'i..^ /»> ^jl' CD- SI i tr • =o I rr I O i ° D m a TEXT-BOOK OF THE EMBRYOLOGY OF INVERTEBRATES. K^-^ 0- TEXTBOOK OF THE EMBRYOLOGY OF INVERTEBRATES BY Dk. e. koeschelt, PROFESSOR OF ZOOLOGY AND COMPARATIVE ANATOMY IN THE UNIVERSITY OF MARBURG. Dr. K. HEIDER, PROFESSOR OF ZOOLOGY IN THE UNIVERSITY OF BERLIN. TRANSLATED FROM THE GERMAN BY MATILDA BEENARD. REVISED AND EDITED WITH ADDITIONAL NOTES BY MARTIN E. WOODWARD, DEMONSTRATOR OF ZOOLOGY, ROYAL COLLEGE OF SCIENCE. Vol. II. PHORONIDEA, BRYOZOA ECTOPROCTA, BRACHIOPODA, ENTOPROCTA, CRUSTACEA, PALAEOSTRACA. LONDON: SWAN SONNENSCHEIN AND CO., Ltd. NEW YORK: THE MACMILLAN CO. 1899. PREFACE. In presenting a second "instalment of the translation of the Lehrhuch der vergleichenden EntvjicklungsgescMchtc der ivir- belloseii TJiierc to students of Zoology, I feel that an apology is due to them for the long interval of time that has elapsed since the appearance of the first part. Professors Mark and Woodworth, the translators of Vol. I., found themselves unable to spare time to continue the translation, and others who subsequently undertook the work were for similar reasons obliged to relinquish it. Consequently the book came into the hands of the present translator and editor only last year ; and although we have endeavoured to push the work forward as quickly as possible, the task of translating, printing, and editing has necessarily consumed a considerable time. The whole work, however, is now well in hand, and, in issuing Vol. II., we can confidently hope that no long period will elapse before the production of Vols. III. and IV., the former being already in the press. It will be noticed that the order of the subject-matter has been changed. In its original form the Text-book is divided into three parts of unequal size. That already translated (Part I.) is much smaller than either of the remaining parts. We have consequently found it necessary to divide the second and third parts of the original into three. In doing so we were faced by a difficulty. If the original sequence of the chapters was to be retained, each part could not be made complete in itself unless the relative size of the volumes was left out of consideration. It was therefore thought preferable to alter slightly the original sequence of the chapters. Those dealing with the Phoronidea, the Bryozoa Ectoprocta and VI PREFACE. Entoprocta, and the Brachiopoda have therefore been removed from the end of Part III. (Vol. IV. of the English edition), and placed at the commencement of Vol. II., which part contains, in addition, the Crustacea proper and the Palaeostraca. I do not think there can be any serious objection to this change, inasmuch as the four groups mentioned probably find a natural resting-place near to the Annelida, which were dealt with in Part I., while they have little, if any, relation to the Molluscan phylum treated in Part III. In all other respects the original text has been adhered to as closely as possible ; but it has occasionally been found necessary to rewrite certain paragraphs which, when translated, appeared somewhat in- volved and obscure. In such cases the motive of the orig-inal has been adhered to as far as possible, the sentences, however, being entirely recast. Additional notes have been added relating; to the most important of the many fresh observations which have been made since the original work appeared. Such additional matter has been placed in footnotes, distinguished from the footnotes of the authors by being enclosed in square brackets. A considerable number of additional references have also been given ; and these are placed with the Literature at the end of each chapter under the form of Appendices, and, as in Vol. I., numbered with Eoman numerals. But even these additions could now be added to. For instance, since going to press a short note on the early development of Phoronis by E. Schultz* has appeared, which might with advantage have been incorporated in this work. Although desirous, as far as possible, to use the same terminology as the translators of Vol. I, I have been compelled to differ from them in the rendering of the ever-recurring word " Anlage." This important term in Vol. I. is rendered by the word " fundament." Exception, with which I concur, has already * "Ueber Mesoderm-bildung bei Phoronis," Otdycln. Oppisk., Tryd. imp. St. Petersburg, Obshch. estestv., T. xxviii., Vuip. 1. Tlie autlior deals mainly with the formation of the mesoderm, and controverts all Caldwell's observations relating to the origin of this layer, which he maintains is a mesencliyme, and arises in the blastnla and gastrula stages, most conspicuously in the latter, as proliferations from the entoderm. He could not iind Caldwell s primitive groove, and thinks that this observer described as his posterior coelomic sacs the ventral invaginations which form the body-wall of the adult. PREFACE. VU been taken to the use of this term,* on the ground that the word fundament implies the solid basis or foundation upon which a structure rests or is built, whereas an " Anlage " is essentially a changing, growing structure, which, though at one time the foundation, when only the foundation exists, eventually gives rise to, or rather itself becomes transformed into, the fully-formed organ. Having thus decided against the continued use of this term, I found myself face to face with the responsibility of selecting one of tlie numerous terms which have at one time and anothei' been put forward as the English equivalent of " Anlage," at the same time knowing full well that, whichever word was adopted, I should find a large number of biologists against me, as nearly every teacher of note has proposed at least one word wdiieh he believes to be the only correct rendering of " Anlage." Realising, then, the impossibility of satisfying everyone, I thought it advisable to pass over all the numerous terms which have been recently suggested, none of which are really satis- factory, and to revert to that much-abused word rudiment. Most biologists will agree that the term rudiment, if it had not been misused by some of our most eminent zoologists, would undoubtedly be the best word by which we could render the German term " Anlage." Unfortunately, following the lead of Darwin and others, we have acquired the habit of applying the terms rudiment and rudimentary to certain structures present in the adult which, in consequence of their small size and frequent loss of function, have retained a somewhat embryonic stamp, thus preserving the outward appearance of a rudiment but losing its essential character, viz., its inherent tendency to further growth. Tliese, then, are not rudiments, but arrested, reduced, vanishing, or vestigial structures, and should be spoken of as vestiges. Why, because Darwin unfortunately * See Nature, 1896, p. 361. P. C. Mitchell, " Anla^en," Nat. ScL, vol. v., 1894. Dr. Willey {Nature, 1898, p. 390), who supports the use of the term primordium, objects to the word rudiment on the ground that tlie latter has been regarded as the tirst visilile "Anlage" of an organ. But who is to decide when a growing structure is first visible to the eye ? Professor Wilder [Science, 1898, p. 793), in a reply to Willey, advocates the use of the term protuii; but to define this term he has to make use of the word rudiment. Thus he states that proton was em[)loyed "to design(ite the primitive, undiH'erentiatcd mass or rudunent of a ])ai't." Vlll PREFACE. misapplied the word rudimentary, should we necessarily regard this misuse as hallowed, and ever after refuse to use the word in its common sense ? To such an extent has this misuse of the word been carried, that even encyclopaedic dictionaries, after defining the word rudiment in such a manner as to prove that it is the very word we are seeking, as a rendering of the idea expressed by "Anlage," give us, under the technical use of the word, " In Zoology, a part or organ, the development of which has been arrested (see Vestige)." It would require but little trouble on the part of teachers of Biology to reinvest the word ludiment with its proper meaning. By carefully insisting on the use of tlie words vestigium and vestioial, or their equivalents, for all abortive or reduced structures met with in the. adult animal, and restricting the terms rudiment and rudimentary to all growing and developing tissues and organs, tliey could insure this result in a few years. We have by no means always rendered "Anlage" as rudiment, for we find tliat the German use of the term is not at all precise, and it was often possible to express the meaning better by another English word. The extreme looseness with which some other terms, sucli as Vorderdarm, Mitteldarm, and Hinterdarm, are used in German Avas unfortunately not recognised until too late. These words are for the most part translated by the equally vague expressions fore-, mid-, and hind-gut. Throughout the Crustacea the fore-gut is co-extensive with the anterior ectodermal in- vagination, the stomodaeum, and the liind-gut is similarly related to the corresponding posterior invagination, the procto- daeum ; the term mid-gut, however, is used in a more varied sense. AVhile for the most part it is applied to the entire entodermal rudiment, it is at times used for the entodermal tube after the separation of the hepatic rudiment. As more precise terms, Professor E. Eay Lankester proposed the words enteron for the entodermal rudiment before the separation of the various entodermal derivatives, and metenteron for what is left of the enteric sac as the central element of the alimentary canal after the separation of the outgrowths,* and * See Preface to English translation of Gegcnbaur's Elements of Comparative Anatomy, 1878. PREFACE. IX these terms might well be adopted in the Crustacea for the three divisions of the alimentary canal. In the Phoronidea, Bryozoa, and Brachiopoda, it is impossible to be so precise in our terminology, for the origin and homology of the various divisions of the alimentary canal in these forms are more or less obscure. This is notably the case in the adult Bryozoon, where it is quite impossible at present to interpret the parts of the canal, all of which appear to arise from the ectoderm. Tlie use of the terms fore-, mid-, and hind-gut must not, consequently, in this group be regarded as implying any morphological homology with the similarly-named parts in the Crustacea, but only an analogy with those parts. All through this woi'k tlie genital glands are spoken of as arising from, or in connection with, the mesoderm ; and in places even the splanchnic or somatic layer of the mesoderm is specially mentioned as giving origin to the genital cells. These statements arose from the incompleteness of our know- ledge, at the time when this work was piiblished, concerning the early history of the genital rudiment. A number of observations have, however, since been made on this point in various groups of Invertebrata, especially in different orders of worms and Arthropoda, which tend to show that the primitive genital cells are separated from the somatic portion of the embryo at an extremely early period in the cleavage of the egg. In fact, in some cases, directly after the first cleavage plane has appeared, it can definitely be pointed out which of the two spheres will give rise to the genital rudiment, although this cell will also form certain other rudiments, the actual isolation of the genital cell taking place somewhat later (as early as the 32-celled stage in Cyclops, according to Hacker). All these observations lend support to the belief that the invisible genital rudiment is from the first quite distinct from the somatic rudiment ; the former, at all events, attains a visible distinction at a much earlier period than is assigned to it in this work (see also Vol. i., p. 12, footnote). MARTIN F. WOODWARD. Royal College of Science, London. December, 1898. CONTENTS OF VOL. II. PAGE Chapter XV. PHORONIDEA. By K. Heider 1 I. Embryonic development ..... 1 The Actinotroclia larva . . . . r, II. Metamorphosis ..... 5 Rise of the organs . . ... 8 III. General considerations ..... 8 Literature . . . . . . 10 Chapter XVI. BRYOZOA ECTOPROCTA. By K. Heider 12 Systematic . ..... 12 I. Formation of the egg, fertilisation, position of tlie embryo 13 II. Embryonic development ..... 15 General account of the form of the larva . 17 III. Metamorphosis ..... 18 1. Type which develops an alimentary canal 19 a. Larva of Alcyonidium . . . . 19 b. Larva of Tendra . . . . 21 c. Larva of Membranipora (Cyphonautes) and Flustrella 21 2. Type of the intestineless Chilostomatous larva with slightlj developed corona . . . . 24 3. Type of the intestineless Ghilostomatons larva with highlj developed corona . . . . 25 a. Structure of the larva . 25 h. Metamorphosis 27 4 Type of the Vesieularian larva . 31 5. Type of the Cyclostoniatous larva 31 6. Type of the Phylactolaematous larva 32 IV. Development of the polypide 37 V. Asexual reproduction of the Ectoprocta 44 A. Budding 44 B. Development of Statoblasts . 49 C. Winter buds (hibernacula) . . 54 VI. Regeneration ... 55 VII. General considerations 55 Literature 60 Chapter XVII. BRAGHIOPODA. By K. Heider 65 I. Testicardines . . . . . . 65 A. Embryonic development . 65 B. Metamorphosis . 68 II. Ecardines • . . 73 III. Changes in the shape of the shell . 75 ^o^yo xu CONTENTS. Chapter XVII. BRACHIOPODA— co7i. Ascothoracida . . . . E. The morphological derivation of the complemental males 6. Metamorphosis of the Copepoda General considerations . . . . A. Gnathostomata . . . . Character of the Nauplius stage Metanauplius . . . . Cyclops stage . . . . £. Parasita ..... General considerations regarding parasitic Copepoda Chondracanthidae . . . . Philichthyidae . . . . Dichelestiidae . . . . PAGE 178 180 181 190 190 191 192 192 193 193 195 196 196 196 199 201 202 203 205 205 208 209 209 209 209 213 214 216 219 219 220 221 221 221 223 226 228 228 229 231 231 232 233 234 235 237 237 239 239 239 XIV CONTENTS. Chapter XIX. CRUSTACEA— coo^mxerf. Caligidae (Chalimus stage) Leniaeidae Lernaeopodidae C. Branchima .... 7. General considerations regarding the segmentation of th and the metamorphoses of the Malacostraca Segmentation of the body Typical form of the limbs Metamorphoses of the Malacostraca . The Nairplius stage The Metanauplius stage The Protozoaea stage The Zoaea stage The Mysis and Metazoaea stages Final stages of nietamorpliosis (Mastigopus, Megalopa) Abl)reviation of metamorphosis Metamorphosis of the Schizopoda and Stomatopoda 8. Metamorphosis of the Leptostraca 9. Metamorphosis of the Schizopoda Nauplius and Metanauplius of tlie Euphausiidae Calyptopis stages . Furcilia and Cyrtopia stages Development of the Mysidae and Lophogastridae 10. Metamorphosis of the Decapoda A. Sergestidae .... Metanauplius of Lucifer Protozoaea and Zoaea of Lucifer Mysis and Mastigopus larvae of Lucifer Metamorphosis of Sergestes . Elaphocaris (Zoaea of Sergestes) Acanthosoma (Mysis larva of Sergestes) Metamorphosis of other Sergestidae B. Penaeidae .... Metamorphosis of Penaeus Ceratapsis Metamorphosis of Stenopus . C. Caridea .... Abbreviated development of some Caridea Amphion • D. Astacidea .... Homarus .... Nephrops . . . Astacus and Cambarus K Loricata .... Embryos of Scyllarus Structure of the Phyllosoma . Metamorphosis of the Phyllosoma F. Thalassinidea Metamorphosis of Gebia Larva of Galliaxis (Trachelifer) bodj' PAGE 239 •240 241 244 246 246 247 247 249 249 249 249 2.')0 251 251 252 252 253 253 255 256 257 257 257 258 259 261 263 265 265 266 267 267 271 271 271 274 275 275 276 277 277 278 278 281 281 284 284 285 CONTENTS. XV Chapter XIX. CRVSTACEA—continncil. G. Aiionmra .... Zoaea and Metazoaea of the Anoniiira . Larva of Porcellana (Lonchophorus) Larvae of the Hippidae IT. Brachyura .... Zoaea of tlie Brachyura Metazoaea . . Megalo])a Abbreviated development of some Brachyura 11. Metamorphosis of the Stomatopoda Erichthoidina stage Erichthus stage .... Alima stage .... Connection of the larval forms with definite genera 12. Metamorphosis of the Cumacea 13. Metamorphosis of the Anisopoda 14. Metamorphosis of the Isopoda Larvae of the Anceidae Larvae of the Bopyridae . Larvae of the Entoniscidae 15. Metamorphosis of the Amphipoda 16. General considerations regarding the development of the Crustacea .... Phylogenetic significance of the Zoaea Phylogenetic significance of the Nauplius Hypothetical derivation of the Crustacea from the Annelida Relations of the Protostraca to the Palaeostraca, Pantopoda, and Pcripatus .... Characteristics of the primitive Phyllopoda Relationships of the Entomostraca . Relationship's of the Malacostraca Significance of Nebalia Literature of the metamorphoses of the Crustacea Chapter XX. PALAEOSTRACA. By K. Heider General considerations regarding the Palaeostraca L Metamorphosis of the Trilobita II. Development of the Xiphosura 1. Oviposition, cleavage, and formation of the germ-layers in Limulus .... 2. Development of the external form of the body . Trilobite stage of Limulus 3. Formation of the organs A. Nervous system and sensory organs £. Alimentary canal C. Formation of the mesoderm D. Respiratory organs 4. General considerations regarding the Palaeostraca Relations of the Palaeostraca to the Araclmida Literature of the Palaeostraca PAGE 285 286 289 290 290 291 293 294 296 297 297 301 301 303 303 304 304 305 306 307 309 309 310 313 313 315 316 317 318 318 319 332 332 334 340 340 343 347 348 348 352 353 355 356 356 360 CORRIGENDA. Page 10, line 2 from bottom— 1897 change to 1S6T. „ 57, iu lettering of Fig. 2S— ma cliange to riuj. „ 84, line 10— Denicker change to Deniker. „ 209, line 5 from bottom— Sluier change to Sluitc-r. „ 290, footnote *— enclose in square brack-ets. CHAPTER XV. PHORONIDEA. Although investigators early drew attention to the many structural features in which the genus Phoroms resembles the Bryozoa, this form has nevertheless hitherto usually been classed with the Sipunculids. Recently, however, greater stress has been laid on its relationship to the Bryozoa and the Brachiopoda (Ray Lankester, Caldwell (No. 1), CoRi (No. 4a)). It is chiefly in the anatomy of the adult that the resemblance between these groups is found, but the larval forms may also M'ithout difficulty be compared with one another. I. Embryonic Development. Our knowledge of the first ontogenetic stages of Phoronis is due to the researches of Kowalevsky (No. 6), Metschnikoff (No. 9), FoETTiNGER (No. 5), RouLE (No. 9a), and Caldwell (No. 2). The latter author, whose description we shall folloAv in all important points, arrived at results which frequently difi^er from those of earlier investigators, so that many points seem to require re-examination. The eggs of Phorojiis, according to Kowalevsky, are fertilised while still in the body-cavity of the parent.* They reach the exterior through the nephridial canals which open near the anus and function as genital ducts, and then, enveloped in a vitelline membrane, become attached to the tentacles of the parent, Avhere the young develop uj) to the time of hatching. Cleavage is total and unequal ; the difterence in size between the blastomeres of the animal and those of the vegetative pole is, how- ever, inconsiderable. As early as the four-celled stage, two smaller blastomeres can be distinguished from two larger ; the eight-celled stage shows four smaller and four larger cleavage-spheres symmetri- cally arranged In the further course of this very regular cleavage, * CoRi considers this statement inijirobable, and thinks that fertilisation takes place outside the parent, in the water. PHORONIDEA. a small central cavity develops, and in this way a l)lastula forms which at first is spherical and then becomes elongated in the direction of the future longitudinal axis ; in this l^lastula, a vegetative (ento- dermal) portion consisting of larger cells can be distinguished from a small-celled animal portion. The longitudinal axis coincides with the plane which divides the animal half from the vegetative half. A true invagination-gastrula then develops (Fig. 1), the blastopore being originally oval (Fig. 1 C), l)ut its posterior portion becomes slit-like and soon closes. The anterior part that has remained open persists as the oral -'■' ■t) /■'-■'"""■-^ aperture of the larva (or, more cor- rectly, as the oeso- phageal aperture). The cells near the posterior closed part of the blastopore are found still later to be actively dividing, and, according to Caldwell, they take part in the formation (if the mesoderm, this l)eing especially the case near a dei^ression (Fig. 1 1), (i) at the most posterior part of the blastopore. Caldwell compares Phoronis in this re- spect to the Yerte- brata, and calls the whole of this closed portion of the blasto- pore (Fig. 1 D) the pnnnfiv^ streaJc, and the depression appearing in the latter the ^}r^^»^Y^i'e groove. Meanwhile, the embryo changes somewhat in shape. Its anterior part becomes swollen (Fig. 1 D) as the first indication of the future pre-oral lobe of the larva. The region of the primitive streak noA\- lengthens greatly, so tliat the depression mentioned al)o\'e shifts quite to the posterior end of the embryo. The surface upon which Fio. 1.— GastiMila staye of Plwronis. A, younger, B, older stage ill optical section (after Metschnikoff) ; x, proto- plasmic bodies in the blastocoele (niesenchynie cells ?). C and D, ventral superficial aspects (after Caldwell). C, stage with oval blastopore. D, stage with slit-like, partly closed blastopore ; m, anterior open part of the same, from which the mouth of the larva is derived ; g, posterior pit-like dejiression of the ijriniitive groove ; a pre-oral lobe is distinctly marked off at this stage. EMBRYONIC DEVELOPMENT. the blastopore is situated may be defined as the ventral surface and the opposite as the dorsal surface. The mesoderm, according to Caldwell (No. 2) forms in a very peculiar way. In the region corresj^onding to the most anterior part of the jjrimitive streak, the entoderm-sac shows two lateral j)ocket-shaped outgroAvths (Fig. 2 A, d), at the base of which meso- dermal elements {m') become detached through the proliferation of the entoderm -cells. "When a considerable number of these meso- dermal elements have been formed, they arrange themselves into a c ,eiL Fio. 2.— Fonnatiou of the niesodenn in Phoronis (after Caldweil). A, transverse sectiim through the anterior part of the embryo, showing the origin of the mesoderm. B, trans- verse section through the oral aperture of an older stage. C, horizontal section through an embryo, showing the formation of the posterior coelomic sacs (in") ; only parts of the anterior sacs (m'} are cut through, ec, ectoderm ; en, entoderm ; d, paired archenteric diverticula; g, small pit at the posterior end of the piimitive streak ; m', anterior, m", posterior pair of coelomic sacs. pair of sacs, each enclosing a coelomic cavity (Fig. 2 B, in'). This formation of paired coelomic sacs by the proliferation of cells from the lateral walls of the archenteron may, perhaps, be traced back to the type in which the enterocoeles arise through folding of the jirchenteron. Further back, single mesoderm-cells become separated from the primitive streak and pass into the space between the 4 PHORONIDEA. ectoderm and entoderm. At the most posterior ])art of the streak lies the small pit mentioned above (Fig. 1 D, g ; Fig. 2 C, g) ; this dej^ression soon gives rise to two lateral diverticula that extend anteriorly between the ectoderm and entoderm (Fig. 2 C, m'), and l)ecome the posterior coelomic sacs. The two pairs of coelomic sacs (m' and m") that thus arise are connected together by the isolated mesoderm -cells which have arisen from the anterior ])art of the primitive streak. After the formation of the posterior pair of coelomic sacs, which may, perhaps, be in some way connected with the formation of the nephridia, a shallow ectodermal depression forms posteriorly and fuses with the wall of the archenteron, and here the anal aperture arises. According to Metschnikoff (No. 9) and Foettinger (No. 5), the mesoderm' forms much earlier tlian is stated by Caldwell, by a kind of mesenchyme- formation, single cells appearing in the cleavage-cavity of the blastula-stage (Fig. 1 A, B, a-). These elements, according to the authors just named, are rather small, so that Caldwell's assumption that particles of protoplasm lying iu the blastocoele have here been mistaken for mesoderm-cells appears .somewhat probable. It is, however, possible that the formation of the coelomic s.ics is ])receded by the rise of a mesenchyme. According to Roule also (No. 9a), single mesenchyme-cells are found as early as the gastrula-stage in the primary body-cavity. Later, after the anal aperture has formed, the cells of the primary entoderm are said to increase ill number at the .sides of the anus, and in this way to produce two solid mesoderm-bands, while single cells that become detached mingle with the rest of the mesenchyme. The anterior pair of coelomic sacs above described, which might be called cephalic cavities, now grow out anteriorly, and soon com- ])letely fill the interior of the pre-oral lobe. These sacs seem to yield only the lophophoral cavities (Fig. 5, Ih) and the connected (?) cavity in the epistome (Fig. 5, eh) of the adult, which are separated by a transverse septum from the posterior part of the body-cavity.* This latter is yielded by the posterior pair of coelomic sacs, from which also is derived the median mesentery suspending the intestine (Fig. 5, ms) which is retained throughout life. Secondary, lateral mesenteries, however, are also found. With the development of the pre-oral lobe, the principal sections of the alimentary canal and the coelomic sacs, the chief parts of the embryo are represented in rudiment, and now the whole surface becomes clothed with cilia. Thickening of the ectoderm now takes * [According to Mastermax (No. II.), there are three perfectly distinct coelomic cavities in the larva, viz. — a pre-oral or epistomal cavity, a collar- or lophophoral-cavity, and a trunk-cavity. Their origin is not described. — Ed.] EMBRYONIC DEVELOPMENT. , O place at two definite points. The anterior thickening which occurs at the apical end of the embryo (Figs. 3, 4 B) may lie regarded as the homologue of the neural plate of other larvae, and yields the ganglion known as the brain,* which here, throughout life, retains its original epidermal connection. A second thickening of the ectoderm •appears behind the mouth as a semicircular, ciliated swelling. From this, which may be regarded as the equivalent of the post-oral ciliated ring, the roAv of larval tentacles develops (Figs. 3, 4 B), and also the nerve -strand that runs along their points of insertion. The tentacles begin to appear early as outgrowths of the body- wall ; they rapidly increase in number, fresh pairs being added dorsally (Figs. 3, 4 B). If this row of tentacles is to be traced back to a transformation of a post-oral ciliated ring, we may regard a strongly ciliated swelling appearing at the edge of the pre-oi'al lobe as the pre-oral ciliated ring. The posterior part of tlie body which carries the dorsally displaced anal aperture now grows out into a large cone, the end of which is surrounded by a circum-anal ciliated ring (Figs. 3, 4 (7 and D). Three body-regions can bo distinguished in the larva, Avhicli must now be called the Actvioiroclia (Figs. 3, 4 i? and C) — (1) the pre-oral lobe; (2) the post-oral section which carries the crown of tentacles and covers the jjosterior part of the body like an apron; and (3) the posterior or anal section. Tlie changes just described, which lead to the development of the Adinotroclia, take place after the commencement of free larval life. The youngest larva, just hatched from the egg (Fig. 4 A), is still without the crown of tentacles, two small projections lying near the anus being the only indications of the tentacle-rudiments, f II. Metamorphosis. The Flioronis larva (Figs. 3 and 4 B), the shape of Avhich has been described above, was discovered by Joh. Muller and descril)ed as Actinotroclta hranchiata, and was afterwards more carefully examined by Wagener, Gegenbaur, and others. Krohn (Xo. 7) ami Schneider (]^o. 10) investigated the metamorphosis which led to the production of a Gephyrean-like form, Kowalevsky (Xo. 6) being * The pre-oral lobe, together with the brain of the larva, is, as we shall see below, thrown off during nietaniorpliosis. Since, however, these jiarts are regenerated later, we may still theoretically trace them lack to the corre- sponding parts of the larva. t [For the structure of the Adinotrocha, see Mastehman's recent work (No. II.) on this larval form. These observations, if contirmed, would show that Phoronis is directly related to Balanoglossus, Cejihalodiscus, and llhabdv- plcura. — Ed.] 6 PHORONIDEA. the first to prove that the Adinotrorlia was the young form of the PJioivnis discovered by Wright. Since that time, the meta- morphosis of Phoronis has been more accurately investigated by Metschnikoff (No. 8), Wilson (No. 11), and Caldwell (No. 1). The first changes that take place in the Adinotrocha consist of a simple increase in size tlirough growth and a continual increase in number of the tentacles. At the same time, a sensory organ develops in front of the neural plate, four eye-spots being added to it in one species. The pigment-spots characteristic of the different species now also develop on the pre-oral lobe and on the tentacles. Rudiments now appear of the de- finitive structures that are destined to replace the larval organs. Of these, the first to develop is an invagination of the body-Avall on the ventral surface of the posterior region of the larval body (Fig. 4 C, iv), and in this tlie two layers of the body-wall (the ectoderm and the somatic mesoderm) can be distinguished. This invagination, Avhich soon grows as a much-coiled tube into the larval body-cavity, represents, as Ave shall see, the rudiment of the greater part of the body-wall of the adult. Small truncated processes now develop at the base of the crown of tentacles, and from these are derived the adult tentacles (Fig. 4 D). When these structures have appeared as rudiments, the Adino- trocha sinks to the bottom, the critical moment of the commencement of metamorpliosis having arrived, the whole process being accom- plished Avithin a quarter of an hour. ISIetamorphosis is introduced by the evagination of the tube mentioned above (Fig. 4 D), this being ju-otruded like the tentacle of a snail. Since the alimentary canal, together AA'ith its mesenteiy, is attached to the inner end of this tul)e, it soon has to folloAV the movement thus begun, and so comes to lie inside the completely evaginated tube (Fig. 4 E). During these changes the rest of the larval body loses its turges- cence. The oral and anal apertures therefore come to lie remarkably near one another. The pre-oral lobe of tlie larva is noAv throAvn off, and the same fate overtakes the larval tentacles and the circum-anal ari' FiQ. 3. — Larva of I'horonis (Aclino- trocha), after Metschnikoff, from Balfour's Text-hook, m, mouth ; an, anus. METAMORPHOSIS. 