investigated ; it is certain, however, that here also the ampullae arise as diverticula of a common central cavity (F. E. Schulze). III.— Type of Development through a Parenchymula Stage. The superficial layer of the swarming larva consists of a cylinder epithelium, composed of long flagellate cells, and encloses an internal space filled with embryonic connective tissue. (a) The superficial layer presents on all sides cells of nearly uniform condition. Ascetta. — By total and equal cleavage there is first produced a coeloblastula-stage, which is similar to that of Oscarella. But even before the hatching of the embryo, which attains an ovate form in later stages, the immigration of cells into the internal cavity takes place ; and this occurs at the posterior pole of the larva. In this manner the primitive body cavity becomes filled with a connective-tissue mesenchyma, the common fundament of the mesoderm and entoderm, in which the permanent gastral cavity subsequently appears as a fissure. The entoderm cells arrange themselves about this in the form of a unilaminar epithelium (0. Schmidt, Metschnikoff). (6) The superficial layer in the region of the posterior pole of the larva presents an altered condition of the cells. ^ 1 [The investigations of Delage (No. I., Appendix to Literature) and those of Maas (Nos. IV. and V.) on the development of the Cornacuspongia (Ceratosa and Silicea) are of great importance. A satisfactory founda- tion for the interpretation of the development of sponges in general has PORIFERA 23 Ceratosa. — The embryos of Spongelia pallescens (Fig. 7) which are ready to swarm possess a cyhndrical form with one end convexly rounded and with a shallow depression at the other. In the region of this shallow invagination the flagellate cells are pigmented brownish red. The inside of the embryo is filled with a gelatinous connective tissue. The embryo Euspongia officinalis before swarming presents a very similar struc- m4 Fig. 7.— Longitudinal section through a larva of Spongelia pallescens (a'"ter F. E. Schulzk). a, pigmented epithelial cells of the posterior pole of the body ; ms, gelatinous connective tissue inside the larva. (The superficial covering of flagella has been omitted in the illustration.) thereby been acquired, inasmuch as the development of the Cornacu- spongia fits in completely with that of Sycandra. The free-swimming larval stage corresponds to the pseudo-gastrula stage of Sycandra. The superficial flagellate epithelial cells of the larva subsequently become the collared entoderm cells of the adult sponge ; whereas the inner cell- mass, which in many cases is exposed at the posterior pole of the larva, represents the common fundament of the ectoderm and mesoderm. The ectoderm and mesoderm of sponges must be regarded as an undivided whole. The larv® attach themselves by the anterior pole, and at the same time occurs that inversion of the layers which must be compared with the process of actual gastrulation in Sycandra.] 24 EMBRYOLOGY ture, save that the parenchyma filling the internal cavity is of a different histological character; it consists of a tissue comparable to hyaline cartilage. The cleavage in this form is total and equal, although a cceloblastula-stage is never developed (F. E. Schulze, I^os. 23 and 24). Chalinida. — The egg of Chalinula fertilis segments totally and un- equally, according to C. Kellek (No. 9). The first act of cleavage results in a larger and a smaller blastomere. The next stage which has been observed shows three small and one large blastomere. By the succeeding division of the small cells there results a stage in which six small cells rest like a cap upon the large undivided blastomere, which also soon divides. The small cleavage spheres are said to represent the fundament of the ectoderm, the large ones that of entoderm and mesoderm together. In the further course of cleavage, the small cells, having the form of a shell, grow around the large ones in such a manner that there results an epibolic gastrula, the mouth of which is entirely filled by the solid cell-mass of the primary entoderm, which is plainly visible at this point. The embryo now develops flagella on its entire surface, acquires a more elongated form, and emerges as a planula larva. At its posterior pole it presents a darker -coloured area of the superficial flagellate layer, and this corresponds to the superficial entodermic portion. The earliest spicules are very soon developed in the cells of the internal parenchyma. The larva now attaches itself by its posterior pole, but very soon turns over, so that it adheres to the support by the entire broad side of the body. It now acquires the form of an irregular flat cake. The forma- tion of the ampullae within is said to proceed in a different manner from that described for Oscarella (p. 22) ; that is to say, individual entodermic cells unite into compact groups, within which a cavity appears later. The fundaments of the ciliated chambers (ampullas), which at first were independent, then establish relations with a large central cavity arising in the parenchyma, which soon breaks through to the outside at the apex of the larva, thus producing the osculum. The larvae of most siliceous sponges appear to belong to a type similar to that of the swarming larva of Chalinula, thus the larvae of Esperia, Amorphina, Easpailia, and Keniera (0. Schmidt, Metschnikoff), further those of Isodyctia and Desmacidon, made known by Ch. Bareois. Reniera. — The larva of Keniera filigrana resembles the ones previously described. It consists of a flagellate columnar epithelium and an in- ternal cellular parenchyma. In the course of the further development the layer of columnar cells ruptures at the anterior and posterior ends, so that the internal parenchyma is exposed. The larva attaches itself by the anterior pole, loses the covering of flagella belonging to the super- ficial layer, and takes on the form of a flattened cake, while in the internal parenchyma, the common fundament of entoderm and meso^ derm, a cavity appears in the form of a fissure, about which the nearest cells group themselves in the form of an epithelium. In this manner the entodermic epithelium is separated from the mesoderm. The first PORIFERA 25 k Later the osculum breaks through, and siliceous spicules are formed in the cells of the mesodermic layer (W. Marshall, No. 10). Hallsarca. — By total and equal cleavage (F. E. Schulze, No. 20) there originates a blastula, into which migrate cells that entirely fill the cleavage cavity, and there form a connective-tissue mesenchyma. Upon emerging the larva presents at the posterior pole an area consisting of large granular flagellate cells. After the larva has attached itself and assumed a cake-like form, the ectoderm loses the flagella, and is changed into a pavement epithelium. In the internal parenchyma there now arise fundaments of the ampullae and canals which at first are separate, but subsequently unite into a common system (Metschnikoff, No. 14). A unique type of development, which perhaps most resembles Reniera, appears to exist in Spongilla. The de- En Fig. 8.— Late stage of cleavage (beginning gastrulation) of Spongilla {Ephydatia) jluviatilis (after Goette). Ec. ectoderm cells; En, entoderm cells; d, central entodermic cavity. velopment of this fresh- water sponge has been described by Ganin (Nos. 4 and 5) and Goette (No. 6) ; but in many points the statements of these investigators do not altogether agree. In our presentation we follow the detailed descrip- tion of Goette without forming any preliminary judgment as to the way in which the development of Spongilla is related to that of other sponges. A final judgment will be possible only when new observations have been made on the development of sponges of the most widely differing groups. The egg of Spongilla (Ephydatia) fluviatilis undergoes 26 EMBRYOLOGY total unequal cleavage, bj means of which there arises an embryo consisting for the greater part of large blastomeres rich in yolk (entodermic portion, Eig. 8 En), and presenting only at its npper pole a cap of small blastomeres containing little yolk (ectodermic part. Fig. 8 Ec). Since the cap of ectoderm cells gradually grows around the entire embryo, there is formed in this way a kind of epibolic gastrula. At an early period there appears in the entodermic mass an eccentrically placed irregular cavity, which can be referred neither to a cleavage cavity nor to an archenteron, and is known as the entodermic cavity. The chief axis of the embryo is recognized by the eccentric position of this cavity, for it always lies close to the apical pole (the subsequent anterior pole). The embryo, which at first is comparable to a plano-convex lens, now elongates in the direction of the chief axis, and is covered on the surface with a coat of cilia. The swarming larva (Fig. 9) is generally ovate in form, and in floating in the water has the broadened anterior end of the body, in which the spacious entodermic cavity is situated, always directed upwards. The larva possesses a superficial, unilaminar, flagellate epithelium, all the cells of which present the same character. Within the posterior half of the body is found a solid entodermic mass, which, in the course of the further development, takes on the character of embryonic connective tissue. The cells lying in the vicinity of the cavity are flattened, and form a reticulated layer of amoeboid elements. A similar layer of flattened cells is found at the surface of the solid entodermic core, where it is in close contact with the ectoderm. At an early period spicules are developed in certain cells of this core. The larva attaches itself by the apical pole, the layer covering the entodermic cavity being split open, and the edges of the breach thus formed bending outward. In this way the cells of the entodermic layer come in contact with the support, to which they adhere by means of pseudopodia- like processes. In the majority of cases the larva after this first attachment bends over in order thus to attach itself by a broader surface. According to GtOette, there ensues the PORTFERA 27 complete loss of the ectoderm, which breaks up and is detached from the surface in shreds, so that the entire body of the young sponge consists of entoderm. In this now solid mass (for the entodermic cavity has disappeared in the process of attachment) ampullae arise as separate fundaments ; they are produced by groups of cells (each of which has arisen from a single cell) acquiring cavities within them. The canals and cavities of the body develop in the same manner in many separate regions, and afterwards unite with one another and with the ampulla3. The most superficial layer of the body acquires the character of a pavement epithelium, ^c '-'- En Fig. 9.— Free-swimming larva of 'Si^onqilla fiuviatilis (after Goettk). The covering of flagella has been omitted. Ec, ectoderm; En, entoderm; d, ento- dermic cavity. and forms the permanent epidermis of Spongilla. All or- gans therefore arise by histological differentiation from a single layer, the primitive entoderm. In Goette's account, which we have given here, several points must appear to be still doubtful, thus, above all, the alleged complete casting off of the ectoderm. To be sure, a rupture and partial loss of the ectoderm has also been maintained for other sponges (Eeniera, Esperia) by earlier observers, to which attention has already been called (p. 24). But such appearances are probably for the most part to be referred to pathological or abnormal processes. Ganin has assumed that in Spongilla the ecto- derm of the larva becomes the permanent epidermis of the sponge ; and 28 EMBRYOLOGY that statement has recently been confirmed by the observations of 0. Maas {Zool. Anz., 1889), who has convinced himself on the same object that the ectoderm of the larva is not cast off, but gradually passes over into the superficial pavement epithelium of the adult sponge. Also in regard to the origin of the ampullfe and canal system Ganin does not agree with Goette. According to Ganin, the entodermic cavity is to be explained as an archenteron, and represents the earliest fundament of the canal system, from which the ampullae arise as diverticula. A mode of origin for the ampullas similar to that communicated by Goette for Spongilla has recently been maintained by Dendy (Quart. Jour. Micr. Sci., 1888) for one of the horn sponges (Stelospongia). The distribution of the recognized types of development among the different groups of sponges, then, is represented as follows : In the calcareous sponges (Calcarea) the amphiblasiula is found in most of the cases observed up to the present time. Perhaps this larval form is confined to the Calcarea. The cceloblastula appears in Oscarella and in the family of the Plakinidffi, whereas the parenchymula seems to be present generally in the Ceratosa and siliceous sponges. Furthermore Halisarca and Ascetta exhibit a parenchymula stage. As is to be seen from the preceding, a uniform plan of development for sponges cannot at the present time be formulated.^ The statements differ too widely. In certain cases we find a ccelogastrula-stage, w^bich attaches itself to some support by the circumference of the large gastrula mouth. This is of importance as a distinguishing charac- teristic in contrast with the Cnidaria, in which the attach- ment is always effected by the aboral pole of the two-layer planula larva. In other cases there is formed a parenchymula, the genesis of which for many forms is as obscure as the further development of this stage into the adalt sponge. We can only assume conjecturally that this stage is in all cases brought about by epibolic gastrulation or by the process of migration of certain cells from the entodermic pole. The most obscure point in the development of sponges is the metamorphosis accomplished at the moment of the attachment of the swarming larva. Even the state- ments regarding the pole by which the larva attaches itself vary for the different forms. Authors likewise differ in regard to the organogeny, above all as regards the origin of the canal system and the flagellate ampullae. In certain 1 Compare footnote, p. 22 [Tbanslators] . PORIFERA 29 common fundament of the ampullae and exhalent system is affirmed to have the form of a central ^ levity (archenteron), from which the ampullae and exhalent anals arise by repeated foldings of the wall. Opposed to this are the observations of other authors, according to whom the different ampuUee are formed independently, and become united by means of the canals which appear later on, and which only gradually unite into a common system of canals. In such a condition of affairs it is hardly possible to arrive at any general conclusions without in one way or another doing violence to the individual statements. This much, however, seems with some certainty to result from all the observations : that we have before us in the sponges an independent stem of the Metazoa, which is connected with the other types only at its roots. We adhere to the view that the sponges have a common origin with the rest of the Metazoa. In the embryology of the sponges we find true blastula- and gastrula-stages, which appear to point to an ancestral form common to the Porifera and all other Metazoa. Characteristics of histological differentiation (the formation of columnar and pavement epithelium, of con- nective tissue and cartilaginous tissue) likewise point to this community of origin. As opposed to these characteristics, the single fact of the occurrence of the collar-bearing flagellate cells of the entoderm does not seem sufficient to warrant the derivation of the Porifera as an independent group from the Choanoflagellata and the denial of their phylogenetic relationships to the rest of the Metazoa (Sollas, No. 15 ; BiJTSCHLl). That the Porifera do not stand in any close relationship to the Cnidaria (Coelenterata in the stricter sense) (Marshall, No. 11) appears to be beyond doubt. We lay less stress upon the absence of nettling capsules, as being a purely histological character, than on tectonic points. The attempts to reduce the structure of sponges to the fundamental form of the Polyp must lead to contradictions. Above all must be emphasized the fact that the exhalent opening of the canal system, the so-called osculum, is not homologous with the 30 EMBRYOLOGY moath of Coelenterates, farfchermore that the Porifera in general are derived from a monaxial, heteropolar funda- mental form in which the production of secondary axes in definite numbers has not yet taken place, whereas the radial type with four rays lies at the foundation of the Cnidaria (compare F. E. Schdlze, No. 12 ; A. Goette, No. 6 ; Heider, No. 8). The absence of movable processes of the body (tentacles) and the low grade of histological differentiation serve as substantiating facts in support of this view. The Porifera possess no true muscle fibres. The property of con- tractility appears rather to belong to all the cells in about the same degree, and the " contractile fibre-cells " occurring in the mesoderm of many sponges are distinguished from true muscle fibres by the fact that the contractile substance in them has not yet become separated as a distinct portion of the cell. The absence of a nervous system has not yet been proved, it is true ; but the presence of such a system does not appear to be firmly established, for up to the present time the groups of cells claimed by Lendenfeld as the nervous system of sponges have remained doubtful as regards this interpretation. In respect to the origin of the canal system of sponges, reference must be made to those primitive forms which are found more especially among the calcareous sponges, and by comparison of which it is most clearly proved that the complicated canal system of the siliceous and horny sponges has been evolved by a continuous process of folding of the wall of the sacular olynthus-like primitive form, whereby the collar epithelium eventually becomes localized in particular parts (ampullae) of the canal system. If the diagram Fig. 10 A represents the wall of a simple ascon perforated by pores, and if we bear in mind that the entire inner surface is covered with collar-cells, then Fig. 10 B shows the origin of the radial tubes of a sycon by means of a folding of this wall. At the same time the entoderm lining the common central cavity is transformed into a pave- ment epithelium ; and alternating with the evaginations of the radial tubes, enfoldings (a) of the outer surface of the body lined with ectoderm have also been formed, the funda- ments of the inhalent canal system. Fig. 10 C shows how, by a repeated process of folding, diverticular spaces (6) of mm, -^^^^^^f ;;--::-.:aVv.,. i'.i ■ T.;.'-- .<'■'■ .-.'v'i Fig. 10. — Diagram of the development of the canal system in various sponges. The ectoderm is indicated by a continuous, and the entoderm by a broken, line. The collar epithelium of the entoderm is expressed by perpendicular striation«. A, cross section through a part of the wall of an ascon ; B, cross section through the wall of a sycon ; 0, cross section through the wall of Leucilla connexiva (after Polejaeff) ; D, vertical section through Oscarella ; a, spaces of the inhalent canal system; b, spaces of the exhalent canal system ; on, osculum. 32 EMBRYOLOGY the central cavity arise, in which the fundaments of an exhalent canal system, lined with entoderm, are to be recognized, so that in this manner we arrive by gradual transitions at the distribution of the canals shown in Fig. 10 D, which serves as a plan for most sponges. This series, resulting from the comparison of different genera, makes it probable that the ontogenetic origin of the gastro-canal system by the formation of diverticula from a single common central cavity, as it has been observed in many forms, represents the original mode of development. As to the development of the calcareous or siliceous spicules, it appears to be certain that they are formed within skeleto- genous mesoderm cells. The circumstance that the forms of the spicules frequently merge into one another, and that in many of the rod-like spicules an axial cross has been ob- served in connection with the central canal, suggesting their origin from triaxial spicules, gave rise to certain specula- tions on the fundamental form of the spicules occurring in the different groups and their derivation from the soft parts of the body through simple mechanical conditions. Thus F. E. ScHULZE (Ahh. kgl. Acad. Berlin, 1887) found that the fundamental form of the regular triaxial spicules of the cal- careous sponges was determined by the regular alternating position of the pores in the wall of the original ascon-like animal, and that the peculiar initial quadriradiate form of the Tetractinellidae (as well as that of the sponges derived from them, the Monactinellidse and the horn sponges) were explain- able by the closely crowded position of the spherical ampullae and the resulting forms of the soft parts of the body, whereas in the Hexactinellidoe the arrangement of the trabeculas of the soft parts led to the fundamental form of the regular sexi- radiate spicules. Whereas the hard parts just mentioned develop within cells, the horn fibres must be considered as cuticular secre- tions, for, according to F. E. Schulze, they are deposited on the inner surface of mesoderm cells (so-called spongioblasts) arranged like an epithelium. Non-sexual Reproduction of Sponges. — In this con- nection is to be mentioned the formation, by means of budding, PORIFERA Of new individuals, which remain united to the parent organism throughout life, thereby permitting the formation of extensive colonies. In many cases (Sympagella nux, Hexact.) the single individuals of the colony can readily be recognized as such (No. 4), whereas in the majority of cases their connection becomes so intimate that it is only the pre- sence of the oscula which makes the recognition of the individuals to a certain extent possible. Fig. 11. — Lophocalyx (Pohjloplius) pliiUppinensis with buds (after P. E. Schulzk). a, young bud ; h, older bud constricted ofif from the parent and attached to it by the siUceous spicules of the parent only. Furthermore, a kind of reproduction occurs in many sponges by means of buds, which separate from the parent in a partially developed condition, and grow up into a new sponge organism. A comparatively simple case of this kind appears to exist in Leucosolenia (Vasseur, No. 36.) The young bud is here a simple outfolding of the body- wall, which is soon separated as an independent sac-like body, and'after becoming attached, grows up, and by the production of an osculum becomes a young Leucosolenia. In a similar man- K. H. E. D 34 EMBRYOLOGY ner the budding in Tethya, Tetilla, Rinalda, etc. (Deso, No. 29; Mekejkowsky, No. 31 ; Selenka, No. 34), and also in Lopho- calyxi. e. Polylophus (F. E. Schulze, No. 33), appears to de- pend upon the outgrowth and abstriction of a portion of the parent body, in which is included a part of the canal system of the latter, while the tissues exhibit an active cell- proliferation. The separation from the parent frequently takes place in these cases by an emigration along projecting siliceous spicules (Fig. 11). After separation has taken place, the bud grows up into a young animal resembling the parent organism. To this class belong also the transportable brood- buds of Oscarella described by F. E. Schulze (No. 32), which are very similar in structure to the larva of this form (Fig. 6 D), since they contain a considerable internal cavity. These vesicular bodies, after they have become separated from the parent, are for a time carried about by currents, and then fall to the bottom, where they grow into young sponge- crusts. Reproduction by budding in these forms depends upon the fact that the superficially located parts of the sponge tissue become separated and acquire the power of reproducing the entire form of the parent organism. If we consider a simi- lar process to take place within the tissues of the sponge, whereby the separation of such a group of cells is accom- plished by an encystment, then perhaps the manner is indi- cated by which we are to consider the first formation of gemmulae to have taken place. Reproduction by means of gemmulaB occurs principally among the fresh-water sponges (Spongilla), although the occurrence of gemmula-like structures has also been affirmed for a few marine forms (TOPSENT, No. 35). The fully developed gemmula (Fig. 12) consists of a multicellular germ (d), whose large cells, poly- gonally flattened by mutual pressure, are filled with yolk particles, and present one, frequently two or more, nuclei (Petr, Weltner). This germ-body is surrounded by an envelope often very complicated in structure, which opens to the exterior by means of a pore (p) provided with an oper- cular apparatus. There is always found a thick cuticular layer (c), to which there is generally added externally a PORIFERA 35 porous mesliwork containing air (air-cliamber layer), "in which skeletal elements (needles or amphidiscs) are often found embedded (6), while outside of all there may be added still another cuticular layer (a). Furthermore the germ- body is said to be immediately enveloped in a delicate membrane (Carter). f The gemmulse are found in the midst of the mesodermal tissue of the parent body. Many views have been advanced in regard to their origin. According to Goette (No. 6), there is a kind of cell-prohferation that affects a particular territory and also involves the ampullas and canals of this region, whereas, ac- cording to Marshall (No. /i 30), certain mesodermic cells, filled with reserve food-stuff, creep together in groups to form the gemmulse. The earliest fundament of the gem- ula, which is essentially mass of cells having embryonic characters, soon exhibits a differentia- tion of two layers (Goette, WiEEZEJSKi). The central mass is composed of large cells in which yolk par- ticles become embedded in ever-increasing quantities. The cells of the outer layer, according to Goette, become club-shaped, and arrange themselves into a kind of elongated epithe- lium enveloping the central mass. This layer, like the spongioblasts, at first secretes a thick cuticula on the inside (the fundament of the inner cuticular membrane, Fig. 12 c) ; the amphidiscs are then formed in the cells of this layer, whereupon it moves outward in order to secrete, likewise from its inner surface, the outer cuticular membrane. Fig. 12 a (Goette). According to Wierzejski (No. 39), the amphidiscs are not formed in the layer of cylindrical cells mentioned, but in the surrounding tissue, and only later move into this epithelial layer, in which they become arranged in a definite manner. The formation of the gemmulae takes place principally in the fall in parts of the sponge which generally die after gemmulation has taken Fig. 12.— Gemmula of Spongilla (Ephydatia) fluviatilis (after Vejdovsky). o, outer cuticular layer; h, amphidisc layer; c, inner cuticular layer ; d, germ-body ; p, pore. 36 EMBRYOLOGY place. In the following spring the germ-body crawls out hrough the pore, to become attached and metamorphosed into a young sponge, by means of processes of development which have not yet been accurately studied.* From what has been said it follows that in the gemmnlee we have to do with an encysted portion of the parent body, provided with food-yolk and retrograded to an embryonic condition, which acquires the power to regenerate the parent organism. On this account gemmulation has also been designated as a kind of internal budding. In support of another current interpretation, according to which the gemmules would represent the winter eggs of Spongilla, evidence is as yet entirely wanting, for in this case it would be necessary to prove that the germ- body originates by the division of a single original cell, but all observations made up to the present time are opposed to this. Literature. Earlier writings by Geant, Liebeekuhn, and Miclucho-Maclay. 1. Balfouk, F. M. On the Morphology and Systematic Position of the Spongidse. Quart. Jour. Micr. Sci. Vol. xix. 1879. 2. Baerois, C. Memoire sur I'embryologie de quelques Eponges de la Manche. Ann. sci.nat. {8er.4). Tom. iii. 1876. 3. Cartee, H. J. Development of the Marine Sponges, etc. Ann. and Mag. Nat. Hist. Vol. xiv. 1874. 4. Ganin, M. S. Zur Entwicklung der Spongilla fluviatilis. Zool. Anzeiger. Jahrg. i.. No. 9. 1878. 5. Ganin, M. S. Materialien zur Kenntniss des Baues und der Ent- wicklung der Spongien. Warschau. 1879 (Russian). 6. GoETTE, A. Untersuchungen zur Entwicklungsgeschichte von Spon- gilla fluviatilis. Hamburg u. Leipzig. 188Q. Also Zool. Anzeiger. Jahrg. vii. u. viii. 7. Haeckel, E. Die Kalkschwamme. Berlin. 1872. 8. Heideb, K. Zur Metamorphose der Oscarella lobularis. Arheiten a. d. zool. Inst. z. Wien. Bd. vi. 1886. 9. Keller, C. Studien iiber die Organisation und die Entwicklung der Chalineen. Zeitschr.f. wiss. Zool. Bd. xxxiii. 1880. 10. Marshall, W. Die Ontogenie von Reniera filigrana. Zeitschr.f. wiss. Zool. Bd. xxxvii. 1882. » [Compare Zykoff (Nos. VIII. and IX., Appendix to Literature on Porifera).] PORIFERA 37 Marshall, W. Bemerkungen iiber die Ccelenteratennatur der Spongien. Jena Zeitsckr. Bd. xviii. 1885. Metschnikoff, E. Zur Entwicklungsgeschichte der Kalkschwam me. Zeitschr. f, wiss. Zool. Bd. xxiv. 1874. [13. Metschnikoff, E. Beitrage zur Morphologie der Spongien. Zeit schr. f. iviss. Zool. Bd. xxvii. 1876. Metschnikoff, E. Spongiologische Studien. Zeitschr. f. wiss Zool. Bd. xxxii. 1879. Schmidt, 0. Zur Orientirung iiber die Entwicklung der Schwamme Zeitschr. f. wiss. Zool. Bd, xxv. Suppl. 1875. '16. Schmidt, 0. Nochmals die Gastrula der Kalkschwamme. Arch. f. mikr. Anat. Bd. xii. 1876. 17. Schmidt, 0. Das Larvenstadium von Ascetta primordialis und A clathrus. Arch. mikr. Anat. Bd. xiv. 1877. 18. ScHULZE, F. E. Untersuchungen iiber den Bau und die Entwick lung der Spongien. I. Mitth. Ueber den Bau und die Entwick lung von Sycandra raphanus. Zeitschr. f. wiss. Zool. Bd. xxv Suppl. 1875. 19. Schulze, F. E. Untersuchungen, etc., II. Die Gattung Halisarca Zeitschr. f. wiss. Zool. Bd. xxviii. 1877. 20. Schulze, F. E. Untersuchungen, etc., IV. Die Familie der Aply sinidae. Zeitschr. f. wiss. Zool. Bd. xxx. 1878. 21. Schulze, F. E. Untersuchungen, etc., V. Die Metamorphose von Sycandra raphanus. Zeitschr. f. wiss. Zool. Bd. xxxi. 1878. 22. Schulze, F. E. Untersuchungen, etc., VI. Die Gattung Spongelia. Zeitschr. f. wiss. Zool. Bd. xxxii. 1879. 23. Schulze, F. E. Untersuchungen, etc., VII. Die Familie der Spon- gidae. Zeitschr. f. iciss. Zool. Bd. xxxii. 1879. 24. Schulze, F. E. Untersuchungen, etc., IX. Die Placiniden. Zeit- schr. f. tviss. Zool. Bd. xxxiv. 1880. 25. Schulze, F. E. Untersuchungen, etc., X. Corticium candelabrum. Zeitschr. f. iriss. Zool. Bd. xxxv. 1881. 26. Schulze, F.E. Ueber das Verwandtschaftsverhaltniss der Spotigien und Choanoflagellaten. Sitzungsb. d. preuss. Akad. d. iVissensch. Berlin. 1885. 27. Sollas, W. J. Development of Halisarca lobularis. Quart. Jour. Micr. Sci. Vol. xxiv. 1884. 28. VosjuAER, G. C. J. Porifera in : Bronn's Classen und Ordnungen des Thier-Reichs. Leipzig u. Heidelberg. 1887. Non-sexual Reproduction of Sponges. 29. Deso, B. Die Histologic und Sprossenentwicklung der Tethyen. Arch.f. mikr. Anat. Bd. xvi. 1879 u. Bd. xvii. 1880. 30. Marshall, W. Vorl. Bemerkungen iiber die Fortpflanzungsverhalt- nisse von Spongilla lacustris. Sitzungsb. Naturf. Ges. Leipzig. Jahrg. xi. 1884. 38 EMBRYOLOGY 31. Merejkowsky, C. de. Reproduction des ifeponges par Bourgonne- ment ext^rieur. Arch. d. Zool. exper. et gen. Tom. viii. 1879—1880. 32. ScHULZE, F. E. Ueber die Bildung freischwebender Brutknospen bei einer Spongie, Halisarca lobularis. Zool. Anzeiger. Jalirg. ii. 1879. 83. ScHULZE, F. E. Report on the Hexactinellida collected by H.M.S. Challenfjer, etc. '' CJiallenger'* Reports. Vol. xxi. 1887. 3i. Selenka, E. Ueber einen Kieselschwamm von achtstrahligem Bau und liber Entwicklung von Schwammknospen. Zeitschr. /. wiss. Zool Bd. xxxiii. 1880. 35. ToPSENT, C. Gemmules of Silicispongise. Ahstr. Jour. Roy. Micr. Soc, London, 1888, p. 596. 36. Vasseur, G. Reproduction asexuelle de la Leucosolenia botryoides (Ascandra variabilis H.). Arch. d. Zool. exper. etgen. Tom. viii. 1879—1880. 37. Vejdovsky, F. Revisio faunae Bohemicse P. I. Die Siisswasser- schwamme Bohmens. Abh. d. k. B'dhni. Ges. d. Wiss. z. Prag. Bd. xii. 1883. 38. WiERZEJSKi, A. Ueber die Entwicklung der Gemmulae, etc. (Polish). Ahh. Acad. Krakau. Bd. xii. 1884. 39. WiERZEJSKi, A. Le developpement des gemmules des Eponges d'eau douce d'Europe. Arch. d. Biol. Slav. Tom. i. 1886. Appendix to Literature on Porifera. I. Delage, Yves. Embryog6nie des Sponges ; Developpement post- larvaire des eponges silicieuses et fibreuses marines et d'eau douce. Arch. Zoul. exper. et gen. Tom. x. 1892. II. Dendy, a. On the Pseudogastrula- stage in the Development of Calcareous Sponges. Proc. Roy. Soc, Victoria. 1890. III. Maas, 0. Ueber die Entwicklung des Siisswasserschwammes. Zeitschr. f. iviss. Zool. Bd. 1. 1890. IV. Maas, 0. Die Metamorphose von Epeira lorenzi 0. S. nebst Beobachtungen an Anderen Schwammlarven. Mitth. Zool. Sta. Neapel. Bd. x. 1892. V. Maas, 0. Die Embryonalentwicklung und Metamorphose der Cornacuspongien. Zool. Jahrb., Abth. /. Anat. Bd. vii. 1893. VI. Weltner, W. Spongilledenstudien I. (contains bibliography of 487 numbers) and II. (on gemmules, etc.). Arch. f. ^atltrg. Jahrg. 59. Bd. i., p. 209. 1893. VII. Wilson, H. V. Notes on the Development of some Sponges. Jour. Morph. Vol. v. 1891. VIII. Zykoff, W. Die Entwicklung der Gemmulae der Ephydatia fluviatilis. Zool. Anzeiger, Jahrg. xv. 1892. IX. Zykoff, W. Entwicklungsgeschichte von Ephydatia miilleri aus den Gemmulae. Biol, Centralbl. Bd. xii. 1892. CHAPTER II. CNIDAKIA. JYSTEMATIC : I. HyDROZOA. 1. Hydroidea 2. Siplionophora. II Anthozoa. III. SCYPHOMEDUS^. I. HYDROZOA. I. Hydroidea. The sexual products of the Hydroidea are usually matured in specially organized individuals, which are either free- swimming, and then attain to the high degree of organization of the hydroid medusa, or remain united throughout life with the polyp colony, and then, as sessile medusoid gonophores (Sporosacs), exhibit that organization only in a degenerated condition. In Hydra, on the contrary, the sexual products develop in the ectoderm of the body-wall of the polyp. The eggs of the hydroid medusae are generally extruded, by the dehiscence of the wall of the gonad, into the sea- water, where they are fertilized and undergo development. Bat in those forms which possess sessile gonophores the first stages of development take place within the gonophore, and the embryo does not. become free until it attains the stage of a planula or actinula. In the following account we separate as metagenetic forms those which produce free-swimming medusae (forms with alternation of generations) from those whose sexual indi- viduals remain sessile, as medusoid buds (forms with masked alternation of generations, Hatsghek). A third group em- braces those Hydroids in which there is developed out of the 40 EMBRYOLOGY egg not a polyp, but a free-swimming larva, which passes into the form of a mednsa by a simple metamorphosis Qiypo- gcnetic forms with suppressed alternation of generations). Metagenetic Medusae. — We begin with the description of the better-known cases of development of the eggs of hydroid medusce, and follow principally the accounts of Claus (No. 3) and Metschnikoff (No. 12). The spheroidal eggs of the craspedote medusae are for the most part colour- less, transparent, and destitute of a membrane. There may be distinguished in them a layer of ectoplasm, consisting of a viscid formative yolk, and an endoplasm filled with coarse Fig. 13.— Development of the e^g of Rathkea fasciculata ("after Metschnikoff). A, an egg immediately after deposition ; r, polar corpuscles ; B, stage of first divi- sion ; C, eight-cell stage ; D, blastula-stage in optical section ; E, planula-stage vrith formation of entoderm, en. i granules of nutritive yolk (Fig. 13 A). After fertilizatio they undergo a total, and in most cases a nearly equal, cleavage. By the formation of the first two meridional and mutually perpendicular furrows (extending from the animal to the vegetative pole, Fig. 13 JB), there arise four blastomeres placed in the form of a cross, and by a succeeding equatorial furrow there is produced an eight-cell stage (Fig. 13 0), while two additional meridional furrows lead to the formation of a sixteen-cell stage. Only in certain cases (^quorea) is the cleavage more unequal, the blastomei^s of the animal CNIDARIA 41 zone being less Toluminous than those of the vegetative. In early stages the blastomeres move away from the centre, so that there is formed a gradually enlarging cleavage cavity within. By additional but less regular cleavages the blasto- meres increase in numbers, and arrange themselves into a single layer of epithelial cells surrounding the cleavage cavity, thus attaining the typical hlastula- stage (Fig. 13 D). This cell- vesicle now elongates, so that it becomes ovoid, or sausage-shaped ; and its surface becomes covered with flagella Fig. 14.— Formation of the entodenn by polar ingressioa in the planulae of Aequora (after Glaus, from Hatschbe's Lelirhuch). (Fig. 13 B), by the motions of which it swims about with the broader end of the body directed forward. The forma- tion of the entoderm now takes place by polar ingression, at first a few, and then numerous, cells migrating from the posterior end of the body into the cleavage cavity, so that, advancing from behind forward, they gradually fill it up (Fig. 13 E, Fig. 14). In this way there arises a larva which is eminently characteristic for the Hydroids, and was named by Dalyell the planula (Fig. 15 A)-, it has also been called a parenchymula, on account of the embryonic cell-mass filling its interior (Metschnikoff). During further development there are formed in the ectoderm nettle-capsules, which seem to be especially concentrated about the posterior pole, while within the mass of entoderm cells there arises a fissure, the first trace of the gasiral cavity^ around which the entodermal cells assume an epithelial arrangement. Preparation is now 42 EMBRYOLOGY made for the process of attacliment.i The larvae sink to the bottom, their motions become slower, and finally they attach themselves by means of the disc-like enlargement of the anterior end of the body (Fig. 15 G). They then lose their cilia ; and the surface becomes covered with a thin cuticular secretion, the perisarc (Fig. 15 D). The disc-shaped region of attachment often acquires a lobed appearance, due to the formation of notches (Fig. 15 E). The disc constitutes the first fundament of the hydrorhiza, while the posterior end of Fig. 15.— Attachment and growth of the larval planula of Eiidendrium (after Allman). a, planula ; B and C, stages of attachment by means of the disc-like enlargement of the anterior end ; D, beginning of the formation of the hydranth and perisarc, p ; E, formation of the tentacles ; F, expansion of the hydranth. the larva, now directed upward, grows up into the first hydranth of the young polyp colony. The fundaments of the tentacles (Fig. 15 jEJ) arise as lateral diverticula, and the mouth-opening is formed at the apex by a breaking through * The development of the eggs of medusee and of hydroid polyps with gonophores is from the planula-stage onward alike, so that Eudendrium may serve in this place as an example. CNIDARIA ' 43 of the body -wall. Finally, the perisarc is ruptured at this place, and the polyp becomes fully expanded (Fig. 15 F). The origin of the polyp stock does not always take place exactly accord- ing to the plan here given. In certain cases the larva becomes attached by its side, and is almost wholly employed in the formation of the hydro- rhiza, while the first hydranth grows out of it by a kind of budding (Mitrocoma, Metschnikoff). The hydroid polyp thus produced propagates principally by means of lateral budding. There is formed at first a hernia-like protrusion of the body-wall, the cavity of which communicates with the gastral cavity of the parent animal, and the wall of which consists of the same layers as that of the parent (ectoderm, entoderm, and the sustentative lamella between them).^ The bud is converted into a hydranth by the progressive abstriction of this protrusion from the parent animal, by the production of a crown of tentacles, and by the acquisition of a mouth-opening through dehiscence at the anterior end. Only rarely do the hydranths thus produced detach themselves from the parent, and become independent (Hydra) ; in most cases extensive polyp colonies are formed by continued budding. The laws of budding, by which the extraordinarily manifold shape as well as habit of the polyp colonies is determined, have recently been subjected to a critical study by Weismann (No. 49) and by H. Driesch {Inaug.-Diss., Jena, 1889). The individuals produced by budding do not always have the same form as that of the first hydranth. A more or less pronounced poly- morphism often arises among the individuals of a colony. There are produced defensive polyps (spiral zooids), abundantly supplied with nettle-capsules, but destitute of tentacles and oral opening, hard-shelled spiny protective polyps, nematophores, etc. The so-called blastostyle, which occurs in many Hydroids, is also to be interpreted as a meta- morphosed, non-tentacular polyp, upon the lateral walls of which the gonophores are produced by budding. The formation of the individuals destined to become sexually mature (medusas, sessile medusoid buds) is the ' [Recently Alb. Lang (No. IX., Appendix to Literature on Cnidaria) has maintained that the whole of the bud in Hydroids is derived from the ectoderm of the parent.] 44 EMBRYOLOGY result of a lateral budding, which during its earliest stages pursues a course quite similar to that described above. Here too there is formed at first a small spheroidal bilaminar bud (Fig. 16 A), between the two cell-layers (ectoderm and entoderm) of which there can be recognized a hyaline sus- tentative membrane. The next change, which takes place at the same time with the progressive abstriction of the bud, is the formation of an ectodermal thickening at the free distal pole, which is developed (Fig. 16 B) into a knob, the so-called nucleus of the bud (nucleus of the hell). By the growth of the latter into the interior of the bud, the ento- dermal sac is invaginated, so that it now assumes the form of a cup (Fig. 16 B). While a fissure (the beginning of the cavity of the bell) makes its appearance (Fig. 16 D) in the nucleus of the bud, the two facing layers of the cup-shaped entoderm sac come into close contact, and fuse along four meridians (Fig. 16 E), so that of the cavity of the entoderm sac only four perradial regions (i.e. places cor- responding with the four chief radii) remain open ; they are the fundaments of the four radial canals. It is to be observed that these four radial canals are connected with one another through the remnant of the obliterated entoderm sac (Fig. 16 ^, ^), in other words through the originally bila- minar so-called vascular lamella (cathammal plate) (L. Agassiz, No. 2 ; Glaus, No. 62). While with the further enlargement of the bud the bell-cavity increases in size and breaks through to the outside, and while the manubrium grows out at the bottom of it, the sustentative lamella of the wall of the bell is converted into the gelatinous substance of the umbrella. The radial vessels have become relatively nar- rower and farther separated from one another. The ring- canal at the margin of the bell appears to arise by a secondary separation of the two layers of the vascular lamella. With these metamorphoses and with the breaking through of the mouth-opening and the formation of the velum,^ the medusa is essentially completed and ready to be detached (Fig. 16 F). * The velum does not arise by the formation of a fold, but directly from that ectodermal membrane which in early stages (Fig. 16 D) sepa- CNIDARIA 45 " Alternation of generations " in Hydroids depends upon the regular alternation of non-sexual generations (hydroid polyps), increasing by lateral budding, with a sexually developed generation (hydromedusa or sessile gonophore). The account of the development thus far affords some S| o I ^ ■•-> ^ a) m pq -- o a » CM S o .5 -5 ° -I B ^ i I i s g o " i .2 2 =" S «M § £ — a lie at the lowermost end of the stem are the oldest. In nearly all of the Caljcophoridae and some of the Physo- phoridae (Apolemia) the individuals of the stem are arranged in definite groups (coi'midia), which are separated from one another by free portions of the stem (internodes) . In many other forms, on the other hand, the limits of the successive internodes are indicated merely by the attachment of the polypites with their tentacles (Fig. 29 A, B, C, D), whereas the parts of the stem lying between the polypites are occupied by groups of individuals (consisting of hy- drophyllia, dactylozooids, and gonophores). (In the accompanying figure, for the sake of simplicity, instead of these groups of individuals, only their dactylozooids are indicated.) Here the law that growth pro- gresses uniformly from above downwards applies only to the polypites (J, B, C, D), whereas each internode presents its own zone of gro^vth for the groups of indivi- duals (a, b, c, d) belonging to it, for which in turn the uppermost end of the internode must be looked upon as the budding point, so that likewise in the series of groups of individuals in each single internode the lowermost (a) is the oldest. Each inter- node of the stem is divided by these groups of individuals into internodes of the second order {Aa, ab, he, cd . . .); and each such internode of the second order may, in the further growth of the stem, become a zone of growth for a series of new groups of in- dividuals (a, (3, y . . .) (Chun, No. 57). For the other groups [of Physophoridae] the details of the laws of budding" are as yet little known. In the Vellelidse the for- mation of the individuals takes place in concentrically arranged circles. -J a Fig. 29.— Diagram of Chun's law of budding of the groups of indivi- duals on the stem of Halistemma. In place of the groups of individuals only the corresponding dactylozooids are shown. 08 EMBRYOLOGY Calycophoridae. — The development of Epibulia auran- tiaca (family of the Diphyidae), which has been very accurately followed by Mrtschnikoff (No. 1H), will be described as the type. The ovate planula larva exhibits a thickening of the ectoderm at the posterior pole and on one side (the subse- quent ventral side). Here the fundaments of the first nectocalyx (Fig. 30 J5, nc) and of the tentacle (Fig. 30 B, t) are developed. The former is developed by the invagination of a solid bud-nuclens (Knospenkern), in which the cavity Fig. 30.— Three larval stages of Epihulia ourantiaca (after Metschnikoff, from Balfour's Comiiarative Emhryology). A, planula; B, stage six days old with funda- mems of nectocalyces (nc) and tentacles (t); C, somewhat older stage with gastral cavity ; nc, nectocalyces ; t, fundament of tentacle ; po, polypite ; c, nutritive cells ; to, fundament of the so-called eomatocyst ; hy, entoderm ; ep, ecioderm. of the bell is soon formed ; the fundament of the tentacle at first consists of a simple invagination of the body-wall, in which two layers take part, the development of an ectodermic layer (Fig. 30 B, hi/) along the ventral side, con- sisting of small cells, having already taken place at this stage. The next important change consists in the establish- ment of the gastrovascnlar cavity (Fig. 30 0), which is CNIDARIA 69 correlated with the disappearance of the nutritive cells. Bj means of it the posterior part of the larval body (Fig. 30 C, po) is characterized as the fundament of the first poljpite, whereas the upper dorsal part is retained for a considerable time as an embryonal remnant, which gradually diminishes and is converted into the stem (like the yolk-mass of the Fig. 31. — Older larval stage of Eiiihulia aurantiaca (after Mkischvikoff, from BA.r-FOCK'8 Comparative Embryology), so, somatocyst ; nc, second nectocalyx bud ; hph, hydropbillium ; po, polypite ; t, tentacle. Agalmidoe). At thq same time the fundament of the necto- calyx (Fig, 30 C, nc) has made considerable progress. The hollow core of the bud is enveloped by a layer of entoderm (hy), into which a part of the gas tro vascular cavity is pro- longed as the fundament of the vessels of the bell. Another entodermal process (Fig. 30 C, so) becomes the so-called somatocyst {Saftbehdlter). Between the entoderm and the 70 EMBRYOLOGY outer ectoderm mesogloea has been secreted. In general the development of the nectocalyx is quite like the budding of a Hjdromedusa described above (p. 43). On the funda- ment of the tentacle (t) the individual nettling tubercles can be seen developing as secondary evaginations (Fig. 30 C). The further changes (Fig. 31) consist in a considerable enlargement of the first nectocalyx, which now, after the reduction of the nutritive cells, is the most voluminous structure of the young colony. The polypite (po) now acquires its permanent structure loy the breaking through of the mouth at its distal end, while the tentacle (t), in this case persisting (not larval),' attains its complete develop- ment. Of interest is the appearance of new buds on the rudiment of the stem, first of all that of a hydrophillium (Fig. 31 hph), with the development of which is established the first group of individuals (cormidium) of the subse- quently elongated stem — consisting of a polypite, a dactylo- zooid, and hydrophillium — which is afterwards developed into the Eudoxia. At the same time we see two smaller buds arising, one of which must be considered as the second nectocalyx (Fig. 31 »/c), whereas from the other the elements of the second group of individuals of the stem bud. forth. In the stage Fig. 30 B, which in Fig. 30 C and Fig. 31 undergoes its further development, is shown a larval stage exceedingly characteristic of the Calycophoridse, which has been designated by Haeckel as the Calyconula, and which represents essentially the Siphonula-stage of the Calycophuridce. Haeckel (No. 70) regards this stage as an individual of the second degree (person), and recognizes in its component parts the constituent organs of an Anthomedusa, which here present a remark- able dislocation. For if the nectocalyx corresponds to the umbrella and the polypite to the manubrium of the medusa, then it is evident that the polypite is here attached to the ex-umbrellar side of the medusa-bell. Haeckel explains this dislocation by the assumption of a ventral fissure in the umbrella of the ancestral forms, through which a gradual emigra- tion of the manubrium was possible. Furthermore the only marginal tentacle of the medusa present has moved from the margin of the nectocalyx to the base of the polypite. The assumption that the Siphonula thus characterized actually corre- sponds to an ancestral form acquires an apparent support from the circumstance that the same type of form is found again in the groups of individuals of the stem {conitidia). For the individuals of the stem in CNIDARIA fe Calycophoridse are united into groups and separated by intervals of the stem {inte modes). They bud in such a way that the group of indi- viduals {cormidium) occurring at the lowermost end of the stem is the oldest. In many cases (Polyphyidas, Desmophyidae, Praya, Galeolaria, etc.) the groups of individuals, even if they produce sexual products, remain united with the entire corm. In most of the Diphyidae, on the contrary, the oldest cormidia separate from the parent stock before they arrive at sexual maturity, and either as Eudoxiae or Ersseae lead an independent life. In this way a kind of alternation of generations is brought about, since the parent stock does not itself produce sexual products, but separates into secondary stocks, which do not reach sexual maturity until later on. Such a detached Eudoxia proup (so the cormidia have usually been called) consists of a polypite with tentacles, a hydrophillium, and a gonophore \vhich develops the sexual products in its manubrium, and at the same time by the rhythmical contractions of its swimmmg sack produces the locomotion of the detached Eudoxia. Haeckel explains the hydrophillium, the polypite, and the tentacles as the constituent parts of a sterile person, in which the bilaterally symmetrical hydrophillium would represent the umbrella of the medusa. The Eudoxia cormidium would accordingly in the simplest case be composed of two persons : a sterile one and a fertile one (the gonophore or sexual bell). It is to be observed that the two persons named would represent two essentially heteromorphous medusa of the same corm. Whereas the sterile person exhibits a bilaterally symmetrical structure and the above-mentioned dislocation of the parts, nothing of this kind can be recognized in the fertile person. The structure here is that of an ordinary quadriradial Anthomedusa, and the manubrium has retained its usual place. Leuckart and Gegenbaur have shown that in various Eudoxiae the gonophore, after the discharge of its sexual products, is replaced by the outgrowth of a new gonophore, and Chun showed it to be probable that in all Eudoxiae a quite regular replacement of the gonophores takes place, so that each Eudoxia has quite a number of gonophores developed one after another. Now let us imagine that the first one of these gonophores remains sterile, functioning merely as an organ of locomo- tion ; that would lead to the form of the Ersffiae (in Haeckel's sense). As Ersaeae are designated the cormidia which bud on the stem of Lilyopsis and Diphyopsis, and which, in addition to the parts described for the Eudoxiae, possess a so-called special swimming bell, so that these cormidia, according to Haeckel's interpretation, embrace at least three persons, two sterile and one fertile. The different parts of the nectosome are also subject to a quite similar replacement by supervening buds. Even in the Diphyidae the two necto- calyces are not retained throughout life. Leuckart had already observed the presence of two to three bud-like supplementary bells in Epibulia, and Chun showed that the nectocalyces of the Diphyidae are subject to a con- 72 EMBRYOLOGY stant replacement by reserve nectocalyces of the same form. This replacement also plays a considerable role, as we shall see directly, in the metamorphosis of the Calycophoridse. The metamorphosis of the Caljcophoridse has been made known chiefly through the investigations of Chun (No. 54). These refer prinoipallj to the development of the Mono- phjida?, i.e. those forms which are characterized by the possession of a single nectocalyx on the nectosome. In a small Monophyid called by Chun Muggigea Kochii, and characterized by its tall pentagonal nectocalyx, Chun was able to prove that the larvae arising from the eggs at first possess a quite differently shaped, cap-like nectocalyx. By the casting off of this primary provisional nectocalyx and its replacement by the permanent heteromorphous one, these larva3, designated as Monophyes primordialis, pass into the Muggiaea form, from the stem of which the groups of individuals at sexual maturity detach themselves as Eudoxia Eschscholtzii. Since Chun has recently been able to prove the presence of these primary, heteromorphous, deciduous nectocalyces in the case of the Polyphyidse, it may be considered probable that such nectocalyces also belong to the larval stages of all Calycophoridae. According to Chun's theory, which Haeckel has adopted, the fundament of the pneumatophore in the Physophoridee would be homologous to the deciduous, pri- mary nectocalyx of the Calycophoridae. General Considerations.— As regards the derivation of the Siphonophora, there are at present two views, as yet directly opposed to each other, underlying both of which is the conception that the Siphonophore is a polymorphous animal stock that has arisen by budding. But while some authors (Leuckart, Claus, Chun) assume the starting-point of this stock to be a floating hydroid-polyp stocklet, which already had the power of producing medusae {hydroid theory), others (Balfour, Haeckel) derive the Siphonophore from a medusa, which, by the budding of its manubrium (like Sarsia or Hybocodon), was able to produce new medusse {medusa theory). The former authors accordingly have two funda- mental forms from which they are able to derive the mani- CNIDARTA 73 fold parts of the Siphonophore body. They can consider certain parts (polypites, dactylozooids, etc.) as metamorphosed polypoid individuals, other parts (nectocalyces, hydrophillia, gonophores) as metamorphosed medusoid individuals, which remain united with the colony. The adherents of the medusa theoi'y, on the other hand, have at their disposal only the hydroid medusa as a fundamental form for the derivation of all the numerous polymorphous parts of the Siphonophore organism, for only new medusae can ever be produced from a medusa by budding. Since by this explanation the poly- pites are homologized with the manubria, and the tentacles with the marginal tentacles of a medusa, the adherents of this theory find it necessary to assume an ancestral form in which the medusa exhibited a bilaterally symmetrical structure, while a single tentacle was advanced to the base of the manubrium, and both these parts had emerged upon the ex-umbrellar side of the medusa-bell through a fissure in the umbrella — conditions which, as a matter of fact, do not exist in any Hydromedusa. As a further consequence the partisans of the medusa theory must assume the possibility of a considerable dislocation of these different primary organs and an extensive capacity of the individuals to multiply different organs. With all these assumptions, there arise certain difficulties which are not encountered in the hydroid theory.^ Even if the ancestral form of the Siphonophora assumed by the medusa theory, and described above (which is re- capitulated in ontogeny by the Siphonula-stage and by the sterile person of the Eudoxiae), were to be derived from bilaterally symmetrical Anthomedusce with only one marginal tentacle (for example, from the Hybocodon belonging to Corymorpha), it would still be difficult to point out in any way the causes for the appearance of the fissure in the umbrella and the described dislocation of the organs. The difficulty is increased by the circumstance that these cha- racters are lacking in the sexual individuals of the Siphono- ^ It should be stated that recently Hatschek (Lehrbuch der Zoologie) has introduced modifications into Haeckel's medusa theory, by means of which a part of these difficulties seem to be set aside. 74 EMBRYOLOGY phora, so that, in pursuance of the medusa theory, we are required to distinguish in the Siphonophora two highly heteromorphous genei*ations, the first, produced from the egg, constructed upon the Siphonula type, and reproducing by budding only, the second a generation of fertile indi- viduals not bilaterally symmetrical, and without dislocation of the primary organ. Still sharper, perhaps, is the contrast between the Disconula of the Vellelidae, which is referred by Haeckel to certain TrachomedusaB, and the structure of the Chrysomitras. On the other hand, for the hydroid theory there is the difficulty of explaining how a firmly attached hydroid stocklet could detach itself and become metamorphosed into a free- moving, pelagic organism. If, however, we assume that a hydroid stocklet attached itself by means of a broadened basal plate to the surface of the water, instead of to a fixed body, as may occasionally be observed in the Scyphistomas, and acquiredunder favourable circamsfcances the power to live on in this condition, then with this conception the transition from the attached to the free mode of life is brought about by a floating at the surface of the water, a form of locomotion which has been retained in Physalia and Vellela. Nay, we need only to conceive that the flattened basal part of the stem, which attached itself to the under-surface of the water, curved inwards like a canoe, and finally became, with its peri- sarc-covered face, completely invaginated,^ in order thus to make the phylogenetic origin of the pneumatophore conceiv- able, and to support this conception by the consideration that such a course of development must have been constantly accompanied by certain advantages to the entire colony. Not until after the development of this hydrostatic apparatus would a separation from the surface of the water and a descent into greater depths become possible. The pneuma- tophore would accordingly be the first, most primitive organ by the development of which the characteristic peculiarities ' It has actually been observed in the planula of various Cnidaria that the future point of attachment, which has undergone glandular alteration, is more or less invaginated, as in the Scyphomedusffi and in Eutima (Brooks). I I C NT D ARIA 75 of the Siplionophore organism were establislied. We might perhaps be led by such considerations to recognize in those forms with a persistent apical stigma (Rhizophysas, Physa- lias) the most primitive of the now existing Siphonophores. In this proposed hypothesis of the derivation of the pneumatophore we are opposed to the conception, shared in by most investigators (comp. p. 72 1, that it is a modified medusa-bell. The latter view is founded partly on the structure of the fully developed pneumatophore and partly on its development. Even though the spaces of the gastrovascular system in the vicinity of the pneumatophore, divided as they are by septa, challenge a comparison with the radial canals of a medusa, and even though the bud-like fundament of the pneumatophore is uncommonly like a medusa bud, as has been stated by Metschnikoff (pp. 62, 63), these resemblances do not appear to us to present proofs of a compulsory nature, the more so since the transition from a medusa into a hydrostatic organ involves a change of function that is somewhat difficult to compre- hend. According to our way of looking at it, on the contrary, the apical position of the pneumatophore, sunk into the uppermost end of the stem, and its early appearance in the ontogeny of many forms, are most easily explained. According to our notion, the pneumatophore would be the most primi- tive locomotor organ of the Siphonophora, to which a nectosome would be added only secondarily. Accordingly the Physophoridse would repre- sent the more primitive forms, and the Calycophoridae derived forms, with divergent development caused by the loss of the pneumatophore and the in part higher differentiation of the nectocalyces. Among the PhysonectsB (Haeckel) the Apolemidae, the nectosome of which is still provided with heteromorphous individuals, would perhaps represent the most primitive branch. Opposed to this theory, however, is the fact that histologically the Calycophoridae exhibit the simplest conditions (Korot- neff) ; but these might have been simplified secondarily. When with the above statements we adopt the hydroid theory founded by Leuckart, it is hereby to be understood that, according to our point of view, the existing facts are most easily explained by this theory. Nevertheless we can as yet ascribe even to it only a certain degree of proba- bility. M. ANTHOZOA. Alcyonarla. — The sexual products of the Anthozoa, which arise from the entoderm (Hertwig, No. 9), undergo the process of ripening in sexual organs which belong to the *JQ EMBllYOLOGY mesenterial septa. It is here also that the eggs in most cases are fertilized, and frequently undergo the first stages of development, viz., cleavage and the formation of a spheroidal embrjo consisting of two germ-lajers. The embryo is afterwards set free in the gastral cavity of the parent, from which it is ejected through the mouth-opening, usually in the stage of a ciliated planula. While thus many Alcyonaria are viviparous, cases have also been observed in which the eggs, either unfertilized or immediately after fertilization has taken place, are extruded through the mouth- opening of the parent, either singly or united into large masses by means of a slimy substance (Alcyonium, Renilla, Clavularia crassa). The early development of the Alcyonaria has become known chiefly through La caze-Dl THIERS (No. 88, Corallium), KovvALEVSKY (No. 10, Alcyonium, Gorgonia), v. Koch (No. 86, Gorgonia), E. B. Wilson (No. 98, Renilla), and Kowa- LEVSKY ET Makion (No. 87, Clavularia, Sympodium). The ripe egg of the Alcyonaria is usually rather rich in granules of food-yolk, which, mixed with oil drops, is accu- mulated especially in the inner parts, so that in certain cases there is a sharp separation of a finely granular ectoplasm from an endoplasm rich in food-yolk. Cleavage has been quite variously described for the forms so far observed ; in fact, in Renilla it exhibits remarkable individual variations. In general it follows the total and equal type, and finally leads to the development of a solid so-called morula- stage, consist- ing of cells more or less uniform in size and exhibiting even at an early stage a difference between the more finely granular cells of the superficial layer and the coarsely granular ones of the inner mass. An interestinor modification of the o cleavage process is met with frequently in Renilla, and con- stantly in Clavularia crassa. Here a multiplication of the cleavage nuclei first takes place, corresponding to which there is only an indentation of the surface, not a real cleav- age of the egg. This does not take place until there are sixteen cleavage nuclei, when it results in the formation of the same number of separate blastomeres. We see that we here have to do with a variation which forms a transition CXIDARIA « '^^^^^F "^^ to the type of superficial cleavage, which is wide-spread among the Arthropoda. In general the cleavage stages of the Alcyonaria are characterized by the absence of the cleavage cavity. Monoxenia forms an exception. Here, according to Haeckel (No. 78), there are produced in the course of a very regular cleavage a typical coeloblastula-stage and a gastrula in- vaginata. In the mornla a difference can early be recognized (Fig. 32 A) between a superficial cell-layer (ectoderm) and an inner cell-mass (entoderm). This difference becomes more marked in later stages (Fig. 32 B, C). The ectoderm cells by progressive division are mefamorphosed into prismatic elements, which constitute a columnar epithelium (Fig. 32 C). Those of the inner cells lying next to the ectoderm also arrange themselves (Fig. 32 C, en) into an epithelial layer (the permanent entoderm), whereas the elements lying at the centre undergo a process of degeneration. The cell boundaries here become indistinct ; vacuolar spaces make their appearance, and soon coalesce into a common internal cavity (the beginning of the gastral cavity, h) ; finally, this entire cell-mass is metamorphosed by fatty degeneration into a kind of detritus (d), which is gradually resorbed. At the same time a fine, structureless, hyaline membrane (the sus- tentative lamella) is secreted between the ectoderm and the permanent entoderm. While these internal changes are taking place, the body elongates and gradually assumes an ovoid or, with increasing length, a vermiform shape, and its surface becomes covered with close-set cilia; thus the swarming planula-stage is developed (Fig. 32 T)). The planula exhibits a somewhat broadened (aboral) end, which is directed forwards during motion, and a posterior (oral), more pointed pole. At the expiration of the swarnting stage the larva attaches itself by means of its broadened anterior end to some convenient support. By a gradual shortening in the direction of the longitudinal axis the larva passes from the elongated into a low placenfciform shape (Fig. 33). At about the same time with the attachment, the in- 78 EMBRYOLOGY vagination of the oesophageal tube (si) takes place, and also the formation of the eight mesenterial septa. The oesophagus arises in general as an ectodermal invagination (Fig. 33), the bottom of which in later stages breaks through toward the gastral cavity, thereby establishing the inner opening of the oesophagus. The formation of the mesenterial septa is referable to a folding of the entoderm, in which the sustentative lamella also takes part. It seems that in the Fig. 32.— Stage in the development of Sijmpodium coralloides (after Kowalevbkt KT Marion). A and B, cleavage stages; C, embryo with permanent entoderm (en) and inner mass of detritus (d), in which beginnings of the gastral cavity {h) can be recognized; D, ciliated planula; ec, ectodei-m ; en, entoderm. Alcyonaria all the eight mesenterial septa always make their appearance at the same time. As regards the origin of the musculature of the septa, especially the longitudinal muscles, authors agree that they arise from epithelio-mus- cular cells (myoblasts) of the entoderm lamella. CNIDARIA 79 In Renilla the oesophagus is developed in the form of a solid ectodermal ingrowth, in which a fissure makes its appearance and opens to the exterior, whereas the development of the inner opening of the oesophagus idoes not take place until later. Fig. 33.— Attached stage of Sympodium coralloides (after Kowalevsky bt Marion). Ec, ectoderm ; en, entoderm ; si, oesophagus. By the formation of the mesenterial septa the gnstral space is separated into a central stomach cavity and eight peripheral gastral pouches. At the upper ends of the latter hollow, bud-like elevations now arise, in which we recognize the earliest fundaments of the eight (subsequently pinnate) tentacles, which accordingly owe their origin to simple evaginations of the body-wall. The dev^elopment of the septa, the formation of the oeso- phagus, and even the establishment of the tentacles may take place before attachment. In general, however, the attachment of the swarming larva precedes the formation of these organs. By means of the developmental processes mentioned the typical structure of the polyp is established. During these metamorphoses important changes in the struc- ture of the ectoderm take place. By multiplication of the cells this layer becomes changed into a multi-layered epithelium. The secretion of a hyaline gelatinous substance [mesogloea] now takes place between the cells of the deeper layers, which thus lose their connection with one another and assume more and more spindle or stellate shapes (Fig. 34). By these processes two different layers arise from the primary ectoderm : a superficial one, which from now on preserves the name of ectoderm, and the cells of which have retained the enithelial continuity, and a lower layer, which assumes more and more the character of a gelatinous connective tissue, and which will be called henceforth mesoderm. This layer accordingly is a product of the 80 EMBRYOLOGY ectoderm (Kowalevsky et Marion, No. 87). In it are secreted the first calcareous spicules (sclerites) (v. Koch, Nos. 82 and 84, Kowalevsky). These arise as small, highly refractive bodies (sp) within the mesodermal elements, which resemble migratory cells, where they soon grow into small needles having lateral outgrowths. The ectodermal axial skeleton of the Gorgonidoe arises later than these mesodermal parts of the skeleton. It must be regarded as a cuticular secretion of the ectoderm of the basal foot-plate (v. Koch), and at its first appearance consists of a thin yellowish pellicle, which may be compared to the sheath of Cornularia and Clavularia. There is soon noticeable on this basal plate a small prominence, which grows up into a process composed of concentrically arranged corneous lamellce, and extends up between the mesenterial septa of the Fig. 34.— Section through the body-wall of a young attached stage of Si/;»i- •pod^ium coval\o{i.es (after Kowalevsky et Maeion). ec, ectoderm ; en, entoderm; g, mesogloea; &y), earliest fundament of the calcareous spicules in cells of the developing mesoderm. primary polyp. Thus the ectoderm of the foot-plate must be correspondingly invaginated, and thus it comes about that the axial process of the ectoskeleton contained within the polyp is covered by a continuous ectodermal lamella (the axial epitheliuin), from which the further development of this part of the skeleton takes place. In the further coarse of development, during the progressive gi'owth in length, the young polyp and the axial skeleton do not take the same direction; the latter thereby acquires greater independence, and represents the earliest fundament of the whole axial skeleton which lies at the foundation of the entire colony subsequently produced by budding (Fig. 36 B) (v. Koch, No. 86 1). 'On the other hand, Stucer -{Arch. f. Naturg. Jahrg., 1887) has CNIDARIA 81 I To explain the phylogenetic development of this axial skeleton of the )rgonidae, v. Koch (No. 85) has for comparison made use of the iteresting discoveries on Gerardia (Antipatharia, Hexacorallia). These colonies of Gerardia form flat membranoid coverings over foreign bodies, and for this purpose commonly select the axial skeleton of dead Gorgonidse as a support. A lamella of horn, which coats the support, is now secreted by the ectoderm of the lower surface of these colonies. The lamella surrounds the axis of the Gorgonia within it like a sheath. When at length the colony of the Gerardia by growth acquires an extent which stretches beyond the limits of the original support, then out- growths covered with young polyps are produced, into which extend horny skeletal processes, produced by the common basal lamella, but no longer enclosing within them any foreign body. It is seen that here is produced the first trace of an independent free axial skeleton, while the basal plate of the skeleton, which in the higher forms is much reduced and attached to a foreign support, arises from the basal lamella. Fig. 35,— Two transverse sections through a polyp of the Alcyonarian type (diagram after v. Koch, from Lang's Lehrb\xc\i); that at the left is at the level of the oesophagus, that at the right at the level of the gastral cavity, ab, plane of symmetry. The ventral side is directed upwards. The polyps of the Alcjonaria present a typical bilaterally symmetrical structure, which is evident in the first place from the position of the longitudinal muscles in the mesen- terial septa. Here the plane of symmetry (Fig. 35 ah) passes through two unpaired chambers (gastral pouches), which are distinguished from each other by the fact that the two septa which bound the ventral chamber exhibit the muscle ridges on the sides which are turned toward each other, whereas this condition is reversed in the dorsal cham- ber. On the remaining septa, in fact on all the septa, the longitudinal muscle ridges are so arranged that they face recently defended the interpretation of the axis of the Gorgonidse as a mesodermal growth. K. H. E. G 82 EMBRYOLOGY toward the ventral side of the polyp, whereas the surfaces of the septa which are without longitudinal muscle bands face toward the dorsal side. The bilateral symmetry can also be recognized by the presence of a ventral ciliated groove running along the laterally compressed oesophagus (siphono- glyphe, Hickson), and above all by the condition of the mesenterial filaments. Of these the pair belonging to the dorsal septa diifers from the others in structure, function, and development. The filaments of the dorsal pair of septa exhibit an epithelial band consisting of tall flagellate cells, and produce a powerful upward ciliary current, whereas the filaments of the other six septa are charac- terized by their richness in gland cells, and they play an important role in digestion. E. B. Wilson (No. 97) was able to show that the latter take their origin as simple out- growths of the entodermal epithelium of the septa, whereas the dorsal filaments belong to the ectoderm, and are con- tinued on to the margins of the septa as direct outgrowths of the oesophageal epithelium. An observation by Wilson is of general interest : that the develop- ment of these dorsal filaments is retarded in the larvae produced from the egg, whereas in the bud they actually outstrip the other filaments in development. Wilson explains this by the conditions of nutrition in the bud, which requires a powerful upward stream of nutritive fluid for its development. Of the various kinds of non-sexual reproduction in the Alcyonaria, budding is the most prevalent ; by means of it extensive colonies (stocks, corms) are developed, owing to the fact that the newly arising individuals remain united with the parent. In the simplest case a lateral " runner " arises from the parent animal and grows out at its end into a daughter individual. The portion remaining between the two as a connective is called a stolon (Fig. 36 A). These stolons, issuing from the base of the polyp, may form a net- work (Cornularia), or fuse into a basal plate (Rhizoxenia). We have seen above (p. 81) how, owing to the formation of a basal skeletal plate upon which an axial skeleton arises, the dendritic stocks of the Gorgonidse can be derived from such flatly extended colonies (Fig. 36 B). In other cases CNIDARIA stolons do not belong exclusively to the basal part of the polyps, but arise at various levels. In this way the peculiar colonj^ of Tubipora (Fig. 36 C) arises by the de- velopment of stolonic plates in higher positions, from which new buds grow out. In other forms, by the intimate fusion and irregular branching of the stolons, there is developed an intermediate tissue (coenenchyma) traversed by numerous nutritive canals (Fig. 36 D), which unites the different indi- viduals. In this way the antler-like colonies of Alcyonium ai"e developed, and by the formation of a mesodermal axial Fig. 36.— Diagrams of budding and stock-formation in the Alcyonaria (after V. Koch, from Lang's Lehrhuch). A, formation of the basal stolon ; B, type of the Gorgonidae ; C, type of Tubipora ; D, type of Alcyonarium ; s, oesophagus; se, septa ; mf, mesenterial ridges ; dh, gastral cavity ; sfc, axial skeleton growing upwards by means of successive layers. skeleton the more slender forms, such as Corallium, Scleor- gorgia, Melitheea, etc. (v. Koch). The development of colonies by budding is of special interest in those forms in which, owing to the regular orien- tation of the daughter individuals to the parent polyp, there is established a regular bilaterally symmetrical structure of the entire colony (Pennatula, Renilla). In these forms a well-marked polymorphism of the individuals is exhibited, g4 EMBRYOLOGY inasmuch as polyps which bear tentacles and become sexually mature [autozooids] can be distinguished from sterile individuals lacking tentacles and having only two septa, the so-called zooids [siphonozooids], which provide for the inflowing of the water (Wilson). The development of Eenilla has been investigated by E. B. Wilson (No. 98). Attachment is here suppressed, and by the invagination of the oesophagus and the development of the septa and tentacles there is produced from the planula- ^^^^^.^ dx- iJ Fig. 37.— Two stages of development of Eenilla (after E. B. Wilson). A, young polyp with two polyp buds (p^) and the terminal zooid {z) ; B, central portion of a somewhat older stage ; pi, p^^ p^, p*, polyp buds; z, terminal zooid; mz, marginal zooid ; d2, dorsal zooid. larva a free-moving polypoid form (Fig. 37 A)^ which, in view of the development of the colony, can be called the axial individual. The upper portion of this individual persists as the terminal polyp, whereas the stem of the entire colony (rachis) and its lower free part, the stalk {peduncle), arise from its middle and lower portions. We may also retain for Renilla these terms, which are borrowed from the Pennatu- lida^, because a striking similarity between these tw^o forms is established in their embryology. The eight septa of the CNIDARIA 85 axial individual are developed in the anterior part of the polyp, and grow from in front backwards ; nevertheless they are restricted, even in late stages, to the anterior parts of the individual, whereas in most Alcyonaria the septa extend as far as the posterior end of the body. On the other hand, another wall is developed in Renilla by a transverse infold- ■wz Fig. 33.