7 ciliated ring. The lost larval organs in the oral region consequently have to be replaced liy extensive regeneration. It has already been mentioned that the rudiment of the permanent croAvn of tentacles which soon assumes the form of a horseshoe -shaped lophophore appears early. Fio. 4. — A Sfi'ies of stages in tlie metamorphosis of I'horonU from the Ai'tiiwtrocha (after Metschnikofk, from Balfour's Text-book). A, young larva. B, larva after the develop- ment of the post-oral ring of tentacles. C, larva in which the invagination (ir) is com- mencing, from which is derived the greater part of the body of the Phoronis. D, invagination partially, and E, completely everted, an, anus ; m, mouth. By means of this remarkable metamorphosis, an animal is pro- duced the body of which is derived for the greater part from a prolongation of the ventral side of the larva. The dorsal side, on the contrary, has undergone considerable abbreviation, and can be recognised in the short tract lying between the mouth and the anus. 8 PHORONIDEA. There are a few points connected with the rise of the organs which must be described a little more in detail. The nephridia of Phoronis which have recently been carefully studied by Caldwell (No. 1) and Com (Nos. 4 and 4a. and Ectoproc. Bryoz. Lit., No. 46) are, in the adult, paired, looped, ciliated canals opening externally near the anal aperture (Fig. 5, n). Each nephridiuni consists of a curved tube starting from the nej^hridiopore, passing outside the transverse septum, and opening through two ciliated funnels into the posterior part of the body-cavity. These organs thus belong essentially to the posterior section of the body. These nephridia, which in structure may be compared to the segmental organs of the Annelida, arise, according to Caldwell, through the metamorjthosis of the larval nephridia discovered by him in the Actino- trocha. The latter, in their structure, recall rather the head-kidney of the Annelida. They are paired canals which open externally behind the transverse septum or diaphragm on either side of the invaginated ventral sac, tlie inner blind ends being connected with a number of excretory cells. These are star- shaped, a line canal leading from each of them into the common duct. Witli regard to the origin of this larval kidney, Caldwell thinks that we may regard the paired ducts as the remains of the communication established between the posterior coelomic sacs and the surrounding medium by means of the pit- like depression mentioned above (Fig. 2 C, g). The excretory cells, on the other hand, would have an independent origin in the .somatic mesoderm-cells. Phoronis is specially distinguished by the possession of a closed blood-vascular system which ramifies on the intestine and in the tentacles, and in which a fluid containing red blood-corpuscles circulates. The details of the origin of this blood-vascular system are as yet not known, but it appears that the vessels oiiginate by dehiscence in the splanchnic layer of the mesoderm. While, according to Cori, the vascular system of the adult is completely closed, there seems, in the larva, to be a communication between it and the cephalic part of the body-cavity. In the latter, the blood-corpuscles are said to arise in large agglomerations. III. General Considerations. The Adinotroclia may without difficulty be regarded as a some- what modified Trocliophore. Indications are found in it of a pre-oral and of a jjost-oral ciliated ring, the latter being transformed into a row of tentacles, and of the characteristic neural plate. The development of the mesoderm is of special importance. This is arranged as two pairs of coelomic sacs which, however, do not appear to be fully equivalent to one another, since, according to Caldwell, they differ in their origin. The anterior pair of coelomic sacs yields the cephalic cavity (lophophoral cavity),* the posterior pair the whole body-cavity of the adult trunk. A transverse diaphragm dividing the two parts of the body-cavity is found in the Adinotrocha on a level with the crown of tentacles. The nephridia belong to the posterior coelomic section. We can tlius distinguish, in the body of the AHiitufrorha, as well * [In the light of Masterman's observations this point requires reinvesti- gation.— Ed.] GENERAL CONSIDERATIONS. 9 as in that of the adult Phoronis, a cephalic section and a trunk section. Further, the body is uiLsegmented, and there are no indica- tions of the descent of Phoronis from a segmented ancestor. The metamorpliosis of Phoronis affords important data for the orientation and interpreta- tion of the adult animal. AVe have seen that the principal parts of the body owe their origin to an excessive growth of the ventral side. The longi- tudinal axis of the adult therefore lies at right angles to that of the Adinotrocha. The dorsal surface is shortened, being restricted to the short tract lying between the mouth and the anus (Fig. 5, ni-a). Al- though the larval organs are cast off, we may regard the organs that appear in their place as their full equivalents. We shall therefore have to derive the tentacle-crown of the adult from the post-oral ciliated ring of a Troclio- pTiore-\\\iQ ancestor, and to regard the epistome as the transformed pre-oral lobe. The remarkable meta- morphosis of Phoronis certainly does not corre- spond to any phylogenetic condition. We shall have to assume that a very gi-adual shifting of the anal aperture along the dorsal middle line took place in the worm-like ancestor of Phoro7iis, which perhaj)s lived in I Fig. 5. ^Diagram representing a median longitudinal section of Phoronis (constructed after Cori). a, anus; eli, epistomal cavity; cp, epistome; g, gan- glion ; ?, intestine ; I, fenestrae in the mesentery ; Ih, lophophoral ca\'ity ; in, mouth ; mg, second stomach [intestine of Benham]; ms, dorso-ventral mesentery ; w, nephridium ; of, oesopliagus ; r, circular nerve ; t, tentacles ; vm, first stomach. 10 PHORONIDEA. sand or even in a tube, a sliifting similar to that wliicli takes place in the Siinmrulidae {YoX. i., p. 363). The fact that, in the trunk region of the adult Phoronis the symmetrical arrangement of the body begins to be disturbed, is no doubt connected Avith the tubicolous manner of life. There is thus- a longitudinal nerve-strand on the left side, and, according to CoRi (j^o. 4), a new plane of symmetry, different from the original one, is evident in transverse sections in the grouping of the longitudinal muscles in Plioronis psammojjhila. Indeed, in the presence of secondary and tertiary mesenteries, which leave between them distinct intervals containing the groups of longitudinal muscles, there is a tendeircy to the development of the radial t3'pe. It should here be pointed out that, in Plioronis, the cephalic region is often lost and again regenerated. According to CoRi (No. 4a), not on]}' the tentacle crown, but Avith it a part of the oesophagus^ the epistonie, the ganglion, the so-called lophophoral organs, the blood-vascular ring, and perhaps also the nephridia are lost. The spontaneous throwing off of the cephalic section is specially liable to occur Avhen the animals are placed in unfavouralile conditions- of existence. This process closely resembles the throwing off of tlie head in Pedicellina and in the Tubularia ; and the same process- is found in the Ectoproctous Bryozoa, in the disintegration and subsequent regeneration of the polypide (p. 55). LITERATURE. 1. Caldwell, W. H. Preliminary note on the structure, develop- ment and affinities of Phoronis. Proc. R. Soc. London. Vol. xxxiv. 1882-1883. 2. Caldwell, W. H. Blastopore, Mesoderm, and Metameric Seg- mentation. Quart. Journ. Micro. Set. 1885. Vol. xxv. 3. Claparede, E. Beobachtungen liber Anatomie und Entwick- lungsgeschichte der wirljellosen Thiere. Leip':ig. 1863. 4. CoRi, C. J. Beitrag zur Anatomie der Phoronis. Inaug.-Diss. Pray. 1889. 4a. CoRi, C. J. Untersuchungen liber die Anatomie und Histologie der Gattung Phoronis. Zeitschr. f.Wiss. Zool. Bd. li. 1891. 5. FoETTiNGER, A. Note sur la formation du mesoderme dans la larve de Phoronis hippocrepia. Archiv. de Biol. Tom. iii. 1882. 6. KowALEVSKY, A. On the anatomy and development of Phoronis (Inaug.-Diss. Kussian). Petersburg, 1897. See Leuckart's. Summary in Jahresh. Archiv. Naturg. Vol. xxxiii. 1867. LITERATURE. 11 7. Krohn, a. Ueber Pilidium uml Actinotrocha. Archiv. f. Anat. u. Phijs. 1858. 8. Mbtschnikoff, E. Ueber die JMetaniorpliose einiger Seethiere. Zeitschr. /. Wiss. ZooJ. Bd. xxi. 1871. 9. Metschnikoff, E. Vergl. embiyol. Studien. (3) Ueber die Gastrula einiger Metazoen. Zeitschr. f. Wiss. Zool. Bd. xxxvii. 1882. 9a. RouLE, L. Sur le developpement des feuillets blastodermiques chez le.s Gephyriens tubicoles (Plioronis Sabatieri n sp.). Cornpt. Rend. Acad. Set. Paris. Tom. ex. 1890. 10. Schneider, A. Ueber die Metamorphose der Actinotrocha branchiata. Archiv. f. Anat. u. Phi/s. 1862. 11. Wilson, E. B. The origin and significance of the metamor- phosis of Actinotrocha. Quart. Journ. Micro. Sci. Yol. xxi. 1881. See besides the older treatises of Joh. Muller, Wagener, Gbgenbaur, Leuckart, Pagenstecher, etc. APPENDIX TO LITERATURE. 1. Benham, W. B. The Anatomy of Phoronis Australis. Quart. Journ. Micro. Sci. Vol. xxx. 1890. II. ]Masterman, a. T. On the Diplochorda. Qtcart. Journ. Micro. Sci. Yol. xl. 1897. CHAPTER XVI. BRYOZOA ECTOPROCTA. Colonial Bryozoa with the anus outside the lophophore, with a well-developed introvert and a spacious coelom. Systematic : — A. PHYLACTOLAEMATA. Fresh- water forms, with horse- shoe-shaped lophophore prolonged into two arms ; with epistome overhanging the oral aj^erture (Cristafella, Plumatella, Pedinatella, Fredericella). B. GYMN"OLAEMATA. Mostly marine; with circular lopho- phore ; without epistome. I. Cyclostomata. Apertures of the zooecia terminal, circular and devoid of closing apparatus; without appendages (Crisia, Diastojwm, Hornera, Tuhuliijora, Frondipora). II. Ctbnostomata. Apertures of the zooecia terminal, closable by means of tooth-like or seta-like projections of the tentacle-sheath. 1. Haley onellea {Alcyonidium, Pherusa, Flustrella). 2. Stolonifera (Vesicularia, Amathia, Boicerhanlda, Farella, HypoplioreJla ; the fresh-water forms Paludicella and Vidorella also belong here). III. Chilostomata. Apertures of the zooecia not, as a rule, terminal, usually closable by a movable lid. 1. Stolonata {Aefea, Eucratea). 2. Radicellata. a. Cellularina (Cellularia, Scrupocellaria, Biigida). h. Flustrina {Flustra, Memhranipora). c. Escharina {Retep)ora, Microporella, Eschara, Lepralia, Schizoporella). FORMATION OF THE EGG. 13 I. Formation of the Egg, Fertilisation, Position of the Embryo. The genital products of the Ectoproctous Bryozoa arise in cell- masses which originate as growths of the mesodermal parenchy- matous tissue (marine Ectoprocta), or of the peritoneal epithelium corresponding to the latter (Phylactolaemata), on the inner side of the hodj'-wall {endocyst), or else in strands of the so-called funindar tissue. The ovaries are very generally found on the neural wall (dorsal side) in the anterior or middle part of the hody ; the testes lie at the Ijase of the zooecium {i.e., in the proximal part). In the fresh-water Bryozoa, the genital rudiment frequently bears a relationship to that part of the mesenterial strand which is known as the funicidus. In Paludicella, for instance, the eggs lie on the body-wall near the point of insertion of the upper funiculus, while the spermatozoa arise on the basal portion of the lower funiculus. In the Phylactolaemata, on the contrary, the ovary lies on the oral body-wall, Avliile the spermatozoa, as a rule, develop in aciniform masses at the upper part of the funiculus. In Cristatella, the spermatozoa arise on the mesodermal septa of the body-cavity. The genital products pass into the body-cavity, where, in some forms, fertilisation takes place. Since the Bryozoa are, as a rule, hermaphrodite, and as it is difficult to state in what Avay foreign spermatozoa can reach the body -cavity, self -fertilisation has been assumed to occur. The fertihsed eggs either pass through the whole of their embryonic development, up to the time when the ciliated larva is formed, within the body-cavity of the parent, or else, through the dehiscence of the body-wall, reach the tentacle-sheath {Valkeria, Joliet, ISTo. 17; Lepralia and Vesicularia, Ostroumoff, No. 26), in the cavity of which they pass through the embryonic stages till the larva hatches, or else the eggs (as in many Chilo- stomata) are received into special outgrowths of the zooecium which serve as brood-cavities. These are the ooecia and ovicells which have been described as individuals of the polymorphous Bryozoan stock peculiarly metamorphosed for the care of the brood.* In those cases in which the larvae pass through their earliest ontogenetic stages in the body-cavity of the mother, they escape either through the aperture of the zooecium, after the polypide to which they belong has undergone degeneration, or else there is a special aperture near the base of the tentacles for the passage of the * [In some Cyclostomata (Crista) tlie ovicells are undoubtedly modified zooecia in which the polypide is rudimentary or degenerate. — Ed.] 14 BRYOZOA ECTOPROCTA. embryos. Such an aperture was found in Farella by Van Beneden, and in Hupojiliorella by Ehlers. In a few Gymnolaemata {Alcyonidium gelatinosum, Memhraiiipora pilosa) Farre, Schmidt, and Hincks found a flask-shaped ciliated canal [nepliridium V) connecting the body-cavity Avith the surrounding medium and opening out between the tentacles. Prouho (No. 28a) recently observed in some species of Alcyonidium, that this inter- tentacular organ was connected with oviposition. In Alcyonidium alhidum, tlie eggs probably undergo fertilisation Avithin the body- cavity and become surrounded by a soft shell, being then ejected into the surrounding medium through tlie intertentacular organ while the parent polypide is extended, further development taking place in the water. More complicated conditions are found in Alcyonidium duplex. In tliis form, at the period of sexual maturity, the (male) polypide of a zooecium wliich is without an intertentacular organ develops spermatozoa. At the same time, on the aboral side of this zooecium, a second (female) polypide, provided with an ovary and an intertentacular organ, is developed ; soon after fertilisation the male polypide degenerates. The fertilised eggs, which are provided with a shell, probably reach the tentacular sheath through the intertentacular organ ; here, each attached by a fine stalk, they pass through their further development. AVhen the polypide is extended, the part of the tentacle-sheath carrying the eggs is evaginated. In this position the egg-shell bursts and the larva swims about freely. Within the ovaries, which originally consist of small, indifferent cells, a few (2-5) young egg-cells soon appear, the remaining cells becoming grouped round these to form a follicular epithelium (ViGBLius). Of these young egg-cells, two at first develop more than the others ; but, as a rule, only one egg becomes fully mature. This egg remains at first connected witli the ovary l)y means of a strand, Avliile what remains of the ovary draws back to the body-wall so as to serve later as the place of origin of another egg. (On the conditions of the maturation of the egg and the ooecia of the Phylactolaemata, see below, p. 33.) Remarkable conditions of enibryogenesis were found by Hahmeu (No. 15) in Crista, a form in which the rijie ovicells contain a large number of embryos.* Beside these is found a protoplasmic nucleated network which sends out finger- * [This condition is brought about by a fission of the primary sexually- produced embryos. As many as one hundred secondary embryos produced l)y budding are found in a single ovicell. Hakmeu, Quart. Journ. Micro. Sci., Yols. xxxiv. and xxxix. — Ed.] EiMBRYONIC DEVELOPiMENT. 15 shaped processes, from the free ends of which tlie embryos are developed like huds and eventually cut off. In quite young ovicells, on the contrary, a single egg-cell, surrounded by a follicle, is found ; this appears to give rise to the finger-shaped Inidding organ mentioned above as producing the embrj-os. In Crista, therefore, the number of embryos produced by the early division of the primary embryo is larger. The mature larvae swim out through tlie tubular aperture of the ovicell. According to Van Beneden and Peruens, the maturation of tlie egg is connected in a certain regular way with the disintegration of the polypidc* that produces it, and, in some forms {Flustra truncata, Microporella malusii, Bugv/a simplex and turhinata) with its later regeneration, so that, when the egg is fully mature, the polypide undergoes histolysis, and becomes changed into a bro^\■u body. While the ovary brings another egg to maturity, a new polypide forms. For details of these processes of regeneration see below, p. 55. Among these must also be reckoned the above observations of PiiouHO on Alcyonidiiim duplex. In the Phylactolaemata also, as a rule, during the development of the embryos and the statoblasts, the polypide to which they belong degenerates, but, in this case, there is no subsequent regeneration. 11. Embryonic Development. The mature, sjDherical, or ellipsoidal eggs are surrounded by a hyaline membrane called by Pergens the chorion. In the ovum can be recognised a vesicular nucleus with spherical nuclear bodies, and a granular yolk often yellow or brown in colour. The two polar bodies, Avhich are usually of unequal size, correspond in position to the animal pole of the egg. The first ontogenetic j^rocesses in the egg of the marine Ectoprocta are best known in Lepralia (Barrois, Nos. 6 and 1), in Tendra zostericola and Bowerhanlcia (Kepiachoff, Nos. 32 and 34), in BiKjuIa calathus (Vigelius, No. 39), and in Microporella mahmi (Pergens No. 27). In these eggs cleavage is total and almost equal (Fig. 6). The two-celled stage is attained Ijy means of a meridional furrow, and the four-celled by means of another meridional furrow at right * The expressions " polypide " and " cystid '' correspond to an older view, according to which the cystid forming the wall of a chamber represents an individual which gives rise asexually through budding to the polypide. Each chamber of the Bryozoan stock, consisting of a polypide and a cystid, would then represent a double individual or a miniature colony. This view was founded on the great independence of the polypides as shown in the processes of degeneration and of regeneration above-mentioned. Althongli we do not share this view, we still retain in use these expressions which luive becohie established. The cystid, then, means to us the lower part of the body- wall, while the polypide rejn'esents the retractile anterior section of the body with the intestinal canal attached to it. These two are merely parts of one individual. [Acting on the suggestion of Dr. Harmeu, we have sulistituted the term " zooecium " for cystid in referring to the outer parts of the ordinary adult individual, as this term is of much more general use in English works on the Bryozoa. — Ed.] 16 BRYOZOA ECTOPROCTA. angles to the first ; this stage, in consequence of an equatorial furrow, passes over into the regular eight-celled stage (Fig. 6 A). A small hlastocoele can frequently he observed as early as this stage, and the animal jDole can he distinguished from the vegetative pole by a differ- ence of size — often only very slight — in their respective hlastomeres. In the further segmentation there is a peculiarity worthy of note. The sixteen-celled stage is reached through the appearance of two A J e^ xc Fig. 6. — Cleavage and embryonic stages of Biir/ula calathtis (after Vigelius). J, stage with eight blastomeres. £, stage with sixteen blastonieres. C, stage witli thirty-two blasto- meres. All seen from above. The dotted lines io A and B indicate the direction of the next cleavage planes. D, blastnla-stage in vertical section E and F, gastrula-stages. G, a later ontogenetic stage c, corona-cells; ec, ectoderm; en, entoderm; /, cleavage- cavity ; z, central tissue derived from the primary entoderm. furrows lying on either side of the first meridional plane and running parallel to it. An embryo thus results, each half of which consists of eight cells (in two rows of four. Fig. 6 B). A similar division leads to the thirty-two-celled stage (Fig. 6 C), two furrows forming parallel to and at the sides of the second meridional plane. Each half of this embryo consists of four rows, each containing four cells. This stage is followed by less regular divisions and results in a EMBRYONIC DEVELOPMENT. 17 ])Iastula liearing a general resemljlance to a liicoiivex lens (Fig. 6 D). A more active division of the cells of the animal (later alioral) pole can now be remarked, while, from the vegetative pole, four entoderm- cells wander into the blastocoele (Fig. 6 E). As soon as the entoderm-cells have shifted inwards, the surrounding ectoderm-cells seem to join together. This type of entoderm-formation may be traced back to polar ingroAvth. A shallow depression in the region of the vegetative pole is often noticeable during these processes. After the ectoderm has completely closed, the four entoderm-cells multiply, their derivatives becoming grouped so as to enclose a small slit-like cavity (archenteric cavity, Fig. 6 F). This cavity is usually obliterated later, the entoderm-cells multiply still more, and finally yield a mesenchyme or embryonic connective tissue (the inner mass, central tissue, "FiillgeAvebe," Fig. 6 G, z), which completely fills the blastocoele. This tissue, Avhich seems to be of little impor- tance in the later development, playing merely a jmssive role, soon assumes a reticular character (Fig. 9 B, /). While these processes have gone on Avithin, tAvo roAvs of cells at the edges of the biconvex embryo have become distinguished by their larger size (Fig. 6 IJ and F) ; one of these roAvs at a later stage becomes very conspicuous, forming the rudiment of the ciliated ring or girdle of the larva, and is knoAvn as the corona or velum (Fig. 6 G, c). After the larva has attached itself, the central tissue gives rise to the meso- dermal portion of the first sedentary individual (primary zooecium), i.e., to the parietal mesoderm-layer on the inner side of the endocyst, and to the outer layer of the sac-like rudiment of the polypide (Fig. 12, b, p. 29). The rest of the central tissue unites with the cell-mass yielded by the degeneration of the larval organs to form a sfjlierical mass (the so-called brown body), Avhich is composed of remains of cells, pai'ticles of food-yolk and detritus, and which, when the first polypide develops, is used up as food material. The central tissue is derived direct from the primary entoderm of the gastrula-stage. This germ-layer does not, as a rule, separate into mesoderm and definitive entoderm. Barrois (No. 7), indeed, maintains that such a separation does take place in Lepralia unicornis, the mesoderm there becoming arranged in paired mesoderm-bands. His observation, however, has received no further confirmation through researches made in connection with various other genera. In those few forms in which an alimentary canal is developed, there nmst of course be a separation of the mesoderm from the definitive entoderm. In quite early stages (corresponding to that in Fig. 6 F), a separation of the mesoderm from the entoderm was observed Ity Prouho (No. 2Sb) in the embryos oi Alcyonidium and Membranipora pilosa {Cyplwnautes), but no further details as to the rise of the mesoderm could be observed. The embryo Avhich originates in this manner (Fig. 6 G), in Avhich^_^__ two body-layers (ectoderm and central tissue ^ primary entodpqi\) y\- A/ ^ i' ^^^^ > ) is remarkable, and contrasts with the somewhat thin ectodermal layer of the retractile disc. Closely con- nected witli the latter layer, in the anterior part of the larva, there is a cell-mass (Fig. 7 B, (J) which Harmer (No 13) regards as the l:)rain. This author arisnig Fig. 1.—A, larva of Alcyonidium mytili (after J. Babrois). B, longitudinal vertical section through the larva of Alcyonidium (diagram constructed after the figures of Harmer). c, coronal cells; g, brain (?) ; m, oral aperture; o, pyriforni organ; i>, mantle-cavity; r, retractile disc; s, sucker. observed, within this cell-mass, fibrous strands, some of which are said to become connected with cells of the ectodermal furrow, while others can be followed into the neighbourhood of the large coronal cells. Should these statements be confirmed, we should appear justified in considering the pyriform organ as a sensory organ, Avhereas until now its histological constitution seemed to point to the conclusion that it functioned as a gland. The presence of a larval brain comparable in position to the neural plate would be of importance in tracing the Ectoproctous larvae back to the Troclioplwre. The statements of Harmer have recently been confirmed by the researches of Prouhq (Xo. 28) in connection with Flustrella, but on account of the LARVA OF TENDRA. 21 difficulty of investigating this point, we are unable as yet to accept with certainty the interpretation that these fibrous strands are nerves. b. Larva of Tendra. If Ave leave out of account Cyplionautes, to which reference will shortly be made, Ebpiachoff (ISTo. 32) deserves the credit of having been the first to prove the existence of a well-developed alimentary canal in a typical Ectoproctous larva. This author found in the larva of Tendra ( Memhraiiipora) zostericola, which in other respects follows the type of the ordinary Eschar ina larva, and also resembles the Alcyonidiura larva, an intestinal canal, the oral aperture of which lies in front of the sucker, while the anal aperture lies behind that organ. The Tendra larva Avould thus approximate to the larva of Alcyonidium in the topography of its internal organs. c. Larva of Membranipora and Flustrella. The larva of Membranipora has frequently been investigated Ehrenberg placed it under the name of Cyplionautes among the Rotatoria; later, JoH. INIuller compared it with the larva of Mitraria; and Semper, Leydig, and CLAPARiiDE (Xo. 1) declared it to be a Lamellibranchiate larva, Schneider (Xo. 5) being the first to establish its true affinities as the young of Memln-anipora. Later observations were made by Allman, Mbtschnikoff (No. 19), Hatschek (No. 2), Barrois (No. 6), Repiachoff (No. 4), and Prouho (No. 28) as to the structure of Cyplionaides, while its attachment and transformation were investigated by both Schneider and OsTROUMOFF (No. 3). It appears from the statements made by these authors that Gyplionautes agrees closely in all important points with the typical Ectoproctous larva {e.g., Alcyonidium larva), the apparent differences between them being of an unessential kind. Cyplionautes (Fig. 8.) resembles in shape a laterally compressed bell. The lower surface is bordered by cilia. If we compare this with the corona of the Alcyonidmm. larva, we find that we must regard the external surface of the bell as the aboral side, and the under surface of the bell as the oral side partly invaginated to form a vestibule or atrium. The whole body is covered externally by two triangular shell-valves connected together along their anterior edges, which run parallel to the fibrous strand (Fig. 8, /"). The two valves that come into contact along this hinge-line cover the laterally compressed body on the two sides. Their posterior edges, which run parallel to the mid gut and the hind-gut (the intestinal edges), are 99 BRYOZOA ECTOPROCTA. shai'i^ly bent round towards the middle line of the body. The shell- valves meet at this point like pincers. N'ear the lower aperture (the entrance to the atrium) the valves gape. At the aboral pole, a prominence (r) covered Avith stiff cilia projects from the shell, but can be withdrawn into it. In this we recognise the homologue of the " retractile disc." Some uncertainty still prevails as to the ciliated ring that runs round the edge of the bell. According to Ehrbnberg and Hatschek, there is a single com- plicated ring, forming many coils, but Claparede and Schneider, whose view has been adopt- ed Ijy Prouho, think they can distinguish Fig. 8. — Larva of Membranipora = Cypho)icmtes (comhincA from the figures of Claparede, Schneider, and Hatschek). a, atrium ; an, anal aperture ; /, fibres connecting the pyri- form organ (o) and tlie retractile disc (r); in, mouth; s, glandular thickening (liomologue of the sucker); s', similar, smaller organ (shell-adductor muscle).* two distinct rings of cilia, an anterior ring in the neighbourhood of the pyriform organ (o) and a posterior, circum-anal ring, part of which sinks into the atrium. On the oral invagi- nated surface we notice lirst, in the anterior angle, a cell- mass provided with a tuft of cilia; this is the pyriform organ (o). According to Repiachoff, two parts may be distinguished in it ; an invaginated, sucker-like epithelial thickening surrounded by a ring of cilia (ecto- dermal furrow), and a cell-accumulation lying above this (the actual pyriform organ). As in Alcyonidium, this organ is connected with the retractile disc (r) by fibrous strands (/), some of which have been claimed as muscle-fibres and some as nerve-fibres (Prouho). Cyplwnautes has a fully-developed and functional alimentary canal. The oral aperture (»?) is found in the base of the atrium, and leads * [The organs s and s' are not isolated structures as represented in the figure, but are situated underneath the Ijody-wall, which latter should be indicated as a line extending forwards from the anus below s and s' and joining the ciliated ring near the mouth. — Ed.] LARVA OF MEMBRANIPORA AND FLUSTRELLA. 23 into a very expansible oesoi^hagus, which ascends towards the aboral surface ; here it bends sharply downwards and joins the dilated stomach which passes into the hind-gut. The anal aperture {an) opens into the atrium. In the space between the mouth and the anal aperture there is a paired glandidar thickening («). Near this kidney-shaped organ lies a similar smaller organ [s). Ostroumoff has shown that Avhen the larva begins to change into the sedentary form, fixation takes place by means of this paired disc and the cementing substance secreted by it, so that Ave may regard it as the homologue of the sucker of other Ectoproctous larvae. The smaller organ seen near the sucker was assumed by Schneider and H. Prouho to he the transverse section of the adductor muscle of the shell. It is evident from the above description that CijiiJioncuites agrees in all important points witli the typical Alcyonidmm larva, from -which it is distin- guished chiefly by the slighter development of the retractile disc and by the absence of a circular mantle-cavity surrounding that organ. In place of this cavity we find an extensive surface (mantle-surface) on the aboral half of the body, which secretes the two cuticular shell-valves. The oral surface does not here, as in Alcyonidium, bulge out downwards, but is withdrawn into the interior of the body, forming an atrium such as is also found in the Entoproctous genus Pedicdlina. Intermediate between the C'yphonautcs and the Alcyonidium type we may place the larva of Flustrella Mspida, known through the researches of Metschnikoff (No. 20), Baurois (No. 6), and Prouho (No. 28). In this larva, the rudiment of the alimentary canal appears in the embryonic stages, but degenerates later. In other respects, the development of the Flustrdlcv larva agrees closely with that of the Alcyonidium larva. An embryo develops in which the central part of the aboral surface becomes marked off from the corona liy a circular mantle-cavity. Only secondarily does the corona bend round downwards in the form of two lateral lobes, apjiarently enclosing the oral surface in an atrium, while the circidar mantle-cavity disappears. Paired shell -valves then apjiear and cover the whole aboral surface with the exception of the very small retractile disc. A similar though less complete trans- formation in the same direction is found iu the larva of Eucratea chelata- (Barrois). The researches of Prouho have led to interesting revelations as to the internal structure of the Flustrella larva. With regard to the presence of a nervous system, Prouho's researches agree essentially with those of Harmek. In the Flustrella larva, the retractile disc is connected, as in Cyphonautes, with the pyriform organ by a strand of muscle-fibres. Along this strand runs a bundle of nerve-fibres ; these start from the retractile disc ami can be followed as far as to the cells of the ectodermal furrow and of tlie corona. In their course, single fibres of this bundle become connected with ganglion -cells. Prouho thinks that he can trace back the elements regarded by Harmer as ganglion-cells to simple mesoderm-cells, and considers certain elements which are provided with large luiclei as the true ganglion-cells of this region. When the complex organ connected with the ectodermal furrow is more carefully examined, an inner cell-complex (glandular organ, pyriform organ 24 BRYOZOA ECTOPROCTA. in a restricted sense) is found to be distinct from the ciliated cells of the ectodermal furrow. In this organ, which is evidently of glandular structure, three parts can be made out The lateral parts show radially arranged cells opening into tlie ectodermal furrow, while the cells of the middle part run longitudinally and open into a small pit lying in front of the ciliated tuft of the ectodermal furrow (Prouho). Further, there is in this larva a highly developed musculature, consisting of lateral parietal and longitudinal muscles, and of a shell-adductor, already noticed by Metschxikoff. A mesodermal cell-layer lying below the ectoderm plays an important part in metamorphosis, the mesoderm-layer of the developing polypide and of the primary zooecium being formed in it. The transformation of the larva of this type into the sedentary form (primary zooid of tlie colony) takes place in the same way as in other Ectoprocta ; the reader is therefore referred to the description of the metamorphosis of Bugula and Lepralia, given on p. 27. It need here only be mentioned that Cyplionautes, whose metamorphosis has been made known through the researches of Schneider and Ostroumoff, attaches itself by means of the paired, disc-like organ mentioned above on the oral side of the body, and on the degeneration (histolysis) of the internal organs, changes into a spherical mass covered by the shell-valves, this being the rudiment of the primary zooecium. While the ectocyst of the zooecium is secreted on the surface of this mass, the polypide of the first chamber is derived from the retractile disc through ontogenetic processes to be described later (Schneider and Ostroumoff). With regard to the last point, the recent researches of Prouho in connection with Flustrella have led to the remarkable discovery that the retractile disc, like all other larval organs, becomes invaginated and undergoes degeneration. During this process, a thickened disc forms out of the surrounding ectoderm, mesodermal elements joining it on its inner side. This bilaminar plate, which lies in the place of the degenerated retractile disc, now in its turn becomes invaginated, and thus gives rise to the sac-like rudiment of the polyjiide. Even though most other authors regard the polypide as derived directly from the retractile disc, it must be considered probable that the ontogenetic process here described in connection with Flustrella is the rule in the development of the poly])ides in all Ecto2)roctous larvae. Pherusa tubxdosa, which, according to Pkouho (No. 28a), is also distinguished by the possession of a bivalve shell, is allied to the above tyjie, and the larva of Uyiwphorella, according to the same author (No. 28b), also approaches this type. 2. Type of the intestineless Chilostomatous Larva with slightly- developed corona. The larvae of this type, in outward appearance, almost exactly resemble the Alcyonidium larva. They chiefly occur in the Escharina tribe, an examjDle being found in LepraUa Pallasiana, Avhich TYPE OF THE INTBSTINELESS CHILOSTOMATOUS LARVA. 