— Older stage of development of the colony of 'RemWa (after E. B. Wilson). J), terminal polyp ; z, terminal zooid ; mz, marginal 20oid; dz, dorsal zooid. ing of the entoderm from the posterior end of the body, tlie so-called 'peduncular septum^ by means of which the gastral cavity is divided into a ventral and a dorsal half. The peduncular septum grows from behind forwards ; and since it grows more actively at its lateral parts, its anterior margin assumes a curved form. Between the two entodermal layers of the peduncular septum is found a cell-mass which subse- quently degenerates, and which is apparently homologous to 86 EMBRYOLOGY the skeletogenous layer of the Pennatalidae, but which is said by Wilson to arise from the entoderm. At an early period the budding of the daughter individuals begins ; these are formed strictly in pairs on the dorsal side of the axial individual (Fig. 37 A, p^). The second pair of polyp buds arises immediately behind the two first ones, the third pair in front of and somewhat ventrad from the Fig. 39.— Young colony of Pennatula phosphorea (after Jungersen). A, youngest stage, seen from the right ; B, older stage, from the ventral side ; C, the same, Irom the dorsal side ; j), terminal polyp ; z, terminal zooid ; p^, pi, polyps of the first pinnate leaflet ; p2, p2, polyps of the second pinnate leaflet, etc. first pair, the fourth pair in the angles between the third pair and the axial polyp (Fig. 37 B, p^, p^, p^, p^). The buds arise separately at first, but subsequently their basal parts fuse. The individuals that have arisen in this way very soon assume a radial position ; and since the buds that appear later are formed in alternating positions and ventrad to those first formed, and since in the further course of de- velopment they grow so actively that they project beyond I CNIDARIA 87 these at the periphery, it follows that the oldest individuals are more and more crowded toward the dorsal side (Fig. 38). The terminal polyp also shares this fate. In this way is developed a discoid colony, the marginal individuals of which are the youngest. The zooids are formed at the same time as the sexual polyps. Even immediately after the appearance of the first pair of polyp buds, a large termi7ml zooid (Fig. 37 z) can be recognized ; this functions as an excurrent opening, and is soon followed by the so-called marginal zooids (mz), arranged in two lateral dorsal rows, while dorsal zooids {dz) make their appearance on the dorsal side of each of the individual polyps. As far as the development of Pennatula is at present known, it is strikingly similar to that of Renilla. Lacaze- DuTHiERS (No. 90) has made some statements on the earliest stages of Pteroides (Pennatula) griseum; the later stages, relating to budding, have been described by Jungersen (No. 81). Here also we find lying at the foundation of the colony an axial individual which is retained for a considerable time as the terminal polyp, and on the sides of which bud forth the daughter individuals, which appear in pairs, but alter- nating with one another. At the bases of these lateral polyps, that are the first to appear, and in positions corre- sponding to the ventral side of the axial individual, new buds continue to arise, thereby introducing the development of the pinnate leaflets, of which accordingly the dorsal indi- vidual exhibiting the greatest length is the oldest. On the dorsal side of the axis we find an unpaired terminal zooid and other zooids which are arranged in two rows. The lateral zooids, which belong to the ventral surface, are not developed until later. In the young stages the terminal zooid probably functions as the only excurrent opening. In the older stages, on the other hand, there is found at the upper end of the axis a group of apical zooids, among which are probably to be found the terminal zooid and the degene- rated terminal polyp, as well as the adjoining polyps, these having assumed the function of the terminal zooids. In the peduncular septum^ which here also divides the gas- 88 EMBRYOLOGY trie space of the axis into a dorsal and ventral canal, there is found a calcareous axis (ectodermal according to v. Koch's conjecture), surrounded by an axial epithelium, and two lateral canals lying at the sides of the former, which, as nutritive or sap canals, belong to the gastrovascular system. From the embryology it appears that the older authors have employed the expressions " ventral " and " dorsal " for the Pennatula colony in the opposite sense to that which is admissible according to the orientation of the axial polyp (Jungersen). Zoantharia. — In the majority of cases fertilization and cleavage take place inside the mesenterial septa, and the further development, as far as the complete formation of the planula, in the gastral cavity of the parent. In this stage the larvae are cast out through the mouth-opening. On the other hand, Cerianthus membranaceus and Actinia parasitica (Adamsia Rondeletii), according to Kowaletsky, eject the spawn in an unsegmented condition. Considerable uncertainty still prevails regarding the earliest developmental processes, the knowledge of which we owe chiefly to Kowalevsky (No. 10), Jourdan (No. 80), and H. V. Wilson (No. 99). In many cases cleavage and the differentiation of the entoderm seem to take place in connection with the formation of a solid morula, therefore in a manner similar to that which has been described for the Alcyonaria. At least there is in support of this Kowa- levsky's observation on Actinia parasitica (Adamsia Ronde- letii), which is described in the following manner: "Cleav- age is regular, but as the result of it there arises not a blastodermic vesicle, but only an aggregation of cells, which becomes covered with cilia, and swims about as a larva; subsequently a small depression is formed at one spot. The opacity of the eggs made a further pursuit of the develop- ment impossible." The author is convinced that the ento- derm in this case is not formed by invagination, but by a splitting off from the blastoderm, as in the Corallia. In sections through ciliated larvae of Astrsea Kowalevsky found the two layers, ectoderm and entoderm, composed of cylin- drical cells, and an inner contained mass, which had obvi- CNIDARIA ^HBP ^^ ously arisen from cells, biifc whieli now showed that it was composed of nuclei and fat spherules only. A similar struc- ture of the planula is also described for Actinia aurantiaca and Balanophyllia regia ; Jourdan's observations show, how- ever, that from the presence of an inner mass filling up the planula we are not at all justified in inferring the origin of the mass from a solid morula. Balfour refers to observa- tions of Kleinenberg according to which the cleavage of the Zoantharia is frequently unequal ; this would allow one to infer the formation of an epibolic gastrula. Accordingly the formation of the entoderm by delamination from a solid morula in this case still appears doubtful. In another series of cases the development of a unilaminar ciliated blastodermic vesicle has been observed, from which the gastrula-stage is produced by invagination ; thus in an edible Actinian from Faro (Messina), closely related to Actinia mesembryanthemum, observed by Kowalevsky. Here the blastopore does not close completel}^, but is directly converted into the inner opening of the oesophagus, while the oesophagus, lined with ectoderm, is developed by the en- folding of the margins of the mouth-opening. In Cerianthus also the formation of a coeloblastula and an invaginate gas- trula following total unequal cleavage was observed by Kowalevsky. Probably Caryophyllia also belongs here. In Actinia equina, according to Jourdan, there is formed a typical in- vaginate gastrula, whose gastral cavity is at first completely empty, and whose entodermal cells contain but little food-yolk. Nevertheless the stomach of the planula larva is filled with coarse yolk granules It still remains uncertain whether these are produced by secretion or by the partial disintegration of the cells of the entoderm. According to the observations of H. V. Wilson on Manicina areolata, first a coeloblastula is formed by total cleavage. Then, by the transverse division of the tall cells of the blastosphere — consequently by delamina- tion—coarsely granular cells are repeatedly constricted off, and finally fill completely the cleavage cavity. While the ectoderm becomes some- what more sharply marked off from the inner cell-mass, the oesophageal invagination arises. The larva now becomes covered with cilia and swims about. The permanent entoderm arises, as in the Alcyonaria, from the inner cell-mass, the cells lying next to the ectoderm arranging themselves into an epithelium, while the central mass is finally resorbed. At any rate, through these various processes of develop- 90 EMBRYOLOGY menfc there always arises the same larval form, with identical structure : a bilaminar, thickly ciliated, oval, pyriform or more elongated vermiform planula, which possesses an ectoderm composed of prismatic or columnar cells, an entodermic epithelium consistino^ of large cubical elements, and a homogeneous membrane (sustentative lamella), which is secreted between the two layers at an early period. The internal cavity of this larva (gastral cavity) is in most cases still filled with masses of food-yolk. In this swarming stage there can be recognized a broader, anterior, aboral end of the body, which subsequently serves for attachment, and is fre- quently characterized by a long tuft of cilia and a narrower posterior end ; here the oesophagus is formed by invagina- tion, and at its deepest part a communication with the gastral cavity is produced by resorption of the cells. The further development takes place principally by the formation of the mesenterial septa, the filaments, the tentacles, and, finally, in the Corallia (Madreporaria), the calcareous skeleton. As regards the sequence in the development of the septa, the views expressed by Milne-Edwards et Hatme, based chiefly upon the condition of the tentacles and calcareous septa of the adult animal, were formerly generally accepted. According to them, first six primary septa are simultaneously developed, then six of the second order in the interspaces between these, then twelve septa of the third order, twenty- four septa of the fourth order, and so on, the septa of each newly appearing cycle being interpolated, as was maintained, between those already present. On the other hand, we owe to the investigations of Lacaze-Duthiers (No. 89) the know- ledge that this regular arrangement, which is based on the number 6, is a secondary one, and that the septa of a cycle are formed at difl^erent times, becoming equalized only sub- sequently. Most important of all in the earliest stages is a well-marked bilaterally symmetrical condition, and the stages with four and with eight septa are to a certain extent well marked, whereas the intermediate stage, with six primary septa, is a very transitory one. As regards details, the statements of Lacaze-Duthiers on the sequence in the CNIDARIA 91 development of the pairs of septa first to appear must be modified in accordance with the conjectures of 0, UNr> R. Hertwig (No. 9), which have been confirmed by the observa- tions of H. V. Wilson (No. 99) and others. The sequence in the development of the different pairs of primary septa is consequently as follows. At first a pair of septa arises which is placed nearly at right angles to the elongated oral fi^ssure which marks the plane of symmetry (Fig. 40 i). This pair of septa is formed as a longitudinal fold of the entoderm, inside of which there extends a process of the gelatinous sustenta- tive lamella. By the development of this first pair of septa, which lies nearer to one oral angle than to the other, the peripheral part of the gastral cavity is separated into two gastral pouches, one of which is smaller than the other. By means of the second pair of septa (Fig. 40 2) the larger of the two pouches is separated into three parts. The third pair of septa is developed in the smaller of the two primary gastral pouches, and divides this in like manner into three parts, whereas the fourth pair of septa is developed in the unpaired pouch which is enclosed by the septa No. 2 (Fig. 40 3 and 4). This stage with four pairs of septa marks a kind of resting phase in the development. Up to this time the septa were always established in pairs, and in such a way that each new pair was developed in one and the same gastral chamber. For the pairs which now follow, Nos. 5 and 6, the statements of H. V. Wilson (No. 99) and A. C. Haddon (No. 77) agree with those of Lacaze-Dqthiers to the effect that they take their origin in the two pairs of chambers which lie next to the pair of septa first formed. Accordingly the septa of these two pairs would make their appearance independently in four different gastral pouches (Fig. 40 B). On the other hand, the brothers Hertwig (No. 9) have observed in Adamsia diaphana another mode of development of these two pairs of septa, both of which here arise in the chambers lying between septa 1 and 2 (Fig. 41). Accordingly even in the Hexactinise alone dift'erent conditions seem to prevail regarding the arrange- ment of the longitudinal muscles on the first eight septa and the development of the fifth and sixth pairs of septa.^ ^ [The recent investigations of Boveri (No. III., Appendix to Literature 92 EMBRYOLOGY The twelve primary septa now arrange themselves in six pairs, each of which encloses an intraseptal chamber (Fig. 42). Two pairs of septa, called directive septa, lying opposite to each other and corresponding in position to the angles of the month (Fig. 42 3 and 4), bear the longitudinal muscles on the sides which are turned away from each other, all other pairs of septa on the sides which face each other. The gastral pouch lying between any two intraseptal chambers is called an interseptal chamber. New septa are never developed in the intraseptal chambers. They always appear in pairs, and from now on in the interseptal chambers and in cycles based on the number 6. ^3 ^^ 3 Fig. 40.— Diagram of the growth of the septa in Hexactinians. A, stage of Manicina areolata with eight primary septa in cross-section (after H. V. Wilson) ; 1, oldest pair of septa, which is in connection with the cesophagus ; ec, ectoderm; en, entoderm ; s, sustentative lamella ; /, mesenterial filaments ; rf, part of the ectoderm of the oesophageal tube tha'. is bent outwards and backwards at the free end of the tube ; B, stage of Aulactinia stelloides with twelve primary septa (after McMubbich). on Anthozoa) are especially important in this connection. Boveri confirms the existence of both the above-mentioned types of septal growth in the Hexactinia, of which the one was made known by Lacaze-Duthiers, the other by Hertwio. In agreement with Haddon, McMurrich, and Dixon, Boveri places special importance on the presence of an Edwardsia stage in the ontogeny of the Hexactinia, and is inclined to regard the Edwardsia type as the phylogenetic starting-point of all the groups of Actinia, an opinion against which doubts have recently been raised, so far as re- gards the Ceriantheae and Zoantheae, by E. van Beneden (Nos. I. and II., Appendix) and Carlgren (No. IV., Appendix).— H.] CNIDARIA 93 The condition described in regard to the growth of the septa applies to the Hexactinia and probably to all Hexacoralla. On the other hand, there are a number of groups among the Actiniaria in which other laws of septal growth prevail, which furnish characters of systematic importance (R. Hertwig). In the Paractinice (Sicyonis, Polyopsis) there are found two pairs of directive septa, as in the type described above, and the rest of the septa also make their appearance in pairs. On the other hand, the num- ber of the septa is not fixed by the numeral 6. The Edicardsidee (Fig. 43 A), like the Hexactiniag, exhibit two oesophageal grooves [siphonoglyphes] and two pairs of directive septa ; never- theless the arrangement of the longi- tudinal muscles on the septa indicates a bilaterally symmetrical structure, as opposed to the biradial condition of the adult Hexactiniae. Of the eight septa present, of which only the directive pair exhibits a paired grouping, six bear their longitudinal muscle bands on the side directed 4> J 3 Fig. 41.— Transverse section of a young Adamsia diaphana (after O. UND R. Hektwig), diagrammatic. The pairs of septa 5 and 6 are iu process of development. 3 3 Fig. 42.— Diagram of the further growth of the septa in the Hexactiniae. Of the numbers at the left 1 to 5 refer to the type of development of Adamsia (comp. Fig. 41), the numbers (I.) and (FI.) to the type of development of ^ulactinia (comp. Fig. 40 B). At the right, I. to IV. indicate the pairs of septa of the first to the fourth cycle ; r, r, oesophageal grooves [siphonoglyphes]. 94 EMBRYOLOGY toward the ventral surface of the animal, whereas the ventral pair of directives exhibits the longitudinal muscles on the opposite side. It is worthy of consideration that, according to the coinciding observa- tions of A. C. Haddon (No. 77) on Halcampa and Peachia and J. P. McMuRRiCH (No. 91) on Aulactinia, the position of the muscles on the first four pairs of septa agrees with the arrangement in the Edwardsidse (comp. Fig. 40 B), so that accordingly in the ontogeny of some Hexactiniae an actual Edwardsia stage is passed through. A bilaterally symmetrical type is also developed in the gi'oups which now follow. In the Monaulea (Fig. 43 B) the dorsal pair of directive septa is lacking, whereas in the paired arrangement of their septa they approach the Hexactinise. The Zomithece (Fig. 43 C) also exhibit a paired arrangement of the septa, but each pair consists of two unequal septa : a small microseptum, not reach- Fig. 43.— Diagram of the position of the septa— ^, in the Edwardsidae ; B, in the Monaulese : C, in the Zoantheae : D, in the Cerianthese. ing to the oesophagus, and a larger macroseptum, extending to the oesophagus. The two pairs of directive septa constitute the only excep- tion to this, the dorsal pair exhibiting only microsepta, and the ventral only macrosepta. The remaining mixed pairs of septa are so arranged that they fall into a dorsal and a ventral group. In the dorsal group, which always consists of only four pairs, each pair turns its macroseptum toward the dorsal pair of directive septa. The number of pairs of the ventral group is usually considerably greater, and is increased by the appearance of new pairs next to the pair of ventral directive septa (at x in the two adjoining interseptal chambers). Here, therefore, only two interseptal chambers function as formative seats of new pairs of septa. CNIDARIA 95 Each pair of these ventral groups turns its macroseptum toward the ventral pair of directive septa. Finally, in the Gerianthea (Fig. 43 D) only one oesophageal groove [siphonoglyphe] is found. Here the numer- ous septa are not arranged in pairs ; two particularly small septa attached to the base of the oesophageal groove (A. von Heider) may be called direc- tive septa. The septa lying at either side of them are the largest, and from here the septa continually decrease in size toward the dorsal side, so that it is probable that the zone of growth of new septa is situated at this place (Hertwig). That the number of groups is possibly not concluded with the types described, is proved by Gonactinia, which represents a peculiar type allied to the Zoantheae (Blochmann und Hilger, No. 74). With respect to the development of the mesenterial fila- ments, H. V. Wilson (No. 99) has proved, at least as far as concerns the filaments of the twelve primary septa, that they take their origin as outgrowths from the ectodermal epithe- liam of the oesophagus. Even earlier A. vON Heider, on the basis of histological agreement, had argued for the ectoder- mal nature of the filaments in Cerianthus, and E. B. Wilson had conjectured that at least the lateral ciliate bands (Flim- erstreifen) belong to the ectoderm. A. Andres also believed that he had convinced himself that the filaments of the six principal septa take their origin by means of outgrowths from the ectoderm of the oesophagus. According to the observations of H. V. Wilson on Manicina areolata, it is to a certain extent probable that not only the lateral ciliate bands, but also the nettle- and gland-cell bands {Nessel- drmenstreifen), arise from the ectoderm. With respect to the more detailed processes of development, the mesen- terial filaments of the first pair of septa differ from those appearing later. The establishment of the first pair of septa and the filaments belonging to it takes place in Manicina areolata at a time in which the space between the oesophagus and the body-wall is still filled throughout by a solid mass of entodermal cells. This cell-mass encircling the oesophagus ^Kis divided into two parts, corresponding to the two primary gastral ^^■>ouches, which are subsequently hollowed out. This division is effected ^^py the formation between the oesophagus and the body-wall of two par- ^^Kitions of the sustentative lamella, which constitute the foundation of the ^^^nrst pair of septa. It takes place in this way : the oesophagus approaches HHue body-wall until it comes in contact with it, then its sustentative lamella fuses with that of the body-wall ; when subsequently the oeso- Iphagus again separates from the body-wall, a bridge of the sustentative jlamella is preserved between the two. While the fundament of the first ! 96 EMBRYOLOGY pair of septa is formed in this way, the development of the filaments takes place by simple downgrowth of the ectoderm of the oesophagus, in the direction of the two primary septa. The two gastral pouches first to appear are now completely hollowed out. The new pairs of septa next arise as foldings of the entodermal lamella of the body-wall, and their upper ends seem to be at some distance from the ectoderm of the oeso- phagus, so that no direct outgrowth of the latter can lead to the formation of the filaments. In order to establish the connection between the ecto- derm of the. oesophagus and the newly formed septa, the former must bend around at the inner opening of the oesophagus, and grow upward on the outer surface of the oesophagus, until it reaches the uppermost part of the newly formed septa, on to which it now advances to form the filament. This bent-over part of the ectoderm is seen in Fig. 40 A, rf. H. V. Wilson conjectures that the mesenterial filaments of all subsequently appearing septa are formed after this type. In general the development of the mesenterial filaments takes place in the same sequence as that of the septa, so that the oldest pair of septa bears the most developed filaments. The tentacles arise as simple evaginations of the body-wall over the different gastral poaches. The sequence of their origin has been described by Lacaze-Duthiers (N^o. 89), especially for Actinia mesembryanthemum. For the early stages it is closely connected with, the sequence of the appear- ance of the different mesenteries and the formation of the chambers dependent on it. In this connection ought speci- ally to be mentioned the fact that the tentacle which arises over the larger of the two first-formed gastral pouches con- siderably outstrips the others in development, so that for a long time the bilateral symmetry of the larva is marked externally by the presence of this one large tentacle (Fig. 44/1). Haacke (No. 76) has called attention to the fact that in attached stock- building forms, as in the blossoms of many Phanerogams, the bilaterally symmetrical fundamental form may be expressed by the position of the buds in relation to the parent animal, i.e., to the axis of the entire colony, since the parts of the bud near to the axis undergo a different development from those remote from it. Moseley had already shown that in Saccophyton and Heliopora the polyps always have their dorsal sides turned towards the axis. We may conclude from such observations that the bilaterally symmetrical structure of the Anthozoa is caused by the formation of stocks. The solitary forms (Actinians) would then have to be derived from those forming colonies. Finally, we may assume that CNIDARIA 97 the bilaterally symmetrical type, which at first is developed only in connec- tion with budding, became so firmly established that it also f omid expres- sion in the first stages of development from the egg (comp. above, p. 52). After the formation of the first twelve tentacles, a rearrangement, according to the number 6, takes place, so that there are two cycles of six tentacles each. The larger ones, those of the first cycle, correspond to the six primary intraseptal chambers, whereas the smaller ones, those of the second cycle, alternate with them. Six large tentacles of the first cycle thus alternate regularly with six smaller ones of the second cycle. The appearance of new tentacles does not take place by the interpola- tion of one in each of the twelve intervals between the elements of the first and second cycles, but by the appearance of six pairs, which occupy only one half of these intervals, as is represented in Fig. 44 B. We Fig. 44. — Two larvae of yldtuia mesemhryanthemum fafter La.caze-Duthiebs, from Balfour's Com'parative Enibryology). J., bilateral ciliated stage, with one large and several small tentacle buds ; m, mouth; B, view of an older stage fiom above. There are twenty -four tentacles around the mouth. The sequence in the origin of the twelve primary tentacles is a', a, h, c, d, /, e. here see that three tentacles lie in the intervals between every two ten- tacles of the first circle, one belonging to the second cycle and two being new ; but these are arranged in such a way that the middle one of the three everywhere belongs to the cycle of the youngest generation. This one now increases greatly in size, and outstrips the individuals of the former second cycle, which in this way lose their rank, and are classed in the third cycle. In later stages, cycles which differ in size (six ten- tacles of the first, six tentacles of the second, and twelve tentacles of the third cycle) actually alternate regularly with one another in position. It must be observed, however, that the present third cycle does not contain uniform elements, but six tentacles of the youngest stage of development and six which previously belonged to the second cycle. A rearrangement therefore has taken place. In the same way the number of the tentacles increases from twenty-four to forty-eight and to ninety-six by the appear- K. H. E. H 98 EMBRYOLOGY ance of new pairs of tentacles, half of the intervals being left empty. Thus by rearrangement a fourth cycle of twenty-four and a fifth one of forty-eight tentacles are developed ; but these, like the third cycle before, consist of elements of heterogeneous origin. Ordinarily the attachment of the hitherto free-swimming ciliated larva takes place in the stage in which the number of tentacles is increased from twelve to twenty-four. It is to be expected that in those forms which exhibit a special law of septal growth the sequence of the appearance of the tentacles is corre- spondingly modified. In a larva called Arachnactls by Sars and A. Agassiz (No. 72), conditions of organization are found which, as has recently been shown by C. Vogt (No. 96), connect it with the Cerianthece.^ The develop- ment of the tentacles also recalls the development of the Cerianthus larva, made known by Haime. In Arachnactis the tentacles do not grow out in cycles between those already present, but there is a dorsal budding zone (as in the case of the septa; comp. p. 95), where the youngest tentacles are formed in pairs. The tentacles of the inner circle also are formed in the same manner. It follows from this that the tentacles of the ventral side must be the largest and oldest. The unpaired, perpetu- ally dwarfed tentacle of the directive chamber, which is found between the longest paired tentacles, forms an exception. The development of the calcareous skeleton of the Madre- poraria has been studied by Lacaze-Duthiers (No. ^S) and V. Koch (Nos. 83 and 85) in Astroides calycularis. It takes place at the stage in which the first twelve tentacles of the larva have been developed, and in which attachment usually occurs. The calcareous skeleton is formed as a secretion on the outer side of the ectoderm of the body-wall (Fig. 45). At first a delicate circular basal plate arises as a secretion from the ectodermal cells of the pedal disc. This hasal plate, by means of which the larva attaches itself to some suitable ^ [In regard to the development of Arachnactis, the adult form of which has been found by Hertwig and Boveri, consult the recent statements of E. van Beneden (No. II., Appendix to Literature on Anthozoa) and BovERi (No. III., Appendix to Literature). A long time ago a very remarkable Actinia larva was described by Semper, and recently by E. van Beneden more in detail. This larva is characterized by the presence of a highly iridescent ciliate ridge running lengthwise of the body. Van Beneden is inclined to refer it to the group of the Zoantheae. Comp. Semper, "Ueber einige tropische Larvenfor- men," Zeitschr. wisa. ZooL, Bd. xvii., 1867, and Van Beneden (No. I., Appendix to Literature on Anthozoa).— H.] CNIDARIA 99 support, consists of roundish crystalline bodies, which sub- sequently fuse with one another. The earliest fundaments of the calcareous septa [sclerosepta] soon make their appear- ance. It was shown by Milne-Edwards et Haime, and afterwards by Lacaze-Ddthiers, that the calcareous septa correspond in position each to a gastral pouch, and therefore that they occur between every two mesenterial septa. The earliest fundaments of the twelve primary sclerosepta are called radial ridges (Sternleisten), and at first are V" ^^' Y-shaped (Fig. 46). The fundament of the theca (Mauer- hlatt) arises by the peripheral ends of the radial ridges soon becoming fused with one another. All of these are struc- tures which are secreted by the ectoderm of the pedal disc, Fig. 45. — Development of the calcareous skeleton of AsUoides calycularis (after V. Koch), diagrammatic. The section is made perpendicular to the pedal di^c in the direction of a secant. At the bottom the fundament of the basal plate ; to the left the epitheca ; to the right two radial ridges [sclerosepta] growing upwards from below, alternating with two mesenterial septa [sarcosepta]. and naturally the more these skeletal parts rise upwards the more the ectodermal layer of the pedal disc must undergo a kind of invagination. It follows from this that in later stages also those parts of the skeleton which apparently lie inside the body of the polyp are covered by an epithelial lamella belonging to the ectoderm of the pedal disc (calycohlast lai/er^ V. Heider). But the lateral walls of the body in its lower portions also deposit externally a calcareous layer, which constitutes the fundament of the so-called epitheca (Fig. 45). The so-called columella is formed by the fusion of the ladial ridges [sclerosepta] with one another at their inner, central ends. Six of the twelve radial ridges soon become more prominent, so that there is established an arrangement in two cycles. Subsequently other cycles make their appear- 100 EMBRYOLOGY ance by the interpolation of new small septa in regular order between the existing ones. Non-sexual reproduction in the form of fission and budding is found widely distributed in the Zoantharia ; by this means extensive colonies are developed in the skeleton-forming Corals (Scleroderraata), whereas in the group of non-skeletal Ac- tiniaria (Malacodermata) the individuals produced by fission or budding usually separate entirely, so that, with few ex- ceptions (Zoan these), the forms in this case remain solitary. Fig. 46.— Basal plate of a larva of Astroides calycularis, soon after at- tachment, with twelve radial ridges (after Lacaze-Duthikrs, from Bal- yocR's Comparative Emhryology). Budding in the Actiniaria has been observed more rarely — Epiactis (Verkill, ?), Gonactinia (Blochmann UND Hilger), Zoanthus. More fre- quently reproduction takes place by fission. This may divide the parent animal into two nearly equal parts : either as longitudinal fission, which begins at the oral disc and progresses toward the base, or takes the oppo- site direction, or as transverse division, a kind of reproduction which has been described in detail for Zonactinia prolifera by M. Sars and by Fig. 47.— Two stages of transverse fissicn of GonacLinia prolifera. Sabs (after Blochmanic uwd Hilgkk). Blochmann und Hilger (No. 74), and which in its outcome presents strik- ing resemblances to the divisions in Flabellum and Fungi a described by Semper, and to the process of strobilization in the Scyphozoa. Inr Go- nactinia it is always young animals that undergo transverse division. CNIDARIA 101 Somewhat below the middle of the parent animal is formed a circle of bud-like projections, out of which is developed the circle of tentacles of the lower individual. While the upper part is being constricted off, the oral disc and the oesophagus of the lower off- spring of the division are developed. Finally, the upper part detaches itself. It appears that both parts have the power to divide again. Another remarkable, more widely distributed kind of division, which had already been ob- served by DicQUEJURE and by Dal yell (No. 4), has recently been studied in detail by A. Andres (No. 73), and has been called laceranon (Fig. 48). This consists in the abstriction of frag- ments of a basal expansion. At the margin of the base of an Actinian a small part is character- ized by the opacity of its entoderm and by its rm adherence to the support, the latter being used by a secretion of the ectoderm. By the ontraction of the parent animal, the modified marginal part is torn away from it. This can now be metamorphosed either directly into a small Actinian, or after further separation into smaller fragments. Both kinds of non-sexual reproduction, fission and budding, are widely distributed among the Corallia. They here lead to the formation of extensive stocks of various shapes. In many cases (Oculinacea and Astraeacea) in which it was formerly believed that lateral budding occurred, Studer (Nos. 94 and 95) was able to show, upon closer investigation, that there exists a reproduction by fission, one of the resultants of division coming with further growth to occupy a position on the lateral wall of the other part. A similar kind of reproduction has been observed among the Fungiaceae in Her- petolitha limax. Genuine basal budding is found, for example, in Turbinaria, where the base of the colony exists as a common plate of coenenchyma, at the margin of which new individuals bud ; like- wise in Galaxea. The form of longitudinal fission occurring in the Corallia, which Usually begins with a constriction of the oral disc, may remain more or less incomplete, so that the individuals remain united with one another in series. This arrangement can be recognized even in the skeleton, since Fig. 48— Reproduction in Ai^tasia lacerata by means of abstriction of a basal part (after A. Andres). A to C, advanc- ins? abstriction ; D, E, metamorphosis of the fragment into a small Actinian, 102 EMBRYOLOGY a whole series of individuals remains enclosed by a common theca, whereas the septa are placed perpendicular to the direction of the tortuous valleys extending between the thecae (Meandrina). In the stone corals also, budding and fission may lead to the formation of individuals which separate from the parent and live independently. In Blastotrochus there are lateral buds that separate, whereas in Flabellum a kind of transverse division occurs. The young stages of the Fungidae form small coral stocks from which the solitary forms, which become sexually mature, are abstricted by transverse division. Since one and the same branch may undergo this process of transverse division several times, the resemblance to the strobilization of the Scyphozoa is very striking. Here also there is a true alternation of generations (Semper, No, 93). III. SCYPHOMEDUS/E. Of the forms belonging here the Luceniaridce and Charyh- deidce are contrasted with the Discophora proper. While the embryology of the latter has been repeatedly investigated, we have as yet only a fragmentary knowledge of the two groups first named, Lucernaridae.— FoL and Korotneff have given accounts of the larvie of the Lucernarians, The development from the egg has been more thoroughly investigated by Kowalevsky (No, 108), whose results have recently been confirmed by R, S, Bergh (No, 101), After the egg and sperm have been discharged into the water fertilization takes place, at the completion of which the egg retracts somewhat from the vitelline membrane. Two polar globules are formed, and then the first cleavage furrow arises. By means of total and equal cleavage a multicellular stage is formed, which presents no cleavage cavity. The pointed ends of the pris- matic cells meet at the centre. An accumulation of entoderm cells now takes place inside this so-called morula ; this is accomplished by a contri- bution of elements from a definite region of the egg, so that the production of the entoderm here seems to approach the type of polar ingression. Kowalevsky believes that it is chiefly a transverse division of the pris- matic cells in this region that leads to the contribution of entodermal elements ; however, simple ingression is not wholly excluded. The bilaminar stage resulting from this is at fu'st completely spherical (Fig. 49/1), but soon elongates in the direction of the future chief axis (Fig. 49 B). The entoderm cells meantime become vacuolated, and arrange themselves more and more in a single row, so that there results from this a rod-like planula, which, like that mentioned for iEginopsis (p. 67), resembles a detached hydroid tentacle (Fig, 49 C). This planula of the Lucernaridae is not ciliated, but creeps slowly about with worm-like movements. The first nettling cells are developed a its posterior end. Preparatory to assuming the polypoid form, it eventually attaches itself CNIDARIA 103 by means of its anterior end. The further development could not be followed. K. S. Bergh, however, mentions a young stage in which the tentacles were not yet united into groups, but were distributed along the margin of the bell, while the arms were not yet developed. Eight ten- tacles lying in definite radii could be recognized as fundaments of marginal papillae. Charybdeidae. — W. Haacke (No. 106) has given a description of some young forms of the Australian CharybdEea Rastonii, which already con- siderably resembled the adult animal. These accounts are thus far the only ones on the embryology of this genus. As contrasted with the cubical form of the adult animal, the young Acalephs showed an approach to a pyramidal shape, and the apex of the umbrella was more strongly arched than in the adult. The youngest stage that was observed exhibited a canal somewhat excentrically situated, and extending from KiG. 49.— Three stages in the develop- ment of Lucernaria (after R. S. Bbrgh), Fig. 50.— Non-sexual reproduction of the Scyphistoma (after M. Sa.eS) —A, by the formation of stolons; B, by lateral budding. the central stomach to the dome of the umbrella, where it ended blindly. Haacke regards it as the remains of a communication with a Scyphis- toma nurse, and therefore maintains the probability of an alternation of generations in the Charybdeidae. From the egg of most Discophora (Discomedusoe) a fixed polypoid creatare is first developed, which is attached bj one pole, and has the mouth at the opposite end, at some distance from which a circle of tentacles is developed (Fig. 51, 3, 4). The Lncernaridae are essentially a more highly developed form of these scyphopolyps, which become sexually mature. In all other Scyphomedusae the polypoid form 104 EMBBYOLOGY (ScijpMstoma) appears to lack the power of generating sexnal products, exhibiting only non-sexual reproduction, which occurs in two modifications : (1) as budding (lateral budding and formation of root-runners or stolons) (Fig. 50), by means of which a scyphopolyp is always produced again — this either separates from the parent and attaches itself independ- ently, or may remain united with the parent, thus tem- porarily producing small colonies (scyphopolyp stocks) — (2) nr^ 7 8 9 12 Fig. 61.— Cycle of development of Aurelia aunia (from Hatschkk's Lehrhiich). ], planula; 2, attached larva; 3, young Scyphlstoma with four tentacular buds; 4, Scyphi>toma with stolonic growth ; 5, beginning of the strobilization, indicated by a circular furrow ; 6, 8, 9, 10, various strobilae polydiscse ; 7, Scyphistoma from above ; 11, Ephyra from the side ; 12, Ephyra from below. as strobilization, in reality a transverse division with subse- quent regeneration. By means of transverse constrictions the scyphopolyp (Fig. 51, e) separates into superposed dis- coid parts (strohila stagp, Fig. 51, 5 — 10), each one of which, by the production of marginal lobes and corresponding internal metamorphoses, is changed into a young medusa, which at first shows the characteristic form of the Ephyra stage (Fig. 51, n, 12), and is not converted into the permanent CNIDAEIA 105 orm of the sexually mature medusa until after a metamor- ^^^phosis, which in most cases is rather complicated. ^B In most of the Discophora hitherto studied development ^^ takes place in the form of an alternation of generations : already described. This is wanting in the Lucernaridse ^fconly, they representing a sexually mature scyphopolyp ^^stage, from the eggs of which individuals of the same form arise. On the other hand, among the free-swiming acraspe- dote medusae cases (Pelagia) of direct development are known, in which a larva developed from the egg of the medusa changes directly into the Ephyra stage. This is looked upon as a case of coenogenetically abbreviated develop- ment, since the formation of a non-sexually reproducing nurse-form (Scyphistoma) is suppressed. Development of the Scyphistoma.— The develop- ment of Aurelia (A. aurita and A. flavidula) is that of which we have the most complete knowledge ; it has been made known through numerous investigations — those of M. Saks (No. 112), V. SiEBOLD (No. 114), L. Agasstz (No. 2), Glaus (Nos. 102 and 103), Haeckel (No. 107), and Goette (No. 105). In the following, we adhere chiefly to the description of Goette, by whose investigations a number of new points of view have been gained. The eggs of Aurelia aurita pass from the ovary into the gastral cavity of the parent, and from there through the mouth into the folds of the oral arms, where, enveloped by a slimy secretion from the entoderm, they undergo embryonic development as far as the stage of the swarming planula. They are enveloped by a delicate vitelline membrane, which is lost in the later stages of cleavage. By total and equal cleavage (Glaus) the egg divides into a number of equal-sized blastomeres, which arrange them- selves in a single layer about a comparatively small cleavage cavity (coeloblastula). While, according to Glaus (in har- mony with the statements of Kowalevsky), the gastrula- stage is reached by means of a process of invagination,^ in which ^ [The observations of Claus have been fully corroborated by the recent investigations of Fbank Smith (No. VII., Appendix to Literature on Scyphomedusas) on Aurelia fiavidula. In this species the entoderm is pro- 106 EMBRYOLOGY the lumen of the archenteron can be recognized only as a linear fissure in the plug-like ingrowth, another method of formation of the lower germ-layer, that may be called polar ingression, has been maintained by Goette. According to GOETTE, the cells of the blastula have not the same form in the entire circumference, but are somewhat shorter and broader in one hemisphere. From this region there is a migration of individual cells into the blastoccele, until finally this cavity is completely filled with a solid cell-mass (entoderm). The archenteron arises in this in the form of a fissure, which soon breaks through to the exterior at the region from which the immigration of entoderm cells took place, thereby forming the primitive mouth (prostoma). Even during this process the embryo, originally spherical, elongates, so that the longitudinal axis passes through the primitive mouth and the apical pole lying opposite to it. But the primitive mouth very soon closes completely. At the same time the larva becomes narrowed at this end, so that it is pyriforra. The swarming out of the ciliated embryo (planula, Fig. 52 ^1) now takes place ; the broader apical pole is directed forwards in swimming, whereas the narrower pole, at which the closure of the primitive mouth took place, comes to lie behind. Nettling capsules very soon make their appearance on the swarming larva ; these arise in great numbers at the posterior pole, whereas they are almost wanting at the anterior end. Even during the swarming stage a shallow depression is developed at the anterior (apical) pole of the larva, and at this point the epithelium acquires a glandular nature. The larva now attaches itself by the apical pole to some support, duced by the formation of a distinct invagination gastrula. A migration of cells into the blastoccele, as described by Goette, was also occasionally observed in A. flavidula. However, these cells appear to disintegrate without taking any part in the formation of the entoderm. On the other hand, the entoderm is formed in Cyanea arctica, according to McMuRRicH (Appendix to Literature on Scyphomedusas, No. V., p. 314, and No. VI., p. 90), by an inward migration of certain cells of the blasto- sphere, and in Cyanea capillata, according to Hamann (No. IV., Appendix to Literature), by the ingrowth from one pole of the embryo of a solid rod, which subsequently becomes hollowed out to form the gastral cavity.— H.j 107 tluis the former anterior end becomes the foot of the scyphopoljp ; this soon contracts a little, whereas the posterior end widens, so that in this way the body acquires the goblet shape characteristic of polyps (Fig. 52 B). During the attachment a cement, which soon hardens into a plate with upturned margins (Figs. 54 and 55 k), is secreted from the foot. The secretion of a mesogloea begins early between the two layers of the larva (Fig. 52 B, g). The next change is the formation of the permanent mouth, which arises by a process of invagination. The ectodermal layer of the prostomal pole invaginates into a gradually deepening ectodermal pocket (Fig. 52 B, s), at the bottom of which a perforation, leading into the gastral cavity, soon arises. In this way an oesophagus, lined with ectoderm, is produced (Fig. 52 (?) ; the outer opening is known as the viouth, the inner, communicating with the gastral cavity, as the inner opening of the oesophagus (Fig.. 52 (7, sp). By means of this process of invagination the entodermal sac becomes crowded downwards, but not throughout its entire extent. Since the larva is compressed laterally, two glove-like ento- dermal processes, corresponding in position to the longer of the secondary axes, are preserved. These project upward, and are the first two gastral pouches (Fig. 52 G m and D m). Very soon, however, in a plane at right angles to this, a second pair of gastral pouches grows upward as diverticulae of the central stomach (Fig. 52 E), so that now the radiate type with four rays is reached. We now have an oesophagus invaginated from the ectoderm, in the circumference of which, at the four radii, lie gastral pouches in connection with the gastral cavity.^ At these places, where two neigh- bouring gastral pouches come in contact, a partition, or ^ [Our knowledge of the first processes of development in the Scyphis- toma stage has been materially increased by recent investigations, which have advanced information in several directions. Nevertheless it is not possible as yet to pronounce final decision concerning these develop- mental processes. The observations of Goette have been only partially confirmed by Claus (Nos. I. and II., Appendix to Literature on Scypho- medusee). As the result of his most recent observations, Claus (No. II.) denies totally the presence of an ectodermal pharynx. Of special import- ance are the statements of Claus (No. II.) concerning the formation of 108 EMBEYOLOGY septum (Fig. 52 E, st), is produced hy their contiguous lateral walls. These four septa lie in the interradii, whereas the four Fig. 52. — Diagrammatic sections through various successive stages of Aurelia (after Go ettk). .4, plannla; ec, ectoderm ; en, entoderm ; JB, attached larva with forming oesophageal invagination, s ; g, mesoglcea ; C, completed rupture of the oesophageal invagination ; sp, inner opening of the oesophagus ; m, gastral pouches ; n, transverse section through the stage represented in C at the level of the oesophagus; E, transverse section through an older stage at the level of the oesophagus ; F, transverse section through the same stage at a part nearer to the stalk ; s, oesophagus ; st, septa ; t, tseniolse. primary gastral pouches lie in the four chief radii (perradii). the proboscis in the developing Ephyra of the strobila stage, a point which had not hitherto received any careful attention. Eecent observations by Goette (No. III.) on Cotylorhiza tuberculata and Pelagia noctiluca have led to the astonishing result, that of the four primary gastral pouches, — although the first pair is of entodermal origin, being produced from diverticula of the archenteron, — the second pair is ectodermal in origin, since it arises by evagination from the ectodermal pharynx.— H.] Consult also Ida H. Hyde, No. VII. 6.— Tkakslators. CNIDARIA 109 » P The structures occurring between these [eight] radii are designated as adradial. The lower free margins of the four septa soon become continuous with the wall of the central stomach in the form 1^^ of four longitudinal folds, which ultimately extend through ^Hthe entire length of the scyphopolyp, even into the foot. ^'' These folds are known as the longitudinal folds, or tceniolce (Fig. 52 F, t), and the sinuses of the central stomach limited by them as gastral furrows. In the further metamorphosis of the larva the form changes, approaching more and more the shape of a goblet (Fig. 53). Fig. 53. — Diagi-a-mmatic loTi<;itudinal section through a Scyphistoma (based on Goettk). ^, perradial longitudinal section ; B, interradial longitudinal section; pb, proboscis; f, tentacle ; tr, septal funnel ; m, gastral pouches; g, mesogloea; s, septum. The entoderm is represented as a dark layer. The lower narrow portion is called the stalk or peduncle (Fig. 55 st), the prolongation of the central stomach extend- ing into it the peduncular canal. The upper part of the body becomes flattened, and thus forms the oral disc or peri- stome, in the middle of which rises the cone-shaped proboscis (Fig. 53 ph) with its central four-sided mouth-opening. The four corners of the mouth are placed perradially (Figs. 54 0 and 55). The first four tentacles now arise over the four gastral pouches, and, in keeping with the successive appearance of the pouches, those over the first pair of pouches arise first, and then those over the second pair. A cylindrical ento- 110 EMBRYOLOGY dermal cord grows frorn the apex of each gasfcral pouch diagonally upwards and outwards, pushing before it the ectoderm of the outer margin of the peristome. The ento- derm cells in the tentacular buds soon arrange themselves in a single row (Fig. 53 0- Other important fundaments of organs are represented by the septal funnels which are now established. Four funnel- like invaginations arise from the ectoderm of the peristome in the interradii ; these sink into the septa, and extend down- wards as solid cords of cells, which are continued along the teeniolae and even beyond these into the stalk (Fig. 53 B ; Fig. 54 A and (7, tr). In this solid portion, the cells appear to be fused with one another, and on their surface longi- tudinal muscle fibrillge are differentiated, so that the four [septal] longitudinal muscles extending in the taeniolse arise in this way (Fig. 54 A, B, sm). The young Scyphistoma thus produced is characterized therefore as a goblet-shaped polj^p, with four longitudinal folds (teeniolse) of the entodermal sac extending upwards as four septa, w^hich are stretched between the body- wall and the invaginated ectodermal oesophagus. The stomach is accordingly divided into a central cavity and four gastral pouches (peripheral intestine \^Kranzdarm']\ which lie be- tween the septa and are directly continuous with the gas- tral furrows. Four perradial tentacles are attached to the margin of the peristome, while four interradial septal fun- nels extend from the peristome into the septa and taeniolsB (Fig. 53). The metamorphosis into older Scyphistomse (Figs. 54 and 55) takes place by an increase in the number of the ten- tacles and other changes, which efface more and more the original characters, and lead by a gradual transition to the structural plan of the Ephyra. The budding of the tentacles presents many irregularities. Heretofore it has been believed that normally after the formation of the first four tentacles, radial in position, the development of four interradial ones {i.e., placed over the septa) took place, and then, after all these eight tentacles had reached the same length, ensued the development of eight others, lying between them (therefore adradial), and so on. According to Goette, CNIDARIA 111 owever, the four tentacles that immediately succeed the four primary ones are not placed over the septa, but bud out from the corners of the gastral pouches of the second pair, which lie next to the septa, and only gradually move into positions over the septa. In this way, their axial, entodermal cords acquire connections with the gastral pouches of the second pair. Since the gastral pouches of the first pair soon follow with the formation of four new tentacles, the equivalence of the first four Fig. 54. — Diagrammatic representation of the structure of an older Scjpbistoma (based upon Gobtte, from Hatschek's Lehrbuch). A, longitudinal section : at the left, perradial; at the right, interradial ; AB, chief axis; o, mouth ; s, inner opening of the oesophagus; gt, gastral pouches; gr, gastral furrow; so, septal ostium ; tr, septal funnel ; sm, septal muscle (the dotted line does not quite reach to it) ; B, transverse section through the lower part of the body ; gr, gastral fur- row; 8, septum ; sm, septal muscle; C, view of the oral side (references as in A). primary gastral pouches is established for the first time in the stage with twelve arms. Goette, therefore, maintains that the numerical series, 4, 12, 20, 28, etc., is the primitive one for the budding of the tentacles, whereas the actually observed series, 4, 8, 16, 24, 32, etc., corresponds to a coenogenetically modified condition. It should be mentioned that the formation of each new tentacle takes place by an outfolding of the cor- responding part of the gastral pouch, so that in reality a small secondary gastral pouch is produced with every tentacle. 112 EMBRYOLOGY The further metamorphosis of the developing Scyphistoma consists in a widening of the central stomach, whereby the oesophagus gradually moves into the proboscis (Fig. 54 A), and the gastral pouches tend to become obliterated. At the same time the entrances to the four funnels, which are widely open toward, the peristome, produce a circular, groove-like depression, involving the entire circumference of the origin- ally flat peristome, which thereby approaches the bell shape of the sub-umbrella of the medusa ; the proboscis, which has become more elevated, corresponds to the oral tube [manu- brium], while the gastral pouches, separated by the septa, represent the peripheral in- testine {Kranzdarrri) of the medusa. The Scyphistoma, by gradual 'metamorphoses, has approached in the most essential features the structure of the medusa (comp. Figs. 54 A and 67). In most of the other Discophora, the development of the Scyphistoma seems to take place in quite the same way, especially in Cotylorhiza borbonica (Kowalevsky, Goette) and Cyanea capillata (Sars, Van Beneden, Agassiz), where the eggs likewise undergo the first stages of development attached to the oral arms, and enveloped in a slimy jelly. On the other hand, the early development in Ckrysaora, a form which is also striking on account of its hermaphroditism, presents notable de- viations (Claus, No. 102 and No. 3). Here fertilization and the entire embryonic development take place within the ovary, so that the larvae are not born until they reach the planula stage. The very small mem- braneless eggs are surrounded in the ovary by a pedunculated follicle, which owes its origin to the cells of the germinal epithelium. Fer- tilization and cleavage are transferred to an early stage in the de- velopment, so that at the same time with the embryonic development there is a considerable growth of the embryo as the result of a con- tinual supply of food material on the part of the parent. This food supply is provided by the follicular cells. In these particulars the de- velopment of the egg and embryo of Chrysaora recalls that of the vivi- parous Aphidce and the Polyphemidae among the Cladocera. In other Fig. 55. — Scyphistoma of Aurelia aurita. ph, proboscis ; tr, entrance into the septal infundibulum ; t, taeniolae; 8f, stalk; fc, adhesive mass. ICC m CNIDARIA 113 respects the phenomena of the embryonic development are essentially e same as those we have described for Aurelia. By means of total and ual cleavage, a coeloblastula arises, out of which by infolding an vaginate gastrula develops, whose prostoma reinains open for a con- siderable time, but finally closes completely. From observations by BuscH, it appears as if reproduction of the embryo, by means of longi- tudinal division, frequently took place at this stage. This recalls the occurrence of fission in the blastula of Oceania armata, according to Metschnikoff (p. 49). In the stage of the ciliated planula the larvae of Chrysaora pass from the ovary into the gastral cavity of the parent, and thence to the outside world through its mouth. A glandular modifi- cation of the ectoderm of the anterior pole of the larva, by means of which the attachment subsequently takes place, can be recognized, whereas the posterior (oral) pole is characterized by the appearance of nettle capsules (Claus). The opaque whitish or yellowish eggs of Nausithoe are laid singly, and are characterized by a gelatinous envelope, provided with nettle capsules (0. Hertwig). Cleavage is here total, and in the first stages unequal, though finally a coeloblastula with walls of nearly uniform condition is produced by the gradual obliteration of the great differences in size be- tween the blastomeres. The blastula changes into an oval, ciliated swarm- ing larva, the cells of which are thickened at the posterior pole, where the gastrula invagination takes place. After invagination the blastopore becomes completely closed. Metschnikoff (No. 12), to whom we owe the knowledge of these processes, was able to observe the attachment ox the planula, which is accompanied by the development of a discoid basal expansion, and its metamorj)hosis into a small scyphopolyp pro- vided with four tentacles and covered with a thin layer of periderm, so that metagenesis has been proved for this form also. Metschnikoff believed that he was justified in assuming that the Spongicola fistularis of F. E. ScHOLZE (Stephanoscyphus mirabihs, Allman), which is para- sitic in sponges, and in which Kowalevsky seems to have observed a kind of strobilization, is the Scyphistoma form of Nausithoe. Strobilization. — The simplest form of reproduction of yoang medusae is represented by the monodisc strohila (Fig. 59 A)^ occasionally observed even in Aurelia. In this case only one young medusa (fi/p/iz/^'a) separates from the Scyphis- toma. While the adoral tentacle-bearing portion of the scypho- polyp is by gradual changes converted into the form of the Ephyra, it becomes separated by means of a circular, trans- verse furrow from the basal portion of the body, and finall}^ detaches itself completely. The basal remnant can, by re- generation of the oral portion, grow again into a complete K. H. E. 1 114 EMBRYOLOGY scjphopoljp, and subsequently go through the process of strobilization again, and so on. In most cases, however, new transverse furrows make their appearance on the basal part before the detachment of the first Ephjra, so that on the elongated cup of the scjpho- polyp a whole set of Ephjras (ten to thirty) are developed at approximately the same time ; but of these any one that is nearer to the base of the polyp is younger than those distal to it (polydisc strobila) (Figs. 56 and 51 e — lo). In this case also the basal portion finally reproduces both a circle of tentacles and the oral part of a scyphopolyp, and is thus enabled to continue its existence as a scyphopolyp when the production of Ephyrae ceases. A polydisc strobila can be derived from a monodisc. In the former new transverse divisions follow one another so rapidly that a large number of Ephyrae are in process of development at the same time. The oral portion of a scyphopolyp, in metamorphosing into an Ephyra, must undergo certain changes, part of which make their appearance before the first indication of an abstriction is produced by the circular furrow. The most important internal change is introduced by the disappearance of the septa and the peripheral communication between the four gastral pouches which is thus brought about. Since the entodermal columns of the four septal tentacles are con- tinuous with the walls of both the gastral pouches adjoining the septum (p. Ill), there is produced at this place a con- nection between the neighbouring gastral pouches. At this point a small perforation now arises in the septum (Fig. 54 so), but this very soon widens to such an extent that only the thickened inner margin of the septum, w^hicli is tra- versed by the septal infundibulum, is preserved (Fig. 57 so). By the formation of these septal ostia, the four gastral pouches coalesce into a common peripheral gastral chamber {peripheral intestine) [Kranzdarm]. The four septal infundi- bula, clothed by an entodermal covering (remains of the septa), now traverse the gastral space in the form of four columns {colnmellce), which are not attached to the wall of the central stomach except at its bottom. A further change is brought about by the disappearance CNIDARIA 115 the Scyphistoma tentacles and the growing out of the margin of the peristome into a lobed crown consisting of eight (four perradial and four interradial) marginal lobes I^Hi^ig. 56). Since the marginal lobes are not formed by the' ■^^xiter body- wall alone, but contain a corresponding diverti- culum of the entodermal sac, the peripheral gastral chamber in this way acquires eight blind sacs : the lohe-pouches (Fig. 58 /). The marginal or primary lobes \^Siammlappen\ soon develop three processes at their ends, of which the middle one buds forth from the sub-umbrellar part of the lobe at some distance from the margin, and becomes the sensory body \_Sinneskolbe^ (sk), while the two lateral processes bud forth from the margin and become the alar bes. Inside of these are found the alar ^pouches (Fig. 58 /) as prolongations of the lobe-pouches ; in the sensory body also there is found a prolongation of the gastral entodermal layer, which is destined to pro- duce the otolith crystals. The peripheral gastral chamber up to this time was simple and undivided, and at its periphery ran out into the eight lobe- pouches. Since the disc of the developing Ephyra is always flat, the upper and lower (ex-umbrellar and sub-umbrellar) walls of the peripheral gastral chamber are very close to each other, and these two walls now grow together from the margin inwards at sixteen regions of the circumference [the shaded areas in Fig. 58], and thus form sixteen radial fusions or concrescence-bands, which are sitnated sub-radially (i.e. between the per-, inter-, and adradii). In this way the marginal part of the peripheral intestine is separated into sixteen marginal pouches (radial peripheral pouches. Fig. 58 m), which are separated from one another by the concrescence-bands (cathammata). Eight of these marginal pouches are situated in the per- and interradii, and are continuous with the lobe-pouches, while eight others are adradial and interpolated between the former. In later Fig. 56.— atrobila polydiscaof Aurelia aurita. On the up- permost Epliyra the Scyphistoma ten- tacles in process of degenerating. 116 EMBRYOLOGY stages the marginal pouches become narrower and move apart; the sixteen regions of fusion [cathammata] thus spread out into a bilarainar plate, which connects all of the marginal pouches with one another : the medusoid, vascular, or cathammal plate. The detachment of the Ephjra now takes place, and from this time on it moves about freely by rhythmical contractions of its discoid body, the former point of attachment being directed upwards, and the manubrium downwards (Fig. 51, ii). The columellae, which are frequently the means of the final connection with the nurse form, now degenerate. It Fig. 57.— Tnterradial longitudina section through an Ephyra monodisca, with the Scyphistoma tentacles still retained (diagram modified from Goette). pb, proboscis; tr, septal funnel; gf, gastral filament ; so, septal ostium ; r/, constrict- ing annular groove. is probable that the last metamorphosed remnant of the septal infundibula can be recognized in the four sub-genital cavities (p. 122) lying on the sub-umbrellar side below the gonads of the medusa. With the degeneration of the columellae the boundary between the central stomach and the peripheral intestine entirely disappears, and the boundary at which the ectoderm of the oesophagus is continuous with the entodermal lining of the peripheral intestine is indicated only by means of four tentaculoid gastral filaments (Figs. 67, 58 gf), which have budded forth at the bases [oral ends] of the columellae. ( CNIDARIA 117 The Ephyra (Fig. 51, ii and 12, and Fig. 58), accordingly, possesses a flat, discoid body, from the nnder-side of which the manubrium hangs down. The margin is prolonged into bifid marginal lobes, each one of which bears a sensory body between its alar lobes. Four of these are perradial, and correspond to the radii of the oral cross, whereas the four interradial ones fall in the radii of the "^astral filaments. The Fig. 58.— Diajjrammatic figure of an embryo of an Ephyra. 0, cruciform mouth- opening ; gf, gastral filaments ; I, lobe-pouches ; /, alar pouches ; c, cathammata, or regions of fusion of the peripheral intestine ; sh, sensory bodies. broad, flat gastral space is prolonged into sixteen peripheral marginal pouches, which are connected by means of the scular plate. Of these pouches the eight perradial and terradial ones are directly continnous with the lobe-pouches and alar pouches. The ectoderm on the oral side of the disc (sub-umbrella) forms a broad, band-like circular muscle, while paired longitudinal muscle-bands stretch along the marginal lobes and into the alar lobes. Hypogenetic Development of the Larvae of Pelagia. Schneider (iS"o. 113) and Haeckel (No. 107) have already 118 EMBRYOLOGY observed that the seyphopoljps of Aurelia aurita, when they are placed in unfavourable conditions (for example, in aquaria), show little inclination to form poljdiscous strobilae, but frequently develop only monodisoous strobilae (Fig. 59 A). In fact, Haeckel observed in certain cases that the transverse division of the scyphopolyp metamorphosing into an Ephyra is altogether sup- pressed, so that the entire body of the larva is converted into the adult animal. This is Haeckel's so-called Ephyra peduncuJata (Fig. 59 B), which was observed in the attached as well as in the free-swimming condition. Here therefore the alternation of generations is omitted, and a simple metamorphosis (hypogenesis) has taken its place. The latter condition is the normal and only one in Pelagia noctiluca, the development of which has been made known through Krohn (No. 109), Kowalevsky, and Metschnikoff (No. 12). In this case there is first formed a blastula which has a large cleavage cavity, and soon becomes covered on the surface wnth flagella. At the same time an invagination from the posterior pole is formed, which leads to the development of a gastral cavity which does not by any means completely fill the space of the primitive cleavage cavity (Fig. 60 A). The blastopore does not close, but becomes the mouth of the larva. A shallow depression is very soon noticeable at the posterior end of the free-swimming larva, in the middle of which the part sur- rounding the mouth projects in the form of a cone (Fig. 60 B). This projection becomes the oral cone of the Ephyra (Fig. 60 C, w), and the circular depression surrounding it the umbrellar cavity, while on the peripheral margin a divi- sion into eight marginal lobes is soon noticeable, into which Fig. 59.— ^, Strobila mo- nodisca of Cyanea capillata (after P. J. van Beitbdkn) ; e, lobes of the Ephyra; *, newly formed circle of Scyphistoma tentacles on the basal portion. B, Ephyra pedunculata of Aurelia aurita (after Haec- ksl). CNIDARIA the gastral cavity is continued in the form of lobe-pouches I^XFig. 60 (J). After the Ephyra shape has thus found ex- ^fcession in the region of the oral pole, the larva shortens m the direction of the chief axis, and gradually assumes the flat, discoid form of the Ephyra. Meantime the larva loses the covering of flagella, and from now on moves like a medusa by the regular contractions of the margin of the disc. In Pelagia accordingly the larva coming from the ep;g passes directly into the Ephyra, although Goette has pointed out that, owing to its structure, we must regard the first stages of this metamorphosis as free-swimming Scyphistoma stages. r ^ ^ hi ^ Fig. 60. — Three stages of development of the free-swimming larva of Pelagia noctiluca (after Krohn). r, marginal lobes; s, sensory bodies; m, mouth- opening. Metamorphosis of the Ephyra. — The metamorphosis of the Ephyra is accompanied by a constant increase in the size of the body. The sensory bodies of the Ephyra become the eight marginal bodies [rhopalia] of the medusa. Since the adjacent alar lobes, from which the ocellar lohes arise, do not continue to grow with fche same rapidity as the rest of the body, new structures, corresponding in position to the adradial regions, are developed in the margin (adradial or intermediate marginal lobes). The simplest conditions directly referable to the Ephyra are found in the Ephyropsidae (Nausithoe), in which the sixteen alar lobes of the Ephyra are retained comparatively well developed, while eight adradial (intermediate) tentacles alternate with these. The pocket-like marginal pouches separated by narrow concrescence-bands (Claus) and the absence of arm-like prolongations of the angles of the mouth are so many characters derived directly from the Ephyra. In the families of the 120 EMBRYOLOGY Pelagidffi and Cyanidffi also the original character of the gastrovaseular system is preserved, the sixteen marginal (radial) pouches being retained as broad spaces separated by only narrow concrescence-bands, and not communicating by any circular sinus. More complicated conditions are found in the Aureliidae, the metamorphosis of which from the Ephyra has been accurately described by Claus (No. 102 and No. 3) for Aurelia and Discomedusa (Umbrosa). In Aurelia the enlargement of the disc is accompanied by the development of eight intermediate (adradial) marginal lobes, on the ex-umbrellar surface of which numerous short tentacles, arranged in a single series, are developed (Fig. 61 i). While Fig. 61.— Development of the margin of the disc and the canal system of Aurelia aurita (after Glaus). A, quadrant of an Ephyra disc 3 mm. in breadth ; -B, quadrant of a young Aurelia with a disc 9 mm. in diameter; i, intermediate (adradial) marginal lobes; o, ocellar lobes; sic, sensory bodies; t, tentacles (some- what removed to the ex-umbrellar side). the disc thus gradually enlarges, the sixteen marginal pouches grow out into elongated, narrow radial vessels, between which the concrescence- bands extend as broad areas of the vascular plate. By the separation in places of the two lamellae of this plate, secondary canals are developed, by means of which there is formed first a zigzag, and subsequently a peripheral circular communication between the different radial vessels (ring simis^, besides numerous lateral branches of the radial vessels (Fig. 01). The four oral arms, beset with papillae, arise as outgrowths from the four angles of the mouth. That which especially interests us in the metamorphosis of the Ephyra of Ehizostoma, made known by Claus (Nos. 3 and 103), is the metamorphosis of the oral stalk (manubrium). The broadened ends of the four oral arms grow out into bifurcate lobes, thus producing the fundaments of the eight oral arms, while by a similar process of growth at the ends of these the fundaments of the CNIDARIA 121 dorsal tufts [Dor'salcrispen] of the lower part of the arm arise. The fundaments of the shoulder tufts [Schulterkraiisen] , or scapulettes, arise in pairs as papillary elevations in the eight radii, and only subsequently assume an adradial position (Claus). The lateral margins of the arms, which are bent under, now grow together, so that there arise from the brachial furrows closed canals, which open to the outside world by means of the so-called funnels, rhizostomes or oscula suctoria (originally lateral folds of Ihe margins of the arms). As the last trace of the mouth, closed by concrescence, we find the central cruciform oral raphe. Fig. 62. — Diagram of an interradial longitudinal section through a Scyphome- dusa (based on a figure by Claus). R, radial vessel; Rk, marginal bodies [rho- palia]; ol, ocellar lobes; Gs, genital sinus; G, genital band; Gf, gastral fila- ments ; Gni, gastro-genital membrane ; S, sub-genital sinus. An account of the metamorphosis of one of the Versuridae (Stylorhiza punctata) has been given by v. Lendenfeld (No. 110). We have still to mention certain organs which are de- veloped at the places originally marked by the four degenerating columellas, i.e. in the interradii. These are, first of all, the gastral filaments and the genital band. In the youngest Ephyr86 only one of the gastral filaments, which originally budded forth as tentacle-like growths at the base of the columellEe (Fig. 57 gf), is found in each inter- radius. However, their number is soon increased (Fig. 58 gf), and finally numerous filaments occur, usually arranged in a curved line corresponding to the inner side of the 122 EMBRYOLOGY genital hand (Fig. 62 G), which is now developing as a fokl of the gastral wall. The sexual products arise from elements of the wall of this fold, are ripened (Fig. 62) between its two lamellae, and by the dehiscence of the wall pass into the gastral cavity, whence they reach the outside world through the mouth. The space underlying this fold and communi- cating with the gastral cavity is called the genital sinus (Fig. 62 Gs). The genital band, which is usually horse- shoe-shaped, is often interrupted at the interradius, so that we then find- in the four interradii eight paired gonads, which are often more or less adradially placed, a condition which probably must be looked upon as the primitive one. With the progressive increase in the thickness of the mesogloea, which grows, principally at the four corners of the mouth, into massive pillars, there is in the interradial region a more and more marked development of an invagina- tion of the outer surface of the body, which is called the sub-genital cavity (Fig. 62 S), and in its earliest beginnings is perhaps to be referred to the cavity of the septal infundi- bula. While, accordingly, the body-wall of the medusa is thickened by an increase of the mesogloea all around this place, it here remains as a very thin g astro- genital membrane (Fig. 62 Gm), which in many forms (e.g. Pelagia) shows a tendency to protrude outwai'ds like a hernial sac, so that in this way a genital sac {g astro- genital pouch) projecting into the sub-genital sinus is developed. While one might conclude from the structure of the adult genital band that it was developed by a simple folding of the sub-umbrellar wall of the stomach, the investigations of v. Lendenfeld and Hamann show that the earliest fundament of the genital band is merely a thickening of this wall, and that an elevation of this thickening in the form of a fold does not take place until later, when an invagination pushes forward more and more from the distal side, thus producing the genital sinus. General Considerations. — The fact that in the eggs of most Scyphomedusae a Scyphistoma stage is first de veloped, and that this stage is also indicated in the modified development of Pelagia (Goette), shows that we must imagine the ancestral form of the Scyphozoa as an attached Anthozoa-like polyp, which originally possessed, in addition CNTDARIA 123 production, that of non-sexaal production by budding and division. In reproduction by means of transverse division, the basal peduncular end of the divided polyp must have reproduced a new oral part, I^^i^hereas the detached oral portion had to move away from ^|pie place of its origin and seek a new place of attachment. ^^Before it could attach itself, however, it must by growing have reproduced the apical part of the goblet-shaped body and the stalk, so that in this way there arose two individuals of the same form as the parent. In this migration of one of the offspring of the division was furnished the motive for its metamorphosis in the direction of an increased power of locomotion, whereby a difference between the form of the attached polyp and that of the free-swimming medusa was initiated. From what has been said it is not to be wondered at, that the two forms are connected by gradual transitions; nevertheless we shall have to adhere theoretically to the differences of these two morphological conditions. The medusa is therefore a morphological phase of the Scypho- zoa which has proceeded from the scyphopolyp ; but, owing to the assumption of free locomotion, it is more highly deve- loped, the presence of sensory bodies and marginal lobes and the more highly organized musculature of the sub- umbrella, with concomitant increase of the elastic mesogla^a of the umbrella, being characteristic of it. In the Calycozoa the scyphopolyp reaches its highest phase of development, whereas the Peromedusoe are to be looked upon as the most primitive medusa forms. The lat- ter still reproduce the elongate bell-shaped form of the umbrella, with its apical, stalk-like process, derived from the attached polypoid ancestral forms, and in the possession of large septal funnels resemble the scyphopolyps, whereas the development of the margin of the umbrella marks them as medusse. In contrast to them, the EphyropsidaB and the corresponding larval Ephyra form appear as a further stage in the developmental series, in which the apical, elongate bell-shaped part of the umbrella and the peduncular rudi- ment have been lost, and in which the septal funnels have degenerated. We must explain the four interradial points 124 EMBRYOLOGY of adhesion (septal nodes, Haeckel), which are present in the Ephyropsida3 on the external side of the row of gastral filaments, as the remains of the columellae corresponding to these funnels. The Semaeostomae and the Rhizostomae are derived by further metamorphosis from the Ephyra form. If we imagine that in the above-supposed attached an- cestral form a division of labour made its appearance of such a nature that the power of non-sexual reproduction was retained by the attached scyphopolyp form, while the generation of the sexual products was confined to the free- swimming (medusa) forms, resulting from the transverse division, the origin of the kind of alternation of generations characteristic of the Scyphomedusae would in this way be explained. Whereas the tendency formerly was to unite the Hydrozoa and Scyphomedusae into a common group, in more recent times our conception has led to a complete se|)aration of these two divisions. 