25 has been described in detail by Barrois (No. 9). Here also we find the oral and aboral surfaces separated by a corona of somewhat large ciliated cells. The aboral surface shows the well developed retractile disc surrounded by a circular mantle-cavity ; on the oral surface, the ectodermal furrow with its ciliated tuft and the sucker- invagination are situated. Strong pigmentation is often found in the larvae of this and the following types, paired patches of pigment, like eye-spots, being found in individual cases. The most essential distinction between the larvae of this type and the Alcyonidium larva is the entire absence of the alimentary canal. We assume that the inside of this larva contains only the so-called central tissue. In the very imperfect state of our knowledge of the anatomy of these larvae, it must still be regarded possible that many of them may possess a well-developed alimentary canal. We are therefore unable at present to define exactly the distinctions between this type and the one last described. The metamorphosis of the larvae of this type will be described together with that of the following type. 3. Type of the intestineless Chilostomatous Larva with highly-developed corona. a. Structure of the Larva. We shall take as type of this group the larva of one of the Cellularina, the Bugula larva (Fig. 9), which has become well known to us through the researches of Nitsche (Xo. 22), Claparede (Xo. 10), Salbnsky (Xo. 37), Barrois (No. 9), and A^igelius (iS^os. 39 and 40). The intestinal canal here also is altogether wanting. The primary entoderm changes direct into the so-called central tissue {Fig. 9 i?, /). The distribution of the ectodermal organs is the same as in the larvae of the Alcyonidium and the Escliarina types; the Avhole form of the larva, however, is somewhat different, chiefly on account of the great growth of the coronal cells (c) in the direction of the principal axis. These cells lengthen greatly, the whole of their surface becoming covered with cilia, while in the larvae of the preceding types these cells usually bear only one row of cilia. This extension of the corona encroaches greatly on the aboral and oral surfaces. The retractile disc (r) is consequently smaller, and the ectodermal furrow (e) is so grown over by the increasing coronal cells that it at last appears enclosed by the coronal region. In this way a more or less spherical or rounded larva is 26 BRYOZOA ECTOPROCTA. produced, and in this tlie plane of symmetry is denoted by the position of the ectodermal furrow ; the larva is uius jDeach-shaped. For details of the structure of the Bwjula larva we are indebted chiefly to Yigelius (No. 39). Taking first the organs of the oral surface, the sucker originates as an ectodermal invagination. In the later stages the aperture of the invagination appears to close, at least it was no longer visible. The upper part of the wall of the sucker becomes invaginated into the cavity of the organ itself, thus greatly compressing it. The whole rudiment in this Avay becomes cup-shaped. In the anterior part of the oral surface a second ectodermal invagina- tion appears ; this is found later between the cells of the corona, and is known as the ectodermal furrow (e). The ectodermal cells of this Fio. 9. — v4, larva of Ilugula plumosa (after J. Barrois). B, median section through a Bugula larva (diagram constructed from figures by Vigelius). c, cells of the corona (a single cell of this kind on each side) ; c, ectodermal furrow ; /, central tissue ; o, pyriform organ ; p, mantlft-cavity ; r, retractile disc ; s, sucker. region are elongated, columnar, glandular cells. In direct connection with this modified glandular epithelium, there is a cell-mass which extends towards the interior of the larva and ends posteriorly in three processes. This is the pyriform organ (o). Vigelius is inclined to regard the ectodermal furrow and the pyriform organ as a common glandular complex derived from the ectoderm. The retractile disc (r) arises on the aboral surface as a simple ectodermal thickening, the cells of Avhich lengthen greatly and become arranged in several layers. At the centre these elements seem to be wanting, while at the periphery of the organ they show a radial arrangement. This disc is protrusible beyond the surface of METAMORPHOSIS. 27 M tlie larva. Its covering of stiff setae probably serves some sensory purpose, perhaps transmitting tactile impressions. b. Metamorphosis. The first indication that the larva is ready to attach itself is an alteration in its mode of progression. The larva noAV constantly moves in circles, with its retractile disc protruded to its utmost limit. By a sudden contraction of the body the sucker is shot out, and the fixation of the larva is accomplished by means of a sticky secretion from its lower sur- face. The evagi- nated sucker (Figs. 10 and 11, s) in Bugula shows at first on its under surface a circular furrow, which sepa- rates an outer liroader part from a narrower central portion that pro- jects downwards. At a later stage, the lower side of the sucker flattens into a broad cell- plate {plaque ad- hesive, Barrois), from Avhich is de- rived the basal sur- face of the endocyst of the primary zooecium. The whole of the rest of the body- wall of the primary zooecium, i.e., its upper part and its lateral walls, are yielded by the aboral portion which, in the larva, is comparatively small. This takes place chiefly by an extension of those parts of Fig. 10. — Metamorphosis of the larva of Lepralia unicornis (after J. Bakrois). A, first stage of metamorphosis. The sucker (s) is evaginated, the corona (c) is beginning to bend round. B, next (so-called umbrella) stage. The corona (c) is comi>letely bent round, c, coronal cells ; o, ijyriform organ ; p, matitle-cavity ; )•, retractile disc ; s, protruded sucker ; V, vestibulum ; x, paired organ of doubtful significance (according to Barrois, the mesodermal rudiment of the polypide). 28 BRYOZOA ECTOPROCTA. the ectoderm which previously formed the lining of the circular furrow (mantle-cavity, p), and which in the larva are distinguished by the great depth of their cells (Fig. 10, A). The mantle-cavity {})) is in this way completely obliterated, the circular fold that bordered it externally, and was called the mantle, bending over downwards (Fig. 10, B) ; its inner thus becomes its outer surface, and gives rise to the body-wall of the primary zooecium, Tlie corona at the same time bends round or i;nfolds (Fig. 10, B). The lower ends of the coronal cells (c) retain their position, while the upper ends first shift outward (Fig. 10, A) and then downward (Fig. 10, B), so that each coronal cell at the end of this process has rotated through an angle at first of 90° and finally of 180° (Fig. 1 1 c). We thus have a stage to which Barrois has given the name of the umhrella- sliaped stage (Fig. 10, B). The upper surface is formed by the future body-wall, the lower by the downwardly rotated coronal cells. The edges of the umbrella become applied to the adhesive plate of the sucker (s) and fuse with it, the body- wall thus completing itself (Fig. 11). The displaced coronal cells at the same time fuse with the upper surface of the sucker, and in this way a circular cavity arises {vesfibuhwi, Barrois, Fig. 11 v), the walls of which soon become completely detached from the body-wall and undergo degeneration. This is the fate not only of the corona, but of tlie ectodermal furrow, the pyriform organ, and (where such is present) the larval intestine. There is then found inside the sac-like larva a cell-mass derived from the degeneration of the most important larval organs; this unites with tlie mass that resulted from the degeneration of the central tissue, and forms the so-called brown body. The rudiment of the body-wall of the primary zooecium is formed in the way just described. Its surface soon becomes covered with a chitinous secretion (often impregnated with lime salts) ; this is the ectocyst. During these ontogenetic processes, other changes can be seen to occur at the aboral pole which result in the development of the polyjjide of the primary individual. The retractile disc that lies Fio. 11. — Loiigitiuliiial section f of attached larva of Biigula fl'ihellata (after J. Barrois). c, coronal cells ; r, rudiment of the polypide (which has arisen by the invagination of the retractile organ); s, lower tliickened surface of the evaginated sucker (adhesive plate) ; v, vestibuluni. METAMOIU'KOSIS. 29 here has changed into a plate formed of deep cells Avhich becomes invaginated (Figs. 11, r and 12 A, a). The sac which thus arises soon becomes completely cut ofi' from the body- wall (Fig. 12 B), and represents the rudiment of the ectodermal and also of the entodermal parts of the polypide. A second cell-layer {h) appears on the outer side of this sac, this being the mesodermal layer of the polypide. Various conjectures have been made as to the origin of this layer. According to Barrois (Xos. 7 and 9), there are, in Lepralia, two ectodermal thickenings at the sides of the anterior ectodermal furrow (Fig. 10, .r) ; when the corona bends over, these reach the interior of the primary zooecium, and they alone are unaffected by the subsequent degeneration of the larval Fig. 12. — Two ontogenetic stages of the primary zooecium of Bugvia calathus (after Vigelius). A, invagination of the retractile disc to form the rudiment of the polypide (a); b, cells of the external layer of the polypide ; e, epithelium of ectocyst. B, a somewhat older stage. The invagination forming the rudiment of the polypide has closed. In .-1 the degenerating mass derived from the larval organs is omitted. organs, and remain to form the mesodermal layer of the polypide. These statements have not been confirmed by later researches. Far more probable is the conjecture of Ostroumoff (Nos. 25 and 26) and Vigelius (Xo. 40), Avhich has recently been confirmed by Prouho (Xo. 28), that this cell-layer is derived from the mesodermal larval tissue (central tissue). Of the remainder of this tissue, part seems to yield the somatic mesodermal layer (the so-called j^arenchymatous tissue), while another part joins with the granular masses which have arisen through degeneration in the formation of the so-called brown body. This body becomes connected with the stomach of the newly- formed polypide and is finally absorbed. These processes Avill be 30 BRYOZOA ECTOPROCTA. more fully discussed in a special section on the further development of the rudiment of the polypide (pp. 43, 44, and 55). In considering the metamorphosis of the Biujula larva, which is identical with that of the types hitherto described, we are at once struck l)y the fact that the larval intestine degenerates in all cases, and tliat the intestine of the primary individual is formed anew from an independent rudiment on the ahoral surface. This fact, however, is explained by the cajDacity possessed by the Bryozoa in certain cases of degenerating and again regenerating the alimentary canal or even the Avhole polypide (see p. 55). We are therefore not obliged to regard the polypide of the primary individual as a second generation of the Bryozoan colony, derived by budding from the larva. The production of the primary individual from the larva rather comes under the category of metamorphosis, although this latter becomes very complicated in consequence of the far-reaching degeneration of the larval organs, and on account of the incon- spicuous form assumed by the rudiments of the future jDarts of the permanent body in the larva, these rudiments being somewhat comparable to the imaginal discs of the Insecta. With regard to these rudiments, that of the polypide is repre- sented in the larva by the retractile disc. The rudiment of the endocyst of the primary zooecium lies partly in the mantle-cavity and partly in the sucker-invagination. The body -wall of the primary individual is thus present in the larva in an invaginated condition, so as to afford the corona, as the locomotory organ of the larva, more room for development. While most of tlie parts of the primary zooecuim are thus waiting in tlie larva ill an invaginated and apparently functionless condition, it is a striking fact that the retractile disc, which has hitherto been regarded as the rudiment of the polypide, seems to be of functional importance to the larva (sensory organ ?). On this account, tiie observation of Pkoxjho recorded above (p. 24) seems significant, according to which tlie retractile disc does not pass directly over into the rudiment of the polypide, but after its invagination undergoes a degeneration similar to that of the other larval organs, while, at the place formerly occupied by it, a new two-layered rudiment develops, viz. , tliat of the polj'pide. The bending over of the corona and the formation of the vestibule, in the wall of which are included the coronal cells, the pyiiform organ, and a part of the sucker, there to undergo, degeneration, will appear less remarkable if we consider liow often larval parts which have become useless, instead of being tlirown out, sink into the interior of the individual undergoing metamorphosis, there, after degenerating, to be utilised as food-material. Under this aspect, these processes ajtpear comparable to the retrogression of tlie embiyonic mem- branes and the formation of the so-called dorsal organ in Insect embryos. TYPE OF THE VESICULARIAN LARVA. 31 4. Type of the Vesicularian Larva. In the larvae of this type an excessive lengthening of the coronal ■cells is found (Fig. 13). The whole larva in this way becomes greatly elongated; the aboral and oral areas are exceedingly circumscribed, and the ectodermal furrow (er), as in the previous types, is covered by the cells of the corona. Another characteristic of these larvae which was observed by Barrois in Serialaria lendigera (N"o. 9) and by Ostroumoff in Vesi- cular ia (ISTo. 26) is the small size of the retractile disc (is) and the un- usual depth of the mantle-cavity (vi) ; posteriorly, i.e., opposite to the ecto- dermal furrow, it is so deep as to extend beyond the middle of the body. Observers are not quite unani- mous as to the sucker. According to Barrois, only a functionless vestige of this organ (.';) is retained ; Ostrou- moff, on the other liand, found a well-developed though not very large sucker. Fig. 13. — Larva of Serialaria lendigera (after Barrois). rs, retractile disc ; s, vestige of sucker ; cc, anterior ecto- dermal furrow ; m, mantle-ca\-ity. In the same way, authorities differ a.s to the metamorphosis of tlie sucker, though ill other respects metamorphosis here takes a similar course to that described for Bugula. Ostkoumoff maintains tliat in this type also meta- morphosis commences with the evagiuation of the sucker, wliile Bakhois holds that fixation 'takes place by means of two lobes Uiat grow out from tlie lower end of the ectodermal furrow and which belong to the corona. For details of the remarkable manner in which the corona is reversed, its long cells being rotated and bent round, we refer the reader to Barrois (Nos. 8 and 9). The polypide is here said to arise not by invagination, but throitgli tlie separation from the retractile disc of a cell-plate which becomes invaginated later (Ostroumoff). 5. Type of the Cyclostomatous Larva. The metamorphosis of the marine Cyclostomatous Bryozoa has been investigated chiefly by INIetschnikoff (No. 21), Barrois (JSTos. 6 and 9), and Ostroumoff (No 25). The larvae of this type (Fig. 14, A) are distinguished by the presence of a large sucker- invagination (s), and by the vestigial condition of the retractile disc (?•), which is found as an inconspicuous cell-mass at the base of the 32 BRYOZOA ECTOPROCTA. large mantle-ca-\'ity (p). A further peculiarity is exhibited l)y the corona, Avhich here does not consist of long cells extending from the oral to the ahoral area, .but is composed of numerous small elements. The ectodermal furrow is hardly i^ercejDtible, but Barrois (No. 9) was able to identify it. This larva, in its metamorphosis, follows the usual course. Here also (Fig. 14, B) the first act of metamorphosis is the evagination of the sucker and its transformation into the basal plate (a) of the primary zooecium ; the bending over of the mantle then takes place, and the fusion of its periphery with the edges of the basal plate, by means of which the annular vestibular space (v) is closed in. While V Fig. 14. — LongitudiDal sections of two ontogenetic stages of a Cyelostoniatous Bryozoon (after Ostroumoff). a, longitudinal section through the larva of Frondipora. B, metamorphosis of Tubulipora serpens, a, adhesive plate ; /, central tissue ; p, mantle-cavitj' ; r, vestige of the retractile disc ; s, sucker ; v, vestibulum ; x, cell-p'ate (rudiment of the polypiile derived from the retractile disc). subsequently this whole structure degenerates, the polypide forms, not by invagination, but by the separation of a cell-plate (x, Ostroumoff). This plate sinks in and becomes covered with mesoderm derived from the central tissue. Ostroumoff's observation of the appearance of a cavity lined with endothelium in the neigh- bourhood of the developing nutritive tube is worthy of mention. In this cavity, Avhich degenerates later, we have perhaps a homologue of the body-cavity of the fresh-water Bryozoa. 6, Type of the Phylactolaematous Larva. The embryonic development and the metamorphosis of the Phylac- tolaemata has been investigated by Metschnikoff (No. 20), Nitsche (No. 52), Rbinhard (Nos, 54-56), Ostroumoff (No. 53), Korotneff TYPE OF THE PHYLACTOLAEMATOUS LARVA. 33 ec (No. 48), and Jullien (No. 47a), and more recently by Davenport (No. 46a), Kraepelin (No. 50), and Braem (No. 45a).* The ovary (Fig. 15 o?;) is an aciniform growth of the inner meso- dermal layer of the body-wall, lying on the oral side of the polypide, immediately beloAV the so-called duplicating bands (posterior parieto- vaginal muscles of Allman) and above the last daughter-bud. Single mesodermal cells increase in size, and become changed into young egg-cells, while other mesoderm-cells become grouped round them as the follicular epithelium. (_)ne embryo only develops in each individual (Eeinhard, Kraepelin). Fertilisation and the commencement of cleavage take place within the egg follicle. Later, however, the young embryo passes into a peculiar brood-sac {ooeeiurn, Fig. 15 X, 16 o), which functions as a uterus, and in which the further embryonic development takes place. The ooecium, when it first appears (Fig. 15 ./■), closely resembles a young polyp-bud, being a pocket-like invagination of the bilaminar body-wall. The outer or meso- dermal layer of this bilaminar invagination is destined to enveloj^ the embryo, while the inner ectodermal layer undergoes degeneration in the course of further development. It was formerly somewhat doubtful whether the ooecium was altogether distinct from the primary egg-follicle. According to Reinhard, it was derived directly from the epithelium of the ovary. Whereas, according to Metschnikoff, the egg passed out of the original follicle into the body-cavity, where the ooecium grew round it secondarily, according to Kraepelin and Braem, the rudiment of the ooecium presses closely on the egg while still in the ovary, and receives it from the egg-follicle. According to the statements of most authors, cleavage is quite regular. Kraepelin, however, thought that he found certain differences in the blastomeres in the first stages. A cleavage-cavity early appears within the embryo, and a coeloblastula (Fig. 16, A, e) develops, which, by a solid ingrowth of cells at one pole (Fig. 16, B), passes into tlie gastrula stage. The point of ingrowth, according to av Fig. 15.— Section through a portion of the zooecial wall of Plumatdlafungosa (after Braem). ec, ectoderm ; m, me.soderm ; ov, ovary ; x, rudiment of the ooecium. * [The form in which the dc-velopiuent of tlie larva and of the polypide has been best worked out is Plumatella (Alcyonclla) fimgosa (= P. ijolymorphuy Kraepelin). — Ed.] 34 BRYOZOA ECTOPROCTA. ^^I^Co O -M £ the statements of Davenport, Braem, and Kraepelin, invariably corresponds to that pole of the embryo which is turned towards the wall of the zooecium. At the same point, at a later time, the first rudiment of the polypide is said to develop, so that Ave have here an important contrast to the Gymnolaemata, in which, so far as we know, the primary rudiment of the polypide always develops at the aboral pole of the larva. While the ingrowing cell-mass more and more fills the blastocoele, a new cavity develops with- in the former; analogy with other animals might incline us to consider this cavity as the primary enteric cavity, but we are led by the study of the further course of develop- ment to regard it as the body-cavity. The ingrow- ing cell-mass represents the mesoderm, and the cavity that develops within it the coelom. We thus find here a gastrulation in which the entoderm proper is apparently want- As the coelomic cavity increases in size, the mesodermal epithelial layer becomes closely pressed against the ectoderm, and in this way a bilaminar, vesicular embryo develops (Fig. 16, C). We shall only be able to interpret this stage correctly if we compare it witli the stage that follows the gastriila in most marine Ectoprocta (Fig. 6, G, p). The primary entoderm tliere yields the so-called central tissue which represents the enteric rudiment of the larva, out of which, however, important mesodermal organs of the primary zooecium are also derived. In the embryo of the Phylactolaemata this tissue is represented by the inner epithelial layer. We must here assume that the larval intestine has undergone excessive reduction, so that not a vestige of it can be seen, and, taking into consideration the further development, we must regard this inner layer as mesoderm, and the central cavity enclosed by it as the coelom. Fig. 16. — Three stages in the embryonic development of PlumaicUa (after Kkaepelin). A, blastnla stage in the ooecium. B, stage of ingrowth of the meso- derm. C, bilaminar stage, e, embryo in the blastula stage ; ec, invaginated ectodermal part of the ooecium (c/. Fig. 15) ; 0, ooecium. mg TYPE OF THE PHYLACTOLAEMATOUS LARVA. 35 K We shall have to call the external cell-layer of the embryo the ectoderm, but it must not be forgotten that this layer contains indifferent embryonic material from which will be produced the future rudiment of the polypide, including the «ntodermal organs. The next ontogenetic phenomenon is the development of the rudiment of the first polypide. The larvae of some fresh -water Bryozoa {Plumatella) at the free - swimming stage contain two well - developed polypides (Fig. 19 A), one of Avhich, however, always seems to develop as the older primary polypide, and the other as a precocious daughter bud. The larva of Plumatella fruticosa, on the contrary, contains, according to Allman, only one primary polypide, and in Cristatella also, the second polypide does not seem to ■develop so early as in the above- named form. The polypides (Fig. 17, p) arise as simple invagi- nations of the bilaminar wall of the embryo. Their further development will be described more in detail below (p. 37). The first indi- cations of the polypide - invagi- nation are found In the form of Q, simple thicken- ing of the body- wall. Daven- port states that, in Plumatella^ the rudiment of the second polypide appears from the first independent of the primary polypide. The conditions in Cristatella are, according to Davenport, somewhat different. We shall see below (p. 47) that in this form every new bud is intimately connected at its commencement with an older bud, and that the Fig. 17.— Two later ontogenetic stages of the embryo ot Pluviatella (after Kraepelin). ec, ectoderm of the embryo ; /, mantle-fold ; 1)1, mesodermal layer ; o, ooecium ; p, polypide-rudinient ; pi, placenta. 36 BRYOZOA ECTOPROCTA, inner budding zone of the polypide is derived from a layer of embryonic tissue capable of gi'owth, lying below the ectodermal epithelium. The ectodermal layer of the polypide -rudiment in the embryo also is overgrown by the surrounding ectoderm. The ectoderm at this part is therefore bilaniinar. The inner layer represents the tissue capable of producing buds, from •which th& inner layer of the polypide-rudiment is derived. From this tissue is developed the inner laj'er of the first polypide-rudiment, and soon after, by its side, the invagination of the second polypide, which thus, in Cristatella, is directly connected with the primary polypide. While the polypicle-rudiments are developing, we find, in Pluma- fella, a zone of ectodermal cells encircling the middle of the embryo and fusing with the wall of the ooecium (Fig. 17, pi). This is the girdle-shaped placenta already recognised by Korotneff, by means- of which the embryo is suspended in the uterus-sac and nourished. In Cristatella this is Avanting ; nourishment here takes place through the ectodermal cell-jDlug which closes the neck of the ooecium, and which is in close juxtaposition to the anterior pole of the embryo. In the later stages a circular fold of the body -wall develops (Fig. 17 B, f), Avhich surrounds the anterior half of the body, and may be regarded as the equivalent of the mantle-fold of other Bryozoan larvae. An ectodermal thickening, including the posterior pole (Fig. 18 A, x), was regarded by Ostroumoff as a vestigial suclier, while the anterior swollen part of the body in which the polypide-rudiments appear may be considered as the equivalent of the retractile disc. Although, in this waj^, the larvae of the Phylactolaemata may, without undue forcing, be compared with those of the marine forms, a difficulty arises from tlie above-mentioned circumstance that the primitive gastrula- ingrowth here occurs at the anterior pole, which has been regarded as tlie equivalent of the retractile disc. We are unable at present to overcome this obstacle in the way of comparing the larvae of the Phylactolaemata with those of the Gymnolaemata. The embryo, after developing in the manner described above,, becomes covered externally with cilia, and, escajiing from the parent, swims about freely. According to Kraepelin, it emerges through the aperture of a degenerated polypide (usually that of the parent). Braem, on the contrary, holds that the cavity of the ooecium opens directly upon the exterior in order to allow of the passage of the embryo. At this stage the larva is oval (Figs. 18 ^ and 19 A), and the whole surface of its body is ciliated. At the anterior pole there is an aperture leading into the large mantle-cavity, from the base of which the two polypide-rudiments project. The attachment of the larva takes place at first by means of DEVELOPMENT OF THE POLYPIDE. 37 modified glandular ectodermal cells at the posterior pole (Fig. 18 ^, x) — sucker-rudiment of Ostroumoff. The projection -which carries the polypides is evaginated from the mantle -cavity, the mantle-fold (/") at the same time bending back (Figs. 18 i? and 19 B). The posterior pole of the body now becomes detached from the substratum, the edges of the mantle-fold (/) approach each other till they fuse (Fig. 18 C), and in this way a closed sac forms in which is included the greater part of the external ciliated surface of the larva, which then undergoes degeneration. The essential distinction between the metamorphosis here described and that of the marine forms consists in the absence of the sucker. In consequence of this a basal adhesive plate does not develop. The whole wall of the young colony is derived exclusively from that part of the body-wall which, in the larva, lined the mantle-cavity. The condition of the mantle-fold consequently here resembles that in the marine forms. -/ ■A yf— Fig. is. — Three consecutive stages in the fixation of the larva of Plumatdla (after Beaem). /, mantle-fold (in Ji and C reflected); k, buds; p, the two first developed polyjndes ; ,«, glandular part at the posterior end of the body, by means of which the first attachment takes place. / IV. Development of the Polypide. We must now describe more in detail the development of the polypide, i.e., of the retractile cephalic section of the animal plus the intestinal canal appended to it. For our knowledge of the metamorphosis of the larva and the development of the polypide in the primary zooecium, we are indebted chiefly to the observations of Repiachoff (Xo. 29) on Tendra, of Barrois (No. 9) on Lepralia, and of Prouho (Xo. 28) on Fhistrella. The development of this primary polypide takes place in exactly the same manner as that of the polypides in the buds, which develop later in the colony, or in those zooecia in which the alimentary canal and polypide, as 38 BRYOZOA ECTOPROCTA, already mentioned (p. 15), have undergone degeneration, the polypide having therefore to be regenerated. In the forms mentioned above the regeneration of the polypide has repeatedly been investigated; notably by Xitsche (jS^os. 23 and 52), Repiachoff (No. 30), Joliet (jSTo. 17), Haddon (K'o. 12), Ostroumoff (No. 26), and more recently by Seeliger (No. 37a), Davenport (ISTos. 11 and 46a), Braem (Xo. 45a), Oka (K"o. 52a), and Kraepelin (No. 50).* The way in which the polypide develops in the statoblasts has been specially described by Braem (No. 45a) and Oka (No. 52a). n Fia. 19. — Two ontogenetic stages of Phimatdla fungosn (after Nitsche). A, an advanced embryo. Jl, larva with two polypides in tlie act of attacliing itself; the mantle-fold (f) is already reflected, the apertures of evagination have moved far apart and the polypides are retracted. i>, the two first formed polypides ; /;, k', buds of future polypides ; /, mantle-fold. We have already seen the rudiment of the polypide arising in the primary zooecium of Bugula (Fig. 12, B, p. 29) in the form of a double-walled sac, the inner cell-layer of which is said^ to be derived through invagination from the retractile disc (see, however, the statements of Prouho, p. 24), while the outer layer has probably arisen from the cells of the central tissue. In the same way the first rudiment of the polypides in the buds (/r, /', Fig. 19), in the regenerated individuals, and in the statoblasts are bilaminar sacs, which arise through invagination at one point of the bilaminar body-wall (endocyst). In the marine Bryozoa and in Plumatella, such an invagination has, from the first, a lumen which passes direct * Cf. Haemeu, Quart, Joiirn. Micro. ScL, Vol. xxxiii., p. 123. DEVELOPMENT OF THE POLYriDE. 39 over into the lumen of the polypitle-sac. In other cases, e.g., in Paludicella (Davenport), we find at first a solid ingrowth within which a lumen forms secondarily. The bilaminar sac always remains connected by a longer or shorter strand with that part of the body- wall at Avhich the invagination occurred. This strand (only a portion of which is represented in Fig. 20, J. and B), is derived from the neck of the primitive polypide invagination, becomes hollow again at a later stage, and then represents the aperture of the tentacular sheath (Fig. 20, C, is). jr B Fig. 20. — Diagram illustrating the development of the polypide (median sections, after figures by J. Barrois and Nitsche). A, sac-like rudiraeLt of the polypide. The separation of the mid-gut rudiment (d) is marked by a constriction. B, development of the tentacle-buds (() and the oesophageal invagination (oe). C, older ontogenetic stage in connection with the endocyst (c, e). a, external, i, internal layer of the polypide-rudiment ; d, mid-gut rudiment ; c, e, endocyst ; m, mouth ; n, in- vagination for the formation of the ganglion ; oc, oesophagus ; r, retractor muscle ; t, tentacle ; ts, ten- tacular sheath ; a (in C), anus. The outer layer of the polypide-sac (Fig. 20, A, a) is continuoi;s with the mesodermal layer of the body-wall (endocyst), and yields the mesodermal part of the polypide (the lining of the body-cavity, and the larger groups of muscles, &c.). The inner layer (Fig. 20, A, i) is derived from the primitive ectoderm, and yields the ectodermal epithelium of the whole polypide and of the tentacle sheath, the nervous system, and the lining epithelium of the whole of the digestive tract. We should be justified in calling this layer the ecto- derm of the polypide, did not the mid-gut epithelium also originate from it.* * [Braem, in a recent work on the development of Plumatclla (Bibl. ZooL, H. 23, 1897), states that entodennal tissue can be made out in the embryo. — Ed.] 40 BRYOZOA ECTOPROCTA. According to the more recent researches of Seeliger, Davenport, Kkaepelin, Braem, and Oka on the development of the buds, there can be no doubt that the inner layer of the bilaminar sac is to be derived by invagination from the ectodermal layer of the parent, the outer layer, on the contrary, owing its origin to the mesoderm of the latter. Some of the earlier observers had derived the whole bud from the ectoderm of the parent (Nitsche, Ehlers, and Clapar1<;de), while Haddon held that all the three germ-layers of the parent take part in the formation of the bud. According to Joliet (No. 17), on the contrary, the first rudiment of the polypide iu Eucratea is derived exclusively from the funicular tissue (endosarc), i.e., from a layer which we regard as belonging to the mesoderm. The first rudiment of the polypide is said here to consist of a mass of similar cells, which become arranged only secondarily into two layers, the inner cells acquiring an epithelial character and surrounding a central cavity, while the superficial cells yield the outer layer. The first change which can be remarked in the bilaminar polypide- sac consists in the formation of a diverticuhim (Fig. 20 A, d ; 23 B, d), which represents the first rudiment of the enteric canal, especially that of the stomach and hind-gut. The origin of this diverticulum can be traced to a constriction or infolding of the wall of the primitive sac taking place from each side. In conse- quence of this, the lumen of the enteric rudiment at first remains connected along its whole length through a narrow slit Avith the lumen of the rest of the sac. As the intestine becomes more differentiated this slit becomes narrower, but never completely closed, for it persists as tlie anal aperture (Fig. 20 C, a). In the following stages a second diverticulum develops at the opposite (oral) side of the polypide-sac (Fig. 20 B, oe) ; this diverticulum, which represents the rudiment of the oesophagus, grows out towards the blind end of the first diverticulum and fuses with it (Fig. 20 C). After communication has been established at the point where the two diverticula come into contact, the intestinal canal is essentially complete, the so-called stomach-caecum developing at a later period. The description here given of the development of the enteric canal is founded upon the statements of Barrois (No. 9), Prouho (No. 28), Braem (No. 45a), Davenport (Nos. 11 and 46a), and Kraepelin (No. 50). A modification of this process was described by Nitsche (No. 23) in Fluviatella and Flustra ■niembranacea, and more recently by Ostroumoff (No. 26), the formation of the intestine taking place through two ingrowths which extend into the interior of the sac-like jjol^'pide-rudiment, from right and left, meeting and fusing in the median plane. The apertures left in front of and behind this fusion become the oral and anal ajicrtures. " Tliis process resembles what takes place if we hold an indiarubber ball with a double wall in our two hands and ])ress it on each side with a finger until the finger-tips are separated only by the four-fold wall of the ball." The oesophagus in this case would not have an independent origin. Seeliger observed a similar origin for the alimentary canal in the buds DEVELOPMENT OF THE POLYPIDE. 41 of Bugula. It should here be pointed out that the two different types of development of the intestine are not fundamentally opposed to one another ; in both cases a constriction pressing in from each side brings about the separation of the enteric canal from the atrium. AVhether the rudiment of the oesophagus is completed at the same time is of comparatively little importance, and we jnay well imagine, even in nearly related forms, that in some cases the first and, in others, the second type of bud- development may be found. In Pectinclla, on the contrary (Oka, No. 52a), in the statoblasts as well as in the buds, the diverticulum that forms first is said to represent the common rudiment of the oesophagus and the stomach, so tlrat here the oral aperture seems to form first. The intestine is said here to grow out as a caecum from the -stomach and to open towards the atrium (the upper cavity), the anal aperture seeming to form in this way. The ganglion can be traced back to an invagination which forms at the base of the upper cavity or atrium between tlie oral and the anal apertures (Fig. 20 C, n). The lumen of this invagination gives rise to the future brain-cavity. After the rudiment of the ganglion has been completely cut off from the surface layer, an ear-shaped diverti- culum forms on each side of it (Braem) ; these diverticula represent the rudiments of the lophophoral nerves. Each grows out as two nerves, one of which runs posteriorly into the corresponding arm of the lophophore, while the other runs forward and spreads out upon the oesophagus. The portion of the primary sac which remains after the intestine has been separated from it is known as the atrium (Fig. 23 B, at) or the cavity of the tentacular sheath. The greater part of its wall becomes modified into the tentacular sheath. The rudiments of the lophojDhore and of the tentacles (Fig. 23 B, at) develop early in the base of this cavity. The first rudiment of the lojjhoj^hore takes the form of a swelling projecting into the atrium, and forming a semicircular border to the oral aperture. In the Gymnolaemata, this swelling closes in front of the anal aperture to form a ring, encircling the mouth, which carries the tentacles. Corresponding to the inwardly projecting swelling of the lophophore, there is, on the outer side of the polypide-sac, a groove which soon changes into a closed canal. This is the so-called circular canal, which must be- regarded as part of the body-cavity. The tentacles arise as out- growths of the lophophore resembling the fingers of a glove. In the Gymnolaemata they are situated in a row on either side of the body (Davenport, Prouho). Only later do these two rows become connected through the development of the tentacle-buds in front of the mouth, while the last tentacles to arise close the ring on the anal side. 42 BRYOZOA BCTOPROCTA. In the Phylactolaemata the posterior ends of the semicircular rudiment of the lophophore grow out as large finger-like structures- (Fig. 23, C, I), projecting into the atrium. Tliese are the rudiments of the two arms of the lophophore. The cavities witliin these processes (lophophore-cavities) are to be regarded as a part of the body-cavity. They communicate with one another through the semicircular canal that encircles the oesophagus. On the anal side, according to Braem, they are connected by the so-called forked canal, so that here also the system of lophophore-cavities forms a ring surrounding the oesophagus. Another significance has been attributed to this forked canal by Cori (p. 56). In the Phjdacto- laemata the outer tentacles make their appearance before those on the inner side of the lophophore. The former appear first in the neighbourhood of the mouth and gradually fresh tentacles are added, each new one being nearer the free end of the lophophore, those at the apex forming last. The tentacles of the inner row develop in the reverse order, i.e., those nearest the apex arise first, whilst those near the epistome and above the forked canal are the last to appear. The epistome arises as a projecting fold on the anal side of the oral aperture, and into it is continued an outgrowth of the body- cavity — the epistomal cavity. It should be mentioned that several authors (Sbeliger and others) maintain that a rudimentary epistome is to be found in the Gymnolaemata. The outer or mesodermal layer of the polypide-rudiment gives rise to the following parts : the peritoneal covering of the enteric canal, the muscle-layer of the intestine, and some of the body-muscles, especially the retractors. The development of the latter has been studied chiefly by Braem and Davenport. Groups of mesoderm- cells become detached at the neck of the bud, and become inserted at one end upon the polypide and at the other on the zooecial wall. These rudiments of the retractor muscles are originally inserted at a point on tlie zooecial wall quite near the neck of the bud, and only at a later stage, as the Avail develops, do the points of insertion shift further from the aperture of the polypide- rudiment. In the Phylactolaemata, the outer mesoderm-layer of the polypide- bud also takes an important part in the development of the funiculus, Avhich will be dealt with later (p. 50). The nephridia of the Phylac- tolaemata also, the existence of which has been maintained by Verworx and Cori, and recently confirmed by Blochmann, although DEVELOPMENT OF THE POLYPIDE. 43 denied by Braem and Krabpelin, are probably to be traced back to this layer.* It is thus evident that the inner layer of the bilaminar polypide- rudiment yields the ectodermal epithelial layer of the jjolypide, and also the epithelial lining of the enteric canal and those parts Avhich we have been accustomed to attribute to tlie entoderm. From the outer layer of the polypide-rudiment are derived the mesodermal structures (the splanchnic layer of the mesoderm, the retractor- muscles, the lining of the tentacle-cavities, etc.). When the rudiments of the organs described above have fully developed, the atrial cavity of the polypide-rudiment becomes connected witli the exterior, and through the aperture so formed the anterior portion of the polypide with its crown of tentacles, the introvert, can be protruded and extended. It must strike the reader as very remarkable that, according to the above statements, tlie whole lining epithelium of the alimentary canal (both the part usually derived from the ectoderm and the entodermal part) is derived from one and the same rudiment, the inner layer of the sac-like polypide-rudiment. In the primary zooecium of Bugula it was possible to trace this layer back to an invagination of the ectoderm of the larva ; and in the same way, in the polypide-rudiment of the buds and of the regenerating individuals it may be traced back to the ectoderm of the zooecia. If these observations are correct, we should be forced to assume that the whole enteric canal here originates from the ectoderm. Several attempts have in consequence been made to find some other origin for the middle (entodermal) part of the alimentary canal. A suggestion of a distinct origin for the enteron is yielded by the constant connection discovered by Repiachoff (No. 30), at later stages, between the enteric rudiment of the polypide and the so-called hroivn body. In the primary zooecium, the brown body contains the mass which has arisen by the degeneration of the larval organs and of the central tissue, while the brown bodies found in the degen- erated zooecia of the adult colony enclosed in cellular envelopes of their own nmst be regarded as remains of the degenerated parent-polypide. The rudiment of the intestine, whicli in the bud of the newly-forming polypide is originally connected with the brown body through strands of funicular tissue, at a later stage comes into contact with it, and is said finally to grow round it and to receive it into the interior of the enteric cavity. The last remains of the brown body are then said to be expelled through the anal aperture of the newly- formed polypide. During this circumcrescence, according to Ostroumoff (No. 26), the epithelial layer of the stomach-caecum is yielded by the cells * [Oka, Jour. Sci. Coll. Japan, Vol. viii. , p. 339, states that the Ecto- proctous Polyzoa have no nephridia. He does not regard the ciliated portions of the coelomic ei^ithelium, which ajjparently open by a pore under the median tentacle, as nephridia. He states that the excretory function is carried on by free mesodermal cells, which leave the body through the above-mentioned pore, their passage to the exterior being facilitated by the presence of cilia on the cells of the j)eritoneal epithelium adjacent to the pore. The tube-like character of these modified peritoneal areas is due to the presence in the Phylactolaemata of an epistome-lophophoral partition. — Ed.] 44 BRYOZOA ECTOPROCTA. of the brown body. A certain difficulty attends this assumption, in consequence of the development of the intestine in the young bails, in ■which there is no brown body. Ostkoumoff, however, tries to escape these difficulties by pointing out the relations established by the funicular tissue between the parent-zooecium and the bud, as well as by assuming that in this way ento- dermal cell-masses pass from the parent into the bud. In this respect the views of Ostkoumoff have something in common with those of H.\1)D0N, who held that all the three germ-layers of the parent-zooecium took part in the formation of the bud. According to Joliet (No. 17), the alimentary canal of the developing polypide does not originate from the inner layer of tlic bilaminar sac-like polj'pide-rudiment, but from a distinct cell-mass derived from the outer layer of that rudiment. The inner layer of the polypide-rudiment would then yield only the ectodermal parts of the polyjjide, while the outer layer would contain the mesodermal and entodermal parts. In any case, according to Joliet, the enteric canal (mid-gut) has an origin distinct from that of the ectodermal rudiment. The most recent researches, how^ever, confirm the view that the whole intestine of the polypide originates from the inner layer of the double-walled sac, i.e., from the ectoderm, but Prouho differs to some extent from the de.scription given above and founded on the statements of more recent authors, for he regards a small mass of irregularly-arranged cells, lying at the end of the diverticulum d in Fig. 20, as the rudiment of the mid-gut, while the hind-gut arises exclusively from the diverticulum itself. According to Prouho, indeed, this cell-mass is to be derived from this same diverticulum ; but it cannot be denied that this statement is likely to strengthen the doubt that lia.s long been felt as to the common origin of the fore-, mid-, and hind-guts. V. Asexual Keproduction of the Ectoprocta. A. Budding. In the Biyozoa, the colonies are produced by the continuous budding of the primary individual. The nature of this budding has been carefully studied by Nitsche (Nos. 23 and 52), and more recently by Braem (Xo. 45a) and Davenport (Nos. 11 and 46a).''' Two kinds of buds may usually be distinguislied, according to the direction of their growth: — (1) those that continue to grow in the same direction as tlie parent-zooecium, sucli buds serving for the direct prolongation of the branch or branchlet to which the parent- zooid belongs ; and (2) those that, in growing out from the parent-zooecium, take a new direction and thus give rise to new branches. In many cases the new branches grow out laterally. The buds of the second kind are then lateral huds, while the median buds provide for the continuation of the branches, such individuals usually continuing in the same axial direction as the parent. It should, however, be mentioned, that in many cases median buds may also give rise to new branches, since, while retaining the same plane as the mother, they may take a new direction. * [For the develojimcnt of the colony in the Gymiiolaemata, see important papers by Hakmeu on Lichcnojwra, Crisia, and Tuhidipora. Quart. Journ. Micro, ,'i'ci., Vols, xxix., xxxii., and xli. — Ed.] BUDDING. 45 The development of the buds and the branching of the stock thereby determined takes place in the different forms of Bryozoa according to definite laws which were long since reduced to certain written formulae. For the laws of growth of tlie colonics, which cannot here be entered ujion in detail, we refer the reader chiefly to Davenport's treatise (Xo. 11). Braem, for Phimatella, has composed a formula which also applies to the other Phylactolaemata. JJi C B C^ C D a C' C J) E a In this, the jiarent and offspring are always connected by a bracket | | thus. This formula will be exjilained by comparing it with Fig. 21. The primary individual A has given rise by budding to the individuals B, B^, B"- ; B, on its side, has produced O and 6'\ and so on. The individuals A, B, C, D, i.e., the so-called princii)al buds resulting from the first division, serve for the continuation of the principal stem in a centrifugal direction, while the inter- mediate buds B^, B", B'\ etc., become intercalated between the oldest bud and the mother A, and lead to secondary branching of the stem. The order of succession in the two cases is reversed. Of the principal buds, the distal indi- vidual whicli denotes the tip of the branch (in this case G) is the youngest, while of the intermediate buds B^, B^, B^, the last-formed individual {B^) lies nearest to the mother-individual A. The number of buds that each indi- vidual is capable of j^roducing is often limited. In Oristafdla, J'ov instance, only two buds are, as a rule, produced, the elder being a lateral bud and the younger a median bud. In Bahcdicella, on the contrary, each individual is able to produce a median l)ud and two lateral buds, and so on. The variations in the appearance of the Bryozoan stocks are caiised by the constitution of the zooecial wall, which through stronger chitinisation or impregnation with lime salts may become stifTened, or, on the other hand, may be soft {Cristatella) or even gelatinous {Alcyonidium, Flustrella), and by the more or less close juxtaposition of the single branches. When the latter retain their independence, moss-like colonies with serrated brandies are formed. When the separate branches lie so close to one another in the same plane that the neighbouring branches fuse together, leaf-like, fan-shaped, or encrusting colonies are produced, while close crowding of the branches in various planes leads to the development of fungoid forms {PlumateUa fungosa). The fusion of the separate zooecia has gone furthest in Cristatella, in which the originally distinct character of the branches is indicated only by the mesodermal septa growing in from the edge of the colony. Fio. 21.— A, branch of Plumaidla frvticosa (after Braem). B, diagram illustrating tlie brandling of Ihis form. 46 BRYOZOA ECTOPROCTA. The youngest buds lie, as a rule, at the apex of the branch. In encrusting' colonies, the edge of the colony thus represents the budding zone, from which the further growth of the colony proceeds. In the same way, in Cristatella , the youngest individuals, those in the act of forming, are at the edge of the colony (Fig. 22 kz), while the oldest [dp) lie at its centre. BiiAEM has pointed out an important distinction between the relations of the bud to the jiarent in the Phylactolaemata ami in the Gymnolaemata. The colony in the Phylactolaemata arises in such a way that the oral side of each individual is directed towards the distal end of the branch to which it belongs. The younger individuals thei'efore bud out on the oral side of the older (Fig. 22). In the Gymnolaemata, on the contrary, the individuals are placed in the opposite way, each new bud arising on the anal side of the parent {ef. Fig. 24). A similar distinction between the two groups is found in the position of the separate individuals witli resjiect to the substratum. In the Phylactolaemata the individuals, in a withdrawn condition, turn the oral side to the substratum, while in the Gymnolaemata the anal side is thus turned. The attempts made by BiiAEM and Davexpout to trace back these different relative positions of the bud and the mother in the Gymnolaemata and the Phylactolaemata to a common type do not seem to us altogether successful. BiiAEM presupposes, in the Gymnolaemata, a degenerate primary individual which cor- responds to the distal end (apex) of the stem. Through this assumption he makes the prin- cipal buds of the Gymnolaemata agree with the intermediate buds of the Phylactolaemata. According to Davenport, on the otlier hand, in both groups, each bud turns its anal side to the budding zone from which it has originated. In the Gymnolaemata, this zone lies at the apex of the stem, i.e., distally, but in the Phylactolaemata it lies proximally. To us it appears that in this very point of the position of the budding zone there is no essential distinction between the Gymnolaemata and the Phylactolaemata, since, in both cases, the youngest buds appear distally, that is, at the edge of the colony. Consequently the budding zone must, in both cases, have the same position. It has already been pointed out by Nitsche {^o. 52) that two distinct types of budding occur in the Ectoprocta, the Gymnolaemata being in this way opposed to the Phylactolaemata. In the first of these groups the type of budding in which the zooecium forms first is common. In this case the zooecium of the bud arises first as an outgrowth or diverticulum of the parent-zooecium (Fig. 24), and only after this has attained a certain size and independence does the first rudiment of the polypide (^j) appear in it as a bilaminar invagination of the zooecial wall. In the Phylactolaemata, on the contrary, the polypide of the bud appears first, in close proximity to Jix Fio. 22. — Transverse section through a colony of Cristatella (after Bra em), dp, eldest polypide of the colony, in the act of degenerating ; kz, budding zone. BUDDING. 47 the parent-polyiiide, as tlie direct descendant of the latter (Fig. 23). Only later does it shift further from the parent-polypide, newly- formed parts of the zooecial wall being interposed between them. This latter is the type of budding in Avhicli the polypide forms hrst. The zooeciuni of the bud is not always a direct outgrowth of the jiarent- zooecium, for budding may be brought about by special basal extensions (stolons) as, for instance, in the Cyclostoniatous Bryozoa (Ostroumoff, No. 25), and in some of the Ctenostomata (the group of the Stoionifera). Fig. 23.— a and B, two ontogenetic stages of the lateral buds of Crista'ella. C, development of the median buds of the same form (after Braem). a, developing polypide ; b, the bud developing on it ; at, atrium ; d, rudiment of the intestine ; ec, ectoderm ; I, rudiment of the lophophore. It appears that in all Bryozoa budding takes place only at definite jiarts of the parent in which the original capacity for regeneration has been retained. In the Phylactolaemata, in which asexual reproduction according to the type in which the polyjiide precedes the zooecium is retained, the rise of a new polypide - rudiment is, as first shown by Hatschek, and more recently by Braem and Davenport, always connected with an already developing polj'pide-rudiment (Fig. 23). While the parent-polypide, which was originally a bilaminar invagination of the body-wall, develops in the way described above (Fig. 23 A, a), tliei-e is often found at the neck of this invagination, on its oral side, an outgrowth ; this is the rudiment of the daugliter-bud {0). As the two rudiments 48 BRYOZOA ECTOPROCTA. develop further, they become more distinctly marked off from one another (Fig. 23 B), and finally shift quite apart, a portion of the tissue of the neck of the hud being used for the development of the adjacent parts of the zooecial wall (Braem). Xitsche had already observed this form of origin of one polypide-rudiment from another, and defined sucli forms as double buds. In Pliimatella, as veil as in Cristatdla, the first bud to develop in each polypide {B, C in the diagram, Fig. 21) forms after the type of the double bud. The buds that develop later {B\ B^, C'\ etc., in the diagram) form after another type which, however, is not essentially different. The rudiment of the bud here arises (Fig. 23 C, h) in the zooecial wall itself on the oral side of the parent ))olypide-rudiment ; the young bud-rudiment is, however, from the first, in direct connection with the germinating tissue of the parent-rudiment, so that here also Ave can recognise the connection of each new rudiment with an older one. "We can see here very clearly that budding is to be traced back to a process of division. Each newly arising individual becomes cut off from an older individual already present, so that finally all the individuals of a colony can be derived from the first individual produced from the larva. a Fio. 24.— Biuldiiii; in Palvdicella Ehreniergii (after Davenport). Median section through the fjrowing apex of a branch, a, b, region of the branch belonging to most distal of the polypides as yet developed ; h, c, region belonging to the polypide-bud (p) that is beginning to develop ; c, d, region of the growing tissue at the tip of the branch ; d, growing tissue ; ed, hind-gut of the polypide ; g, ganglion ; p, young polypide-bud. Authors still differ as to the development of the zooecial wall in the Phylacto- laemata. Brap:m is inclined to derive it exclusively from the germ-tissue of the neck of the bud, but Davenport maintains that the zooecial wall grows independently in Cristatella, at least at the marginal parts of the colony. Another method of budding, in which the rudiment of the polypide has from the first a certain independence, forms the transition to the type of budding in which the zooecium develops first, a type common in the Gynino- laemata. Fig. 24 illustrates the rise of new individuals at the apex of a branch of raludicella (Davenport). The apex of the branch is here occupied by actively growing tissue {d), which gives rise.first of all to the zooecium of the new individual. While the wall of this new zooecium undergoes the general histological transformation by means of which it attains the special character of the adult form, the tissue, at one definite point, retains its embryonic character BUDDING. 49 and its capacity for regenerating, and at this point the polypide-rudiment arises (p). Two islands of germinating tissue which persist at the two sides of this rudiment rejiresent tliose parts of tlie zooecial wall which, later, form the stai'ting point of the lateral buds. It Avas pointed out by ISTitsche (No. 23), and more recently by Pergens (No. 27), that the oldest individuals of the Bryozoan colony, and above all the primary zooecium derived from the larva, may, in many cases, be distinguished from the other normal zooecia of the colony by their shape and size as Avell as by their method of budding. For example, ISTitsche found that the primary zooecium in Flustra memhranacea is remarkably similar to that of Merabranipora, and with respect to budding agrees with Memhranipora j^i^osa as described by Schneider (N'o. 5), i.e., a number of buds here appear, whereas the secondary zooecia of Flustra memhranacea, as a rule, produce only one bud each at the distal end. Pergens also found that the primary zooecium of Microporella passed through a Mem- braniporan stage, Avhile the buds produced from it grew into zooecia of normal shape. Heteromorphous development of single individuals often occurs in the Ectoprocta. Thiis in the polymorphous stock there may be found, besides the usual individuals, root-processes, and specialised organs known as ooecia, oviceUs, avicularia, and vihracula which are regarded as polypides modified in adaptation to a special function. B. The development of Statoblasts. A special kind of asexual midtiplication is Ijrought about in the Phylactolaemata by the production of peculiar reproductive bodies, the so-called datoUasts. This form of development may be traced back to budding. The recent researches of Krabpelin and Braem prove that the statoblasts are undoubtedly encysted persistent buds, intended to secure the continuance and distribution of the fresh- water Bryozoa during the winter months. The lenticular statoljlasts or winter eggs (Fig. 27 B) consist of a mass of cells, the germ-body, enclosed in a thick cuticular envelope (c), the latter is usually provided Avith a ring of air-cells which serve as a float (sr). In the germ-body Ave can distinguish a superficial epithelium, discovered by Reinhard (Fig. 27 A, ec), and a granular cellular mass rich in yolk (d). The structure of the germ-body can therefore be compared to that of the Ectoproctous embryo described above (p. 15), if Ave regard the superficial epithelium as the ectoderm and the granular inner mass as the equivalent of the central tissue. e 50 BRYOZOA ECTOPROCTA. The statoblasts arise in connection with a mesodermal strand known as the funiculus, which runs from the end of the stomach- caecum to the zooecial wall (Fig. 28, /). Along this strand they are arranged in a chain, the youngest visible rudiment of a stato- blast appearing at the end of the funiculus nearest to the zooecial wall, while the most developed are found near the stomach -caecum. To obtain a clear view of the origin and significance of the stato- blasts, we must, following Braem, examine the earliest stages in the development of the funiculus. This strand first arises as a fold-like prominence in the outer (mesodermal) layer on the oral side of the neck in the young polypide- rudiment. In the fur- ther course of develop- ment, this fold separates from the neck of the polypide-rudiment (Fig. 25 A), so that it as- sumes the form of a short strand (/), this beins; inserted at its upper end into the zooecial wall, and at its lower into the j^olypide- rudiment. Through the further growth of the zooecial wall, the upper point of insertion of the funiculus shifts further and further from the aperture of the polypide, and may finally reach the basal surface of tlie Fio. 25.— A, young polypide-bud of Plumatella with the rudiment of the funiculus (/) (after Braem). x, inner (ectodermal), y, outer (mesodermal) budding layer. B, longitudinal section through a funiculus of Cristatella {after Braem). hm, "formative mass"; eg, "cystigen part" of the statobla-st rudiment (st) ; e, ectoderm ; ec, inner (ectodermal) layer of the funiculus ; m, outer (mesodermal) layer of the funiculus. zooecium. The funiculus, in its origin, is, as we have seen, purely meso- dermal. Very soon, however, a conical mass of ectoderm-cells, capable of further growth by division (Fig. 25 B, ec), grows out from the zooecial wall into the funiculus, which in this way becomes THE DEVELOPMENT OF STATOBLASTS. 