0. und R. Hertwig's (N"o. 9) doctrine of the diphyletic origin of the medusa form and their dis- tinguishing between Ectocarps and Entocarps first prepared the way for this separation. Although various persons, especially Claus (No. 102), had previously placed importance upon the presence or absence of the taeniolae, which are also possessed by the polyps, as characteristic differences, nevertheless the sharp separation between the scyphopolyps and the hydropolyps was first established by Goette (No. 105). On the other hand, the observations of Goette, especially the discovery of the ectodermal nature of the oesophagus in the Scyphistomae, have led to approximating this group to the Anthozoa, so that recentlj^ various authors (Lang, Hatschek), in accord with Goette, have united the two groups as Scyphozoa. It must be mentioned, however, that the scyphopolyps are separated from the Anthozoa by the possession of septal infuudibula, and by the ectodermal origin of the longitudinal muscles, to which are to be added as distinctive characters differences in the origin of the first four gastral pouches and many differences in general histological character — greater development of the mesodermal tissue in the Anthozoa. Even though we CNIDARIA 125 jsume, then, that Scyphomedusse and Anthozoa are de- jended from a common polypoid ancestral form which was already characterized by the possession of an ectodermal oesophagus, still the direct union of the two groaps does not seem to be as yet sufficiently established. General Considerations on the Cnidaria. — The Cni- daria constitute a very homogeneous, well-defined branch of the animal kingdom. We assume that the fundamental and ancestral form from which they are derived was a polyp similar to Hydra, the chief axis of which was the same as in the preceding free-swimming ancestral form. A free oral pole and a pole of attachment can be distinguished. The latter corresponds to the anterior pole of the free- swimming ancestral form. The radial type in the structure of the Cnidaria has arisen in connection with the attached mode of life, whereas in many Cnidaria, as the result of stock-formation, a bilaterally symmetrical type is second- arily developed. It appears that even the ancestral form of the Cnidaria had developed the quadriradial structure, so that those forms in which, on account of the arrangement of the tentacles, no definite secondary axes can be recognized, like the Corynidce and Clavidcp., would represent a secondary modification. The growth of the Cnidaria frequently takes place by the typical intercalation of new radii between those already present (Hatschek) . The Hydrozoa are derived directly from this Hydra-like ancestral form (Archihydra), whereas the common ancestral form of the Anthozoa and Acraspeda is developed from it by the formation of an ectodermal oesophagus and radial septa. The presence of the longitudinal muscles in these septa indicates that they were developed in connection with the attached mode of life. In the ontogenetic series, it is true, the septa often make their appearance before the attachment and before the development of the tentacles, from which Goette concluded that there was an ancestral form, called a Scyphula, common to the Anthozoa, Acra- speda, and Ctenophora, which was characterized by a free- swimming mode of locomotion and by the possession of an oesophagus and radial septa. It is possible, however, that 126 EMBRYOLOGY ontogeny does not represent the primitive condition in this regard. The attached polyp form recurs ih the ontogeny of most Cnidaria. In the Anthozoa and Luicernaridae it constitutes the adult animal; in the Hydrozoa it is co-ordinate with the Hydromedusa ; wherea? in the Acraspeda, in comparison with the highly developed medusa form, it must instead be considered as the young stage. Many medusae (Tracheo- medusae, Pelagia) develop directly from the free-swimming larvae into the medusa (corap. pp. 53 and 118). But here also certain conditions of development can be interpreted as modified polypoid stages. ^ The development of free-swimming sexual forms (medusae) did not take place until after the separation into hydropolyps and scyphopolyps, and therefore occurred in the two groups independently. The differences in or- ganization between the hydroid- and scyphopolyps are explained by the different structure of their polyp forms and by their independent development. The hydroid- medusa is developed as a lateral bud, whereas the strobila- tion of the Scyphomedusae is to be explained as a process of transverse division. The medusa must be explained as a polyp which acquired powers of free movement, and as a result of this underwent certain changes in form. The first cause for the evolution of such locbmotion we have re- cognized in the migration which in non-sexual reproduction (division, budding) the detached portion must undertake before attaching itself. An opposite explanation, which is based chiefly on the occurrence of hypogenetic forms, and which sees in these the more primitive conditions, starts from a free-swimming medusoid ancestral form, the larvae of which, also at first leading a pelagic life, had secondarily acquired the attached mode of life and reproduction by budding or division. The polypoid forms would then have to be considered as ccenogenetically interpolated larval conditions (C. Vogt, No. 115 ; Brooks, No. 17). How- ever, the entire structure of the medusae points to a primitive attached ancestral form too clearly for us to grant this interpretation. In the search after those hypothetical free-swimming ancestral forms which preceded the attached Hydra-like form, we must first think of such creatures as are repre- CNIDARIA 127 sented in the ontogeny of Pelagia, for example, bj the stage of the gastrula invaginata, i.e., a ciliated, ovoid, free- swinxming form, in which an archenteron opening to the outside world by means of the prostoma was developed by an invao-ination at the posterior end. It can easily be explained how an ovoid blastula-like heteropolar an- cestral form happened to develop the earliest beginnings of the archen- teric invagination at the posterior pole of its body. In the case of monaxial, heteropolar blastular larvae which are allowed to swim through water containing particles of carmine, it can be seen that these particles are repulsed at the anterior and lateral parts of the body by the move- ments of the larvae, whereas they are crowded together at the posterior pole. Here accordingly was a favourable place for the reception of particles of food, and by a flattening or shallow invagination of the posterior pole these favourable conditions were increased. The archen- teron therefore in its earliest beginnings was a pit in which to catch particles of food. v^K If we incline to the view that the hypothetical ancestral form of the Cnidaria was similar to the gastrula invaginata, then in most Cnidaria we must assume a secondary change in the ontogeny, for the typical larcal jorm of the Cnidaria is the jola7iula, a form in which we can recognize a ciliated ectoderm and a more or less compact entodermal mass within. The taking of food is here suppressed. This form serves exclusively for locomotion and the consequent dissemination of the species over a larger territory. In attached forms such larval conditions are of great importance for the pre- servation and distribution of the species. In the interest of this function, the archenteric cavity appears to have degenerated in the planula. It is probable that the transition from the free-swimming gastrula-like ancestral form to the attached polypoid form was brought about by means of an interpolated creeping stage, which would be recalled by the creeping planula of many existing forms (e.g. Lucernaria). Literature. Cnidaria. in General. ^^1. Agassiz, a. Illustrated Catalogue of the Mus. Comp. Zool. Harvard ^B Coll. No. II. North American Acalephae. Cambridge, U.S. 18f5. L 128 EMBRYOLOGY 2. Agassiz, L. Contributions to the Natural History of the United States of America. Boston. Vol. iii., 1860 ; vol. iv., 1862. 3. Glaus, C. Untersuchungen iiber die Organisation und Entwick- lung der Medusen. Prag. u. Leipzig. 1883. 4. Dal YELL, J. G. Eare and Eemarkable Animals of Scotland. London. 1847. 5. Gegenbaur, C. Zur Lehre vom Generationswechsel und der Fortpflanzung bei Medusen und Polypen, Wilrzhurg. 1854. 6. Grobben, C. Doliolum und sein Generationswechsel, nebst Bemer- kungen iiber den Generationswechsel der Acalephen, Cestoden, u. Trematoden. Arheiten Zool. Inst. Wien. Bd. iv. 1882. 7. Haeckel, E. Das System der Medusen. Denkschr. Med.-natunc. Gesell. Jena. Bd. i. 1879 — 1880. Also separate. Bd. ii. Jena. 1879—1880. 7a. Haeckel, E. Die Tiefseemedusen der Challenger-Eeise u, der Organismus der Medusen. Jena. 1881. Also : Deep-sea Medusae. " Challenger " Beports. Vol. iv., part 12. 1882. 8. Hertwig, 0. UND E. Der Organismus der Medusen und seine Stellung zur Keimblattertheorie. Denkschr. Med.-natunc. Gesell. Jena. Bd. ii. 1878. Also separate. Jena. 1878. 9. Hertwig, 0. und E. Die Actinien, etc. Jena Zeitschr. Bd. xiii,, 1879 ; bd. xiv., 1880. 10. KowALEvsKY, A. Investigations on the Development of Ccelente- rates (Eussian). Mem. Bog. Soc. Friends of Nat. Set., Anthropol., and Ethnogr. Moscoic, 1873 (1874). See Jahresb. Anat. u. Phys. (Hoffmann u. Schwalbe), 1873. 11. Leuckart, E. Ueber den Polymorphismus der Individuen o. d. die Erscheinungen d. Arbeitstheilung in der Natur. G lessen. 1851. 12. Metschnikoff, E. Embryologische Studien an Medusen. Ein Beitrag zur Genealogie der Primitiv Organe. Wien. 1886. 13. Metschnikoff, E. Studien iiber die Entwicklung der Medusen und Siphonophoren. Zeitschr. tciss. Zool. Bd. xxiv. 1874. 14. Steenstrup, J. Ueber den Generationswechsel o. d. Fortpflanzung u. Entwicklung durch wechselnde Generationen. Uebers. von Lorenzen. Kopenhagen. 1842. Hydroidea. 15. Allman, G. J. A Monograph of the Gymnoblastic or Tubularian Hydroids. Bay Society. 1871—1872. 16. Brooks, W. K. On the Life-history of Eutima, and on Eadial and Bilateral Symmetry in Hydroids. Zool. Anzeiger. Jahrg. vii. 1884. 17. Brooks, W. K. The Life-History of the Hydromedusse : A Dis- cussion of the Origin of the Medusae, and of the Significance of Metagenesis. Mem. Boston Soc. Nat. Hist. Vol. iii., p. 359, plates 37—44. 1886. CNIDARIA 129 18. Brooks, W. K. The Life-history of Epenthesis McCradyi, n. sp. Stud. Biol. Lab. Johns Hopkins Univ. Vol. iv. 1888. 19. Ci.uiiciAN,. J. Ueber den feineren Bau u. die Entwicklung von Tubularia mesembryanthemum. Zeitschr. wiss. Zool. Bd. xxxii. 1879. 20. Claus, C. Beitrage zur Kenntniss der Geryoniden und Eucopi- den-Entwicklung. Arbeiten Zool. Inst. Wien. Bd, iv. 1882. 21. Claus, C. Entwicklung des iEquoriden-Eies. Zool. Anzeiger. Jahrg.w. 1882. 22. Conn, H. W. Development of Tubularia cristata. Zool. Anzeiger. Jahrg. v. 1882. 23. Davidoff, M. Ueber Theilungsvorgange bei Phialidium variabile Haeckel. Zool. Anzeiger. Jahrg. iv., No. 98. 1881. 24. Dujardin, F. Memoire sur le developpement des Meduses, etc. Ann. Sci. Nat. Ser. 3. Tom. iv. 1845. 25. FoL, H. Die erste Entwicklung des Geryoniden-Eies. Jeiia. Zeitschr. Bd. vii. 1873. 26. Haacke, W. Zur Blastologie der Gattung Hydra. Jena. Zeitschr. Bd. xiv. 1880. 27. Hamann, 0. Beitrage zur Kenntniss der Medusen. Zeitschr. wiss. Zool. Bd. xxxviii. 1883. 28. Hamann, 0. Der Organismus der Hydropolypen. Jena. Zeitschr. Bd. XV., p. 480. 1882. 29. HicKsoN, S. J. On the Sexual Cells and the Early Stages in the Development of Millepora plicata. With two plates. Phil. Trans. Roy. Soc. London. 1888. 30. HiNKs, T. On the Development of the Hydroid Polyps, Clavatella and Stauridia, with Eemarks, etc. Brit. Assoc. Bep. 1861. 31. Jung, H. Beobachtungen iiber die Entwicklung des Tentakelkranzes von Hydra. Morph. Jahrb. Bd. viii. 1882. 32. Kerschner, L. Entwicklungsgeschichte von Hydra. Zool. An- zeiger. Jahrg. iii.. No. 64, p. 454. 1880. 33. Kleinenberg, N. Hydra, eine anatomisch- ent wicklungsgeschicht - liche Untersuchung. Leipzig. 1872. 34. Koch, G. v. Vorl. Mittheilung iiber Coelenteraten. Jena. Zeitschr. Bd. vii. 1873. KoLLiKER, A. Ueber Stomobrachium mirabile. Zeitschr. wiss. Zool. Bd. iv., p. 326. 1853. KoROTNEFF, A. Zur Kenntniss der Embryologie von Hydra. Zeitschr. wiss. Zool. Bd. xxxviii. 1883. KoROTNEFF, A. Cunoctantha and Gastrodes. Zeitschr. wiss. Zool. Bd. xlvii. 1888. 38. Lang, A. Gastroblasta Raffaelii, eine durch eine Art unvoUstandiger Theilung entstehende Medusen-Colonie. Jeiia. Zeitschr. Bd. xix., 1886, und Bd. xx., Suppl. Heft 1. K. H. E. 130 EMBRYOLOGY 39. Lendenteld, E. v. Ueber eine eigenthiimliche Art der Sprossen- bildung bei Campanulariden. Zool. Anzeiger. Jahrg. vi. 1883. 40. LovftN, S. Beitriige zur Kenntniss der Gattungen Campanularia u. Syncoryne. Arch. Naturg. Jahrg. in., Bd. i., pp. 249 u. 321. 1837. Translation. 41. Marshall, W. Ein neues Susswasser-Coelenterat von Nordamerika, Microhydra Eyderi Potts. Biol. Centralbl. Bd. vi., p. 8. 1886— 1887. 42. Merejkowsky, C. de. Histoire du developpement de la meduse Obelia. Bull. Soc. Zool. France. Tom. viii. 1883. 43. Metschnikoff, E. Vergl. embryologische Studien. Zeitschr. wiss. Zool. Bd. xxxvi. 1882. 44. Moseley, H. N. On the Structure of the Stylasterids. Phil. Trans. Boy. Soc. London. 1878. 45. MuLLER, JoH. Ueber eine eigenthiimliche Meduse des Mittelmeeres und ihren Jugendzustand. Arch. Anat. u. Phys. 1851. 46. Schulze, F. E. Ueber den Bau und die Entwicklung von Cordylo- phora lacustris. Leipzig. 1871. 47. TiCHOMiROFF, A. On the Embryology of Hydroids (Kussian). Mem. Boy. Soc. Friends of Nat. Sci., Anthropol., and Ethnogr. Moscoic. 1887. 48. Ussow, M. Eine neue Form von Siisswasser-Coelenteraten. Morph. Jahrb. Bd. xii. 1887. 49. Weismann, a. Die Entstehung der Sexualzellen bei den Hydro- medusen, zugleich als Beitrag zur Kenntniss des Baues u. der Lebenserscheinungen dieser Gruppe. Jena. 1883. 50. Weismann, A. Die Entstehung der Sexualzellen bei den Hydro- medusen. Biol. Centralbl. Bd. iv. 1884. 51. Wilson, H. V. The Structure, of Cunoctantha octonaria in the Adult and Larval Stages. Stud. Biol. Lab. Johns Hopkins Univ. Vol. iv. 1886. Appendix to Literature on Hydroidea. I. Brauer, a. Ueber die Entwicklung von Hydra. Zeitschr. iciss. Zool. Bd. Hi. 1891. n. Brauer, A. Ueber die Entstehung der Geschlechtsproducte und die Entwicklung von Tubularia mesembryanthemum Allm. Zeitschr. wiss. Zool. Bd. lii. 1891. III. Gerd, W. Zur Frage iiber die Keimblatterbildung bei den Hydro- medusen. Zool. Anzeiger. Jahrg. xv. 1892. IV. Haecker, V. Die Furchung des Eies von Aequorea. Arch. mihr. Anat. Bd. xl. 1892. V. Heider, K. Ueber Gastrodes, eine parasitische Ctenophore. Sit- zungsber. Gesellsch. ^aturf. Freiaide Berlin, 1893. CXIDARIA 131 VI. HiCKSON, S. J. On the Maturation of the Ovum and the Early Stages in the Development of Allopora. Quart. Jour. 2Iicr. Sci., n.ser. Vol. xxx. 1890. VII. HicKSON, S. J. The Medusae of Millepora Murrayi and the Gono- phores of Allopora and Distichopora. Quart. Jour. Micr. Sci., n. ser. Vol. xxxii. 1891. Vll.a. HiCKsoN, S. J. The Early Stages in the Development of Disticho- pora violacea, etc. Quart. Jour. Micr. Sci., n. ser. Vol. xxxv., p. 129. 1893. VII b. Hyde, Ida H. Entwicklungsgeschichte einiger Scyphomedusen. Zeit-^ &)^f ^- 184 EMBETOLOGY contrary, this stage of development happens in the warm season, then creatures of another shape are developed from the germ cells : the tailed Cercarice. (Fig. 89 E [contained individuals] and F). In other cases the Cercariaa arise only in the Redise of the second generation. The mode of origin of the Cercarice has been thoroughly studied by ScHWABz (No. 9). As has already been mentioned, this investigator finds in their origin a great resemblance to the development of the embryo. The THoruZa-like heap of cells which arose from the germ cell is further developed in such a manner that there are difi'erentiated a peripheral cell-layer, a central compact mass of cells, and a layer between the two. The first supplies the dermal layer, which is to be considered as a meta- morphosed epithelium ; from the central mass arise the genital organs, whereas the intermediate parts of the embryonal tissue (the " meristem " of ScHWAEz) give rise to the other organs. Anterior to the central cell- mass a number of cells are arranged in a regular manner. This is the fundament of the intestine, which later becomes hollowed out and con- tinuous with the two branches of the intestine, which have arisen in the same way. The central part of the excretory apparatus is also formed by means of such a regular arrangement of cells in the posterior part of the body. The dermo-muscular layer and the fundament of the nervous system arise nearer the periphery. The remaining part of the "meri- stem " becomes the parenchymatous tissue of the body. The Cercaria already exhibits to a certain extent the organization of the adult Bistomum^ e.g., in the presence of an anterior sucker and one situated on the ventral side (Fig. 89 F). In the centre of the former lies the mouth, which leads into the muscular pharynx, and thence into the forked intestine. The oesophageal ganglion, with the two lateral stems, and also the bipartite excretory system are present. But a long muscular tail is attached to the posterior portion of the body. In this condition the Cercaria leaves the Redia through the birth aperture, which lies at the anterior end (Fig. 89 E, G), and seeks an escape by work- ing its way through the tissues of the host by means of its suckers and tail. Its free life in the water lasts for only a short time. It soon attaches itself to plants which are found at the water's edge. It casts off the tail, and secretes about itself a cyst. A large number of glands which lie on either side in the body of the Cercaria, and which give a PLATHELMTNTHES I aracteristic appearance to tlie animal, serve for this pur- (Fig. 89 F). These glands appear on the free Cercaria as white opaque masses ; but when their contents have passed out during the encjstment, the body of the young worm becomes entirely transparent (Fig. 89 G). If the cyst, together with the plant to which it is attached, is swallowed by a sheep, the envelope is dissolved in its stomach ; the young worm becomes free, and finally reaches the liver, where, in the course of about six weeks, it develops into the sexually mature Distomum hepaticum. The different Distomidae present great differences as to the course of their de- velopmental processes. The eggs from which embryos are to emerge do not always become free, but may be taken up directly by the intermediate host, and hatched out only when they have reached its intestine (Distomum ovocaudatiim, ac- cording to Leuckart). It is not neces- sary that a sporocyst should be first developed out of the embryo, and a Redia out of it, as in Distomum, hepaticum, but the sporocyst may become metamor- phosed directly into a Redia. Sporocyst and Redia in most cases beget directly Cercariae. The sporocyst in Distomum macrostomum and Gasterostomum fimhri- atum is very aberrant in shape. In this species it develops tubular processes, which serve for the reception of the Cercariae. The sporocyst of Distomum m,acrostomum, known as Leucochloridium^ which inhabits the liver and other organs of Succinea amphibia, attains an extra- ordinarily large size, for it sends out processes into the antennae of the snail, where, on account of their external re- semblance to insect larvae, they are seen I and eaten by birds (Zellee, Heckert). I Fig. 90.— Cercaria Vil- loti, Monticelli (after Vir.LOT). 186 EMBRYOLOGY The Cercarise produced in germ tubes present a variety of forms. This applies chiefly to the caudal appendage, as can be recognized in the peculiarly formed Cercaria represented in Figs. 90 and 91. One of these, Cercaria setifera Villot} a marine form, which arises from a sporocyst inhabiting Scrobicularia tenuis, possesses an extraordinarily large tail, beset with bristles. The other (Fig. 91) has two tails, which are directed forwards, however, in swim- ming. This is the Cercaria of G aster ostomum jimbr latum, and is known under the name of Bucephalus polymorphus. Fig. 91.— Cercaria of Gasterosfoinum/mbriafum (after Zieglejr). Under certain conditions the tail is entirely wanting in the Cercaria stage. This is the case when the Cercarise are not compelled to undertake a migration, but remain in their host until, along with it, they are consumed by another animal, the final host. Since they do not pass through a free stage, they do not require any special organs of loco- * The Cercaria setifera of Villot is called Cercaria Villoti by Monti- CELLi, for the term setifera occurs in another species (Monticelli, " Sulla Cercaria setifera Miiller," Bolletino di Natiirahsti in NapoH, vol. ii., 1888). PLATHELMINTHES motion. The tailless Cercarise of Distomum macrostormtvi (produced in Leucochloridium paradoxum) , together with parts of the germ tubes, arrive in the intestine of the final host [birds], in the cloaca of which they become sexually mature (Zeller). As a rule the Cercaria passes, by an active migration, from its first intermediate host into a second, which naturally is also an aquatic animal, either another snail or a worm, crustacean, insect, mollusc, fish, or amphibian. In this second intermediate host it casts off the tail and becomes encysted. The young worm awakens to new life only after its host has been taken as food and digested by some other, usually higher, animal. In this way the cyst is dissolved, and the young Distomum now reaches the stage of the sexually mature animal. But we have seen that in Distom.um hepaticum the second intermediate host may be omitted, and that the Cercaria, after becoming encysted in the free condition, passes directly into the final host. The state- ment, often made, that tailed Cercariae could migrate directly into the final host (for example, the Cercaria macrocerca of Bistomum cygnoides into the urinary bladder of the frog), has not been sub- stantiated. On the contrary, these Cer- cariae appear to be obliged to pass through the encysted stage. "P R Fig. 92.— Embryo of JUonostomum mutabile, shortly after hatching (after v. Siebold). R, Redia. A most remarkable condition is presented by the embryos of Monostomum mutahUe and M. flavum, two Distomidce, which are found in the thoracic and orbital cavities of various aquatic birds. The embryos abandon the egg-membrane when still in the uterus of the parent. These Distomids are therefore viviparous. In each embryo a Eedia-like creature is already present (Fig. 92). In this case, therefore, the embryo produces the new generation even before it has time to find an intermediate host, within which to develop into a sporocyst. There is scarcely a doubt but that the bud is formed from the germ cells of the embryo. 188 EMBRYOLOGY II. Polystomidj:. The eggs in the Polystomidae also are composed of the egg-cell proper and yolk-cells (Fig. 93). Their egg-raera- brane is provided with an operculum, and occasionally with a long filiform, and twisted process, which serves for the attachment of the eggs (Diplozoon). The course of develop- ment is simpler than in the Distomidae, for the embryo while still in the egg-membrane attains nearly the form of FrG. 93.— Egg of Microcotyle Mormyri. Within its operculated shell lies an egg- cell surrounded with yolk-cells (after LoBENz, from Hatschek's Lehrhuch). Fig. 94.— Embrvo of Polystovuim integerrimum, shortly after hatch- ing (after Zki.ler). the parent (Gyrodactylus), or at least passes through only a single metamorphosis, not an alternation of generations (heterogony). The early development has been but little studied. We are best acquainted with it (Zeller, ISTos. 16 and 17) in the case of Polystomum integerrimum, which inhabits the urinary bladder of the frog. The eggs of this species are voided into the water, where cleavage soon begins. The result of this PLATHELMINTHES 189 is a spherical mass of cells, which subsequently becomes elongated, and thereby exhibits, even at this stage, the form of the embryo. The fundaments of the eyes, the sixteen hooks of the clasping disc [retinaculum], the cavity of the intestine, and the pharynx soon make their appearance (Fig. 94). The newly hatched embryo possesses in addition five rows of cilia, of which the three anterior belong to the ventral surface, the two posterior to the dorsal surface. Furthermore there is a fringe of cilia in front on the head (Fig. 94). The embryo, leaving the egg at this stage, now seeks the tadpole of the frog, to the gills of which it attaches itself by means of the hooks and suckers. Here the ciliated cells, which are no longer of any use to the animal, degene- rate, and the Polystomum larva approaches more and more the form of the parent. In extraordinary cases it can attain this condition even in the branchial cavity, but as a rule this is not the case ; on the contrary, the young Polystomum, upon the degeneration of the gills of the tad- pole, penetrates into its mouth-cavity, migrates through the entire length of its intestine, and finally passes from the cloaca into the urinary bladder, where it attains sexual maturity. DipJozoon paradoxum, which is remarkable on account of its subsequent habits, also leaves the egg as a ciliated larva (Zeller, No. 18). The larva, known under the name of Diporpa, bears suckers and hooks, by the aid of which it attaches itself to the gills of fresh- water fishes (Phoxinus Uevis, for example). It can remain here for weeks and months, gradually approaching the organization of the adult. But before it arrives at this condition it is necessary for one individual to unite with a second, and, in fact, for the rest of their existence. This takes place by the larva seizing with its ventral sucker a knob-like outgrowth situated on the back of the other animal. Then the second individual turns and twists its body, so that it too may grasp the dorsal prominence of its mate with its ventral sucker. In this position the two animals grow together firmly, and in this I condition reach sexual maturity. i 190 EMBRYOLOGY the PolystomidcR, also living on the gills of fishes, is very re- markable. Its reproduction approaches that of Monostomum, already described, for in this species also the embryo while still in the body of the parent contains another embryo; indeed, the latter already exhibits within itself traces of a new individual, so that four generations are included one within the other (Wagenee, Metschxikoff). Accordingly here, as in Monostomum, the germ cells produce the new generation very early; but otherwise this developmental process is not very different from that of the other Trema- toda. In order to understand the cause of this accelerated production, one Avould have to know more accurately the processes themselves, as well as the habits of the animal. III. CESTODA. The eggs of the Cestodes exhibit a close resemblance to those of the Trematodes. Like these, they are composed of the egg-cell proper and a number of yolk-cells ; where the latter are wanting, an accessory yolk-mass corresponding to them appears to be present. The eggs are surrounded by a thin egg-membrane, which occasionally possesses a movable lid. The development of the eggs takes place for the most part in the uterus of the parent, but in many forms it occurs only after the eggs are laid. In the- latter case the mem- brane is thicker. The investigations of E. VAN Beneden and Villot on the TcBiiiadoe, and especially those of Schauinsland on the Bothrlocejphalidoe, have shown that the embryonic develop- ment of the Cestoda takes place in a manner quite similar to that of the Trematoda. According to Schauinsland, the development of the Bothrio- cephalidcB is accomplished in two different w^ays, depending upon whether the embryos are developed before, or after ovi- position. The undeveloped eggs which are deposited in the water are thick-shelled, operculated, and provided with a large number of yolk-cells. From them emerge larvae which bear a thick coat of cilia. The eggs of the second kind are thin-shelled, without an operculum, and provided with only PLATHELMINTHES 191 a relatively small amount of yolk material. The embryos contained in them are naked. The embryonic development of the Bothriocepha- lldaB approaches closely that of the Distomidse. Cleavage takes place in much the same way as there. At an early period two cells are differentiated at the two poles of the elongated germ, upon which they rest like a cap. They then grow around it, and constitute the enveloping membrane (Hilll- tnembran) . Afterwards another cell is separated off from the spherical cell-mass surrounded by the enveloping mem- brane, and this at first also covers the germ like a cap, and then grows around it. Later this external layer consists of several cells. It is in this way that the ectoderm is formed. The embryo now consists of a single layer of ectoderm and a solid entodermal mass (Fig. 95). Six chitinous hooks make their appearance in the latter. With this the forma- tion of the embryo is completed. It is said to be composed of the inner (entodermal) mass only. The ectoderm separates from it, so that a space arises between the two. The embryo is now surrounded by two envelopes in addition to the egg-membrane, the ectodermal mantle, and the enveloping membrane. In this respect, too, the conditions described for the Distoraidse are repeated, and a comparison of Fig. 95 with Fig. 88 (on p. 180) shows without further comment the close agreement of the two groups at this stage of development. Whereas the embryo quitting the egg leaves the enveloping membrane behind in the egg-shell, it takes the ectodermal mantle with it (Fig. 95). The latter either serves actively in locomotion when it possesses cilia, or it swells up so much in the water that it serves the larva both as a protective Fig. 95,— Embrj'O of BotKrioce- phalus latus pressed out of the egg. Ec, ectjderm ; Km, enveloping membrane (after ScHA.TJiifSLA.ND). 192 EMBRYOLOGY envelope and as a means of making it of nearly the same weight as water, thereby enabling it to float. Where cilia are present, they are at first short, and only gradually in- crease in length. In Bothriocephalus latus the exceedingly delicate cilia attain a very great length. After the larva has floated about in the water for a time, under certain con- ditions for several days, it divests itself of the mantle, whether ciliated or not. In many cases (as sometimes even in Bothriocephalus latus) it may at the very beginning strip off the mantle with the enveloping membrane. Even in this naked condition the larva may live free for a time, but finally perishes, if it finds no suitable host. ScHAUixsLAND explains the circumcrescence of the germ by the cap- shaped cells, which occurs twice in nearly the same way, as an epiboly. Accordingly he is compelled to assume a complete loss of the ectoderm in the casting off of the superficial layer. The embryo is developed out of the entoderm alone. He finds a support to this view in the fact that up to the present time no actual body epithelium has been found either in the Cestoda or in the Trematoda. This fact is in his opinion an argu- ment that ectodermal structures are not present in these cases, a view that Leuckart (No. 8) also maintains. In any event the origin of the cuticula-like dermal layer merits a thorough investigation. If, as is to be conjectured, it arises by the metamorphosis of a superficial cell-layer (E. ZiEGLER, ScHWAEZE, ct alii), then it would correspond to the body epi- thelium. The question whether in the casting off of the outer layer the entire ectoderm is removed, or whether certain of its cells still remain behind, must be difficult to determine on account of the small size of the ^ [As is well known, a distinct epithelium could not be found on the external surface in Cestodes and Trematodes. It was natural to connect this fact, the absence of the body epithelium, with the casting off of the external cell-layers in the embryo, and thus to assume that the entire ectoderm was lost. A body epithelium, therefore, could not be present. This question has often been considered, and even recently has been re- sumed. While some investigators assume that the cuticula which covers the body is secreted by the subcuticular layer, and that the latter is a part of the body parenchyma (Bkaxdes, Loos), others maintain that it is a metamorphosed epithelium, and believe they see, more or less distinctly, cell nuclei retained in it (Braun, Monticelli). The most of these obser- vations refer to the Trematodes, although investigations in this direction have also been made on Cestodes (Zograff, Grassi ; see Appendix to Literature on Cestoda). Zograff in particular finds that in various Ces- PLATHELMINTHES 193 The formation of the larval membranes in the Trematodes and Ces- todes recalls in a striking manner the Amnion and Pilidium in the Nemer- teans. Since, however, similar processes do not occur in the Turbellaria, — to which relationships are shown by the Trematodes and Cestodes on one side, and by the Nemerteans on the other, — and since the Turbellaria are to be considered as the more primitive forms, we have here to do with only analogous phenomena. The embryonic development of the Taeniadae differs to some extent from that of the Bot]irioce2)halidoe, but leads finally to a similar result (Leuckart, No. 8 ; Moniez, No. 9 ; E. VAN Beneden, No. 2). A difference is caused from the very beginning by the yolk-material bestowed upon the egg being less abundant, or not in the form of distinct cells. In Tcenia serrata the egg- cell lies embedded in this yolk-material. In other cases the yolk appears to enter into still more inti- mate relations with the egg-cell ; however, it appears from the somewhat various statements of the authors concernino^ the different forms that even in these cases the nutritive material becomes separated as early as the first divisions of the egg. There are one or several rather voluminous, gene- rally granular cells, which are thus at first constricted oft* and then consumed, while the other cellular matter multiplies further. In Tcenia cucumerina, it is true, the entire egg is said to be transformed directly by means of a rather regular cleavage into the embryonic cell-mass (Moniez). In the further development of the Taeniadae we can find again the characters which we observed in the Bothriocephalidae, al- though the details of the process are somewhat different. In the Taeniadae also certain cells detach themselves at an early period, and grow around the germ as its enveloping membrane. In the Taeniadee known as the Bladder-tape- worms, the second membrane may present an appearance somewhat different from that with which we have thus far acquainted ourselves. It becomes cuticalarized, assumes a radially striated appearance, and thus finally forms a firm membrane about the embryo, which even in this stage is todesthe subcuticular matrix is independent of tbe connective-tissue body parenchyma, and explains how in the embryo an ectodermal cell-layer still remains behind after the casting off of the ciliated mantle.— K.] K. H. E. () 194 EMBRYOLOGY equipped with three pairs of hooks. Furthermore, according to VAN Beneden, a cortical layer can early be distinguished from the differently constituted internal cell-mass ; and ScHAUiNSLAND also spcaks of smaller peripheral cells and larger central ones. It is natural to regard this as a differ- entiation into the two germ-layers, though Schauinsland be- lieves that such is not the case. According to him, the entire ectoderm, with the two membranes, is excluded from further participation in the formation of the embryo, which consists exclusively of a homogeneous cell-mass : the entoderm. This point, and especially the origin of the layers of the embryo, appears to us in urgent need of renewed investigation. With Schauinsland, we regard the homology of the embryonal mem- branes of the Bothriocephalidce, Tceniadce, and Distomidce as unquestion- able. The different development of the second membrane — in the one case into a ciliated layer, in the other into a chitinous layer — is determined by the mode of life of the particular worms. Some of them inhabit animals which continually come in contact with water. In these the deposited eggs develop very quickly and require no special protection. The others inhabit land animals. Their eggs reach the outside world while still within the proglottis, and the more the already developed embryos are protected against desiccation, the better their prospects for existence. Hence the development of the chitinous membrane. In such Taniadce, on the contrary, as inhabit aquatic animals, the chitinized embryonal membrane may be absent, and in place of it there may appear a thin membrane, similar to the non-ciliated ectodermal mantle of many Bothriocephalidce (ScHAtriNSLAND, No. 12). The further development of the six-hook embryo (Fig. 96 A) takes place only after it has migrated into an intermediate host. Either this may take place directly, — when the embryo, as in the Bothriocephalidce, is a free- swimming larva, and so at once migrates into an aquatic animal, — or the embryos, still enclosed in the egg-membrane, may enter by passive means into the intermediate host. Generally this happens by the segment of the tapeworm, which crawls about on plants, being swallowed with the food. The proglottis is digested in the stomach, the eggs thereby become free, their membrane ruptures, and the embryos now find themselves within the intestinal canal. They do not remain there long, but penetrate into the in- PLATHELMINTHES 195 testinal wall by means of the boring movements of their booklets. In this way apparently they arrive in the blood- vessels, and are probably carried along by the blood current, finally to take up their permanent abode in various organs, very frequently in the liver, sometimes in the brain, in the musculature, etc. There a vigorous growth soon begins ; this is connected with a simultaneous activity of the sur- rounding tissues, which form a membrane about the intruded foreign body. The latter now casts off its hooks, and on its surface there appears a rather thick cuticula, under- neath which circular and longitudinal muscle fibres are differentiated. Beneath these there follows a cortical layer resembling connective tissue, which differs from the central parenchymatous tissue (Fig. 96 B). The latter soon exhibits spaces, in which an aqueous fluid makes its appear- ance. By the coalescing of these spaces with one another, a large cavity filled with fluid finally arises within the body. Herewith the development of the tapeworm has reached the stage which is known as the Cysticercus, hladder-worm, or hydatid. It has been compared to the sporocyst of the Trematoda, although it presents no particular resemblance to it either in structure or in regard to its further develop- ment.^ The excretory system has the same organization in the bladder-worm as in the tapeworm. It is composed of capil- laries which arise in ciliated funnels in the tissues, and discharge into larger stems. The latter .unite into the chief trunks, which may fuse to form a short sac at the posterior end and there open to the exterior (Gr. Wagener, Leuckart).^ ^ [In many cases the formation of a cavity in the Cysticercus is greatly reduced or becomes entirely suppressed. There are found in the lungs of crows and in the body-cavity of Lacerta vivipara, for example, Cysticerci of this kind {Pietocystis variabilis and P. dythiridium Diesing), the body of which is filled with a continuous connective tissue (Leuckart). Such Cysticercus stages of Cestodes have been designated by the name Plerocerci and Plerocercoids (M. Braun), — by the latter when the scolex is only slightly marked off from the bladder. Such, to a certain extent aberrant, bladder-worms are found in the Tseniadse, as well as in the Bothriocephalidae and other Cestodes. — K.] " [The Cysticerci with long caudal appendages, which occur in in- 196 EMBRYOLOGY The Cysticercus may remain for a longer or a shorter time in the condition described, but may increase meantime in '■J*?^,, ^CsE. Systematic : van Beneden distinguishes four genera : Hcyema, Dlcyemella, Dicyemina, and Dicyemopsis, which are listributed among four genera of Cephalopods : Octopus, Uedone, Sepia, and Sepiola. They are found in the append- Lges of the branchial veins. Whitman admits only two genera : Dicyema (with eight cells in the head region) and Hcyemennea (with nine cells in the head region). The body of the Dicyemidce is elongated. It consists of tn outer layer of ciliated cells and a single large axial cell, the latter surrounded by the former (Fig. 99 D). At the interior end the outer cell-layer is differentiated into a kind ►f cap [polar calotte]. Elsewhere the outer cells are nearly ilike. K. H. E. 210 EMBRYOLOGY A certain difference in individuals is manifested in the manner of their reproduction. The latter consists in the production of embryos in the axial cell. But these are of different shapes ; vermiform and infusoriform (rhomboid) embryos can be distinguished (Figs. 99 and 100). They arise in different individuals, which, according to VAJf Bene- DEN, are recognizable even by their outward form. The nematogenous individuals are longer and more slender, the rhomhogenous shorter and more compressed. According to Whitman, in addition to the forms that bring forth only vermiform embryos, and which he designates as primary Nevmtogevs, there also occur forms in which at first infusoriform and later vermiform embryos are produced {secondary Nematogens). Development of the Vermiform Embryos.— There can hardly be any doubt that the cells which constitute the earliest fundament of the reproductive elements, and which correspond to the genital cells of the other Metazoa, take their origin by the division of the axial cell of the parent. The products of this process of division are, however, not equivalent ; moreover, the newly formed cells remain in the axial cell (Fig. 99), whereby the appearance of an endoge- nous cell-proliferation is produced. The production of the germ cells begins very early, for even in embryos there is to be seen inside the axial cell and behind its nucleus a new cell undergoing differentiation, the first germ cell (Fig. 99 A), and a second one soon arises in its anterior part (Fig. 99 B and C). Their nuclei have very probably arisen by division from the nucleus of the axial cell. Subsequently the latter takes absolutely no part in the formation of new nuclei. It appears to preside over the other cell functions only. The two germ cells, on the contrary, begin to increase by division, and soon furnish a large number of genital cells, from which the embryos subsequently arise. The development of the germ cells, which are eventually present in large numbers within the axial cell of the parent, takes place in situ after the manner of cleavage. An epibolic gastrula is formed here, as in the Orthonectidce, except that its inner large cell remains undivided. It becomes the axial cell. ORTHONECTID^ AND DICYEMIDiE 211 r ^^Bj increasing in length the embryo becomes vermiform, ^^jvhence its name (Fig. 99 B and C). These embryos are not essentially different from the adult animal, whose shape is soon fully assumed by the accomplish- ment of the slight differentiations in the outer layer of the body and in the head region, and by the elongation becoming more pro- nounced (Fig. 99 (7 and D). Then the formation of new germs in the axial cell begins very early, in fact while the embryo still remains within the parent. The processes described apply there- fore to embryos which are still found within the parent (Fig. 99 AtoD). When they have arrived at maturity, they break through the outer layer of the parent, but remain in the venous appendages of the Cephalopods, where they still grow considerably and pro- duce other embryos. Structure and Develop- ment of the Infusoriform Embryos. — The infusoriform embryos differ widely from the vermiform in shape. Of a shorter, more compressed form, they also present numerous internal differentiations (Fig. 100 BtoF). In swimming, the broader end of the embryo is directed forwards. Whereas the an- terior end is naked, the rest of the body is ciliated (Fig. 100 C and D). The entire embryo is constructed on the bilateral plan, for two lateral parts as well as a dorsal and ventral side can be distinguished. Anteriorly and more dorsally lie two highly refractive bodies (Fig. 100 D, r), somewhat behind them, and lying more ventrally, the organ called by van Beneden the "urn." This peculiar organ, the function of Fig. 99.—^ to D, stages in the development of the veimiform embryos of Dicyema ; A, of Di- cyemennea eledones (after Whii- man); B to D, of Dicyema typus (after E. van Bknedeh). Ax, axial cell ; K, nucleus of the axial cell ; Kz, germ cells. 212 EMBRYOLOGY which is not clear, is composed of a shell-like envelope, a granulated body contained in it, and a lid. The shell lies with its cavity toward the ventral side TFig. 100 F). It consists of two parts and, owing to small comma-shaped bodies embedded in its free edge, acquires a striated appear- ance (Fig. 100 D, U). Its contents consist of four large cells of nearly equal size, which lie close together, and are granular. Finally, the lid, which corresponds to the ventral part of the urn, consists in turn of four cells, which unite, at the point where they all abut on one another, to form the knob of the lid (Fig. 100 D to G, I). Within the urn van Beneden sometimes obser.ved a ciliation, which he ascribed to the granulated cells. Fig. 100.—^ to G, infnsorif orm embryos and their development— .4 to D, of Dicyema typws ; EtoG, of DicyemelJa Wagnerii (after van Benbdkn, from Balfoub's Compara- tive Embryology). A to C, stages of development ; D, embryo seen from the ventral side; E, from the right side ; F, from the front ; G, the " urn " isolated; gr, granu- lated cells contained in the urn ; I, its lid ; u, the shell, which forms the floor of the urn; r, highly refractive bodies at the anterior end of the embryo. The origin of the infusoriform embryo, although at first sight quite different from that of the vermiform embryo, can perhaps be referred to this. It takes place in the axial cell of the rhombogenous individuals, though not directly, being introduced by a preparatory process (Whitman). Near the nucleus of the axial cell there arise two new cells, the nuclei of which in all probability originate from the nucleus of the axial cell. These two cells soon multiply, but not so rapidly as in the formation of the vermiform embryos. They never exceed eight in number, and often only a few are present. Before these cells develop further they undergo ORTHONECTIDiE AND DICYEMID^ I a process which Whitman compares to the formation of the polar globules in the eggs of the Metazoa. As the result of a process of division a considerable portion of the nucleus is said to be cast out of them, which, as the " paranucleus," j^^ can be recognized for a long time in the axial cell (Fig. 101 I^H^). Then ensue a cleavage of the cells and, as its result, ^^ the formation of cell-masses which have quite the appear- ance of an epibolic gastrula with a central cell. Such stages ihad already been observed bj van Beneden (Fig. 101 A). Thej are entirely like those which occur in the development of the vermiform embryos. Whitman compares them directly to these, and looks upon them as special individuals, which appear early in the course of reproduction. For in their central cells there are soon formed new cells (Fig. 101 J. and B), which subsequently give rise to the infusoriform embryos. On this account Whitman calls this gastrula stage an Infusorigen. Compared to the nematogenic developmental series, the gastrula stages would correspond to the vermiform embryos, which, as we saw, also produce embryos at a very early period. From the central cell (cellule ger- migene of van Beneden) of the gas- trula stage, which increases in size, arise several generations of germ cells, which surround it in the form of a rosette.^ The larger of these cells become infusoriform embryos ; the smaller ones are said subsequently to divide repeatedly, and vermiform embryos are said to arise from them when, after the formation of the in- ^ The central cell itself is to be looked upon as the homologue of the axial cell of the vermiform embryos. Fig. 101. — ^, "Infuso- rigen embryo " (after van Beneden) ; B, the same lying in the axial cell (^a;) of the rhombogen indi- vidual (after Whitman). A, of Dicyema typus ; B, of Dicyemennea eledones. C, the central cell of the " In- fusorigen embryo," which has already produced new germ cells ; K, nucleus of the central cell ; Ke, nuclei of the outer layer of the rhombogen individual ; Pn, paranucleus. On the right side of Fig. 101 B the re- ferences K and Ke are transposed. 214 EMBRYOLOGY fusoriform embryos, the rliombogen individuals have entered upon the second phase of their development {secondary Nema- togens according to Whitman). The formation of the infusoriform embryos from the germ cell also begins with a process of cleavage, the result of which is an epibolic gastrula (E. van Beneden). However, in this case several cells make their appearance in the centre, at first four large ones (Fig. 100 A, 21). Two of these be- come the shell and two the lid of the urn ; whereas four smaller cells, which arise later, supply the four granular cells contained in the urn (Fig. 100 B and G, gr). In the meantime the two highly refractive bodies have made their appearance in the outer layer of the embryo (Fig. 100 A, 1>, r), and its posterior portion has become covered with cilia. Whereas at first the embryonal cells which become the urn lie side by side, they subsequently alter their position so that the granular cells become enclosed above and below by the lid and shell of the urn. Nothing definite is yet known about the significance of the infusoriform embryos. From the fact that they can be kept alive in sea-water for days (E. van Beneden), it was thought that these forms were probably for the purpose of transferring the species from one cephalopod individual to another, where they would develop into a form which, like the vermiform embryos, produces new germs. Besides this view, there is a second one, which compares the infusoriform embryos to the male of the Orthonectidae. Van Beneden is inclined to see in the granular and vibratory contents of the urn the homologue of the testis of the Orthonectidae. Whitman several times observed the penetration of infusori- form embryos into nematogen individuals, which is perhaps to be compared to a process of fertilization. Related to the Dicyemidae are the Heterocyemidce {Conocyema and Microcyema), observed by van Beneden (No. 2), which also inhabit the appendages of the veins of Octopus and Sepia. Their shape differs from that of the Dicyemidas inasmuch as they do not nearly attain the length that these do, and wart-like structures are present at the anterior end, which can be extended and withdrawn. Nematogen and rhombogen individuals are also distinguished here. Although the vermiform embryos 215 differ somewhat from those of the Dicyemidae, yet on the whole they develop like them. The infusoriform embryos of Conocyema resemble those of the Dicyemidae. General Coxsiderations. There are so many common features in the structure and development of the Orthonectidae and Dicjemidae that we cannot doubt the relationship of the two groups. Their relations to the other divisions of the animal kingdom, on the contrary, are more difficult to understand. In view of th6 fact that they are said to be composed of only two germ- layers, an attempt was made to create out of them a new division of the animal kingdom, that of the Mesozoa, which would be interpolated between the Protozoa and the Metazoa (E. VAN Beneden, Julin). Since it is only parasitic forms with which we have to do, such an explanation seems to us venturesome at least, and we consider it more probable that these simply constructed animals are Platyhelminthes which have become degenerated through parasitism (Leuckart, Metschnikoff, Whitman). The resemblance of the female of the Orthonectidae to the embryos of the Distomidce has already been pointed out. The theory that such embryos have reached sexual maturity has nothing improbable about it, for such cases are also known elsewhere in the animal kingdom. Thus Dinophihis is evidently to be regarded as an annelid larva which has become sexually mature, and it is appropriate for comparison here, inasmuch as its males have degenerated to nearly the condition of the Orthonectidae and Dicyemidae (comp. infra, p. 313). The intestine and other features of a higher organization having been lost, they present within the body only a large testicular sac, similar to the males of the Orthonectidae, which, to be sure, remain at a somewhat lower stage. If we regard the Orthonectidae and Dicyemidae as degenerated forms, then the Orthonectidae, with their central cell-mass, would represent the higher grade, whereas the Dicyemidae, in which only one central cell is present, are more degenerate. However, in these also the inner por- tion becomes multicellular as soon as the formation of the germ cells by the division of the axial cell begins. Literature. Orthonectid^. 1. Braun, M. Die Orthonectiden. Centralbl. Bakt. u. Parasiteiikunde. Bd. ii. 1887. 216 EMBRYOLOGY 2. GiARD, A. Les Orthonectida, classe nouvelle du phylum des vermep. Jour. Anat. et Physiol, Norm, et Path. Tom. xv. 1879. 3. JuLiN, C. Contribution a I'histoire des Mesozoaires : recherches sur I'organisation et le developpement embryonnaire des Ortho- nectides. Arch, de Biol. Tom. iii. 1882. 4. Keferstein. Beitrage zur Anatomie und Entwicklungsgeschichte einiger Seeplanarien von St. Malo. Abh. Gesell. Wiss. Gottivgen. Bd. xiv. 1868. 5. McIntosh, W. C. a Monograph of the British Annelids. Part I. The Nemerteans. London {Ray Society). 1874. 6. Metschnikoff, E. Untersuchungen iiber Orthonectiden. Ztitschr. wiss. Zooh Bd. XXXV. 1881. 7. Spengel, J. W. Die Orthonectiden. Biol. Centralbl. Bd. i. 1881— 1882. DiCYEMIDJ;. 1. Beneden, E. van. Recherches sur les Dicyemides survivants actuels d'un embranchement des Mesozoaires. Bnixelles. 1876. 2. Beneden, E. van. Contribution a I'histoire des Dicyemides. Arch. de Biol. Tom. iii. 1882. 3. Bkaun, M. Ueber Dicyemiden. Zusammenfassender Bericht. Centralbl. Bakt. u. Parantenkunde. Bd. ii. 1887. 4. Kollikee, a. v. Ueber Dicyema paradoxum, den Schmarotzer der Venenanhange der Cephalopoden. 2tes Bericht der Zool. Anstult in Wilrzburg. 1849. 5. Krohn, a. Ueber das Vorkommen von Entozoen in den Venenan- hangen der Cephalopoden. Froriep's " Neiie Notizen.'" Bd. xi. 6. Leuckart, R. Die Parasiten des Menschen. 2te Auflage. 1879 — . 7. Wagenee, G. Ueber Dicyema Kollikeri. Arch. Anat. u. Phys. Jahrg. 1857. 8. Whitman, C. O. A Contribution to the Embryology, Life-history, and Classification of the Dicyemids. Mitth. Zool. Stat. NeapeL Bd.' iv. 1883. CHAPTER VI. NEMERTINI. The Nemerteans laj their eggs enclosed in gelatinous enve- lopes, either singly or balled into large masses of spawn. It appears that fertilization may take place either outside or inside the body of the female. In the latter case the sperma- tozoa penetrate into the female genital organs (egg sacs) through their efferent ducts. In many forms (Monopora vivipara) the eggs are there developed up to the maturity of the embryo. The development takes place either directly or by means of a metamorphosis. The latter is of various kinds, according as a free-swimming larva differing very much from the ultimate shape of the animal is produced, or merely a larval form which does not differ essentially from the young animal, but which nevertheless produces the latter within it. In the first case, in view of the shape of the larva, one speaks of a Pilidium larva, in the latter case, of development after the type of Desor, thus named from its discoverer. I.— Development through the Pilidium Larva. As the result of the equal cleavage a regular blastula arises from the egg of Lineus lacteus. It loses its regular shape, owing to a considerable increase in the size of the cells of the lower half and to the occurrence at the same time of a flattening on the under-side of the blastula (Fig. 102 A). The outer and inner germ-layers can be distinguished on the blastula as early as this, for the cells of the ectoderm are smaller than those of the entoderm. The first trace of the middle germ-layer is likewise already present. In the cleav- age cavity and in contact with the entoderm are found a number of cells (Fig. 102 A) which to all appearances take their origin from the entoderm (Metschnikoff, No. 20), and 217 218 EMBRYOLOGY subsequently prove to be mesencbymatous mio^ratory cells (Fig. 102 B and C), like tbose which arise in the develop- ment of the type of Desor. After the blastula becomes covered with cilia, has assumed its characteristic shape, and has acquired a large flagellum at its apex (Fig. 102), it may break through the egg- mem- brane to swarm about at large. More often, however, the larva reaches the outside world only after invagination has taken place, i.e. as a gastrula. This is accomplished by the symmetrical invagination of the already-formed entoderm Fig. 102.— 4 to C, blastula, gastrula, and pilidium of Linens lacteus (after Metsch- jfiKOFF) ; C, combination of two of Metschnikoff's figures; s, ectodermal in- vaginations, which subsequently grow around the intestine as the prostomial and metastomial discs. (Fig. 102 5). The blastopore is circular, and the entire larva presents a radiaLform. This is soon changed, however, for the blastopore elongates somewhat and becomes oval, while the archenteron bends to one side, and its blind end grows more and more toward one wall (Fig. 102 5). In this way the form of the larva becomes bilaterally symmetrical. The larva assumes its permanent shape — i.e., the one which its discoverer, Joh. Muller, designated as pilidium— hj the NEMERTINI 219 downgrowth of a lobe on either side of the blastopore (Figs. 102 C and 103). It now consists therefore of an upper bell- shaped part, which we call the umbrella, and the two pen- dent lateral lobes. Between the two latter, on the under- side of the umbrella, lies the wide mouth-opening (Figs. 102 C and 103). It leads into the oesophagus, which corresponds to an ectodermal invagination, whereas the real entoderm is represented by the intestinal sac back of it (Fig. 102 C). The intestine of the larva, the cells of which are provided with cilia, remains closed. Like Turbellarian larvae, the pilidium is encircled by a continuous band of cilia, which fringes the periphery of the umbrella and the margins of the lateral lobes. The ciliation of the band is distinguished from that of the rest of the body by its longer cilia (Figs. 102 G and 103). The par- ticularly stout flagellum already mentioned takes its origin in a depression at the apex, corresponding to which there is a thickening of the ectoderm. The latter has been compared to the apical plate of the Trocliopliore larvae of the Annelida (comp. infra, p. 266). As in the annelid Trochopliore, two muscle strands, which also seem to contain nerve fibres, issue from the apical plate (Salensky, No. 25). The presence of these cords would not constitute, however, the only resem- blance to the annelid larva, but, according to SAiiExsKY, the ciliated band is also accompanied by a nerve cord, which would correspond to the ring-nerve in the ciliated zone of the Trochophore. This nerve cord, which is composed of nerve fibres and ganglionic cells, is said indeed to present a more varied histological differentiation than the ring-nerve of the an- nelid larva. At the point where the nerve cord passes from the lateral lobes to the umbrella, it forms ganglionic swellings, which Salensky interprets as the central organ of the nervous system. The inside of the larva, between ectoderm and entoderm, is filled with a gelatinous mass, in which the variously shaped mesenchymatous cells are found embedded (Fig. 102). These become at first the muscle-bands which traverse the larva at regular intervals ; subsequently they become in part the mesodermal elements (connective tissue, musculature, etc.) of the adult animal (Butschli, No. 2). 220 EMBRYOLOGY The pilidia of different Nemerteans differ from one another in shape as the typical form described above is more or less distinctly developed. In place of the flagellum, Pilidium gyrans bears a tuft of cilia at the apex (Fig. 103). In Pilidium auriculatum (Letjckaet und Pagenstecher) the two lateral lobes are only very slightly developed, and the Pilidium brachiatum described by E. B. Wilson, which resembles P. auriculatum, possesses, in addition to the two slightly developed lateral lobes, three Fig. 103.— Pilidium gyrans, with completely formed worm inside (combined from two of BuTSCHLi's figures). Am, amnion; D, intestine of the pilidium already sur- rounded by the worm ; Ec, ectoderm of the worm ; M, mouth of the pilidium ; N, fundament of the nervous system ; B, proboscis ; So, lateral organs. which have arisen by indentations of the edge of the additional ones, umbrella. The Pilidium recurvatum found by Fewkes (No. 5) at Newport exhibits a very aberrant form, which, by the absence of the lateral lobes, by the lateral curvature of the upper part, and by the presence of a row of cilia at the posterior end, acquires a striking resemblance to the Tornaria larva NEMERTINI 221 of nalanoglossus. Moreover, the metamorphosis of this larva is said by Fewkes to be accomplished in a manner different from that of other Pilidia. Whereas usually the larva remains intact even after the maturity of th worm, and in this condition is abandoned by it, in the present case the collapsed Pilidium, after the withdrawal of the Ne- mertean, is said to hang to its posterior end, where it is gradually resorbed, in the same way as the Pluteus larva is drawn into the body of the young sea-urchin. After Gegenbaur had expressed the view that possibly a new animal was developed within the Pilidium, this idea was more precisely defined by Krohn, who maintained that regularly a young Nemertean arises from the pilidium. Leuckart und Pagenstecher were able to raise this view to a certainty, for they (No. 17) followed the development of the Nemer- tean inside the pilidium. The accompanying processes were then fully elucidated by Metschnikoff (No. 19) and Butschli (No. 2). The formation of the IS'emertean in the pilidiam is initi- ated by the appearance of foar pit-like depressions of the ectoderm in the region of the mouth. Externally these pre- sent the appearance of round suckers, for w^hich at one time they were mistaken by JoH. Muller. As the depressions become deeper they become sac-like in shape (Fig. 102 C), and the wall directed toward the intestine of the larva is much thicker than the outer one. The further changes of the invaginations consist in their being constricted off from the ectoderm, becoming considerably expanded and growing around the intestine of the larva (Fig. 104 A and B). They have now assumed more of a discoid shape. At the points where they come together the discs fuse, and their thicker wall, the one directed inwards, constitutes the superficial layer of the body of the Nemertean, whereas their thin outer layer forms around the body an envelope, which is known as the amnion (Fig. 103 Am). This separates from its con- nection with the body of the worm, which it surrounds as a delicate membrane. The anterior pair of discs becomes the head of the Nemertean (as far back as the lateral grooves), whereas the posterior pair gives rise to the ectoderm of the rest of the body (Fig. 104 A and B). Consequently the an- terior discs, which, moreover, are the first to fuse, are known as the head- [^prostomial'] discs, the posterior as the trunk- [metastomiaV] discs. The union of the anterior with the 222 EMBRYOLOGY posterior pair does not take place until the components of each pair have completely united with each other. At the point where the two prostomial discs first come together an invagination is formed, the fundament of the proboscis, which soon grows backward a long distance (Fig. 104 A and ^-' '^-^ 1 So Fig. 104.— Diagrams to show the formation of the Nemertean (after Salensky). A, evaginatioiis of the oesophagus (considered by Hubkecht to be the fundaments of the nephridia) ; D, intestine ; If, mouth ; N, fundament of the nervous system ; B, proboscis; Es, sheath of the proboscis; S,, prostomial [head-] discs; S„, meta- stomial [trunk-] discs ; So, lateral organs. The position of the young worm in the pilidium is illus- trated by Fig. 103. The larval intestine is entirely included within the worm. Meanwhile the oesophagus continues to pass through the body-wall of the worm, still terminating in a wide opening, until at a later stage it fuses with the ecto- derm of the worm and is displaced to a position relatively further forward. * The lateral organs are said by Salensky and Hubrecht to arise in the same way as the somatic discs. They originate as two invaginations of the wall of the pilidium, one on either side of the oesophagus (Fig. 104 A, So), then grow out toward the somatic discs, and finally separate from NEMERTINI 223 their connection with the primary ectoderm of the pilidium, in order to fuse with that of the somatic discs (Fig. 104 B). Thus they are said to be formed directly as parts of the pilidium. The nervous system of the young worm makes its appearance in the form of two ectodermal thickenings (Fig. 104 N), which arise in the region of the anterior pair of discs on either side of the invagination of the proboscis. At this place the ectoderm cells are differentiated into several layers, of which the more superficial are said to become the body epithelium and the ganglionic cells, the deeper, the Punktsuhstanz . The anterior thickened parts of the fundaments correspond to the brain, and their backward prolongations to the lateral nerve-trunks (Fig. 104 A and B). According to this, the fundament of the central nervous system would have nothing to do with the apical plate of the larva. Even before the discs had separated from the ectoderm, mesenchyma cells were applied to their inner (deeper) layer ; and since such cells were also found in the region of the larval intestine, a considerable number of them came to be enclosed within the worm (Butschli, Salensky). Like the separate fundaments of the cephalic and somatic parts, the fundament of the mesoderm is double. In the first place, a mass of mesenchyma cells is formed on each of the two prostomial discs, and a similar one at the apex of the invagination of the proboscis. It could not be determined whether the latter originated from the former. Then each disc has its own mesenchyma layer, which likewise has arisen by an accumulation of mesenchyma cells. The anterior and posterior parts of the body are established, therefore, quite independently. The mesen- chyma of the trunk is said by Salensky to split into two layers, one of which is applied to the intestine as the splanchnic layer, the other to the body-wall as the somatic layer. A kind of coelom thus arises, which, to be sure, subsequently becomes reduced and breaks up into small cavities, owing to the cells of both layers sending out processes which unite with one another. In the head that part of the mesoderm which is applied to the prostomial discs becomes the musculature, whereas the layer in con- tact with the proboscis splits into two cell-layers, one of which is applied to the proboscis, while the other forms the sheath of the proboscis. Accordingly the cavity of the proboscis-sheath would be a portion of the coelom (Salensky). The proboscis and its sheath attain their subsequent great length by growing backwards (Fig. 104 B). [The results of a recent investigation by Bukgbr (Appendix to Literature on Nemertini) differ in several particulars from the account given above. The formation, and especially the differentiation, of the head- and trunk- discs, the formation of the head itself, and the development of the nerv- ous system are there described quite differently. The musculature of the dermo-muscu^ar sac appears to be of double origin, inasmuch as the outer layer of it arises from the ectoderm, but the remaining portion from the mesoderm. Similar statements are also made regarding the Annelids. — K.] 224 EMBRYOLOGY When the development of the worm has progressed as far as this, it breaks through the amnion and pilidium and swims about free in the water by means of its covering of cilia. At this stage it lacks the anus, which arises only later. In some cases eye-spots are present ; in others they are absent. II.— Development after the Type of Desor. For the more intimate knowledge of the mode of develop- ment in the type known as that of Desor, we are chiefly indebted to J. Barrois (No. 1). Recently Hubrecht (Nos. 9 to 11) has reinvestigated the subject. Here also, as in the development of the pilidium, an invagination gastrula arises, which at first is radial, subse- A Fig. 105. — A to C, formation of the somatic plates by in- vagination in Lineus ohscurus (after J. Bareois). Fig. 106.— Section of an em- bryo of Lineus ohscurus (after Hubrecht). D, intestine; if, mouth, which, however, like the oesophagus, is closed by cells ; Mes, mesenchyma cells ; S, discs which subsequently form the ectoderm of the worm. quently bilaterally symmetrical. According to Hubrecht, cells (the mesenchyma cells) are said to migrate into the blastocoele from both ectoderm and entoderm (Fig. 106). On the ventral surface of the ectoderm Barrois found a pair of invaginations in front of the mouth and another behind it. He saw that these invaginations were closed by the growth of the ectoderm over them, and that finally their floor became separated from the rest of the ectoderm (Fig. NEMERTINI 225 I 105 A to C). In this way there arose under the ectoderm four cell-plates, corresponding to the prostomial and metastomial discs of the pilidium, but distinguished from them by the fact that they are not composed of two layers, but of only one cell-layer (Fig. 106). When subsequently they grow around the embryonic intestine, there is thus formed only one cell-layer, the body-wall. The amnion is wanting (Fig. 107 B). The body, of course, is still surrounded by the larval skin, the original ectoderm. At the place where the Fig. 107. — A, gastrula stage of Linews obscurus, seen from the side ; B and C, older embryos of Lineus seen from the ventral surface (after Baeeois, from Bal- foue's Cord'parativa Embryology) ; ae, arclienteron ; cs, lateral organs ; Is, larval skin ; m, mouth ; me and ms, mesenchjma ; pr.d, prostomial disc ; po.d, meta- stomial disc; pr. proboscis ; st, stomach. prostomial discs come together, the proboscis arises as a solid ingrowth of the ectoderm, which subsequently becomes hollow (Fig. 107 0). HuBRECHT describes a fifth plate, derived from secondary ectoderm, in addition to the four plates found by Barrois (Fig. 106 S). It is said to be formed on the dorsal side of the embryo, but in a different manner K. H. E. Q 226 EMBRYOLOGT from the four ventral plates, namely by delamination. Hubrecht derives the epithelium of the young Nemertean from the fusion of these five plates. Hubrecht likewise differs from Barrois and Salensky in his descrip- tion of the mode of origin of the proboscis. The latter authors derive it from the secondary ectoderm, but according to Hubrecht it arises from the primary ectoderm and subsequently separates from this and fuses with the secondary ectoderm. When one considers that the lateral organs, according to the concordant statements of Hubrecht and Salen- sky, take their origin from the primary ectoderm, then such a mode of origin of the proboscis might perhaps be intelligible. Small argument, it is true, is to be got from the fact that in the pilidium the proboscis arises from the secondary ectoderm. Pilidium evidently represents the more primitive state, and on this account one must also expect in it the more primitive condition as regards the mode in which the organs origi- nate. As to the lateral plates, which are to be considered as sensory organs, it is easier to suppose that they were already present in the larva, whereas this is scarcely probable for the proboscis. The statement of Barrois that the mesenchyma cells are detached from the somatic discs seems to need revision. We have already become ac- quainted with the origin of these cells in the pilidium. In the cephalic part of the embryo they are applied in part to the proboscis (as its musculature), and in part they are arranged in the vicinity of it to form the proboscis- sheath. In this particular also Hubrecht differs from Salensky, for he considers the pocket of the proboscis to be the remains of the blastocale, whereas Salensky maintains that it arises (as a kind of coelom) by means of a splitting of a mesenchyma-layer. Hubrecht also looks upon the blood lacunae and the cavities of the vessels, the walls of which as well as the body musculature are of a mesenchymatous nature, as remnants of the blastocoele. He likewise derives the fundament of the nervous system from the mesenchymal, a conclusion which is wholly discredited by Salensky, since in the development of Pilidium this observer recognized the nervous system to be ectodermal in nature ; this, moreover, agrees with the ordinary mode of origin of this system of organs. On the other hand, Hubrecht is inclined to derive the genital organs, which make their appearance at an early period, from the ectoderm. While the proboscis and its sheath have grown considerably farther backwards, striking changes have taken place in the intestine. It consists, as in the pilidium, of a posterior wider part and an anterior narrower portion, although in this type even the latter is said to be of entodermal nature. The anterior part becomes solid as the result of cell-growth (Fig. 106), but subsequently is hollowed again, and its lumen NEM-ERTINI 1227 then communicates both with the lumen of the intestine and with the outer world. Therefore the permanent mouth still lies at the place of the blastopore. At about the time of the closure of the blastopore the ciliated embryo breaks through both the embryonal envelope, which is likewise ciliated, and the egg-membrane, to continue its development in the free condition. According to Hubrecht, the paired fundaments of two nephridia (?), which only later would come into connection with the outer world, arise as vesicular structures from the oesophagus, and consequently from the entoderm (for the oesophagus is said to be of entodermal nature). In the development from the pilidium these structures are found on the sftiterior intestine (Fig. 104 A), which here is of ectodermal nature.^ III. — Direct Development. I A transition from the indirect to the direct development is afforded by the Nemertean studied by DiecK: Cephalothrix galathece. Here a ciliated hlastula arises as the result of the tolerably regular cleavage. Dieck is inclined to look upon a wide cup-shaped depression, which makes its appearance on the blastula, as evidence of relationship with the pilidium form, for an extension of the edges of the depression would result in t-he lateral lobes of the pilidium. But a process which is accomplished later recalls far more the indirect mode of development of other Nemerteans than this outward shape of the embryo does. After the embryo has elongated and has assumed a rather worm-like shape, the layer of ciliated cells covering it begins to be cast off, and under it a new coat of cilia is immediately developed. Apparently here also, as in the type of Desor and in the pilidium, the new covering of the worm is formed under the larval skin ; a great simplifi- cation in the mode of development has, however, taken place. Special plates, which enlarge and unite to form the new body-covering, are no longer formed as the result of invaginations of the larval skin, but the body-covering is * [These organs have also been recently recognized by Burger (Ap- pendix to Literature on Nemertini) to be nephridia, and their origin is likewise referred by this author to the ectodermal anterior intestine.— K.] 228 EMBRYOLOGY split ofP directly from tlie larval skin. This process takes place on the free-swimming larva, for long before this the embryo had broken through the egg- membrane. Even at the time of its be- coming free, it exhibits at its anterior and posterior ends stout cilia (Fig. 108), which are likewise a reminiscence of the pilidium stage. Stout cilia, or tufts of cilia, arise at the ends of the body of the embryo even in those forms in which the development has become quite direct {AmphiporuSj Tetrastemma, Malacohdella) , and in which no other points in harmony with indirect development are still found, — apparently an indication that development by means of a ciliated free-swimming larva Fi«t. 108.— Embryo of is the more primitive mode, direct de- Cephaiothrix gaiathem yclopment, on the Contrary, the derived just hatched (after r ) j > DiBCKj. method. Cleavage and the production of the germ-layers in the forms developing directly do not appear always to take place in the manner which we have hitherto considered. In 3Ionopora vivipara, it is true, after an irregular cleavage there arises from the hlastula an invagination gastrula (Salen- sky) ; other forms, on the contrary {Amphiporus lactifloreus, Folia car- cinophila, Tetrastemma varicolor), are said to possess a delamination gastrula (Barrois, Hoff^iann). The sheet of long prismatic cells which forms the hlastula splits into an outer and an inner layer. The former corresponds to the ectoderm, whereas the latter again separates into a double layer, the outer one the mesoderm, the inner one the entoderm. In Tetrastemma the diJEferentiation of these cell-layers takes place in a solid mass of cells, a part of which remains at the centre and is employed only as food material. In Malacohdella also the germ is said to consist of a solid cell-mass, from which the ectoderm becomes detached. Indi- vidual cells migrate from the inner cell-mass into the cavity thus pro- duced, and constitute the middle germ-layer. The remaining cells correspond to the entoderm, and finally arrange themselves into an intestinal epithelium, which fuses with the ectoderm to form the mouth and anus. The embryonic development is now completed. The ciliated embryo reaches the outside world to develop directly into the Nemertean (Hoffmann). NEMERTINI 229 The origin of the different organs in the Nemerteans with direct deve- lopment has recently been studied by Salensky (No. 24) on Monopora. It corresponds essentially with what we have learned as occurring in the indirect process of development. At the anterior end the central nervous system arises in the form of two ectodermal thickenings, which soon become detached from their connection with the ectoderm. The funda- ments of the two brain ganglia grow backwards as the two lateral nerves. The proboscis and the oesophagus arise in the vicinity of the ganglionic fundaments, both of them as bud-like thickenings of the ectoderm and both of very similar appearance. The proboscis in this form lies, even in the adult animal, very close to the oesophagus, whereby the relation of the fundaments of the two organs is explained. The proboscis, which lies above the oesophagus, opens with it into a common atrium. In spite of this, however, relations [genetic] between oesophagus and proboscis should hardly be sought for here, ^s Hoffmann and Balfour supposed ; but the condition in Monopora is much more likely of a secondary nature. The proboscis, at first located at the anterior end of the body, was only subsequently [phylogenetically] united with the oesophagus by moving backwards. Moreover, the union is a very loose one, for the proboscis and oesophagus do not actually unite, but rather open independently of each other into the common atrium. In the further course of development the fundament of the oesophagus becomes hollow and unites with the intestine. The latter, after the closure of the blastopore, consists of a closed sac. Entodermal cells migrate into its lumen, thereby entirely filling it. Later they become arranged into an epithelium, and then the oesophagus also unites with the wall of the intestine- Afterwards the anus is formed. The proboscis in this case also is composed of ectoderm and mesoderm^ the latter giving rise to the envelope of the ectodermal invagination and to the proboscis-sheath. Salensky argues for a formation of the body- cavity by means of a splitting of the middle germ-layer in Monopora also. Even before the appearance of the body-cavity, two lateral bloodvessels and one dorsal are said to be formed, corresponding to the conditions of the adult animal. To all appearances they owe their origin to the inner part of the mesoderm, for they are located in the vicinity of the intestine. We find no statements in Salensky regarding the formation of the genital system. General Considerations. In conclusion, we must once more point out how closely the different modes of development of the Nemerteans ap- proach one another. In the piUdium the worm arises bj the formation of four vesicular ectodermal invaginations, which assume a discoid shape, and, growing around the 230 EMBRYOLOGY larva, unite to form the epidermis of the worm. Since the discs, owing to their mode of origin, are bilaminar, the body-covering of the worm, formed by the inner layer, is enclosed by an envelope (amnion), the outer layer of the discs. The larva itself goes to pieces. In the type of Besor, the discs, which likewise arise from the ectoderm, are from the beginning unilaminar only; the amnion, therefore, is absent, whereas in other respects the developmental pro- cesses are quite similar. Finally, the discs are no longer formed at all. However, as a suggestion of the former mode of development, the outer ectodermal layer separates from the embryo and is cast off (Cephalothrix). Moreover, the embryo bears rigid cilia, as in Pilidium. This is also the case in those embryos which are metamorphosed directly into the worm without having their ectodermal covering undergo any important changes. Accordingly the pilidium appears as the older type of development, from which the others are derived by becom- ing at the same time simplified. But even the development of the pilidium cannot be the original form. The origin of the worm within the larva is a secondary process, which has probably arisen through adaptation to the conditions of life. Originally the larva was certainly metamorphosed directly into the worm, as is still the case in the Turhellaria and Annelida^ for example. If the statements of Fewkes (No. 5) should be verified, those forms in which the pilidium is said to be resorbed by the body of the worm could best explain the original mode of development (comp. p. 220). The shape of the Nemertean larva, irrespective of the Tornana-like form (Fewkes), points to relationships of two kinds^ One of these con- cerns the Turhellaria. The resemblance is clear without further comment, if one considers the Stylochus-larva of Goette (Fig. 80, p. 168). This larva exhibits the two typical lateral lobes of ihe pilidium. We have shown in treating of the Turhellaria how it can be referred to MuLiiEK's larva. In the comparison of larval forms caution is of course necessary, and especially in the case uni,der consideration, where the first stages of development in the two groups differ very much from each other. Thus one might be inclined to maintain the similarity in the outward shape to be accidental, if the adult animals did not also possess many similar characters in their organization. NEMERTINI 231 The pilidium resembles the Trochophore of the Annelida (comp. p. 266), as well as the Turbellarian larvae. In common with this, it pos- sesses the apical plate, the cords radiating from it, and the ring-nerve which extends under the ciliary apparatus. The apical plate, to be sure, does not give rise here, as in the Annelida, to the oesophageal ganglion, for it is lost with the pilidium. For this reason, it does not seem allow- able to homologize the brain of the Nemerteans directly with that of the An7ielida. Apart from this, the lateral nerves of the Nemerteans arise by the growing out of the cerebral ganglion, which is already separated from its connection with the ectoderm, whereas the ventral ganglionic chain of the Annelida appears to take its origin by means of progressive differentiation of the ectoderm. The nervous system of the Nemerteans is most closely allied to that of the Platyhelminthes, and particularly to that of the Turbellaria, with which the Nemerteans also present other common features, such as the uniform ciliation of the entire surface of the body, the body parenchyma, and the lateral organs. But the shape of the proboscis appears to us to be of particular value for the comparison of the two groups. The proboscis, situated at the anterior end of the body, apparently having arisen by the metamorphosis of the latter into a tactile organ and being withdrawn into the inside of the body, presents in the two groups a position and structure too much alike not to challenge comparison. Other conditions separate the Nemerteans from the Turbellaria. The intestine possesses an anal opening, which is wanting in all Platyhel- minthes. The presence of a differentiated blood-vascular system indi- cates a higher organization of the Nemerteans. The genital organs are constructed quite differently from those of the Platyhelminthes, whereas those of the Turbellaria, Trematoda, and Cestoda show great agreement in structure. In the position of the sexual organs a segmental arrange- ment can be recognized. Whether the indications of segmentation furnished by the presence of the septa which constrict the intestine and the numerous openings of the water system have any higher meaning cannot be said in the present state of our knowledge. What we have hitherto learned in regard to the excretory system (v. Kennel and Oude- MANs) entitles us neither to recognize therein a higher degree of organi- zation nor to place the Nemerteans nearer to the Platyhelminthes, though the presence of two longitudinal vessels might point to the latter. On account of their numerous relations to the Turbellaria (although their organization is much higher than that of the latter), it does not seem justifiable to separate the Nemerteans from the Platyhelminthes altogether and to place them, as has already been done, nearer the segmented worms. The Nemerteans would have to be separated much more sharply from the Turbellaria, if the statements regarding the segmenta- tion of the body and the occurrence of a true body cavity were verified. Finally, we cannot leave unmentioned in this place a theory which places the Nemerteans in relation with the Vertebrata. Hubreoht, the 232 EMBRYOLOGY expounder of this view, compares to the central nervous system of the Vertebrata the dorsal nerve cord found by him. The cerebral ganglia of the Nemerteans are held to correspond to the series of ganglia of the cranial nerves of the Vertebrata, and the lateral nerves to the nervi laterales which occur so constantly in the latter group. In the chorda HuBRECHT sees the metamorphosed sheath of the proboscis, while the remnant of the proboscis itself would be recognized in the hypophysis. For this view Hubkecht finds support in the fact that in certain Nemer- teans the proboscis opens out in the vicinity of the oesophagus, and that in Tetrastemma it is said even to arise from the wall of the oesophagus (Hoffmann, No. 7). For the present these explanations have only the value of a mere hypothesis. Literature. 1. Barrois, J. M^moire sur I'embryogenie des Nemertes, Ann. Set. Nat. (ser, vi.) Zoologie Tom. vi. 1877. 2. BuTscHLi, 0. Einige Bemerkungen zur Metamorphose des Pilidium. Arch. Naturgesch. Jahrg. xxxix., Bd. i. 1873. 3. Desor, E. Embryologie von Nemertes. Arch. Anat. n. PTiys. Jahrg. 1848. 4. DiECK, G. Beitrage zur Entwicklungsgeschichte der Nemertinen. Jena. Zeitschr. Bd. viii. 1874. 5. Fewkes, J. W. On the Development of Certain Worm Larvae. Ball. Mus. Comp. Zobl. Harvard College, Cambridge, Mass. Vol. xi. 1883—1885. 6. GegenbauRj'C. Bemerkungen iiber Pilidium, Actinotrocha, und Appendicularia. Zeitschr. wiss. Zool. Bd. v. 1854. 7. Hoffmann, C. K. Beitrage zur Kenntniss der Nemertinen : I. Zur Entwicklungsgeschichte von Tetrastemma varicolor Oerst. Niederl. Arch. Zool. Bd. iii. 1876—1877. 8. Hoffmann, C. K. Zur Anatomie u. Ontogenie von Malacobdella. Niederl. Arch. Zool. Bd. iv. 1877—1878. 9. Hubrecht, a. a. W. Proeve eener ontwickelungsgeschiedenis van Lineus obscurus Barrois : Prys verhandeling met goud bekroond en uitgegeven door het provinciaal Utrechtsch genootschap van Kunsten en Wetenschapen. Utrecht. 1885. (Only the two following commimica- tions of this author, which treat of the same subject, were accessible to us.) 10. Hubrecht, A. A. W. Contributions to the Embryology of the Nemertea. Quart. Jour. Micr. Sci. Vol. xxvi. 1886. 11. Hubrecht, A. A. W. Zur Embryologie der Nemertinen. Zool. Anzeiger. Jahrg. viii. 1885. 12. Hubrecht, A. A. W. On the Ancestral Form of the Chordata. Quart. Jour. Micr. Sci. Vol. xxiii. 1888. NEMERTINI 233 13. HuBKECHT, A. A. W. Relations of the Nemertea to the Verte- brata. Quart. Jour. Micr. Sci. Vol. xxvii. 1887. 14. HuBRECHT, A. A. W. Eeport on the Nemertea collected by H.M.S. Challenger. " Challenger " Reports. Vol. xix. 1887. 15. Kennel, J. von. Beitrage zur Kenntniss der Nemertinen. Arh. Wiirzhurger Zool. Imtituts. Bd. iv. 1877—1878. 16. Krohn, a. Ueber Pilidium und Actinotrocha. Arch. Avat. «. Physiol. Jahrg. 1858. 17. Leuckart, R., und Pagenstecheb, A. Untersuchungen iiber niedere Seethiere. Arch. Anat. u. Physiol. Jahrg. 1858. 18. McIntosh, W. C. a Monograph of the British Annelids. Part I. The Nemerteans. London (Ray Society). 1874. 19. Metschnikoff, E. Studien iiber die Entwicklung der Echinoder- menund Nemertinen. 3Iem. Acad. St. Petersbourg (ser. 7). Tom. xiv. 1869. 20. Metschnikoff, E. Vergleichend-embryologische Studien. Ueber die Gastrula einiger Metazoen. Zeitschr. loiss. Zool. Bd. xxxvii. 1882. 21. MiJLLER, JoH. Fortsetzung des Berichts iiber einige Thierformen der Nordsee. Arch. Anat. u. Physiol. Jahrg. 1847. 22. MuLLER, JoH. Ueber verschiedene Pormen von Seethieren. Arch. Anat. u. Physiol. Jahrg. 1854. 23. OuDEMANS, A. C. The Circulatory and Nephridial Apparatus of the Nemertea. Quart. Jour. Micr. Sci. Vol. xxv. 1885. 24. Salensky, W. Recherches sur le d^veloppement de Monopora (Borlasia) vivipara Uljan. Arch, de Biol. Tom. v. 1884. 25. Salensky, W. Bau und Metamorphose des Pilidium. Zeitschr. wiss. Zool. Bd. xliii. 1886. 26. Wilson, E. B. On a New Form of Pilidium. Stud. Biol. Lab. Johns Hopkins Univ., Baltimore. Vol. ii. 1882. Appendix to Literature on Nemertini. BiJRGER, 0. Studien zu einer Revision der Entwicklungsgeschichte der Nemertinen. Ber, Naturf. Gesell. Freiburg i. Br. Bd. viii. 1894. CHAPTER VII. NEMATHELMINTHES. Systematic : I. Nematoda s. str. II. GORDIIDJ:. I. NEMATODA S. STR. Embryonic Development. The eggs of the Nematoda, which are usually oval, but occasionally spherical, are laid at very different times. Sometimes they are deposited very early, even before cleavage begins, and then are surrounded by a thick shell {Ascaris lumhricoides, Trichocephalus dispar), whereas thin- shelled eggs begin their development when still in the parent, and may even continue to develop here to a quite advanced stage. Still other Nematodes, as, for example, Trichina spiralis and some species of Ascaris, are viviparous. The embryonic development of a number of forms is known, though, it is to be regretted, not perfectly. As far as ascer- tained, the cleavage appears in general to be fairly alike in all cases. It is total and approximately equal, and leads to the formation of a blastula, which, to be sure, may be some- what variously shaped. It may have the form of a mere cluster of cells, designated by Goette as a sterroblastula (Bhabditis nigrovenosa), or it may be a true vesicle, with, however, only a very small cavity (Ascaris megalocephala), or, finally, it may appear in the form of a bilaminar plate of cells {Cucullanus elegans). At a very early period the fundaments of the germ-layers and the differentiation of the various regions of the body can be recognized on the segmenting egg (Goette, Hallez). As early as the first cleavage the 234 NEMATHELMINTHES 235 regg is divided into an ectodermal and an ento-mesodermal half. In Rhabditis nigrovenosa, according to Goette, the ventral and dorsal sides and the anterior and posterior ends of the embryo can be recognized even at this time. The ento-mesoderm divides first into two blastomeres. The ectodermal blastomere sends out a process dorsally over both of these (Fig. 109 A), and a newly formed ectodermal sphere is then situated at this point. In the further division of the ectoderm and ento-mesoderm the elements of the former push them- selves more and more over those of the latter, and thus as a whole come to lie more dorsally (Fig. 109 B). In subse- quent stages two cells lying close together at the former ectodermal pole of the egg indicate the tail-end of the embryo (Fig. 110 A, B), while the head-end lies oppo- site. Whereas Goette makes the separation of the mesoderm from the entoderm take place later, it occurs, according to Hallez (in Ascaris and Rhabditis aceti), even in the eight-cell stage, in which two" meso- derm cells are constricted off from two entoderm cells. In the twenty-four-cell stage, the blastula, with a small cleavage cavity, is formed, the dorsal part of which is composed of the ectodermal cells, the ventral of the entodermal and meso- dermal cells. Fig. 109.— -4 to D, cleavage stages and formation of the germ-layers in Rhabditis nigro- venosa (after Gobtte). ect, ec- toderm ; ent, entoderm; mes, mesoderm. Gastrulation takes place in vari- ous wajs, according to the form of the blastula. In Ascaris nnegalo- cephala an invagination gastrula is formed, the archenteron of which is very shallow, owing to the shape of the thick-walled blastula (Hallez). The process of gastrulation in Cucullanus elegans takes place in a peculiar manner, as was demonstrated by BtJTSCHLi. In this form the blastula stage consists, as has been mentioned, of a bilaminar cell-plate. This shape is soon lost, however, for the cells of one layer multiply more rapidly than those of the other, and therefore a bending toward the latter ensues. Finally a kind of tube is formed. 236 EMBRYOLOGY which presents an elongated fissure on one side. This con- stitutes the blastopore of this peculiar gastrula. Also in the forms observed by Hallez and in Rhahditis nigrovenosa the blastopore exists in the form of a long slit (Fig. 110 B, hi). In the last-mentioned Nematode the gastrula arises by the more active proliferation of the ectoderm cells, which produce an epibolic overgrowth, embracing the ento-meso- derm (Fig. 109 B, D), whereby a long slit, the blastopore, persists on the ventral side (Fig. 110 B). Subsequently this closes gradually from behind forwards. A transition between Fig. 110. — A to E, different stages of development of Bhahditis nigrovenosa (after Gokttk). hi, blastopore ; d, intestinal canal ; ent, entoderm ; g, fundament of the genitalia; m, mouth; mcs, mesoderm; n, fundament of the nervous system. gastrulation by invagination and by epiboly exists, according to Hallez, in Oxysoma. As regards certain facts of the subsequent embryonic development which are not yet wholly clear, we have to abide chiefly by the statements of Goette for Bhahditis nigrovenosa. According to him, when the circumcrescence is already far advanced, the formation of the mesoderm takes place by the squeezing out of two cells from their connection with the ento-mesoderm at the posterior end of the embryo (Fig. 109 D). A comparison of these two cells with the NBMATHELMINTHES 237 primitive mesoderm cells [teloblasts] of the Annelida is suggested, especially since in multiplying they extend toward the anterior end, and then constitute two rows of cells lying by the side of the entoderm, resembling the mesodermal bands of the Annelida (Fig. 110 A and G). Their subse- quent development, however, is not the same as in that group, for single cells afterwards separate from them and take np various positions between the intestine and the body-wall, without giving rise to a cojlom homologous to that of the Annelida (comp. p. 268). The embryo, which up to this time has possessed an oval form, changes in shape, for it becomes curved toward the ventral side (Fig. 110 D) and more elongated. The shape of the entoderm should be considered in connection with this. At first it forms two layers of cells, between which only a narrow lumen exists (Fig. 110 A and C). The latter soon disappears in the posterior part of the embryo, and the cells now arrange themselves one after the other in a row (Fig. 110 D). The lumen is retained only in the anterior portion; there is formed here a depression of the ectoderm, the fundament of the fore-gut, which unites with the en- toderm (Fig. 110 D and E). The mouth lies in the same place where the last trace of the slit-like blastopore, which closed from behind forwards, was visible. Later a lumen is again formed in the remaining part of the intestine by the splitting of the entoderm (Fig. 110 -E7). The entoderm cells at the posterior end, according to the statements of Goette and Hallez, fuse with the ectoderm to form the anus, with- out any depression of the ectoderm being noticeable, whereas Strubell (No. 10) maintains the existence of such a depres- sion. The central nervous system arises by a thickening of the ectoderm in the region of the mouth (Fig. 110 0 and Z), n) ; the dorsal and ventral parts of the oesophageal ring are said to sever their connection with the ectoderm earlier than do its lateral parts (Ganin). The ventral longitudinal nerve appears to arise from a paired fundament, a condition which has led to a comparison with the ventral longitudinal nerve cords of the Platyhelminthes. In pursuing this idea, there has also been an in3Unation to refer the dorsal lonjritudinal 238 EMBRYOLOGY cord to the dorsal nerves of tlie Platjhelmintliee, and to compare the lateral nerves of the latter to the two nerves of the lateral lines in Nematoda. It must be noted, however, that the facts actually established offer no certainty that this comparison is justified. More uncertain still are the observations on the further changes of the mesoderm. The mesoderm cells multiply greatly, separate from the two cell- rows, and migrate in various directions. They also penetrate between the fundaments of the nervous system and the skin, separating these from each other (Fig. 110^, mes). Finally, the mesoderm forms a rather even layer between the in- testine and the epidermis, so that the originally bilateral arrangement thus disappears. It would be important to know more accurately about the formation of the body cavity in the Nematoda. The origin of the sexual organs, which in the early stages is the same for both sexes, is better known. In each of the mesodermal bands, which at first consists of only a few cells, one of these cells is distinguished by its remarkable size (Fig. 110 D and E, g). It constitutes the fundament of the genital organs. In Bhahditis a cord of cells is developed from it by division, the individual elements of which divide further, and finally become the sexual products (Goette). In other Nematoda the original cell multiplies, it is true, but the protoplasmic bodies of the newly formed cells do not separate from one another ; on the contrary, a syncytium with many nuclei is formed. The sexual fundament, which is at first saccular, grows and differentiates into germ glands and ducts. While in the former the protoplasmic mass with the nuclei persists as the germarium, in the latter a peripheral epithelium is formed (Ant. Schneider). The shape of the ripe embryo resembles on the whole that of the Nematode, although it still has to undergo, especially in the parasitic worms, many changes before it attains the adult form. Several moultings are often necessary for this. In some cases the embryo possesses provisional organs, which appear to be adaptations to its mode of development. Thus in ISpiroptera obtusa and Cucullanus elegans a boring-tooth is found at the edge of the mouth, and the posterior end of NEMATHELMINTHES 239 the larva of the last-named worm is prolonged into the form of an awl, whereas the adult worm possesses a conspicuously blunt posterior portion. Post-embryonic Development. The post-embryonic development in the parasitic Nematoda is very diverse. It appears to be simplest in those cases where the eggs of the Nematode reach the outside world from the place where the parasite lives— for example, from the intestine of the host with its faeces — and then are taken up by another animal of the same species along with its food. The eggs may be more or less developed at the differ- ent stages of this migration, but in any event their envelopes are first destroyed in the intestinal canal of the new host, and the embryo here finds at once the conditions of life suit- able to it. Leuckaet has observed such a direct conveyance of the eggs into the intestine of the host in Trichocephalus affinis and Heterakis vesicularis. The conditions are somewhat less simple when the eggs are enveloped by only a thin shell, from which the embryos hatch as larvae. These then live and take food in damp earth or water, like those Nematodes which always lead a free existence. In general they resemble the members of the genus Bhabditis so closely that they are not distinguish- able from them (Leuckart). During its free existence the worm attains a certain size and development. Only when it arrives in its host do the organs needed for a free exist- ence degenerate ; it now adapts itself to the parasitic mode of life. Such is the case, for example, in Bochmius tri- gonocephalus and D. duodenalis. The RhabdUis-like larvae of these worms undergo several moultings during their free existence, are then swallowed by the dog along with its drinking-water or by man, and, as the result of a gradual metamorphosis, acquire the sharp mouth armature which is peculiar to them in the adult condition. The process of development is somewhat different in the Mermithidce, which are found as sexually immature forms in the larvae of insects. After prolonged periods of residence, they abandon this place of habitation by breaking through the body- wall, 240 EMBRYOLOGY and then remain in the damp earth. Here they moult and metamorphose into the sexually mature animals. These copulate, deposit their eggs in the earth, and the young worms developing from them then migrate into insect larvae again, that of Mermis albicans, for example, into young caterpillars. The mode of development just described for many Nematodes, in which the worms pass through a Rhabditis stage, may well be regarded as most nearly resembling that form in which parasitism in the Nematoda originated ; that is to say, a more or less f ally developed worm resorted to the body of another animal, or at first only became attached to it in order to gain nourishment from its juices. The parasitism only gradually became permanent ; it is precisely the Nematodes that offer all transitions from a partial to a complete parasitic life, which eventually leads to a total transformation of the form of the body. Such a metamor- phosis of the most extreme kind is realized in Sphserularia hombi, which was first investigated by Ant. Schneider and more recently in detail by Leuckart (No. 7). This worm in the adult condition consists of a thick warty sac, which lies in the body cavity of female humble-bees. To it is attached a diminutive worm, which can be recognized as a Nematode only upon careful examination. The entire sac owes its origin to the fact that the vagina of the worm became everted, and, growing to an enormous size, included the other sexual organs in it. The entire animal now con- sists, with the exception of the small attached worm, simply of a sac filled with sexual products. In it the eggs develop. The young worms reach the body cavity of the humble-bee, and from there the outside world, where they attain to sexual maturity. They copulate in the free condition, and probably the fertilized females migrate into the humble-bees when these seek their winter quarters in the ground. Then the remarkable transformation of the female begins, r A transition to Sphserularia is represented by Atractonema gibbosujn, discovered by Leuckart, in which an e version of the vagina likewise takes place, although it attains no greater size than about that of the worm itself. It is NEMATHELMINTHES 241 appended to tlie worm as a hernia-like body. The intestine of the latter degenerates, so that here also nutrition must take place bj endosmosis. The course of development of Atractonema resembles that of Sphasrularia. The eggs arrive in the body cavity of the host, and the young worms from here repair to the outside world, where they develop into sexually mature animals and copulate. The fertilized females penetrate into the larvae of a gall-fly, Cecidomya jpini, where they undergo their further development. To the developmental processes of the forms last con- sidered can be added that of the Beet Nematode, Heterodera Schachtii. Swellings containing a spherical worm filled with eggs, which can be recognized as a Nematode by its develop- ment, are often found on the lateral roots of the sugar-beet. The eggs of the worm develop within it, and pass into a slimy brood-sac, secreted by its genital ducts and attached to its posterior end. From here the larva passes to the out- side world, and undei'goes a development which differs somewhat according as a male or a female arises from it. The female, which is provided with a stiletto-like structure on the pharynx, bores into a beet-root,'moults here, and sucks up such a large amount of nourishment that it becomes swollen into a plump body, and thereby causes the epidermis of the root to burst. In this way the hind end of the female is exposed, and it is probable that copulation does not take place until this time (Strubell). Probably the most profound metamorphosis undergone by any Nematode is exhibited by Allantonema mirabile, likewise discovered by Leuckart (No. 7), a worm of a sausage-like, stubby shape, which lives in the body cavity of one of the Curculionidae (Eylohius pini). Except for the form of the sexual apparatus and its products, resemblance to a Nematode could be recognized neither from the external nor internal conditions of this intestineless structure. This worm is said to be hermaphroditic, and it is maintained that self-ferti- lization takes place. The eggs are developed in the uterus into young Nematodes, which are set free in the body cavity, and subsequently reach the outside world through the intes- tine. For a considerable time the larv» lead a free life, for K. H. E. R 242 EMBRYOLOGY which their organization fits them. Thej develop into males and females, which copulate and lay fertilized eggs. These develop in the free condition, and a generation of Rhabditis- like Nematodes arises from them. The latter most likely migrate into the larvae of the weevil, and here change into the Allantonema form described above. Here therefore the process of development is further complicated by embracing two differently formed generations, of which one is free throughout life ; the other, however, leads in part a parasitic life. This condition, long known as heterogeny, corresponds to the mode of development of llhahditis nigrovenosa, only that in the latter case no such fundamental metamorphosis of the parasitic generation takes place. The hermaphroditic form, ordinarily known as Ascaris nigrovenosa, inhabits the lung of the frog. It produces eggs, the development of which we have described above. The eggs are developed in the parent, from which the embryos emerge in the lung of the frog. From there the embryos pass into the intestine and then out with the faeces, and then develop into males and females, the true Rhabditis form. After copulation, there are developed within the female a small number of young, which abandon its body after they have been nourished by its contents. These young worms likewise exhibit the Rhahditis form, and do not lose it until they have migrated into the lung of the frog, where they are metamorphosed into the hermaphroditic form. The course of development in Rhabdonema strongyloides is, according to the discovery of Leuckart, also similar; the hermaphroditic form, hitherto known as Anguillula intestinalis, inhabits the intestine of man, whereas the dioecious Rhabditis form {Rhab- ditis stercoralis) is found in a free condition. Those forms also which, in order to reach their complete development, must live parasitically in two different hosts, show a very high degree of adaptation to a parasitic mode of life. This applies, for example, to Cucullanus elegans, which is found in the intestine of the perch. The young of this viviparous Nematode pass from the intestine of the host into the water, where they may live free for several weeks, until they meet with some suitable host. This is not the perch, NEMATHELMINTHES 243 'as one miglit imagine ; but the worms migrate into a Cyclops by first penetrating into the intestine through the month and then into the body cavity of the Crustacean. Here they undergo several changes in form, but attain their permanent shape only after the Cyclops which harbors them is swallowed by a perch, and they are set free in its intestine, where they soon become sexually mature, and in turn bring forth young, which undergo the same course of development. Dracunculus medinensis, a Nematode parasitic in the human body, appears to have a quite similar mode of development. Dracunculus inhabits the subdermal connect- ive tissue, and by its pressure against the skin causes a tumor and finally an abscess, through which it is able to pass out. In this way also the embryos, which are present in the worm in countless numbers, may reach the outside world. During the bathing of persons afflicted with the disease they get iato the water, and, like the larvae of Gucullanus, migrate into Cyclopidse ; however, they penetrate directly through the body-covering to the interior of the host. These infected Cyclopidae are probably swallowed along with the drinking-water by the inhabitants of those regions where the parasite abounds. Sjpiroptera obtusa has a development very similar to that of the two forms last considered, only it is still more adapted to parasitic life, for the eggs of this worm do not develop into a free organism, but are directly received by an intermediate host. Spiropfera obtusa inhabits the intestine of the mouse. The eggs, in which the embryos are already developed, reach the outside world with the fa3ces. They are swallowed by the larvas of the mealworm, Tenebrio, which feed on the dung balls. The embryos hatch out in their stomach, break through the wall of the intestine, and become encysted in the fat-bodies of the mealworm. When a mouse devours a mealworm, it becomes infected with the Spiroptera, which wakes up to a new life in the intestine of its host, becomes sexually mature and reproductive. The course of development in Trichina spiralis is one of those which are most completely adapted to parasitic life, for this Nematode accomplishes its entire life-history within the 244 EMBRYOLOGY bodies of two hosts. Its development differs from that of other Nematodes in that the young born of the sexually mature females in the intestine of the host do not reach the outside world, but break through the walls of the intestine and migrate into the muscles of various parts of the body of the host, to become, after sufficient growth, encysted. In order to wake the muscle Trichina into new life and bring about its sexual maturity, it is necessary that the infected muscle be consumed by some other animal, in whose intes- tine the Trichinae then attain their complete development and power of reproduction (Leuckart). II. GORDIID>E. The accounts of the development of the Gordiidse are still superficial. The eggs are not deposited singly, but are united into large balls or strings ; for during oyiposition a tenacious mass is poured over the eggs, which are already surrounded by a shell. The mass hardens in the water. Since the egg- strings are heavier than water, they sink to the bottom and remain thereuntil their embryonic development is completed. This does not begin until after the laying of the eggs, and requires quite a long time, according to Meissner about a month or more. As regards the first stages of development, the statements of the authors (Villot, No. 16, and Camerano, No. 11) do not agree. According to Camerano, the cleavage is nuequal, and leads to the formation of a bilaminar cell- plate, which, by the bending in of the edges, is metamor- phosed into a gastrula with a long slit-like blastopore, in a manner similar to that described above (p. 235) for Cucul- lanus. The gastrula closes, exactly as in Cucullanus, from behind forwards. The observations of Camerano on Gordius Villoti extend up to this stage, and it appears as if the figures given by Villot for Gordius aquaticus could be ap- plied to Camerano's observations. Villot describes the cleavage as regular. A solid heap of cells arises which, after further multiplication of the cells, splits into a central cell-mass and a peripheral layer (Villot). The hitherto spherical embryo elongates somewhat, and a deep depression Fig. 111.— 4 and S, two larvge of Gordius subhifurcus, with retracted and with everted proboscis (after Mkissitkb). now arises at one end. In this the head of the embryo is formed in such a way that it can subsequently be everted. The head is composed of a thicker basal portion and a slender k proboscis. The former bears three circles of six hooks each, the latter three long stylets, so that the embryo appears well armed. 'At the time of hatching, the head, with its armature, is everted (Fig. Ill B), but can at any time be retracted as before (Fig. Ill A). In the mean- time the intestine is formed, and ^ leads from the mouth, at the tip of the proboscis, to the anus, situated somewhat in front of the posterior end. The efferent duct of a remark- ably extensive gland opens into the oesophagus, at the base of the pro- boscis. Externally the embryo pre- sents an annulated appearance (Fig. 111). The embryo after hatching- lives as a larva for a long time free in the water, and then, with the aid of its sharp armature, penetrates through the skin into the interior of Chironomus larvae, as was observed by Yillot. This observer regards as simply an exceptional case Meissner's statement (No. 13) that the larvae of Ephemeridas are also infected with Gordius larvae. The parasite becomes surrounded by an envelope formed from the tissue of the Chironomus larva, and re- mains there nntil the larva happens to be swallowed by a fish (ViLLOT, No. 16). Becoming free in its intestine, the Grordius larva perforates the intestinal wall and again be- comes encysted. It remains here for a long time without I undergoing essential change. Finally (at the beginning of spring) it returns to the intestine, which it leaves with the faeces, and then gradually assumes the form of the adult worm, during which the cephalic armature is lost, the an- nulation of the surface of the body is smoothed out, and the sexual organs are developed. At the same time its intestine suffers a partial degeneration, and the mouth becomes closed. However, it appears to be by no means certain whether I the mode of development described is the one realized, or 246 EMBRYOLOGY whether the life even in a single host may not prepare the Gorclius for further development (Villot, No. 17). In addition to fishes, the Gordius larvae may also get into frogs, insects, spiders, and Crustacea, although, according to Villot, fishes are their most common hosts. As is well known, the Gordiidee are often found also in terrestrial in- sects, for example beetles and grasshoppers ; but nothing is accurately known concerning the conditions of infection in these animals. In insects of prey it can be explained as the result of swallowing infected insect larvae. The extraordi- nary size and development of the Gordiidae in such terres- trial animals is explicable by the fact that they so long lacked the opportunity of reaching the water, the place of their final development. General Considerations. The systematic position of Gordius must be briefly considered here. Vejdovsky has recently reverted to the idea, prevalent in former times, that the Gordiidae present much closer relationships to the Annelida than to the Nematoda, and perhaps are even to be looked upon as degenerate Annelids (Nos. 14 and 15). The segment-like arrangement of the ova- ries, and more especially the structural conditions of the body cavity, give rise to this conception. The latter is said, according to Vejdovsky, to be bounded, on its somatic wall at least, by a well-marked epithelium, and the intestine, as well as the genital organs, is said to be united to the body-wall by means of mesenteries. Villot denies the existence of the mesenteries, and refers the epithelium seen by Vejdovsky to a kind of mesenchymatous tissue, which in young stages of development fills up a large part of the space between the intestine and the body-wall. There- fore the body cavity of the Gordiidae, like that of the Nematoda, is bounded on one side by the musculature of the body-wall, which arises from that [mesenchymatic] tissue, and on the other by the entodermal wall of the intestine itself. On the latter Vejdovsky even could recognize no lining epithelium, a condition which he explained, however, by the great reduction of the intestine. However, v. Linstow has also denied recently the presence of an epithelium lining the body-wall (No. 12), and Came- RANo, in consequence of the early stages of development of the Gordiidce observed by him, argues for their relationships to the Nemathelminthes (No. 11). Nevertheless the Gordiidte are distinguished from the Nematoda by the peculiar condition of their genital organs and their deviations from them in other systems of organs, especially the nerve-ring, which is pro- longed into the ventral cord, and they can be co-ordinated with them as a separate division. What was said in considering the Nematoda must NEMATHELMINTHES 247 be repeated here, namely, that for the determination of the position of this division a better knowledge of the mode of origin and metamorphosis of their middle germ-layer is most desirable. Embryology gives no satisfactory basis for the determination of the position of the Nemathelminthes in the system, and it is hardly possible, in the present state of our knowledge, to decide this. It cannot be deter- mined whether on the one side they have relationships to the Platyhel- minthes or to the Nemertini, and whether on the other to the Annelida. It appears to follow from the structure of the adult animal that there exist resemblances to the organization of the Echinoderes and Gastro- tricha. But the latter are unquestionably related to the Eotatoria, so that in this way relations of the Nemathelminthes to the Trochophore would be brought about (comp. p. 259). More obscure still than the position of the Nemathelminthes is that of the Acanthocephali. We have not, as is customary, considered these along with the Nemathelminthes ; it is for the reason that their associa- tion with this group rests upon grounds of purely external nature, and is required neither by their organization nor by the manner of their develop- ment. Even the musculature, which has ordinarily been made use of for the comparison of the two groups, does not appear to be the same in the Nemathelminthes and Acanthocephali either in its arrangement or structure (Safftigen, Koehler [Literature on Acanthocephali, Nos. 6 and 31). Literature. 1. BuTscHLi, 0. Zur Entwicklungsgeschichte des Cucullanus ele- gans. Zeitschr. wiss. Zool. Bd. xxvi. 1876. 2. Ganin, M. Ueber die Embryonalentwicklung von Pelodera teres. Zeitschr. tciss. Zoul. Bd. xxviii. 1878 (Abstract). 3. GoETTE, A. Untersuchungen zur Entwicklungsgeschichte der Wiirmer. II. Rhabditis nigrovenosa. Leipzig. 1882. 4. Hallez, p. Recherches sur I'embryogenie de quelques Nematodes. Fans. 1885. 5. Leuckart, R. Die Parasiten des Menschen. Erste Auflage. Leip- zig. 1876. 6. Leuckart, R. Ueber Trichina spiralis. Leipzig. 1866. 7. Leuckart, R. Neue Beitrage zur Kenntniss des Baues und der Lebensgesehichte der Nematoden. Abh. Kgl. sacks. Akad. JViss. Bd. xiii. 1887. 8. Schneider, Ant. Monographie der Nematoden. Berlin. 1866. 9. SiEBOLD, T. V. Beitrage zur Naturgeschichte der Mermithen. Zeitschr. wiss. Zool. Bd. v. 1854. 10. Strubell, a. Untersuchungen iiber den Bau und die Entwicklung des Rubennematoden, Heterodera Schachtii. Bibliotlieca Zoo- lo(jica (Leuckart u. Chun). Bd. i., H. ii. 1888. 248 EMBRYOLOGY 11. Camerano, L. I primi momenti dell' evoluzione del Gordii. Mem. Real. Accad. Sci. Torino. Ser. ii. Tom. xl. 1889. 12. LixsTow, 0. V. Ueber die Entwicklungsgeschichte und die Ana- tomie von Gordius tolosanus. Arch. mikr. Anat. Bd. xxxiv. 1889. 13. Meissner, G. Beitrage zur Anatomic und Physiologic der Gordia- eccn. Zeitschr. tvias. Zool. Bd. vii. 1856. 14. Vejdovsky, F. Zur Morphologic der Gordiiden. Zeitschr. wiss. Zool. Bd. xliii. 1886. 15. Vejdovsky, F. Studien iiber Gordiiden. Zeitschr. wiss. Zool. Bd. xlvi. 1888. 16. ViLiiOT, A. Monographic des Dragonneaux. Arch. Zool. exp. et gen. Tom. iii. 1874. 17. ViLLOT, A. Nouvellcs rcchcrches sur I'organisation et le developpe- ment des Gordiens. Anji. Sci. Nat. Ser. 6. Tom. xi. 1881. 18. ViLLOT, A. Sur I'anatomic des Gordiens. Ann. Sci. Nat. Ser. 7. Tom. ii. 1887. 19. Vlllot, a. Sur le developpement et la determination specifique des Gordiens vivants a I'etat libre. Zool.Anzeiger. Jahrg.x. 1887. Appendix to Literature on Nemathelminthes. I. BoTERi, T. Uebcr die Entstehung des Gegensatzes zwischen den Geschlechtszellen u. den somatischen Zellen bei Ascaris megalo- cephala, nebst Bemerkungen zur Entwicklungsgeschichte der Nematoden. Sitzungsber. Gesell. Morph. u. Phys. Miinchen. Bd. viii. 1892. II. Camerano, L. I primi momenti dell' evoluzione dei Gordii. Boll. Mm. Zool. Torino. Tom. iv. 1889. III. Cobb, N. A. Beitrage zur Anatomic und Ontogenie der Nematoden. Jena. Zeitschr. Bd. xxiii. 1889. IV. Ha>iann, 0. Zur Entstehung der Excrctionsorganc der Seitenlinien und der Leibeshohle bei den Nematoden. Centralblatt Bakt. u. Parasitenkunde. Bd. xi. 1892. V. LixsTow, 0. V. Ueber die Entwicklungsgeschichte von Gordius tolosanus Duj. Centralblatt Bakt. u. Parasitenkunde. Bd. ix. 1891. VI. Vejdovsky, F. Studien zur Organogenic der Gordiiden. Zeitschr. wiss. Zool. Bd. Ivii. 1894. VII. ViLLOT, A. L'evolution des Gordiens. Ann. Sci. Nat. Ser. 7. Tom. xi. 1891. VIII. Wandollek, B. Zur Embryonalentwicklung des Strongylus para- doxus. Arch. Naturg. Jahrg. Iviii., Bd. i. 1891. IX. Zur Strassen, 0. Bradynema rigidum. Zeitschr. wiss. Zool. Bd. liv. 1892. CHAPTER yiir. ACANTHOCEPHALI. The eggs of the Acanthocephali are detached from the ovarium as membraneless, usually spindle-shaped cells, and then come to lie in the interior of the body of the female. Here they are fertilized, after which each egg is surrounded by a delicate transparent membrane, and then begins to cleave. When this (in Echinorhynchus gigas) has advanced as far as the formation of twelve blastomeres, a second mem- brane is formed under the first, which has separated some distance from the egg, and to which are added in the course of the development two more protective envelopes, so that finally foar of them are present. This applies to Echino- rhynchus gigas (Fig. 113 A). Ordi- narily three such embryonal mem- branes are formed, the middle one of which acquires a considerable thick- ness and firmness by the deposition of concretions of a brownish colour. These structures are particularly noteworthy, for the reason that they first make their appearance during cleavage, and therefore are not to be looked upon as egg-membranes, but as a kind of embryonal membrane ; still they do not appear to have any cellular structure. They recall the embryonal membranes occurring in the Taeniadae, which may also acquire a considerable firmness. During the formation of the embryonal membranes the cleavage has continued.^ It is unequal, and, according to ^ In this connection we follow, in addition to the older observations of Greeff and especially Leuckart, the newer investigations of Kaiser on 249 ih-'H- Fig. 112. — ^ to D, four cleavage stages of Echinor- hynchus proteus (after Leuc- kart) ; eh, first embryonal membrane. 250 EMBRYOLOGY Leuckart, takes place in Echinorhynchus proteus and angus- tatus in such a manner that the spindle-shaped e^g is divided at right angles to its long axis into a number of cells which are not quite equal in size (Fig. 112 A, B). After five blastomeres have been formed in this way, they are divided in the direction of the long axis, and a rather irregular ar- rangement of the cleavage spheres ensues (Fig. 112 C, D). As the result of cleavage an epibolic gastrula is pro- duced (Kaiser), the outer layer of which is formed of a large number of polyhedral cells, whereas the inner layer consists of much larger cells and encloses a remnant of yolk in the centre. Even at this stage the embryo acquires its armature. In the middle of every group of four contiguous ectoderm cells is formed, as the product of their secretion, a recurved booklet which protrudes into the space bounded by the em- bryo and the innermost protective envelope. The entire surface of an embiyo of Echinorhynchus gig as is beset with small spines, and in ad- dition five larger hooks are found at the anterior end (Fig. 113 A). The anterior end of the body, on which they are located, can be retracted, forming a funnel-shaped depression. In Echinorhynchus angustatus it is truncated, and five to six hooks are always found on the disc thus formed (Fig. 113 B). As in Echino- rhynchus gigas, it also can be drawn in. After the central yolk is entirely consumed, there begins a process called by Kaiser histolysis. This consists in the following changes: the boundaries between the cells dis- appear, the bodies of the cells flow together, and the cell nuclei move to the middle of the embryonic body, where Echinorhynchus gigas, which, however, have as yet been published only in a preliminary way and without illustrations, but which nevertheless afford an insight into the development of these forms. Fig. 113.—^, embryo of Echinorhynchus gigas in the embryonal membranes (eifi) ; B, larva of Echinorhynchus angustatus with the disc (s) at the anterior end bearing the armature (after Leuc- kart) ; efc, •' embryonic nu- cleus." ACANTHOCEPHALI 251 jcumulate to form the structure called by Leuckakt embryonic nucleus or core. Moreover, layers of two kinds can still be distinguished in the syncytium : an outer tough one and a less firm inner one, which encloses the em- bryonic nucleus. Leuckart had already shown that later the greater part of the worm arises from this central portion of the embryo. Furthermore, he compared it to a rudi- mentary intestine, and showed how the solid body Avhich he found lying between the cephalic disc and the embryonic nucleus (Fig. 11.3 B) could be interpreted as a rudimentary pharynx. This conception seemed satisfactory in view of the relationship of the intestineless Acanthocephali to other worms (Nematoda). In the condition above described the embryos, enclosed in their firm envelopes, pass out by means of the complicated mechanism of the sexual conductive apparatus. They now find themselves in the intestinal canal of the host — a fish in the case of Echinorhynchus angustatus and proteuSj the hog in the case of Echinorhynchus gigas — and then reach the outside world with the faeces of the animal. The em- bryos of the latter species are swallowed by the larvae of Cetonia aurata along with their food, whereas those of the two worms first mentioned are swallowed in the same way by Asellus aquaticus and Gammarus puhx. The embryonic envelopes soften in the stomach of the new host, and the em- bryo becomes free ; it immediately penetrates into the intes- tinal wall, and comes to rest either here (Echinorhynchus gigas and angustatus), or in the body cavity of the host. The larva of Echinorhynchus angustatus also reaches the body cavity later, but in a more passive manner as the result of its active growth and the rupturing of the intestinal mus- culature. Here (in Gammarus puJex) are also found the young stages of Echinorhynchus polymorphus, which as an adult worm inhabits the intestine of ducks and other aquatic birds (Greeff). The further development of the larva is connected with a metamorphosis of the external shape of the body, due to the formative processes which take place within. In Echi- norhynchus gigas the middle part of the body swells greatly 252 EMBRYOLOGV as soon as the larva has located itself in the intestinal mus- culature of the intermediate host. The developmental pro- cesses proceed from the central part, the so-called embryonic nucleus, for it is this which contains the formative material. According to the observations of Leuckart, it is differenti- ated into four groups of cells lying one behind the other (Fig. 114 A). The hindermost of these four groups soon acquires a greater volume, sending out a peripheral layer, which spreads out in front and on the sides, and encloses the other groups, with the exception of the most anterior one (Fig. 114 A). The Echinorhynchus is formed for the most part out of these cell groups. The most anterior is said to become the proboscis, the second the ganglion, the third, which soon divides into two, the sexual glands, and, finally, the fourth the sexual ducts. The cell-layer which surrounds the groups subse- quently splits into two layers, which were treated by Leuc- kart as answering to the soma- tic and splanchnic layers of the mesoderm. In the absence of the intestine, the splanchnic layer would be represented by the so-called ligament and the proboscis- sheath only, both of which structures arise from it. The somatic layer, after it has separated farther from the splanchnic, and has left the body cavity between them, would form the musculature of the body, whereas the epidermis arises directly from the larval skin. When the internal formative processes have pro- gressed sufficiently to allow of it, the cuticula of the larva is alone cast off. A new cuticula then arises. During these Fi6. lU.—A, B, two larvae of Echi norhynchus jiroteus, in which the " embryonic nucleus " is already un- dergoing its metamorphosis (after Lkuckabt). r, proboscis; rt, pro- boscis-sheath ; 11, ganglion ; g, funda- ment of the genital glands ; I, ducts of the genital system; m, the cell- layers which are destined for the formation of the musculature. ACANTHOCEPHALI 253 processes the different parts of the " embryonic nucleus " have also enlarged considerably, and thus have once more approached somewhat nearer to the larva in size (Fig. 114 B). At the same time the differentiation and development of the different organs begin. From the preceding description, which contains the chief features of Leuckart's discoveries, it is seen that the largest part of the worm arises from that frequently men- tioned central mass into which the nuclei are said to retreat at the beginning of development. The more recent state- ments of Kaiskr also agree in general with this conclusion. Since it appears that, owing to improved methods, certain processes have been more elaborately worked out by him, and since these are of a most peculiar nature, his obser- vations will be considered here more at length, although it is difficult to obtain a clear idea of the complicated processes from his brief communication, unaccompanied as it is by illustrations.^ After the larva of Echinorhynchus gigas has attached itself to the wall of the intestine, and the middle part of its body has become greatly swollen, groups of cells are said to detach themselves one after the other, to become surrounded with cytoplasm, and thus to form the cells which produce the permanent hooks of the proboscis. The groups of cells move forward and finally unite to form the proboscis, which at length can be everted. At about the same time the permanent body-covering of the worm is formed by the detachment of nuclei from the entire periphery of the "embryonic nucleus" and their migration into the outer layer of the body plasma (Fig. 114). Accompanying an active division of the nuclei, there is soon formed a very regular body-epithe- lium. Here also the cuticula of the larva appears to be cast off, just as its provisional hooks are. The epithelium secretes a new cuticula, and beneath it a colourless, tenacious product, the fibrous tissue of the so- called subcuticula. The muscular elements which are found in the sub- cuticula are said to be formed at the same time from the cells of the ^ [A paper by Hamann (Appendix to Literature on Acanthocephali) and especially a voluminous work by Kaiser have furnished us with a new exposition of the development of Echinorhynchus. These investigations elucidate to a great extent the remarkable developmental processes, which are here only briefly touched upon. We refer to these two works them- selves, since it is not possible to give in this place the results of these extensive studies. — K.] 254 EMBRYOLOGY body-epithelium. They arise as "primitive muscle-fibres" in the epithelial cells, and pass from these into the fibrous tissue of the sub- cuticula. When this process is completed, the body-epithelium de- generates completely and disappears. The formation of the lemnisci agrees with that of the skin. The nuclei, which have separated from the central mass and moved to the anterior end of the body, here form a circular swelling which at two diametrically opposite points is drawn out into slender processes, the fundaments of the lemnisci. In them the formation of the fibrous tissue takes place just as in the skin. Near the anterior end of the body, and immediately behind the rod-like pro- boscis, there also lies an extensive mass of nuclei, the fundament of the central nervous system, from which the peripheral nerves soon grow out to the different organs. The organs the development of which has thus far been described are said to be of ectodermal origin ; this, indeed, is very probable, although sufficient grounds for this conclusion cannot as yet be recog- nized in Kaiser's description. The real body-musculature, the sexual glands, and the ducts of the genital apparatus arise, according to Kaiser, from the entoderm. Leuckart had spoken of a mesoderm, which splits into an outer and an inner layer, but as yet Kaiser has not given attention to this statement. Again, it is layers of nuclei which separate from the central mass to give rise to new structures. Three such layers of nuclei can be recognized, owing to their somewhat different shape. The two outer ones soon migrate to the body- wall, and here, after various metamorphoses, supply the circular musculature and the longitudinal musculature of the body. Behind the proboscis, in the neighborhood of the ganglion, are found nuclei, arranged in definite order, concerning whose origin more accurate knowledge would be important, for out of them arise the proboscis-sheath and the retractors as well as other muscles of the proboscis, therefore structures which would be ascribed to the inner layer of the mesoderm did such exist. The formation of the genital oi'gans takes place in quite a peculiar manner. Behind the proboscis-sheath a prismatic protoplasmic mass makes its appearance, from the edges of which arise four thin plates, which divide the cavity of the body into four sectors. By this descrip- tion one is mvoluntarily reminded of the mesenteries which unite the fundaments of the genitalia with the body- wall, and at the same time recalls the conditions which, according to Vejdovsky, exist in the Gordiidse. In the female the plates unite in the dorsal and ventral sectors to form the ligament ; in the male the plates of one sector degenerate. The germ glands themselves arise from the axial mass of plasma. The resemblance of the thin plates to mesenteries, referred to above, is increased, as far as can be judged from the brief statements of Kaiser, by the two lateral sectors being filled with a cellular mass; subsequently, however, this degenerates and thus gives rise to the body ACANTHOCEPHALT 255 cavity. If then, provided we rightly understand K\iser's statements, a union of two plates were to take place dorsally and ventrally, the re- semblance to mesenteries would indeed be strong. The plates could certainly arise only from the above-mentioned third or inner layer, which separated from the central mass at the time of the formation of the body- musculature. The two outer layers would then be applied to the body- wall, whereas the inner layer would perhaps assume the formation of the internal organs, the proboscis-sheath, and the ligament, in some such manner as was described by Leuckart. This is the way at least in which we should interpret the statements of Kaiser in the absence of his more detailed descriptions. In regard to the origin of the genital organs, especially the extensive conducting apparatus, we refer to Kaiser's communication, or, better still, to the awaited complete work, for it cannot be determined from the former what is the real origin of those elements which constitute the genital apparatus. The Echi7iorhi/nchus, which even in the body of the inter- mediate host attains in general the form of the adult worm, becomes capable of reproduction only when the animal harboring it is consumed by another which is adapted to serving it as permanent host, thus, for example, the Gammarus by a fresh- water fish or a duck, if the species be Echinorhynchus polymorplius. Literature. 1. Greeff, E. Untersuchungen iiber den Bau und die Naturgeschichte von Echinorhynchus miliarius (E. polymorphus). Arch. Natuig. Jahrg. xxx., Bd. i. 1864. 2. Kaiser, J. Ueber die Entwicklung des Echinorhynchus gigas. Zool. Aiizeiger. Jahrg. x. 1887. 3. KoHLER, K. Documents pour servir a I'histoire des Echinorhynques. Journ. Anat. et Physiol. Tom. xxiii. 1887. 4. Leuckart, E. Die Parasiten des Menschen, etc. Bd. ii. Leipzig. 1876. 5. MectNin, p. Eecherches sur I'organisation et le developpement des Echinorhynques. BvlL Soc. Zool. de France. Tom. vii. 1882. 6. Safftigen, a. Zur Organisation der Echinorhynchen. Morph. Jahrb. Bd. x. 1885. Appendix to Literature on Acanthocephali. I. Hamann, 0. Die Nemathelminthen. Beitrage zur Kenntniss ihre Entwicklung, ihres Baues, und ihrer Lebensgeschichte. Jena. Zdtschr. Bd. xxv. 1891. II. Kaiser, F. Die Acanthocephalen und ihre Entwicklung. Bibliotheca Zuologica. Heft 7. 1893. CHAPTER IX. KOTATOKIA. 'The Eotatoria are peculiar in regard to their reproduction. Three different kinds of eggs occur among them : in the first place, thin-shelled summer eggs, which develop par- thenogenetically into females ; then eggs similar to these, but of only half the size, from which arise the simply organized males ; and finally thick-shelled winter eggs or resting eggs, which, as it appears, require to be fertilized. The eggs are either deposited free in the water or cemented to the body of the female. The development of the thin- shelled eggs takes place in many forms even in the body of the mother ; that of the resting eggs occurs only a long time after laying. The expulsion of the polar globules precedes cleavage. The parthenogenetically developing eggs, according to Weismann unb Ischikawa, produce only one polar globule. Little is yet known concerning the development of the Rotatoria. The chief descriptions are from Salensky, Joltet, and Tessin ; they present, however, many gaps. In our presentation of the subject we follow principally Tessin's work, which is occupied chiefly with the development of Eosphora digitata. Cleavage is from the beginning unequal (Tessix, Joliet). In the stage of four blastomeres one large and three small cells can be distinguished (Fig. 115 A). At the time when the latter divide into six, the abstriction of a new portion from the large blastomere takes place, and when those cells which subsequently supply the mesoderm are differentiated from the cells at first produced, a division of the large blastomere is still in progress (Fig. 115 jB). That part of it which is now left as a rather extensive remnant represents 256 ROTATORIA 257 fthe fundament of the entoderm, for it is subsequently over- grown by the other cells. The small blastomeres, which now [divide repeatedly, are, however, to be considered as ectoderm [and, mesoderm. Particularly striking is the statement that the mesoderm (in the form of three dark, granular cells) arises by division of the small blastomeres that were first to appear, and that it still remains united to the ectoderm, whereas even after its differentiation ectodermal elements continue to be separated off from the large blastomeres. According to 0. Zachiakias, however, the mesoderm is supplied directly by the large blastomere, which on the whole corresponds more to the ordinary mode of formation of the mesoderm, but does not, it is true, appear to be well established in the case under consideration. The conditions of for- mation of the meso- derm hitherto known do not allow a com- parison with the An- nelida, as one would perhaps expect from the relationships of the Rotatoria to these forms. The three meso- derm cells lie at the subsequently dorsal side of the embryo (Fig. 115 0). With the progressive division of the ectoderm cells and the commencing circumcrescence Fie. 115.—^ to F, stages of development of Eosphora digita*a (after Tessin). A to C, cleavage stages; D, epibolic gastrula. The large blastomeres are already entirely OYergrown ; the mesoderm cells lie at the blastopore. E, the mesoderm cells have moved in- ward ; an invagination of the ectoderm follows it ; the entoderm cells have multiplied. F, embryo, on which the head-, tail-, and lateral lobes can be recognized. Bl, blastopore ; Ec, ectoderm ; En, ento- derm ; Mes, mesoderm. of the large blasto mere by these, the mesoderm cells are pushed farther forward (Fig. 115 D). Meanwhile their number has doubled. Even before the formation of the epibolic gastrula is completed the enclosed entoderm cell has divided. As the result of K. H. E. s 258 EMBRYOLOGY the forward growth of the ectoderm, the mesoderm cells are now forced inward, and the invagination of ectoderm cells, which succeeds it (Fig. 115 D, E), subsequently pro- duces the trochal apparatus and the oesophagus. The outer form of the embryo is now changed in such a way that an anterior, posterior, and two lateral protuberances can be distinguished (Salensky, Tessin). On the surface which bears the blastopore, these regions of the body are seen to be separated from one another by shallow grooves (Fig. 115 F). The posterior elevation bends forward, and growing further in the same direction, forms the foot (or caudal appendage) of the Rotifer. Tessin seeks to refer the anterior and lateral elevations (cephalic and lateral lobes) to the lobular processes of the Turbellaria, especially to those of the larva of Stylochus. Inasmuch as the Eotatoria do not pass through any real larval stage, the lobular processes would have become rudimentary. In the further course of development the cephalic and lateral processes are again smoothed out, and can no longer be recog- nized as special structures. Concerning the origin of the inner organs even Tessin can give little definite information. We have already mentioned that he derives the trochal organ and the most anterior part of the intestinal canal from an ectodermal invagination. On the other hand, he combats the discoveries of Salensky, for he derives the masticating stomach (pharynx), which is provided with jaws, from the entoderm ; Salensky maintains that this part is of ectodermal origin. According to Tessin, by far the largest part of the intestine (together with the appended glands) arises from the entoderm, for the latter extends far backwards ; it is said even to send a process into the caudal appendage. The hind-gut arises by means of an invagination of the ectoderm (Salensky, Joliet). The further fate of the mesodermal fundament remained obscure to Tessin. The statements regarding the origin of the nervous system and the genital organs are of too doubtful a nature for us to consider them. Nothing is as yet known concerning the formation of the excretory organs. The development of the male of BracMonus urceolaris, which, as is known, is very simply constructed, takes place, according to Salensky, in the same way as that of the female. The degenerative processes which characterize the I ROTATORIA 259 incomplete form of the male begin only after the trochal organ and the foot have been formed.^ General Considerations. The development of the Rotatoria gives us as jet no information as to their doubtful position in the system. Such forms as that of Trochosphcera cequatorialis (Fig. 116), found bj Semper in the Philippines, point with almost imperative force to relationships with the Tro- 1 - ^ I// W, M ^ 1/ ..^^^^^^^wwfw^ \-"''.'W tl. _1 '\M=±t ~^^t^- - '■ Fig. 116. — Trodiosphcera cequatorialis (after Sempke). Ce, cloaca; Dr, appendi- cular glands of the fore-gut; Ex, duct of the excretory organs; G, brain; Ge, female sexual organs wirh duct ; M, mouth ; Mu, musculature ; N, nerve that emerges from the brain; S, oesophagus ; Si, sense organ ; TF^, preoral, If,,, post-oral, ciliated band. chophore larv^a of the Annelida (comp. p. 266). Like the latter, Trochosphcera possesses a complete preoral circle of cilia and an indication of a post-oral one. Both of these are also to be recognized in the trochal organ of other Rotatoria, the form of which is different from that of the Trochophore. The course of the intestine is similar to that of the annelid Trochophore. The structure of the excretory system also argues for a relationship with the Trochophore- like forms. The excretory canals of the Rotatoria begin with blind ciliated funnels in the body cavity, and the same 1 [Our knowledge of the development of the Eotatoria has recently- been much enlarged by the thorough investigation of Zelinka, to which the reader is referred (see Appendix to Literature on Eotatoria). — K.] 260 EMBRYOLOGY is said to be the case in the Trochophore. In the Rotatoria there are two trunks to the excretory system, which in the Gastrotricha, which are related to the Rotatoria, open to the exterior by means of two ventral openings (Zelinka, No. 12), so that this organ thus acquires still greater resem- blance to the so-called head kidney, the excretory system of the Trochophore, for the two head kidneys also open directly and separately from each other to the exterior (comp. p. 266) The agreement of the Rotatoria with the Trochophore was especially advocated by Hatschek (No. 1), with whom recent investigators of the Rotatoria, such as Plate and Zelinka, in the main agree (Nos. 3, 4, 11, 12). Tessin contends against the relationship of the Eotatoria to the An- nelida or their ancestral form, which we have briefly indicated above, because, owing to the origin of the trochal organ from the stomodeal invagination and the position of the brain outside of the area included within the trochal organ, a comparison of the trochal organ with the ciliated rings of the Trochophore larva does not seem to him admissible. Tessin seeks rather relationships to the Turbellaria, being influenced by the lobular structures of the embryo. His comparison of the caudal appendage of the Eotatoria with the post-abdomen of the Crustacea, which he supports with the fact that a process of the entoderm is said to extend into the tail, seems weak. This perhaps indicates that the Eotatoria have a tendency to increase in length. We would here recall the growing out of the Trochophore into the worm. Eelationships of the Eotatoria to the Arthropoda have also been found in forms such as Hexarthra polyptera, to which attention has recently been called by Plate (No. 4). This remarkable Eotifer, discovered by Schmaeda, possesses on the ventral side three pairs of movable setose appendages, which are like extremities, and give to it almost the appearance of a Nauplius. In view of the close relationships of the Eotatoria to the Trochophore, one will certainly not think of a descent of the Eotatoria from the Arthropoda, especially from the Crustacea; it is, however, interesting to see how Trochophore-like beings can vary in the direction of the Arthropod type, even if it be only in their outward shape. Still less justifiable than a comparison of the caudal appendage (foot) of the Eotatoria with the abdomen of the Crustacea is such a comparison with the foot of the Mollusca, which has been attempted by various writers, who have based their conclusions principally upon the position of both organs between the mouth and anus, which is particularly well expressed in embryos and larvte. ROTATORIA 261 Literature. Hatschek, B. Studien zur Entwicklungsgeschichte der Anneliden. Arheiten Znol. Inst. Wlen. Bd. i. 1878. 2. JoLiET, L. Monographie des Melicertes. Arch. Zool. exper. et gin. Ser. ii., torn. i. 1883. 3. Plate, L. Beitrage zur Naturgeschichte der Eotatorien. Jena. Zeitschr. Bd. xix. 1886. 4. Plate, L. Ueber die Eotatorienfauna des bottnischen Meerbusens. Zeitschr. wiss. Zool. Bd. xlix. 1889. 5. Salensky, W. Beitrage zur Entwicklungsgeschichte von Brachionus urceolaris. Zeitschr. wus. Zool. Bd. xxii. 1872. 6. ScHMARDA, L. Zur Naturgeschichte -iEgyptens. Denkschr. Akad. Wiss. Wien. math.-naturw. Klasse. Bd. vii. 1854. 7. Semper, C. Zoologische Aphorismen : Trochosphsera sequatorialis, das Kugelraderthier der Philippinen. Zeitschr. wiss. Zool. Bd. xxii. 1872. 8. Tessin, G. Ueber Eibildung und Entwicklung der Kotatorien. Zeitschr. wiss. Zool. Bd. xliv. 1886. 9. Weismann und Ischikawa. Ueber die Bildung der Richtungskorper bei thierischen Eiern. Berichte naturf. Gesell. Freiburg i. Br. Bd. iii. 1887. 10. Zacharias, 0. Ueber Fortpflanzung von Rotifer vulgaris. Zeitschr. wiss. Zool. Bd. xli. 1885. 11. Zelinka, C. Studien iiber Raderthiere II. Zeitschr. wiss. Zool. Bd. xlvii. 1888. 12. Zelinka, C. Die Gastrotrichen. Zeitschr. wiss. Zool. Bd. xlix. 1889. Appendix to Literature on Rotatoria. Zelinka, C. Studien iiber Raderthiere III. Zur Entwicklungsgeschichte, der Raderthiere nebst Bemerkungen iiber ihre Anatomie u. Biologie. Zeitschr. wiss. Zool. Bd. liii. 1891. CHAPTER X. ANNELIDA. I.— CH/ETOPODA AND ARCHIANNELIDA. The two chief divisions of the Chsetopoda are unlike as re- gards development, this being in the Polychseta generally indirect, and involving a free-swimming larval stage, where- as in the Oligochseta it is considerably abbreviated, and free-swimming larv« are absent. The Archiannelidi are like the Polychaeta. 1. Development through Free-swimming Larvae (Foltjchdeta and ArcTiiannelida). In general the Polychaeta develop from free-swimming larvse which are provided with ciliated bands. Only a few forms bring forth living young. Such is the case in Eunice sanguinea, SylUs vivipara, and in a Cirratulus in which the eggs develop either in the body cavity or in the cavity of a segmental organ which has become a uterus. A kind of brooding also occurs in many forms, as, for example, in Auto- lytus cornutus, an extensive sac, in which the eggs are developed, being formed on the ventral surface by the dis- tension of the skin of the body. In Polynoe cirrata, the eggs, which are stuck together in a single mass, are carried on the dorsal surface under the dorsal scales. Similar to this is Grubea limhata, in the females of which at the time of sexual maturity the entire back is thickly covered with eggs, while Exogone gemmifera and Sphasrosyllis pyrifera carry their eggs on the ventral side, namely on the ventral cirri (Viguier, No. 46). In Spirorhis Pagenstecheri the tentacle which bears the operculum of the tube is enlarged, and thus serves as a brood-chamber ; in Spirorbis spirillwin, on the contrary, the 262 ANNELIDA 263 eggs are deposited inside the tube, and are found here be- tween it and the bodj-wall. Similar to this is the brooding in the genus Capitella, in which one finds the eggs firmly- attached like a mosaic, on the inner surface of the tube. Other tubiculous worms attach their eggs to their habitations on the outside ; thus in Sahella lucullaria the eggs, which are enveloped in a slimy mass, form a thick ring around the tube of the parent. Many Polychgeta deposit their spawn in the form of large gelatinous packets or clumps (Aricia, Psygmo- hranclius) ; others discharge the eggs into the sea-water without any other protection than the egg-membrane (Eupomatus, Pomatoceros et al., likewise Polygordius). In such forms artificial fertilization can be employed with success. Cleavage is unequal, but in some forms may approach very near to the equal: type (Pomatoceros according to v. Drasche). In the latter case it produces a coslohlastula, the entodermic part of which can be distinguished by the greater thickness (Sahellaria, Aricia), or an epibolic gastrula is formed (Nereis, Psygmohranchus). The Polychaeta, which have been studied with this object in view by Hatschek, Goette, v. Drasche, Salensky, and other investigators, offer all transitions between the different types of cleavage, and correspondingly the form of the gastrula also varies from a typical invagination to an epibolic gastrula. In Terebella Meckelii, for instance, we find a blastula with the wall thickened on one side, the cavity of which soon becomes filled by the intruding macromeres^ so that we now have before us a so-called sterrogastrnla (Salen- sky). [Wilson (Appendix to Literature on Annelida, No. XXVII.) has recently given a very detailed account of the early stages of development in Nereis, especially of the cleav'- age, in which the fate of the individual cells is established with great precision. — K.] As an example of the embryonic development of a Poly- chgete, we select that of Eupomatus (according to Hatschek). The spherical e^g is divided by means of the first two meri- dional planes and succeeding equatorial plane of division into eight blastomeres of almost equal size. Soon, however, the divisions at the animal pole take place more rapidly than 264 EMBRYOLOGY at the vegetative pole, and thus the blastomeres at the latter remain more voluminous. Jn the resulting blastula, the cells from which the three germ-layers arise are already differen- tiated (Fig. 117 A). The upper hemisphere, composed of smaller cells, gives rise to the ectoderm, and the greater part of the lower to the entoderm ; however, two cells here are distinguished at an early period from the others by assuming a more spherical shape : they produce the mesoderm, and are called by Hatschek the primitive mesoderm cells [mesodermal teloblasts (Fig. 117)]. The region where they lie corresponds to the anal end of the larva. Even as early as this stage a delicate equatorial circle of cilia makes its appearance, the future preoral ciliated band of the larva. nar Fig. 117. — A, B, blastula and gastrula stages of Evpomatus nncinatus (after Hatschek); eh, egg-membrane; mes, one of the two mesoderm cells. Soon afterwards the tuft of cilia at the apex of the larva makes its appearance (Fig. 117 B). The cilia perforate the egg-membrane, which therefore most probably consists of a soft mass. The subsequent behaviour of the egg-membrane is of a peculiar nature for, according to the concurrent statements of various authors, it is pro- visionally retained, increases in extent with the growth of the larva, and is thus formed into a cuticula-like envelope (Hatschek, No. 20), which, however, is replaced later by the permanent cuticula from the ectoderm. Thus here the embryo is converted directly into the larva. The entodermic part of the blastula now invaginates. At ANNELIDA 265 the same time the two primitive mesoderm cells have moved into the inside, having detached themselves from their con- nection with the other cells. In the lateral aspects seen in Fig. 117 A and B, only one of the two cells can be recognized. It should be mentioned here that later they divide (Fig. 118). The two primitive cells still continue to be dis- tinguished from the newly formed ones by their greater size. Hatschek calls them the two pole cells of the mesoderm. They lie at the ends of the two mesodermal bands formed by cell-proliferation. In the farther development of the Fio. 118.— -4 and B, trochophores of Eupomatus in younger and older stages of development (after Hatschek). a, anal opening ; ab, anal vesicle ; fcii, head kidney; m, mouth-opening; mes, mesodermal bands; mu, muscles; oc, eye-spot; ot, auditory vesicle ; sp, apical plate. larva the intestine bends toward the anal side, in order to fuse in later stages of development with the originally slight depression of the ectoderm, which produces the hind- gut and anus (Fig. 118). Some time before the completion of this process, the blastopore had become considerably narrowed. It assumed the form of a fissure which closed (from behind forwards) and left only a small opening re- maining in front. At the place of this last trace of the 266 EMBRYOLOGY blastopore, the ectoderm becomes invaginated, and forms the oesophagus. This is followed bj an enlargement, the sto- mach of the larva, and this in turn by the small-gut and the hind-gut (Fig. 118 5). The formation of the intestine takes place less simply in the eases where the gastrulation is epibolic and the entoderm at first consists of a compact mass of cells. The intestinal wall is only gradually formed, detaches itself from the central yolk-mass, and finally unites at the fore- and hind-guts with the ectoderm (comp. the figures 128 A and B, p. 280, of Psyginobranchus). Trochophore. — Even during gastrulation the embryos rose, with the aid of their vibratile apparatus, from the bottom, and betook themselves to the surface of the water. Together with the internal changes described, alterations have also taken place on the outer body, the upper portion of which has become bell-shaped, while the under-portion tapers more conically (Figs. 118 jB and 119). The band of cilia which lies in front of the mouth extends around the longest periphery. Thereby is reached the larval stage, de- signated by Ray Lankester as the Trochosphere, but now with ^^ Hatschek more generally known Fig. iiG.-Larva of Poiygordius as the Trochophore. In addition (after Hatschek from Balfour's ^o the Organs already mentioned (jom.'parative Embryology), an, anal . . " . *^ . opening ; m, mouth-opening ; me. p, — ciliated band, intestinal tract, mesodermal band; nph, head kid- and mesodermal bands — still ney ; ol, stomach ; sg, apical plate. • i m others can be seen m the Tro- chophore. An ectodermal thickening at the upper pole, which bears the tuft of cilia, if such be present, is de- signated as the apical plate; it represents the fundament of the super-oesophageal ganglion (Figs. 118 A and 119). The cells of the preoral ciliated hand also form a similar thickening. These may consist of several successive circles of cells, and between or underneath them is placed a ring of fine nerve fibres, which is connected with ganglionic cells, and is considered by its discoverer, Kleinenberg, as the ANNELIDA 267 r central nervons system of the larva. At the base of the ciliated cells there also lies, according to Kleinenberg, a ring of muscle cells, which, like the ciliated band itself, is made use of by the larva in locomotion. In addition, various other muscle strands traverse the inside of the body ; thus some extend from the apical plate to the stomach, others are found in the lovrer part of the body, and one surrounds the intestine at the point where the stomach and oesophagus unite (Fig. 118 B). These muscle cells have become de- tached from the mesodermal bands (Hatschek). From the latter also arises the so-called head kidney, the paired excre- tory organ ; it is formed from a few cells situated near the pole cells, which increase greatly in length and become hollow. The head kidney then extends from the pole cells, that is, from the vicinity of the anus, as far as the oesophagus (Fig. 118 B, hn). It consists of a ciliated canal, which may branch (as, e.g., in Folygordius), and of one or more funnel-shaped terminations (Figs. 119 nph and 120 B, hn). These do not open freely into the blastocoele, but are said to end blindly (Fraipont), and in this regard therefore re- semble more the excretory system of the Platyhelminthes and Rotatoria. The spot where each of the two head kidneys opens to the exterior can be seen from the figures 120 A and B. Of the ectodermal structures of the larva there should still be mentioned as important, in the first place, the eye- spot, consisting of an accumulation of pigment, which in the larva of Eupomatus is located in a cell of the apical region, but asymetrically on the right side (Fig. 118 B, oc). The two ectodermal vesicles which arise symmetrically on the posterior portion of the body, each from one ectoderm cell, also constitute sensory organs (Hatschek). They are pro- vided with fine stiff hairs, which project into their lumina ; highly refractive concrements are also found inside of them. Thus they are to be recognized as otolith vesicles (Fig. 118 B, ot). The large sac which in Fig. 118 B is seen lying at the posterior end of the larva arises by the enlargement of an ectodermal cell. This anal vesicle appears to have no im- portant significance. In Eupomatus is developed another 268 EMBRYOLOGY (perianal) circle of cilia, which is situated on the posterior portion of t!ie body (Fig. 118 B) ; this is lacking in many- other Annelids. Furthermore there is added on the ventral side a ciliated area extending from the mouth backward, the adoral ciliated zone (Figs. 118 B and 128 A, p. 280). The metamorphosis of the Trochophore larva into the adult worm will be followed in Polygordius, a form in which it takes place in a particularly clear manner. The Polygordius larva was first discovered by Loven, and held to be that of a cha3tiferous worm. Ant. Schneider showed that Loy^n's larva belonged to Polygordius. It has the form of a typical Trochophore (Fig. 119). The ciliate apparatus, which encircles the larva at its greatest diameter, is com- posed of two rings, one in front and one behind the mouth. The preoral ring consists of a double, the post-oral of a single, row of cilia. A third ring, the perianal ciliated band, makes its appearance at the posterior end of the larva (Fig. 120 J5), bat it is not formed until the later stages of development. The development of the Trochophore into the perfect worm, which has been thoroughly studied by Hatschek, consists first of all in a growing out of the posterior section of its body and a gradual reduction of the anterior part. At first a segmentation of the larva is noticeable (Fig. 120 A), which depends upon a marked change in the two meso- dermal bands. These, which at first consisted of only a few cells, have become by active cell-proliferation much more voluminous. Each of them is separated into two cell-layers (Fig. 133 A to C, p. 290), and spreads out toward the ventral and dorsal lines. Then a segmentation makes its appear- ance in them, proceeding from in front backwards (Fig. 120 A), and at the same time the two layers of the bands sepa- rate from each other by the formation of a cavity in each segment. In this way the primitive segments arise, the outer and inner walls of which become in each segment of the worm the somatic and splanchnic layers of the mesoderm, and the walls, abutting on one another, form the segmental boundaries (dissepiments) of the body of the worm.