51 "bilaminar, consisting of an inner core of ectoderm and an outer layer of mesoderm. At the lower end of this cone the first statoblast- rudiments form, a group of ectoderm-cells (eg) becoming detached from the central core and arranged round a small cavity, thus forming a vesicle. This part of the statoblast-rudiment has been named by ISTitsche the "cystigen half," as from it are formed the cysts of the statoblasts. A second part of the rudiment, the so- called "formative mass" (bm), arises from the outer mesodermal layer of the funiculus. It represents the rudiment of the inner mesodermal mass which, filled with particles of food-yolk, is found in the statoblasts, while the "cystigen half" yields not only the cell-layer which secretes the cyst but, as Reinhard has proved, the ectodermal layer of the statoblast-germ (Fig. 26 A, a and b). According to Davenport (No. 46a) the fiinicuhis is to be traced back not so much to a fold as to an active independent wandering of its component mesoderm-cells. Kraepelin (No. 50) has stated that the inner layer of the funiculus does not grow in from the zooecial wall, but from the internal lining ■of the stomach-caecum at tlie opposite end of the strand. If we follow the statements by Braem given above, the essential agreement between the rudi- ment of the funiculus and tliat of a polypide-bud is very clear, so that we are justified in regarding the statoblasts as internal buds. A view formerly held by Verworn (No. 57) derived the cystigen half and the formative mass from a single cell, which underwent a process of cleavage. Yerworn was therefore inclined to regard the statoblasts as parthenogenetic winter eggs. As the statoblasts develop further, the complete circumcrescence of the "formative mass" (Fig. 26, b) by the vesicular "cystigen half" (a) takes place. The point last affected by this circumcrescence (jj), where for some time an aperture can be seen, corresponds to the middle point of the lower surface of the lens-shaped statoblast, which is usually somewhat more convex than the other. The ^'formative mass" is in this way enveloped by two layers of the "cystigen half" (a' and a"). The inner layer (a") corresponds to the ectoilerm of the statoblast, while the outer layer (a') is concerned in the formation of the shell (r). The cells of this outer layer first •secrete on their inner ends, i.e., on the side next to the ectoderm of the statoblast, a cuticular cyst (c) which surrounds the statoblast. In the substance of this cyst, a slit or line of demarcation soon ■appears, corresponding to the equator of the cyst; this indicates the division of the cyst into the two watch-glass-like halves which .separate later to allow of the escape of the young colony. After this chitinous envelope has been secreted, the marginal cells of the .secreting layer elongate and form a fold round the statoblast. These 52 BRYOZOA ECTOPROCTA. cells now form the annular float, the whole surface of each of them becoming cnticulariseJ. The protoplasmic body remaining within the cuticular cells then completely disappears. When the annular float is completed, it is grown over from above and below by the margins of the central caps of the chitinous secreting layer, which then secretes an external enveloping chitinous layer (Fig. 27,. ud and od). In Cristatella, the equator of the developing statoblasts is at right angles ta the longitudinal axis of the funiculus. In Plumaidla, on the contrary, it lies parallel to that structure. The elongated form of the statoblasts of Plumafella is explained in this way. The most complicated conditions are found in the annular float of Cristatella. For an account of these we must refer the reader to A''erwoiin and Braem. &2e c- Fig. 26.— Three ontogenetic stages of the statoblasts of Cristatella (after Verworn). a, cysti- gen half; a', external layer; a", inner layer of the eystigen half; b, formative mass;, c, cuticular envelope ; p, point where during circumcrescence a pore forms ; /, funiculus. The germ -body proper consists of the ectodermal layer (Fig. 26 C, a") and of the formative mass (h) which this surrounds. In proportion as the cells of the latter become filled with food-yolk the boundaries between them disappear. But not all the cells of the formative mass undergo this transformation. Some of them which are in close contact with the ectodermal layer remain unchanged. It appears that these cells, which Braem found to be si)ecially plentiful in Cristatella, where they often form a continuous. THE DEVELOPMENT OF STATOBLASTS. 53 .m layer under the ectoderm, are of importance in the development of the mesoderm-layer (Fig. 27 A, m) of the young colony. The completely formed statoblasts which, after the disintegra- tion of the parent colony, become free, are not at once capable of regeneration. The capacity for further development, as a rule, appears only after they have been frozen, or after a long period of rest when air has been excluded (Braem). A higher temperature or contact with air stimulates the statoblasts, which are then capable of germinating, to further development. A transformation of the cells of the inner yolk-mass first takes jjlace. Those of the superficial M /^ layer assume the appearance of ordinary embry- onic cells and become applied to the ectoderm (Fig. 27, ec), thus thickening the layer of meso- dermal elements (?«) mentioned above as lying below the ecto- derm. This me- sodermal layer consequently soon becomes a continuous epi- thelium. The first rudiment of the primary polypide consists of a rounded ec- todermal thick- ening (germ-disc, Fig. 27 A, p) which develops in the middle of that side of the statoblast which, when floating, is turned downwards, i.e., at the point at which the circumcrescence Fig. 27. — Two ontogenetic stages of the germinating statoblasts of Cristatella miicedo (after Braem). A, on the lower side of the statoblast (that turned upwards in the figure) the polypide- rudinient {p) can be seen in the form of a germ-disc. B, the polypide-rudiinent in a more developed condition, a, anus; an, intestinal rudiment; c, cuticular shell (disc); d, yolk substance with nuclei ; cc, ectoderm ; 7.s, budding zone of the future polypide ; m, mesoderm-layer ; n, ganglionic invagination ; o, rudiment of the oesophagus ; p, germ-disc ; od, upper spines ; sr, float ; ud, lower spine. 54 BRYOZOA ECTOPROCTA. of the formative 1iy the cystigen half was eonchided (Fig. 26 C, jj). The germ-disc, in which must also be included the adjacent lower mesodermal layer, now becomes invaginated into the yolk -mass. The closing of the aperture of this invagination gives the polypide- rudiment the form of a bilaminar sac, which develops further according to the t3^pe described above (p. 37 and Fig. 27 B). The only difference in this case is that the l)ody-cavity and all its derivatives (the lophophore- cavity, the circular canal, etc.) are originally completely filled with food -yolk (d), which disappears only gradually through the absorption that takes place during further development. The rudiments of the future buds are early to be recognised as epithelial thickenings (Fig. 21 B, As) at the margin of the stato- blast (corresponding to the oral side of the primary polypide), from which, by invagination, the second and then the third polypide- rudiments are formed. These, from the time of their first develop- ment, are thus independent of the germ-disc. In a similar way, according to Davenport, in the embryos of Phimatella, the rudiment of the second jjolypide appears independently of the first. All the buds that develop later, on the contrary, arise in connection with an older polypide-rudiment, as was described above, for the type of budding in which the polypide develops first. C. Winter Buds (Hibernacula). In the fresh -water Gymnolaemata, Vidorella and Paludicella, statoblasts do not develop. In Paludicella, isolated individuals (zooecia with rudimentary polypides) persist in an encysted con- dition. These individuals, which are enclosed in strong, highly calcified, chitinous envelopes (ectocysts), are known as winter buds [liihernacula, Van Beneden). In the s^jring the envelope bursts, and there emerges from it an individual covered with a delicate chitinous cuticle, which by budding gives rise to a new colony ; the budding may occur while still in the hibernaculum. The winter buds here represent merely the resting condition of the adult form. Their development in the spring is essentially reducible to a j^rocess of ecdysis. In a similar way, in Vidorella, short stolons with closely crowded knob-like rudiments of individuals persist through the winter, and give rise in the spring to the first cylindrical cells of new colonies (Kraepelin). REGENERATION. 55 VI. Regeneration. It has long been known that, in the marine Gyninolaemata, the polypides in the older individuals of a colony are constantly regenerated, the zooecium remaining unaffected by the processes of degeneration. The remains of the degenerated polyjDide are found as the so-called brown body suspended by strands of the funicular tissue in the body-cavity. The regeneration of the polypide takes place from the zooecial wall and, in the Chilostomata, in most cases (according to Ostroumoff and Davenport, 'No. 11) on the operculum. Harmer found recently that the first rudiment of the polypide is paired and appears on the lateral margins of the operculum, the two parts only uniting later to form an unpaired invagination. This explains the abnormal cases in which, during regeneration, two polypides are formed in one zooecium. By the invagination of the two layers of the body-wall that now takes place, a new polypide is produced in the way described above. It has long been known that, during this process of regeneration, the wall of the stomach of the new polypide comes into close contact with the brown body. According to Haddon, the latter even passes through the wall into the stomach of the newly-formed polypide, and Ostroumoff held that during these processes entoderm-elements pass out of the brown body into the wall of the stomach of the polypide, to take part in the development of the latter (pp. 43 and 44). These statements, however, have not been confirmed by Davenport's more recent researches. According to Harmer, indeed, in Flustra, the brown body is actually taken into the newly -formed alimentary canal, although in Bugiila this is not the case. We are still without any explanation of the significance of these regenerative processes, which recall the regeneration of the head in Phoi'onis, Pedicelliyia, and the Tubulai'ia, and which may be com- pared with similar processes in the Tunicata. We should, however, mention Ostroumoff's view that, together with the brown body, certain excreta are received into the intestine, which are afterwards ejected through the anal aperture with the remains of the brown body. Harmer's researches (No. 16) seem to some degree to support this view. VII. General Considerations. Recent anatomical researches, especially those of Caldwell and CoRi, show that there is a great agreement betAveen the structure of Plioronis and that of the Ectoproctous Bryozoa. This similarity 56 ERYOZOA ECTOPROCTA. is specially marked in the Pliylactolaematous Bryozoa (Fig. 28), on account of the horse-shoe-shaped lo^jhophore, the jiresence of the epistome {ej)), and the similarity in the arrangement of the body- cavity. Thus in the latter, we find a septum separating the lophophore-cavity from the rest of the coelom, and having the nephridial funnels* sunk in it, and an epistomal cavity distinct (?) from the lophophore-cavity {rf. Fig. 5, p. 9). The recognition of these points of agreement throws a new light upon the Bryozoa It shows that the Phylactolaemata represent by far the most primitive type, while the Gymnolaemata, Avith regard to the segmentation of the lophophore, have undergone simplification, the condition of the body-cavity also showing degeneration. Further, through a comparison with Pliorojiis, the remarkable form of the Bryozoan larva is to some extent explained, and the task of tracing back the Bryozoan larva to the Trochophoran type is rendered easier. The comparison of the Bryozoan larvae to the Actinotrodia has been made chiefly by Harmer (Entoj)rocta, Lit., No. 4) and Ostroumofp (No. 24). The forms of Bryozoan larvae in which the larval intestine is' retained will lie treated first, as the most primitive. Among these forms, Cijplw7iautes, by the possession of shell-valves and an atrium, shows a secondarily modified condition, the larvae of Tendra and Alcyonidium being most suited for comparison with the larvae of other groups of animals. The most striking organ of the larva is the massive equatorial corona, the locomotory organ. We should be tempted to see in this the homologue of the pre-oral ciliated ring of the TrochojyJwre, but for the difficulty presented by the fact that the circular mantle-cavity which is destined to form the greater part of the body-wall lies in front of the corona. In any case, we are led, liy a comparison with the Adinofroclia, to consider the retractile disc, on account of the similarity of its position, as the equivalent of the apical plate. This view, which is * The investigations made so far on this jioint are not conclusive. The nephridia of the Phylactolaemata (see diagram, Fig. 28, n) were first seen by Verworx, and later were described more in detail by Coui (No. 46), but their occurrence has recently been disjiuted by Buaem and Kiiaepelin [also by Oka]. Braem connects the structures referred to with his " forked canal" (p. 42). It is, however, not impossible that these organs exist side by side. Cori's state- ments have, on the other hand, been supported by Blochmann (Brachiopoda, Lit. No. 4). Harmer (No. 16) was unable to convince himself of the presence of special excretory organs in the Gymnolaemata, but we may well feel inclined to regard as a nej)hridium the intertentacular organ observed in some forms by Farke, SMrrx, HiNcKs, and more recently by Pjiouiio (see p. 14). GENERAL CONSIDERATIONS. 57 / supported by the utilisation of this structure as larval tactile organ, obtains further confirmation from the statements of Harmer and Prouho quoted above (p. 20). Both these authors were able to observe that the larval brain is connected with the retractile disc, and Prouho Avas able to add the information that the latter organ underwent, during metamorphosis, a degeneration similar to that which takes place in all other larval organs (p. 24). Of the larval organs lying on the oral side, the so-called sucker is evidently the homo- logue of the invagination which lies on the ventral side of the Adinotroclia, between the mouth and the anus (Fig. 4 C, w, p. 7)* In the Bryozoa also, during meta- morphosis, this invagination gives rise to a part of the body-wall, though it here produces only the adhesive plate which lies altogether at the base of the primary zooecium. The structures known as the ectodermal furrow and the pyriform organ, on the contrary, seem to be provisional organs peculiar to the Bryozoan larvae, for Avhich no homo- logue can be found in the AdmotrocJia or the Troclio- pliore stage of other groups. We may thus, perhaps, recognise in the Ectoj)roctous larvae a somewhat modified TrochojjJwre stage, which in the possession of a ventral sucker may be connected with the Adinotroclia Fio. 2S. — Diagrammatic median section tlirough a Phylactolaomatous Bryozoan (constructed after CoRi). This should be compared with the dia- grams of Phoronis (Fig. 5, p. 9) and Terehratida (Fig. 41, p. 80). o, anus; ed, hind-gut; eli, epis- tomal cavity ; ep, epistome ; /, funiculus ; g, ganglion; Ih, lophophore- cavity, here termed the circular canal ; m, mouth ; ma, stomach ; n, nephridium ; oe, oesophagus; t, tentacle; tm, tentacular membrane ; ts, tentacle-sheath. Eecently, however. * In position this organ agrees closely witli the rudiment of the foot in Molluscau larvae, with which it lias repeatedly been liomologised. 58 BRYOZOA ECTOPROCTA. obstacles to such a view have arisen from the study of the ontogenj of the Phylactolaeniata. While most authors are agreed in assuming that, in the Gymnolaematous larva, the retractile disc belongs to the aboral region, more recent investigators are of the opinion that, in the Phylactolaemata, the j^oint at which the polypide- rudiments appear (and which corresponds to the retractile disc) occupies the position of the vanished blastopore. We are not, at the present moment, in a position satisfactorily to solve these difficulties, which are due to the incompleteness of our knowledge of the ontogeny of the Bryozoa, and must await further investigations. In the Bryozoa, as in Plioronis, metamorphosis begins with the evagination of the sucker-like organ. This is followed by fixation and extensive disintegration of the larval organs. In Plioronis, only a few larval organs are thrown oflF, and these are replaced by permanent organs {e.g., pre-oral lobe, tentacle-crown, circum-anal ciliated ring), but, in the Bryozoa, the intestinal canal and the whole of the body-wall of the larva undergo degeneration. The latter are not simply thrown off (as are the provisional organs of Plioronis), 1)ut sink by invagination into the interior of the body (through the formation of the vestibulum), and are there transformed into the so-called brown Tjody. This degeneration affects not only the larval integument, but all the provisional organs which lie in it — the corona, the ectodermal furrow, and (according to Prouho) the retractile disc. By means of these transformations, the larva reaches an attached sac-like stage, and already shows on its surface the future ectoderm of the i^rimary zooecium. Within it are found the remains of the original central tissue and the brown body, described above, which is formed of the degenerated larval organs. A primary zooecium, at this stage of development, strongly recalls those individuals of the Bryozoan colony in which, as is often the case, the polypide degenerates. These also consist of a zooecium closed on all sides, and have within them, besides strands of the funicular tissue, a brown body derived through the degeneration of former polypides After the attachment of the Bryozoan larva, the primary polypide very soon commences to form, while the disintegration of the larval organs is still going on. Its rudiment is found at the upper or distal pole of the primary zooecium, having been produced either, as has till now been believed, by the invagination of the retractile disc, or, as Prouho holds, not directly from this, but from a new structure which appears at this point, as to the origin of which GENERAL CONSIDERATIONS. 59 nothing is known. We have further seen above (p. 37) that this polypide develops in the primary zooecium exactly as the polypides regenerate in those zooecia of the colony in which a previous polypide has degenerated. In both cases, as it appears to us, the continuity of the individual is retained by means of the persistent zooeciiim. We shall therefore have to regard the rise of the primary zooecium from the lar^a merely as metamorphosis, and to consider the newly-formed polypide as a part of the same individual Avhich is represented by the larva. The cephalic section has been, if we may so express it, I'egenerated in the attached larva. It would be theoretically inaccurate to regard the rise of the polypide in the primary zooecium as the budding of a new individual. We must here bear in mind that in the Phoronid larva also it is chiefly the organs of the cephalic region that are thrown off and regenerated during metamorphosis. The metamorphosis of the Ectoprocta is, indeed, connected with so far-reaching a disintegration of the larval organs that it is not possible to institute a direct comparison betAveen the position of the organs in the larva and in the adult. There is a gap here Avhich may be filled up by an examination of the metamorphosis of Phoronis. We may in consequence assume for the adult individuals of the Bryozoan colonies also, that the short line between mouth and anus is the dorsal middle line, and that the ganglion lying at this part represents the supra-oesophageal ganglion which is derived from the neural plate. A few remarks as to the Bryozoan larvae in general have still to be added. A comparison with the Adinotrocha is here specially instructive. The Adinotroclia leaves the egg when only slightly developed ; it feeds and grows considerably, and during larval life develops the rudiments of many important organs (and this is also the case in Cyplionautes). The Adinotrocha thus passes through important processes of growth and development, and at the same time seeks out a suitable point for fixation, and facilitates the dis- tribution of the individuals over a large area. The, swarming stage of most Bryozoa, on the contrary, serves, as a rule, merely for the last purjjose. The larva does not feed, and consequently, the alimentary canal degenerates. Its one office is to seek out a suitable point of attachment, and for this purpose it is provided with a highly-developed locomotory apparatus and sensory organs. So as to facilitate locomotion as much as possible, the rudiments of the future parts of the body are present in an invaginated condition only 60 BRYOZOA ECTOPROCTA. (pallial cavity, sucker). The internal organs, furthermore, are packed into the smallest space possible. We sliall therefore not be surprised at the absence of the coelom in the larva, although, from a com- parison with the Adinotrocha, we should have expected it to be present. The coelom, even in adult forms of marine Ectoprocta, is, in a certain sense, degenerate (the peritoneal epithelium being wanting); the cause of its excessive degeneration in the larva, however, is to be sought in the conditions just mentioned. A similar temporary obliteration of the interior of the larva is to be found, for instance, in the Platmla-lnrva. of the Cnidaria. If we now consider the Phylactolaematous larva in the light of what has been said above, we shall find that the central cavity Avithin it, taking into account its further development, must be assumed to be the coelom. In this respect, then, the embryo of the Phylactolae- mata, as compared with that of the Gymnolaemata, recalls more primitive conditions. We must not, however, forget that, in other respects, this larva has undergone the most far-reaching degenera- tions. The development of a large central cavity inside the ciliated larva may be connected with its life in a specifically lighter medium (fresh water). We find, for instance, that the larva of Spongilla is distinguished from somewhat similar marine Sponge larvae by the possession of a larger cavity. LITERATURE. Cyphonautes. Besides tlie works of Ehrenbbrg, Joh. Mullbr, Semper, Allman, J. Barrois, and H. Prouho (Xo. 28 and 28b) cf. : 1. Claparede, E. Beobachtungen iiber Anatomic und Entwick- lungsgeschichte wirbelloser Thiere an der Kuste der Normandie. Leijtzig, 1863. 2. Hatschek, B. Embryonalentwicklung und Knospung der Pedicellina echinata. Zeitschr. f. Wiss. Zool. Bd. xxix. 1877. p. 533, etc. 3. OsTROUMOFF, A. Note sur la metamorphose du Cyphonautes. Zool. Am. Jg. viii. 1885. See also below, jS'o. 26. 4. Eepiachoff, W. Bemerkungen iiljer Cyphonautes. Zool. Anz. Jg. ii. 1879. 5. Schneider, A. Zur Entwicklungsgeschichte und systematischen Stellung der Bryozoen und Gephyreen. ArcMv. f. Micro. Anat. Bd. v. 1869. LITERATURE. 61 Gymnolaemata. 6. Barrois, J. RecliGrclies sur rembryogenie des Bryozoaires. Lille, 1877. 7. Barrois, J. ]\lemoire sur la metamorphose des Bryozoaires. Ann. des Sci. Nat. Ser. 6. Tom. ix. 1879-1880. 8. Barrois, J. Embryogenie des Bryozoaires. Journ. Anat. et Phys. Tom. xviii. 1882. 9. Barrois, J. Memoire sur la Metamorphose de quelques Bryozo- aires. Ann. des Sci. Nat. Ser. 7. Tom. i. 1886. 10. Claparede, E. Beitriige zur Anatomic und Entwieklungs- geschichte der Seebryozoen. Zeitschr. f. Wiss. Zool. Bd. xxi. 1871. 1 1 . Davenport, C. B. Observations on Budding in Paludicella and some other Bryozoa Bidl. JShis. Comj). Zool. Harvard Coll. Vol. xxii. 1891. 11a. Ehlers, E. Hypophorella expansa etc. Ahli. K. Ges. Wiss. Gnttinrien. Bd. xxi. 1876. 12. Haddon, a. C. On Budding in Polyzoa. Quart. Journ. Micro. Sci. Yol. xxiii. 1883. 13. Harmer, S. F. Sur I'embryogenie des Bryozoaires ectoproctes. Arrliiv. Zool. E.,-ph: (2). Tom. v. 1887. 14. Harmer, S. F. On the Regeneration of lost parts in Polyzoa. Rep. Brit. Assoc. Adv. Sc, 1890. London, 1891. 15. Harmer, S. F. Origin of Embryos in Ovicells of Cyclostomatous Polyzoa. Proc. Camhr. Phil. Soc. Yol. vii. 16. Harmer, S. F. On the ]N"ature of the Excretory Processes in INIarine Polyzoa. Quart. Journ, Micro. Sci. (2). Yol. xxxiii. 1892. 17. Joliet, L. Contributions a I'histoire naturelle des Bryozoaires des cotes de France. ArcJiiv. Zool. Exper. (2). Tom. vi. 1877. 18. Joliet, L. Recherches sur la blastogenese. Arcliiv. Zool. Ex^x'V. (2). Tom. iv. 1886. 19. Metschnikoff, E. Ueber die ]\[etamorphose einiger Seethiere. GiJtting. Nadir. 1869. 20. ]\Ietschnikoff, E. Beitrage zur EntAvicklungsgeschichte einiger niederer Thiere. 5. Seebryozoen. Bidl. Acad. Sci. Peters- hour g. Tom. XV. 1871. 21. jMetschnikoff, E. Yergl. embryologische Studien 3. Ueber die Gastrula einiger Metazoen, Zeitschr. f. Wiss. Zool. Bd. xxxvii. 1882. 22. Xitsche, H. Beobachtungen liber die Entwicklungsgeschichte einiger chilostomen Bryozoen. Zeitschr. f. Wiss. Zool. Bd. XX. 1870. 62 BRYOZOA BCTOPROCTA. 23. NiTSCHE, H. Beitrage zur Kenntniss der Biyozoen. Zeitsclir. f. Wiss. Zool. Bd. xxi. 1871. 24. OsTROUMOFF, A. Extrait de I'oeuvre sur la Morphologie des Bryozoaires marines. Zool. Anz. Jg. viii. 1885. 25. OsTROUMOFF, A. Zui' Entwicklungsgeschiclite der cyclostomen Seebryozoen. Mittlieil. Zool. Stat. NeajM. Bd. vii. 1886- 1887. 26. OsTROUMOFF, A. Contributions a I'etnde zoologique et niorpho- logique des Bryozoaires. Arcliiv. Slav. Biol. Tom. i., ii. 1886. 27. Pergens, Ed. Untersuchungen an Seebryozoen. Zool. Anz. Jg. xii. 1889. 28. Prouho, H. Eecherches sur la larve de la Frustrella hispida (Gray), structure et metamorphose. Archiv. Zool. E.qxT. (2). A'^ol. viii. 1890. See also Comjpt. Rend. Acad. Sci Paris. Tom. cviii. 1889. 28a. Prouho, H. Sur la reproduction de quelques Bryozoaires cteno- stomes. Coin^f. Rend. Acad. Sci. Paris. Tom. cix. 1889. 28b, Prouho, H. Sur trois cas de developpement libre observes chez les- Bryozoaires ectoproctes. Compt. Rend. Acad. Sci. Paris. Tom. cxii. 1891. 29. Repiachoff, W. Zur Entwicklungsgeschiclite der Tendra zostericola. Zeitsclir. f. Wiss. Zool. Bd. xxv. 1875. 30. Repiachoff, W. Zur Naturgeschichte der chilostomen Seebryo- zoen. Zeitschr. /. Wiss. Zool. Bd. xxvi. 1876. 31. Repiachoff, W. Zur Kenntniss der Bryozoen. Zool. Anz. Jg. i. 1878. 32. Repiachoff, W. Ueber die ersten embryonalen Entwicklungs- vorgiinge von Tendra zostericola. Zeitsclir. f. Wiss. Zool. Bd. XXX. Suppl. 1878. 33. Repiachoff, A^^. Embryologie der Tendra. Zool. Anz. Jg. ii. 1879. 34. Repiachoff, W. Embryologie der Bowerbankia. Zool. Anz. Jg. ii. 1879. 35. Repiachoff, W. Zur Kenntniss der Bowerbankia-Larven. Zool. Anz. Jg. iii. 1880. 36. Repiachoff, W. On the Morphology of the Bryozoa (Russian). Zapisld Novoross. OhshcJi. Estestv. Odessa. Tom. vi. 1880. 37. Salexsky, M. Untersuchungen an Seebryozoen. Zeitschr. f. Wiss. Zool. Bd. xxiv. 1874. 37a. Seeliger, O. Bemerkungen zur Knospenentwicklung der Bryo- zoen. Zeitsclir. /. Wiss. Zool. Bd. 1. 1890. LITERATURE. 63 38. Smitt, J. A. Om Hafs-Biyozoernas utveckling och fettkroppar. bfversigt of Ixoncjl. Vetenslicths Akademiens Fijrhandlingav. Jahrg. xxii. Stocliholm, 1865. 39. ViGELius, W. J. Zur Ontogenie der marinen Bryozoen. Mittlieil. Zool. Stat. Neaj)el. Bd. vi. 1886. 40. ViGELius, W. J. Zur Ontogenie der marinen Bryozoen. Mittheil Zool. Stat. Neapel. Bd. viii. 1888. Phylactolaemata. Cf. Barrois (No. 9), Hatschek (No. 2), and Metschnikoff (No. 20). 41. Allman, G. J. Monograph of the fresh-water Polyzoa. Ray Society. London, 1856. 42. Allman, G. J. On the structure and development of the Phylactolaematous Polyzoa. Journ. Linn. Soc. Zool. Vol. xiv. 1879. 43. Braem, F. Untersuchungen liber die Bryozoen des sUssen Wassers. Zool. Anz. Jg. xi. 1888. 44. Braem, F. Ueber die Statoblastenbildung bei Plumatella. Zool. Anz. Jg. xii. 1889. 45. Braem, F. Die Entwicklung der Bryozoencolonie im keimenden Statoblasten. Zool. Anz. Jg. xii. 1889. 45a. Braem, F. Untersuchungen liber die Bryozoen des slissen Wassers. BiUioth. Zoolog. Heft. vi. 1890. 45b. Braem, F. Die Keimblatter der Bryozoenknospe. Zool. Anz. Jg. XV. 1892. vSee also Ein Wort liber H. Prof. Kraepelin, etc. Cassel, 1893. 46. CoRi, C. J. Ueber Nierenkanalchen bei Bryozoen. Lotos. Prag, 1891. New issue. Bd. xi. 46a. Davenport, C. B. Cristatella : the Origin and Development of the Individual in the Colony. Bidl. Mus. Comj). Zool Harvard Coll. Vol. xx. 1891. 46b. Davenport, C. B. The Germdayers in Bryozoan buds. Zool. A71Z. Jg. XV. 1892. 46c. Demade, p. Le Statoblaste des Phylactolemates etudie chez I'Alcyonella fungosa et la Cristatella niucedo. La Cellule. Tom. viii. 1892. (This was published after our MS. Avas, completed, and could not therefore be referred to.) 47. Hyatt, A. Observations on Polyzoa, Sub-order Phylactolaemata, Communications Essex Institute. A^ol. iv., p. 228 ; Yol. v., pp. 97-112, 145-160, 193-232. 1865-1866. 47a. Jullien, J. Observations sur la Cristatella mucedo. Mem.. Soc. Zool. France pour Vannee 1S90. Tom. iii. 1889. 64: BRYOZOA ECTOPROCTA. 48. KoROTNEFF, A. Zuv Entwiclilung der Alcyonella fungosa. Zool. Anz. Jg. X. 1887. 48a. KoROTNEFF, A. On the development of tlie Fresh - water Bryozoa. Zapisid Kievsk. Ohslich. Estedv. Tom. x. 1890. (Russian; Explanations of plates in German.) 49. Kraepelin, K. TJeber die Phylogenie nnd Ontogenie der Slisswasserbryozoen. Tagehlatt der lix. Versammlung cleufsche)' NaturforscTier und Aerzte zu Berlin, 1886. Biol. Centralhlatt. Ed. vi. 1886-1887. 50. Kraepelin, K. Die dentschen Susswasserbr3'Ozoen. Ahliandl. des natunc. Vereins Harnlmrg. Th. i. Bd. x. 1887. Th. ii. Bd. xii. ' 1892. 51. K^iTSCHE, H. Beitrjige zur Anatomie nnd Entwicklungsgeschichte der phylactolamen Slisswasserbryozoen. ArcMv. f. Anat. u. Phys. 1868. 52. !N"iTSCHB, H. Beitrage zur Kenntniss der Bryozoen. Zeifschr. /. Wiss. Zool. Bd. XXV. Suppl. 1875. 52a. Oka, A. Observations on fresh- water Polyzoa (Pectinatella gelatinosa n. s}).). Joxirn. Coll. Sci. Tokyo. Yo\. iv. Part i. 1890. 53. OsTROUMOPF, A. Einiges liber die Metamorphose der Slisswasser- bryozoen, Zool. Anz. Jg. ix. 1886. 54.' Reixhard, '\V. Zur Kenntniss der Sllsswasser-Bryozoen. Zool. Anz. Jg. iii. 1880. 55. Reixhard, W. Embryologische Untersuchungen an Alcyonella fungosa, und Cristatella mucedo. Zool. Anz. Jg. iii. 1880. 56. Reixhard, "W. Zur Kenntniss der Sllsswasser-Bryozoen. Zool. Anz. Jg. iv. 1881. 57. Yerworx, M. Beitrage zur Kenntniss der Sllsswasser-Bryozoen. ZeitscJir. f. Wiss. Zool. Bd. xlvi. 1888. APPENDIX TO LITERATURE ON BRYOZOA ECTOPROCTA. I. CoRi, C. J. Die Nephridien der Cristatella. Zeifschr. f. Wiss. Zool. Bd. Iv. 1893. II. Harmer, S. F. On the Development of Lichenopora verrucaria. Quart. Journ. Micro. Sci. Vol. xxix. 1896. III. Harmer, S. F. Article Polyzoa. Camhr. Nat. Hist. ii. 1896. 1\. Oka, a. On the so-called excretory organ of Fresh -water Polyzoa. Journ. Coll. Sci. Tokyo. Vol. viii. 1895. V. Prouho, H. Contribution a I'histoire des Bryozoaires. Arcliiv. Zool. E.rper. (2). Tom. x. 1892. CHAPTEE XVII. BRACHTOPODA. Our knowledge of the ontogeny of the Brachiopoda is still somewhat incomplete, especially with regard to certain important points. The anatomy of the adult forms is better known, but the close crowding of the organs between the shell-valves in the larva renders it difhcult to ascertain their relative positions. This very crowding has no doubt been the cause of many of the changes that have taken place in the original type of organisation. All that is known of the- ontogeny of the Brachiopoda, however, points to the conclusion that they are closely related to the two groups which have just been considered — the Plioronidae and the Ectoproctous Bryozoa. This view is founded upon the presence of a tentacle-bearing lophophore, originally horse-shoe-shaped, and of an integumental fold (epistome) above the mouth, and, further, on the agreement that prevails in the three groups with respect to the body-cavity and the nephridial system. The characteristic pelagic larvae of the Brachiopoda also, can Avithout difficulty be brought into agreement with those of the Ectoprocta. In our description of the Brachiopoda we shall treat of the Testicardines and the Ecardines separately, beginning with the first of these groups, the ontogeny of which has been more fully investigated. 1. Testicardines. A. Embryonic Development. The first ontogenetic stages of this group have been investigated by Lacaze-Duthiers (jS'o. 10 in Thecidium), Morse (JSTos. 11 and 12 in Terehrahdina), and especially by Kowalevsky (IS^o. 8 in Argiope, Thecidium, Terebndula, and Terehratulina). The more recent researches of Shipley (Xo. 16) have, in all essential points, confirmed the statements of Kowalevsky. With respect to the first ontogenetic processes, the Testicardines F 66 BRACHIOPODA. 2^V,/ may be divided into two groups, the one including the genera An/iojje, Terehrahda and Terehratulina, and tlie other being represented by Tliecidnim. The distinctions between these two groups, however, are only in points of secondary importance, and can be explained by the croAvding of the blastomeres in Thecidium. In Arfjio}je, the matuire eggs pass first into the body-cavity and thence into the nephridial canals that function as oviducts. The latter open into brood-pouches* lying on either side of the body ; these are to be regarded as invaginations of the body- wall, and in them the eggs pass through the first stages of their development. The embryos are attached to the wall of the brood-sac by a tough filament at their anterior ends. It has not yet been definitely ascertained where fertilisation takes place, but it is probable that it occurs after the egg has reached the brood-sac. Cleavage is total and almost equal, and leads to the development of a regular coeloblastula which is followed by a gastrula-stage resulting from invagina- tion. During this stage, the plane of symmetry of the body seems already to be defined. The last ])0\\\t at which the blastopore closes seems to correspond to the anterior part of the ventral side, perhaps to the position of the future oral aperture ('/. similar conditions in Phoronis, p. 2). "While the blastopore is closing, two lateral coelom-sacs become separated from the archenteron by ingrowths of its walls (Fig. 29 A); this abstriction takes place in such a way that the last vestige of communication between tlie three cavities is retained in the most anterior part of the body. In this last point only is there a distinction between this process and that by which mesodermal folds arise in Sagitta (Vol. i,, p. 367) ; in other respects the two processes are somewhat alike. After these coelomic sacs have become completely cut off, we find in the now lengthening embryo an archenteron closed on all * In certain fossil forms, the embryos seem to pass through the whole of their development within these brood-sacs, or at least within the mantle-cavity of the parent, as is indicateil by the discovery by SuEss of quite young shells enclosed in a Slringocephalus {cf. Zittel, No. 17). Fio. 29. — Two ontogenetic stages oiArgiope (after Kowalevskv, from Balfour's Text-book). A, late gastrula stage showing the origin of the coelomic sacs (pu). B, stage after the sepa- ration of the three regions of the body. 6, provisional setae ; bl, blastopore ; lae, mid-gut ; pv, coelomic sacs. EMBRYONIC DEVELOPMENT. 