^ Since ^ A more detailed description of these conditions will be found in the discussion of the formation of the body cavity (comp. p. 289). ANNELIDA 269 r ^^m to each segment of the body a pair of primitive segments be- i|» longs, these meet in the middle line of the ventral and dorsal surfaces, and form there a ventral and dorsal mesentery (Fig. 133 G, p. 290). In the figures 120 A and B (lateral views of the larva) the primitive segments can already be recognized in the form of an internal segmentation of the larva. The most anterior primitive segments are the oldest, the posterior ones younger. The body is seen to be already considerably grown out backwards, although the head por- tion has not yet diminished in circumference. Two ciliated tentacles, which are still very small, now make their appear- ance on the apical plate (Fig. 120 B). The originally sac- like mid -gut has grown in length with the body, and is now cylindrical in shape. Very near the posterior end of the body, a short distance in front of the anus, is the pos- terior ciliated band (Fig. 120 B). While the pos- terior part of the body of the larva is gradually passing from the earlier conical into the cylindrical form, the head portion attains its greatest volume, but thereafter soon diminishes. The metamorphosis of the voluminous cephalic vesicle into the slender cephalic process of the worm is effected by the thickening and conical outgrowth of the apical plate (Fig. 121 A), and by the contraction of the wall of the head generally. The previously flat cells become considerably thickened, whereby the circumference of the R. \ Fig. 120.—^ and B, larvae of Polygordius (after Hatschek). a, anus ; m, mouth-opening ; fcn, head kidney ; raes, mesodermal bands j sp, apical plate. 270 EMBRYOLOGY entire head is diminished, until it is not much larger than the trunk. The apical plate has grown out forward in the form of a cone. The ejes are more conspicuous than in the larva. In the trunk the primitive segments have in- creased in number, and made the segmentation of the body still more distinct, since thej have enlarged more and have applied themselves more closely to the intestinal and bodj-walls. At the posterior part of the trunk they are less clearly expressed. These changes are much more evident in the last stage of development (Fig. 121 B), which we introduce for comparison. There the segmental constrictions of the intestine cause the metame- rism to be still more distinct. The cephalic vesicle and the vibratile apparatus have already entirely disappeared in this stage ; and we have now before us in its chief features the adult worm, although it has not yet reached its complete development. The worm gives up the larval mode of life, that of floating upright in the Avater, and adapts itself to locomotion by creeping. The papillae, which make their appearance in front of the posterior band of cilia, which has now disappeared (Fig. 121), serve the worm for the purpose of attachment. The Different Larval Forms. — PoJygordius was selected as an example because it shows in a par- ticularly instructive manner the transition of the larva into the worm. It does not show, however, the ordinary condition Fig. 121.—^ and B, larvse of Polygordius (after Hatschkk). a, anal opening; m, mouth- opening ; fcn, head kidney. ANNELIDA 271 of the Trochophore larva, for the anterior bell- shaped part in the majority of cases is not retained unaltered for so long a time. Generally also it does not surpass the trunk so considerably in size, and it soon comes to be even smaller than the trunk. Since in many forms the typical shape of the Trochophore is not so strongly expressed, and, on the other hand, the segmentation of the body of the worm makes its appearance at an early stage, many deviations from the shape of the Trochophore are realized. The larvae of Annelids are very variously shaped, for some of them, owing to the early appearance of the segmentation, are found in phylogenetically younger stages than the Trochophore^ and others, although they stand at the same level with it, may be modified by the occurrence of various kinds of locomotor organs and by other external changes in form. The principal difference in the larvce consists in the presence or absence of segmentation of the entire larva, not includ- ing that of the trunk part, which is acquired only during the metamorphosis. To be sure, this difference should not be overrated, for the segmented forms likewise must pass ontogenetically through an unsegmented stage. The Annelid larvae have usually been distinguished according to the dis- tribution of their cilia : as Atrochse when a ciliated band is lacking ; Monotrochae with a preoral band of cilia, to which, as in the Telotrochae, there may be added a post-oral band lying directly behind the mouth ; Telotrochse with an anterior and posterior (perianal) band of cilia ; Mesotrochse, in which the ciliated band is situated in the middle of the body ; and, finally, Polytrochse, which possess a greater or smaller num- ber of ciliated bands, and as a result of this exhibit at an early stage a segmentation of the body. The ciliated bands of the Polytrochas may form either closed rings, or only half- rings. In the latter case, according to their position on the dorsal or ventral surface, Nototrochas and Gastrotrochse are in turn distinguished. They are called Amphitrochse when ventral and dorsal half-rings alternate with one another. This classification has been made use of by different investi- gators for distinguishing the larvae. However, ClaparIide and Metschnikoff themselves, to whom we owe the most 272 EMBRYOLOGY thorough knowledge of the Annelid larvae, point out that the characters cited have no great morphological value, for larvae occur in the same family, and even in the same genus, which belong to more than one of these types. The differ- ences in shape are probably due to differences in the mode of life. Variations in regard to the development of the locomotor apparatus — i.e., in the distribution and stoutness of the ciliation — would easily follow, if the larvae of closely related forms adopted different modes of life, as is actually the case. Terebella larvce (Terebella conchilega) are known which must be placed among the Nototrochse, while others belonging to this genus entirely lack the ciliated rings {Tere- bella Mechelii). The former are good swimmers, and lead a pelagic existence ; the latter, on the contrary, never move far away from the masses of eggs from which they hatch, and may sometimes develop into young worms, even in the jelly surrounding the eggs. The presence or absence of the preoral band of cilia may well be im- portant in the interpretation of Annelid larvae, for (according to Klei- nenbebg) this alone possesses a ring-nerve, which is said to be lacking in all other bands of cilia that make their appearance, with the exception of the so-called post-oral band, which stands in close relation to the preoral. Even where posterior ciliated bands appear without the existence of an anterior one, as in the Mesotrocha, the ring-nerve is said to be absent. But these conditions are as yet too little studied to allow one to base on them a distinction between the larvse. Out of the multitude of variously formed Annelid larvae, only a few of the particularly characteristic forms can be chosen. We shall first consider the unsegmented larvae. The simplest larvae of the Annelids are un- doubtedly those whose body is covered with a Fi6. 122.— A and B, so-called atrochal Annelid larvae — ^, of Lumhriconereis (?) (after CLAPAKioB und Metschnikoff); B, of Sternaspis scutata (after Vkjdovsky). cu, cuticula ; d, intestine: eni, entoderm. ANNELIDA 273 P ^^J uniform coat of cilia, and which at the most possess at the ^^K anterior end of the body a tuft of cilia, which serves for ^B steering (Atrochce, Fig. 122 A, B). ^^B CLAPABfeDE UND Metschnikoff describe atrochal larvsB of Lumbnco- nereis (?), and Vejdovsky those of Sternaspis.^ Both larvse are at first spherical, but later become elongated (Fig. 122 B). The former possess eye-spots ; the latter do not. A differentiation in the ciliation appears even in these larvae, for in Lumhriconereis narrow zones, one toward the anterior and one toward the posterior end, remain free from cilia, and in Sternaspis the entire posterior end is without cilia (Fig. 122 A, B). Inside one recognizes in the first form the sac-like fundament of the digestive tract, in the latter, on the contrary, only a compact mass formed of large entoderm cells. The further development of Lumhriconereis is marked by the ap- pearance of setae in pairs at the posterior end of the body, thus giving expression to the segmentation. At the same time the degeneration of the cilia begins. In Sternaspis the entire ciliation of the body dis- appears, and the larva continues to live in this naked condition for some time, the segmentation of the body being as yet unrecognizable (Vejdovsky, Eietsch). Its further development was not followed in detail. It is difficult to say whether in the evenly ciliated larvae we have to do with phylogenetically older stages than those represented by the Trocho- phore. The incomplete development of the intestinal canal, especially in the larva of Sternaspis, and also the subsequent development of this worm, make it appear as more probably a derived form. Although in Lumhriconereis the cilia in later stages are arranged into an anterior and posterior region, this distribution is altogether too indistinct to be referred to the anterior and posterior ciliated bands of the Trochophore. Although we are not justified in looking upon these atrochal larvae as primitive forms, still it appears to follow from the development of another Annelid that the larvae which are provided with ciliated bands represent a stage succeeding the atrochal forms. In Terehella Mechelii, which was studied by Milne-Edwards, Clapar^de und Metschni- koff, and later by Salensky, there arises from a larva, which * Sternaspis has been classed with the Echiuridae ; nevertheless in this form, which is provided with setae, such a distinct segmentation is ex- pressed, both externally on the body and internally, in the matter of the arrangement of the muscles and blood-vessels (Kietsch), that this group of Annelids — very aberrant, it is true— must still be placed among the Chaetopoda. K. H. E. T 27-4 EMBRYOLOGY at first is rather evenly ciliated, one having a preoral and a perianal band of cilia, which is substantially in the Trochophore stage. The young larvse of Terebella Meckelii are at first spherical, then elongate a little, and become covered with an even coat of cilia, which leaves bare only the small part of the anterior end of the body lying in front of the eyes. Later they become pyriform, and the cilia now cover only the voluminous anterior part of the body, whereas the posterior region is destitute of them. It is only in a later stage of development that a perianal row of cilia makes its appearance. In this stage it re- sembles the previously described larva of Lumhriconereis, The ciliation is gradually confined to a preoral band, a perianal band, and a ventral ciliated groove (Salensky). To be sure, the outer form of the larva is in this case, on account of the small size of its bell, not that of a typical Trochophore ; nevertheless nothing prevents us from comparing it to one that has already begun its metamorphosis into the worm. In front of the mouth lies the preoral band of cilia ; the intestine has the usual shape ; at the posterior end, in the vicinity of the anus, is found the perianal band of cilia. About midway between the anterior and posterior ciliated bands appear indications of the two first segments, behind which others soon follow. They become noticeable externally by the development of pro- tuberances, which are studded with setae. The worm grows in length ; evaginations at its anterior end form the tentacles ; it secretes the tube and attaches itself. The larvae of the Chcetopteridce, known as Mesotrochce, also arise from uniformly ciliated embryos. In Chcetopterus pergamentaceus, which at first is even- ly ciliated, there is formed a tuft of cilia at the anterior end of the larva, and gradually a ring also of cilia, en- circling the body at about the middle (Fig. 123). The inside of the larva is pretty well filled by the large sac-like intestinal canal. The larva of Telep- FiG.i23.-So-caiiedme- «^^'^« costarum is similar, only that it sotrochai larva of Chmtop- lacks the anterior tuft of cilia. On terns pergamentaceusi. iter ^^^ ^^^^^ ^^^^ ^^^ ^^^^^ ^f p;^^;^. Wilson), m, mouth. _ ' _ _ ^ '' chcetopterus socialis also exhibits stouter cilia at the anterior end ; it possesses two ciliated bands which lie close to the posterior end. A preoral band of cilia is not present in these larvae, and the middle one cannot be ANNELIDA 2'?5 [directly compared to the perianal ciliated band of other An- nelid larvae, for it is not, like that, situated at the posterior [end, but a number of segments make their appearance be- tween it and the hind end. The anus in these lartaeJ is placed dorsally, for a pointed prolongation is formed pos- [teriorly on the ventral surface, a condition which also recurs in polytrochal larvae (corap. Fig. 127). Noteworthy is the tuft of cilia at the anterior end which we met with in the atrochal and mesotrochal larva, and which is also recalled by the stout cilia found at the head end of many telotrochal and polytrochal larvffi. Such a ciliation of the apical area occurs also in Turbellarian, Nemertean, and Molluscan larva, and has perhaps a higher significance than that of a merely secondary acquisition, connected with the larval mode of life only. Apparently aberrant larval forms, such as those of Mitraria (Fig. 124 A, B), are referable to the Trochophore. Mitraria would, therefore, have to be classed with the Monotroch(E, in which a preoral, but not a perianal, band of cilia is de- veloped (comp. the larva of Psygmobranchus, shown in Fig. 128, p. 280). Monotrochce and Telotrochce cannot be separated from each other, inasmuch as in the beginning the larvae frequently possess only a preoral band of cilia, are therefore monotrochal, whereas later a perianal ciliated band, which gives them the character of Telotrochce^ is developed on them. Mitraria, the Annelid larva discovered by Joh. Mullee, and subse- quently more thoroughly studied by Metschnikoff, can easily be recog- nized in its young stages as a Trochophore, with a well- developed bell, but much-reduced posterior portion (Fig. 124 A). As a result of this, the anus and mouth are brought close together. The ciliated band lies in front of the mouth. Later the posterior part of the body grows out more, and the ciliated band, which acquires many outfoldings, there- fore comes to lie more anteriorly (Fig. 124 B). In this figure the beginning of the worm, which is gradually developed out of the Mitraria, can be easily recognized. On the lower area, surrounded by the ciliated band, two lateral protuberances, which bear long cilia, can be recognized in the young larva. In the older larva they are seen lying dorsally. The metamorphosis of the larva into the tubicolous worm is due to the vigorous growth of the segmented posterior portion and the degeneration of the chief part of the Mitraria, together with its lobes and setiferous papillae. Thereupon the larva sinks to the bottom, secretes the tube, and becomes attached. 276 EMBRYOLOGY In Mitraria we recognized a larva which possesses pro- visional larval appendages in the form of long bristles, Fig. 124.— Lateral views of Mitraria larvse (after Mbtschnikoff, from Balfour's Comparative Embryology), an, anus ; b and br, the lateral elevations with the pro- ] visional setae ; m, mouth; pr. h, preoral ciliated band ; sg, apical plate. ] which also occur in other Annelid larva8. Trochophore larvas ' are known which exhibit a number of long denticulate j bristles on both sides of the bodj, thus, e.g.., in the genera ', Sahellaria, Spio, etc. Figs. : 125 and 126 show larv86 I more advanced in develop- ' ment with richly developed, '. and in part extraordinarily | long, provisional setae. Setae 5 of this kind do not appear I in adult recent Chae^topoda, | but, on the other hand, are | found in fossil forms. It | has been conjectured that they might have been inherited from unsegmented ancestors of the existing Chsetopoda. This suggestion appeared to be supported by the fact that they are mostly found on the anterior unsegmented part of the larva (Alex. Agassiz). Fig. 125.— Larva of ^ferine (after Albx. Agassiz, from Balfoue's Comparative Embryology), ANNELIDA 277 I ^^H^ The larvre of Ophryotrocha puer His (Fig. 127) 8i.re Polytro- ^l^cJice — segmented larvae in the proper sense. They possess a ^ number of segments, each one of which is provided with a ciliated band. In addition, stouter tactile hairs are found at the anterior and posterior ends of the body. The first ciliated band belongs to the head region of the larva. Next to this is situated the mouth-opening, which leads into a large pharynx provided with a chewing apparatus. The intestine extends straight backwards, and opens to the exterior at the end of the last seg- ment. The anus is situated dorsally, in- asmuch as the last segment possesses a pointed process on the ventral side (Fig. 127). The next stages of development in Opliryo- trocha remain much Hke the larva described, since the new segments formed in front of the anal segment are also provided with bands of cilia. Knob-like parapodia then bud out on the segments, and in them the setag are developed. The number of the segments is considerably in- creased, yet this small Annelid, which never becomes over 2*5 mm, long, remains, as it were, in a larval condition, since the segments retain their ciliation throughout life. Still another ciliated band has been developed on the head, and two small knob-like feelers have arisen there, which bear long cilia, just as do the two cirri which have made their appearance near the unpaired process on the anal segment. The two most anterior segments remain without appert- dages (Claparede und Metschnikoff). Fig. 126. — Annelid larva with provisional setae (after Albx. Agassiz, from Balfoub's Comprti-a^ tive Embryology). In Ophryotrocha the ciliated rings surround the entire segment. They appear to be arranged in the same way in Arenicola marina ; in other larvae, on the contrary, they have the form of half-rings only, and are confined to the dorsal or ventral surfaces (Nototrochce and Gasterotrochce) . Noto- trochal larvae are found, for example, in Terebella conchilega. 278 EMBRYOLOGY gasterotrochal in members of the genera Magelona, NeHne, and Spio. In the two last-named genera there are found amphitrochal larvae— i.e., such as possess dorsal as well as ventral half-rings — in addition to the gasterotrochal, just as atrochal and polytrochal larvee appear in the genus Terehella. The polytrochal larvae sometimes appear as a stage succeeding other larval types. Thus those of Arenicola marina arise from larvae which at first were monotrochal, later became telotrochal, and finally, by the ap- pearance of new ciliated rings between those already present, assumed the stage of poly- trochal larvae (Max Schultze). Also the true polytrochal larvae — i.e., those which possess only the ciliated bands, but do not yet, like many other polytrochal larvee, ex- hibit the fundaments of the setae and other parts of the body of the worm — appear as the stage succeeding the Trochophore. Thus we have just ndted a stage entirely resembling a Trochoplwre, which preceded the polytrochal larva of an Ophryotrocha. This condition warrants the assumption that the segmented forms are to be looked upon as the younger, the unsegmented, on the other hand, as the phylogenetically older. As may be inferred from the manifold shapes of the Annelid larvae, their metamorphosis into the worm is also extremely varied. This has already been' briefly discussed in some forms while considering the larval stages. The segmentation may be expressed on the body of the larva in various ways. In some cases the body elongates and divides into segments, while the ciliated bands are still retained. In other forms the setae alone first make their appearance in pairs, and indicate the segmentation of the body, or at the same time the parapodia are established in the form of protuberances. Thus larvae are found which have still preserved the entire form of the Trochophore, and yet exhibit already the two lateral rows of setae or parapodia. At first only a few Pio. 127.— Poly- trochal larva of Ophryotrocha pue- rilis (after Glapa- lEDK UWD MkTSCH- xikoff). d, intes- tine; fe, jaws. ANNELIDA 279 segments are present ; new ones are, however, continually being interpolated behind. Since, moreover, the parapodia acquire more and more their permanent shape, and the larval organs, on the other hand, degenerate, the larva approaches the shape of the adult animal. The segmental appendages do not arise in a uniform manner in the different divisions of the PolychaBta. In the Errantia the dorsal and ventral parapodia arise from a common fundament, which afterwards separates into the dorsal and ventral parts. This has been observed, for example, in Nereis. Such a separation, however, does not take place in the Sedentaria, but their dorsal hook-bearing segmental appendages arise independently of the ventral parapodia (thus in Terebella). Accordingly it is maintained that only the dorsal appendages of the Sedentaria cor- respond to the common parapodial fundament in the Errantia, whereas the ventral appendages are to be considered as new formations of a secondary nature (Salensky). The cirri and tentacles arise as elevations and evaginations of the ectoderm, into which continuations of the somatic mesoderm may also extend. Of these the unpaired median tentacle, as it occurs, for example, in Terebella, Pileolaria, and Psygmobranchus, usually extending forward beyond the head, is of an especially peculiar nature. It attains at first a large size, and is provided with a considerable cavity (Terebella), but may soon become reduced in size again [Psygmobranchus). When it is present, there are found near it, and on either side of the head, the lateral tentacles, the number and form of which are very variable in the different Annelida. The tentacles may attain a peculiar development by putting forth bud-like evaginations, which enlarge and become the gills. In Psygmobranchug the larva, by means of the trifid gills, acquires quite a peculiar shape (Fig, 128 B). ^ The median tentacle, which was present somewhat earlier and extended forward beyond the head, has already in this stage degenerated. As sometimes the tentacles, so also may the eyes, degenerate in the Sedentaria, since these sessile forms can scarcely have further need of them. In Psygmobranchus a peculiar organ is seen lying behind the gills (Fig. 128 B, kr), which is also developed in other Annelida (Pileolaria, Splrorbis). This is an annular fold of the 1 The larva which is here figured has developed from a so-called monotrochal larva (Fig. 128 A), which exhibits the form of the Trocho- phore, provided with a preoral and post-oral ciliated band, while the preanal band is wanting. The post-oral ciliated band is continued into the ventral (so-called adoral) ciliated groove. The Trochophore already possesses two eye-spots, but still lacks the intestinal canal, which is represented by only an entodermal mass of large cells. The mouth is already indicated. 280 EMBRYOLOGY skin, which grows out backwards, and surrounds like a mantle the part of the body lying behind the head. The first two segments lying behind the head are conspicuous in many Polychaeta {Terehella, Oph- ryntrocha) by the fact that they are destitute of seg- mental appendages ; this fact has caused them to be reckoned as belonging to the head, which is thus sup- posed to arise from several segments. However, the manner of formation of their internal organs (neural and mesodermal), which are begun like those of the body segments, is an argument against this (Salensky). Differences of opinion exist among the different investi- gators concerning the origin of the head itself, for some of these maintain that it is formed from the preoral por- tion of the Trochophore alone ; while, according to others, post-oral parts of the larva also enter into its for- mation. In forms which, like JExogone gemmifera and Gruhea limhata, brood their eggs, the stage of the free larva may be altogether omitted, the embryo breaking through the egg en- velope in the form of the young worm al- ready provided with a number of segments, parapodia, and cirri (Vigcter). It is these conditions which recall those in the Oligochaeta, more especially since one of Fig. 128.— yl, B, larval stages of Fsygmobran- chus protensus (after Salensky). ^.Trochophore with pre- and post-oral ciliated bands and adoral ciliated groove (seen from the ventral surface); m, region of the subsequently formed mouth- openiug; ent, still undifferentiated entodermal mass within ; B, later stage with gills (fc) ; b, fundaments of the setas ; fcr, collar ; and vd, md, ed, fore-, mid-, and hind-gut. ANNELIDA 281 the forms mentioned {Exogone) is said to pass through a stage of development which, according to the description of ViGUiER, strongly resembles the " larva) " of the Oligochseta. 2. Development without Free-swimming Larva (Oligochmta) . The Oligochaeta lay their previously fertilized eggs in firm cocoons, consisting of a chitin-like substance. The cocoons vary greatly in shape in the different genera, and, according to the life-habit of the worm, are found either in the earth or attached to aquatic plants. The slender, spindle-shaped cocoons of Criodrilus attain a considerable length (as much as 5 cm.). In the Lumbricidce they are rounded or ovate, and of different sizes in one and the same species, being about as large as a pea or a bean. Correspondingly the number of eggs which they contain is also variable. Some- times only a very few eggs are found, while in other cases the number may reach as many as twenty or thirty. Usually not all of these eggs develop, but, as appears, some of them develop at the cost of the others. Ordinarily the eggs float in an albuminous mass. Their development is different according as they contain a small amount of food-yolk (Lumbricus, Criodrilus)^ or possess abundant yolk (Rhyn- chelmis, Tuhifex). Cleavage is always unequal, but in the first case an invaginate gastrula is formed, while in the second an epibolic gastrula occurs. Cleavage and the formation of the germ-layers in the Oligochaeta have been thoroughly studied by various investigators (Kowalevsky, No. 27 ; Hatschek, No. 18; Kleinenberg, No. 24; Vejdovsky, No. 45). In Lumbricus a blastula is formed which is thicker on one side, and which may be flattened so that the cleavage cavity is small ; and the gastrula, which soon arises by invagination, is also at first rather flat (Kowalevsky, Fig. 130 A). These characters are less marked in the case of Luvibricus trapezoides (Fig. 129 A), in which occurs the peculiar phenomenon of the division of the embryo in the gastrula stage, producing in this way two embryos, which, separated from each other, develop further. Fig. 129 A represents such a stage of division of an embryo, and shows the two embryos (which are in the same stage of development) only slightly united. 282 EMBRYOLOGY In case the egg is very rich in yolk, as in Rhynchelmis, there arise, according to Vejdovsky, as the result of the first divisions, four blasto- meres, from which four much smaller blastomeres are constricted off, so that now four micromeres and four macromeres are present. While the micromeres increase rapidly by division, the hindmost and largest of the four large cleavage spheres buds off three cells of medium size : the mesomeres. Now the macromeres also divide further ; the micromeres, which, as well as the mesomeres, have meantime increased in number, grow over the latter, which thereby come to lie inside. Between micro- meres and macromeres a small cleavage cavity arises, which is soon obliterated, when the small cells grow over the large ones further, in this way forming an epibolic gastrula. Ordinarily several, nsuallj two, blastomeres are differen- tiated before the formation of the two primary germ-layers, but apparently exhibiting relations to both of them ; these withdraw from connection with the other cells and enter the cleavage cavity (Fig. 129 A). They constitute the fundament C Fig. 129.— J. to C, sections through embryos of Lumhricus trapezoides (after KLBitrKWBRRG, from Balfour's Comparative Embryology). A, horizontal longi- tudinal section of an embryo in the gastrula stage, which is about to divide into two embryos : between ectoderm and entoderm the two large pole cells of the mesoderm (m') can be recognized, with the mesodermal bands (ms) arising from them on either side ; al, archenteron ; B and C, cross-sections of somewhat older embryos, which show how the mesodermal bands (ms) move toward the ventral side, and how the cavity (pp) makes its appearance in them. ANNELIDA 283 of the mesodermal bands. These arise by the division of the two cells, and by the smaller cells thus produced moving away from them. This process can best be understood from the figures of Lumhricns given by Kowalevsky and Klei- NENBERG (Figs. 129 and 130). The two large cells (pole cells) from which the smaller cells of the mesodermal bands have arisen by division are seen in the posterior part of the embryo (comp. the interpretation given by Kleinenberg, p. 286). The mesodermal bands extend on both sides of the embryo towards the mouth ; v^hereas they at first diverge, later they move from the lateral position toward the ventral surface, and now lie on either side of the median line (Fig. 129 2?, C). mes, m£*r. Fio. 130. — A, B, optical longitudinal sections of two embryos of Lumhricun rubellus (Allolohophorafoetida [?], Vejdovsky) of different ages (after Kowalevskt). bt, blastopore ; ect, ectoderm ; ent, entoderm ; m, mouth-opening ; mes, mesodermal bands ; p. pole cells of the mesoderm. In the figures 130 A and B, the mesodermal bands are seen in side view, and their first appearance (Fig. 130 A) can be recognized. In this case they consist from the begin- ning, not of one, but of several, cell-rows ; but even here the pole cell of each band can be seen at the posterior end. The bands extend further and further, and finally acquire the considerable length which is represented in the figures 130 B and 131. Together with the elongation of the mesodermal bands already described, which are also, though inappropriately, called germ bands, the formation of the embryo as a whole has progressed further (Figs. 130 B and 131). It has en- larged by the rapid multiplication of its cells, and now con- 284 EMBRYOLOGY sists substantially of a bilaminar cellular vesicle, between the two layers of which are lodged on the ventral side the mesodermal bands. The blastopore has become the per- manent mouth, in the neighborhood of which a lip-like thickening of the ectoderm makes its appearance. The cells lying around the mouth appear to be of a contractile nature, and accordingly execute swallowing movements, in consequence of which the intestine becomes filled with the albumen in which the embryo floats, and which serves it as food. As a result of this nutrition the embryo becomes more and more distended, and increases in volume. The embryo in this condition may be compared with the free- swimming larvm of other Annelida, especially as it bursts the vitelline membrane at about this stage, and now floats, as has been mentioned, in the albumen of the cocoon. The larva-like appearance of the embryo is increased by the fact that in Lumhricus trapezoides (according to Kleinenberg) there is found a ring of delicate cilia, surrounding the mouth and continuing into a ventral ciliated groove, which extends in the middle line between the mesodermal bands. Hatschek also found an adoral ciliated zone in the embryos of Criodrilus. Farthermore Yejdovsky proved the existence in Rhynchelmis of a paired head kidney, which Bergh also found in Criodrilus. It consists of a long, semicircular tube, the blind inner termination of which lies in the vicinity of the mouth, whereas the external opening of the ciliated canal is situated laterally at about the middle of the body. In view" of all this, the embryos of the Oligochaeta appear to be degenerate larval forms, which float free in the albumen of the cocoon, and here feed independently. The absence of the anus does not enter much into the question, for we see that in many Trochophore larvae also — for example, in Psygmo- branchus — the anus, and even the mouth, may be wanting in the early part of its free existence (comp. Fig. 128, p. 280). The metamorphosis of the larva- like embryos into the worm is accomplished principally as the result of the further development of the mesodermal bands. These at first con- tinue their growth forwards and surround the fore-gut, which has been formed out of an invagination of the ecto- ANNELIDA 285 derm (Figs. 130 B and 131). Also in the parts of the embryo lying further backwards the mesodermal bands advance from the ventral side, to which they were at first confined, to the dorsal side, and thus separate the ectoderm from the entoderm. The separation of the mesodermal bands into primitive segments and the splitting of these into somatic and splanchnic layers occurred even before this (Figs. 129 G and 131). The posterior part of the embryo ..][nivs wg i„^ "liiiiii.iiii t Y^ 6'fv t»a Pio. 131.— Optical longitudinal section of an embryo of Lumhricus olidus (after WiLsow, from Lang's Lehrhuch). hm, fundament of the ventral nerve cord; e, ectoderm ; en, entoderm ; g, fundament of the supra-oesophageal ganglion ; kh, head cavity ; m, mesodermal bands ; md, cavity of the mid-gut ; nh, neuroblast cells ; 0, mouth ; pm, parietal (somatic) layer of the primitive segments ; pms, pole cells of the mesoderm ; sh, cavity of the primitive segments ; st, stomodseum (fundament of the fore-gut) ; ug, suucesophageal ganglion; vm, visceral (splanchnic) layer of the primitive segments. is greatly distended by the albumen taken into the intestine, and bulges out like a yolk-sac on the embryo, which in the meanwhile has grown longer (Fig. 132 h). Also in the posterior distended part of the embryo the primitive seg- ments are ultimately formed and grow around the ento- dermal sac towards the dorsal side, so that finally it is entirely surrounded by mesoderm ; thus the most im- portant parts in the development of the worm, as far as regards its outer form, are completed. The anus is formed later by an ectodermal depression at the posterior end of the worm. 286 EMBRYOLOGY 064. Both the origin and the further development of the mesoderm are dis- puted points in the development. In some cases, as, for example, in the LttffiftricMS studied by Kowalevsky (and in Nereis, according to Goette), it appears as if the first mesoderm cells had been derived from the ento- derm cells, whereas in other cases they seem rather to have belonged to the ectoderm. Usually their origin cannot be referred to either one or the other of the two primary germ-layers, for they were established before the formation of these, or on the border-line of the two. Such is the case in various Oligochaeta and also in Polycheeta. In Rhynchelmis (comp. p. 282) the so-called mesomeres are separated off from the large blastomeres, which subsequently become the entoderm, and, to- gether with the micromeres, overlie these ; apparently therefore they belong to the ectoderm. It is only later that they are overgrown by the ectoderm, and move to the inside here to develop into the mesoder- mal bands (Vejdovsky). Just as the opinions of authors are divided in regard to the deri- vation of the mesoderm, so also are they in regard to the manner in which the mesodermal bands are formed. Whereas some authors derive them from proliferations of the primitive mesoderm cells (Kow- alevsky, No. 27; Hatschek, No. 18 ; Goette, No. 15), others are of the opinion that the parts of the ectoderm which lie over the meso- dermal bands also supply cells for the reinforcement of these bands, and that as a result ectoderm and mesoderm are in this region con- tinuous (Fig. 129 B). Kleinenberg (No. 24) thus describes the condi- tions in Lumbricus trapezoides. Salensky (No. 87) agrees with him. Kecently Kleinenberg (No. 26) has gone still further, for he considers that the entire mesoderm — the existence of which as a separate layer he, moreover, denies — has been gradually split off from the ectoderm. This point will be referred to again in considering the organogeny. \. Fig. 132. — An embryo of Lumbricus agricola, already far advanced in deve- lopment (after Kowalevsky). ?i, pos- terior part of the embryo, resembling a yolk-sac ; its wall is formed of ecto- derm and entoderm, and it gradually becomes overgrown by the mesodermal bands ; mes, upper limit of the left mesodermal band ; ces, cBsophagus. 3. The Formation of the Organs. So much of the formation of the individual organs as has not been considered in the two preceding sections upon the general form of the body will be added here. However, it ANNELIDA 287 should be noted at the beginning that upon these matters there prevails as yet among authors but little clearness, and agreement to only a limited extent. Ectodermal Structures. The epidermis of the larva and of the adult worm arises directly from the embryonal ectoderm, its cells multiplying greatly, and becoming much flattened. The setigerous sacs arise, according to the concurrent state- ments of KowALEVSKY, Vejdovsky, and Bergh, as club-shaped ingrowths of the epidermis, inside of which the setse are secreted. According to other descriptions, the setigerous sacs originate from the mesoderm. Nervous System and Sensory Organs. — In considering the origin of the nervous system it seems necessary to separate the supra-oesophageal ganglion from the ventral cord. Both arise as thickenings of the ectoderm (Fig. 133 C), the ven- tral chain of ganglia either as a longitudinal, unpaired or as a paired thickening, which detaches itself from the ectoderm, and moves to the inside, where it may be sur- rounded by mesoderm (Kowalevsky). The further develop- ment proceeds from in front backwards. Opinions are very far apart regarding the origin of the supra-oesophageal ganglion, and especially its connection with the ventral chain of ganglia. In Hatschek's opinion, there first arises an ectodermal thickening at the head end of the embryo : the apical plate. From this the ectodermal thickening progresses backwards in the form of two cords, which extend on either side of the mouth. From the oesophageal connectives thus formed, the thickening process continues further and further. In this way the two lateral cords of the ventral nerve-trunk are formed, and in addition a groove-like invagination, lying in the longitudinal median line (similar to the medullary tube of vertebrates), takes part in the formation of the ventral chain of ganglia. Hatschek defends the view that the entire nervous system arises from a single fundament. In this he relies mainly upon his embryological investigations on Criodrilus and Polygordius, and furthermore on the comparative anatomical conditions in Protodrilus, in which Archiannelid the oesophageal connectives are said to remain throughout life in connection with the body epithelium as ectodermal thickenings. 288 EMBRYOLOGY In opposition to this theory, Kleinenberg — with whom Goette, Salensky, Bergh, and Fraipont agree — espouses the view that the nerv- ous system is composed of two separate fundaments. The supra-oeso- phageal ganglion arises as a preoral ectodermal thickening independ- ently of the two longitudinal thickenings of the ventral side, which represent the fundament of the ventral nerve cord. (A longitudinal furrow corresponding to the medullary tube of the vertebrates does not exist.) Later it puts forth lateral processes, the oesophageal con- nectives, which unite with the already-formed ventral nerve cord. Such is the condition in Luvibricus. The origin of the nervous system in Lopadorhynchus, likewise studied by Kleinenberg, depends upon much more complicated formative processes. Lopadorhynchus develops from a monotrochal larva, the posterior portion of which grows out into the worm in the manner already described. A ciliated pit, the so- called apical organ (Fig. 135, p. 293), and the two apical tentacles arise in the vicinity of the apical pole as provisional sense-organs. Behind these the two pairs of permanent antennae and the olfactory pits are formed, also as ectodermal growths. From these organs, which later degenerate in part, the formation of the supra-oesophageal ganglion proceeds. Ordinary ectodermal cells are metamorphosed into ganglionic cells, which accumulate in the region of these organs, later sink in deeper, and unite to form the supra-oesophageal ganglion. This finally detaches itself from connection with the ectoderm and appears inside the body as an independent organ. Just as the formation of the supra-oesophageal ganglion, according to Kleinenberg, starts from the sensory organs, so the origin of the ventral nerve cord is also referred by this investigator in part to the influence of the sensory organs. In the main, however, the impetus to its formation proceeds from the locomotor organs. From the inner [deep] side of the ventral ectoderm, the outer surface of which bears tufts of sensory hairs (Fig. 134 C), a plate is separated off, the so-called neural plate, in which a right and left portion can be distinguished (Fig. 134, p. 292). Along this plate arise segmental, paired ingrowths, the setigerous sacs (Fig. 134 C). Dorsad and ventrad from these are formed as ectodermal ingrowths the dorsal and ventral cirri. The parts of the neural plate situated nearest to the median plane supply the ventral cord. They approach more and more the middle line, and here fuse with each other. The union of the ventral nerve cord with the supra-oesophageal gan- glion is secondary. It is brought about by the neural plates extending forward and sending out processes to the ring- nerve of the ciliated band. But processes of the supra-oesophageal ganglion also pass into this, and in this way the oesophageal connectives arise, whereas the ring-nerve itself, together with the ciliated band, disappears. Therefore, according to Kleinenberg's description, here reproduced briefly, the brain and ventral cord appear to have a separate origin, the impulse to which comes through sensory and locomotor organs. ANNELIDA 28,9 The origin of the sensory organs has already been touched upon several times, as, for example, the formation of the auditory vesicles in Eupomatus (comp. p. 267). The eyes of the Alciopidoe are formed, according to Kleinenberu, as invaginations of the ectoderm, which are constricted off and unite with the brain, this union constituting the optic nerve. The retina arises as the result of the differentiation of the inner [deep] wall of the vesicle, whereas the outer wall becomes very thin. Within, the lens and, by the activity of certain gland-like cells, the vitreous body are secreted. Mesodermal Structures. Body Cavity ; Musculature ; Blood-vessels. — The differentia- tion of the mesodermal bands, which results in the formation of the segmental cavities, and thereby causes the segmenta- tion of the body, takes place in a simple manner. The meso- dermal bands have extended forward and in the anterior part of the body become several rows wide and several layers deep. Then a segmentation begins at their most anterior end, individual parts becoming differentiated in groups, and finally separated from one another by transverse boundaries (Fig. 131, p. 285). These box-like, quadrangular cell-plates, which succeed one another along the course of the meso- dermal bands, and therefore lie side by side in two rows, are the primitive segments, the influence of which on the segmentation of the body we have already briefly men- tioned in considering the development of Polygordius and Lumhricus (pp. 268 and 284 — 286). We saw there also that the development of the primitive segments takes place from in front backwards. When the primitive segments are already well formed in the anterior part of the body, the mesodermal bands are still entirely undifferentiated in their posterior portions, and new cell material continues to be formed here in the vicinity of the primitive mesoderm cells (Fig. 131). A fissure soon makes its appearance in the primitive segments, owing to the fact that the two or more cell-layers out of which they are composed separate from each other at the middle of each primitive segment (Fig. 133 B and G). Thus the segmental cavity — that is to say, the be- K. H. e. U 290 EMBRYOLOGY efinning of the body cavity — is formed in eacli segment of the body of the worm. The cavity enlarges while the walls of the primitive segments are more and more distended and apply themselves to the body-wall and to the wall of the intestine as the somatic and splanchnic layers respectively (Fig. 133 C). But of course every two of the segments abut on each other with their anterior and posterior walls, and thus arise the septa (dissepiments), which separate the different segments of the body. Since each segment of the body requires for its formation a primitive segment on the right side and one on the left, there are formed a dorsal and a ventral mesentery (Fig. 183 (7). These mesenteries dis- til Fig. 133. — A to C, transverse sections of Polygordius larvae (after HiTSCHBK). A, optical cross-section of the body of an unsegmented larva, immediately in front of the anus, showing the two primitive mesoderm cells (mes) ; B, C, two cross-sections of an older larva, the former from the posterior, the latter from the anterior, part of the body ; ect, ectoderm ; ent, entoderm ; mes, mesoderm ; u, fundament of the nervous system ; so, somatic, sp, splanchnic layer of the meso- derm. appear in most of the Chaetopods (just as the septa also are frequently perforated), but they persist in some of the lowest Annelids, such as Polygordius among the Archiannelida and Saccocirrus among the Chsetopoda. The body musculature arises from the somatic layer of the primitive segments, the ventral longitudinal muscles being the first to be formed. By the arrangement of the ANNELIDA 291 masculature its segmental origin can be recognize 1 even in the fully developed animal. The peritoneal epithelium is also derived from the primitive segments. The splanchnic layer of the mesoderm produces so much of the wall of the intestine as is not of entodermal or of ectodermal origin, and the walls of the 'vessels also arise from it. According to Salensky, the formation of the blood- vascular system begins (in Psygmobranchus and Terebella) in the form of canals, which lie between the entoderm and the splanchnic layer, and which are therefore really parts of the segmentation cavity. Later these blood-filled cavities are surrounded with a cellular wall, which comes from the splanchnic layer. According to Kowalevsky's observations also, mesodermal cells, which collect between the ectoderm and splanchnic layer, form the walls of the vessels ; more- over, the dorsal vessel (Lmnbricus and Criodrilus, according to Vejdovsky) arises from separate paired fundaments. These extend along the boundaries of the mesodermal bands as they grow dorsad, and, advancing with them, finally fuse with each other to form the dorsal vessel. This con- dition is of especial interest from the fact that in Pleurochceia (Megascolex) the separate fundaments of the dorsal vessel are retained in certain parts of the body throughout life (Beddard). It appears questionable whether the cephalic cavity is formed in the same way as the segmental cavities of the body, or whether it is to be distinguished from these. On the first assumption, the two most anterior primitive segments would unite for its formation, and therefore outgrowths of the mesodermal bands must have crowded forward past the oesophagus even into the head region. The outer wall of the head and the muscula- ture of the oesophagus would then be formed from the somatic and splanchnic layers respectively in the usual way. The conditions were described in this way by Kleinenberg in his earlier work, and Vejdovsky also derives the cephalic cavity from the two " anterior ends " of the mesodermal bands, which he describes as being fused {Rhynchelmis). Opposed to this is the view, supported especially by the free-living larvee, that the cephalic cavity arises by means of a separation of the two primary germ-layers and by an immigration of mesodermal elements from the trunk (Hatschek). According to this explanation, the first pair of primitive segments lies behind the head, and the mesoderm of 292 EMBRYOLOGY the head arises from one wall, or from the still undifferentiated meso- dermal bands. The absence of the mesenteries in the head of Pohj- gordius supports the view that the cephalic cavity is not paired, but unpaired, in its origin (Hatschek). The difference between head cavity and body cavity vanishes when — as is the case, according to Kleinenberg, in many Annelids, for instance Fig. 134.— il to C, parts of frontal loncjitudinal sections of the larva of iopod- orJiyncJius, showing the splitting off of the muscle-plates (after Klkinkn bebg). bs, fundaments of setigerous sacs ;ect, ectoderm; ent, entoderm; g, ganglionic cells of the larval nervous system; mp, muscle-plates; np, neural plates; », larval sensory organs. in Lopadorhynchus — a regular splitting of the mesoderm into a somatic and a splanchnic layer does not take place, but the investment of the entoderm is effected by separate cells detached from the mesodermal ANNELIDA 293 The body cavity of the trunk in such cases therefore does not represent the cavity between the two layers of the mesoderm, but corre- sponds to the blastoccele (traversed by mesodermal cells), just as the cephalic cavity does in the case mentioned above. Moreover, in Lopad- orhyuchus this also arises by an immigration of mesoderm cells into the head portion. From this description it follows that the formation of the body cavity does not always take place in so regular a manner as has been described above ; in fact, according to Kleinenberg's statements, the formation of the entire mesoderm may be effected in another manner. It has already been mentioned (p. 286) that in Kleinenberg's opinion the ectoderm, in addition to the pole cells, takes part in the formation of the mesodermal bands of Lumhricus trapezoides. Cells are separated off from the ectoderm and added to the germ bands lying under them. In Pig. 135— Sagittal section of a larva of Lopadorhynchus (after Kleinenbebg)/ d, intestine ; mp, muscle-plate ; np, neural plate ; ces, fundament of the (permanent) oesophagus ; sg, fundament of the supra-oesophageal ganglion ; so, apical organ ; st, stomodseum (temporary fore-gut of the larva) ; w, preoral band of cilia. Lopadorhynchus Kleinenbero derives the entire mesoderm from the ectoderm. According to him, mesoderm does not exist as a separate layer. The musculature of the Lopadorhynchus larva arises by means of an emigration of cells from the ectoderm (Fig. 134 A to C). The so- called muscle-plates (Fig. 135 mp) are formed by a splitting of the thickened ventral ectoderm, first at the hind end of the larva and then successively further forwards. The course of this cell growth, leading to the formation of the muscle-plates, is evident from the figures 134 A to C. The muscle-plates of the two sides are separated by a fold of the ento- derm. The segmentation of the muscle-plates takes place after the fundaments of the setigerous sacs have grown in from the neural plates (comp. p. 287 and Fig. 134 C). The limits of the segments arise by the 294 EMBRYOLOGY loosening of the texture of the tissue in successive planes at right angles to the long axis of the body (Fig. 135). As was mentioned, individual cells separate from the muscle-plates, in order to apply themselves (like the splanchnic layer) to the intestine, whereas the remaining part of the muscle-plates supplies the musculature and epithelium of the body-wall. Blood-vessels and segmental organs were not observed in Lopadorhynchus. According to Kleinenberg's explanation, with which, as regards the ecto- dermal origin of the mesoderm, Salensky also concurs, the primitive mesoderm cells occurring in other Annelids must be looked upon only as early differentiations of ectodermal parts. But when several organs of altogether different kinds, such as the musculature, the blood-vascular and excretory systems, can be referred back to such a common funda- ment, then the theory which considers this fundament as a germ-layer is not unwarranted, even when the fundament at times, as in Lopad- urhynchus, makes its appearance in somewhat later stages and in a less primitive manner, namely by the splitting off of a cell-layer from one of the two primary germ-layers. Wilson's view should also be mentioned here, according to which, in addition to the two pole cells from which the mesodermal bands arise, three other pairs of similar pole cells are present on the ventral side of the embryos of Lumbricvs. The three large cells mentioned, from each of which a row of cells extends toward the anterior end of the embryo, lie on either side of the middle line somewhat farther forward than the pole cells of the mesoderm and somewhat more superficial, therefore more in the region of the ectoderm. The innermost of these three rows is said to constitute the fundament of the nervous system, and the middle one that of the nephridia, whereas the significance of the outer one re- mained unknown to the author of this theory.^ The formation of the mesoderm and body cavity in Enchytraoides takes place, according to Koule, in a very singular manner, as far as can be judged from his brief communication. In the "morula," which re- sults from an irregular cleavage, an outer layer, the ectoderm, is split off from a central mass, while the latter separates by means of a similar process into the centrally situated entoderm and the surrounding meso- derm. The former (entoderm) by the appearance of a cavity in it becomes the intestine, while spaces are formed in the mesoderm, which become confluent, and thus give rise to the body cavity between the two layers * [The conditions of the mesoderm have been further elucidated by the recent works of Wilson (Appendix to Literature on Annelida, Nos. XXIV. and XXV.) and Bergh (No. VII.). The origin of the pole cells and their relations to the organs have been followed out further. As a result more definite relations of those parts which were previously held to be exclu- sively mesodermal have been disclosed. A portion of the body muscula- ture appears to be of ectodermal origin. On this point the papers of Bergh and Wilson should be consulted. — K.] ANNELIDA ' 295 I ^^m cells from the somatic layer, which unite with the splanchnic layer 1 Head Kidney and Segmental Organs. — In the larva of Eupomatus the head kidney arises by the outgrowth of a cell lying in front of each pole cell of the mesodermal bands. Some other mesodermal cells take part in their formation, for they supply the spheroidal cells which rest upon the inner blind end of the head kidney and elongate into the ligament-like threads of attachment of the organ (Fig. 118 B, p. 265). In addition to the formation of the head kidney, the few cells of which the mesodermal bands at first con- sist are further employed for the formation of the larval muscles. Only the two pole cells remain. These, by re- peated division, supply the new mesodermal bands, which Hatschek designates as secondary in contrast to those primary ones which were early put to use. Later they reach the great development which has already been described. According to Hatschek, the remaining segmental organs originate from the head kidney, for a small ciliated canal (in Polygordius), running in the somatic layer of the meso- derm, branches oU from each head kidney at the junction of its two arms. The nephridia are said to be given off from this canal, one to each segment (Fig. 136). While the head kidney degenerates, they reach their final development. Hatschek's description has met with little recognition, for it could not be substantiated by subsequent investigators (Fraipont). However, the discoveries recently made by E. Meyer on certain Terebellidae {Lanice, Loimia) show the observations of Hatschek in a new light. In the two Annelids mentioned the nephridia are united by means of a [longitudinal, blindly ending, common] duct, which extends far backward. The dis- charge takes place through as many [successive lateral] canals as there are nephridia present, but these canals are connected with the nephridia only indirectly, i.e. by means of the common duct. In the Capitellida also, according to Eisig, connections between the different nephridia exist in the form of ciliated canals. We do not intend to assert that any great importance is to be ascribed to these conditions, for in the first place their development is not known, and then the nephridial system of the Terebellidae (and CapitellidsB) is shown to be essentially modified. 296 EMBRYOLOGY A connection between all the segmental organs is found by Hatschek in Criodrilus also, for they are said to arise from a cord-like thickening of the somatic layer, which extends the whole length of the body dorsal to the ventral longitudinal muscles. These cords are then separated segmentally into loop-like parts, the fundaments of the nephridia. The latter acquire lumina, and open into the body cavity in front of the segment to which they belong through the future funnels, and finally fuse at their posterior ends with the ectodermal wall to form the external openings. The funnels and nephridial ducts arise separately. But JliV AjZr 4^ -J^ / Fig. 136.— Diagram of the development of the excretory system of Polygordius (after Hatschek, from Balfoue's Comparative Embryology). these statements of Hatschek also find opponents in Vejdovskt and Bergh, according to whom the segmental organs of the Oligochseta arise from separate fundaments by the growth of the cells in the somatic layer and in the partition walls of the primitive segments. Fig. 137 A shows that at the boundary of the septum and somatic layer there is a considerably enlarged cell (tz). It con- tributes especially to the formation of the funnel. Behind it other cells of the somatic layer arrange themselves into a cord of cells (Fig. 137 B), which constitutes the funda- ment of the nephridial duct. In it, as also in the funnel, a lumen makes its appearance later ; the entire structure ANNELIDA 297 becomes covered with peritoneum (Fig. 137 (7, pt), and presses its way out towards the ectoderm, in order to fuse with it directly, or with an invagination of it (Bergh, No. 7), which forms the terminal portion of the duct, or collective vesicle, when such is present (Yejdovsky, No. 43). According to the observations of E. Meyer (No. 31), the thoracic nephridia of Psygmobranchus are composed of separate parts. The nephridial ducts arise from large mesoderm cells, which are found in the blastocoele of the larva ; the funnels, on the contrary, and the peritoneal covering of the nephridia are supplied later by the primitive segments. The ends of the nephridial tubes open to the exterior by means of provisional pores, which later occupy the floor of a ciliated groove, which closes, and represents the unpaired ectodermal efferent duct of these two so peculiarly con- stituted nephridia of Psygmo- branchus. Genital Organs. — The development of the sexual glands is very simple in both the Polychceta and Oligochceta. They arise as growths of the peritoneal epithelium on the septa, or, as frequently in the Polychceta, on the invest- ment of the blood-vessels. The genital gland, which in Lumhricus is distin- guishable even during cocoon life (Bergh), sepa- rates from the peritoneum as the result of a rapid prolifera- tion of the cells, and gradually assumes its permanent form (Fig. 138 A to jD, after E. Meyer). The genital products are Fig. 137.—^ to C, parts of longitudinal sections through embryos of Crwdrilus, showing the developmetit of the nephridia (after Bebgh). ed, ectoderm ; 7i, cavity of primitive segment ; pt, peritoneum (of the nephridia) ; s, septa ; so, somatic, sp, splanchnic layer of the mesoderm; t, funnel ; tz, funnel cell. 298 EMBRYOLOGY liberated one hy one (Fig. 138 D), and either undergo their further development while floating free in the body cavity, or, as in the case of the testicular cells of the earthworms, reach special vesicles (vesiculae seminales), which, according to Bergh (No. 5), arise on the septa by means of a process of growth and invagination. The ducts of the sexual organs are to be looked upon as more or less modified segmental organs. They arise in the Fig. 138. — A to D, diagrammatic representation of the structure and development of an ovary of Amphitrite rubra (after E. Meybh). g.dr, sexual glami ; g.e, genital epithelium; gf.z, genital cells in the act of breaking away; pm, peritoneum; V.v, vas ventrale. same way as the nephridia themselves, except that the funnel is formed earlier than in the actual segmental organs (Vejdovsky). The entirely independent origin of the efferent sexual ducts from the nephridia and the simultaneous occur- rence of both organs in the same segment, as happens in the earthworms, form no argument against the origin of the ducts from nephridia, since in some Annelida {Gapitellidcej according to Eisig) several pairs of nephridia occur in the same segment. In the earthworms especially, many things ANNELIDA 299 seem to suggest that two pairs of nephridia originally be- longed to each segment (Benham). The metamorphosed segmental organs function in the following manner : the funnel takes up the genital products out of the body cavity, the nephridial ducts pass them onward, and the part dis- tended into a terminal vesicle serves as a genital atrium. But the terminal portion in the male apparatus of the Oligochaeta may be metamorphosed into an evertible copu- latory organ (thus in Stylodrilus, No. 43). The receptacula seminis have also been traced to nephridia, of which only the ectodermal vesicular end portion is assumed to develop ; but Bergh prefers to consider them metamor- phosed dermal glands. They arise as tubular invaginations of the epidermis into the interior of the body cavity, and are surrounded by the other layers of the body- wall (Vejdovsky, No. 43; Bergh, No. 5). Entodermal Structures. Intestinal Canal. — In the Polychaeta and Oligochaeta we saw that the intestine arises from portions of all three germ- layers. The permanent mouth is generally found at the place of the blastopore, a depression of the ectoderm taking place here, so that the fore-gut (just like the hind-gut, which arises later) is an ectodermal structure. In those cases in which the larva arises from an epibolic gastrula, and the blastopore does not become the mouth, as in Rhynchelmis and Psygmobranchus, the fundament of the intestine at first consists of a solid entodermal mass rich in yolk (Fig. 128 A and B, p. 280). The entodermal wall of the mid-gut, by means of which the yolk-mass that still remains is absorbed, arises by the disintegration of the more central cells, while smaller cells at the periphery with less food-yolk are separated from the rest of the mass and form an epithelium. In this condition the intestine consists of a sac closed on all sides. The fore- and hind-guts are formed by its union with the ectoderm in front and behind. The share which the two ectodermal invaginations take in the forma- tion of the fore- and hind-guts is said to be very variable in the different 300 EMBRYOLOGY Annelida (Salensky). Thus the oesophagus may be formed of ectoderm {Pileolaria, Lumbrlcus), but it is maintained that it may also be for the greater part of entodermal origin [Psygmohranchus, Rhynchelmis). The conditions in Lopadoi'hynchus are peculiar ; here the wide ciliated stomodaeum (the larval fore-gut) is not directly converted into the oesophagus, but constitutes a transitory structure. On the wall of the stomodaeum two cushion-like thickenings make their appearance, which become hollow, and form two small sacs (Figs. 139 and 135 oes). These become considerably enlarged, surround the stomodaeum, and finally grow together, after the stomodaeum has closed and separated from the ectoderm. The detached stomodaeum is now seen as a ciliated sac, surrounded by the likewise saccular oesophagus. This finally unites permanently with the ectoderm and entoderm (Kleinenberg). ocs. Fig. 139(135). — Sagittal section of a larva of Lopadorhynchus (after Kleinenbebg). d, intestine ; mp, muscle-plate ; np, neural plate ; ces, fundament of the supra-oeso- phageal ganglion; so, apical organ ; at, stomodaeum (fore-gut of the larva); tc, pre- oral ciliated band. The nmsculature and peritoneal covering of the intestine are supplied by the splanchnic layer of the mesoderm. The chlorogogenous cells which surround the intestine are con- sidered as excretory organs, and have the same origin and significance as the so-called pericardial glands occurring on the blood-vascular system, and are also outgrowths of the same layer (Grobben). In the Lumbricidse the typhlosole arises in the dorsal median line of the intestine as a more or less deep groove-like infolding of its entire wall. ANNELIDA 301 4. Non-sexual Reproduction Generations. Alternation of The Cheetopoda possess to a high degree the power of restoring parts of the body that have been lost ; not only the less important parts of the body, but also the more important ones, such as the head region, includ- ing mouth and brain, can be formed anew by them. This power of regeneration passes into a kind of non-sexual reproduction (schizogeny), when the body, as in Lumbriculus, separates spontaneously into several pieces, each of which is able to regenerate itself into a perfect worm. Approaching closely to this is the reproduction of one of the marine Annelida, closely allied to the Oligochaeta, which is found actively multi- plying throughout the year, without at any time developing genital organs {Ctenodrilus monostylos, according to Count Zeppelin). This worm repro- duces in the most primitive way : a constriction is formed on the body immediately behind a septum, and be- comes deeper and deeper until the worm falls into two parts (Fig. 140 A). Of the two resulting parts the anterior is thus without an anus, the posterior without a head. This primitive kind of division may go so far that parts arise which are destitute of both head and anus, and at times con- sist of only a single segment (Fig. 140 B), Head and terminal parts are formed by the thickening of the integument (hypodermis), which sends inward plug- like ingrowths that unite with the intestine. In this way the mouth and anus arise. The new segments that are to be formed are interpolated between the newly formed anus and the preceding segment. Less primitive is the condition in another worm belonging to the same genus, Ctenodrilus pardalis, which likewise was found only in a state of non-sexual reproduction (v. Kennel). In this case thickened zones, corresponding to the anterior and posterior ends of the worms about to be produced (Fig. 140 C), are formed— even here in a simple way, it is true— before the division; that is to say, a so-called budding zone is Fig. liO.—A, Ctenodrilus monostylos dividing transversely (after Count Zeppelin). B, a por- tion of the same worm consisting of only a single segment ; c, cirrus ; d, intestine. C, Cteno- drilus pardalis (after v. Kennel); fcn, budding zone, where the worm later separates into the different parts ; d, intestine. 302 EMBRYOLOGY found at this point, where the division is about to take place. In Ctenodrilus pardalis each portion includes only one segment, and thus the budding zones are seen to be repeated segmentally. While they still remain united to one another, the cephalic lobe [prostomium] and the brain, as well as the oral and anal invaginations, are developed on the several portions. The degree of development in which the zones are found increases from behind forwards (Fig. 140 C). Some Polychffita and those Oligochseta in which non-sexual reproduc- tion is known are like Ctenodrilus in so far as they also divide in a condition in which no genital organs are present. In the Protula described by Huxley a budding zone arises between the sixteenth and seventeenth segments ; then follows the formation of the prostomium of a new individual in the seventeenth segment. In this case, however, after the separation both individuals become sexually mature. The conditions are similar in the Naidida, in which they were thoroughly studied by Semper. These worms also reproduce by division in the sexually immature condition only. The body of the worm may be first separated by a budding zone into two regions ; then new budding zones are interpolated ; that is to say, fundaments of younger animals arise in the individuals already established. This process is kept up, not, how- ever, serially from in front backwards, but in such a way that individuals of quite different ages come to lie one behind another. When the chain has reached a certain stage of development, it separates into the different individuals, which now reach their final shape by growing considerably, by increasing the number of their segments, and by maturing their sexual organs. The cases last considered were, it is true, those of animals without sexual organs, which multiplied non -sexually ; finally, however, all in- dividuals acquire sexual maturity, and are not distinguishable from one another in shape. The conditions are different in those Polychaeta in which new individuals, which become sexually mature, are continually being separated off from the hinder part of the body of an individual that remains [sexually] sterile, a process that is to be placed alongside that of strobilization in the Scyphomedusas. Thus in Autolytus (according to Krohn and Agassiz) there are formed, by the budding of the parent, male and female animals, which lie in a chain one behind the other ; of these the most anterior, the one lying nearest to the parent animal, is the youngest. They separate from the chain according to their ages. The sexually mature animals are essentially different in shape from the budding forms, so that the two were formerly assigned to different species. The sexual animals appear to copulate ; for in a brood-pouch of the female the eggs develop into the worm which subsequently repro- duces by budding. This is therefore a genuine alternation of generations. Similar conditions of reproduction are found in some Syllidce, from the budding individuals of which are detached sexual animals, which, by the great development of their parapodia and setae and by their well- ANNELIDA 303 developed orienting apparatus, are especially well adapted to a free life. They swarm about and secure the necessary distribution of the sexual products, whereas the less active budding form remains at the bottom of the sea. The extensive development of the parapodia is completed even while the buds are still connected with the parent. This recalls the condition of certain NereidcB, in which new setae, better adapted to swimming, make their appearance on the posterior part of the body at Fig. 141. — Part of a stock of Syllis ramosa (somewhat diagrammatic, after McIntosh, and from a preparation of the Challenger material), d, intestine, which branches throughout the entire stock. The stock is injured in some places. the time of sexual maturity (epitokal form) ; these give to the sexually mature animals an entirely different appearance from that of the young form (so-called atokal form), so that here also the sexually mature and the young forms were assigned to different species and genera. In the 304 EMBRYOLOGY Nereida, however, the hinder part of the body, thus equipped, does not become detached, but its better equipment serves only to facilitate the locomotion of the sexually mature animal (Ehlers). But in any event the conditions existing in the Nereidce and Syllidce are comparable with each other. Under the influence of special conditions of life the reproduction of the Syllidce may assume a very peculiar form. In the sponge, Aulochone, and other Hexactinellids a Syllis has been found (McIntosh), on which new individuals arise, not only one after another, but also by lateral budding {Syllis ramosa, Fig. 141). A genuine stock is formed in this in- stance, the branches of which extend without limit in the canal system of the sponge, for the branches in turn have the power of producing new buds. These detach themselves from the stock as male and female sexual animals (Fig. 142) ; and since they are provided with better swimming apparatus, and with especially well developed eyes, one can readily as- sume that they abandon the sponge, and, swarming about free in the sea, secure a wider distribution of the sexual products. Their descendants, which they produce by sexual means, then migrate back, it is to be as- sumed, into sponges. In this in- teresting case the alternation of generations combined with stock-formation is particularly evident. Fig. 142.— Anterior part of a female individual, such as is found in the sponge stocks inhabited by Syllis ramosa. The animal is filled with eggs. The large eyes can be recog- nized on the head (after McIktosh). II. ECHIURID>qE. JSchiurus ; Thalassema ; BonelUa. While BonelUa viridis lays its eggs in the form of a thick, tortuous cord, consisting of a gelatinous mass, in which the eggs lie in several rows (Spengel), the Thalaasema mellita observed by Conn discharges its eggs and spermatozoa free in the sea, so that with this animal artificial fertilization could be undertaken. 1. Cleavage and Formation of the Germ-layers. The cleavage of the egg was closely studied by Spengel in BonelUa. An animal portion of the egg, consisting of finely ANNELIDA 305 granular protoplasm, can be distinguished from a yolk-laden vegetative pole. The cleavage corresponds with this condi- tion. At first the egg is divided into four large spheres by cleavage along two meridional planes, but then four small blastomeres are constricted off from these at the animal pole. By division of the latter and the formation of new micromeres from the macromeres, the number of the small cleavage spheres rapidly increases ; they spread out over the four large spheres, and finally envelop them completely, producing an epibolic gastrula (Fig. 143 A). The small cells that are now being constricted off from the macromeres no longer reach the surface, but remain lying under the layer of micromeres. They form the entoderm. Inside of these the four macro- meres are still retained for a while. In the region of the Fig. 14S.—A and B, embryos of Bonellia (after Spengbl, from Balfour's Com- parative Embryology). ^, epibolic gastrula; £, formation of the mesoderm ; hi, blastopore ; ep, ectoderm ; me, mesoderm. blastopore there appears a layer of cells (Fig. 143 B), which surrounds the blastopore in the form of a circle. Spengel believes this to have arisen by a migration inward of the micromere layer. He interprets it as the fundament of the mesoderm (Fig. 143 B, me). Cleavage and the formation of the germ-layers in Thalas- sema (Kowalevsky, Conn) do not take place in the same manner as in Bonellia. In Thalassema cleavage is equal, and its result is a blastula from which an invagination gastrula arises. To be sure, the latter is not altogether typically ex- pressed in Thalassema mellita, for the invagination merges K. H.E. X 306 EMBRYOLOGY into an ingrowth of cells (Fig. 144 A). As in the later stages of the epibolic gastrula of Bonellia, here also the entoderm consists of a solid mass with an outer diiferentiated cell-layer and an inner yolk-mass (Fig. 144 A). 2. Larval Form and Metamorphosis of Echiurus and Thalassema. The EchiuridsB possess free-swimming larvae, which exhibit more or less clearly the form of the Trochophore. The development of the gastrula into the Trochophore takes place in Thalassema by the appearance of a thickening of the ecto- derm at the upper pole, the fundament of the apical plate. This place becomes covered with a tuft of cilia in the same way as in the larvae of the Polychaeta. At an early stage cilia make their appearance on the outer surface of the embryo, and these are said to traverse, as in Eupomatus, the egg-membrane, so that the latter would become the cuticula of the larva (Conn). Even during the gastrula stage a circle of long cilia makes its appearance at the equatorial circum- ference of the larva (Fig. 144 A). Below it lies the blasto- pore. The cavity of the intestine arises as the result of the absorption of the central yolk-mass by the rapidly multiply- ing entoderm ceils. It becomes connected with the outside world by means of the mouth, which is formed at the place of the blastopore. By the outgrowth of the hinder part of the larva the mouth comes to lie more on one side (on the future ventral surface) immediately under the band of cilia. The latter is then differentiated into a row of cilia lying in front of the mouth and one lying behind it (Fig. 144 B). In addi- tion a ciliation makes its appearance in the middle line of the ventral surface. The intestine grows so much in length that it lies in loops. This is particularly true of the anterior part. The terminal portion fuses at the hind end of the larva with the ectoderm, thereby giving rise to the anus (Fig. 144 B). Fore- and hind-guts, according to Conn, are entodermal formations (?). Mesenchymatous muscle-cells extend be- tween the ectoderm and entoderm of the larva, and at the hind end lie two band-like cell complexes, the mesodermal ANNELIDA 307 r ■Brands (Fig. 144 B, mes). Consequently the larva of Thalas- sema possesses the greatest similarity to the Trochophore i^Hof the other Annelida. The same applies to the larva {■^fof Bchiurus (Figs. 145 and 146), the structure and meta- ''"' morphosis of vi^hich were thoroughly studied by Hatschek, who established the presence of a head kidney. This paired organ consists at first of a simple canal, which opens to the exterior on the ventral R III, I side at the anterior end of the mesodermal bands. Later there is added to this primary head kidney a secondary branch, which is much ramified (Fig. 145) . Altogether the larva undergoes a number of changes, until it arrives at the height of its develop- ment, and the larval organs begin to degenerate. This is true of some other meso- dermal structures as well as of the head kidney. In addition to the muscles characteristic of Annelid larvae, which extend through the blastocoele, there appears in EcMurus under the ectoderm a fine membrane, which arose by the union of branched mesodermal cells and is characteristic for this larva. The mesodermal bands are developed in the anner typical for the Annelida. The pole cells lie at their osterior ends, whereas the differentiation begins at the anterior end. It is here that they are first many-layered, and that they separate into the primitive segments. The ^-on — m Fig. 144.—^ and B, gastrula stage and Trochoplxore larva of Thalassema mellita (after Cown). a, anus; d, intestine; m, mouth; mes, mesodermal bands; ces, oosophagus; sp, apical plate. 