67 sides (rudiment of the mid-gut), which soon grows oiit posteriorly as the rudiment of the intestine, and two lateral coelomic sacs (rudiments of the middle germ -layer and the body-cavity) — Fig. 29 i?. During the whole of larval life the alimentary canal remains closed, the rudiments of the mouth and anus being wanting. The coelomic sacs grow completely round the mid-gut, the walls of the two sacs becoming applied later to form a dorsal and ventral mesentery. The embryo now grows somewhat in length and becomes marked oif into two parts by a circular furrow. The anterior part of the body is soon cut up by another circular furrow into two regions, so that the animal now consists of three parts not quite of equal size (Fig. 29 B).* These three parts have repeatedly been called segments, but Ave shall see that they are in no way true segments, and we shall therefore name them the cephalic, the tlioracic, and the pedal regions. The latter, defined by Kowalevsky as the caudal segment, contains only the posterior pro- longations of the coelomic sacs (Fig. 29 B), the mid-gut belonging to the two anterior regions. The cephalic region gives rise later to the umbrella-like cephalic section of the bodj^ which is surrounded by a ring of cilia (Fig. 30), and carries at its apex four symmetrically placed eye-spots, the dorsal pair of which appears first. In the thoracic region, a fold is soon found growing out posteriorly {m) ; this fold, which at first is circular and then becomes divided ujJ into a dorsal and a ventral lobe, must be regarded as the rudiment of the two mantle-lobes, and will here be called the mantle-fold. It carries ventrally two pairs of provisional tufts of setae {h), and almost completely encircles the pedal region * Authors differ with regard to the origin of the middle region of the body. According to Hoyer (No. 7), it becomes cut off from the anterior section of the embryo when the latter consists of two sections. This is confirmed by Shiplf-y's statements in connection with A7-giope. and Lacaze-Duthiers' in -connection with Thecidium. Oehlert and Deniker (No. 9), on the contrary, agree with Koavalevsky in the view that " Le segment median s'est probable- nient forme par la division du segment caudal." It cannot be denied that tliis is a point of some importance. Fig. 30.— Free-swimming larva of Argiope (after Kowalevskv, from Gegenbadr). 6, setae ; Hi, mid-gut ; m, mantle. 68 BRACHIOPODA. of the larva. The latter region gives rise to the peduncle of the adnlt. The eggs of Thecidium, which are distinguished by their com- paratively large size, after leaving the oviduct, pass into a brood-sac which develops as a median invagination of the ventral mantle-lobe, into which two of the cirri of the ring of tentacles hang down. To these the eggs are attached by means of fine filaments (cf. Pliownis). Cleavage here also is total and equal, but the cleavage-cavity is from the first small. An invagination of the blastoderm does not here take place, but the second embryonic layer arises " through the simple and irregular formation of its cells from the cells of the blastoderm," and thus proliably by polar ingression. The whole of the cleavage-cavity is soon filled with cells of the primary entoderm, which become arranged into three masses, in each of Avhich a cavity soon appears. The part that lies between the other two becomes the mid-gut, and the lateral parts represent the coelomic sacs, so that we now have a stage equivalent to that described above for Argiope. The further development of tlie two forms also agrees. The embryo first lengthens and becomes divided up into transverse regions. According to Lacaze-Duthiers, the middle region here arises by abstriction from the anterior half. The most anterior part of the cephalic region becomes marked off later by a circular furrow, so that the ciliated larva is finally com- posed of four distinct regions separated from each other by circular furrows. B. Metamorphosis. Our knowledge of the metamorphosis of the Brachiopoda we owe- principally to the investigations of Morse in Terehrafulina, and KoWALEVSKY in Avgiope and Thecidmm, but these investigations are far from complete. The attachment of the larva, the form of which has been briefly described above, is brought about by means of a cement secreted by tlie posterior pole of the body (pedal region). The mantle -fold now bends anteriorly (Figs. 31 F-K, 32 A), so that it soon completely envelops the cephalic segment. The former external surface of the mantle-lobes now becomes the inner surface and vice versa. In this way the points of insertion of the four larval tufts of setae come to lie on the inner surface of the mantle (Fig. 32). The setae now soon fall off (Fig. 33), and, in those forms which, in the adult condition, possess setae on the margin of tlie mantle, are replaced by the permanent setae. In METAMORPHOSIS. 69 Argiope, these latter are wanting. The two shell-valves now soon form on the outer surface of the mantle-lobes as cuticular secretions. The pedal region of the larva becomes the peduncle of the adult, and two large groups of muscles which can be recognised even in the larva (Fig. 32 B) become changed into the ventral peduncular muscles. In Liotliyriiia and Terehrcdulina there is also a pair of dorsal peduncular muscles in the larva. The two pairs of muscles which end at the bundles of setae become the shell-adductors. The pair of muscles which, in Fig. 32 i?, lies on either side of the posterior end of the digestive tract, represents the rudiment of the divaricators. This pair divides later into a pair of dorsal and a pair of ventral divaricators. Fig. 31.— Stages in the attacliiuent and metamorpliosis of the larva of Tciebratiiliria (after Morse), h, tufts of provisional setae ; c, cephalic region ; th, thoracic region ; p, peduncular region of the body. The metamorjDhosis of the cephalic region of the larva is the most obscure. In comparing this larva with that of PJioronis, we should expect that this section would give rise merely to the integu- mental fold above the mouth (epistome) and to the supra-oesophageal ganglion (see diagram. Fig. 34). According to Kowalevsky, however, it api^ears, on the contrary, that the rudiment of the oesophagus (Fig. 33 A, oe) develops as an ectodermal invagination in the region of the cephalic lobe, and the latter, with its eye-spots, is for some time longer recognisable within the attached larva (Fig. 33 B). If this is the case, we should have to attribute to the ciliated ring, which runs round the cephalic lobe of the larva, a postoral position. 70 BRACHIOPODA. It is, hoAvever, not impossible, indeed, taking into account the dorsal position of the lophophore-rudiment, it is probable that even before the oesophageal invagination appears in the cephalic lobe of the attached larva, certain displacements have taken place Avhich alter the primitive conditions {cf. Fig. 34). wi ^i| '^ whir'' m III! / Fig. 32. — Two stages in the iiietainorphosis of Argiopc (after Kowalevsky). b, provisional setae ; d, rudiment of the mid-gut ; k, cephalic lobe ; m, mantle-fold ; st, peduncle. The rudiment of the lophophore (Fig. 33 A, t) arises in the form of an almost circular thickening on the inner surface of the dorsal mantle-lobe. As the mouth is later encircled by the lophophore, this circular thickening of the cephalic lobe must extend ventrally. The first tentacle-rudiments [t) are soon seen in the form of four swellings. At a later stage these develop into hollow tubular outgrowths METAMORPHOSIS. 71 (Fig. 33 B), their number being increased by the addition of neAv rudiments, the point at which the budding of tentacles first takes place being the most anterior or distal part of the lophophore, which later becomes its dorsal part. Fig. 33.— Two older ontogenetic stages of Argiope (after Kowalevsky). a, eye-spots ; d, mid- gut ; oe, oesophagus ; st, peduncle ; t, tentacle-rudiments. It may at first sight appear remarkable that the tentacles grow out on the inner side of the mantle-lobe. But if we consider the free-swimming larva, in which this inner side still functions as the outer side of the thoracic section, we shall see that the crown of tentacles here has a similar post-oral position as in the Adino- trocha {cf. diagram Fig. 34, t). K in -■' V - / V Fio. 34. — Hypothetical scheme of the metamorphosis of the Brachiopoda for comparison with the Actinotroclia. A, free-swimming larva. The oral aperture (m,) is depicted, although not actually present at this stage. The position in which it here appears, below the cephalic lobe (A:), is not in accordance with Kowalevsky's statements. The tentacle-buds also are not actually present at this stage. B, an imaginary transition stage. C, younger Brachiopud after the reversal of tlie two mantle-lobes. Tlie epistome has arisen from the cephalic lobe (k). The row of tentacle-buds (0 belongs for the most part to the dorsal lobe of the mantle. d, dorsal lobe of the mantle ; ep, epistome ; k, cephalic lobe ; m, moutli ; st, peduncle ; t, tentacle-buds ; v, ventral lobe of the mantle. 72 BRACHIOPODA. The later transformations in the lophophore have been investigated in Terehratulina by ]\Iorsb. The lophojDhore here is originally circular, but, later, the anterior edge becomes indented. In this indentation, close to the median line, new tentacles form. The tentacular apparatus, which in this way has become horseshoe- shaped (Fig. 35), resembles the similar- organ in Plioronis and the Phylactolaemata, this resemblance being heightened by the fact that here also a dorsal fold, the lip (e), can be seen over the mouth ; this must be regarded as the epistome which, in later stages, is continued along the whole length of the row of tentacles. The mouth (/h) lies in a ciliated furrow, the buccal groove, between the ventral row of tentacles and this fold, this furrow being con- tinued on to the arms as the brachial groove. The two processes Fig. 35. — Two ontogenetic stage.s of tlie lophophore of Terehratu- lina scjitentrionalis (after Morse), a, outer, b, inner arm ; d, alimentary canal ; e, epistome ; I, hepatic appendage ; in, mouth. of the horseshoe -shaped lophophore develop into the large lateral arms of the adult, while the little, spirally-coiled, inner arms (h) grow out only later in the dorsal indentation. The details of the development of the lophophore in later stages must vary on account of the varying shape of the adult organ in the different forms ; but on this point we are still without accurate information. In Argiopc, the lophophore retains the simple original horseshoe -shape throughout life. Both tliis form and Tliccidium show a primitive condition, inasmuch as the connection of the arms with the dorsal mantle-fold is per- manently retained. With regard to the external alteration of form in the attached larva, it should further be mentioned that, according to jNIorse, tlie ECARDINBS. 73 youngest stage of TerehratuUna septentrionalis strongly recalls the shape of the shell in Megerlia and Argiope. Later, a stage develops which, on account of the long, flattened shell- valves and the long peduncle, strikingly recalls Lingula, and this leads finally to the adult form. The only details known as to the development of the inner organs are such as are easily understood. The division of the intestine into sepa- rate sections and the development of the hepatic tubes as lateral diverticula of the anterior section of the intestine r vestibule (Fig. 45 A, v) forms in the Pedicellina larva, the oral ciliated surface becoming more and more depressed. The floor of the cavity thus formed soon shows a deep depression extending between tlie mouth and the amis (Fig. 45 A), from the wall of which the sub-oesophageal ganglion is said to originate as a simple thickening of the ectoderm (Harmer) The lateral walls of this depression separate it from an outer groove in the floor of the vestibule; this latter groove runs round the anal cone and passes over anteriorly into the funnel-shaped oral aperture. This groove corresponds to the tentacle-groove of the adult. The outer thickened edge of the cup that has thus arisen now becomes separated by a furrow from the rest of the body-surface, and develops a massive ciliated ring (Fig. 45 B), Avhicli functions as the locomotory organ of the larva. As the muscle -fibres develop, the diff"erent parts of the body become markedly retractile. The floor of the vestibule, especially, can be protruded far beyond its aperture and again withdrawn. When the larva is in the condition of greatest expansion, two conical processes can be seen projecting from the aperture of the vestibule (Fig. 45 B). The posterior cone carries at its apex the anal aperture (a, anal cone), while the anterior process is distin- guished by a tuft of long flagella ; this latter process has the same position as the epistome, as it lies behind the oral aperture at the anterior edge of the deep depression mentioned above. A feature of morphological importance is the presence of an excretory apparatus {7iph) consisting of two small ciliated canals which, in position and structure, agrees with that of the adult (F'ig. 48, ex). We may compare this with the head-kidney of the Trochophore larva of the Annelida. It opens externally aita,^i:^ppx\ between the epistome and the ganglion. /C'^ '"^ '^ /\ [ujI LIBR AR Yj3o 96 ENTOPROCTA. The structure of these excretory canals has been repeatedly investigated in the adult of Loxosoma, Pedicellina, and Asc(ypoda7'ia, a certain importance being attributed to these forms in connection with the significance of this organ as the head-kidney (protonejihridium, Hatschek). The results of these investigations must not as yet be regarded as conclusive. "While Harmer, Foettinger, and Hatschek regard the cells of this organ as perforate, and its lumen thus as being intracellular, Ehlers, and more recently Prouho (No. 9), do not agree with this view. According to the latter authors, the lumen of the canal lies between the cells. According to Ehlers, the canal, ends blindly internally, but does not, as Harmer stated, terminate with a flame-cell. The most recent investigator of this subject (Prouho) even throws doubt on the blind termina- tion of the internal end. According to Foettinger and Ehlers, the two canals unite to form a common unpaired duct opening by a single aperture. Metamorpliosis. According to the statements of the older investigators (P. J. van Beneden and others), the fixation of the Pedicellina larva seemed to take place by means of the ciliated disc. This organ was then said to pass over into the pedal gland of the Loxosoma stage. It was therefore usually called, even in the larva, the "cement-gland, or sucker," or" received some other similar appellation. Metamorphosis would then simply have consisted in the growing out of the apical part of the larva as the peduncle, while the organisation of the adult would be attained by the budding out of tentacles in the vestibule. On the other hand, the more recent researches of Barrois (No. 1), which were confirmed by Harmer (No. 5), have revealed the sur- prising fact that the fixation of the larva takes place, as in the Ectoprocta, at the oral side, by means of the edge of the vestibule. During fixation the larva is in its most retracted condition (Fig. 46 A), even the cell-row of the ciliated ring being Avithdrawn into the vestibule. The aperture of the latter is very much narrowed, and soon closes completely, and its marginal cells fuse (Fig. 46 B). In this way the vestibule becomes a sac closed on all sides. In the course of further metamorphosis, the body of the larva changes in shape. The lower part narrows somewhat to form the future peduncle (Fig. 46 B and (7), while the upper part swells out into the head, this swelling being at first more marked posteriorly. The body in this way becomes almost pipe-shaped (Fig. 46 G). At the same time the alimentary canal, together with the vestibule, changes its position, its posterior portion shifting upwards. The stomach, the principal axis of which originally lay horizontally, has now assumed a more oblique position, which finally passes into a perpendicular position. The oesophagus then lies below, while the hind-gut with the anal aperture (a) lies above. These two parts METAMORPHOSIS. 97 of the alimentary canal, the oesophagus and the hind-gut, now lie almost horizontally. While the intestine has shifted in the way just described, the vestibule has considerably lengthened (Fig. 46 B). At a later stage only a part of this cavity is retained, the lower part, lying in the peduncle, disintegrating through histolysis of its walls. Further disintegration and histolysis take place also in other parts of the Fig. 46.— Three stages in the metamorphosis of Pedicellina (after Harmer). J, larva just attached. B, commencement of the rotation of the alimentary canal, and partial disintegra- tion of the organs. C, the breaking through of the vestibular aperture and the development of the tentacles (t). a, anal aperture ; d, dorsal organ ; m, oral aperture ; p, pedal gland ; (, rudiments of tentacles ; lu, ciliated disc (sucker of Harmer). vestibule and of the intestinal wall. These broken down cell-masses are found in numbers in the lumen of the vestibule and in the stomach. These processes, however, do not here lead to the destruction of the whole organ, but only affect the portion nearest the peduncle. The two larval organs, the dorsal organ (d) and the ciliated disc (w), on the contrary, are completely destroyed. After the rotation of the intestinal canal round its transverse axis, 98 ENTOPROCTA. above described, has taken place, the vestibule develops a new external aperture, its outer wall fusing with the inner surface of the body-wall, and a slit-like fissure arising at the point of fusion. In close proximity to this perforation of the wall, the rudiments of the first tentacles (Fig. 46, t) appear as outgrowths of the vestibular wall extending into the vestibular cavity. The stage in the development of Pedicellina attained by means of these transformations recalls the appearance of certain species of Loxosoma in the oblique position of the vestibular aperture, and the presence of a pedal gland (j>) which has arisen as an ectodermal thickening. Only later are these characters lost, the head becoming more distinctly marked off from the peduncle and assuming an upright position, the aperture of the vestibule being apical. During the histolysis of the inner organs, many of the isolated cells pass into the body-cavity ; these are gradually absorbed, and the cavity then appears filled with star-shaped mesenchyme cells. The metamorphosis of the larva of Loxosoma still remains unknown. There can, however, be no doubt that it is in essential agreement with that here described for Pedicellina. Asexual Reproduction. Asexual reproduction through the formation of buds plays an important part in the life-history of the Entoprocta. In Loxosoma the buds arise in large numbers at the ventral side of the cup, on either side of the parent-animal. The buds here seem to form alternately, those of the right side alternating in age with those of the left. When the buds have attained a certain grade of develop- ment they become detached, and continue their lives as solitary individuals. In Pedicellina ecJmiata, the formation of new buds takes place from a basal stolon ; since the buds retain their connection with the parent, colonies of variable size develop. These colonies are essentially bilaterally-symmetrical in their development. The stolon forms from the lower end of the peduncle of the oldest individual, on its anal side (Seeliger). The buds are arranged on the stolon in such a way that the median plane in them corresponds with that in the parent- animal. The youngest buds are found at the growing end of the stolon. Each bud develops on the oesophageal side of the bud next older than itself. In rare cases a lateral branching of the stolon has been observed in Pedicellina echinata (Seeliger, Ehlers). In other forms (Ascojyodaria), such branching is common. ASEXUAL REPRODUCTION. 99 £ The stolon is formed by the simple growth of the two germ-layers present in the peduncle, i.e., the ectoderm and the mesenchyme (Harmer, Seeliger). Consequently the buds formed on the stolon also owe their origin exclusively to these two germ-layers. Hatschek's view that, in addition to these constituents, an entoderm-sac also enters into the formation of each bud, this sac having arisen by abstriction from the mid-gut rudiment of the next older individual, must, according to Harmer and Seeliger, be regarded as erroneous. According to these two authors, the formation of the buds recalls that of the polypides of the Ectoprocta. The first rudi- ment of the bud is a bulging ■caused by the growth of the two germ-layers (Fig. 47 A, St), an ectodermal invagination soon forming at the apex of this swelling (Fig. 47 B). The sac which has thus •arisen soon be- •comes divided by constriction into two parts (Fig. 47 A), the larger outer part representing the rudiment of the vestibule (a), and the smaller inner part that of the whole alimentary canal (^). In this case, as in the Ectoprocta, the raid-gut has not an independent entodermal origin. The point at which communication between the two parts of the sacs is retained becomes the future oral aperture, while the development of the hind- ,gut and of the anal aperture follows only later. In this respect the formation of buds in the Entoprocta deviates from that in the Ectoprocta, where the anal communication of the vestibule with the mid-gut is the first to be established (p. 40). An ectodermal invagination which forms at the base of the vestibule, between the oral and anal apertures, gives origin to the ganglion, which, as a Fig. 47. — Budding in PedicelUna (after Seeliger). A, portion of a stolon with a very young rudiinent of a bud (s?) and an older rudi- ment in which can already be seen the separation of the vestibule (a) and the intestinal rudiinent (i). B, bud, with simple polypide- invagination (p, coiauion rudiment of the vestibule and the alimentary canal), a, vestibule ; ec, ectoderm ; i, rudiment of intestine ; ms, mesenchyme : o, mouth ; p, polypide-rudiment ; St, young bud-rudiment. ] 00 ENTOPROCTA. solid cell-mass containing within it a dense network of fine nerve- fibrils (Punkt-substanz), soon becomes detached from the ectoderm. The tentacles arise as outgrowths projecting into the vestibule. The voluntary throwing oft' of the head, and its subsequent regeneration from the end of the peduncle, has also recently been more carefully investigated by Seeliger. Extensive processes of degeneration occur in the "head" before it is separated. The end of the peduncle, after the detachment has taken place, sliows the same composition out of two germ -lasers which was found in the stolon, and the development of the new " head " actually takes place under exactly the same conditions and bj' just the same processes as the development of the buds on tlie stolon. The spontaneous throwing otf of the " head " and its regeneration recall the degeneration and new formation of the polypides in the Ectoprocta. The budding in Loxosojna takes place, according to Seeliger (No. 13), iu just the same way as described above for Pedicellina. Here also an ectodermal invagination j'ields the common rudiment of the vestibule and the alimentary caiial, while the mesoderm is derived from the immigrating mesenchyme-cells- of the parent-animal. As a rule, the young bud projects early as an outgrowth om the parent, but there are certain modifications in the manner in which the buds are attached in the different species. While, in Loxosoma singularc, Raja, cochlear, and phascolosomatum, the buds appear attached to the parent by the peduncle, in L. Kefersteinii, according to Nitsche and Claparede, the attachment to the parent is in the region of the dorsal side of the bud at the boundary between tlie peduncle and the trunk. The peduncle therefore- does not here grow out as a free projection from the body. According to Prouho, the budding in Loxosoma (Cyclatclla) annelidicola is at first internal, the young bud developing in an ectodermal depression forming a kind of amniotic cavity. The buds appear to develop in the same way in Loxosoma Raja; this fact led 0. Schmidt to trace back the formation of buds in Loxosoma to a imrthenogenetic development. General Considerations. In treating of the aflinities of the Entoprocta, we must start with the free-swimming Pedicellina larva. This larva may, without mucli difficulty, be traced back to the Troclwphore type. The ciliated rings of the Ped(celli?ia larva would then correspond to the pre- oral ciliated ring of the Troclwphore, while the region lying behind the ring becomes invaginated to form the vestibule. In such a com- parison, we have to regard the short line extending between the oral and the anal apertures as the ventral median line of the Ento- proctous larva, the correctness of this view being confirmed by the position of the blastopore. In comparing the Entoproctous larva with the Troclwphore, we have left the apical plate out of consideration. Whether Harmer's view that the dorsal organ (.'■) is the brain of the larva and the equivalent of the apical plate is suHiciently corroborated by the facts to be observed, must be left to tlie decision. GENERAL CONSIDERATIONS. 101 of future investigators. Beside the points of agreement with the Trochophore, the Entoproctous Larva shows many points of comparison with the Ectoprocta, which may also be traced back to the Trochophore type. We have, however, already shown (p. 94) that such a comparison meets with considerable difficulties in the details of the organisation of the two larvae. It is of great importance for a comprehension of the adult Pedi- cellina (Fig. 48) that we should carefully consider the metamorphosis of the larva, made known by Barrois and Harmer. Such an examination shows that the relative positions of the mouth, the anal aperture, the alimentary canal, the ganglion, and the nephridial canals undergo no alterations during metamorphosis. This latter consists solely in the rotation m fv /x n (t 4 ms 7 of the whole complex of organs now under consideration round a transverse axis. The vesti- bule alone undergoes any con- siderable change, part of it being dilated and part degen- erating, a new aperture also forming. It thus results, from a careful consideration of the metamorphosis of Pedicellina, that the adult is here at essen- tially the stage of organisation of the TrocJwpJiore. We are thus compelled to regard the short line extending from the mouth to the anal aperture in the adult Entoproctan as the ventral median line, for this still directly corresponds to the region at which the slit-like blastopore closed. We must, then, for the sake of consistency, regard the ganglion lying at this point (n) as a sub-oesophageal ganglion belonging to the ventral side. In the same w^ay, Ave would then, perhaps, be able to trace back the ring of tentacles (t) of the adult to the pre-oral ciliated ring of the larva. The above considerations seem to show that the Entoprocta, in spite of many remarkable points of agreement, cannot in any way be compared with or combined into the same group as the Bryozoa (Ectoprocta). We regard them as an entirely independent group, and in this respect agree with Hatschek Fig. 48.— Diagrammatic median section through an adult Pedice'lina (after Ehlers).- a, anus; ep, epistome-like oral fold ; ex, excretory organ ; g, genital organ ; -Hi, mouth ; ms, fibres of the sphincter muscle ; n, nervous system ; t, ten- tacles. 102 ENTOPROCTA. (Lehrb. d. Zool., p. 40). Just as surely as the orientation given above seems established for the Entoprocta, does that entirely ditferent interjiretation of the Eetoprocta, sujiported by a comparison with Phoronis, seem justified. In this group the oral and anal apertures are dorsal, the ganglion is supra-oesophageal, and the ring of tentacles is post-oral. The ganglia and the crown of tentacles would thus not be homologous in the two groups. If other divergences of structure be added (tlie want of a body-cavity in the Entoprocta, the position of the anal a^ierture, the characteristic nephridia formed on the type of the head-kidney), sufficient grounds are found for completely separating the Ento- procta from the Eetoprocta. We must, however, bear in mind the fact that the foundation upon Avhich the above view of the systematic position of the Entoprocta rests, contains a considerable number of hypothetical elements. Among the decisive reasons for separating the Entoprocta from the Eetoprocta are two observations of a very difficult nature (that of the metamorphosis and that of the structure of the nephridia), and, from these, errors may not altogether have been excluded. Until further corroborating investigations have been made, any attempts to decide the affinities of the Entoprocta must be regarded as provisional. "We liave, therefore, appended them to the Mollus- coida, though as an independent group.* LITEEATUEE. 1. Barrois, J. Memoire sur la Metamorphose de quelques Bryo- zoaires. Ann. Set. Nat. (7). Tom. i. 1886. 2. Ehlers, E. Zur Kenntniss der Pedicellineen. Ahhandl. der kgl. Gesellscliaft der Wissenscli. Giittingen. Bd. xxxvi. 1890. 3. FoETTiNGKR, A. Sur I'Anatomie des Pedicellines de la cote d'Ostende. Arcliiv. Biol. Tom. vii. 1887. 4. Harmer, S. F. On the structure and development of Loxosoma. Quart. Journ. Micro. Sci. (2). Vol. xxv. 1885. 5. Harmer, S. F. On the life history of Pedicellina. Quart. Journ. Micro. Sci. (2). Vol. xxvii. 1887. * [In the foregoing account of the Brj'ozoa, considerable stress has been laid on the su])posed relationship of the Eetoprocta to Phoronis, and, as a con- sequence, the Eetoprocta are completely severed from the Entoprocta. These conclusions are liy no means accepted Ity all students of the Bryozoa. Harmer (Eetoprocta Lit., No. III.) does not consider that the Eetoprocta have any connection with Phoronis, and he would regard any structural resemblances as the result of coincidence rather tlian of close relationship. Both Harmer and PiiOUHO (Ectojnocta Lit., Xo. VI.) regard the Eetoprocta and the Entoprocta as nearly related. An important article dealing with the Bryozoa has recently appeared in Sedgwick's Ted -hook of Zoology (1898). Here the whole question as to the morphology and relationship of the Bryozoa is discussed, and Harmer's con- clusions are adojjted. Important criticisms of our authors' interpretation of the surfaces of the adult and larval Eetoprocta are given. — Ed.] LITERATURE. 103 6. Hatschek, B. Embryonalentwicklung unci Knospung der Pedi- cellina echinata. Zeitsch. f. Wiss. Zool. Bd. xxix. 1877. 7. JoLiET, L. Organe segmentaire des Bryozoaires entoproctes. ArcJiiv. Zool. Exper. Tom. viii. 1879-80. 8. NiTscHE, H. Beitr. ii. s. w. II. Ueber den Ban und die Knospung von Loxosoma Kefersteinii. Zeitsclir. f. Wiss. Zool Bd. XXV. Suppl. 1873. 9. Prouho, H. Contributions a I'histoire des Loxosomes. Arcliiv. Zool. Exper. (2). Tom. ix. 1890. 10. Salensky, M. Etudes sur les Bryozoaires entoproctes. Ann. Sci. Nat. (6). Tom. v. 1877. 11. Schmidt, 0. Die Gattung Loxosoma. Arcliiv. f. Mikro. Anat. Bd. xii. 1876. 12. Seeliger, 0. Die ungeschlechtliche Vermehrung der entoprocten Bryozoen. Zeitschr. f. Wiss. Zool. Bd, xlix. 1889. 13. Seeliger, 0. Bemerkungen zur Knospenentwicklung der Bryozoen. Zeitschr. f. Wiss. Zool. Bd. I. 1890. 14. Uljanin, V. Zur Anatomie und Entwicklung der Pedicellina. Bull. Soc. Imp. des Natural. Moscoiv, 1870. 15. VoGT, C, Sur le Loxosome des Phascolosomes. Archiv. Zool. Exper. Tom. v. 1876. APPENDIX TO IITERATURE. 1. Davenport, C. B. On Urnatella gracilis. Bull. Mus. Comp. Zool. Harvard. Vol. xxiv. 1893. XL Harmer, S. F. Article Polyzoa. Camhr. Nat. Hist. Vol. ii. 1896. III. Prouho, H. Contributions a I'histoire des Bryozoaires. Archiv, Zool. Exper. (2). Tom. x. 1892. CHAPTEE XIX. CRUSTACEA. Systematic : — A. ENTOMOSTRACA. C T T> 1 • J r Apus, Brandiipus, 1. iirancliiopoda { . I Estheria. II. Cladocera. a, Phyllopoda ...uis!] h. OSTRACODA. C. CiRRIPEDIA. 1. Eucopepoda -{ I 2, Parasita. II. Branchiura (^r^wZws). B. :malacostraca. a. Lbptostraca (Nebalia). b. Thoracostraca (Podophthalmata). I. Schizopoda {Eupliausia, Mysis). ' 1. Macrura. II. Decapoda < 2. Anomura. . 3. Brachyiira. III. Stomatopoda. lA^. Cumacea. c. Arthrostraca (Edriophthalniata). I. Anisopoda (Tanais, A2)seudes). II. Isopoda. y III. Amphipoda. The Development of the Embryo. 1. Oviposition, Care of the Brood. The eggs of the Crustacea are, as a rule, spherical in form, although in a few cases they are somewhat ellipsoidal {Oniscus, Gammarus, Ligia, Palaemon, Atyephyra, Grangon, etc). Where the OVIPOSITION, CARE OF THE BROOD. 105 eggs are crowded together in a brood-cavity {e.g., in the Arthro- straca), their shape, during the first stages of development, may be rather irregular, on account of mutual pressure. It has been observed that, in various Crustacea, the female under- goes ecdysis before laying the eggs. This is the case with some Cladocera before the laying of the summer eggs (Jurine, Grobben), also with Gammarus (Della Vallb) and Alyephyra (Ischikawa). The adaptations for the protection of the eggs vary greatly. The eggs are laid singly {Cypris among the Ostracoda and Cetochilus, Dias, Centropages among the Copepoda), or in bands (Arguhis), or united into masses (Stomatopoda). The winter eggs of many of the Cladocera are either, when laid, enclosed merely in their own envelopes, or are further protected by a cuticular, saddle -shaped structure, the so-called ephippium, which is a cuticular thickening of the dorsal integument of the brood-chamber of the mother. The summer eggs of this sub-order, on the contrary, undergo their entire development within a brood-cavity, covered by the shell of the mother, and a similar cavity shelters the eggs of the Notodelphyidae (Copepoda) whilst they undergo development. In the Branchiopoda there are many difterent adaptations for the protection of the eggs, which are carried about by the mother until they reach a certain stage of development. In Ajms, for example, the eggs are carried in a watch-glass-shaped receptacle formed by processes of the eleventh pair of limbs ; in Branchipus, in a pocket-like cavity of the abdomen ; in Estheria, they are attached to filamentous appendages of the ninth and tenth pairs of legs, situated between the valves of the mother's shell. They are only deposited in the mud after the formation of the blastoderm is completed and the outer germ-layer has developed. Whereas, in the Ostracoda, the eggs are, as a rule, laid singly (Cypridae), in Cypridina, they are retained within the shell of the mother until they are hatched ; this is also the case in the Lepto- straca {Nehalia) and in the Cirripedia. In the latter, the eggs are cemented together in lamellae {Lepadldae) or enclosed in branched ovisacs {Rhizocephala). In the Copepoda, except in the cases just mentioned {Cetochihcs, Notodelphyldae), the eggs are carried in ovisacs formed by a secretion of a special cement-gland, and are attached to the genital segment. In the Arthrostraca, Cumacea, and Mysidae, the eggs lie in a brood-chamber on the ventral side of the thorax, externally protected by lamellate appendages (oostegites) of the coxal joints of the thoracic limbs belonging to this region. In the Decapoda, on the contrary, the eggs are usually attached 106 CRUSTACEA. to the limbs of the abdominal segments (pleopoda) by nieans of the secretion of a special cement-gland. 2. Cleavage and Formation of the Blastoderm. The Crustacean egg is, as a rule, distinguished by the laige amount of food-yolk contained in it. The latter consists of spherical particles interspersed with fat-drops. In most cases the food-yolk is found equally distributed throughout the egg, although, as a rule, the yolk-spherules are of smaller size at the surface of the egg. In a few cases, in eggs that contain less nutritive yolk, a superficial layer of protoplasm (formative yolk) is developed (e.g., many Cladocera and CetocTiilus). As a rule, part of the formative yolk is evenly distributed between the particles of food-yolk, while the rest is massed near the first cleavage -nucleus. Rarely, a& in Moina, the polar differentiation of the egg is made evident by the unequal distribution of the food-yolk, which accumulates in the vegetative half of the egg. In Moina also, the first cleavage-nucleus is found (as also in Cetochiius) not exactly in the centre of tlie egg, but in an excentric position, somewhat nearer the animal pole. The first cleavage -nucleus usually lies, together with an accumulation of protoplasm, near the centre of the egg ; and even in those forms (e.g., Mysis) in which discoidal cleavage takes place, it originally occupies a similar position. The Crustacean egg, after the ejection of the polar bodies and subsequent fertilisation, is usually at hrst surrounded by a homo- geneous cuticular envelope, Avhich is probably secreted by the egg itself, and must therefore be called the vitelline memhrane. It is not as yet luiiversally admitted that vitelline membrane is the correct designation for this envelope. The formation of this membrane takes place either in the lower portion of the oviduct, or only after the egg is laid (fertilisation occurring simultaneously). Claus considered that it arose as a secretion or hardening of the external layer of the yolk, and therefore assumed that it was a vitelline membrane, while Van Beneden (No. 1) thought it probable that it originated from the cells of the follicles or from the epithelium of the oviduct (in those cases wliere no follicles are developed), and accordingly called it the chorion. This latter name has been adhered to by many recent authors. H. Blanc (Xo. 35), in support of this view, has shown that in Cuma the membrane adlieres more closely to the follicle-cells than to the surface of the egg. The view that it is a vitelline membrane, -whicli Ludwig also held, receives its strongest support from the observations of Claus, who was able to prove by means of measurements that, in Chondr acanthus, a decrease in the volume of the egg took place simultaneously with its appearance, and of Gkobben, who showed that in Ccfochilus (No. 21) this membrane is formed after the egg is laid, when a similar contraction of the egg takes place. These CLEA.VAGE AND FORMATION OF THE BLASTODERM. 107 observations agree with tliose of Weismanx, who watched tlie passage of the naked egg into the brood-cavity in various Cladocera, and tlie subsequent forma- tion of the membrane. More recently, Bella Valle (No. 76) has shown that in Gammarus also the eggs are passed on into the brood-cavity without an external envelope, and only secrete the vitelline membrane after fertilisation. Secondary external envelopes are often present in addition to the vitelline membrane. Among these must be enumerated the external hard shell of the winter eggs of the Phyllopoda (Fig. 50, d, p. 108), the ovisacs of the Copepoda and the Cirripedia, and the membrane of attachment (stalked egg-shell) of the Decapoda, which does not always completely surround the Qs^g. There is much variation in the cleavage of the egg in dilferent forms of the Crustacea, but these various types of cleavage cannot be assigned with any exactitude to the different subdivisions of . the group, since distinct kinds of cleavage are to be found in nearly FiQ. 49. — Three stages in the cleavage of the egg of Lucifer (after Brooks). A, stage showing division into eight cells. B, blastula stage with central cleavage-cavity. C, gastrula stage. d, yolk-containing portions arising from the cell c. related forms. Gammarus affords an example of this, the different species of this genus showing variations of cleavage which, however, according to Della Vallb (No. 76), are not so remarkable as we were led to believe by the earlier investigations of La Valette St. George (No. 77), Van Beneden (No. 1), and Bessels (No. 2). Similar examples might be cited from among the parasitic Copepoda and the Cladocera. The latter group exhibits particularly clearly how the type of cleavage is influenced by the quantity of food-yolk present, and by the possibility of the egg being otherwise provided with nutritive material. In many forms of this sub-order, the winter %gg, which is rich in food-yolk, differs in the type of its segmentation from the summer egg, which is poor in yolk, and which, during the whole course of its embryonic development, receives from the mother fluid nourishment through the albuminiferous contents of the brood- cavity (Weismann, Claus). 108 CRUSTACEA. The following four types of cleavage* may be distinguished among the Crustacea : — Type I. Eggs witli complete^'and equal cleavage. This type is of very rare occurrence among the Crustacea. It is, however, to be found in the egg of Lucifer (Brooks, No. 43, Fig. 49), which is very poor in yolk. In this egg, after cleavage has taken place in the most regular manner, there is formed a coeloblastula consisting of few cells (Fig. 49 B) and having a spacious central cleavage-cavity, n ■}■■ 'V^ Jim B '*/^;*l*' -vV^Cii Flo. 50.— Fertilisation and cleavage of the egg of Bmncldpus (after A. Brauer). A, fertilisa- tion stage. B and C, early stages showing total cleavage. D, older stage with superficial cleavage, c, vitelline membrane ; d, secondary egg-shell ; /, cleavage-cavity ; pf, female pronucleus ; pi/i., male pronucleus ; r, polar body. this stage giving rise to a very primitive invagination-gastrula (Fig. 49 C). All the cells at first appear to be similar in form and equally provided with spherules of yolk. At the commencement of the process of invagination, however, one cell lying at the vegetative pole (Fig. 49 B, c) is to be distinguished by the greater accumulation * It should be mentioned that J, Nusbaum (No. 39) similarly distinguishes four types of cleavage among the Crustacea. His Types I. and II., however, do not agree with those here given. CLEAVAGE AND FORMATION OF THE BLASTODERM. 109 of yolk within it. First two and then four portions become separated from this cell (Fig. 49 C, d), and these, shifting out of connection with the entoderm, come to lie within the primary body-cavity at the apex of the archenteric invagination. The significance of these portions of this cell is still doubtful (rf. below, p. 127). Type II. Eggs with total cleavage in the first and superficial cleavage in the later stages. This type is very common among the Crustacea. The cleavage here begins with a total and, in most cases, an equal division (cf. Fig. 50 B and C), the egg dividing up first into two, then into four, eight, and sixteen cleavage-spheres of equal size, which are similarly filled with spherules of yolk. Within each of these cleavage-spheres lies a nucleus surrounded by a star-like mass of protoplasm which sends out numerous processes. The further cleavage proceeds, the more do these separate nuclei approach the surface of the egg. As a consequence of this they lose control of those portions of the now prismatic cleavage-cells which stretch inwards. The result is a stage in which cell-regions can be dis- tinguished at the surface divided by furrows, while in the interior of the egg the originally distinct cells have become secondarily fused together (Fig. 50 D). The cleavage has become superficial. At the same time an even sharper distinction between formative and nutritive yolk takes place. The superficial cells finally contain only formative yolk, and become separated from the nutritive yolk by a distinct line. A blastula-stage (Fig. 50 D) is thus finally reached ; this consists of a superficial layer of cells of equal size, and of an inner mass of yolk (now apparently filling the cleavage-cavity*). In the latter, no distinct demarcation between the portions belonging to the separate blastoderm-cells can, as a rule, be made out. There are, however, indications of such demarcation in the form of radial furrows, which are to be seen specially distinctly in the egg of Astacus (Fig. 55, p. 114) belonging to the next type of cleavage. In this egg, the central mass of yolk breaks up into the so-called primary or Rathke's yolk-pyramids (observed later by Lereboulet, Xo. 58, and BoBRETZKY, Xo. 41) and a spherical central body (Reichenbach, Nos. 64, 45). The yolk-pyramids here represent the yolk of the separate blastomeres, while the central mass represents the unseg- * Strictly speaking, the food-yolk does not lie in the cleavage-cavity, Init occupies a considerably larger space than did the original cavity. We must therefore distinguish two parts in the food-yolk : a central portion which fills the original cleavage-cavity, and a peripheral portion corresponding to the fused inner ends of the blastomeres. Only the distal portions of the blastomeres have entered into the formation of the blastoderm. 110 CRUSTACEA. mented mass of yolk which fills the actual cleavage-cavity. Simihxr yolk-pyramids were observed in Palaemon by Bobret^cky ; here, however, in the centre of the egg, they appear to be fused together. This is also the case in Alphetis, Palaemonetes, and Hippa (Herrick). It has been observed that, iu individual cases, not all the cleavage-nuclei shift to the surface to form the blastoderm, but that some of them may remain in the central mass of yolk {Atycphyrci, Ischikawa ; Crangon, Kingsley, Ko. 53). The significance of these cells is not yet clearly understood. Kingsley believes that they are belated cells, left behind in the process of blastoderm-formation. It is, however, possible that they ought to be regarded as early representatives of the vitellophags (c/. below, p. 134). A very ^irimitive form of cleavage, to be ranked with the type above described, is found in Branchipus, according to the unpublished researches of Brauer (Fig. 50). This form is distinguished by the fact that total cleavage is followed for a long time, giving place only in the late stages to the superficial method,* and that a gradually increasing _4. -rj blastocoele (/) appears early. Such a blastocoele is not, as a rule, to be observed in Crus- tacean eggs, save in the type of cleavage now under considera- tion, the prismatic or pyramidal blastomeres being usually in contact at the centre. Careful examination of Crus- tacean eggs that abound in yolk and belong to this type reveals the fact that, even in the first stages, the blastomeres are hardly able to contain the mass of food-yolk supplied to them, or to avoid fusing with the neighbouring blastomeres. Ischikawa (No. 51) found in Atyephyra, that, after division into two the blastomeres once more fused to form a single D Fic;. 51. — Four stages in the cleavage o Enpagums Prideavxii (after P. Mayer, from Balkodr's Text- book, hi, the completely formed blastoderm. ellipsoidal mass. In a similar manner division into four is inaugurated by the separation of the four blastomeres, which, however, soon lose their independence and completely fuse. Only in the later stages do the blasto- meres become independent. Another instance of such disturbance seems to l)e afforded by the peculiar type of cleavage observed by Mayer (No. 59) in Eupagurus Pridenuxii (Fig. 51). Here the first cleavage-nucleus divides into * According to Dr. Brauer's more recent researches, carried on since the above was written, the process in Branchipus seems to be somewhat different, inasmuch as the last stage, which was regarded as showing superficial cleavage (Fig. 50 I>), has in reality i)assed into the stage of the formation of the germ- layers, and the cleavage-cavity has become filled by the immigration of entoderm ceils. CLEAVAGE AND FORMATION OF THE BLASTODERM. Ill two, foiir, and eight nuclei without separation of the individual blastomeres ; the cleavage of the egg, which is at first complete, taking place only after these preliminary divisions of the nucleus. From the sixteen-cell stage onwards the egg then follows the superficial method of segmentation. Besides the forms already mentioned {Bra)ichi2)us, Atyephyra, Eupagurus), the following Crustacean eggs belong to this type of cleavage: — (1) The summer eggs of many Cladocera {Polyphemus and Bytliotrephes, accoi'ding to Weismanx and IscHiKAWA, No. 6, the latter form having a blastocoele). (2) The eggs of the Ostracoda (C'ypris reptans, Weismann and Ischikawa, No. 6). (3) The eggs of the free-living Copepoda (Glaus, Xos. 18 and 19, Hoek, No. 22 ; Cetochilus, ■Grobben, No. 21 ; Cetochilus and Harpacticus, Van Beneden and Bessels, No. 2). (4) Chondracanthus among the parasitic Cope])oda (Van Beneden and Bessels, No. 2), most of the Amphipoda (Uljanin, No. 75 ; Pereyas- LAWZEWA and Rossljskata, Nos. 70 to 73). From the observations of La Valette St. George (No. 77), Van Beneden and Bessels (Nos. 1 and 2), there appeared to be a considerable difference in the methods of cleavage of the •dilferent species of Gammarus. G. locusta was said to belong to the type under consideration, but the fresh-water species (G. pulex and fiuviatilis) to the type here ranked as third. Della Valle (No. 76), however, in confirmation of the older observations of Leydig, has proved that in the latter cases also cleavage is total in the first stages, so that we must class all the species of Gammanis under the present t3'pe. (6) Some of the Decapoda are, perhaps, also to be classed here ; besides Eupa- gurus and Atye- 'phyra, Pcdaemon (BOBRETZKY, No. 41) and Palaemo- netcs (W. Faxon, No. 46) may pos- .sibly belong to this type. This type of cleavage also, perhaps, includes the Cirripedia, whose first stages of development seem to follow a fairly simple course. In Balanus (Lang, No. 28 ; Hoek, No. 27 ; Nassonow, Nos. 13 and 29 ; Nusscalm, Nos. 30 and 31) the cleavage appears to be total, but somewhat \inequal (Fig. 52), in which case we should have an example of unequal cleavage in the Crustacea. The somewhat elongated egg has one pole rounded and the other pointed. By the first cleavage, which takes place horizontally or somewhat obliquely, the egg breaks up into two dissimilar spheres ; the one near the rounded, i.e., the future anterior pole, consisting entirely of formative yolk («), yields the ectoderm, while that near the pointed or posterior pole (i), which is rich in food-yolk, yields the elements of the mesoderm and entoderm. The next division takes place in the ectoderm- sphere and leads to the formation of a cap-like mass of cells (Fig. 52 B, C), which gradually grow round the sphere containing food-yolk (Fig. 59 A, b). Fig. 52.^Three consecutive stages in the cleavage of Balanus (after Lang). A, stage of division into two cells. B, the upper cell a has divided into two. C, the same after division into four. 112 CRUSTACEA. This circumorescence has Leeu described as epibolic gastrulatioii (Lang), but it must still be considered doubtful whether this is the correct interpretation. According to Nassonow's drawings (Xo. 13), it appears that when the formation of the blastoderm is completed, cell-elements are ejected by the central food-yolk sphere also, and become massed more superficially near the point last affected by the circumcrescence of the blastoderm (blastopore, Lang, Nassonow). Even this may, perhaps, be only a modification of a primarily complete and later superficial cleavage. The gastrula-stage would have to be sought later, when a small depression of the svu'face appears (Fig. 59 B, bl^^, p. 126) at the above- mentioned point, and a simultaneous immigration of the entoderm-cells (en) into the mass of food-yolk takes place. This method of formation of blastoderm would belong to that order of superficial cleavage in which the blastoderm originally appears as a disc, but would be distinguished from the typical examples of that method by the fact that the point at which the disc appears here lies opposite to the lilastopore, while in other cases the two points agree in position (c/. p. 115).* Fig. 53.— Diagram of the cleavage of Callianassa siihtcrranea (after Mereschkowski). In the stages F-II, the food-yolk is limited to the central portion of the egg. Cleavage takes place somewhat differently in Sacculina (Van Beneden, No. 25 ; Kossmann). In this egg, the formative and the nutritive yolk become separated only in the four-celled stage which is reached by total and regular cleavage. We then have four micromeres consisting of formative yolk and four macromeres consisting of food-yolk. "Whilst the micromeres increase by fission and cover the surface of the egg with a blastoderm -layer, the macromeres fuse to form a simple central mass of food-yolk. The cleavage of the yolk, Avhich * [According to Groom (No. L, Airp. to Lit. Cirripedia), the smaller of the two cells resulting from the first cleavage, which he terms the first blastomere, forms only a portion of tlie ectoderm. The second l)lastomere is formed from the yolk, and the blastoderm is formed partially from the blastomeres already present, but largely also from fresh blastomeres {incrocytes) yielded by the yolk. The first division consequently does not divide the egg into ectoderm and entoderm, but simply into a macromere (yolk-cell) and a micromere, tlie latter being but one of the future blastomeres, of which the macromere gives olf a number. He further concludes that the gastrula is formed by cpiboly, and that there is no superficial cleavage during the later period of division. — Ed.] CLEAVAGE AND FORMATION OF THE BLASTODERM. 113 lias been observed by Kossmanx, seems here also to be a secondary process occurring in the later stages. As the cleavage of Saccxdina aj^pears to belong to the Type II. b, described below, p. 115, it tends to confirm the view of the cleavage of Balamis to which preference was given above. Type III. Eggs with purely superficial cleavage. In this type, the formative yolk, from the very beginning, 1ms no control over the mass of food-yolk. The first cleavage -nucleus, lying in the centre of the egg (Fig. 53 A), divides regularly into two, four, eight, etc., nuclei (Figs. 53 B-D and bi A), which are surrounded by radiate masses of protoplasm. The areas of the separate cells, however, do not become marked off by furrows cutting right through the egg, although, in a few cases, such furrows are indicated even in early stages as grooves on the surface (Fig. 53 E). The larger the number of cleavage -nuclei becomes, the more do they shift towards the surface (Figs. 53 D and 54 B), and a regular blastoderm is formed in the same way as that described under Type I. (Fig. 53 F-H). Fig. 54. — Two stages in the cleavage of the egg of Astacus (after Morin). A, younger stage with a few cleavage-nuclei within the egg. B, older stage with superlicial distribution of the cleavage-nuclei and a correspondingly wavy surface. The processes here described as taking place within the egg, i.e., the division of the cleavage-nnclei and the shifting apart of the radiate islands of protoplasm which surround them have often been called cleavage. Indeed, these islands of protoplasm have even been named cleavage-cells, which then appear in a certain contradistinction to the masses of food-yolk. In so far, however, as we regard the wJiole egg as the equivalent of a cell, and the two, four, eight, etc , .separate spheres produced by its total cleavage as cells, it is evident that we ought not to consider these protoplasmic islands, which arc designated " cleavage- cells," as fully equivalent to blastomeres. They represent merely the centres of blastomeres, whose areas, owing to the absence of furrows, have not been demarcated. In tlie first stages of superficial cleavage, the egg is on the level of a multinucleate cell. Tiiis view is supported by the frequently observed fact tliat the so-called "cleavage-cells" are connected together by a reticulum of 114 CRUSTACEA. fine protoplasmic processes. If, tlierefore, we understand by cleavage here, as elsewhere, the act of marking off separate cell areas, it follows that the term superficial cleavage is applicable to the present type, since cleavage actually occurs here only in the superficial parts of the egg. This type of segmentation is also very common among the Crustacea. It is found in the following eggs: — (1) In the summer eggs of many Cladocera {Moina, Daphnia, Sida, Leptodora, DaphncUa, Weismanx and Ischikawa, No. 6), and in all ivinter eggs (Moina, Daphnia, Sida, Bgthobxplies, Pohiphemus, Leptodora, Weismaxn and Ischikawa, No. 16). There is thus among the Cladocera a group of forms {Bi/thotrejjhes, Polyphemus), the summer eggs of which belong in their cleavage to Type II., while the winter eggs belong to Type III. (2) In the eggs of several Isopoda [Asellus,* Van Beneden, No. 79; Porccllio, Reinhaud (No. 91) and Roule (No. 92). This type of cleavage may, perhaps, be more common among the Isopoda than has hitherto been thought. (3) In the eggs of Penaeus (Haeckel, No. 47), Callianassa sub- terranea (Mekeschkowski, No. 60), Astacus (MoiiiN, No. 61), Homarus (Hekrick, No. 50a). B Fio. 55.— Later stages in the cleavage of the egg of AstacMs (after Reichenbach, from Hatschek's Text-hook). A, section of one of the stages. The formative yolk has collected at the surface. The food-yolk has divided into separate yolk-pyramids. The central body is found within it. B, later stage in which the layer of blastoderm-cells {1) has become distinct from the yolk-pyramids (2). It is very important, in the formation of tlie blastoderm Avhich follows the superficial cleavage, to distinguish clearly two subsidiary methods : (a) Tlie formation of the Uastoderm talces place all over the surface simultaneously, e.g., in Astacus, Brancliipus, and the free-living Copepoda. * According to more recent statements by Roule (No. 92), which are not very clear, it appears as if the cleavage of Asellus were at first total and only later superficial. Van Beneden, on the contrary (No. 79), emphasises the fact that at first a mere increase in the number of nuclei within the yolk takes place, that these nuclei become distributed later at the surface of the egg, and that it is at the surface that the limitation of their areas occurs, while nearer the centre the mass of yolk remains unfurrowed. In this case, the egg of Asellus undoubt- edly belongs to Type III. CLEAVAGE AND FORMATION OF THE BLASTODERM. 115 (h) The lilastoderm develops first on the ventral side of the egg. Its formation begins at one point on the surface and proceeds gradually from this point, which always represents the future ventral side of the egg; in the Decapoda the point denotes the most posterior ■end of the ventral side, or the spot at which, later, the gastrula invagination appears. This has been observed in Palaemon and Eriplda, in which the formation of the blastoderm is completed over the whole surface of the Qgg and closed on the dorsal side, only when the rudiments of the embryo are already to be seen on the ventral side. These modifications in tlie forniatioii of the blastoderm are found both in Type II. and Type III., and thus result in four subsidiary methods of cleavage, uhich deserve detailed description : Type II. a. , The cleavage is at first total, and later superficial, being folloAved liy the simultaneous formation of the blastoderm all over the surface ; Branchipus, free-living Copepoda, summer eggs of Polyphemus and Bythotrephes, Eupagurus. Type lib. The cleavage is at first total, later superficial, the blastoderm developing first on the ventral side. This method is exceedingly common among the Amphipoda. As, in this method, the cells on the future ventral side ole at which the small cells are found belongs to the vegetative half, that jiortion of the surface at which, later, the formation of entoderm takes place, -vvhile the larger cells belong more to the animal region (the later dorsal side). In any case the two axes here compared (that jtassing through the animal and vegetative poles, and that from the small-celled to the large-celled pole) do not Appear to coincide, but to cross each other obliquely {cf. below, p. 142). Type I [I. a. The cleavage is purely superficial, the subsequent formation of hlastoderm taking place all over the surface simultaneously ; many Deca]ioda ■{Penaeus, Asfacus, CitHianassa), all the winter eggs and many summer eggs •of the Cladocera. Type Ill.b. The cleavage is purely superficial, the blastoderm developing first on the ventral side of the egg. In this case a few of the numerous •elements to be found within the yolk first shift to a definite point at the surface, there to be transformed into blastoderm -cells. There thus arises a small blastoderm-disc, corresponding in position to the future veiltral side of the egg ; this disc gradually increases in size, new elements from within the yolk con- stantly rising to its periphery and becoming changed into blastoderm-cells. This method of cleavage consequently strongly resembles the discoidal cleavage to be descriljed^ presently. The distinction between the two types consists in Ihe fact that in Type Ill.b, the increase of the blastoderm is due to the accession •of new elements from the inner portion of the egg, while in true discoidal •cleavage (IV.), the increase takes place exclusively by the division of the elements 116 CRUSTACEA. already present in tlie blastoderm -disc. It is probable that these t\vo types have- been confused one with the other, since in many eases in which the occnrrence of discoidal cleavage in Crustacea has been maintained, the observations were not confirmed by the systematic examination of sections. It therefore appears probable to us that most parasitic Cope})oda, to which Van Beneden and Ressels (No. 2) ascribed discoidal cleavage, in reality develop according to- Type III. h. The same is possibly the case with the Isopoda [Onificus, Ligia*), for which discoidal segmentation was asserted by Bobhetzky (No. 80) and Van Beneden (No. 1). We are, indeed, justified in raising the question whether true discoidal cleavage ever occurs in the Crustacea, and whether more careful examination might not lead to the complete absorption of Type IV. in Type III. h. Among the Decapoda also there are some representatives of this, last type, si;ch as Homarus, Eriphia, and perhaps also Palacvion. It is possible, however, that the last-named genus, on account of the total (?) cleavage that takes place in tlie first stages, should, like Atiic2)}ujra, be assigned to Type II. 6. Type IV. Eggs with discoidal cleavage. In tlie types of cleavage which we have so far considered, two processes go on side by side simultaneously, viz., (1) the increase of the elements,^ and (2) the separation of the blastomeres from the food-yolk {i.e., the separation of the plastic portion of the egg from the nutritive). In the blastula stage, which finally results in Types II. and III., we then find a superficial layer of epithelium and an inner mass of food-yolk, in which, as a rule, no cell nuclei or other plastic elements are still to be found. If we now imagine this process of separation between the blastomeres and the food-yolk to be shifted back to the earliest stages, we obtain an explanation of discoidal cleavage, as it has been observed in Mijsis (Van Beneden, No. 37 ; Nusbaum, Nos. 38 and 39) and Cuma (No. 35), as well as in some other forms, -j- Here the very first cleavage-cell becomes entirely separated from the food-yolk, on the surface of which it comes to lie. The food-yolk from tliis time onward contains no more cleavage-nuclei. The formation of the blastoderm begins with the superficially placed cleavage-nucleus, which divides (Fig. 56 A), and thus yields a cap of blastomeres (Fig. 56 B) ; these increase in number by continual division, and gradually grow over the whole surface of the sphere of food-yolk. The starting-point in the formation of the blastoderm here corresponds to the spot where, later, gastrulation takes place (posterior end of the ventral side of the embryo), Avhile the blasto- derm is finally completed on the dorsal side. At this ventral starting-point of the blastoderm there is from the first a thicker * Quite recently Nusbaum has again maintained the presence of discoidal cleavage in Ligia (No. 85rr). t [BouTCHiNSKY [Zool. Anz., XX., 1897, p. 219) describes a discoidal seg- mentation in Nchalia.