308 EMBRYOLOGY latter acquire cavities and enlarge in the well-known manner. Just as in the other Annelida, so in the Echiuridce (Fig. 145), there is established an internal segmentation, corre- sponding to which there is an outer one, in so far as a large number of segmental ciliated bands make their appear- ance on the posterior part of the Echiurus larva. But this segmentation is only temporary, for, like the bands of cilia, the septa between the cavities of the segments also degenerate. Of the fifteen primitive segments which were begun, only the somatic and splanchnic layers remain, and, as the resnlt of the disappearance of the septa, the secondary body cavity of the trunk unites with the z—T^' ■V'' sc. tn. primitive head cavity. Like the cavity of the trunk, the head cavity is also tra- versed by branched cells (Fig. 146), and since these are in part applied to the ectoderm, the dermo-muscu- lar sac, which was estab- lished in the trunk at an earlier period, is also deve- loped in the head region. The nervous system is also established in the larva. On either side of the ventral ciliated groove arise thick- enings of the ectoderm, from which small nodular pro- cesses grow inward, and unite segmentally into large masses of cells, the ventral ganglia (Figs. 145 and 146). In this way the lateral cords arise, to which a middle cord is added. The latter sepa- rates from the ectoderm of the ciliated groove. At first the entire ventral cord is still intimately united with the ectoderm, but the latter gradually detaches itself, and the ventral cord thereby acquires a deeper position. The funda- ment of the supra-oesophageal ganglion, which is small in the Fig. 145. — Trocliophore larva of Echiimis (after Hatschek). a, anus; ah, anal vesicle ; d, intestine ; fen, head kidney ; m, mouth ; mes, mesodermal bands ; n, ven- tral chain of ganglia; sc, oesophageal con- nective; sp, apical plate. The ciliated bands of the posterior part of the body are indicated by the cilia at the margins only. ANNELIDA 309 -jyi -vb illj formed animal, we have already identified as the apical From it two cords extend backwards, embrace the louth-opening, and unite with the ventral cord. In this ray are formed the oesophageal connectives, which are un- usually large in the Echiuridm (Figs. 145 and 146 sc). The anal vesicles, which open to the i>xterior along with the intestine (Figs. L45 and 146), do not arise, as was sup- )osed, from the intestine, but are formed in the somatic layer of the lesoderm. They lie here in the ter- linal segment of the body as two Compact cylindrical structures, which later become hollow and unite with bhe ectoderm on either side of the anus, ^t the same time they grow inward. 'heir middle part is distended, and the fnner end opens free into the body Lvity by means of a ciliated funnel (Fig. 146 ah). Their entire mode of origin proclaims the anal vesicles to be nephridia, which only secondarily entered into connection with the hind- gut. The intestine is no longer so wide in comparison with the entire body, as is to be seen in Fig. 146 ; on the con- trary, it has grown more in length and makes several turns, which subse- quently are still more emphasized. In the meantime the larva has also altered externally, in that its transverse dia- meter has decreased in proportion diameter (Figs. 145 and 146). On the surface the rows of dermal papillae become noticeable, and, just as in the Chaetopoda, the uncinate setae which are formed in the setigeroas sacs (immediately under the ectoderm) break through to the outside (Fig. 146). The further development consists first of all in an active growth of the hinder portion Fig. U6.— Larva of Echiurus (after Hats- chkk). a, anus ; ah, anal vesicle ; b, circle of setae at the hind end of the body; d, intestine; m, mouth; n, ventral chain of ganglia; sc, oesopha- geal connective; sp, api- cal plate ; vh, ventral or uncinate setae. to its longitudinal 310 EMBRYOLOGY of the body, which thus approaches the form of the adult worm, as can be seen from the larval stage of Thalassema shown in Fig. 147. Even in this stage the ciliated bands are still present. With the gradual reduction of these, which soon takes place, the larva approaches /^ ^ more and more the shape of the adult ^ animal. The oral and preoral parts of ^ the larva are transformed into the pro- stomium, whereas the hind part still grows in length. ^ 3, Larval Form and Metamor- phosis of Bonellia. The larva of Bonellia resembles the Trochophore less than do the larvae of Echiurus and Thalassema, although it also can be referred to the Trochophore. Evidently it is much modified, as is proved by its internal organization, which is less adapted to a free life. We follow Spemgel in describing the development of Bonellia. The larva of Bonellia, which is at first spherical, possesses, in addition to the anterior band of cilia, a posterior one (Fig. 148^). m^^m^,^ The anterior to all appearances corresponds to ^^^K. - - flj the anterior ciliated band of the larva of \ Thalassema and Echiurus, especially since in the latter a band of cilia is also found in the region of the anus. Anteriorly two eye-spots make their appearance (Fig. 148 A and h). The intestine is not yet differentiated as such, but is developed later. The larva increases in length, becomes flattened dorso-ventrally and covered with cilia (Fig. 148 B). It now has more the appearance of a Turbellarian, and, as its shape indicates, it moves by creeping. The further development of the larva affects first its internal organi- zation. The supra-cesophageal ganglion has become differentiated from the ectoderm, and later the oesophageal connectives and the ventral nerve cord begin to develop. The entoderm has become a single layer P! Fig, 147.— Late larval stage of Thalassema mdlita (after Conn), a, anus ; ah, anal vesicle ; d, intestine ; m, mouth ; vh, ventral er uncinate setae. ANNELIDA 311 of cells, which surrounds the central yolk-mass like a sac. On the front end of this fundament of the intestine a conical appendage becomes noticeable, the first indication of the oesophagus, which subsequently breaks through to the exterior in the region of the anterior ciliated band (probably behind it). Lying between ectoderm and entoderm is the mesoderm, which has split into a somatic and a splanchnic layer in most parts of the body, whereas in the head region it has the form of a com- pact mass of vesicular cells. Besides the mesodermal elements which are transformed into musculature and peritoneum, there exist still others, which lie in the body cavity. These are transformed into structures like blood cells, which float in the body fluid. It is only at this time that the formation of the spacious body cavity is accomplished. The vessels arise from the peritoneal lining of the body cavity. al^ Fig. 148. — Stages of development of JBoneilia (after Spengki,, from Balfour's Comparatiue Embi-yologyj. A and B, larvae with anterior and posterior bands of cilia. C, yoang Bonellia. oX, intestinal canal ; an.v, anal vesicle ; m, mouth j «, fundament of the ventral hook j ze, excretory organs. At the same time with the internal processes of development described, an external change of form takes place. The ciliation of the body disappears for the most part ; the anterior portion of the body grows considerably in length, and its ventral side, which is still ciliated, be- comes depressed, in this way acquiring a spoon-like form (Fig. 148 C), and thereby realizing a stage like that in Ectdurus. Later, projections are formed on the prostomium where the eye-spots are situated. By their further growth is brought about the bifurcation of the prostomium which characterizes the female of Bonellia. Of the internal changes, there is still to be considered the further development of the intestine, whose central yolk-mass becomes absorbed. The mouth-opening breaks through the base of the prostomium, while 312 EMBRYOLOGY the anus is formed at the posterior end of the ventral side. The anal vesicles are said by Spengel to arise as evaginations of the hind-gut (Fig. 148 C). "We saw that in Echiurus their origin was described differently, and that therefore they are preferably to be considered as nephridia (p. 309). — A pair of tubes which make their appearance behind the mouth are considered by Spengel to be the provisional excretory organs (Fig. 148 C). Immediately behind these the ventral setse are formed (Fig. 148 C). The earliest fundament of the ovary was also observed by Spengel. It is formed, in the same way as in other Annelids, from the peritoneal covering of the blood-vessels, in this case on the posterior part of the ventral vessel. The duct for the sexual products is a tube, which is to be looked upon as a nephridium, although it is not quite evident whether it is connected, and if so by what means, with the provisional excretory organs observed by Spengel. The description of the development of Bonellia up to this point applies to the female only. The development of the small male, living in the uterus, is much simpler, since it remains in the state of the ciliated larva. The larvae which develop into males seek the ciliated groove on the prostomium of the female, and there attach themselves. They lose the two ciliated bands, but retain the uniform coat of cilia. Their internal organization corresponds on the whole to that of the female, only certain simplifications arise ; thus, for example, the mouth and anus are wanting. In the male also the genital products arise from the cells of the peritoneum. Balls of spermatic cells are detached from this and fall into the body cavity, subsequently to be taken up by the funnels of the spermatic duct. After the males have remained for a short time on the prostomium of the female, they migrate into the oesophagus, in order to complete there their metamorphosis. Spengel found as many as eighteen males in the CBSophagus. Subsequently they abandon the oesophagus and repair to the uterus, where ordinarily six, eight, or more males are found. General Considerations. — As regards the position of the EchiuridaQ, we agree with Hatschek's view (No. 51) ; he sees in them a division of the Annelida, and brings them into relation with the Chcetopoda. The form and internal organization of the larva, as well as the mode of origin of the setse, seem fully to substantiate this view. Even though a segmentation [metamerism] no longer exists in the adult animal, it was nevertheless established in the larva, just as in the Chsetopoda and Archiannellda. The loss of the segmentation and the reduction of the setse, as well as the enormous extension of the prostomium, or so-called proboscis, make the Echiurldse appear as somewhat modified forms. ANNELIDA 313 II. DINOPHILUS. Although the development of Dinophilus is not yet known in detail, we include this aberrant form in the course of our present account, because the adult animal itself remains to a certain extent at the stage of an Annelid larva. In its outward shape (Fig. 149) Dinophilus presents a striking resemblance to certain poljtrochal Annelid larvae, e.g., to those of Ophryotrocha and a larval Syllis not yet de- scribed, which we observed in the "tow" at Trieste. This applies not only to the ciliated bands, the distribution of the sensory hairs, and the ventral tail-like appendage, but more especially to the entire habit of the worm. The caudal appendage of Dinophilus, like that of the Annelid larva in question {Ophryotrocha sp.), is segmented. For this reason, as also on account of its ven- tral position, it recalls the foot of the Rotatoria, a comparison which indeed does not appear altogether without foundation when one considers the simi- larity in the organization of Trochophore larva to that of the Rotatoria, which has al- ready been emphasized (p. 259). Should the statement prove to be correct that the five pairs of nephridia possessed by Dino- philus (Fig. 149) end blindly in the body cavity (B. Meyer), there would exist in this particular also conditions such as are found in the Annelid Fig. 149. — Female of Dinophilus gyrociliatus (after E. Mbyer, from Lang's Lelirhuch). a, eye; au, anus; ed, hind-gut ; m, mouth ; md, mid-gut ; n, nephridia; o, ovary; jjJi, pharynx; pJid, pharyngeal glands ; wfc, ciliated band. 814 EMBRYOLOGY larva and in the Rotatoria. As regards the rest of the organi- zation— for example, the structure of the nervous system — Dinophilus has been compared directly to the Archiannelida. Dinophilus lays its eggs, several united, in transparent gelatinous capsules. lu Dinophilus apatris (gyrociliatus) there are found in the capsules, in addition to the large oval eggs, spherical ones, which are several times smaller ; the number of the former compared to the latter is about as two to one. From the larger eggs arise the females, from the smaller the males, when, as in D. apatris (gyrociliatus), there is a great difference in the size and form of the two sexes (Korschelt). Cleavage is unequal in both kinds of eggs, and, according to Korschelt's statements, which are corroborated by Harmer, leads to the formation of an epibolic gastrula. Repiachofp, on the contrary, describes the occurrence of an invagination gastrula for the large eggs of D. gyrociliatus, which arose from a blastula thickened on one side. The blastopore appears to become the mouth. Two large cells are differentiated near it, which, according to Repia- CHOFP's observations, move into the blastocoele, and there, as in other Annelida, produce the two mesodermal bands. The supra-oesophageal ganglion is to be seen lying close to the ectoderm. However, the accounts last mentioned do not seem to be well established ; but what is known of the development of Dinophilus harmonizes with Annelid development, and the entire organization of the worm points to relationships with the Annelida. Prom the fact that a most striking sexual dimorphism exists in Dinophilus, — in so far as the males are much smaller and more simply organized than the females, lacking the intestine, the eyes, and the segmental bands of cilia (Kor- schelt),— relationships of this genus with the Rotatoria have also been contended for ; but these conclusions do not appear to be justified when one reflects that, while sexual dimor- phism makes its appearance in certain species (D. apatris, i.e., D. gyrociliatus), in very similar species, such as D. voi'ti- coides, gigas, and tceniatus (according to 0. Schmidt, Weldon, ANNELIDA 315 and Harmer), the males, apart from the sexual characters proper, are formed just like the females. IV. MYZOSTOMA. Myzostoma, the discoid parasite of the Crinoids, provided with hook-bearing parapodia arranged in pairs, deposits its eggs in large masses without bestowing on them any special care. The eggs, which are enclosed in a structureless mem- brane, are fertilized outside the parent, after the formation of the two polar cells. The fertilized eggs sink to the bottom. Their development was studied by Metschnikoff iu Mi/zostoma cirrifenim, and afterwards somewhat more fully by Beard in Myzostomum glahrum. The unequal cleavage leads to the formation of an epibolic gastrula, of whose six inner cells the two lying nearest to the blastopore are said to give evidence by their darker appearance of being mesodermal cells. Tufts of cilia soon afterwards make their appearance on the ectoderm cells all around the ovate embryo, which now breaks through the egg-shell. Its shape soon becomes pyriform. An ectodermal invagination, the fundament of the mouth and fore-gut, makes its appearance in the region of the blastopore. It grows inward and unites with the stomach, which in the meantime has been formed out of the entoderm cells, which have increased greatly in number. The anus arises at the posterior pointed end of the larva. It is not easy to determine whether it also is formed by an ectodermal invagination or merely by a fusion of ectoderm and entoderm. In front of the anus there appears a papilla [on the ventral side], which subsequently becomes quite large and constitutes the end of the body, the anal opening thus becoming displaced dorsad (Fig. 150). The subsequent stages are characterized by the fact that the cilia distributed over the entire body become restricted to certain regions. These are, first the anterior end of the preoral part of the body, which constitutes the apical area and bears a tuft of rigid cilia (Fig. 150), then a band of cilia lying immediately behind the mouth and a second one in the region of the anus, and finally a bundle of rigid cilia at the tip of the caudal appendage (Fig. 150). At the same time with the changes in 316 EMBRYOLOGY the ciliation, there appear on both sides of the head, behind the post oral band of cilia, the fundaments of setae, which soon elongate considerably, and finally reach approximately the length of the entire larva. It is not to be denied that this Myzostoma larva possesses a very great resemblance to the larvae of Annelids, even though the absence of the pre- oral band of cilia, which could not be found by Beard, inter- feres with a complete resem- blance to the Trochophore. The thickened apical area, the two bands of cilia, the sensory hairs at the anterior and posterior ends, as well as the internal organization of the larva are quite Annelid-like. A caudal appendage, covered in the same way with tactile hairs and con- stituting a prolongation of the ventral surface, is found in the larvae of Telepsavus and Ophrijo- trocJia. In the same way the provisional setae of the Myzo- stoma larva, which probably arise in ectodermal sacs, point to the corresponding structures of certain Annelid larvae (comp., for example, the drawing of Mitraria, Fig. 124, p. 276). In any event the similar characters in the larvae of Myzostoma and other Annelids are very many, and the further develop- ment also presents other common features, e.g., the formation of the parapodia and their bristle- or hook-bearing, stump- like processes. After the larvae have moved about free at the bottom of the aquarium for some seven days, they cast off the pro- visional setae and betake themselves to an Antedon, on which they are found crawling about like worms, for the larval Fig. 150 —Larva ftf Myzostoma glabrum (after Beaed). a, anus; b, setae; m, mouth; s, caudal ap- pendage ; sp, apical plate. ANNELIDA 317 ciliation meanwhile has degenerated. The further develop- ment of the larva is quite simple. The body, which up to this time was broad in front and narrow behind (Fig. 150), changes its shape to such an extent that it becomes broader behind than in front. The principal changes in the shape of the larva are brought about by the develop- ment of parapodia, which takes place from in front back- wards, as in the Polychaeta. Like the setae in Polychasta, the hooks in Myzostoma are said to arise in ectodermal sacs. A segmental differentiation of the compact mesodermal mass lying between the integument and the intestine, a differen- tiation which proceeds from in front backwards, might be compared with the segmentation of the mesodermal bands in the Annelida. A large part of these mesoderm cells become connected with the parapodia as musculature. Others are applied to the mid- and fore-guts. The latter effect the formation of the evertible proboscis. Up to this time the intestinal canal has retained its simple character ; but by the time the development of the parapodia is completed evaginations are seen in it, and in this way its branched character takes its origin. As regards the formation of the nervous system, the apical plate, which is to be looked upon as a larval central organ, is said by Beard to degenerate ; but since this author did not recognize the presence of a supra- oesophageal ganglion and oesophageal ring, which nevertheless are present, it is quite possible that the former arises from the apical plate, and that, as in other Annelids, a union with the ventral cord, which arises as an ectodermal thickening, also occurs in the development. The ventral cord, which exhibits, according to Nansen and v. Wagner, the usual form of a chain of ganglia with transverse com- missures, has thus a segmental arrangement. The statements concerning the origin of the mesodermal structures are less certain. A true body cavity is not present, but its place is occupied by parenchymatous tissue, which is traversed by muscle fibres, and yet the authors (Nansen, Beard) speak of an epithelium of the body cavity, from which the sexual products arise. It appears, then, as if the hollow spaces which contain the sexual products constitute remnants 318 EMBRYOLOGY of the body cavity. Segmental organs have not been identified, though the oviducts were held by Beard to be remnants of such, and Nansen believed the same of the paired ciliated depressions of the outer surface of the body formerly called sucking discs ; but up to the present time su fficient grounds for this view have not been produced . The sexual organs in Myzostoma are not always developed in the same way. In addition to the hermaphroditic individuals, there are living on them very mach smaller male animals (complement al males). The fact that oviducts were also found in these (Nansen) indicates that we have to do, not with individuals of really separate sexes (Beard), but only with incompletely developed hermaphrodites. The place which we assign to the genus Myzostoma appears to be justified by the manner of its development. This characterizes it as a branch of the Annelid stem, which, to be sure, is rather aberrant, and has suffered great changes, probably as the result of the parasitic mode of life. The place previously ascribed to it, supplementary to the class of Arthropods, was necessarily given up when the development became better known. The form and internal organization of the larva, as well as its ciliation, which is also a feature of the adult animal, separate it sharply from the Arthro- poda. V. HIRUDINEA. The Hirudinea, like the Oligochseta, lay their eggs in cocoons, which are formed in the same way in the two cases, namely, by a secretion from dermal glands, which hardens. The cocoons themselves are of various sizes, according to the size of the animal. In the medicinal leech they become more than 2 cm. in length. Their shape also varies in different species and genera. Those of Hirudo and Aulastoma are ellipsoidal, and exhibit outside the shell proper a layer of spongy substance, Avhich probably serves to protect them against desiccation (Leuckart). They are deposited in the earth. The flattened cocoons of Clepsine and Nephelis are found in water, firmly glued to some fixed object. Clepsine covers the cocoon with its body, and ANNELIDA 319 further cares for tlie brood by carrying about with it, at- tached to its ventral side, the young after they have hatched from the cocoon. The cocoon ordinarily contains a large number of eggs, as many as twenty in the medicinal leech. The Gnatliobdellidse and Rhynchobdellidoe are distinguish- able by the fact that the cocoon of the former is filled with albumen, in which the eggs are found embedded, whereas in the Rhynchobdellidas the cocoons lack the albumen, and the much larger eggs lie in rows and in layers alongside and above one another in great numbers, in Glepsme, for example, as many as 200. Correspondingly the eggs of the Gnathob- dellidse are small, and contain little yolk ; the embryos leave the eggs at an early stage of development, and, like the Oligochaeta, float as larval forms in the albumen of the cocoon, by means of which they are nourished. Only after several weeks do they quit the cocoon. The Rhynchob- dellidse, on the contrary, whose large, richly yolk-laden eggs furnish to the embryos sufficient nourishment, do not break through the egg-m.embrane until a much more advanced stage of development and soon after also abandon the cocoon. 1. Cleavage, Formation of the Germ-layers, and Development of the Outward Form of the Body. A. Rhynchobdellid^. The process of cleavage can best be followed in the Bhyn- chohdellidde, on account of the larger size of the eggs, and has been repeatedly studied in Glepsine. According to Whitman, three small blastomeres and a single larger one are first produced by the formation of two vertical cleavage planes, whose position indicates the subsequent orien- tation of the body of the worm. The three smaller ones mark the anterior end, the larger one the posterior end, of the worm. Then four small blastomeres bud out at the animal pole from the four large ones, whereby the familiar stage of four macromeres and four micromeres is reached (Fig. 151 A). The further metamorphosis consists in the separa- tion of the posterior large blastomere into two of nearly the 320 EMBRYOLOGY same size (Fig. 151 5), one of which Whitman designates as neuronepliroblast, and the other as mesoblast, in accord- ance with their subsequent fate. The mesoblast soon divides into two cells, which at first do not occupy bilaterally sym- metrical positions, as would be expected of the primitive cells of the mesoderm. One of them lies more behind, the other more in front beneath the micromeres, the number of which soon increases, first at the expense apparently of the Fig. 151.-/4 to C, cleavage stages of Chpsine, diagrammatic (after Whitmvk). I. and II. indicate the direction of the first and second planes of division ; a, h, c, the macromeres which become entoblasts ; k, the macromere which supplies the germ bands ; a', h', c', k', micromeres which arise from the macromeres a, b, c, and k ; m, mesoblast ; nn, neuronephroblast ; n, neuroblast; np, nephroblasts ; I, pole cells of the lateral cell-row of the germ bands, ect {mikr.), which are descendants of the micromeres. macromeres (Fig. 151 B). With the exception of this pro- duction of micromeres at the animal pole, the anterior three macromeres take no further part in the subsequent cleavage. They contain the nutritive yolk of the egg^ and later supply the cell material for the formation of the mid-gut ; they are therefore to be designated as entoblasts. At the time when the neuronephroblast divides into eight cells symmetrically ANNELIDA 321 w i^Bplaced at tlie posterior pole (Fig. 151 0), additional nuclei make their appearance in the entoblasts without any corre- sponding division of the entoblasts. In addition, however, certain cells that are from the beginning distinct are con- tricted off from the entoblasts; they lie under the layer of icromeres, and are to be looked upon as the earliest ntoderm cells. Later, cells that have been differentiated ithin the entoblasts are added to them, so that a distinc- ion between the two kinds can no longer be made. The embryo, up to the stage to which we have followed it, onsists of a solid mass of cells formed of the three macro- eres (entoblasts), which become partly covered over by the isc of micromeres, which have now become very numerous Fig. 151 C). Under the latter, consequently between them nd the entoblasts, there already lie a number of entoderm ells, while the nuclei that appear within the entoblasts provide for the formation of further cell material to be added to the entoderm. At the posterior pole appear the two symmetrical groups of neuronephroblasts, each com- posed of four large cells (Fig. 151 G), and below them, sunk somewhat deeper, lie the two mesoblasts ; these, too, are now almost symmetrically arranged, although that re- lation cannot be recognized in a surface view, such as Fig. 151 C. The two groups of five cells each at the posterior end are of great importance from the fact that the greater part of the body of the leech arises from them. Since in their origin they can be traced back to the hindermost of the four original macromeres, it follows that this is the one which is responsible for the development of by far the largest part of the body. These two groups of cells undergo the follow- ing change : from the anterior face of each of the ten cells new cells are constricted oft' by repeated cell division, a pro- cess which can be compared with the multiplication of the pole cells of the mesoderm in the Chaetopoda, and also leads to the same result. On each side, then, there arise four adjacent rows of cells, those of the neuronephroblasts, and one lying somewhat deeper, that of the mesoblasts. All of them together constitute the two germ bands, which, how- K. H. E. Y 322 EMBRYOLOGY ever, as will be shown later, are not directly comparable to the mesodermal bands of the Chsetopoda. As a result of the rapid cell-proliferation, the germ bands grow forward, and the layer of micromeres, which have in the meanwhile increased considerably in numbers, advancing at the same time with them, covers a greater extent. Thus the entoblasts gradually become overgrown by the germ bands and the descendants of the micromeres. Whereas the two germ bands at first diverge, their ends subsequently unite at the anterior part of the embryo (Fig. 152 A). They now elongate to such an extent that they occupy approxi- mately the greatest periphery of the egg (Fig. 152 B). Their further change in position takes place in such a way that Fig. 1.')2. — Embryos of Clepsine, elucidating the development of the germ bands (after Whitman). Between the germ bands (fcs(r) the portion already overgrown by ectoderm is dotted; the entoblasts [ent) are shaded by parallel lines, m, region of the mouth; p, pole cells of the cell-rows constituting the germ bands. their anterior and posterior ends appear to be fixed, but they themselves move down toward the ventral side, and thus approach each other, so as finally to unite in the ventral middle line (Fig. 152 F). Fig. 152 C, which is a view at an angle of ninety degrees with that of Fig. 152 B, shows the beginning of this process, whereas in Fig. 152 D the fusion of the germ bands, which takes place from in front back- wards, has progressed still farther. Fig. 152 E exhibits the other hemisphere, and shows that here the germ bands are p ANNELIDA 323 not yet completely united. This has occurred, however, in Fig. 152 F, which shows the embryo in profile. Since the layer of small cells arising from the micromeres follows the growth of the germ bands, the embryo becomes surrounded by a superficial cell-layer, which, according to Whitman, produces the epidermis. Furthermore, the head portion of the worm is said to arise from these cells, and perhaps in the same way as the trunk is formed from the germ bands, for the trunk alone owes its origin to these bands (Whitman, Bergh). During the processes described certain changes, which give rise to the formation of the mid-gut, have also taken place in the entoderm. Even at an earlier stage certain cells had separated from the upper part of the entoblasts. Others succeed these, for the nuclei move out to the surface of the entoblasts, surround themselves with plasma, and in the form of an epithelium — the cells of which at first are flat, but later become cubical — are added to the cells already present. The formation of the mid-gut begins at the anterior end, and progresses from there backward on the ventral side with unusual rapidity. Finally the completely formed mid-gut surrounds the entoblasts, which have now sunk to the value of mere food-yolk. At the anterior end the pharynx, which has arisen as an ectodermal invagina- tion, unites with the mesenteron. A shallow depression makes its appearance at a very early period in the region of the ectoderm cells which are first formed (micromeres) ; in later stages this comes to lie at that point where the two germ bands meet (Fig. 152 m). This depression indicates the future pharyngeal invagination. The latter makes its appearance as a solid growth of the ectoderm, which lies in the depression. Later it becomes hollow and fuses with the entoderm. This (entoderm) lines a part of the proboscip, whereas the remaining part of the proboscis and the pro- boscis-sheath are formed of ectoderm. The anus does not arise until later. When the embryo has developed thus far — that is, when the circumcrescence is completed, and its surface is entirely closed — it abandons the egg and soon afterwards the cocoon 324 EMBRYOLOGY also to undergo its further development while attached to the ventral surface of the parent. As regards the external form of the body, a segmentation [metamerism] can be recog- nized, which is to be referred to that of the germ bands (Fig. 152 F). This segmentation makes its appearance in the same way as in the Chaetopoda, namely, progressing from in front backwards.^ Furthermore the shape changes, in that the body, which was first flat on the dorsal and strongly curved on the ventral , s. d.- 0-- *■-• surface (Figs. 152 i^ and 153), becomes straight and flat on the ventral surface, while its \ ^^^^^^Hl^/ dorsal side assumes the fami- l^^f^ liar arching owing to the more active growth of that part. Fig. 153.-Kmhryo of clepsine (after At this stagO the body COnsists RATHKKandWHiTiiAM). d, intestine; of thirty-three Segments, the fc."*, g-erm bands; s, pharynx; sn, , . • •, , p i • i -i gupjjgj. posterior eight ot which unite to form the posterior sucker (Fig. 153). The anus arises dorsad of this by the fusion of the entoderm and ectoderm. It is a question how the formation of the germ-layers is to be ex- plained. Whitman assigns to the ectoderm the cells that have arisen by ^ [Whitman (No. XXII., Appendix to Literature on Annelida) has recently investigated the metamerism of the Hirudinea and its origin. He endeavors to explain the segmentation of the adult animal by means of embryological facts, and further supports his opinion by the anatomical conditions, especially that of the nervous system. The principal question concerns the interpretation of the head, which is composed of the primary head segment and several trunk segments united with it, a condition similar to that which is also assumed for the Chaetopoda. The mouth may be placed as far back as the fourth segment. The segmented body is derived from an unsegmented. The origin of the segmentation is to be sought in the reproduction by division of the originally unsegmented worm. The individual segments therefore really correspond to separate individuals. The increase in the number of segments is caused by the method of life, which necessitates such an increase. (Comp. in this connection the statements made under " General Considerations regarding Annelida," p. 348.) — K.] ANNELIDA 325 the division of the neuronephroblast (Fig. 151 B, C) ; in that case the circumcrescence of the macromeres by the layer of small cells and the germ bands would seem to produce an epibolic gastrula, an interpretation that was in fact given by Balfour. The deep layer of the germ bands arising from the mesoblasts would then be the mesoderm, though its superficial layer also, by becoming overgrown with the small cells, soon comes to lie inside. These processes recall to a certain extent those in Jihynchelinis, in which Oligochsete the mesomeres at first lie in the region of the ectoderm and give off to it products of their division. Perhaps more detailed observations on this point will yield greater evidence of agreement. At present the germ bands of the Hirudinea and the meso- dermal bands of the Oligochata are not to be looked upon as homologous structures, for they are composed of different kinds of elements. How- ever, Kleinenberg argues for a participation of the ectoderm in the formation of the mesodermal bands, and Wilson likewise finds these same bands of cells, which form the germ bands of the Hirudinea, even in the Oligochaeta (comp. supra, p. 294). If such stages of the embryos of Glcpsine and Lumbriciis as are shown in Fig. 153 (p. 324) and Fig. 132 (p. 286) are compared, the conclusion is natural that processes which led to such similar structures must have ibeen at the beginning of like nature, even though they are now changed [in their details. B. GNATHOBDELLTDiE. A detailed investigation of the cleavage of the e^g of \Nephelis has been given by Butschlt. Nevertheless, owing to the small size of the egg, we are not as accurately informed about the cleavage and formation of the germ-layers in the GnathobdelUdoi as about the corresponding processes in the Bhynchohdellidce. At all events, certain differences between the groups seem to exist. In Nephelis there also occurs a cleavage stage of four viacromeres and [four micromeres, though the latter are said not to arise from all four, jbut from only three, of the macromeres, whereas the fourth, posterior [blastomere remains for a time passive. These three macromeres then .again give rise to three small cells, which are arranged, as in Glepsine, under the micromeres first formed, and constitute the first entoderm cells. The fourth of the four macromeres now divides into two large blastomeres, which Whitman interprets as corresponding to the neurone- phroblast and mesoblast in Clepsine. According to this view, to which Bergh also inclines, the superficial layer of the germ bands would be derived from the former, the lower layer, on the contrary, from the latter. The fact that the "neuronephroblast" is said to form two 32G EMBRYOLOGY additional small cells, which are added to the four ectoderm cells already present, does not agree with the processes in Clepsine. The " neurone- phroblast " and the " mesoblast " each divide into two cells, which are placed symmetrically in respect to the middle line. The edges of the macromeres arch up more or less over the small blastomeres, so that these at times appear to be embedded in them, a process that also takes place in like manner in Clepsine. The fate of the different blastomeres ^,^^ ^^, ---fs^adi. Fig. 151.—^ and B, cleavage stage and an embryo at the time of hatching oE Nephelii vulgaris (after Butschli). ect, ectoderm; ent, entoderm; fcstr, germ bands; mnfcr, macromeres ; mifcr, micromeres ; m, mouth -opening; «, pharynx. could not be followed farther than this, though it is to be assumed that the further differentiation is the same as in Clepsine. At all events, two " germ bands " are also formed here (Fig. 154), which extend from behind forwards and there (in the region of the future mouth) unite. The metamorphosis of the entoderm is important, and in determining the entire shape of the animal, significant. To the entoderm cells first I ANNELIDA 327 »rmed from the three macromeres have been added others, which like- ise have probably been furnished from the same source. The entoderm now Hes in the form of two rows of cells upon the macromeres (Fig. 154 A), which now represent a kind of food-yolk. They are surrounded and partly covered in by the germ bands, while the ectoderm, now increasing I lore rapidly, covers over the anterior part of the embryo. A central ssure (the fundament of the cavity of the mesenteron) soon arises be- Rreen the entoderm cells, which enlarge at the expense of the yolk-cells macromeres] (Fig. 154 A). The latter are forced more toward the hind end of the embryo, and are finally overgrown by the ectoderm, which also spreads out backwards (Figs. 154 B, 156). In this case, there- fore, the macromeres are not taken into the intestine, as in Clepsine, but remain outside of it ; but in this position they too are gradually absorbed. The mouth and pharynx finally arise at the anterior end of the embryo in the form of an ectodermal invagination, which unites with the intestine (Fig. 154 B). 2. The Larvae of the Gnathobdellidae. The embryos of the Gnathobdellidae break through the egg- membrane at a stage in which thej are spherical or oval and have attained about the condition represented in Fig. 154 B. The pharynx, still very simple in structure, leads into the intestine, which now begins to enlarge. The ecto- derm has not yet quite grown over the macromeres. The " germ bands " lie between it and the entoderm. It is seen that the development is not so far advanced as that of the hatching embryo of Clepsine. Whereas the latter is converted directly into the worm, the embryo of the Gnathobdellidee undergoes a protracted larval existence. Like the larvae of the Oligochasta, those of the Gnathobdel- lidae float in the albumen of the cocoon, and take this into the intestine by means of deglutitory movements. For this purpose a provisional pharynx (Figs. 154 B and 156 s) is developed, which is provided with a powerful musculature. The larva possesses still other provisional structures which, are entirely wanting in Clepsine. In Nephelis a cephalic process is developed, which is thickly covered with cilia (Fig. 156). This ciliation recalls that which occurs in the larvae of the Oligochaeta, especially since, as in Lumhricus trape- zoides, it extends on to the ventral side, where it is found in 328 EMBRYOLOGY the median line of the entire ventral surface (Robin). The larvae of the Gnathobdellidfe also possess provisional excretory organs which are comparable to those of the Oligochaeta, even though in number and form they are different. In Nephelis tliere are two (Fig. 156 ?*% and un2), in Hirudo three, and in Aulastoma four pairs of provisional kidneys. In the last form they are found lying on the ventral surface of the larva, on either side of the germ bands from which, according to Bekgh, they take their origin as cell-growths, composed at first of one, then of several rows of cells (Fig. 155). Sub- sequently they separate from the germ bands, and then con- sist of structures somewhat annular in shape and composed of two rows of cells (Fig. 156 un^). These two cell-rows subsequently differentiate in such a way that they consist of two adjacent canals ; one of them becomes the stouter, chief canal, and the other one winds several times about it (Fig. 156 unj) . At the turn- ing point the two are continuous with each other, and therefore really constitute only one canal. A cilia- tion has not been observed in the canals. Not only is there in Ne- phelis the ring-like canal, which is wound about itself, but this is pro- longed into a duct, which to a cer- tain extent constitutes the efferent duct of the organ, and has been compared to such by Bergh, in the sense that the two primitive kidneys would correspond to the two arms of the primitive kidney of Polygor- dius, and that the duct would lead to the point of union of the two. As has been stated, there are in Hirudo three and in Aulastoma four pairs of primitive kidneys, and Leuckart even describes in the medicinal leech their opening to the exterior. Bergh, however, could not confirm this. The primitive kidneys of the Hirudinea are said by Bergh to have nothing to do with the permanent excretory -«<«.. un /23' Fig. 155.— Origin of the primitive kidneys (uiii to un^) from the germ bands (Riaiip/- fceim) of Aulastoma gulo (after Bkbgh). m, mouth ; pz, pole cells terminating the cell-rows of the germ bands. ANNELIDA organs, for these are not formed in the germ bands until the primitive kidneys have already separated from them (comp. ilso p. 332). Like the primitive kidneys, other organs of the larva also degenerate during its metamorphosis into the adult worm. |A musculature consisting of longitudinal and circular fibres, which in the region of the mouth enlarges into a powerful circular muscle, is found under the epidermis of the larva. Between the muscle fibres Bergh finds spindle-like and branched cells, which he takes to be of a nervous nature. This entire larval skin is said by Bergh to be cast ofE in the metamorphosis, and the whole body of the leech, with the f single exception of the mid-gut, arises from the so-called trunk and head germs {Bumpf- und Kopfkeime), of which more will be said later. At this time the mouth closes. The provisional pharynx of the larva is replaced by a per- manent one. Details about these processes will be given in considering the formation of the organs. 3. The Further Development of the Body; For- mation of the Head and Trunk. A distinction between head and trunk was apparent even in the ChsBtopoda ; it was recognizable by the condition of the mesoderm, and also probably found expression in the mode in which the nervous system was formed. In the Htrudinea this contrast is still more decided, for the funda- ments of the nervous system of the head and trunk are separate, and the so-called germ bands probably take no part whatever in the formation of the head. According to the investigations of Bergh, which to a certain degree con- firm and extend the earlier discoveries of Leuckart and Semper, there are two so-called head germs (Kopfkeime) in addition to the germ bands, which we have already learned about, and which are designated by Bergh as trunk germs (Bumpfkeime) . These head germs, the origin of which is still obscure, lie between the pharynx and epidermis as two cell-masses, which become united by a connecting cord 330 EMBRYOLOGY of cells, extending over from one to the other (Fig. 156). From these head germs the whole head portion is said to be formed, including even the epidermis, for the epidermis which is now present (Fig. 156 ep) is of only a provisional nature. In like manner the entire trunk portion (with the exception of the mid-gut) is said to arise from the trunk germs. The head germs and trunk germs unite in the region of the mouth. Thus in the formation of the body a decided difference would exist between head and trunk. Fig. 156. — Longitudinal section of a larva of Nephelis (after Bergh). ent, ento- dermal elements; ep, provisional epidermis; kk, head germ {Kopfkeim) ; m, mouth-opening ; mex, individual mesoderm cells; mu, musde cells; pz, pole cells of the germ band (i.e. trunk germ) ; rfc, trunk germ {Rumpfkeim) ; s, provisional pharynx; urii and uu„ primitive kidneys or their fundaments. Whitman also assumes a fundamental distinction between head and trunk portions, and is inclined to refer the origin of the former to the four micromeres first formed. However, the difference in the Clepsine observed by Whitman is not so striking as here, for in that case the epidermis is not cast off. If the permanent body of the medicinal leech is really formed from four fundaments, then the comparison with the formation of the Nemertean from the Pilidium, which was attempted by Bebgh, is a natural one. In Pilidium also the larval skin is cast off, and the body arises from several separate fundaments, of which the mesodermal are four in number (two in the head and two in the trunk portion) (comp. p. 223). Yet these processes, as far as they are known, appear to take place in ANNELIDA 331 the Nemertini and Hirudinea in a manner that is too inharmonious to warrant a comparison.^ Also the further development of the "head germs and trunk germs," which in the Nemertini takes place by means of ectodermal invaginations and additions of mesenchyma cells, but in the Hirudinea as early differentiations of embryonal cells, shows little similarity, apart from the fact that the Annelidan and Nemertean larvae themselves have only a very slight resemblance to each other. 4. Formation of the Organs. The Body -covering. — At an early stage of embryonic de- velopment the layer of small cells grows over the germ bands and macromeres, and thus forms the epidermis. This epi- dermis, beneath which muscles have already been developed, probably from the germ bands, becomes in Clepsine the epi- dermis of the adult worm, whereas in the Gnathobdellidas it, together with its musculature, is said gradually to dis- integrate, and to be replaced by a new epithelium, which is formed from the superficial layer of the " head germs and trunk germs." These have united in the region of the mouth, and thus the entire body is covered by the new epi- dermis. At the same time the body musculature is formed from the head germs and trunk germs. The remnants of the larval skin are finally cast off. According to the description given by Whitman for Clepsine, and by Bergh for Aulastoma and Nephelis, the epidermis does not appear to be homologous in the two groups, which differ from each other to the extent that the larval skin in one group is directly transmitted to the adult animal, but in the other is cast off, being replaced by a layer of different origin. However, an intermediate condition is said to exist in Clepsine, for, according to Whitman, two cells of the germ bands take part in the formation of the epidermis, though Whitman explicitly denies that it arises from these alone. In Clepsine there is developed out of the epidermis, between it and the ganglion underlying the pharynx, a mass composed of numerous large gland-cells, whose secretion serves to attach to the mother the newly hatched young until their suckers are developed (Whitman). The Nervous System. — In the development of the Hirudinea it is difficult to separate ectodermal and mesodermal ele- 1 It should be added that Bergh himself afterwards ceased to place any value on this quite natural comparison. 382 EMBRYOLOGY ments from each other. Thus the germ bands can be in- terpreted as being formed of both kinds of elements (Whitman). As we have seen, each germ band is composed of four superficial rows of cells and a more deeply located one (Fig. 151 C). The ventral chain of ganglia arises from the innermost row of each germ band. The cells multiply, and in this way a cord of cells consisting of several layers is formed from the single row. A segmentation takes place in this from in front backwards. In addition, median and lateral parts are differentiated in the separate cell-masses, and both cords unite in the middle line. In this way arise the ganglia and the commissures. Bergh, like Whitman, also derives the ventral chain of ganglia from the germ bands, but, according to him, the permanent epidermis arises from the same source, and consequently the nervous system takes its origin beneath this. Ncsbaum's (No. 75) theory of the origin of the nervous system differs from that described. He derives the ventral chain of ganglia, as well as the brain, from a thickening of the ectoderm — that is to say, from the primitive epidermis — and thus adopts an interpretation that (more in harmony with theoretical considerations) was also espoused by Kowa- LEvsKY and Balfour. The statements of Nusbaum on this and other developmental processes of the Hirudinea harmonize so little with the statements of the other authors on this subject that any further con- sideration of them must be omitted. The development of the supra- oesophageal ganglion is initiated in the head germs, underneath the layer which supplies the epidermis, by the segregation of a compact mass of cells, in which the Fuiiktsuhstanz can soon be recognized (Bergh). The fundaments of the brain and ventral chain of ganglia would then be distinct, and not until after the concrescence of the head germs and trunk germs would they be united by the development of the oeso- phageal connectives. The Nephridia. — According to Whitman, the nephridia arise from the two middle cell-rows of each germ band,^ and, in fact, Whitman, in opposition to Bergh (comp. supra, p. ^ The fate of the fourth, outer row of cells remained unknown to Whitman. ANNELIDA 333 328), finds a certain resemblance between the primitive kidneys and the permanent nephridia, in that both of thpm arise from the same parts, i.e. from the middle rows of the germ bands. However, in the Rhyncobdellidae themselves, which. Whi TMAN studied, primitive kidneys are not present. The origin of the nephridia from a continuous cord of cells, which, moreover, is described by Wilson in the same way for Lumbricus, recalls the theory advanced by Hatschek that in Criodrilus the permanent nephridia arise from a cord of cells in the somatic layer of the mesoderm (comp. supra^ p. 296). The development of the nephridia takes place from in front backwards by the cord of cells becoming many- layered and undergoing a segmental division. How the nephridia arise from the cell-masses thus produced has not yet been accurately determined. A pair of nephridia is begun in each segment, though all of them do not develop, for in the adult worm there are only sixteen pairs. ^ The Body-cavity and its Lining ; Musculature ; Blood-vessels. — The peritoneal lining of the body cavity and the somatic and intestinal musculature arise from the two more deeply located cell-rows of the germ band, the pole cells of which we have learned to designate as the mesoblasts. The two cell-rows have changed into voluminous cords of cells by the rapid multiplication of their elements. These cords undergo a segmentation from in front backwards. The primitive segments thus produced extend out dorsally, and cavities make their appearance in them. The latter correspond to the segmental (metameric) cavities of the Chsetopoda. After growing completely around the intestine, they are said to become confluent, and to form the marginal sinus, which belongs to the lacunar portion of the blood-vascular ^ [The formation and differentiation of the rows of cells produced by the teloblasts has been again traced by Bergh and by Apathy, as well as in the works of Whitman (see Appendix to Literature on Annelida). The subjects involved are the formation of the nervous system, the body musculature, and the nephridia. These organs have been traced back to definite parts of the so-called germ band, though as yet complete agree- ment on the part of the authors has not been reached. — K.j 334 EMBRYOLOGY system (Whitman). According to another view, however, \ the two remain separate, and constitute the lateral sinuses of the two sides. The other processes — the formation of the septa and that of the intestinal and body musculature — appear to take place in the same way as in the Cheetopoda. By the growth of the mesodermal elements, the body cavity may undergo a greater or less reduction. In the Rhyn- chobdellidae the body cavity is still well developed, and is provided with a distinct peritoneal epithelium, whereas in the Gnathobdellidag it is almost entirely suppressed (Bourne). It has already been mentioned that portions of the body cavity are metamorphosed into parts of the blood-vascular system. It has been stated that the dorsal and ventral trunks of the blood-vessels take their origin from the splanchnic layer, owing to a splitting of it.^ The Genital Organs are doubtless of mesodermal origin, though the statements which are made concerning their formation are little to be trusted. ^ The Intestinal Canal. — In both the Rhynchobdellidae and the GnathobdellidsD we have already become acquainted with the origin of the mesenteron from the three entoblasts. These give rise to a vesicle composed of large cells, which gradually resorbs the entoblasts whether enclosed within or lying outside it, and becomes connected with the outer world by means of an ectodermal invagination (comp. pp. 323 and 327). The pharynx which is formed in this way presents different conditions, according as the development is direct or indirect. In the first case the pharnyx, pro- duced by the collaboration of entodermal, ectodermal, and ^ [Burger (Appendix to Literature on Annelida) has made an ex- tensive investigation of the formation of the body cavity, the blood- vascular system, and the nephridia. He traced the establishment of the coelom, its differentiation, and its relation to the circulatory system. In regard to the nephridia, considerable agreement with the Oligochaeta has been found. — K.] [However, Burger (Appendix to Literature on Annelida) has re- cently given a detailed account of their origin, according to which they are referable on the whole to proliferations of the peritoneal epithelium. Not only the sexual glands, but also the efferent ducts, arise in this way.— K.] ANNELIDA 335 probably mesodermal parts, is converted directly into the oesophagus, pharynx, and proboscis-sheath of the adult animal. The intestinal canal attains its final shape as the result of the ingrowth toward it of the dissepiments, which thus cause the caecal diverticula of the intestine. At the same time the intestine is provided with its musculature. In Glepsine there are six pairs of such diverticula ; the seventh pair grows backward through five segments, and consequently acquires constrictions similar to those of the intestine itself. The terminal portion of the intestine extends straight backwards and unites with the ectoderm to form the anus. The conditions are not so simple in the Gnathobdellidse. Here the pharynx first formed is of a provisional nature, and functions only in the reception of the albuminous nourish- ment. After this office is discharged it degenerates ; the mouth closes as the result of the concrescence of the head germ and trunk germ (Bergh). At the same point there is formed an invagination of the united head and trunk germs, the fundament of the permanent pharynx, which grows into the larval pharynx and unites with the entoderm, while the tissue of the old pharynx is gradually absorbed. The jaws arise in the pharynx as fold-like elevations covered by a firm cuticula (Leuckart). The oral sucker is formed as a circular elevation of the superficial layer of the body. The development of the mid-gut takes place in a manner similar to that already described above for Glepsine. On the other hand, according to Bergh's observations, the hind- ut is formed as a solid outgrowth of the tissue of the " trunk germ," which subsequently becomes hollow, and unites with the entoderm. Such a mode of origin agrees with Bergh's entire theory, according to which the whole body of the leech, with the single exception of the mid-gut, is formed from the so-called head and trunk germs. The degeneration and regeneration of the pharynx in the Gnathob- dellidae recall the replacement of the larval stomodffium by a permanent pharynx in Lopadorhynchus as described by Kleinenberg, even though the metamorphosis takes place there in a different way. ir> 336 EMBRYOLOGY General Considerations. — The development of the Hirudinea doabtless points to the fact that in dealing with them one has to do with Annelida. Although differing in details, the entire process of development is similar to that of the Chaetopoda, and especially of the Oligochaeta. The so-called germ bands of the Hirndinea and the mesodermal bands of the Chaetopoda, it is true, do not appear to be homologous structures, but the entire manner of their formation and their relation to the embryonic body in general, as well as their subsequent development, indicate that both are to be referred to like structures, and that in the Hirudinea a modification has appeared only in so far as the more simple mesodermal bands have there acquired a more complicated structure by the addition of ectodermal parts. In their mode of development the Hirudinea appear to be less primitive forms than the Chaetopoda. Just as the mode of origin of the individual organs, especially the body cavity, the nervous system, and the excretory system, shows the leech to be an Annelid, so, too, does its anatomical structure. This is mentioned only for the reason that direct relationships between the Hirudinea and Platyhelminthes have been sought for in various direc- tions. In this connection it is only the structure of the genital organs and their resemblance to those of the dendrocoelous Turbellarians that appear to be remarkable. It would be desirable to know more than we do at present regarding this point. In brief it must be said that, as compared with the Chaetopoda, the Hirudinea show themselves to be in struc- ture and development higher forms, which exhibit many secondary modifications. VI. BRANCHIOBDELLA. The systematic position of Branchiohdella is not yet established. There are anatomical grounds for the view that this worm is to be assigned to the Oligochaeta, and that it is only in consequence of its parasitic mode of life that it has acquired certain characters — for example, the posterior ANNELIDA 337 sucker — which cause it to resemble the Hirudinea (VoiGT, Vejdovsky). The development exhibits in some features a resemblance to that of the Hirudinea, but otherwise it is so peculiar — provided we can rely on the statement of Salensky — that the relationship of Branchiobdella to either branch of the Annelida cannot be inferred from it. Branchiobdella lays its eggs, each surrounded by a firm envelope, on the gills of the crayfish, where they are attached by means of a stalk, a prolongation of the envelope. A cocoon proper, as in the Oligochaeta and Hirudinea, does not exist, although the egg is surrounded, as in these, by a special envelope ; perhaps therefore the outer envelope is equivalent to a cocoon. In the cleavage and the formation of the germ-layers, conditions are exhibited which do not resemble those of the Hirudinea or Oligochaeta, but can perhaps be referred more jadily to the latter. We begin with the stage in which one large and three small blastomeres are formed. All four are bo be called macromeres, for soon four micromeres are ibstricted from them. By division of the micromeres and fehe formation of new ones on the part of the macromeres, a ipid increase of these small (ectoderm) cells takes place. 'hey soon form an irregularly defined cell-plate, the sides of 'hich grow out and overlie the macromeres in the form of a iap. The striking thing in this is that the micromeres are [said to correspond to the ventral side of the worm ; however, [it is also stated that for Clepsme the mouth breaks through in the region of the first four micromeres, and it has a jimilar position in Nephelis (comp. Fig. 154, p. 326). A ither small cleavage cavity makes its appearance between bhe micromeres and macromeres; it is subsequently forced tway from the macromeres by the production of new ^ells. The macromeres have likewise divided and arranged themselves as two pairs of large cells at the posterior end [Fig. 157 A). A cord of ectoderm cells forces its way between the two pairs. The double-row arrangement of the lacromeres persists even vrhen they divide further (Fig. L67 B). K. H. E. Z 338 EMBRYOLOGY These wo rows of macromeres have been compared to the macromeres of the Hirudinea, though, as far as can now be seen, this comparison is ijot warranted, for the macromeres in Branchiobdella are said to continue dividing, and to give rise to the mesoderm and entoderm. But in both the Hirudinea and the OHgocheBta the separation of the two germ -layers takes place much earlier. The division of the macromeres advances steadily from behind forwards. In this way two different groups of cells arise, one of which lies next to the ectoderm, and constitutes the mesoderm, while the other, lying next to the macro- meres, is the entoderm. What is left of the macromeres themselves divides uninterruptedly, so that the cells arising in this way become like the ectoderm. They cover the posterior part of the embryo (Fig. 157 D). Even before the macromeres separated into the different elements in the manner described, a depression of the ecto- derm (Fig. 157 A), which does not persist long, and perhaps represents the fundament of the supra-oesophageal ganglion, makes its appearance in front of them, and therefore on the dorsal side of the embryo. This originates independently of the ventral chain of ganglia. The latter arises, according to Salensky, in the form of an extensive groove on the ventral side (Fig. 157 C). The groove is very broad at the posterior end of the embryo. It is bounded here by the large cells still remaining, which, continuing to divide, contribute to the formation of the margins of the groove. The groove becomes narrower anteriorly, extends on to the dorsal side of the embryo, and here bifurcates (Fig. 157 D). The part of the ectoderm which is encircled by the two branches probably corresponds to the above-mentioned ectodermal depression, and produces the supra-oesophageal ganglion, which secondarily unites with the two processes of the ven- tral chain of ganglia by means of two ridge-like processes, the oesophageal connectives. The ventral cord itself is said to originate in a manner quite like that of the medullary tube of vertebrates. The groove becomes closed by the bending together of its upper edges (in this case, however, from in front backwards), and in this way forms a tube, which finally separates from the overlying ectoderm, loses ANNELIDA 339 uraen, and lies as a cell alar cord in the ventral median of the embryo. On each side of the nerve cord lies a ribbon-like cord of cells, the mesodermal band. The two mesodermal bands are united to each other by a median part. They have arisen from the common ento-mesodermal mass, the origin of which we have previously traced, by the separation of a ventral layer, the mesoderm, from the dorsal layer, the entoderm. A segmentation, like that in the ventral cord, also makes its appearance in the mesodermal bands, which separate into the primitive segments. The processes thus effected, as ell as the formation of e body cavity and the septa, take place in a manner similar to that described for the other Annelida. The internal segmentation is late in finding expression on the exterior of the body, and is suppressed in its an- terior and posterior parts, where the primitive seg- ments for the present acquire no cavities, and therefore remain in an embryonic condition. Each segment exhibits externally a division into a broader and a narrower portion (Fig. 157 E). The former corresponds to a ganglion, the latter to a septum. In front of the anterior end of the ganglionic chain lies a part of the mesoderm, which forms the head cavity ; but regarding this, Salensky could not determine whether it Hkewise arose from the mesodermal bands. Fig. 157.—^ to E, embryos of Bratichio- bdella in vadous stages (after Salenskt). ect, ectoderm ; gr, pit in the entoderm on the dorsal side; ma, macromeres ; m, mouth- opening ; n, neural groove j «, sucker. ' 340 EMBETOLOGT At the time of the appearance of the outer segmentation a peculiar change in the position of the embryo occurs. Up to this time its ventral side was greatly curved, for both the anterior and the posterior ends grew toward the dorsal side. Later it assumes the reverse position. This is effected by a rotation of the embryo on its own axis. The movement begins at the anterior and at the posterior parts of the em- bryo, and gradually extends to the middle portion. Where- as the ganglion chain at first lay on the convex side of the embryo, it is now found on its concave surface. In the course of this process the anterior and posterior parts of the body assume their permanent shape (Fig. 157 E). The posterior end is abruptly truncated. A depression on it, which soon makes its appearance, represents the fundament of the sucker. The absence of segmentation at the anterior end is noticeable; the head, however, is distinct from the anterior part of the body (Fig. 157 E). The mouth-opening makes its appearance as a shallow depression of the ecto- derm far in front, and probably at the place where the medullary groove bifurcated. It unites with the fore-gut, which, as well as the hind-gut, is said to arise from the en- toderm. The entoderm for a long time consists of a compact mass of cells, which increases in length with the growth of the embryo. In the formation of the epithelium, the cells withdraw to the periphery of the mass; and the nutritive material, which is surrounded by them, remains at the centre just as in the formation of the intestine in Bhynchelmis. The fore- and hind-guts are the first to be hollowed out. The latter unites with the very short tube which forms the anal invagination located on the dorsal side of the sucker. The entire oesophagus, even the jaws, are said by Salensky to be of entodermal nature ; and only the lips, with their in- ternal lining, are formed of ectoderm. Last of all follows the development of the mid-gut. Even in the hatching em- bryos, which have approximately the development described (Fig. 157 E), the mid-gut is still filled with an undigested yolk- mass. To enumerate once more the chief points in the develop- ment of this unique group, which it has hitherto been im- ANNELIDA 341 possible to unite satisfactorily either with the Oligochaeta Por Hirudinea, the following points, in addition to the al- ►gether aberrant cleavage phenomena, are remarkable: the formation of the mesodermal bands and the very peculiar manner in which the nervous system is formed. A germ band in the sense of the Hiradinea is not present, but, on the other hand, the fundament of the nervous system differs from that of the Chaetopoda. To be sure, the origin of the chain of ganglia from a ventral ectodermal invagination has been repeatedly described for the Annelida, but this conclu- sion has not gained currency. At all events, the origin of the nervous system and mesodermal bands of Branchiobdella merits renewed investigation. General Considerations regarding Annelida. The embryology of the Annelida affords us some hints regarding the phylogenetic derivation of the Annelid stem Lud its genetic relationships to other groups of animals, and Jso regarding the origin of metameric segmentation. These mggestions are significant, even though they do not as yet [furnish a foundation of positive knowledge, but serve only [to support theories of greater or less probability. If we consider the larval forms of the Annelida, we see |that their different shapes, however variously they may be pexpressed, are referable to the Trochophore. The Trocho- ihore is the typical larval jor'm of the Annelid stem. Even in the derived and much-modified groups, such as the OUgo- Ichoeta and Echiuridas, as Avell as in the aberrant genus Myzo- Istoma, the larval Trochophore form can be recognized more ►r less distinctly. Dinophilas corresponds in its shape fand organization to a so-called polytrochal larva, which it was possible to derive directly from the Trochophore (comp. :p. 278). The embryos of the Hirudinea exhibit the greatest [resemblance to those of the Oligochaeta. However, they are [much more modified than these, and consequently cannot be [traced directly to the Trochophore, though this may be [accomplished through the mediation of the Oligochaeta. Most likely the Trochophore of the Annelida embodies 342 EMBRYOLOGY the ontogenetic recapitulation of an ancestral form which was common to the Annelida, Mollusca, and Molluscoidea, and from which these animal stems branched off as inde-, pendent groups. The assumption that the Trochophore cor- responds to an ancestral form is supported, not alone by the circumstance that the larval forms in the groups men- tioned can all of them be traced more or less directly to the Trochophore : it acquires a further and important sup- port from the fact that in the division liotifera we see before us forms which in their adult condition remain essentially at the stage of organization of the unsegmented Trocho- phore. We have already mentioned (p. 259) that not only the Rotifer known by the generic name of Trochospheera, but also the rest of the Rotatoria, can readily be referred to the plan of the Trochophore. The Rotatoria accordingly are organisms which still exhibit the closest relationships to the Trochophore-like ancestral form whose mode of locomo- tion and plan of organization, with some secondary changes, they have retained. If we take into comparison the rest of the groups of the so-called Vermes, there is apparent, in the first place, a striking resemblance between the Trochophore of the An- nelida and the larval form of the Nemertini known as Pilidium (comp. p. 231), even though in their further de- velopment the two groups pursue different paths. By means of the Pilidium we are also led to bring certain larvas of the Turbellaria into remote comparison with the Trochophore (comp. pp. 168 and 230). In searching for the ancestral forms from which the Trochophore-like archetype arose one meets with great diffi- culties. In order to arrive at an idea of this ancestral form, the Trochophore has been compared to a Medusa. Its pelagic mode of life, its shape, and, above all, the nerve- ring of the ciliated band discovered by Kleinenberg, were the things which led this author and Balfour to assume its descent from a medusoid form. Derived in such a way, the preoral band of cilia is, from its position, referred to the margin of the umbrella, and the aboral dome of the Trocho- phore to the ex-umbrella, whereas the part of the larva lying ANNELIDA 343 behind the ciliated band must be considered as the sub-um- brella, made to bulge downward. A more careful considera- tion, however, offers considerable difficulties to a derivation of this kind. Even if we disregard the fact that the Medusa represents the most divergent and most highly developed form of the Cnidaria type, and that forms which are highly developed in one direction ordinarily do not become points of departure for new developmental series, still the difficulty of the derivation suggested is evident from a comparison of the mode of locomotion of, the two forms. The Medusa moves by means of oar-like strokes of a complicated loco-- motor apparatus, depending upon muscular action. On the other hand, the Trochophore, with its trochal organ operated by ciliary motion, represents a much more primitive loco- ^motor apparatus, directly comparable in it^ mode of action to the ciliated planula (comp. p. 154, et seq., the grounds hich have been advanced against the derivation of the fCtenophora from Medusse). A chief difficulty in the derivation under discussion is found in the presence of the I central nervous system at the apical region, where important Forgans are never developed in the Medusas. We should then [Jiave to look upon the nerve-ring of the Trochophore as the ihief part of the central nervous system, and the apical plate LS a subsequently acquired secondary part of it ; but in the iresent state of our knowledge we are not justified in this. We recognize that the two parts of the nervous system be- long together, and have probably been developed in close ^relation with the locomotor apparatus, as regulators of the movements. Thus perhaps the apical plate in its earliest [origin is to be traced back to a tuft of cilia functioning as a rudder (such as is met with at the apical pole of many ; Actinian larvae), whereas the ring-nerve, it is to be assumed, has been formed in connection with the development of the I trochal organ, both of them as localizations of a system of [nervous internuncial fibres, distributed under the entire ectoderm. It might be mentioned here that, in addition to _,the apical plate, a nerve-ring is also met with in the Pili- dium. We have above adduced the difficulties which, according 344. EMBRYOLOGY to our point of view, are opposed to a derivation of the Tro- chophore from the medusoid form, and have already made some suggestions respecting a derivation of the Trochophore which, although based upon hypothetical grounds, neverthe- less appear to be better supported by the facts of compara- tive anatomy and embryology than the former view. This view brings the Trochophore into relation with the ancestral forms of the Nemertini, Turhellaria, and Gtenopliora, and regards it as having arisen tolerably directly from much more primitive coelenterate forms than is possible on the assumption of derivation from Medusae. It should be ex- pressly noted here that we necessarily abandon the realm of positive demonstration in making these statements, which scarcely have any higher value than that of mere conjec- tures. To as the facts appear to indicate that the ancestral form arose rather directly from a uniformly ciliated gastrula- like archetype by a change in the mode of locomotion. Such a primitive, completely and uniformly ciliated organism exhibited an anterior apical and a posterior oral pole. Secondary axes had not yet been developed ; the form presented at first the monaxial heteropolar type. It is possible, and in view of the ancestors of the Ctenophora probable, that on this form certain differentiations made their appearance without causing an abandonment of the monaxial, heteropolar form, or the radial form that arose from it. Among these differentiations we reckon a tuft of cilia at the animal pole functioning as a rudder (the earliest fundament of the apical plate), an ectodermal pharyngeal tube, and the formation of diverticula of the entodermal portion of the intestine, by the regular distribution of which around the chief axis the first impetus to the formation of definite secondary axes was probably given. It must be mentioned that many Actinian larvse present exactly the structure described {Scyphuia). However, this resemblance is probably founded merely on analogy, for in the Cnidaria we assume that the formation of radial gastral pouches took place only after attachment and the development of an Archhydra stage, whereas the Cfcenophora and Bilateria probably never had ^n attached ancestral form. ANNELIDA The original mode of locomotion of the uniformly ciliated, radial ancestor described, which had arisen from the gas- trula stage by means of some further differentiations, was spiral, such as we may still see in the ciliated planulae of many lower animals. It depended upon a combination of a prog-ressive movement in the direction of the longitudinal axis with a rotation of the entire body about this axis. The ancestral forms of the Platyhelminthes have perhaps been directly developed from such a uniformly ciliated stock- form by the assumption of the creeping mode of life, and the ancestors of the Ctenophora may have been developed by a change in the method of pelagic swarming and by the formation of rows of ciliary plates. Whereas in the latter case the rotation around the longitudinal axis sank into insignificance, and the combined force of the ciliary plates was concentrated on propulsion in the direction of the longi- tudinal axis, in those forms which effected the transition to the Trochophore-like ancestor a change of movement took place. In these cases, though the body as a whole ceased to rotate, the rotatory movement was retained in the trochal organ in the form of a regular circular wave of contraction, Avhereby this organ was in position to under- take a function in relation to the body (now progressing in a constant position) similar to that of the ship's screw in relation to the hull of the ship. Hand in hand with this alteration in the mode of locomotion went a higher differentia- tion of the ciliary locomotor apparatus, by means of which the passage from the original uniform coat of cilia to distinct locomotor organs was brought about. As such are to be mentioned the tuft of cilia functioning as the rudder and the rows of cilia, but especially the preoral ciliated band. It is possible that the bilaterally symmetrical distribution of the body-masses was directly developed in connection with the above-mentioned changes in the form of motion by means of which the body was balanced in its forward movement. At any rate, one of the most important factors for^the development of bilateral symmetry is to be sought in the shifting of the mouth-opening, which now moved 346 EMBRYOLOGY forwards from the posterior pole of the body, thus deter- mining as ventral the side of the body on which this shifting took place. The first cause of this forward migra- tion of the mouth, during which, the opposite edges of the posterior parts of the blastopore successively approached each other until the slit thus produced was at last entirely closed, is to be sought in the significance of the trochal organ as a food-procuring apparatus and the importance of the approach of the mouth toward it. By such a change in the position of the mouth, the relations of the primary axes were disturbed, so that henceforth the chief axis of the body can no longer be referred directly to the primary axis ; for this reason the Bilateria are also designated as Heteraxonia (Hatschek). Owing to the resemblance which the Pilidium and many Turbellarian larvte present to the Trochophore, one might also be inclined to derive the Platyhelminthes and Nemertini directly from a Trochophore-like ancestor. The complete ciliation of these forms would then be not primitive, but re-acquired after the transition to the creeping mode of life (therefore as in Coeloplana, comp. p. 157). On the other hand, it must be pointed out that ciliated bands are developed in great variety in pelagic larvas, and we are certainly not in position to prove the homology of all these bands. Hence the resemblance of these larval forms to the true Trochophore of the Annelida is perhaps to be referred merely to analogy in development, which would have its cause in a developmental tendency in this direction inherited from the common ancestor. As a result of the development of the most important locomotor organs in the anterior half of the body, it came about that the organs of the animal functions arose in this region. It is this part of the body which, as head, we place in contrast with the posterior, subsequently elongated por- tion of the Trochophore, which is called the trunk, and gives rise particularly to organs of vegetative functions. The fruitfulness of the conception, in the interpretation of the Annelid body, that head and trunk are distinct, has only been equalled by the difficulty of determining the precise boundaries between these two primary regions of the body. In the first place, the question arises whether the mouth ^ ^ ANNELIDA 347 to be assigned to the head or to the trunk. In the >lution of this problem the condition of the mesoderm plays a particalarlj important role. No real coelom appears be formed in the portion lying in front of the mouth ; m the contrary, the first pair of primitive segments is said surround the pharynx. If this were so, then a distinction rould really be established between the preoral and the )ral portions, and the latter would have the greater re- jemblance to the body segments. However, this distinction later obliterated owing to the fact that portions of the lesoderm from the most anterior primitive segments migrate [into the preoral region and form its musculature. Thus interpretations differ, inasmuch as the preoral part alone |(Kleinenberg) and also that together with the oral segment 'Hatschek) have been taken to be the head portion. More- >ver, induced by the peculiar phenomena in non-sexual ^production, some authors have gone farther than this and 'considered a greater number of segments (as many as six) as the head portion of the worm (Semper, v. Kennf.l). The first and last theories seem to us to go too far. Until the final settlement of the question how the mouth and the pharynx are related to the first primitive segment, we would reckon the mouth region as belonging to the head of the worm. The transition from the Trochophore-like stock-form to the real Annelid ancestors (Archiannelida) took place by considerable growth in length, whereby the trunk portion of the worm became larger, and the primary head portion less and less conspicuous. At the same time a change in the mode of life took place, the pelagic being exchanged for the creeping habit. The larval stages belonging to this transition are dis- tinguished especially by the terminal growth of the body. Near the posterior end of the body, which we can henceforth distinguish as the anus-bearing terminal segment, is found a zone of proliferation, from which new cell material is continually being given off forward to the elongating trunk portion. Since, at the same time with this growth in length, the segmentation of the trunk is established, it 348 EMBRYOLOGY follows tliat the most anterior segments of the trunk are formed first, and therefore are the most differentiated ones in the developing larva, while behind follow younger and younger ones. The growth of the Annelid body therefore does not depend upon growth of the body in all directions, but upon a partial (terminal) growth, since new segments are always being supplied from a zone of proliferation lying near the posterior end of the body (in front of the terminal segment). This productiveness of a restricted portion of the body strongly recalls certain kinds of non-sexual re- production, and therefore the process has been called, even in this case, a " budding of the segments." That, however, is an inaccurate mode of expression. The most natural comparisons are those with the tapeworm chain and with the strobila of the Scyphomedusae. The point of comparison in all these cases lies in the production, from a certain zone ot proliferation, of homodynamic portions of the body, which become to a certain extent independent. For this reason the view has been expressed that in the segments of the Annelid body we have befbre us single individuals (which do not arrive at complete separation), and accordingly in the entire body of the Annelid a s^ock or corm. It seems scarcely favorable to this^^ theory that the degree of in- dependence which the individual segments present is com- paratively slight. The most ' important organs (nervous system, body musculature, blood-vascular system) show themselves to be single fundaments of the entire body, and are also developed as such even though they also exhibit evidences x)f metamerism. Even the excretory canals may give up vtheir segmental isolation and become united to one another by means of longitudinal canals. The comparison with the single fundaments of the othei* systems of organs inclines us to the opinion that the development of the nephridia from separate fundaments (Bekgh) represents a coenogenetically altered condition, and that the nephridial system was originally developed by separation from a com- mon cord (Hatschek). By such an assumption the com- parison of the nephridial system of the Annelida as a whole with the excretoiy organs of the Platyhelminthes would ANNELIDA 349 become possible, since the longitudinal stems in the two i cases could be looked upon as corresponding to each other | (whereby we even have in mind a former connection of the 5 permanent nephridia with the head kidney). At all events, ^ the anatomy and development of the Annelid body permit the establishment of the interpretation of the entire body ^ as an individual. Just as in the consideration of the tape- \ worm chain we were induced by the comparison with un- i segmented forms to refer the entire chain to an unsegmented ; individual/ and, on the other hand, to see in the proglottis, ] not a complete individual, but only the abstricted hinder portion of the body of the Cestode, in the same manner, | and with much more reason, we adhere to the individualitv i of the Annelid body. We can accordingly recognize in | metameric segmentation only the regular repetition of certain I groups of organs in the trunk at uniform intervals. i In the question of the origin of the metameric segmenta- - /^ \ tion we shall have to ascertain whether the synchronism J of the terminal growth of the body and the appearance of ,' ; metameric segmentation correspond to a palingenetic con- . ■ dition. In other words, in the hypothetical ancestral form - { were new segments successively added behind during in- ] crease in length, so that forms with many segments arose ' i from those with few by gradual increase in the number of ' J segments ? The fact that the growth of the bo^iy in length " j by the formation of new segments at the posterior end is ' l typical in all Annelids and the forms tierived from them . j (Arthropoda) is an argument in support of this theory. In j that case w^ might perhaps be inclined to the opinion, as j, stated by Hatschek, that in ancestral forms enlarging by terminal growth the differentiation, originally progressing ^ '] continuously, became intermittent, and thus reached the type ■ of the metameric animal. But another view may also be maintained, and, as it seems to us, with quite as much justice — a view which is based upon the assumption that at first an unsegmented, elongated ancestral form was pro- duced by terminal growth, whereupon the entiue iSody be- -; ^ There is a considerable difiference between this and the process of i strobilization. :; 350 EMBRYOLOGY came separated at once into a large number of segments by a rearrangement of the individual organs. This assump- tion is supported bj the consideration that with the lateral sinuous movement of the body, and with the rigidity of the tissues caused by increasing differentiation, the formation of alternating regions of greater and less motility was of considerable advantage to the individual, and rendered possible a further elongation of the body. The first cause for the appearance of metameric segmentation would then be sought in the manner of locomotion and in mechanical conditions. However, this latter view is not supported in any way by embryology. Even though we have not been able to give a positive decision on these difficult questions, yet it seems to us ap- propriate, in the present state of knowledge, to indicate the direction of future inquirj'- by which a possible solution of the questions is to be sought. Literature. I. Ch^topoda and Archiannelida, 1. Agassiz, a. On the Young Stages of a Few Annelids. Annals Lyceum. Nat. Hist. New York. Vol. viii. 1867. 2. Agassiz, A. The Embryology of Autolytus cornutus. Boston. Jour. Nat. Hist. Vol. vii. 1863. 3. Beddard, F. E. On the Anatomy and Histology of Pleurochseta Moseleyi. Trans. Roy. Soc. Edinburgh. Vol. vii. 1883. 4. Benham, W. B. Studies on Earthworms. Quart. Jour. Micr. Sci. Vols. xxvi. and xxvii. 1886 and 1887. 5. Bergh, R. S. Untersuchungen iiber den Bau und die Entwicklung der Geschlechtsorgane der Regenwiirmer. Zeitschr. wiss. Zool. Bd. xliv. 1886. 6. Bergh, R. S. Die Entwicklungsgeschichte der Anneliden. A compilation. Kosmos. Bd. xix. 1886. 7. Bergh, R. S. Zur Bildungsgeschichte der Excretionsorgane bei Criodrilus. Arheiten zool. Imtitut. WUrzhurg. Bd. viii. 1888. 8. BiJLow, G. Ueber Theilungs- und Regenerationsvorgange bei Wiirmern (Lumbriculus variegatus Gr.). Arch. Naturg. Jahrg. xlix., Bd. i. 1883. 9. CLAPARfeDE, E., UND Metschnikoff, E. Beitragc zur Erkenntniss der Entwicklungsgeschichte der Chatopoden. Zeitschr. wiss. Zool. Bd. xix. 1869. ANNELIDA H. 15. 16. 17. 21. 22. 23. 24. 25. 26. 27. 23. 29. >RASCHE, K. VON. Beitrage zur Entwicklung der Polychaten. Wien. 1884—1885. Ehlers, E. Die Borstenwiirmer (Annelida chaetopoda). Leipzig. 1864— 18G8. EisiG, H. Die Capitelliden des Golfes von Neapel : Fauna und Flora des Golfes von Neapel. Monogr. vi. Berlin. 1887. Fewkes, J. W. On the Development of Certain Worm Larvae. Bull. Mu8. Gomp. Zool. Harvard Coll., Cambridge. Vol. xi. 1883—1885. Fraipont, J. Le Genre Polygordius : Fauna und Flora des Golfes von Neapel. Monogr. xiv. Berlin. 1887. GoETTE, A, Untersuchungen zur Entwicklungsgeschichte der Wiirmer. Leipzig. 1882—1884. Grobben, C. Die Pericardialdriise der chatopoden Anneliden, etc. Sitzungsb. Akad. Wiss. Wien. Math.-naturw. CI. Bd. xcvii. 1888. Hatsciiek, B. Beitrage zur Entwicklungsgeschichte und Mor- phologic der Anneliden. Sitzungsb. Akad. Wiss. Wien. Math.- naturw. CI. Bd. Ixxiv. 1877. Hatschek, B. Studien iiber Entwicklungsgeschichte der Anne- liden. Arbeiten zool. Inst. Wien. Bd. i. 1878. Hatschek, B. Zur Entwicklung des Kopfes von Polygordius. Arbeiten zool. Innt. Wien. Bd. vi. 1886. Hatschek, B. Entwicklung der Trochophora von Eupomatus. Arbeiten zool. Jnst. Wien. Bd. vi. 1886. HoRST, E. Over Bevruchting en Ontwikkeling van Hermella alveolata. Versl. en Mededeeling Akad. Amsterdam. Afdeeling Naturk. Bd. xvi. 1881. Huxley, T. H. On a Hermaphrodite and Fissiparous Species of Tubicolar Annelid. New Phil. Jour. Edinburgh. Vol. i. 1855. Kennel, J. von. Ueber Ctenodrilus pardalis, etc. Arbeiten zool. Bd. V. 1882. The Development of the Earthworm Lumbri- Quart. Jour. Micr. Sci. Vol. xix. 1879. Suir origine del sistema nervosa centrale degli Inst. Wilrzburi}. Kleinenberg, N. cus trapezoides. Kleinenberg, N. Annelidi. Mem. Accad. Lincei Roma. Tom. x. 1881. Kleinenberg, N. Die Entstehung des Annelids aus der Larve von Lopadorhynchus : Nebst Bemerkungen iiber die Entwicklung anderer Anneliden. Zeitchr, wiss. Zool. Bd. xliv. 1886. Kowalevsky, a. Embryologische Studien an Wiirmern und Arthropoden. Mem. Acad. St. Petersbourg. Ser. 7, tom. xvi. 1871. Krohn, a. Ueber die Erscheinungen bei der Fortpflanzung von Syllis prolifera und Autolytus prolifer. Arch. Naturg. Jahrg. xviii., Bd. i. 1852. LovEN, S. Beobachtungen iiber die Metamorphose einer Annelide. Arch. Naturg. Jahrg, viii., Bd. i. 1842. 352 ' EMBRYOLOGY 30. Metschnikoff, E. Ueber die Metamorphose einiger Seethiere (Mitraria und Actinotrocha). Zeitschr. wiss. Zool. Bd. xxi. 1871. 31. Meyer, E. Studien iiber den Korperbau der Anneliden. 1. und 2. Theil. Mitth. Zool. Station Neapel. Bde. vii. and viii. 1887 32. Milne-Edwards, H. Observations sur le developpement des Ann61ides. Ann. Sci. Nat. Ser. 3 : Zoologie. Tom. iii. 1845. 33. M'Intosh, W. The Keport on the Annelida Polychseta. " Chal- lenger " Reports : Zoology. Vol. xii. 1885. 34. QuATREFAGES, M. A. DE. Memoire sur I'embryogenie des Annelides. Ann. Sci. Nat. Ser. 3 : Zoologie. Tom. x. 1848. 35. EiETSCH, M. 6tude sur le Sternaspis scutata. Ann. Sci. Nat. Ser. 6 : Zoologie. Tom. xiii. 1882. 36. EouLE, L. Sur la formation des feuillets blastodermiques et du coelome chez un Oligochtete limicole (Enchytraoides Marioni). Compt. Bend. Acad. Sci. Paris. Tom. cvi. 1888. 37. Salensky, W. Etudes sur le developpement des Annelides. Arch. Biol. Tom. iii., 1882 ; tom. iv., 1883 ; and tom. vi., 1887. 38. Schneider, Ant. Ueber Bau und Entwicklung von Polygordius. Arch. Anat. u. Phys. Jahrg. 3868. 39. ScHULTZE, M. Ueber die Entwicklung von Arenicola piscatorum. Halle. 18$6. 40. Semper, C. Die Verwandtschaftsbeziehungen der gegliederten Thiere. Arheiten zool. Inst. Wiirzhurg. Bd. iii. 1876 — 1877. 41. Semper, C. Beitrage zur Biologic der Oligochaten. Arbeiten zni,l. Inst. Wurzburg. Bd. iv. 1877—1878. 42. Vejdovsky, F. Untersuchungen iiber die Anatomic, Physiologic, und Entwicklung von Sternaspis. Denkschr. Akad. Wiss. Wien. Math.- naturw. CI. Bd. xliii. 1882. 43. Vejdovsky, F. System und Morphologic der Oligochaten. Prag. 1884. 44. Vejdovsky, F. Die Embryonalentwicklung von Khynchelmis. Sitzungsb. b'dhm. Gesellsch. Wiss. 1886. 45. Vejdovsky, F. Keifung, Befruchtung, und die Furchungsvorgange des Khynchclmis-Eies, Entwicklungsgeschichtliche Untersuch- ungen. Prag. 1886. 46. ViGuiER, C. ifctudes sur Ics animaux inferieurs de la bale d' Alger. Arch. Zool. exp. et gin. Ser. 2, tom. ii. 1884. 47. Wilson, E. B. Observations on the Early Developmental Stages of some Polychagtous Annelids. Stud. Biol. Lab. Johns Hopkins Univ. Baltimore. Vol. ii. 1882. 48. Wilson, E. B. The Germ-bands of Lumbricus. Jour. Morph. Vol. i. 1887. 49. Zeppelin, Graf M. Ueber den Bau und die Theilungsvorgange des Ctenodrilus monostylos. Zeitschr. wiss. Zool. Bd. xxxix. 1883. ANNELIDA 353 ECHIURID^. Conn, H. W. Life-history of Thalassema. Stud. Biol. Lab. Johns Hopkins Univ. Baltimore. Vol. iii. 1886. Hatschek, B. Ueber Entwicklungsgeschichte von Echiurus, etc. Arbeiten zool. In$t. Wien. Bd. iii. 1881. KowALEVSKY, A. Mittheilungcn uber die Entwicklung von Tha- lassema. Zeitschr. wiss. ZooJ. Bd. xxii. 1872. KiETScH, M. Etude sur les Gephyriena armes ou Echiuriens. Eecueil zool. Suisse. Tom. iii. 1886. Salensky, W. Ueber die Metamorphose des Echiurus. Morph. Jahrb. Bd. ii. 1876. Spengel, J. Beitrage zur Kenntniss der Gephyreen. Mitth. zool. Station. Neapel. Bd. i. 1879. DiNOPHILUS. Harmer, S. F. Notes on the Anatomy of Dinophilus. Jour. Marine Biol. Assoc. Vol. i. 1889. KoRSCHELT, E. Ueber Bau und Entwicklung des Dinophilus apa- tris. Zeitschr. wiss. Zool. Bd. xxxvii. 1882. Meyer, E. Studien tiber den Korperbau der Anneliden. Mitth zool. Station. Neapel. Bd. vii. 1887. Kepiachoff. Ueber Bau und Entwicklung des Dinophilus gyro- ciliatus. (Kussian.) Odessa. 1886. Schmidt, 0. Neue Beitrage zur Naturgeschichte der Wiirmer, etc. Jena. 1848. Weldon, W. F. K. On Dinophilus gigas. Quart. Jour. Micr. Sci Vol. xxvii. 1885. Myzostoma. Beard, J. On the Life-history and Development of the Genus Myzostoma. Mitth. zool. Station. Neapel. Bd. v. 1884. 63. Graff, L. von. Das Genus Myzostoma. Leipzig. 1877. 64. Graff, L. von. Eeport on the Myzostomida. '' Challenger" Re- ports : Zoology. Vol. x. 1884. 65. Metschnikoff, E. Zur Entwicklungsgeschichte von Myzostomum. Zeitschr. wiss. Zool. Bd. xvi. 1866. 66. Nansen, F. Bidrag til Myzostomernes Anatomi og Histologi. Bergens Museum. 1885. 67. Wagner, F. von. Das Nervensystem von Myzostoma. Graz. 1886. K. H. E. . A A 354 EMBRYOLOGY HiRUDINEA. 68. Bergh, K. S. Ueber die Metamorphose von Nephelis. Zeitschr. iciss. Zool. Bd. xli. 1885. 69. Bergh, E. S. Die Metamorphose von Aulastoma gulo, Arbeiten zool. Imt. Wiirzhurg. Bd. vii. 1885. 70. Bergh, E. S. Ueber die Deutung der allgemeinen Anlagen am Ei der Clepsine und der Kieferegel. Zool. Anzeiger. Jahrg. ix. 1886. 71. Bourne, A. G. Contributions to the Anatomy of the Hirudinea. Quart. Jour. Micr. Sci\ Vol. xxiv. 1884. 72. BuTSCHLi, O. Entwicklungsgeschichtliche Beitrage. Zeitschr. ■wiss. Zool. Bd. xxix. 1877. 73. Hoffmann, C. K. Zur Entwicklungsgeschichte der Clepsinen. Niederl. Arch. Zool. Bd. iv. 1877—1878. 74. Leuckart, E. Die menschlichen Parasiten, etc. Leipzig und Heidelberg. First edition. 1863. 75. NusBAUM, J. Eecherches sur I'organogenese des Hirudinees. Arch. Slav. Biol. Tom. i. 1886. 76. Eathke, H. Beitrage zur Entwicklungsgeschichte der Hirudineen, herausgegeben von E. Leuckart. Leipzig. 1862. 77. EoBiN, C. Memoire sur le developpement embryog^nique des Hirudinees. Paris. 1875. 78. Whitman, C. 0. The Embryology of Clepsine. Quart. Jour. Micr. Sci. Vol. xviii. 1878. 79. Whitman, CO. A Contribution to the History of the Germ-layers in Clepsine. Jour. Morph. Vol. i. 1887. Branchiobdella. 80. Salensky, W. Etudes sur le developpement des Annelides — He. partie : Developpement de Branchiobdella. Arch. Biol. Tom. vi. 1887, and Beitrage zur Entwicklungsgeschichte der Anneliden. Biol. Centralblatt. Bd. ii. 1882—1883. 81. Vejdovsky, F. System und Morphologic der Oligochaten. Frag. 1884. 82. VoiGT, W. Untersuchungen iiber die Varietatenbildung bei Branchiobdella varians (and other works on Branchiobdella by the same author). Arbeiten zool. Inst. WUrzburg. Bd. vii., 1885, and Bd. viii., 1888. Appendix to Literature on Annelida. | I. Andrews, E. A. Compound Eyes of Annelids. Jour. Morph. [ Vol. V. 1891. I n. Apathy, S. Keimstreifen und Mesoblaststreifen bei Hirudineen. i Zool. Anzeiger. Jahrg. xiv. 1891. I ANNELIDA 355 III. Beddaed, F. E. Researches into the Embryology of the Oligochffita. 1. On Certain Points in the Development of Acanthodrilus multiporus. Quart. Jour. Micr. Sci. Vol. xxxiii. 1892. IV. Benhaji, W. B. The Post-larval Stage of Arenicola marina. Jour. Marine Biol. Assoc. Vol. iii. London. 1893. V. Bebaneck, E. Etude sur I'embryogenie et sur I'histologie de I'oeil des Alciopides. Bevue Suisse Zool. Tom. i. Geneve. 1893. VI. Bekaneck, E. Quelques stades larvaires d'un Ch^toptere. Bevue Suisse Zool. Tom. ii. Geneve. 1894. VII. Beegh, R. S. Neue Beitrage zur Embryologie der Anneliden : Zur Entvvicklung und Differenzirung des Keimstreifens von Lumbricus. Zeitschr. wiss. Zool. Bd. 1. 1890. VIII. Beegh, R. S. Die Schichtenbildung im Keimstreifen der Hirudineen. Zeitschr. wiss. Zool. Bd. Hi. 1891. I IX. BouENE, A. Certain Points in the Development of the Earth- worms. Quart. Jour. Micr. Sci. Vol. xxxvi. 1893. X. Braem, F. Zur Entvvicklungsgeschichte von Ophryotrocha puerilis Clprd. Mecz. Zeitschr. iciss. Zool. Bd. Ivii. 1893. XI. BiJEGER, 0. Beitrage zur Entwicklungsgeschichte der Hiru- dineen : Zur Embryologie von Nephelis. Zool. Jahrb., Abth. f. Anat. u. (Jntog. Bd. iv. 1891. XII. Burger, 0. Zur Embryologie von Hirudo medicinalis und Aulastoma gulo. Zeitschr. wiss. Zeol. Bd. Iviii. 1894. XIII. Haecker, V. Ueber die Metamorphose der Polynoinen. Ber. naturf. Gesell. Freiburg. Bd. ix. 1894. XIV. KoRscHELT, E. Ueber Ophryotrocha puerilis Clprd. Mecz. und die polytrochen Larven eines anderen Anneliden (Harpochaeta cingulata nov. gen. nov. sp.). Zeitschr. tciss. Zool. Bd. Ivii. 1893. XV. Malaquin, a. Etude comparee du developpement et de la morphologic des parapodes chez les Syllidiens. Compt. Bend. Acad. Sci. Paris. Tom. cxiii. 1891. XVI. Racovitza, E. Zur la Micronereis variegata Clap. Compt. Bend. Acad. Sci. Paris. Tom. cxvi. 1893. XVII. Randolph, H. The Regeneration of the Tail in Lumbriculus. Zool. Anzeiger. Jahrg. xiv. 1891. XVIII. Roule, L. ;^tudes sur le developpement des Annelides et en particulier d'un oligochaete limicole marin (Enchy- traeoides Marioni). Ann. Sci. ^at. Ser. 7, Zool., tom. vii. 1889. XIX. Vejdovsky, F. Entwicklungsgeschichtliche Untersuchungen : Die Entwicklungsgeschichte von Rhynchelmis und der Lumbriciden. Prag. 1888—1890. XX. Vbjdovsky, F. Die Organogenic der Oligochteten. Prag. 1892. 356 EMBRYOLOGY XXI. Vejdovsky, F. Zur Entwicklungsgeschichte der Nephridial- apparates von Megascoiides australis. Arch. mikr. Anat. Bd. xl. 1892. XXn. Whitman, C. 0. The Metamerism of Clepsine. Festschrift fiir Lnickart. Leipzig. 1892. XXIII. Whitman, C. 0. A Sketch of the Structure and Development of the Eye of Clepsine. Zool. Jahrb., Ahth. f. Anat. v. Ontog. Bd. vi. 1893. XXIV. WiiiSON, E. B. The Embryology of the Earthworm. Jour. Morph. Vol. iii. 1889. XXV. Wilson, E. B. The Origin of the Mesoblast Bands in Anne- lids. Jour. Morph. Vol. iv. 1890. XXVI. Wilson, E.B. Some Problems of Annelid Morphology. Biol. Led., Mar. Biol. Lab. Woods Holl. Bnfttou. 1891. XXVII. Wilson, E. B. The Cell Lineage of Nereis. Jour. Morph. Vol. vi. 1892. XXVIII. WisTiNGHAusEN, C. V. Untersuchungen iiber die Entwick- lung von Nereis Dumerilii. Mitth. zool. Station. Neapel. Bd. X. 1891. CHAPTER XI. SIPDNCULID^. • UR knowledge of the development of the SiptmcuUdse is still |very meagre. Concerning Sipunculus and Phascolosoma, Ihe embryology of which has been studied, w^e know that bhey possess larvae which may be compared to the Trocho- thore. The development of Sipunculus^ which has been thoroughly dealt with by Hatschek, presents many peculiari- ties, above all the formation of an embryonal membrane, 'he Trochop}iore-Vik.e larva does not arise directly from the [embryo, but the latter is surrounded by a cellular mem- ►rane, as if by an amnion. I.— The Development of Sipunculus. The first stages in the development of Sipunculus are not [known. By pelagic fishing, Hatschek captured the embryos an the blastula stage. In these embryos the fundaments of khe three germ-layers can already be recognized (Fig. 158 A). 'he thickened part of the spherical blastula consists of tall jells, the entoderm; there is prominent among these a )articnlarly large cell, which, in contrast to the other [prismatic) cells, retains a more spherical shape. This is |tlie first mesoderm cell. It lies in the median plane between bhe ectoderm and entoderm, and marks the posterior part of bhe embryo (Fig. 158 A). The free space which existed )etween the embryo and the egg-membrane — the latter being traversed by radial pores — disappears during the blastula jtage, owing to the fact that the cells apply themselves to bhe egg-membrane. They send out cilia through the pores )f the membrane, so that the embryo, together with the egg- 357 358 EMBRYOLOGY membrane, now begins to rotate. The vegetative pole, which begins to flatten, and then to invaginate, remains free from cilia. In the formation of the archenteron a small part of that portion of the blastula which is still to be assigned to the ectoderm (Fig. 158 B) is also invaginated. The boundary between it and the entoderm is marked by the mesoderm, which has increased to two cells (the primitive mesoderm cells) and now moves inward. The two cells are symme- trical in relation to the median plane. The depression of the ectoderm already mentioned, which follows that of the entoderm, gives the first impulse to the development of the permanent larval skin. It sinks in deeper and deeper and bends over forwards, forming in this way a lamella of thick cells (Fig. 158 G and D, rp). The plate remains temporai-ily united, by means of an amnion-like fold that is not exten- sive, it is true, with the primitive ectoderm, which soon appears only as the serosa (Fig. 158 D). Hatschek calls the plate the trunk-plate in contrast to the head-plate, which is also differentiated from the primitive ectoderm. This differentiation takes place as follows : — In the region of the animal pole (corresponding to the apical plate), which has now also become thickened, the cell plasma retracts from the egg-membrane in an annular furrow, and thus gives rise to a circular space (Fig. 158 B to D, ka) be- tween the permanent ectoderm and an outer layer (serosa). The head -plate, therefore, corresponds to the apical plate. The space between it and the serosa (the amnion is only slightly developed here) Hatschek calls the head amniotic cavity (ka), and that between the trunk-plate and serosa the trunk amniotic cavity (ra). The fold w^hich would corre- spond to the amnion on the trunk-plate is retained for a short time and is then included in the formation of the trunk-plate. The trunk- and head-plates alone supply the ectoderm of the larva. All the rest of the ectoderm of the embryo is employed in the formation of the embryonal membrane (serosa). The serosa grows over the trunk-plate and completely encloses it (Fig. 158 D and E, se) ; however, this is not the case at the opposite (animal) pole. The sipunculidj: 359 head-plate is not overgrown by the serosa, and consequently a circular opening is always retained here. In consequence of a complicated process of growth, regard- ing the details of which the reader must be referred to Hatschek's descriptions, the at first median band-like trunk-plate spreads considerably and grows out towards the sides and then dorsal ly, finally to unite, at the termination of the embryonic period, with the head-plate, which has likewise enlarged somewhat. During this circumcrescence of the internal parts by the trunk-plate, a change in the position of the embryo takes place. The posterior part of the trunk-plate had even before this turned toward the animal pole, and thereby was in a position to supply the dorsal part of the ectoderm (comp. Fig. 158 D and E). In the region of the blastopore, which closes, the oesophagus had been formed from the ectoderm at an early period (Fig. 158 D, hi). This part also changes its position, for it moves more toward the apical plate, whereas the entodermal sac is crowded further backwards (Fig. 158 E). As a result of this, the position of the mesoderm is necessarily altered (Fig. 158 G, D, and E). It is moved from the posterior part of the larva further forwards. Its cells have meanwhile greatly multiplied, and two mesodermal bands have arisen from it (Fig. 158 E, mes). The latter do not undergo a segmentation ; on the contrary, a fissure makes its appear- ance in them, which separates them into a splanchnic and somatic layer. This diiferentiation is first noticeable in the anterior part of the mesodermal bands, and proceeds from in front backwards. The complete development of the embryo is reached by the gradual closing of the permanent ectoderm. We saw that the primitively band-like trunk-plate curved over toward the dorsal side, and that its end moved toward the apical plate. Since the band-shaped trunk-plate lies in the median line, the embryo of such a stage appears in a median section, almost enclosed by the permanent ectoderm (Fig. 158 E) ; however, this is not actually the case, for, although the ventral and dorsal parts of the trunk-plate also spread out laterally, yet they remain separated on either 360 EMBRYOLOGY side hy a broad space. The ventral and dorsal parts of the trunk-plate now grow out more and more on the sides, and Fig. 158.—^ to F, stages of development of Sipuncidus nudus (after Hatschek). A, blastula ; B, gastrula ; C to E, other stages, in which the development of the head- and trunk-plates takes place ; F, embryo during hatching ; a, anus ; hi, blasto- pore ; dr, glandular appendage of the fore-gut; ect, ectoderm; ent, entoderm ; fca, head amniotic cavity; m, mouth; mes, mesoderm; n, nephridium; ra, trunk amniotic cavity; rp, trunk-plate; s, pharynx; se, serosa; so, somatic, sp, splanch- nic, layer of the mesoderm ; w, ciliated band. The ectoderm is finely and the ento- derm coarsely punctate ; the mesoderm is cross-hatched. T I I SIPUNCULIDiE 361 finally fuse in the lateral lines ; f arthermore, a complete fusion takes place with the head-plate (Fig. 158 E and F). An ectodermal invagination at the posterior dorsal part of the larva produces the hind-gut, and fuses with the entoderm. The deep and voluminous fundament of the fore-gut now oes the same. Two invaginations make their appearance on the oesophagus : an anterior, which is developed into a gland with a ciliated efferent duct, and a posterior, the fundament of the pharynx (Fig. 158 F, dr, and s). Stout cilia make their appearance in the circumference of the body behind the mouth-opening, and form the post-oral cili- ated band (Fig. 158 F). The embryo is now ready to hatch. It has up to the present retained its spherical shape; but at the time of hatching it passes into the permanent shape of the larva, owing to the appearance of a constriction behind the ridge that bears the circle of cilia (Fig. 159) ; this marks off the broad anterior part of the body from the conical pos- terior portion. At the same time the entire body enlarges, and its cellular walls consequently become thinner. Hatch- ing takes place by the pointed end of the larva breaking through the serosa and egg-membrane at the pole opposite the apical plate and the emergence of the embryo at this point (Fig. 158 F). The connection with the serosa, as far as it still exists, breaks, and the tuft of cilia of the apical plate is withdrawn through the pores of the egg-membrane, to be retained by the larva. The egg-membrane itself re- mains for a while on the larva like a helmet. The larva of Sipunculus strongly resembles the Trocho- phore, but differs from it in the absence of the preoral ciliated band and the great reduction of the preoral part of the prostomium (Fig. 159). As a result of this, the apical plate comes to lie in the vicinity of the mouth, which is shifted well toward the anterior end of the larva. The usual three regions of the intestine can be recognized, though the hind-gut opens to the exterior on the dorsal side (Fig. 159) ; this, however, is frequently observed in Annelid larvae. A head kidney has not as yet been observed. The internal organization is of a higher grade than is general in the Trochophore, and in part already corresponds to that of 362 EMBRYOLOGY 7^, the adult worm. This applies, for example, to the arrange- ment of the mesoderm, which is seen clothing the walls of the body and intestine as the somatic and splanchnic layers, though, according' to Hatschek, the somatic layer also sup- plies the covering of the fore- and hind-guts, so that only the covering of the entodermal part of the intestine (the mid-gut) arises from the splanchnic layer. From the somatic layer arise also the four retractors of the anterior part of the body, which are developed even in the larva, and extend from the head region to the anus (Fig. 159 r). In conse- quence of this arrange- ment, the anterior part of the larva can be in- vaginated into the pos- terior part. A circular muscle lying behind the ciliated band (Fig. 159 rm) serves to close the opening of the invagina- tion in the larva, which in this retracted condi- tion is almost spherical. The paired nephridia, which in their structure correspond to those of the Annelida, are also produced from the soma- tic layer of the meso- derm. At quite an early stage of the embryo a mesoderm cell was dis- tinguishable from the rest by its strikingly- yellow colour. Some other cells were then added to it. The entire structure assumed a looped form, and a lumen was excavated inside it (Figs. 158 F, 159 n). There are cells, likewise of mesodermic origin, resembling blood corpuscles, which were Fig. 159.— Larva of Sipunculus nudus (after HiTscHKK). a, anus; 61, cells resembling blood corpuscles; dr, glandular appendage of the fore-gut ; in, mouth ; n, nephridium ; r, retractors; nti, circular muscle ; s, pharynx ; ap, apical plate. sipdnculidj: 363 detached from the peritoneal epithelium, and are found ^Jloating free in the body cavity. ^P The metamorphosis of the Sipunculus larva into the adult animal is first indicated in the considerable growth of the dj and the reduction of the head portion. Connected ith this is the complete displacement of the mouth to the nterior end and the further shoving forward of the anus, he latter being brought about by the more rapid growth of the extreme posterior portion of the body. The ciliated band gradually atrophies. It has nothing to do with the development of the tentacles, which take their origin as evaginations of the margins of the mouth. The brain arises from the lower layers of the apical plate, which has become several layers thick. The ventral nerve cord arises in the ventral middle line from an ectodermal thickening, which rogresses from in front backwards. The oesophageal con- ectives grow from its anterior end towards the brain, therefore in a direction opposite to the growth of the ventral cord, and contrary to the method of outgrowth in the Anne- lida, where the apical plate grows out into the connectives. Two additional pigment spots are added to the two which had already arisen in the larva in connection with the apical plate (Fig. 158 F). The provisional organs of the intestine — the glands and the so-called pharynx — atrophy ; the intes- tine itself increases in length and is thrown into several loops (Fig. 159). On the dorsal side of the intestine there arises from its mesodermal covering a blood-vessel ; but this does not take place until quite late. The longitudinal and circular muscle-layers of the dermo-muscular sac are differ- entiated much earlier. The nephridia are said to undergo a peculiar change, terminating internally in vesicular en- largements, while their external openings are retained. The condition of the nephridia recalls the statements made by Schau- INSLAND that in the Priapuhdae the nephridia are closed and, on the one hand, function with their blind ends as excretory organs, while, on the other hand, they are directly united with the germaria, and, in fact, accord- ing to Schauinsland's description, even arise from the latter. Thus even in closed nephridia their function as an efiferent apparatus of the genital products would be explicable. It should be mentioned, however, that 364 EMBRYOLOGY the nephridia of the adult Sipiinculus are described as opening towards the body cavity, and that the sexual glands are explained as growths of the peritoneal epithelium, the products of which are set free in the body cavity, and from there pass into the funnels of the nephridia. II.— The Development of Phascolosoma. According to Selenka's description, the development of Phascolosoma elongatum is much simpler than that of tSipun- culus. Following an unequal cleavage, there arises an epibolic gastrula, which is said, however, to be converted into a kind of invagination gastrula after the division and inva- gination of the macromeres, which soon ensue. Cilia, which, as in Sipunculus, perforate the egg-membrane, make their appearance early. They form a tuft at the apical pole and a post-oral ciliated band ; however, a preoral band is also present, so that in Phascolosoma the resemblance to the Tronhophore is greater. The blastopore is said to be directly converted into the mouth ; the anus in this case, too, lies on the dorsal side. The formation of an embryonal mem- brane is not described by Selenka ; on the contrary, this in- vestigator states that the egg-membrane becomes the cuticula of the larva, as has already been described for some Annelids. The embryo (the larva) then simply elon- gates, so that here a stage quite similar to that in Sipunculus is reached, but in a simpler manner. The larva, which, as one of the last stages, was observed by Selenka, is elon- gated (Fig. 160). The trunk, which is much the more voluminous, is separated from the small head portion by means of a thick, collar-like ridge, which bears the post-oral ciliated band. A large portion of the head is occupied by the broad preoral band of cilia, and at the anterior end the ciliated tuft of the apical plate projects. The head bears two eye-spots. The hooks which constitute the armature of the adult animal make their ap- FiG. 160.— Larva of Phascolosoma elonga- tum (after Selenka). b, setae ; m, mouth ; w, preoral, w^, post- oral, ciliated bands. SIPUNCULID^ 365 pearance in front of the collar. Two pairs of rigid bristles (Fig. 160) arise on the trunk, each of which belongs to an ectodermal cell. A third pair is subsequently added to these. Selexka is inclined to compare them to the setae of the Annelida. The latter arise, it is true, as ectodermal structures, though not in so simple a way as here. General Considerations. — With the limited knowledge that we possess of the development of the different genera of Sipunculidce, it is difficult to pass judgment on the syste- matic position of this group. Until quite recently the Sipunculidce, with the Echiuridce, were usually united into the group of Gephyrea. The grounds which led to this association were rather of an external nature. A comparison of the anatomical and embryological data proves that the two groups exhibit no special resemblances. The so-called proboscis of the Echiuridce corresponds to the elongated cephalic lobe of the larva; the mouth lies at its base, but in the Sipunculidse at the tip of the proboscis. The cephalic lobe entirely degenerates even in the larva. (Comp. Fig. 159, p. 362, and Fig. 145, p. 309.) The differences in the structure of the nervous system, and especially the musculature, which separate the Sipunculidce from the Echiuridce and also from the Annelida, are striking. It seems very doubtful whether these differences can be main- tained after a comprehensive knowledge of the development of the Sipunculidce, and if so, to what extent. The chief point is whether or not the Sipunculidce are to be derived from segmented forms, i.e., whether they are related to the Annelida. In the Echiuridae we saw that a segmenta- tion was indicated in the larva, and through this and the remaining structural conditions of the larva we acquired an insight into their relationships to the Annelid stem. In Sipunculus such indications are lacking. To be sure, the mesoderm here also splits into two layers, progressing from in front backwards, and the differentiation of the nervous system, which, however, is aberrant in being produced from an unpaired fundament, takes place in the same direction ; but no transitory segmentation is indicated, a head kidney is not present, the preoral ciliated band is lacking, and the 366 EMBRYOLOGY head portion sinks into complete insignificance (Fig. 159, p. 362). As regards the formation of the embryonal membrane, this might be a later acquisition, especially since it is said to be wanting in Phascolosoma. Moreover, Selenka argues that the pairs of so-called setae in the latter form might indicate a segmentation ; but this evidence cannot be considered as conclusive. Finally, it should be stated that the structure and develop- ment of the Sipunculidce do not disprove relationships to the Annelida, but that as yet no justification exists for uniting them with these. We place them here next to the Annelida, because definite relationships to any other branch of the animal kingdom are not demonstrable and because in the shape of their larva they are most nearly related to the Annelida. A closer relationship to Phorcniis and the Mollus- coidea appears to us not yet sufficiently established. Literature. 1. Hatschek, B. Ueber Entwicklung von Sipunculus nudus. Arb. Zool. Inst. Wien. Bd. v. 1884. 2. Selejjka, E. Eifurchung und Larvenbildung von Phascolosoma elongatum. Zeitschr. wiss. Zool. Bd. xxv. 1875. 3. Selenka, E. Die Sipunculiden. Wiesbaden. 1883. 4. ScHAuiNSLAND, H. Die Excretions- und Geschlechtsorgane der Priapuliden. Zool, Anzeiger. Jahrg. ix. 1886. CHAPTER XII. CH^TOGNATHA. HE Chaetognatha occupy an altogether isolated position as regards their structure and mode of development. Though thej most nearly resemble the Annelida in peculiarities of structure, they differ from this group in important embryo- logical features. Among the most characteristic peculiarities of the development of the Chaetognatha are to be mentioned the origin of the mesoderm by the formation of two archen- teric diverticula and the early differentiation of the funda- ment of the sexual organs. Owing to the absence of peculiar larval forms, it is evident that the development of the Chaetognatha is abbreviated. The developmental history of the Chaetognatha has been made known chiefly by Gegen- BAUR, KOWALEVSKY, BtJTSCHLI, and O, HeRTWIG. The eggs of the Chaetognatha (Sagitta) after fertilization has taken place are discharged into the surrounding water. ^ They are spherical, transparent, and contain numerous clear yolk spherules. They are surrounded by a vitelline mem- brane and an outer gelatinous mantle. Cleavage must be considered as total and equal, and leads to the formation of a regular blastula, which is characterized by the tall pris- matic form of its cells, which are grouped about a compara- tively small cleavage cavity. One half of the embryo soon flattens and invaginates, whereby the cleavage cavity is reduced to a fissure. In this way a very regular invagina- tion gastrula is formed (Fig. 161 A), the blastopore of ^ [BovERi states that the eggs at the time of ovipositing in passing through the narrow orifices assume an elongated form, but that they recover the rounded shape in the water. When the eggs are laid the first polar spindle is already formed, and every egg also contains a spermatoi- zoun. — K.] 367 368 EMBRYOLOGY which soon narrows. At an early period two large cells, the genital cells, become noticeable at the bottom of the archeateric invagination, directly opposite to the blastopore. A plane passing between these two cells would correspond to the future plane of symmetry. In the course of farther development the genital cells withdraw from epithelial con- tinuity in the wall of the archenteron, passing into the archenteric cavity. Here they divide so that four genital cells lying in the transverse axis can be distinguished (Fig. 161 B). Of these the two middle ones represent the fundaments of the two testes, the two outer, on the other hand, those of the ovaries of the two sides. In the anterior widened portion of the archenteron the formation of two folds now takes place from in front backwards ; these push the genital fundaments before them (Fig. 161 B), and by their development the archenteron is divided into three spaces lying side by side, the middle one of which represents the cavity of the mid-gut, the two lateral ones, on the other hand, those of the paired coelomic sacs. ^ While the embryo now increases in length, the blastopore closes and the permanent mouth-opening breaks through, the latter being accompanied by the development of the fore-gut, which probably arises as an ectodermal invagina- tion (Fig. 161 0, st). The middle one of these three previously formed diverticula acquires in this way an open- ing anteriorly. In the view from the dorsal side (Fig. 162 A) the blastopore and permanent mouth appear to lie directly opposite each other, but side views show that the blastopore is moved a little, as it seems, towards the ventral side of the embryo, so that accordingly the permanent longitudinal axis occupies^a position oblique to the chief axis of the gastrula. With their further growth in length the two folds are pushed farther and farther backwards (Fig. 161 C). In this ^ [According to Joukdain, the two evaginations of the archenteron do not produce the ccelom, but their cavities disappear, and at the same time between ectoderm and entoderm there is formed in the mesoderm a fissure, which becomes the permanent body cavity. This statement contradicts those of the authors mentioned above. — K. 1 I chj:tognatha 869 way the two primitive cells of the testes are also pushed Lckwards (Fig. 161 C), whereas the primitive cells of the -Sd Fig. 161.— Tliree embryos of Sagitta at the sta','e of the formation of the germ- layers, in frontal section (after O. Hkrtwig, from Lang's Lehrbuch). hi, blasto- pore; ud, archenteron; g, primitive cells of the sexual organs ; vm, visceral (.splanchnic), pr>i, parietal (somatic), laj-er of the mesoderm ; d, fundament of the mid-gut; cs, coelomic sacs ; st, stomodajum (fore-gut). ovaries lie at the sides of the folds and in this way are moved rather into the pair of coelomic sacs, in accordance with their subsequent permanent position (Fig. 162). The embryo bends more and more towards the ventral side, during which a ventral ectodermal thickening becomes Fig. 162. -Dorsal and lateral views of an advanced embryo of Sagitta (after BtJTSCHLi, from Ba-t^vovr's Comparative Embryologx)}. m. mouth j al, intestinal canal; vg, fundament of the ventral ganglion; ey, ectoderm; c.pr, ct pi. alio part of the body cavity ; so, somatic, ?p, splanchnic, layer of the mesoderm ge, sexual organs. K. H. E. B B 370 EMBRYOLOGY noticeable as the fundament of the ventral ganglion (Fig. 162 B, vg). Many obscure points still exist regarding the further development. In a species studied by Butschli two portions of the coelomic diverti- cula lying in the head cavities are constricted off at an early period (Fig. 162 c. pv) ; the walls of these are said to be employed chiefly in the formation of the musculature of the head. In the species studied by 0. Hertwig, on the other hand, the formation of such paired head cavities could not be recognized, for here in the course of the further development the walls of the mid-gut and the coelomic sacs are so closely applied to each other that these organs soon present only a slit-like lumen, which finally disappears entirely. A solid, laterally compressed ectodermal cord and two, likewise solid, lateral mesodermal masses, which contain within them the genital products, can now be distinguished. All three of the cords grow out backwards, not only in the region of the future trunk, but also in the tail region, so that the latter also has an ento- dermal fundament ; it is, however, smaller than that of the trunk region. The rudimentary tail portion of the intestinal canal is subsequently em- ployed in the formation of the sagittal septum separating the two caudal cavities from each other ; here it atrophies without acquiring a lumen. It has not as yet been observed either in what manner the transverse septum between the trunk- and tail-cavities is formed, how the canal opening is developed, or even how the efferent sexual ducts are produced. Of interest is the great extension of the ventral ganglionic mass, which remains united with the skin throughout life (Fig. 162 B, vg, and Fig. 163 bg), extends in the young animal along the ventral side and the lateral parts of the entire trunk region, and does not become relatively more re- stricted until later. The transversely striated fibres of the four longitudinal muscle-bands are differentiated from the cells of the somatic layer of the lining of the coelom after the type of epithelial musculature (Fig. 162 A, so). The fins arise as simple evagi- nations of the lateral parts of the ectoderm, whereas the cuticular skeleton found in them probably arises as a secretion of this ecto- dermal cell-layer at its base. In later stages of development the two coelomic sacs move in the trunk region into close contact above and below the intestinal canal, so that a Hies Fig. 163. — Traueverso section through the trunk of Sagitta (after O. Heetwig, from Lang's Le/irbucTi). Ih, body cavity ; mes, mesentery of the intestine ; md, mid-gut; Im, longi- tudinal musculature ; bg, ventral ganglion. CHJITOGNATHA 871 dorsal and a ventral mesentery are formed by their contiguous walls j^rig. 163 mes). The young Sagitta upon hatching from the egg exhi- Ifl^bs essentially the form of the adult animal. I^BcJeneral Considerations.— The problem as to the position of the Chflstognatha in the zoological system is not brought any nearer to I^^lution even by embryology, and for the time being can be treated ^H|ly with the utmost reserve. The agreement which exists between ^^Se transverse section through Sagitta and that through Pohjgordim has already been pointed out by 0. Hertwig. As a matter of fact, a significant resemblance in the tectonic of the two groups is shown in the presence of paired entodermal sacs lined with epithelium, a dorsal and ventral intestinal mesentery, and the four longitudinal muscle- bands arranged in pinnate lamellae, to which in some cases the indication of a transverse musculature is added. The chief difficulty in arriving at a safe conclusion regarding the position of the Chcetog- natha is our ignorance in regard to the excretory system. The sexual organs, particularly those of the male portion of the body, exhibit an important resemblance to the conditions in the Annelida, and if it is permissible to refer the efferent sexual ducts to metamorphosed nephridia, we should have to ascribe to Sagitta at least two trunk somites, and accordingly explain the Cheetognatha as forms in which, perhaps in connection with the manner of locomotion, a primitive segmentation of the body has been retained in a degenerated form only. Embryologically considered, the ChaBtognatha are distinguished from the Annelida by the absence of a Trochophore-like embryonal or larval stage, and, above all, by the characteristic folding off of the mesoderm. In order to harmonize this kind of mesoderm formation with the de- velopment of mesodermal bands in the Annelida, one would have to assume that [in the Chsetognatha] the mesodermal elements increased considerably by proliferation as early as in the blastula and gastrula stages, so that in this way paired mesodermal bands arose, which at first remained lying at the surface of the walls of the archenteron, retaining an epithelial connection with the entoderm, and only later, by the forma- tion of diverticula, became detached. By this assumption it is compre- hensible how, even in closely related animals, two so apparently different kinds of mesoderm formation might be realized. Literature. 1. BuTscHLi, 0. Zur Entwicklungsgeschichte der Sagitta. Zeitschr. wiss. Zool. Bd. xxiii. 1873. 2. Gegenbaur, C. Ueber die Entwicklung der Sagitta. Ahh. Naturf. Gesell. Halle, 1856. Translation: Quart. Jour. Micr. Sci. Vol. vii., p. 47. 3. Grassi, B. I Cheetognati. Fauna und Flora des Golfes v. Neapel. Monogr. v. Leipzig. 1883. 372 EMBRYOLOGY 4. Hektwig, 0. Die Chatognathen. Jena. Zeitschr. Bd. xiv. 1880. % 5. KowALEv'sKY, A. Embi'yologische Studien an Wiirmern und Arth- '{ ropoden. Mem. Acad. St. Petersbourg. Ser. 7, torn. xvi. 1871. ;^ i Appendix to Literature on Choetognatha. ;] I I. BovERi, T. Zellstudien. Heft 3. Jena. Zeitschr. Bd. xxiv. 1890. ! II. JouRDAiN, S. Sur I'Embryogenie des Sagitta. Comiyt. Rend, t Acad. Sci., Paris. Tom. cxiv. 1892. 1 CHAPTER XIII. ENTEROPNEUSTA. NDER the name of Enter opnetcsta it is customary to place e isolated form Balanoglossus next to the Echinodermaia^ ince it scarcely presents closer relationship to any other division. At the close of this chapter something further will be said on its probable position in the system. In order 13 make oui'selves more easily understood concerning the evelopmental processes, it seems necessary to discuss first 3me morphological conditions. ^ Anatomical. — Balanoglo.9sus possesses an elongated vermi- jrm body, on which different regions can be recognized externally. Anteriorly the so-called acorn [balanus], less ap- propriately called proboscis, is marked off from the rest of the body ; upon this follows the mascular collar, and then the branchial region, which gradually merges into the posterior part of the body (Fig. 164). Acorn and collar are essentially a locomotor apparatus, and therefore are largely composed of muscle fibres, which can be distinguished as external circular and internal longitudinal muscles. The cavities inside both organs which are left between the longitudinal muscles and connective-tissue cells can be filled with water from the out- side by means of one or two dorsal pores lying at the base of the acorn (Fig. 165 p). Similar pores also conduct water into the cavities in the collar (Spengel). These conditions have been compared to those of the water-vascular system of fl 1 [It should be mentioned here that since the publication of our de- scription of the development of Balanoglossus the important works of Spengel (No. VI.) and Morgan (Nos. II., IV.) have appeared, necessitating some modifications in the account which we have given. The most im- portant of these will be pointed out in what follows. — K.] 373 374 EMBRYOLOGY the Echino(^rmata, and it was indeed supposed that the acorn represented a rudiment of this system, especially since the cavity of the acorn in its earliest condition presents a certain resemblance to the water- vascular system as it origi- nates in the Echinodermata. It appears certain that the acorn serves as an organ of locomotion. It was believed that it took in water from the outside by means of the proboscis pore, and that there- ^^ kr aij. fore it operated in the same way as the am- bulacral feet of the Echinoderms (Spen- gel). However, it has been maintained, on the other hand, that particles of pigment distributed in the water are never found inside the acorn, and that the proboscis pore therefore does not serve for the reception, but only for the elimination, of sub- stances from the inside (Batesox). This ob- servation is of parti- cular interest, inasmuch as the acorn contains a glandular structure, which has been inter- preted as an excretory organ. The locomotion of Balanoglossus is effected by peristaltic movements on the part of the acorn, which thus pushes its way into the sand. The collar follows it, also taking part in the same way in the pro- gression of the animal. At the same time the sand enters the mouth-opening, which lies at the base of the acorn, Plj.-' Fig. 164. — Balanoglossus Kowalevsldi (after A. Agassiz). e, acorn ; fcr, collar; fc, branchial region ; g, genital region of the body ; Ch, dorsal, vh, ventral, blood-vessel. ENTEROPNEUSTA 375 gradually fills the entire intestine, and finally passes out through the anus at the posterior end of the body as a sausage-like cord. Thus the animal eats its way, as it were, through the sand. The intestinal canal commences immediately under the Fig. 165. — Sagittal section through the acorn and collar of BnlanogJossus imiensis (made somewhat diagrammatic, after Kohlkr). d, intestine; de, intes- inal epithelium ; db, dorsal blood-vessel ; di, diverticulum of the intestine ; dn, dorsal nerve ; h, the so-called heart; Ih, the body cavity; m, mouth ; p, proboscis pore; sk, skeletal body; vh, ventral blood-vessel; x, the so-called proboscis gland. acorn with the broad mouth-opening, which cannot be closed (Fig. 165). It extends backwards tolerably straight. 376 EMBRYOLOGY The appendicnlar structures which arise from the intestine, and in part remain intimately united with it, are important for the organization of the animal. The intestine produces in its anterior part a dorsal evagination, which extends into the base of the acorn (Fig. 165 di). Between the ventral wall of this evagination and the epidermis of the acorn is inserted the anterior part of the so-called proboscis- or acorn- framework (Fig. 165 sk), a skeletal body, the unpaired part of which has the position described, whereas two arms, which project from it backwards and downwards, embrace the fore-gut like a hoop. They lie in folds of the intestinal wall; they could not be introduced into the figure. The entire skeletal body, according to Kohler, Bateson, and Morgan, is a product of the intestinal epithelium, i.e., the above-mentioned evagination of it, [whereas Spengel (No. VI.) interprets the acorn-skeleton as only a modification of the bounding membrane (Grenzmembrafi), and also makes the coelomic sacs share in its formation. According to Spengel, the histological structure of the acorn-skeleton in no way agrees with that of the chorda dorsalis of vertebrates; notwithstanding, this has been repeatedly emphasized by different observers, and has been regarded as one of the points for comparison between Enteropneusta and Vertebrata (comp. infra). — K.] The gills, which are most important for the whole inter- pretation of the animal, occur somewhat further back on the intestine, behind the collar region of the body. They are paired, pouch-like outfoldings of the dorsal wall of the intestine lying on both sides of the middle line (Fig. 166 k). Each of these pockets, which are lined with cilia, sends upwards a short duct, which opens on the dorsal surface by means of a pore (Fig. 166 p). Externally the rows of branchial pockets can be recognized by the transverse arched bands (Fig. 164 k). Behind, these transverse arches become less extensive, which indicates that the formation of new gills takes place even in the later stages of the animal's life. A skeleton, formed of delicate chitinous hoops, which is embedded in the walls of the branchial pouches, serves as a protection for the gills. The water is taken in by the mouth, ENTEROPNEUSTA 377 to the outside world by means of the dorsal pores, ^ The intestine also presents paired dorsal evaginations in I the parts which lie behind the gill region. These are the hepatic appendages. They also influence the shape of the body, inasmuch as they cause the skin to protrude (Fig. 164), and the musculature is only slightly developed at these places. The hindermost portion of the intestine lacks the appendicular structures, and extends straight to the anus. One mesentery extending in the dorsal and another in the ventral line serve for the attachment of the intestine. By means of the mesenteries the body cavity is divided into a right and a left portion, but the two parts are confluent in 'perforation of the dorsal mesentery. The body cavity of the collar is dis- tinct from that of the trunk, and also differs from it in its mode of origin ; moreover, it is for the most part reduced by being filled with connec- tive tissue and muscula- ture (Fig. 165). In the trunk, on the contrary, the greater part of the body cavity is said to persist, and its wall is composed of the longi- tudinal and circular mus- culature of the somatic and splanchnic layers (Spengel), other statements, even the trunk cavity is said to lose the OJl. Fig. 166. — Transverse section through the branchial region of Balanoglossus minu- tus (after Spengei). d, intestine ; db, dorsal blood-vessel; dn, dorsal nerve; g, genital organ; fc, gill-pockets; Ih, body cavity ; p, pore of the gill-pockets ; so, somatic, sp, splanchnic, layer of the meso- derm ; vh, ventral blood-vessel ; vn, ven- tral nerve. However, according to 1 [A full account of the very complicated structure of the gills is given in Spengel's monograph (No. VI.), to which we particularly call attention in the matter of this and other anatomical conditions, and especially m view of the correction which it has since undergone. — K.] 378 EMBRYOLOGY nature of a true coelom and to be filled with connective tissue and muscles (Kohler). The two chief vascular trunks of Balanoglossus (Figs. 165 and 166 vh and dh) extend in the ventral and dorsal middle lines between the wall of the intestine and that of the body. . The blood flows forward in the dorsal vessel, backward in the ventraL They give off branches at regular intervals, which extend to the body-wall, to the intestine, and to other organs. According to Kowalevsky, there are also two lateral trunks which receive vessels from the intestine and from the gills. Their presence was confirmed by Kohler. It still appears doubtful whether the saccular structure lying at the base of the acorn and at least connected with the vessels of the body, which was maintained by Kohler and Bateson to be the central organ of the blood-vascular system, is to be looked upon as a heart. In Fig. 165 it (Ji) is seen lying on the dorsal side of the intestinal diverticulum. An organ (x) lies above it the significance of which is still less certain. It is a closed saccular body, the epithelial lining of which is greatly thickened anteriorly (Fig. 165 x). Owing to its intimate relation to the blood-vascular system, due to its position, Spengel looked upon the anterior part of this organ as an internal gill (acorn gill), whereas Bateson and Kohler explain it as a gland (proboscis gland), which has an excretory function. To be sure, a difficalty occurs with this explanation, namely, the absence of the efferent duct ; for it is not evident how the proboscis pore conveys away the products of this " gland," which is a closed sac. Apart from this, and in the absence of any other excretory organ, this interpretation is nevertheless natural. A thick cord, which lies in the dorsal mid-line of the collar, is to be looked upon as the central organ of the nervous system (Fig. 165 dn). It is said to possess a cavity which would be comparable to the central canal of the spinal cord of the Vertebrata (Bateson), but this is denied by Spengel. According to both Spengel and Kohler, the cavity is traversed by cords of cells, so that only irregular spaces appear in it. Kohler further states that the following pecu- liar condition exists : the internal space of the nerve cord EJTTEROPNEUSTA 379 opens to the outside at its posterior end, the cells of its walls merging into the epithelial cells of the body-wall. Similar i communications of the inner space are also said to exist at the anterior end (neuropore according to Batkson). A stout nerve, which extends along the entire dorsal mid-line of the body, is given off from the central organ. This in turn gives off two nerves just behind the collar, which extend down- wards (ventrad), unite in the region of the first pair of gill- pockets, and extend backwards in the body as the ventral median nerve (Fig. 166 vn). The genital organs of Balanoglossus either belong to the branchial region of the body or lie behind this. Balanoglossus is dioecious. Male and female organs are entirely alike as- regards form and position. The genital glands lie in the form of simple or branched tubes on both sides of the body, and their external openings are found one behind the other, forming two rows on the dorsal surface (Fig. 166 g). In addition to these lateral rows of sexual organs (Fig. 164 gr), two others (median) lying between the gill-pockets and the dorsal blood-vessel may make their appeai'ance. In many species the part of the body succeeding the gill-pouches can also be called the genital region, for the sexual organs are especially well developed there. Owing to the fact that the parts which contain the sexual glands undergo a great flatten- ing and lateral extension, wing-like extensions of the body are produced in certain species, e.g. B. claviger and B. minuhis-^ studied by Kowalevsky. Development without the Tornaria Larva. — The fertilization of the eggs takes place outside the body in the sea-water, into which in the American species studied by Bateson {Balanoglossus Kowalevslcii) both kinds of sexual products are said to pass by the rupturing of the body- wall. Artificial fertilization could not be under- taken, though Bateson found the eggs in large quantities in the slimy sand inhabited by the adult animals. The eggs are closely enveloped by a delicate membrane, which sepa- rates from the egg when fertilization has taken place. Cleavage is total and tolerably equal. A blastula arises as 380 EMBRYOLOGY the result of it, whicli is at first spherical, but subsequently becomes flattened on one side. On this side an invagination then takes place, and the result is a typical invagination gastrula. Soon, however, the originally wide blastopore contracts to a short, narrow fissure. At the same time the external surface of the embryo becomes covered with short cilia, and a circle of stouter cilia makes its appearance in the vicinity of the blastopore. Subsequently the blastopore entirely closes. The two primary germ-layers remain united at this point, but only by a plug of cells ; finally, they separate entirely from each other, so that the embryo then consists of two sepa- rate cell-vesicles lying one within the other. At the same time the embryo elongates somewhat and then assumes a shape such as is represented in Fig. 168 J.. At about this stage or even somewhat earlier, the embryo breaks through the egg-membrane and be- comes a free larva, which does not, however, lead a pelagic existence, but lives on the bottom and is found in places where the water is not very deep. The internal structure of the larva is soon changed in such a way that an in- ternal segmentation can be recognized. The elongated, com- pletely closed archenteron bulges out at its anterior end, and forms a pair of diverticula, which are directed back- wards (Fig. 167 Cj). These are constricted oif from the archenteron in connection with each other, and lie in front of it as closed vesicles. In the same way two pairs of coelomic sacs are formed further back as evaginations of Fig. 167.— Diagram of a longitudinal section through a larva of BalanoglosKus KowaUvskii (after Bateson). Ci, anterior, Cji, middle, cm, posterior, coelomic sacs ; d, intestine. ENTEROPNEUSTA 381 ^fi the archenteron (Fig. 167 c„ and Cuj) ; these also become detached, and are subsequently found next to the intestine ias flattened sacs. The mouth is formed at a somewhat later stage on the ventral surface at the point where a transverse furrow has made its appearance on the outside of the larva (Fig. 168 A). The anus arises at the posterior end of the larva near the place where the blastopore closed. Both mouth and anus are formed by the fusion of the inner with the outer germ-layer. During these processes the external shape of the larva undergoes important changes. At first a transverse furrow, which gradually deepens, and behind which a second one soon makes its appear- /J ance, arises at about the middle of the body (Fig. 168 A and B). While the first furrow, even as early as this, marks off the anterior portion — namely, the acorn — from the rest of the body, the second furrow, together with the first, bounds the future collar. Behind this the gill region is now also indicated by the appearance of two pores as evidence of the first pair of gill- pouches (Fig. 168 B and C, h). The part of the larva lying behind the ciliated band gradually elongates (Fig. 168 C). Thus the principal parts of the body of the adult animal are established in the larva even at this stage. Development by means of the Tornaria Larva.— Not all species of Balanoglossus, however, develop from the egg into the form of the adult animal in so simple a >Lr ^ A. Fig. 168.—^ to C, free-swimming larvje of Balaywglossus Kovmlevskii in different stages of development (after Batesoit). e, acorn ["pro- boscis"]; fc, branchial openings ; fcr, collar. 382 EMBRYOLOGY manner as that described above, for some species pass through a larval stage, the shape of which recalls the larvse of the Echinodermata. The larva called Tornaria was de- scribed bj JoH. MtJLLER as an Echinoderm larva. Its shape, which, moreover, exhibits modifications in the different species, is illustrated by Fig. 169. On the ventral side of the bell-shaped larva lies the mouth-opening, from which the oesophagus ascends, and then bends backwards, to be- come continuous with the capacious stomach. Upon this Fig. 169.—^ and B, Tornaria and later stage of development of Balanoglossus (after Kowalevsky, from Balfocr's Comparative Emhryology). The broad black lines indicate the ciliated band and the ring of cilia behind it. an, anus; hr, gill-pocket; c, body cavity; Ji^, " heart " ; m, mouth ; iv, the so-called water- vas- cular vesicle. follows the hind-gut, which opens to the exterior through the anus at the posterior end of the larva. The surface of the larva becomes engirdled by ciliated bands, which, how- ever, are distinguished from those of the Echinoderm larvae by their different parts acquiring a greater independence. In the first place, we distinguish a preoral from a post-oral ENTEROPNEUSTA 383 jiliated band, both of which are provided with several lexures (Fig. 169 A). They almost come in contact with mch. other at the anterior end of the larva. At this point Es found an ectodermal thickening, comparable to the apical >late of the Annelid larvae, with two eje-spots lying over it. 'rom this region a contractile band extends backwards, [esenchjmatous cells seem to arise between the intestine md the body- wall. The posterior part of the larva is en- jcircled by a ring of cilia which is independent of the other jiliated band (Fig. 169 A), and in later stages of the larva mother one may also make its appearance behind this. [The ciliated band of the anterior part of the body may undergo more )r less extensive outfoldings, by which the external form of the larva is [considerably modified. These outgrowths are sometimes large, such as ^e shall find in the larvee of Echinoderms, and these influence, as has tbeen said, the entire form of the body. Continual outgrowths of the ciliated band of limited extent result in the formation of tentacle-like structures. — K.] The condition of the larva described is only gradually reached during its free pelagic life. At first the transverse (posterior) rings of cilia are lacking, and the preoral and post-oral ciliated bands have a more simple course. In the further development of the Tornaria its anterior end elon- gates and becomes the acorn of the Balanoglossus. Preoral and post-oral ciliated bands then disappear, and instead of these the entire body becomes covered with cilia (Fig. 169 B). The eye-spots are still retained for a while at the tip of the anterior end. The middle region of the body is encircled by the transverse ciliated band, and it can thus be seen that the parts lying behind it have also elongated. Two openings make their appearance externally on the dorsal side of the anterior part of the body, the external openings of the gill-pockets. With this nearly the same stage is reached which we saw arising by direct means from the larva described by Bateson. The simpler mode of de- velopment is doubtless to be considered as the derived, and that of the Tornaria as the more primitive, since the absence of mouth and anus in a free-swimming larva does not represent a primitive condition. ::384 EMBRYOLOGY Further Developmental Processes of Both Types. — Thus far we have chiefly considered the external shape of the Tornaria. As regards the internal development, we find confirmed the processes which, following Batesox, we have already described. In the youngest stages of Tornaria yet observed (Fig. 170), the archenteron, which in this case never loses its connection with the ectoderm, develops an unpaired evagination. This is said to be the fundament of the so-called water-vascular vesicle, which, like the corre- sponding organ of the Echinoderm larva, opens out by means^ of a pore on the dorsal surface (Fig. 169 A). This in particular has given rise to a comparison with the Echino- derm larvae. In addition to this diverticulum, two pairs of evaginations arise farther back on the intestine (Agassiz). They are the coelomic sacs, which soon become detached from the intestine, and lie close to it on either side as two pairs of flattened vesicles. They soon become considerably enlarged, and then their walls are applied to the wall of the body and that of the in- testine as the somatic and splanchnic layers.! A mesentery is developed dorsally and ventrally, separating the sacs of the body cavity of the two sides ; but, according to Spengel, the dorsal mesentery may afterwards de- generate. The hinderraost pair of coelomic sacs supplies the greater part of the body cavity, — namely, that of the entire trunk, — whereas the body cavity of the collar arises from the anterior pair, and the cavity of the acorn is developed from the so-called water-vascular vesicle (Spengel). The re- semblance of the latter to the paired structures points to the fact that they originally had the same significance, and tbe appearance of two acorn pores in Balanoglossus Kupfferi 1 As is to be seen in Fig. 172 (p. 387), entirely similar conditions appear also in Balanoglossus Kowalerskii, which does not develop by means of a Tornaria. Fig, 170.— Early stage of a Tornaria (after Goette, from Balfotte's Compara- tive Emhryology),. m, mouth ; art, anus ; W, so-called water-vascular vesicle. ENTEROPNEUSTA 385 seems to indicate the paired origin of even this anterior coelomic sac. The acorn pore arises from the dorsal pore of the Tornaria. In the Balanoglossus which does not pass through the Tornaria stage, the anterior coelomic sac, as Batesox unreservedly calls it, forms a pointed process, the end of which fuses with the ectoderm and breaks through to the exterior. After the coelomic sacs have separated from the archen- teron, the remaining entoderm produces, in the form of a forward evagination, the intestinal diverticulum, which lies at the base of the acorn (comp. Fig. 165 di), and from this the formation of the acorn stalk probably takes place. Even earlier than this, the gill-pockets develop as paired evaginations from a portion of the intestine behind the diverticulum. They are directed toward the dorsal surface (Fig. 169 B), with which they soon unite, since they open to bhe exterior by means of pores which are at first rather large (Fig. 168 C). In several forms at first only one pair of gill- pockets is to be observed (Figs. 168 G and 169 B) ; in the Tornaria studied by Agassiz, on the contrary, four pairs of khem make their appearance simultaneously (Fig. 171). [The form of the gill-pockets, which is at first so simple, is later much more complicated, for their walls become Folded, and the skeletal hoops are developed between them. The formation of new gill-pockets continues to take place even when the Balanoglossus has long since assumed its permanent shape. After the posterior part of the body has considerably increased in length, the paired evaginations of [the intestine which have been interpreted as hepatic append- [ages arise behind the gill region. In the Tornaria there is found next to the so-called water-vascular I vesicle, or even sunk into a depression in it, a spherical vesicle, which ordinarily is called the heart of the Tornaria (Figs. 169 B and 171 ht [not h, Fig. 165]). It does not merit this name, for, according to Spenqel, it is developed into the organ which Bateson and Kohler have called the saccular posterior portion of the " proboscis gland" (comp. IFig. 165). In Balanoglossus Koioalevskii the proboscis gland arises by fdelamination from the tissues of the anterior coelomic sac after this has already spread out inside the acorn. This mode of origin seems to us K. H. E. C C 886 EMBRYOLOGY to indicate that even the so-called heart of the Tornaria might arise from the water-vascular vesicle or, what is the same thing, from the anterior coelomic sac.^ The early appearance of the organ in the Tornaria is favorable to the explanation (excretory) which has been given to the fully developed organ. In other divisions of the animal kingdom also we see the excretory system established at a very early period. Likewise the organ which is interpreted as the real heart first appears, according to Bateson, as a fissure in the mesodermal tissue. This fissure makes its appearance between the fundament of the "proboscis gland" and the intestinal diverticulum, and is only gradually surrounded by a firm wall. It has not been determined whether or not it is from the beginning connected with the blood-vessels of the body. The blood- vessels probably arise from the mesoderm in the same way as the supposed central organ. [The observations of authors are not in agreement respecting the origin of the coelomic sacs, for Spengel (No. VI., Appendix to Litera- ture on Enteropneusta) and Boukne (No. I., Appendix) maintain that they arise from the hind-gut, whereas Morgan (Nos. III., IV.) refers them to the entoderm, as was formerly done. Morgan, moreover, assumes a different method of origin in the different species of Balanoglossus, for in one case (Tornaria from the Bahamas) he even refers them to scattered mesenchyma cells of the primary body cavity. Also the origin of the so-called heart-sac is not yet sufficiently clear. Morgan would refer it likewise to an accumulation of mesen- chyma cells, whereas Spengel adheres to his former account of the ectodermal origin of this organ. Moreover, it seems to be im- possible to re«oncile the new results with the earlier account, and for this reason we must refer to the original articles. — K.J The earliest fundaments of the genital organs occur as pyriform sacs, and are found in close connection with the ectoderm, a fact air Fig. 171.— Stage of de- velopment of Balanoglossus (after Agassiz, from Bal- four's Comparative Embryo- logy), an, anus; hr, gill- pockets ; ht, "heart"; m, mouth ; W, the so-called water-vascular vesicle. 1 Spengel, however, affirms that the " heart " is formed as a thicken- ing of the epidermis next the acorn pore ; but perhaps this statement can be harmonized with the opinion expressed above by assuming that in this case the development of the so-called heart took place later, that is to say, when the lining of the internal cavity of the acorn by means of the water-vascular vesicle had already been accomplished. This, however, is only conjecture. ENTEROPNEUSTA 387 Ihich caused Bateson to believe them derh^ed from it, and not from the lesoderm, as would seem more natural, especially since at this time le mesodermal tissue is already found closely applied to the ectoderm, however, Bateson states that the origin of the genital organs is as yet not certainly determined. \^^m [Spengel found the union of the nascent gonads with the ectoderm less ^^^Kntimate, and was inclined to refer their production to the mesenchyma ■"■ of the body cavity. The connection with the ectoderm is only secondary. According to Morgan, they arise as pa-oliferations of the wall of the coelomic sacs of the trunk, which at first have no connection with the ectoderm. — K.] The central part of the nervous system arises, according to Bateson, as follows : — A part of the cells of the deepest layer of the ectoderm in the median line of the collar is X C^' Fig. 172.— Transverse section through the anterior part of the collar of a larva )f Balanoglossus Kowalevskii which is at about the stage of Fig. 168 B (p. 381) J( after Bateson). Above is seen the dorsal ciliated groove, d, intestine ; n, funda- ment of the nerve cord ; cu, cavity of the middle coelomic sac, which is already ipplied to the wall of the body and that of the intestine as somatic and splanch- lic layers. [differentiated in a pecaliar way, and is finally detached from the ectoderm along the entire length of the collar (Fig. 172 n). This process, moreover, is said to be accompanied by a superficial depression of the ectoderm, which is notice- able as a dorsal, longitudinal ciliated groove on the recently leveloped collar of the young larva (Fig. 172, after Bateson). [Spengel also speaks of the development of the nervous system as the result of a dorsal invagination in the middle 388 EMBRYOLOGY line of the collar. The cavities, however, which occur in the central nervous system of the adult animal, are not to be referred to a formation comparable with the neural tube of the Yertebrata, but arise in the cell-layer which was split off from the ectoderm, in all probability by the appearance of fissures. However, at the end of the central cord, where it merges into the indifferent cells of the ectoderm and where the latter is considerably thinner, a kind of folding process seems to take place ; at this point also the lumen of the central cord is said to communicate with the outside world (neuropore ?) . There seems to be no relation between the dorsal groove and the blastopore ; for the groove does not extend so far back. A direct connection with the conditions occurring in the Chordoma is therefore not in- dicated by this (comp. infra)} Like the chief, central parts of the nervous system, its peripheral portions are also differentiated from the lower cell-layers of the ectoderm, which, according to Bateson, everywhere exhibits large accumulations of sensory cells. General Considerations. — The external resemblance of the Tornaria to the Echinoderm larvae and the oc- currence of the water- vascular vesicle, opening out by means of a dorsal pore, have caused Balanoglossus to be brought into relation with the Echinodermata. Corre- spondingly, the acorn, the lining of which is supplied by the so-called water- vascular vesicle, has been explained as the last remnant of the water- vascular system, as the single remaining ambulacral tentacle. The nature of the skin, provided with calcareous structures, is, in addition to the water-vascular system, characteristic of the Echinodermata. The entire absence of calcareous bodies in Balanoglossus and the different condition of the skin, together with the other peculiarities in the entire structure of the body, ^ [According to Spengel's description, it must be assumed that the account given by the previous observers does not at all relate to the first fundament of the nervous system, and that in its formation, which takes place at an early period, there is no invagination. Differentiations of the ectoderm without any invagination give rise to the nervous system, which only subsequently sinks in deeper. — K.] ENTEROPNEUSTA 389 lo not allow ns to put any great weight on such an in- jrpretation of the acorn. A comparison of the Tornaria dth the Echinoderm larva is difficult to carry out, for the jiliated bands so characteristic of the latter present here [uite a different distribution. Moreover, the 'Tornaria ap- )ears to possess a kind of apical plate, which is absent in bhe Echinoderm larvae. The latter likewise exhibit no [eye-spots. The resemblance between the Tornaria and the [Echinoderm larvae is therefore of a rather superficial nature, 'he possession of an apical plate and the cords radiating Prom it point rather to relationships of the Tornaria with the Trochophore. The occurrence in Balanoglossus of paired coeloraic sacs, lying one behind the other, indicates a segmentation. In t;his, it is true, a resemblance to the Echinodermata would jxist, if the statement should be confirmed that in the [latter also several pairs of ccelomic sacs are developed (comp. p. 414). This internal segmentation of the larva subsequently disappears, and the segmentation which can be recognized on the adult Balanoglossus has nothing to [do with it. In searching through the animal kingdom after relationships for balanoglossus, characterized as it pre-eminently is by the possession ►f gills, a comparison with the Chordata has been reached ; but there is IS yet no adequate basis for this comparison. It is Amphioxus which [has been especially in mind, and the comparison has been based chiefly [on the gills, on the intestinal diverticulum, called by authors the [chorda, and its skeletal body, and on the formation of the nervous system. A striking resemblance is noticeable between the anterior 3oelomic sacs of Balanoglossus and the most anterior archenteric diver- ticula of Amphioxus, which also make their appearance very early, and [one of which greatly enlarges and opens to the exterior by means of [a ciliated canal, like the so-called water-vascular vesicle or the anterior [ccelomic sac in Balanoglossus. [To what precedes we append the following : The supposed relations 'between Enteropneiista and Chordata have become, according to recent observations at least, very doubtful. The diverticulum of the intestine in the acorn, or rather the acorn skeleton, is, as it appears, comparable [with the chorda neither in regard to its origin nor structure. The needed igreement in the formation of the nervous system seems, in fact, to be ranting. The gills of Balanoglossus, it is true, are strikingly similar 390 EMBRYOLOGY in structure to those of Amphioxus ; but a detailed comparison reveals in them organs of different phylogenetic origin (Spengel). The ccelom of the Enteropneusta exhibits other conditions than in the Chordata. The probability of the relationship of Balanoglossus to the Echinoderms, on the contrary, has, it may be said, been increased by recent investiga- tions. A comparison of Tornaria with the larvse of Echinoderms, so far as regards the external form, and particularly the structure, is perhaps still possible. The apical plate present in Tornaria is, according to recent observations, also found in the larv«e of Echinoderms — e.g., in Antedon — although in a greatly reduced condition. The coelom seems to be segmented in the Echinoderms as well as in the Enteropneusta— a fact which is certainly of importance. However, it is precisely the interpretation of the coelom in Balanoglossus which is not yet adequate, since the different portions of the intestine, from which it takes its origin, are not fully understood. Unfortunately it must be admitted that even yet the relationships of Balanoglossus are shrouded in dark- ness.— K.] Literature. 1. Agassiz, a. Notes on the Embryology of Starfishes (Tornaria). Ann. Lyceum Nat. Hist. New York. Vol. viii. 1867. 2. Agassiz, A. The History of Balanoglossus and Tornaria. Mem. Amer. Acad. Arts and Sciences. Vol. ix. 1867. S — 5. Bateson, W. Early and Later Stages in the Development of Balanoglossus. Quart. Jour. Micr. Sci. Vols. xxiv. — xxvi. 1884—1886. 5a. Bateson, W. The Ancestry of Chordata. Quart. Jour. Micr. Sci. Vol. xxvi. 1886. 6. Fewkes, J. W. On the Development of Certain Worm Larvse. Bulh Mus. Comp. Zuol. Harvard Coll., Cambridge, 3Iass. Vol. xi. 1883 — 1885 (description of a Tornaria found at Newport). 7. GiAKD,A. Systematic Position of Balanoglossus. Jour. Roy. Micr. Soc. Vol. ii. 1882. 8. Goette, a. Vergleichende Entwicklungsgeschichte der Comatula Mediterranea. Arch. Mikr. Anat. Bd. xii. 1876. d. Koehlek, E. Contributions a I'etude des Enteropneustes. Inter- uat. Monatsschr. Anat. u. Hist. Bd. iii. 1886. 10. Koehleb, Rr Sur la parente du Balanoglossus. Zool. Anzeigcr. Jahrg. ix., No. 230, p. 506. 1886. 11^ KowALEvsKY, Alex. Anatomis des Balanoglossus delle Chiaje. Mem. Acad. St. Petersbourg. Ser. 7, tom. x. 1867. 12. Marion, A. F. Etudes Zoologiques sur deux especes d'Entero- pneustes, etc. Arch. Zool. exper. et gen. Ser. 2, tom. iv. 1886. 13. Metschnikoff, E. Untersuchungen iiber die Metamorphose einiger Seethiere. I. Ueber Tornaria. Zeitschr. u-iss. Zool. Bd. xxii. 1870. ENTEROPNEUSTA 391 14. Metschnikoff, E. Ueber die systematische Stellung von Balano- glossus. Zool. Anzeiger. Jahrg. iv., Nos. 78, p. 139, and 79. p. 153. 1881. 15. MuLLER, J. Ueber die Larven und die Metamorphose der Echino- dermen. Ahhandl. Akad. Wiss. Berlin. 1849, 1850. 16. ScHiMKEWiTSCH, W. Ueber Balanoglossus Mereschk. Zool. Anzei- ger. Jahrg. xi., No. 280, p. 280. 1888. 17. Spengel, J. W. Ueber den Bau und die Entwicklung des Balano- glossus. Amtl. Ber. der 50. Vers. Deutsch. Naturf. u. Aerzte in MUnchen. 1877. 18. Spengel, J. W. Zur Anatomie des Balanoglossus. Mittheil. Zool. Stat. Neapel. Bd. v. 1884. Appendix to Literature on Enteropneusta. I. Bourne, G. C. On a Tornaria found in British Seas. Jour. Marine Biol. A^soc. Ser. 2, vol. i. London. 1889. II. Morgan, T. H. The Growth and Metamorphosis of Tornaria. Jour. Morphol. Vol. v. 1891. III. Morgan, T. H. Balanoglossus and Tornaria of New England. Zool. Anzeiger. Jahrg. xv. 1892. IV. Morgan, T. H. The Development of Balanoglossus. Jour. Morphol. Vol. ix. 1894. V. KiTTER, W. E. On a New Balanoglossus Larva from the Coast of California, etc. Zool. Anzeiger. Jahrg. xvii. 1894. VI. Spengel, J. W. Die Enteropneusten. Fauna u. Flora des Golfes von Neapel. Mem. xviii. 1893. CHAPTER XIV. ECHINODEKMATA.i Development in the five divisions of the Echinodermata offers so much in common that we shall treat of these together as far as possible. In the development of the Echinodermata we distinguish the following four periods : — 1. The formation of the primary germ-layers and the mesenchymal together with the establishment of mouth and anus. 2. The origin of the enterocoele and hydrocoele. 3. The formation of the typical larval form. 4. The metamorphosis of the larva into the Echinoderm. I. THE FORMATION OF THE PRIMARY GERM-LAYERS AND THE MESENCHYMA, TOGETHER WITH THE ESTABLISHMENT OF MOUTH AND ANUS. As far as is known, the cleavage of the Echinoderm egg is always total. In the Holothurioidea (Synapta), more- over, it is strictly equal ; whereas in the star-fishes and sea-urchins it takes place less regularly. Within the mass of cells which has arisen by cleavage there is found even during this period a cavity which, in ,the further course of cleavage, continues to enlarge, and becomes an extensive blastocoele. The result of cleavage is always a ccbIo- hlastula. The next stage of development, too, exhibits an essential agreement in the different groups of Echino- derms, for in all of them it consists of an invagination gastrula. In the details, however, certain deviations from the common plan of development occur in the different forms. ^ The maturation and fertilization of the Echinoderm egg will be discussed in the general part. 392 ECHINODERMATA 393 I'Holothurioidea.— The earliest stages of development re the simplest in the Holothurioidea. We follow Selenka's fcount (No. 54) of the development of Synapta digitata. Cleavage is entirely regular. By means of the first vision the egg is halved. Since the newly formed blas- tomeres always divide into equal parts, this process being repeated nine times consecutively, there finally arises a stage the prismatic cells of which are approximately equal in size, and arranged in the form of a hollow sphere. They have already acquired cilia, although the hlastula is still enclosed by the vitelline membrane. (A similar stage in Holothuria is shown in Fig. 180 ^.) At this stage the Fig. 173.— Blastosphere of Synapta digitata at the beginning of gastrulation, still lying within the egg-membrane (after Sklenka). farther division of the blastomeres is suspended for a con- siderable time, only subsequently to proceed slowly at the vegetative pole, and at first at this pole alone. Gastrulation is initiated by means of this cell-proliferation realized at the vegetative pole (Fig. 173). The result is a regular gastrula with a small archenteron (Fig. 174). In this stage the embryo becomes a free-swarming larva, which moves about by the aid of its long cilia. The gastrula very soon undergoes a change, for the archenteron bends towards the wall of the gastrula, and unites with the ectoderm 2 ■ 394 EMBRYOLOGY (Fig. 175). This region corresponds to the dorsal surface of the larva. After the fusion of ectoderm and entoderm, the lumen of the archenteron communicates with the outer world, thus establishing the so-called dorsal pore (Fig. 176). Johannes Muller, who, even in his time, was acquainted with this process, considered the dorsal pore to be the mouth, of the larva ; but that is not its fate, for the archenteron soon separates into two portions, of which the one connected with the dorsal pore constitutes the funda- FiG. 174, Figs. 174 andl75.— Gastrula stages of Synapta digitata (after Sblenka). Tn Fig. 175 the mesenchyma begins to develop. Bl, blastopore. ment of the water- vascular system and body cavity, while the other becomes the intestine. With the multiplication of its cells the latter acquires a knee-like bend (Fig. 177), and, while increasing in length, turns toward the ventral side. Even before it reaches this, the communication be- tween the upper and lower portions of the archenteron is interrupted (Figs. 177 and 178). Of these two portions ECHINODERMATA 395 only the lower interests ns for the present. Its blind end enlarges a little, and comies into contact with the ventral wall of the larva (Fig. 178). The corresponding part of the wall sinks in, forming a cup-like depression (Fig. 179), the intestine fuses with it, and an opening leading into the intestine now breaks through at this point: the .mouth-opening of the larva. The mouth-opening is there- ifore a new formation. The intestine opens to the outside at the posterior end by means of the blastopore ; the Fig. 176. Figs. 176 and 177.— Larvae of Synapta digitata, showing the formation of the dor- sal pore (P) and the vaso-peritoneal vesicle (after Sblenka). Bl, blastopore. l^astrula mouth consequently has become the anus of the larva. With these changes the larva has also undergone a certain differentiation in external shape. Its bilateral symmetry is already expressed. The mouth and anus mark definite regions of the body. The former lies on the ventral surface, the latter at the posterior end of the larva. Moreover, as we will mention in anticipation of the sequel, the ventral surface 396 EMBRYOLOGT is ordinarily flattened somewhat, whereas the dorsal side is more convex. Before concluding the consideration of the first develop- mental processes of Synapta, we must refer to a process which takes place before the change in position of the arch- enteron already described ; this is the formation of the mesen- clnjma. At the time when the blind end of the archenteron begins to bend toward the dorsal surface, there appear at its apex two cells which project out beyond the other cells (Fig. 175), and are called by Selenka the two primitive cells of the P Fig. 179. Figs. 178 and 179.— Larvae of Synapta digitata, showing the formation of the intestine and vaso-peritoneal vesicle (after Selenka). Bl, blastopore; M, mouth; P, dorsal pore. mesenchyma. These cells then separate from their connec- tion with the archenteron, migrate into the blastocoele, and apply themselves to the ectoderm, but not at predetermined points. Subsequently a large number of such mesenchyma- tous or migratory cells are found in the blastoccele (Figs. 176 to 179). According to Selenka's description, they arise by the division of the two primitive mesenchyma cells ; but the process of mesenchyma formation in other Holothurioidea a «p I EGHINODERMATA 397 shows that it is not two primitive mesenchyma cells that give rise to the entire mesenchyma, bat that a large num- ber of cells separate from their connection with the others and migrate into the blastocoele, where they subsequently increase in numbers (Fig. 180 B). In Cucumaria doUolum and Holothuria tnhulo, development of the AuriculaYia from the fundamental form of the Echinoderm larva^ (diagram after Joh. Mijller, from Balfoub's Com-parative Embryology), The broad black line indicates the ciliat'^d band, the shaded area the depressed part of the surface, an, anus; m, mouth. from the tracts extending lengthwise on both sides, the so-called longitudinal parts of the ciliated band (Figs. 198 A and B). The anus lies near the posterior pole of the larva. In its further development the body of the larva, in front as well as behind, becomes more hollowed out on both sides, while the elevated parts of the ventral surface persist and grow toward each other (Fig. 198 B and C). In this way there results a larval form, on the ventral side of which an 1 The manner of orientation chosen by Joh. Muller has been retained in this and in the following diagrammatic figures (200, 202, and 203) for practical reasons only. It would be better if they were placed with the mouth upwards and the anus down, as has been done, for example, in Figs. 199, 204, 209, and 211. K. H. E. E E 418 EMBRYOLOGY anterior, so-called preoral, and a posterior, anal area can be distinguished as elevated parts from the depressed portions (Fig. 198 G and D). At the anterior and posterior ends the two areas bend around toward the dorsal surface. Fig. 199 ^ shows a larya at about this stage as seen from the side. The further development of the shape of the larva is finally attained by the extension of the depressed, or hollowed- out, region more to the periphery, and by the production of lobular processes at the mar- gins of the body, owing to the outgrowth of cer- tain parts (Fig. 198 D). Calcareous deposits, having the shape of delicate miniature wheels, may make their appearance in these ear -like appendages (Fig. 205, p. 426). Along the periphery of the lobes runs an uninterrupted ciliated band, which borders the two ventral areas as well as the dorsal surface. In each of the two depressed lateral surfaces of the Auricularia larva lies a structure resembling a ciliated band ; but these structures bear no relation to the ciliated band itself. Each of these two bands exhibits the form of a blunt angle opening toward the ventral surface. The cords consist of ciliated cells and fine longitudinal fibres lying under them. Strands of fibres pass from them to the ciliated band. Accordingly Metschnikoff (No. 37) and Semon (No. 55) interpret the two cords as the central nervous system of the larva. They also occur under similar cir- cumstances—to anticipate— in the Pluteus larvae of the Ophiuroidea. On the other hand, corresponding structures do not occur in the larvae of the Echinoidea and Asteroidea. According to Semon, however, fine fibres, similar to those in the nerve cords of the Auricularia larvae, occur in the ciliated band of these larvae, so that the nerve apparatus would be connected with the ciliated bands in the same way as in the larvae of the Annelida (comp. p. 266). [A very large Holothurian larva (Auricularia nudibranchiata), which Fis. 199.— -4 and B, larva of a Holothurioid and an Asteroid respectively seen from the side (from Balfour's Comparative Emhryology] , a, anus ; l.c, ciliated band ; m, mouth ; pr.c, adoral band of cilia of the Bipinnaria ; st, stomach. ECHINODERMATA 419 Attains a length of 6 mm. and is characterized by the complicated form of Kks ciliated band, has been recently described by Chun (No. VII., Appendix ^^mjiterature). The ciliated band of this larva is extraordinarily tortuous, ^nd exhibits arabesque-like foldings. Somewhat similar conditions were previously mentioned in the case of a very large Tornaria. Chun's paper, which is important in many respects, contains an account — to which attention may be called here— of a sac-like invagination of the hind-gut of the larva, from which, according to Chun's conjecture, the respiratory trees of the Holothurian may arise. — K.] The Aaricularia larva does not occur in all the Holo- thurioidea. Thus, for instance, the larva of Cucumaria doliolum at the time of the formation of the mouth assumes a cylindrical form (Selenka). The fiagella disappear zone by zone, until the larva retains only four to five bands of cilia, a ciliated anal area, and a ciliated cephalic zone. With this the so-called pupal stage is reached, which does not make its appearance in the development of other Holothurians until later (corap. p. 427). Another Holothurian, Psolinns brevis, develops, according to Kowalevsky (No. 28), alto- gether without a metamorphosis. The young Holothurians arise directly from the eggs, which are laid in the sea- water. In Phyllophorus urfia the larvae, which are probably com- pletely and uniformly ciliated, are said to swim about in the body cavity of the parent. When they abandon the parent, they already possess five tentacles and two feet. A similar condition is found, according to LuDWiG (No. 33), in the likewise viviparous Chirodota rotifera. Asteroidea. — The larval form of the Asteroidea, like that of the Holothurioidea, can be derived from the funda- mental form. If Figs. 200 B and 198 C, from JoH. Muller's diagrams, are compared, one sees that in the Asteroid larva the preoral area of the ventral surfaces, together with the part of the ciliated band surrounding it, is isolated. The de- pression on the ventral surface is continued farther forward here than in the Holothurian larva. In this way the con- nection of the preoral area with the dorsal surface is inter- rupted, and the ciliated band is separated into two parts. Thus two ciliated bands arise, which, from their positions, may be designated as the adoral and adanal (Figs. 200 A 420 EMBRYOLOGY and 199 B). Of these the latter is much the longer (Fig. 200 A to D). By the bulging and growing outward of the peripheral parts of the larva, there arise longer and shorter processes, which are bordered by the ciliated bands (Fig. 200 G). This larval form received from its discoverer, Sars, the name of ^^ Bipinnaria'' (asterigera), which it continued to bear even after its relation to the starfishes was recognized. Fig. 200.— Development of the Bipinnaria and Brachiolaria from the funda- mental form of the Echinoderm larva (diagram afterJoH.MiJLLEK, from Balfoxir's Comparative Embryology). The broad black line indicates the ciliated band, the shaded part the depressed portion of the surface. (Comp. footnote on p. 417 in regard to the orientation of the figures.) an, anus; m, mouth. The opinion which Semon advances concerning the origin of the Bi- jnnnaria larva does not agree with the description of it which we have just given, Semon finds in the Echinoderm larva a ciliated band surround- ing the mouth, a loop of which occasionally extends into the oesophagus (as in the Auricularia). This " adoral " ciliated band has nothing to do with the continuous ciliated band, but exists independently of it. In the Bipinnaria the "adoral" ciliated band of Semon is said also to supjjly that part of the ciliated band which we called the adoral part, so that the latter is not, as we described it, to be looked upon as a detached part of the continuous ciliated band. As long as strict proof of such a mode of origin is not forthcoming, we are unable to accept this opinion. The agreement of the preoral area in Auricularia and Bipinnaria is too striking for one not to assume that its isolation took place by means of ECHINOBERMATA 421 "deeper and deeper incision on the part of the lateral depressions. ^ (Comp. also in this connection Figs. 200 and 198, as well as Fig, 199 A and B.) Joh. Muller figures Auricularians in which the two depres- sions almost meet at the anterior end of the larva. Furthermore, the breaking up of the ciliated band discovered by Semon and the metamor- phosis of the component parts into the epithelium of the fore-gut men- tioned by him make the band appear to be more probably an oral ciliary apparatus, serving for the capture of food (comp. p. 427). The Brachiolarla arises from the Bipitmaria of the star- fish as a subsequent stage by the formation of two addi- tional processes at the base of the longer (dorsal) appendage (Fig. 200 D). In this way are formed the so-called Brachiolarian arms, which are different from the others. They are not bordered by a ciliated band, but possess wart- like elevations at the ends, which probably serve the larva for attachment in later stages. In the starfi^shes, too, many exceptions to the typical form of the larva are found. This is the case in Asterina gibbosa, the development of which has been made known through the thorough researches of H. Ludwig. The larva, which at first is pyriform, acquires a ridge-like thickening, en- Fis. 201.—^ and B, larv* of Asterina gibhosa (after Ludwig). A, a younjrer stage seen from in front; B, older stage seen from the side; Lo, larval organ; ttt, mouth. closing a depressed area, at the anterior end (Fig. 201 Lo). This thickening finally acquires a volume surpassing that of the rest of the larval body (Fig. 201 B). The peculiar organ consists of two lobes, and since the anterior 1 [This interpretation is confirmed by the investigations carried on since the above statement was written. — K.] 422 EMBRYOLOGY one of these occasionally divides, a certain resemblance to the Brachiolarian arm is produced. For in this way there arise two lobes, which lie symmetrically in front of the mouth, and a third unpaired lobe, which is farther removed from the mouth-opening. But the arms of the Brachiolaria which lie in front of the mouth have a similar position, and therefore Ludwig homologizes the larval organs of Asterina with the latter. The organ is muscular, and serves the larva for attachment. Similar, but multifid, larval appendages have also been described by Sars, Joh. MiJLLER, Agassiz, THOMPSON, and others, for Echinaster and Asteracanthion (Millleri) . Nothing more detailed has been learned of a vermiform Echinoderm larva described by Joh. Mullee, which was divided into five segments by transverse constrictions, and to the under-surface of which a five-lobed star was attached. According to JoH. Muller's account, it develops into a star- fish. Ophiuroidea. — The Fluteus larvae of the Ophiuroidea ex- hibit an essentially different shape from that of the larvae thas far considered. But they also arise from the same fundamental form. As in the cases previously considered, there is a continuous ciliated band, which borders the deep depressions of the body (Fig. 202 A). The subsequent characteristic shape of the larva depends, in the first place, upon the fact that the anal area increases considerably in extent, while the preoral area, on the contrary, almost entirely disappears (Fig. 202 B). Apart from this, the shape of the larva is determined by the long processes into which its peripheral portions grow out (Fig. 202 C and D) . These are bordered by the ciliated band, which is still, and always remains, continuous. As the form of the Fluteus is reached, the anal area becomes pointed (Fig. 202 D). The two ventral, posterior arms are especially well developed. They are also significant for the reason that they are always present, whereas the other arms may be more or less sup- pressed. The Fluteus larvae, unlike the Auricularia and Bipinnaria, possess a calcareous skeleton. As early as the gastrula ECHINODERMATA 423 stage of the larva two triradiate calcareous bodies are secreted by the mesenchyma cells in the vicinity of the blastopore. These soon elongate owing to the activity of the raesenchyma cells. They increase considerably in lume, branch, and send out rod-like processes into the ^ 'ms as soon as the latter are developed. The calcareous rods fuse at the posterior end of the larva, and appear to be united here by a kind of ring having a transverse position (Fig. 211, p. 438). In this way thei-e arises an excellent supporting apparatus for the larva and its appendages. Fig. 202.—^ to 1), evolution of the Ophiuran Pluteus from the fundamental form of the Echinoderm larva (diagram after Joh. Mullkr, from Balfour's Comparative Embryology). The broad black line indicates the ciliated band, the shaded area the depressed part of the external surface- In regard to the orientation, what is said on p. 417 (footnote) applies here also, an, anus; m, mouth. The other letters refer to the nomenclature of the appendages, which is not further considered here. AmpMiira squamata develops without any real meta- morphosis. AmpJiiura is viviparous. The earliest develop- mental processes are nearly the same as those that we have already learned about. There arises an oval embryo, which assumes a bilaterally symmetrical shape, but which does not develop into a ciliated larva, passinof, on the contrary, directly into the five-rayed star. The young, even at the time when they come into the world, exhibit the organization of the parent. It is interesting, however, that, despite this, the larval skeleton of the Pluteus is begun in the embryos. 424 EMBRYOLOGY This points to the fact that even in Amphiura or its an- cestors a metamorphosis took place, which, however, was abandoned owing to a change in the mode of life. Echinoidea. — The larva resembles that of the Ophi- uroidea, and, like it, is called a Pluteus. In it also the anal area preponderates on the ventral surface. The ciliated band is simple. A calcareous skeleton is found inside the body and its appendages (Figs. 204 and 212, p. 440). The derivation of the Echinoid Pluteus from the funda- mental form is nearly the same as in the Ophiurans, and is explained by the diagrams of Fig. 203. The shape of the different sea-urchin larvae is quite varied, according to the greater or less development of the arms. The larvae of Fig. 203.— Evolution of the Echinoid Phi.teus from the fundamental form of the Echinoderm larva (diagram after Joh. Muller, from Balfoue's Comparative Em- bryology). For further particulars consult the explanation of Fig. 202. Echinus and Spata7igus may be distinguished as particularly characteristic forms. On the anal area of the former, after the development of all eight processes, the so-called ciliated epaulettes make their appearance (Fig. 202 D). These are two pairs of ciliated projections of the body, which lie on either side immediately behind the ciliated band, but isolated from it. According to A. Agassiz, they should be interpreted as detached parts of the ciliated band. The larvae of Spatangus do not possess the ciliated epau- lettes, but have three processes on the anal area (Fig. 203 ECHINODERMATA 425 E), which are siapported by calcareous rods, like the other processes of the body. In the Pluteus of Arbacia there are only two processes on the anal area (Fig. 204), but they are particularly long. Furthermore, in addition to the ordinary Pluteus-arms, it possesses two pairs of auricular processes Fig. 204) which, like the arms, are surrounded by the siliated band (Joh. Muller, Fewkes). A pedicellaria of the future sea-urchin can already be recognized on the anal area of this larva. Fig. 204.— Pluteus larva of Arbacia pustulosa (after Joh. MifLLEK). P, pedi- cellaria. The skeletal rods are dark. The skeletal parts are developed very early as products of the me- senchyma (Selenka, No. 53 ; Ludwig, No. 34). There is first secreted between two cells a calcareous concretion, which soon enlarges and becomes triradiate. The skeletogenous cells then migrate along their respective rays, gradually moving farther away, while they continue to secrete calcareous salts. In this way finally arise the long skeletal rods, which may be many times branched and perforated like a network (Figs. 204 and 212, p. 440). The typical larval form may also be omitted in the sea- urchins. Thus A. Agassiz (No. 4) describes a viviparous 426 EMBRYOLOGY Spatangoid, Hemiaster australis, the eggs of which develop within the ovary, and then pass into a kind of brood-cavitj, which lies over the ambulacra! furrow and is formed of close-set and connivent spines. Here the young sea-urchins undergo direct development. IV. THE METAMORPHOSIS OF THE LARVA THE ECHINODERM. INTO The metamorphosis of the larva into the Echinoderm takes place most simply in the Holothurioidea, and therefore we consider this group first. Fig. 205.—^ and B, metamorphosis of the Auricularia larva of Synapta dlgit>tta into the pupal form (after Semon). a, anus; ed, proctodaeum ; eat, enterocoele ; kr, calcareous wheels; m, oral funnel; mg, stomach; n, nerve bands; u', water- vascular ring with the evaginations (tentacular and radial vessels). Holothurioidea. — The metamorphosis of the Auricularia into the Holothurian is manifested in the external shape of the larva by the gradual disappearance of its lobular pro- cesses and the alteration of the ciliated band, which breaks up into several pieces (Fig. 205 A and B). The larger number of these pieces alter their positions by acquiring a transverse position in place of a longitudinal one (Fig. 205 B). At the same time the protuberances of the larval body disappear, and it assumes more of a cylindrical form, whereby, according to Semon, its circumference is strikingly diminished. Finally the different pieces of the ciliated ECHINODERMATA 427 band grow together into five rings, which surround the larva like the hoops" of a barrel (Fig. 206). This is the so-called pupal stage, which is also assumed by some Holo- thurians {e.g. Gucumaria) without passing through the Auricularia form. This stage is remarkable owing to its resemblance to the larva of Antedon, with which it has in common even the number of the ciliated ring's. According to Semon, we may imagine that the rearrangement and loss of continuity in the ciliated band are the result of the migration of the band together with the adjacent body epithelium, probably in con- sequence of internal processes of growth. In the metamorphosis of the ciliated band we have not yet considered the parts lying near the mouth, which do not share in the formation of the external ciliation of the pupal stage. Parts of the longitudinal and transverse portions of the ciliated band approach very close to the 'region of the mouth-opening (Fig. 205 A). After the breaking up of the ciliated band, four parts can be dis- tinguished, which closely surround the mouth, and finally form a continuous ring about it. They gradually move more into the infundibular depression which leads to the mouth -opening. By a marked narrowing of the funnel they come to lie inside the larva, and are employed to clothe the tips of the five anteriorly directed evaginations of the hydrocoele (therefore for the formation of tentacles) (Fig. 205 B). The nerve bands have moved down into the funnel even before the parts of the ciliated band have, and are to a certain extent forced down by them (Fig. 205 A^ n), for the nerve bands occupied a position nearer to the mouth-opening than the ciliated bands. Their free ends then unite at the bottom of the funnel and there form the nerve-ring of the Synapta (Semon). It was precisely those four parts of the ciliated band that moved into the funnel which were united to the nerve bands by means of nerve fibres. Probably this connection is retained during the metamorphosis, and as soon as the ciliated ba|;id has covered the five tentacles the five large tentacular nerves are es- tablished on the nerve-ring, whereas the five radial nerves 428 EMBRYOLOGY do not bud forth from it until later (Semon). The parts of the nervous system, which at first lie superficially, are finally overgrown by the rest of the ectoderm, and since mesenchyma cells crowd in over them, they come to lie at a still greater depth. The ciliated band, which, according to Semon, encircles the mouth, changes its position during the metamorphosis of the larva by coming to lie wholly in the stomodaBum. Here its cells are said to spread them- selves out on the wall and constitute the epithelium. Our previous account of the internal organization of the Holothurian larva was confined to the formation of the intestine, the two enterocoelic sacs, which extended between the intestine and the body-wall, and the hydrocoele. We saw the latter growing around the stomodaeum of the larva as a five-lobed structure. It already represented the fundament of the water- vascular ring and the five tentacles of the Holothurian. Five secondary evaginations of the water-vascular ring arise between the five primary ten- tacles, and at first are also directed upwards (Fig. 205 B). Later, however, five of the ten evaginations of the water-vas- cular ring now present bend over the calcareous arches and grow out backwards (Figs, 206 and 207), so that there are now five tentacles and five radial vessels (Semon). The question now arises whether it is the five vessels first developed (the so-called primary tentacles) which bend over backwards, and thus correspond to the radial vessels of other Echinoderms — as seems most natural, though this has been denied — or whether it is the five vessels of the second group, which correspond to radial vessels. As to the homologies of the ambulacral vessels in the different Fig. 206. — Pupal stage of the larva of Synapta digttata (after Semon). ed, proctodseum ; ent, en- terocoele ; m, oral funnel ; w, water- vascular ring with evaginations forwards (tentacles [tj) and back- wards (radial vessels and Polian vesicles [p]). ECHINODERMATA 429 livisions of Echinoderms, which apparently seem to be so 5lear, the opinions of authors are nevertheless at variance [comp. Semon, No. 55). [Through Ludwig's (No. XV.) recent investigations of Cucumaria planci, these conditions have been satisfactorily elucidated. The five ^invaginations of the water-vascular ring first formed really produce I the radial vessels. They are, it is true, at first directed forward, but [isoon bend backward, thus marking the radii. The tentacular vessels do not arise directly from the ring-canal, but branch off from the radial canals ; however, every radial canal does not give rise to a tentacular vessel, the latter being distributed unsymmetrically to three radial canals only. — K.] The internal condi- tions are more evident in Fig. 207, a longi- tudinal section of the pupal stage of Cucu- maria (after Selenka). The tentacular evagi- nations are seen coming off from the water- vascular ring forwards, and posteriorly those of the radial vessels. The Polian vesicles also take their origin as evagina- tions of the water-vas- cular ring. In the stage under consideration the ring is still in connec- tion with the outside world by means of the stone canal and the dorsal pore. This con- nection is afterwards broken, since a cluster of mesenchyma cells subsequently applies itself to the stone canal Fig. 207.— Longitudinal section of a larva of Cucumaria doliolum, somewhat diagrammatic (after Sklbnka). A, anus; Am, ambulacral (radial) vessels ; D, intestine ; E, enterocoeles ; F, feet; M, mouth; P, dorsal pore, leading through the stone canal to the water-vascular ring ; T, tentacular vessels ; Wr, water-vascular ring. 430 EMBRYOLOGY at about the middle of its course, and here deposits a semi- lunar calcareous ridge, which must be looked upon as cor- responding to the madreporic plate. Where it rests upon the stone canal, the latter is cut through by a constriction, and one half henceforth hangs down from the ring-canal free in the body cavity, whereas the other half is gradually obliterated. Externally the pupal stage now approaches more the adult Holothurian, owing to the fact that the first two feet, the development of which is to be traced to evaginations of the corresponding radial vessel, make their appearance on the posterior part of the ventral surface (Figs. 207 F and 208/). At the fame time the tentacles also advance in their development. We saw in Synapta that a part of the ciliated band moved down into the oral funnel to supply the ecto- dermal covering of the tentacular vessels, which consists partly of sensory cells. The oral funnel then closed to an extremely narrow fissure, and there was thus formed a kind of vestibule (comp. the corresponding processes in the development of the vestibule of Antedon, p. 446). The tentacles, to which the nerve-ring is still joined, lie in the vestibule. The nerve-ring lies at the point where the calca- reous ring, — the supporting apparatus of the tentacles, con- sisting at first of five and later of ten rods, — is attached to the tentacles. When the tentacles have reached the neces- sary development, they are extended out through the fissure, which widens again (Fig. 208), and the young Holothurian now moves both by means of those ciliated bands which still remain, and by means of adhesion with the tentacles and feet when the latter are present. In Synapta, as is known, the feet are not developed, even the radial vessels degenerating. V^. \ The shape of the Holothurian i>d/ W^^'^^l^ ^ would thus be attained if the ^ young animp,! did not lack the Fig. 208.— Holothurian larva with larger number of tentaclcs and ciliated bands, extended tentacles f^^^ ^^^ •£ ^^^ body-COVering (T), and developing feet (;) (after ' . JoH. MijLLBE). already possessed its permanent ECHINODERMATA 431 brucfcnre. Additional tentacles and feet are developed in ie same way as those which we have already learned about, bmely, by evaginations of the parts of the water- vascular rstem already formed. The ciliated bands are said by [emon to disappear owing to their cells spreading themselves \ut over the entire surface of the larva, and substituting lemselves for the originally flat body epithelium, which now lakes way for a thick epithelial layer. With progressing 'growth the mouth and anus of the larva are shifted from the more ventral position toward the anterior and posterior ends of the animal respectively. Now that we have followed the Holothurian larva in its development as far as the young animal, there remain for consideration only a few important internal developmental processes, which relate to the deriva- tion of the middle germ-layer. As we were obliged in the previous description to take into consideration various forms, so are we under the ^■pime necessity in the following statements, which for the greater part we ^^Bake from the works of Selenka and Semon. ^B Having treated of the origin of the mesenchyma cells in the first ^H|i vision of the chapter, we intentionally left the subject for the time ^^Heing. By far the greater part of these cells become connective tissue. BUVhereas they are usually applied to the inner surface of the ectoderm as isolated cells, they are accumulated on both sides of the proctodeeum in larger groups, which give rise to the calcareous spherules and calcareous wheels. They also occur in large numbers in the region of the stone canal and of the water-vascular ring, forming in one place the well- known calcareous deposits and in the other the calcareous ring surround- ing the oesophagus. The mesenchyma cells give rise by multiplication to a kind of cutis under the entire ectoderm (Metschnikoff, No. 37). Underneath the ciliated band they form, according to Semon, groove-like sheaths, which probably serve as supports for the ciliary apparatus. Fissures in the mesenchyma are said to give rise to the blood-vessels of the Holothurian. Thus the vessels accompanying the intestine first make their appearance as lacunar spaces in the mesenchyma lying dorsad and ventrad of the intestine. The blood cells, on the contrary, are said to have been detached from the walls of the hydro-enterocoele and to have taken part in the formation of these vessels. These free cells, which are found in the body cavity as well as in the ambulacra! vessels and blood-vessels, would therefore, according to this view, not arise from the primitive mesenchyma (Semon). Of all the musculature only that of the stomodaeum originates from the mesenchyma. It persists, being carried over from the larva to the young animal (Selenka). The rest of the musculature arises partly I 434 EMBRYOLOGY mesenchyma cells at the sides of the stomach. The com- prehension of this process is rendered more difficult by the fact that the ambulacral and antambulacral surfaces are not parallel, but nearly at right angles to each other. Between the two lies the capacious stomach. In Fig. 209, which, however, corresponds to a somewhat earlier stage, the water- vascular rosette (H) is seen to be partly covered by the stomach, whereas this in turn is partly covered by the funda- ment of the antambulacral surface. The latter develops further in such a way that from the calcareous concretions a number of plates are formed (comp, infra}, which cover a pentagonal area. This grows out then into five processes, thus establishing the dorsal surface of the arms, upon which there appear wart-like elevations, from which the spines arise later. At this stage the starfish already approaches the shape of the adult animal, at least as far as regards its dorsal external surface, and is seen attached to the larva, the posterior end of which it has quite absorbed (Fig. 210). Its anterior portion is still well preserved. Now, however, degeneration also begins here. It gradually shrinks, its substance being consumed by the phagocytic mesenchyma cells, undergoing intracellular digestion, and being doubtless employed in the formation of the new body (Metschnikoff, No. 40). At the same time with these processes the size of the stomach decreases, as a result of which the two surfaces of the star- fish, which were separately developed, are able to approach each other. They cover each other, and finally fuse. The hitherto unclosed water-vascular rosette grows around the oesophagus, and its radii elongate to form the ambulacral vessels, which in their turn give rise to the feet. In this process the distal end of the vascular fundament becomes the so-called tentacle, but the feet are established laterally in pairs. The youngest feet are always found next to the tentacle, therefore at the tip of the arm, whereas the oldest are crowded toward its base. The eye makes its appearance as an accumulation of red pigment at the base of the tentacle at a very early period. Even before this, there have been produced on the antam- ECHINODERMATA 435 bulacral surface secretions of calcareous salts, which at first formed delicate rods and subsequently united into reticular plates. Eleven such plates can soon be recognized, a central one and (arranged about it in a circle) two rows of alter- nating plates, the fundaments of the radial and interradial plates. One of the former, which at first lies at the left next to the dorsal pore, subsequently grows around it, and thus becomes the madreporic plate (Ludwig). According to LuDwiG, the ambulacral or vertebral plates of the arms make their appearance very early as five pairs of calcareous bodies at the base of the five hydroccele pockets. They therefore have even now the position which they retain afterwards, namely, on the outer side of the future am- bulacral vessel. The other skeletal pieces of the arm are not developed until later. The question now is. What relation does the larval intes- tine have to the newly formed starfish ? The older state- ments are not precise on this point ; for this reason we adhere to the recent investigations of Ludwig on Asterina gibbosa, a form, however, which is developed neither from a Bipinnaria nor from a Brachiolaria (comp. supra, p. 421). Yet in this species the two surfaces of the starfish are established independently, and afterwards unite as described above. From this, one may perhaps conclude that the processes in question resemble those in the typical larvd. In Asterina the stomodaeum of the larva separates from the stomach and hangs down from the larval mouth as an internal blind rudiment. For a time the intestine is without any connection with the outside world. The permanent mouth of the starfish is then developed by an outfolding of the stomach, growing out toward the body-wall and finally breaking through to the outside world. The stomach itself is transmitted to the starfish. It subsequently acquires five outpocketings, which bifurcate at their tips, the funda- ments of the five pairs of intestinal caeca. The larval anus is obliterated even before the union of the intestine with the mouth takes place, and the new anus does not arise until after the formation of the mouth-opening. It breaks through at the margin of the central plate, between it and 434 EMBRYOLOGY mesencbyma cells at the sides of the storaach. The com- prehension of this process is rendered more difficult by the fact that the ambulacral and antambulacral surfaces are not parallel, but nearly at right angles to each other. Between the two lies the capacious stomach. In Fig. 209, which, however, corresponds to a somewhat earlier stage, the water- vascular rosette (H) is seen to be partly covered by the stomach, whereas this in turn is partly covered by the funda- ment of the antambulacral surface. The latter develops further in such a way that from the calcareous concretions a number of plates are formed (comp. infra), which cover a pentagonal area. This grows out then into five processes, thus establishing the dorsal surface of the arms, upon which there appear wart-like elevations, from which the spines arise later. At this stage the starfish already approaches the shape of the adult animal, at least as far as regards its dorsal external surface, and is seen attached to the larva, the posterior end of which it has quite absorbed (Fig. 210). Its anterior portion is still well preserved. Now, however, degeneration also begins here. It gradually shrinks, its substance being consumed by the phagocytic mesenchyma cells, undergoing intracellular digestion, and being doubtless employed in the formation of the new body (Metschnikoff, No. 40). At the same time with these processes the size of the stomach decreases, as a result of which the two surfaces of the star- fish, which were separately developed, are able to approach each other. They cover each other, and finally fuse. The hitherto unclosed water- vascular rosette grows around the oesophagus, and its radii elongate to form the ambulacral vessels, which in their turn give rise to the feet. In this process the distal end of the vascular fundament becomes the so-called tentacle, but the feet are established laterally in pairs. The youngest feet are always found next to the tentacle, therefore at the tip of the arm, whereas the oldest are crowded toward its base. The eye makes its appearance as an accumulation of red pigment at the base of the tentacle at a very early period. Even before this, there have been produced on the antam- ECHINODERMATA 435 bulacral surface secretions of calcareous salts, which at first formed delicate rods and subsequently united into reticular plates. Eleven such plates can soon be recognized, a central one and (arranged about it in a circle) two rows of alter- nating plates, the fundaments of the radial and interradial .plates. One of the former, which at first lies at the left next to the dorsal pore, subsequently grows around it, and thus becomes the madreporic plate (Ludwig). According to LuDWiG, the ambulacral or vertebral plates of the arms make their appearance very early as five pairs of calcareous bodies at the base of the five hydroccBle pockets. They therefore have even now the position which they retain afterwards, namely, on the outer side of the future am- bulacral vessel. The other skeletal pieces of the arm are not developed until later. The question now is. What relation does the larval intes- tine have to the newly formed starfish ? The older state- ments are not precise on this point ; for this reason we adhere to the recent investigations of Ludwig on Asterina gibbosa, a form, however, which is developed neither from a Bipinnaria nor from a Brachiolaria (comp. supra, p. 421). Yet in this species the two surfaces of the starfish are established independently, and afterwards unite as described above. From this, one may perhaps conclude that the processes in question resemble those in the typical larvd. In Asterina the stomodaeum of the larva separates from the stomach and hangs down from the larval mouth as an internal blind rudiment. For a time the intestine is without any connection with the outside world. The permanent mouth of the starfish is then developed by an outfolding of the stomach, growing out toward the body-wall and finally breaking through to the outside world. The stomach itself is transmitted to the starfish. It subsequently acquires five outpocketings, which bifurcate at their tips, the funda- ments of the five pairs of intestinal caeca. The larval anus is obliterated even before the union of the intestine with the mouth takes place, and the new anus does not arise until after the formation of the mouth-opening. It breaks through at the margin of the central plate, between it and 436 EMBRYOLOGY an interradial plate. According to tlie observation of Agassiz, the month arises by a shortening of the long oesophagus, and the anus persists. Wm The condition of the dorsal pore and the stone canal, as described by LuDWiG for Asterina, is interesting. In this form, after the separation of the enterocoele and hydroccele, a canal is developed on the latter, which, attached to the water-vascular rosette, opens into the enterocoele quite near the place where the dorsal pore is connected with the entero- coele. This is the stone canal, which does not unite with the dorsal pore until later. Thus there is a stage in which the stone canal does not open directly from the water-vascular ring to the outside world, but, on the contrary, leads into the body cavity. This, however, is in turn connected with the outside world by means of the dorsal pore. Ludwig compares this condition to that which he described for the Crinoids (Nos. 30, 32). In them the water penetrates by means of the pores in the cup [Kelchporen) into the body cavity, to be taken from there and con- ducted into the water-vascular ring by the stone canals, of which there are several hanging down from the water-vascular ring into the body cavity. The earliest fundament of the hlo oil vascular system arises, according to Ludwig, at the place where the intestine grows out to form the oeso- phagus. In the mesenchymatous layer lying between the walls of the hydrocoele, the enterocoele, and the intestine, there is formed a fissure, which exhibits a lining of very flat cells. This is the fundament of the first blood-vascular ring. The structure ordinarily described as the central plexus of the blood- vascular system also arises as a fissure next to the stone canal (comp. General Considerations, p. 456). The nervous system of Asterina is first established in the form of a circular epithelial thickening, which surrounds the region of the future mouth-opening. Its development is certainly similar to that of the central nervous system of the Holothurioidea, with which we are already familiar. The metamorphosis of those starfish larvae which differ from the Bipin- naria- and Brachiolaria-forms, as, for example, that of Asterina gibhosa (Fig. 201, p. 421), is likewise accompanied by the transmission of the greater part of the larval organs to the starfish (Ludwig). Only the mouth and anus have to be formed anew, and the larval organ suffers degenera- tion, being gradually absorbed. Here also the starfish arises from an ambulacral and an antambulacral fundament, which at first are separate. The development of Pteraster militaris seems to resemble that of Asterina (Koken et Danielssen). In this starfish, however, a kind of brooding occurs ; a membrane stretches out over the spines on the back of the animal, forming a brood-chamber. Into this the eggs pass, and there develop into the larvae and young starfishes. ECHINODERMATA 437 Ophiuroidea.— Although the larvce of the Ophiuroidea and Asteroidea are so different in shape, their metamor- phosis presents a certain resemblance. In the Pluteus larva too the ambulacral and antambulaci^l surfaces are begun separately, and only after their subsequent union give rise to the complete star (JoH. Muller, Metschnikoff). In the Pkiteus the five-rayed water-vascular rosette, which opens to the outside world on the dorsal side of the larva, lies ventrad of the oesophagus. It is on this that the first steps in metamorphosis are manifested, for it is over its different radii that the mesenchyma and contiguous larval skin be- come thickened. In this way the fundament of the ambu- lacral surface of the star is produced. Each of the five radii of the rosette, which represent the future ambulacral vessels, produces two lateral evaginations ; thus the larvlae acquire the fundaments of the first feet, which are soon followed by a second and a third pair, etc. While these processes are taking place on the ventral side of the Pluteus, the first indications of the antambulacral surface of the Ophiuran have made their appearance on its dorsal surface, in the form of five outgrowths of the larval skin. They are arranged in a line, so that three of them lie on the larger and two on the smaller part of the umbrella. The five skeletal pieces arise in them as products of the mesenchyma cells. Although the principal parts for the production of the star are now present, nevertheless a total rearrange- ment must take place to accomplish its formation. This begins by the growth of the heretofore semicircular water- vascular rosette, together with its appendages, around the oesophagus, to form the water-vascular ring. With the closure of the ring, the two vessels which at first were farthest apart naturally have come to lie close together, and at the same time the form of a star has now been reached, first on the ambulacral surface. This is, however, not the case on the antambulacral surface. Here also the dermal out- growths (dorsal) undergo considerable changes in position ; but it is not until the larval appendages begin to degenerate that the antambulacral parts cover the ambulacral, and thus complete the star. The internal parts of the larva — the 438 EMBEYOLOGY enterocoelic body cavity, the intestine, etc. — then form a part of the permanent star ; the mouth is said to persist, whereas the anus disappears. Upon the completion of these processes, which result in the establishment of the permanent shape of the animal, the calcareous skeleton of the Pluteus disin- tegrates. The rods break up into pieces; as a result of this, the arms collapse, and the skeleton, together with the larval body, appears finally to be resorbed by the young Ophiuran. "711^ Fig 211. — Pluteus larva with the fun- dament of the Ophiuran (after Joh. Muller). The rods of the larval skele- ton are dark. Like the arms of the starfish, those of the Ophiuran grow at the tips, with the exception of the terminal pieces, which correspond to the skeletal pieces first formed on the dorsal surface. Therefore the new pieces are interpolated between the terminal pieces and the adjacent ones. The skeletal parts thus follow a law quite similar to that of the feet, the de- velopment of which always takes place between the (terminal) ten- tacle and the pair of feet next to it. The origin of the arm-plates is interesting ; according to Ludwig (No. 34), they result from the fusion of two calcareous plates ly- ing on either side of the median line of the arm. Echlnoidea. — According to Metschnikoff's description (No. 37), a difference exists between the metamorphosis of the Echinoidea and that of other Echinoderms, inasmuch as there is developed an invagination of the larval skin, at the bottom of which the earliest fundament of the body of the sea-urchin makes its appearance. Thus it happens that the earliest fundament is not exposed, but is covered by a fold of the larval skin as though by an amnion. Since, how- ever, the larval skin here too becomes directly converted into that of the sea-urchin, this difference does not appear to us to be in any way important. The processes of the metamorphosis of the Pluteus into ECHINODERMATA 439 the sea-urchin are as follows: inside the Pluteus oi Strongylo- centrotus lividus, which is provided with four arms, we tind nearly the same conditions that have been described apropos of the development of the enterocoele and hydrocoele. The enterocoelic sacs lie to the right and left of the stomach ; the hjdroecele lies over the left one of these, and has the form of a retort, the neck of which opens to the exterior on the back of the larva, somewhat as in Figs. 212 and 218, figures of a Spatangoid ; the conditions in these forms are, however, somewhat different, as will be mentioned farther on. Later, when the Pluteus has become six-armed, an invagination of the outer skin is formed over the hydrocoele (Fig. 212). This arises from a thickening of the epidermis, which gradually sinks in and finally rests with its bottom upon the hydrocoele.^ The thickened discoid bottom of the dermal invagination is the earliest fundament of the lower (oral) surface of the body of the sea-urchin (called " Echinoid disc" by Joh. Muller). The much thinner lateral parts of the invagination overlie this as an amnion-like covering (Fig. 213). The opening of the invagination has narrowed, but persists, whereas in the Spatangoids other conditions subsequently make their appearance (comp. infra). The hydrocoele now grows out into five processes, and the Echinoid disc does the same, by developing a dermal cover- ing over each of the five hydrocoele processes. In this way the first five feet of the sea-urchin arise. They extend into the cavity of the invagination, almost filling it. During the changes described in the region of the Echinoid disc, the first indications of the dorsal surface of the sea- urchin also become noticeable. Two roundish dermal eleva- i Figures which Fewkes (No. 13) gives of the developmental stages of Eehinarachnius parma may confirm Metschnikoff's deBcription, although this cannot be gathered from the text of the work. Likewise it seems to us from the figures of Colton and Gakman (No. 11) that the metamor- phosis of Arbacia is like that described by Metschnikoff for EeJdnoids and Spatangoids. A cavity appears on the Pluteus, in which the first formed feet become visible. The work of Coltox and O^H^l^n was un- fortunately not accessible to us, and is known only through the descrip- tion of Brooks [Handbook of Invertebrate Zoology, Boston, 1882). 440 EMBRYOLOGY tions are developed on the umbrella of the Pluteus, one on the dorsal surface and one on the anal area. Each of these soon assumes a trilobed form, and they prove to be the two first pedicellaria of the sea-urchin. Fig. 214 shows this condition in another sea-archin {Arhacia pustulosd) } With the progress of development the disc continually increases Fig. 213. Fig. 212. Figs. 212 and 213.— Parts of a Spatangoid Pluteus (after Mktschnikoff). D, intestine; Ei, invagination of the larval skin, which in Fig. 213 covers the hydrocoele (if). The latter opens to the exterior by means of the dorsal pore. H, hydrocoele ; P, dorsal pore ; Sfc, larval skeleton. in circumference, and at the same time the opening of the invagination also widens again. The contractile feet are finally extruded through the latter, and are now seen to 1 [Very thorough and accurate investigations of these features of de- velopment have recently been carried on by Theel (No. XXIX.) on Echinocyamus, a form in which the larval development and the meta- morphosis could be established almost without a gap. Unfortunately his fine results cannot be stated here in brief form, and we are compelled to refer to the original paper. — K.] I ECHINODERMATA 441 xeciite tactile movements. At this time the larval skeleton begins to break up, and the arms of the Pluteus degenerate I^Has a result of this (Fig 215). The body thereby assumes HRiearly the form of a hemisphere with the disc as the base. The circumference of the disc has increased more and more, nd correspondingly the opening of the invagination has Iso become enlarged. The amnion-like envelope meantime adually diminishes in prominence ; at length it forms only circular fold, surrounding the circumference of the disc, Fig. 214.— Pluteus larva of Arhacia pustulosa (after Joh. Mullee). The skeletal rods are dark. P, pedicellaria. and finally disappears. Thus the amnion seems to become directly converted into the skin of the sea-urchin, and, in fact, would seem to supply that part of the skin which unites the sole-like ventral surface with the arched back. Fig. 215 represents a young sea-urchin which still possesses, in addition to the feet, some of the Pluteus arms. Its feet ai-e already employed as locomotor organs. In Fig. 216 the young sea-urchin spines are already seen making their 442 EMBRYOLOGY appearance by the side of the pedicellariao. They arise as evaginations of the skin in which reticulated calcareous rods are deposited. The first of the dorsal plates to make its appearance is the central one, which is perforated bj the anus. Other plates are then secreted about it, in a spiral line, i.e. in sach a way that the newly arising plates crowd away the older ones from the anal plate, since they are interpolated between it and the older plates (Agassiz). Fig. 215.— Young sea-urchin [Arhacia punctulata) with parts of the Pluteus larva attached (after Colton and Garman, from Brooks's Handhoolc). a, partially de- generated arms of the Pluteus ; /, feet ; St, spines. The internal larval organs become a part of the sea-urchin, although a new oesophagus is said to be developed, which is not grown around by the water-vascular rosette, but grows through the previously formed water-vascular ring (Bury), — a condition, therefore, somewhat different from that which we observed in the development of the other groups of Echinoderms. The dorsal pore persists, and also its connection with the water-vascular ring by means of the stone canal. The outgrowth of the water-vascular rosette into the ambulacral trunks appears to take place, according to Agassiz, in the same way as in the starfishes, for new feet are continually interpolated between the terminal foot and the adjoining pair. ECHINODERMATA 443 In the metamorphosis of the Spatavgoid Phitetis into the sea-urchin the invagination is said by Metschnikoff to close. At the bottom of it the earliest fundament of the sea-urchin then makes its appearance. Furthermore the " amnion " is said to become detached from the larval I kin. In the protrusion of the feet the amnion, as well as part of the irval skin, would therefore have to be broken through. Crinoidea. — We left the larva of Antedon rosacea at a tage in which the nearly ovate form exhibited slight curv- ing toward the ventral side. The further development is characterized by the fact that the larva abandons its free life and grows into an attached stalked form. It therefore passes through a stage in which it resembles a stalked Crinoid. This is known as the pentacrinoid stage. Traces d^^'^-X-'^/^ ^^ Fig. 216.— Young sea-urchin {Avbacia "[ivsiMlo&o) with degenerated Pluteus arms attached (after Joh. Mullbb). /, feet; P, pedicellariaB; ^i, spines. of this stage are already shown in the free-swimming larva through the fundaments of the skeleton, which make their appearance in the mesenchymatous tissue of the larva. They are first seen as small granules, which, however, soon enlarge into triradial and quadriradial forms, and finally become fenestrated plates (Fig. 217). Two rows of five plates each can be distinguished, — the oraXia and hasalia con- stituting the calyx,— and a piece lying below these, the future terminal plate of the stem (Figs. 217, 221, and 222, p. 450). According to Bury, it is this plate of the skeleton which first makes its appearance deep in the body of the larva. Inas- 444 EMBRYOLOGY 1 # Gr mucli as new segments (the stem-joints) are interpolated between it and the basalia, it moves farther downward. The stem-joints take their origin at the base of the calyx. The youngest, therefore, lie next to this, the oldest, on the other hand, next to the terminal plate. At first they form annular plates, but soon alter their shape and become thick segments by the secretion of rod-like calcareous concretions on both their surfaces. Between the uppermost stem-joints and the basalia lies a larger skeletal piece, which has been called the centrodorsal plate (Fig. 223 cd, p. 451). It forms the important foundation of the basal plate of the calyx. According to Bury, it arises by the nnion of several skele- tal pieces. For three sub- basal plates make their ap- pearance below the basalia; ^5. these at length fuse with one another into a five-pointed I star and finally unite with ^ ;^^ the uppermost joint of the N / stem to form the centrodor- ^^— ^ sal plate. This condition is Fig. 217.— Larva of Antedon rosacea, important, bccause certain with ciliated bands and tuft of cilia, as „ ., ^. ., /- t i ±i well as fundan^ents of the skeletal ^^SSll CrmOlds (Ichthyocri- plates inside. Gr, pit, by means of noidse) also poSSesS three which the larva attaches itself; Lm. g^^.^asal plates with the the so-called larval mouth. ^ \ Lth The series of plates of the calyx are at first not arranged in a closed ring, but in the form of a horseshoe, the open side of which corresponds to the position of the " larval mouth." Before the skeleton attains the development described, the larva has already given up its free life. After about twelve hours of swarming it attaches itself by means of the ECHINODERMATA 445 L ^Mhesive disc. At this stage of attachment the larva lies ^^ith its entire ventral surface on the object to which it ^^ttaches itself. At first it still possesses its typical ciliation, but that is soon lost. At the same time its shape changes! the anterior end, with which the larva attaches itself and which subsequently grows out into the stem, diminishing in size and the opposite end becoming broader. The club- shaped larva now rises from its support, to which only the small end remains united. Accordingly we now designate the club-shaped portion, which becomes the calyx, as the upper part, the narrowed portion as the lower part, of the larva (Fig. 218). Fig. 218.—^ to C, early stages of development of the attached larva of Antedon rosacea (after J. Baerois). Development of the vestibule (F) by invagination of the ectoderm (ect). D, intestine; Ls, subambulacral, Lv, visceral, body cavity; S, stalk of the larva ; T, tentacular vessel ; x, stone canal (?). The most important change which takes place in the larva after this metamorphosis of its external shape affects its ventral surface. The wide pit whicK is found there, and which is called the larval mouth, becomes obliterated during the attachment of the larva, but a new invagination of the ectoderm takes place at the same spot, which is deeper than the pre-existing one. Here also, as in the region of the "larval mouth," the ectoderm is greatly thickened (Fig. 218^). The invagination soon enters into relationships with the internal organs, for its upper margin extends out 446 EMBRYOLOGY toward the upper pole of the larva, and thus the floor of the pit comes to lie over against the internal organs (Fig. 218 B). At the same time the opening of the invagination narrows, and is finally entirely closed and [the invaginated layer] detached. In this way the invaginated part of the ectoderm comes to lie inside as a closed sac, and since it follows still more the tendency existing from the beginning, it movies quite to the upper end of the larva (Fig. 218 G). This sac subsequently changes in such a way that its floor overlies the evaginations of the water- vascular system (tentacular vessels), and its roof unites with the mesenchyma and the outer ectodermal lamella to form the roof of the vestibule (Figs. 219 and 220) — the chamber on the floor of which the mouth subsequently arises, and the roof of which disappears to set free the tentacles. However, before the beginning of these processes, which bring the larva nearer to its per- manent shape, important changes take place in the internal organs. Just as the Antedoa larva, with its five ciliated rings, recalls the cask- like shape of the Holothurian larva, so, too, the development of the vestibule and the investment of the tentacles by its floor show a certain resemblance to the formation in the Holothurian larva of the vestibule in which the tentacles lie (comp. p. 427). Here, as there, it is a de- pression of the ectoderm which forms the vestibule and supplies the external covering of the tentacular vessels. In both cases the process takes place in the region of the mouth, which, however, exhibits a different position in regard to the ciliated rings. We left the internal oi-gans at a stage of development at which the two enterocoeles and the hydrocoele lay at the side of the saccular intestine. The latter, which at first lies ventrad of the intestine, moves, v^^ith the metamorphosis of the larva into the pentacrinoid form, to a position over the intestinal sac (Figs. 218 and 219), and grows out into the shape of a horseshoe, its two arms finally uniting into a ring. At the same time there are formed five upward evaginations, which are covered over by the ectodermal cell-layer which forms the floor of the vestibule (Figs. 218 G and 219). The prolongation of the hydrocoele, which was recognizable even ECHINODERMATA 447 m Heated as far as the outer body- wall and fused with it (Fig. 219), forming in this way the stone canal (Barrois). As in he other Echinoderms, so also in the Crinoids, at least while hey are yoang, a communication exists between the water- vascular system and the outside world ; this fact was estab- lished by Perkier and confirmed by Barrois. As is well known, a large number of stone canals hanging down into e body cavity occur in the adult Crinoids. Ludwig (No. 32) had Iready shown that in the penta- crinoid stage of the Antedon larva at first only one stone canal is pre- sent ; but he believed that this also, arising from the water-vas- cular ring, ended free in the body cavity, whence it took up into it the water which entered through a pore in the body -wall. This view corresponds nearly to that which was defended upon embryo- logical grounds by Bury. Ac- cording to him, the free process of the fundament of the hydro- coele, which was considered by Barrois as a stone canal, is rather a third coelomic sac. This en- larges, comes into connection with the body-wall by means of a pro- cess (parietal canal), and thus opens to the outside world by means of the water-vascular pore. It is only secondarily then that the hydrocoele, by means of a stone canal, is united to this part of the body cavity. The descrip- tion of these conditions coincides with that given by Ludwig for Asterina, where the stone canal also opens into the enteroccele, and is connected with the dorsal pore only by means of this (comp. supra, p. 436). According to Pekrier, the pore described by Ludwig, which lies in one of the oral plates, cor- responds to the external opening of the stone canal. The stone canal is said to be easily separated from the pore in dissecting, and then hangs from the water- vascular ring free in the body cavity. Fig. 219.— Longitudinal section of an Antedon larva (after figures by Perbikr). D, intestine ; is, subambulacral, Lv, visceral, part of the body cavity ; N, stalk of the larva; St, stone canal; T, tentacle; V, vestibule; Wr, water-vas- cular ring, from which spring the ten- tacular vessels and the stone canal. 448 EMBRYOLOGY In later stages of the larva still other canals are developed as evagina- tions of the water-vascular ring and the peritoneum which covers it. They grow out toward the body- wall and unite with it. At the time when the larva detaches itself, five such canals are present, all of which communicate with the outside world. Afterwards, however, the forma- tion of the evaginations and the pores of the cup is said no longer to take place simultaneously, so that the former may multiply independently of the latter, and vice verxd. Thus a condition would arise such as was described by Ludwig, — cup-pores leading to the interior, and free appen- dages of the water- vascular ring opening into the body cavity. The canal which was first formed attains an extraordinary development, and Perriek regards it alone as homologous to the stone canal of other Echinoderms, those formed later being of a secondary nature. The conditions of the body cavity of the Antedon larva are complicated. The two coelomic sacs, which at first lie at the right and left of the intestine, subsequently, when the larva passes into the pentacrinoid stage, arrange themselves above and below the intestine (Figs. 218 and 219). Perrier designates the parts of the body cavity as the subambulacral and visceral portions. Where the two come together there is produced a mesentery, extending transversely across the body (Figs. 219 and 220). Furthermore, according to Bury, tw^o longitudinal mesenteries are formed, owing to the facts that the two ccelomic sacs are (in cross-section) nearly horseshoe-shaped, and that the two arms of each sac grow out toward each other. In each case they meet in a longi- tudinal mesentery, the one belonging to the upper entero- coele lying in the anal radius and that of the lower (visceral) enterocoele in the preceding radius (reckoned according to the course of the intestine). As the coelomic sacs enlarge they apply themselves to the intestine and water-vascular ring as the splanchnic layer, and to the mesenchymatous tissue of the body-wall as the somatic layer. The aboral body cavity, as was pointed out by Goette, sends a process into the narrow posterior part of the larval body (Fig. 220) . According to Perrier, this process consists of both layers of the mesoderm (Fig. 219), and Goette's conjecture is con- firmed, viz. that the chambered organ, which in the adult animal lies within the centrodorsal plate as an important part of the blood- vascular system, arises from this posterior ^JrO ECHINODERMATA 449 ■ 'ocess of the body cavity. We shall revert to this part of the body cavity and its derivatives later. An entirely clear insight into the structural conditions of the body cavity, which are obviously difficult to follow, cannot be gained from the authors' statements (Goette, Perkier, Barrois, Bury), since they do not agree. The older statement of Goette, according to which the body cavity also takes part in the formation of the vestibule, appears in another light since the descriptions which Barrois and Bury give of Fig. 220.— Longitudinal section through the cup of an Antedon larva, the vesti- bule of which is still closed (after Goette, from Balfour's Comparative Embryo- logy), ae, depression of the vestibular epithelium to form the mouth (m); al, intestinal canal; an, region of the anus; Ip, subambulacral body cavity; Ip', vestibule; m, mouth; mt, transverse mesentery ; r, roof of the vestibule ; rp, visceral part of the body cavity and its prolongation {rp') into the stalk of the larva ; t, tentacle ; tcr, water-vascular ring. this process. According to these authors, the lining of the vestibule is not mesodermal, but ectodermal. The penetration of an enterocoelic diverticulum into the stem, observed by Goette, Perrier, and Bury, is denied by Barrois. According to him, these axial structures arise rather by an accumulation of mesenchyma cells. On the other hand, according to Barrois, a process of the subambulacral body cavity penetrates axially toward the stem. Bury's contention that, in addition to the right and K. H. E. G G 450 EMBRYOLOGY left portions of the body cavity, there is developed still a third part, was already mentioned in considering the hydrocoele (corap. p. 413). Figs. 221 (and 222).— Pentacrinoid larva (C) and swarming larvae {A and B) of Antedon (after Thomson, from Balfour's Comparative Embryology). The most anterior ciliated ring, described by Buky, is lacking on the swarming larvse, which are placed in the positions in which thev subsequently become attached (comp. Fig. 217, p. 444). ECHINODERMATA 451 kj Hitherto we have learned of the intestinal canal of the ntedon larva only as a closed sac. The month and oeso- phagus do not arise until the formation — on the floor of the vestibule in the middle of the water-vascular ring — of a depression, which fuses with the intestine (Fig. 220 m). The intestine therefore does not even yet open directly to the outside world, but into the vestibule. Its interior at this time is not empty, but filled with cells (Bury) or with a kind of nutritive yolk (Barrois). The entodermal mass elongates backwards (basal- wards) to form the intestine, and winds spirally about the axial part of the body cavity. Its end then moves in the transverse mesentery, at about the height of the upper margin of the basal plates, up to the body- wall (Fig. 220), with which it fuses, subsequently break- ing through to the exterior. The anus comes to lie in the vicinity of the water- vascular pore. Subse- quently it is shifted to its final po- sition on the ventral wall of the cup. As in the rest of the Echinoderms, the anus seems to have no direct relation to the blastopore. Having considered the internal developmental processes, we turn again to the external shape of the larva, which in the meanwhile has essentially changed. These changes are partly due to the metamorpho- sis of the hydroccele. Each of the five primary tentacles, which we have already seen to be an evagination of the water- vascular ring, splits into three Fio. 223.— Diagram of a pentacrinoid larva of Antedon rosacea (after Thomson, from Balfooe's Comparative Em- bryology), cd, centrodorsal plate ; or, oralia ; 4, radialia ; 3, basalia ; 1, terminal plate. 452 EMBRYOLOGY parts, so that fifteen tentacles can now be recognized. By the addition of two new tentacle-buds to each of the five groups, the number of tentacles soon increases to twenty- five, arranged in five radial groups. The tentacles (Figs. 219 and 220 t) project into the vestibule, the roof of which is stretched out between the upper margins of the oralia. This roof is at first thick, but gradually becoming thinner (Fig. 220 r), finally disappears entirely. The gradual dis- appearance of the roof is partly a result of growth, partly brought about by histolytic processes. According to Bury, such processes can also be recognized on the rest of the larval body, and cause a disappearance of the histological differentiation. Probably migratory cells make their appear- ance in this connection as ^phagocytes. After the disappearance of the roof of the vestibule, the tentacles, on which papillae subsequently bud forth, project free to the exterior (Fig. 221 C). The under-part of the larva has elongated into the stalk, and it now rests with its terminal plate on some support. The fundaments of the arms bud forth on the upper part of the cup as five pro- jections (Fig. 221 C). The tips soon split into two branches corresponding to the permanent forking of the arms. One of the radial tentacles, each of which has likewise split into two, unites with the fundament of each of the arms. Sur- rounded by this, it grows out with it and becomes the ambu- lacral canal of the arm. By means of lateral budding it gives rise to the tentacles of the arm. The tentacle which is first formed always remains at the tip of the arm. The new tentacles arise at its base in groups of three. The mode of formation of the tentacles is therefore similar to that of the ambulacral feet of other Echinoderms. The development of the pinnules is the result of a forking of the arms, which occurs alternately to the right and to the left (W. Carpenter, Perrier). This explains the alternating position of the pinnules. Important changes have taken place in the skeleton of the larva. Between the basal and oral plates, alternating with the latter, five new skeletal pieces, the radialia, have made their appearance (Fig. 223^ ) ; these become greatly enlarged, ECHINODERMATA 453 and serve for the support of the arms (Fig. 224 ri-rm). Through the vigorous growth of the radial pieces, to each one of which two other plates are added, the oral plates are crowded on to the oral surface, where they finally undergo resorption. In other Crinoids (e.g., Uhizocrinus) , on the contrary, the oralia are said to persist throughout life. Another change has taken place at the base of the cap : the centrodorsal plate has gradually overgrown the ha