— 'E.D.I CLEAVAGE AND FORMATION OP THE BLASTODERM. li; deposit of cells, which are here deeper and form a rounded thicken- ing (germinal disc). Not ouly the eggs of Mysis and Cuma, but those of several Isopoda are said to show this type of segmentation (Onisctis, according to Boehetzky, No. 80, Ligia, according to A^an Beneden, No. 1). It is also said by Van Beneden and Bessels (No. 2) to occur in many parasitic Copepoda {Anchorella, Caligus, Clavella, Lernaea, Lernaeopoda, BracldeUa, etc.). It must, liowever, be con- sidered probable that the greater number of these cases really belong to Tj'pe III. 6. Tliis view is supported by the observations of Boutchinsky (No. 37a), who, according to the plates given in his Russian treatise, observed -a simple superficial cleavage in Faradopsis cornuta. Our retention for the present of the discoidal type of cleavage for the Crustacea is due entirely to Nusbaum'8 recent descri])tion of Ligia oceanica, according to Mhich a type of segmentation agreeing with that given above for Mysis actually occurs (No. 85a). The type of discoidal cleavage here described shows some superficial resem- blance to that kind of discoidal cleavage which, in many groups of animals (e.g., Cejjhalopoda), is developed from a total, unequal cleavage. Closer examination reveals, however, tliat what we are now considering is a peculiar process of dis- ^ coidal development of the germ, which has evidently developed independently among the Crustacea out of a suj^erficial type of segmentation. For where dis- coidal cleavage has developed out of total, unequal cleavage, we find that the pole of formation of the germ-disc con-esponds to the animal pole, its gradually out-spreading edge to the blastopore, and the plug of yolk to the vegetative pole of the egg. Here, however, in the discoidal cleavage of the Crustacea, it is quite otherwise. The formative jiole of the germinal disc corresjionds to the ventral side of the embryo, and all observations point to the fact that here also the formation of the germ -layers commences, and the important but somewhat obscured process of gastrulation. The circumcrescence of the food-yolk here proceeds from the ventral to the dorsal side, and has evidently in this case nothing to do with gastrulation, since we should otherwise be compelled to assume, in Crustacea with discoidal cleavage, a blastopore closing dorsally, which would be in contradiction to the conditions found in other Crustacea. By the above considerations we are led to regai'd the discoidal cleavage of the Crustacea as an extreme case of that tyjie of cleavage described above as Type III. 6. In the discoidal method, the formation of blastoderm at the jjole where it originates is carried out so early, that its rudiment originally consists of a single blastoderm-cell, which, by subsequent successive division, yields the whole blastoderm. Fig. 56.— Two stages in the segmentation of Mysis (after Van Beneden) as an example of discoidal cleavage. A, two cells are to be seen at the surface of the yolk. B, the two cells have increased by division and form a cap. 118 CRUSTACEA. If we consider the variations found in the different j^'anes of organisation among tlie Vertehrata, we find a certain element of similarity between the discoidal cleavage of these latter and of the Crustacea in the fact that in Loth cases a deposit of food-yolk on one side of the principal axis takes place, and determines the peculiar type of development. In those A''ertebrates in which discoidal cleavage occurs, the dorsal side of the body takes the lead in develoii- ment, while the ventral side is hindered in its develo]>ment by the accumulation of food -yolk. The blastopore is here shifted to the dorsal side. In those Crustacea in which discoidal segmentation is found, on the contrary, the rudiments of the ventral side appear first, aiid the blastopore takes up a ventral position corresponding with the plan of organisation of the group. In this case, the dorsal side of the body is influenced in its development by the accumulation of food-yolk. In many Crustacea, when the formation of the blastoderm is complete, a cuticle is secreted at the surface of the blastoderm cells. We follow Van Benedbn (No. 79) in calling this membrane the blastodermic cuticle (cuticula blastodermica). Its ajDpearance can only be explained by a process of ecdysis shifted back to an early embryonic period. Similar membranes are given ofi' by the eggs of the Arachnids and of Limulu^. The formation of a blastodermic cuticle is to be observed specially among the Malacostraca. It has, however, also been seen in the parasitic Copepoda (Van Beneden, No. 17). In the Malacostraca it is very common, being found in Nehalia (Van Beneden, No. 79), the Cumacea (H. Blanc, No. 35), in many Decapoda (Lereboullet, No. 58, and Reichenbach, Nos. 64 and 65, found it in Astacus ; P. Mayer, No. 59, in Uujxtrjurus ; Bobretzky, No. 41, in Palaemoii ; Kingsley, No. 53, and Van Beneden, No. 79, in Crangon ; DoHRN, in Forlunns), in the Amphipoda (Van Beneden and Bessels, No. 2, in Gammarus lociista; Van Beneden, No. 79, in Caprella; Uljanin, No. 75, OrcJiestia) ; finally, in the Isopoda (Van Beneden, No. 79, in Ascllus ; Bobretzky, No. 80, in Oniscus). Further, it has been observed by Dohrn in Tanais, The formation of this blastodermic cuticle is, in many Crustaceans, followed Ijy other ecdyses in later embryonic stages. This is especi- ally the case where the development is abbreviated, i.e., where many of the stages of development are shifted back into embryonic life. The cuticles formed during these ecdyses are usually distinguished by swellings corresponding to the rudiments of the limbs. These membranes are called larval integuments. As there is considerable variation in the time of appearance of all these cuticular membranes, it is often difficult, in single cases, to distinguish between the actual egg-integument, the l>lastodermic cuticle, and the larval integuments which appear later, and to establish with exactitude the homology of the special cuticular structure in each individual case. There can, however, be no doubt that the cuticle very common among the Arthrostraca, and observed in Ligia by F. MuLLER (No. 4), and furtlier, the integument developed in Mysis and THE FORMATION OF THE GERM-LAYERS. 119 the Decapoda after the completion of the Nauplius stage are equivalent to' larval integuments (Van Beneden, No. 79). In the Decapoda, a second larval integument is often developed in later stages ; this surrounds the hatching Zoaea, and was assumed by Conn to be the cuticle of the Protozoaea stage. In Anchorella and Lernacopoda (Van Beneden, No. 17), the embryo moults three times during the course of its development: (1) When the blastodermic cuticle is formed, (2) when the Nauplius cuticle develops, and (3) during the transition to the Cyclo^^s stage. >^Dohrn has bestowed special attention on the subject of larval integuments in his various works {cf. his treatise on the larval integument of the Cumacea, of Tannis, and on the Nauplius stage in the egg of Daphnia longispina). On account of tl o great variety prevailing among the Crustacea, and the uncertainty in the identification of the cuticles in the different cases, we should be overstepping the limits of this work were we to enter upon all the cases illustrating tliis point. The peculiar processes which were observed in the winter eggs of many Cladocera by Weismann and Ischikawa (No. 16), and described as para- copulation, synchronise with the phenomena of cleavage. Here, after the ejection of the polar bodies and subsequent fertilisation, there is found a body resembling a nucleus surrounded with an accumulation of protoplasm ; this is called the eopulation-ccll. During the first division of the cleavage-nucleus, by which the purely superficial segmentation of the^pe (III.) is introduced, the copulation-cell remains apparently jjassive near the vegetative pole of the egg ; it, however, soon approaches one of the nuclei which result from the division and becomes completely fused with it. The further fate of the cleavage- nucleus thus entering into paracopulatiou has not been traced. The view that it is destined to yield the genital rudiments is a mere assumption. The copu- lation-cell appears first at the time of the formation of the egg. "When the latter is maturing in the ovary, chromatin particles are ejected from the germinal vesicle ; these unite to form the nucleus of the copulation-cell, which, at a later stage, becomes surrounded by a mass of protoplasm, probably arising from the cell-body. So far no hypothesis has been formed as to the significance of the processes of paracopulatiou. The origin of the copulation-cell recalls the ejection of chromatin particles by the germinal vesicle observed by Stuhlmann and Blochmann in the eggs of insects. Similar processes have also been noted in the maturing eggs of Myriopoda (Balbiani) and Arancae (Leydig) and in other groups of animals. 3. The Formation of the Germ-layers. A. Copepoda. Among all the Crustacea, so far as our present knowledge of their ontogeny enables us to judge, the Copepoda most closely resemble the Annelida in their development. In them we have an invagi- nation-gastrula and the formation of the mesoderm through the separation of two primitive mesoderm-cells. The formation of the germinal layers in the Copepoda was made known by the researches of Grobben (No. 21), Hoek (No. 22), and Urbanowicz (Nos. 23 and 24). We have utilised the minute observations of the first of these writers in the account of Cetochilus here given. 120 CRUSTACEA. Cetochilus, like most free-living Copepoda, belongs to our second type of cleavage. At first the cleavage is total, in later stages superficial (c/. p. 109). As early as the thirty-two-cell stage, the transition to the actual blastula- stage commences, the first histo- logical differentiation in the different germinal layers becoming perceptible. At this stage there is found a small segmentation- cavity in which the food-yolk is deposited, and into which the polar bodies also find their way. A similar immigration of the polar bodies was noticed by Weismann and Ischikawa (No. 6) in the - en UM Pig. 57. — Four stages in the development of Cetochilus (after Grobben'). a, thirty-two cell stage seen from the vential side. B, later stage, same aspect; all the germinal layers are already distinct. C, longitudinal section of the gastrula-stage. D, ventral aspect of the gastrula-stage, in which the blastopore is closing, en, central entodenn-cell ; gm, mouth of the gastrula (blastopore) ; m, inesoderm-cell ; sii, lateral entoderm cells ; uvi, primitive mesoderm-cfills ; vn, anterior entodemi-cells. summer eggs of Bythotrephes. It is probable that the small cell observed by Urbanowicz in the cleavage-cavity of Ci/dops is also to be referred to the polar bodies. If we examine the vegetative pole at the thirty-two-cell stage of Cetochilus, a decidedly bilateral arrangement of the blastomeres can be recognised (Fig. 57 A). We find two cells, a larger one {ai), distinguished by the coarse granulation of its protoplasm, and a smaller anterior cell {vn). These two cells lie in the median plane, COPEPODA. 121 and at a later stage yield exclusively entoderm-elenients. They are distinguished as the central {en) and the anterior (vn) entoderm-cells. The four blastomeres (lateral cells) situated symmetrically on each side of these two, yield, by future division, both entodermal and ectodermal elements. The cleavage-sphere [u), which lies behind the central entoderm-cells, also appears to be of importance. This divides later into four elements, two larger anterior cells representing the primitive mesoderm-cells (Fig. 57 B, um), while two posterior •cells become ectodermal elements. Fig. 57 B shows us the central entoderni-cell (en) divided to form two blastomeres ; the lateral entoderm -elements (sn) have also become distinct by the division of the lateral cells. We thus have the rudiment of the entoderm consisting of seven cells, behind which lie the two primitive mesoderm-cells {ain). The immigration of the mesoderm elements towards the centre of the embryo next takes place. The primitive mesoderm-cells yield by division two laterally placed elements (Fig. 57 C, m, and «?n), and these four mesoderm-cells (of Avhich the two median cells are to be considered as the pole-cells of the lateral mesodermal band) shift into the segmentation-cavity (Fig. 57 C). Soon after this there occurs the invagination of the entodermal elements {en), by which the gastrula-stage is reached (Fig. 57 C). The blastopore, a longitudinal fissure (Fig. 57 D), now closes from before backward, and the ento- derm which has sunk inwards thus becomes a closed vesicle. It appears that the blastopore corresponds in position to the future ventral side of the embryo. If so, the part that closes latest would lie in the neighbourhood of the future anal aperture. The stomodaeum and the proctodaeum, according to the researches of Urbanowicz in connection with Cyclops, arise as ectodermal invaginations, the former appearing during embryonic development, while the latter develops only in the earliest larval stage. They both become connected with the archenteric vesicle. The gastrula-stage iu the Copepoda was tirst investigated and described hy HOEK. The statements made by Urbanowicz as to the formation of the germ-layers iu Cyclops do not agree witli Gkobben's views. In Cyclops there is, at first, only one entoderm-cell which sinks inwards, the blastopore closing above it; this cell then yields by division the whole entodermal rudiment. A mesencliyme next arises by the abstriction of ectodermal elements, this mesenchyme giving rise to most of the mesodermal structures of the NaupUus, while the actual mesoderm is a later, secondary structure, probably originating from the entoderm and yielding exclusively the mcsoderm-band. When, however, we take into con- 122 CRUSTACEA. sideration the fact that round the central entoderm-cell in Cctochilus there lie- elements -which divide into ectodermal and entodermal elements, it appears- possible that Urbanowicz has taken tliis process for the formation of the mesenchyme. The later fate of the mesoderm in the Copepoda has not yet heeii clearly made out. It appears, however, that its elements, in the segments of the Ncmplius stage, become divided more irregularly,, after the manner of a mesenchyme, and very soon become grouped as the organs of the Nauplius. Certain cells come to lie along the intestine, and give rise to its musculature, others form the muscles of the limbs, or unite to form the antennal gland. The body-cavity here exhibits the character of a pseudocoele. In the posterior part of the body of the larva, which yields the remaining and greater number of body-segments, on the contrary, a true paired mesoderm- band is developed ; in this, according to Urbanowicz (No. 23) and Fritsch (No. 20), the rudiments of true coelomic vesicles appear. 'J he most anterior pair of these vesicles represents the maxillary segment. The dissepiments between the consecutive coelomic vesicles, which Grobbbn also (No 21) appears to have observed in the abdomen of CetocMhis, disappear in the later stages, whereas a dorsal and a ventral mesentery are said to persist throughout life (Fritsch). The dorsal mesentery is attached to the back by the separation of its two halves, thus leaving a median dorsal sinus Avhich must be regarded as a remnant of the blastocoele and as the homologue of the cardiac cavity. This dorsal sinus is connected with the anterior portion of the body-cavity, which develops as a pseudo- coele. Even in early stages, when the mesoderm-band is still short, one large cell can be distinguished from the rest ; this is the genital cell, which develops on each side into the rudiment of the genital glands. [Of. Hacker (App. to Lit., Copepoda, No. I.)] The food-yolk, in Cetochilus, is present in small quantities, and is of little importance. In the eggs of the parasitic Copepoda, in which it is plentiful, it aj^pears, according to Van Beneden, that the cells of the entoderm at first migrate into the yolk and take it up into themselves, thus bringing about the appearance of a secondary segmentation of the yolk. At a later stage, however, the cells again rise to the surface of the mass of food-yolk, there forming an epithelium which becomes the wall of the mid-gut (Fig. 73 C, e7i, p. 148). The latter thus finally surrounds the remains of the food-yolk decreased by gradual absorption (cf. below the formation of the mesenteron in the Cirripedia, pp. 126, 174). PHYLLOPODA. 123 B. Phyllopoda, The formation of the germ-layers in the PhylloiDoda is best known in the case of the summer eggs of Moina, one of the Cladocera, which have been closely investigated by Grobben ( jSTo. 11). There is considerable resemblance between the processes here and those already described in connection with Cetochilus. We must not, however, lose sight of the fact that two factors in the development of the eggs of Moina have brought about a difference: (1) The nourishment by means of the_ blood-plasma transuding into the brood-cavity, which probably leads to secondary diminution of the food-yolk (in Cetocliilus also the yolk seems to be secondarily diminished, although from other causes), and (2) the paedopartheno- genesis, which is connected Avith the precocious development of a distinct genital cell. Cleavage here, as in most Cladocera, is purely superficial (Type III., (/. p. 113). As early as the thirty-two-celled stage we find the blastomeres at the surface fairly sharply marked off from the central mass of food-yolk. As in Cetochilus, at the vegetative pole of the egg, certain differentiations appear which accompany the formation of the germ-layers. In this region are found those rudiments which, at a later stage (after a certain amount of displacement), lie at the ventral surface of the embryo. There is here a central richly- granular cell, which may be called the genital cell (Fig. 58, g), and from which, later, the paired genital rudiment arises. Behind this lies a cell represented in the act of division, which as the entoderm- cell (en) represents the rudiment of the whole entoderm. At a somewhat later stage, these two rudiments have, by division, become multicellular. An area composed of numerous entoderm -cells (Fig. 5S B, en) can then be distinguished, and, anteriorly to this, four genital cells (g). Around the latter are a number of cells representing the rudiment of the mesoderm (ms). All the remaining cells now form the ectoderm. Even at this stage the rudiment of the mesoderm has a tendency to migrate inwards below the genital cells (Fig. 58 B). In later stages, this process is completed (Fig. 58 C, ms). The mesoderm now lies entirely within the embryo, and at the same time the entodermal area becomes invaginated, the gastnda stage being thus reached (Fig. 58 G). Soon after the mouth of the gastrula has completely closed, the eight genital cells, produced by division from the four above- 124 CRUSTACEA. mentioned, shift inwards and lie upon the entoderm {cf. below, Fig. 72 A, (I, ]). 147). Grobben holds that the point at which the blastopore closes corresponds to the future oesophageal invagination. It would, however, be more in agreement with the conditions met with in other Crustacea, especially in the Decapoda, if we might assume that it lay in the neighbourhood of the future anal aperture. While the embryo lengthens, the Nauplius limbs grow out and the rudiment of the brain becomes distinct at the anterior end of the dorsal surface as an ectodermal thickening {neural plate, Fig. 72 A, p. 147); the rudiments of the internal organs also undergo corresponding development. The entoderm {en) develops into a cylindrical body, the cells of which, in cross section, appear radially arranged ; no lumen is, however, at first to be seen. The stomodaeum and proctodaeum (Fig. 11 B, m, af) arise as ectodermal invaginations ; the former is distinct even in the Nauplius, the latter only at a mf' .en Fig. 5S.— Three stages in the development of the suiniiier egg of Moitni (after Grobben). A, egg ill llie thirty-two-cell stage seen from the vegetative pole. B, blastula stage, same aspect. G, median section of gastrula stage, b, blastopore ; en, endoderin-cells ; layed hy the food-yolk in the eggs of those Cladocera or Branchiopoda in which it abounds. Moina is a striking exception to the rule in being poor in yolk. In young BrancJiipus larvae all the tissues, even the ectoderm, appear to be permeated with granules of food -yolk (cf. Claus, No. 9). In Daphnia shnilis, according to the recent researches of Lebedixsky (No. 11a), a blastoderm, of equal thickness throughout and completely covering the egg, is first formed by superficial cleavage. This only thickens later at points where the cells become elongated in the region of the cephalic lobe and on the ventral side of the egg. The formation of the germinal layers is com- menced l)y the appearance of a very shallow depression (blastopore), from which point immigration of amoeboid cells into the yolk takes place. The latter represent the meso-entoderm. While the mesoderm-eells become arranged into two sj-mmetrical bands running forwards from the blastopore (mesoderm-bands), the entoderm forms a solid strand, in wliich a cavity develops at a later stage. Not all the entoderm-cells, however, take part in the formation of this mid-gut strand. ' ' A few of them form a covering to the food-yolk, and give rise to two large symmetrically-placed provisional hepatic vesicles." (?) In Moina, the breaking-up of the mesoderm into somites and the development of a true coelom has not been observed. In the Brancliiopoda, where the formation of the germ-layers is not yet known, we must fall back on observations recorded in connection with the larval stages of Artemia and Branchipus. In the earliest NaupJii of Artemia, there is, according to the figures of Xassonow (No. 13), a temporary development of paired coelomic vesicles. In Brancliipus, on the contrary, whose youngest or Metanaiiplius stage has been carefully investigated by Claus (No. 9), the process is different. Here the mesoderm in the region of the actual Nauplius segments and of the terminal segment has already become modified for the formation of organs, and has attained definite histological differentiation. The same is the case with the splanchnic layer along the Avhole of the alimentary canal (Fig. 88 A, sp, p. 179). Fn those segments which are interposed, between the mandibular and the terminal segment, and are found ia the act of appearing, the somatic layer bears a more embryonic stamp. It is here arranged in paired mesoderm-bands, whose cells appear to be seg- nientally arranged in a definite manner. This arrangement is due merely to a regular grouping of the mesoderm-cells, Avhich to some extent recalls the arrangement described below (p. 137) in connection with the Isopoda. In the most posterior regions of the body, the mesoderm-bands are united to form a plate lying below the intestine, and here is found the budding zone, from which proceeds the formation of new segments. Grobben thought himself justified in 126 CRUSTACEA. assuming that two cells lying behind this budding zone, at the edge nearest the terminal segment, were primitive mesoderm-cells ; but Claus proved that these cells, of which there are two on each side, ia the stages hitherto examined, do not take part in the pro- duction of mesodermal elements. The most striking peculiarity in Dranchipus appears to be the early development of the splanchnic mesoblast forming the intestinal muscles. C. Cirripedia. After tlie l)lastoderm is fully formed (c/. p. Ill), the embryo of Balanus consists of a layer of cells (ectoderm) completely covering the suiface of a central mass of food-yolk (Fig. 59 B). Near the posterior pole of the egg, where an insignificant depression {bl, the blastopore) is perceptible, tlie blasto- derm consists of several layers. The deeper layers represent the rudiment of Fig. 59.— Longitudinal sections through three embryonic stages of Balanus improi'isus (after Nassonow). a, Interstage of segmentation (cf. above, Fig. 52, p. 111). B and G, stages in which the germinal layers are being formed, hi, blastopore ; cc, ectoderm ; en, entoderm ; ms, mesoderm. the entoderm (en) and the mesoderm {ms). From this point the mesodermal elements are distributed along the ventral side of the egg (Fig. 59 C) in the form of a symmetrical mesodermal plate, the ectoderm being correspondingly thickened, so that in this way a ventral thickening of the superficial layers of the embryo resembling a germ-band arises. The elements of the entoderm, on the contrary, are now evenly distributed througli the food-yolk (Fig. 59 C, en) ; then, by a secondary cleavage of the yolk, the formation of definite cell-areas is brought about. * Finally the nuclei of the yolk-laden entoderm-spheres shii't to the surface to form the epithelium of the mid-gut, whose cavity arises by the absorption of the food-yolk (arts of the optic lobes LEPTOSTRACA, SCHIZOPODA, BECAPODA. 157 becomes the rudiment of the hilinim, liehinri which the stomodaeal invagination soon appears. The latter, as a rule, lies in the space between the first and second pairs of antennae, but Kingslfa' has specially emphasised the post-oral position of the first pair of antennae in C'rangon. Paired swellings arise near the posterior margin of the oral aperture; these are the rudiments oi the paragnatka or bilobed lower lip. The later stages (Figs. 80, 81, and 83) are characterised by the further growth of the thoraco-abdominal rudiment, which soon breaks up into segments, and also by the development of the posterior pairs of limbs which appear from before backward. At the same time the typical biramose character is found in the anterior limbs which were the first to develop, as well as a seg- mentation cor- respondingwith the developing joints (for the number and shape of the limbs of the li a telling em- bryo, which vary greatly in the different groups of the Decapoda, see below, p. 257 : M etamorphosis of the Deca- poda). The optic lobes gain in independ- ence, bulging Fig. 80. — Embryo of Astacus fluciatilis, with the thoracic feet beginning to form (after REiCHF.NBAfH, from Lang's Text-book). A, eyes; Oj, a„, first and second antennae; ah, abdomen; y, rudi- ment of the brain (procerebrum+antennal ganglion) ; go, optic ganglion; /, upper liiJ ; m, mandible; mx^, mx.,, first and second maxillae; T, telson ; t-, t„, thoracic limbs; ?,, t^, maxillipedes ; tf, t^, ambulatory limbs ; Is, rudiment of the thoracic shield. forwards and gradually rising from the surface below them, so that the club-like shape of the stalked eye can be recognised in the rudiment (Fig. 81). The posterior part of the body undergoes important alterations. There is here an early separation of the terminal or anal segment (Figs. 80, 81, T) from a budding zone composed of large cells which lies in front of this segment (in the embryo, however, owing to the forward curvature of the tail, further back). The anal aperture is 158 CRUSTACEA. originally formed on that surface of the terminal segment which afterwards becomes the dorsal surface, later it shifts into the space between the two lobes into which the terminal segment soon divides, Fig. 81. — Embryo of Astacus fluviatilis with the rudiments of all the limbs (Fig. 83, p. I(i3) (after Reichenbach, from Lang's Text-book). The thoraco-abdoinen, which in reality curves forward ventrally, is cut off and laid back, ad, antennal gland ; ab, abdomen ; t^, first pair of ambulatory limbs (chelate feet) ; t^-tg, second to fifth pair of ambulatory limbs ; T, telson. and through this space gradually reaches the later ventral side, and, in this way, its final position. Immediately behind the thoraco-abdominal rudiment, at the point where this is connected with the rest of tlie body, in later stages, an FORMATION OF THE ORGANS. 159 accumulation of mesoderm-cells is discerned, forming the first rudi- ment of the heart (Figs. 63, 64, h, pp. 131, 132). In early stages the thoraco-abdominal rudiment is surrounded by a clear area bounded externally by closely contiguous blastoderm-cells. The latter soon become raised so as to form a fold which is specially distinct in its lateral parts ; this is the first rudiment of the thoracic shield (Fig. 80, is). The clear area thus referred to is the rudiment of the branchial cavity. A remarkable structure lying like a dorsal disc on the embryo and called the *'carapax" has Ijeen described in the early stages of Atycphyra by Ischikawa. The development of the Decapoda is principally characterised by the fact that the whole mass of food-yolk is confined to the anterior part of the dorsal side of the body, while the thoraco-abdominal rudiment is free from yolk. Even in late stages, in which the form of the hatching animal is already quite distinct, the cephalo-thoracic portion appears swollen up ])y food-yolk luitil almost spherical. Formation of the Organs. Our knowledge of the development of the different organs in the Crustacean embryo is still somewhat limited. The Decapoda are in this respect best known to us from the researches of Bobrbtzky (Xo. 41), Eeichenbach (Nos. 64, 65), and Kingsley (No. 52-55). Besides these we have the observations of Nusbaum (No. 39) on Ml/sis, of BoBRETZKY (No. 80) and Nusbaum (No. 85) on Oniscus, of CtRobbbn (Nos. 11 and 21) on Moina and Cetochilus, of Claus (Nos. 8 and 9) on Branchipus, Ajms, and others. A. External Integument. As the superficial ectoderm of the embryo yields on its external surface the chitinous skeleton of the larva, it gradually acquires the character of the ht/podermis or the matrical layer of this skeleton. It has been recently proved by T. Tullberg that, in the Lobster, the origin of this chitinous skeleton may be traced back to a direct transformation (chitinisation) of the body of the cell. It is an interesting fact that the hypodermal cells not only are able to change into chitin at their external ends, which are directed to the body surface, but that occasionally even their basal portions are similarly modified. Thus Eeichenbach observed that, in Astacus, single hypodermal cells lengthen iiiAvards and grow out into chitinised strands and pillars which function partly as supports for the carapace, and partly as points of attachment for the groups of muscles. In IGO CRUSTACEA. individual cases it is often impossible to distinguish between these ectodermal ingrowths and true connective tissue. This inner chitinous supporting tissue belonging to the ectoderm was found by Claus (No. 9) ver}' richly developed in Branchipus. B. Endoskeleton. A further development of internal chitinous structures is brought about by infolding and invagination of the external integument. In this way are developed those tubular chitinised ingrowths which, as affording attachment for the more important muscles, have been called chitinous tendons, and some of which even, as was proved for the mandiljular muscle of Astacus (Baur), are renewed at ecdysis. A striking development of chitinous tendons of this kind, formed from ectodermal invaginations (Eeichenbach), is found, as is well known, in the penultimate joint of the pincer in Astacus fliiviatilis. The inner sternal skeleton (endophragmal system) which bridges over the thoracic ganglia in Astacus develops in a similar manner, as a series of invaginations of the external integument (Bobretzky, No. 41), by a process of infolding of the inner wall of the branchial cavity and of the sternal surface of the thorax. Nusbaum (No, 85) was able to observe in Oniscus the origin from paired lateral invagi- nations of a similar chitinous diaphragm covering over the chain of ganglia in the thorax. A hemispherical, chitinous articular fold also arises as an ectodermal invagination in connection with the movable eye of the Cladocera and of most Branchiopoda (Grobben). 0. Nervous System. Although probably belonging to a common rudiment,* the supra- oesophageal ganglion (brain) and the vential chain of ganglia must be separately dealt with. The whole of the central nervous system arises as an ectodermal thickening. Even in early stages, paired ectodermal thickenings can be recognised on the inner side of the * Most authors agree that, in the Crustacea, the rudiment of the brain, from its lirst appearance, is connected with the primitive swellings of the ventral chain of ganglia by means of paired ectodermal thickenings (rudiments of the oesophageal commissures). This view, however, has been opposed. UiiBANOWicz, for example, found that in Cyc/ops (No. 23) the brain and a suboesopliageal ganglion originate independently, and only become connected later liy the development of the oesophageal commissures. Tliis observation cannot, however, be considered as affording direct proof of Kleinenberg's views as to the original independence of the rudiments of these two parts of the central nervous system (Vol. i., p. 288), for it is easy to understand that the rudiments of the more massive parts of tlie central nervous system should become earlier perceptible as ectodermal thickenings, while those of the more delicate parts {e.g., of the oesophageal commissures) are only visible at a later stage. NERVOUS SYSTEM. 161 rudiments of the limbs ; these represent the rudiments of the paired ventral ganglia belonging to the individual segments. The consecutive pairs of ganglia are, however, connected by thickened ectoderm-bands, the rudiments of the longitudinal commissures, so that we may regard two longitudinal ectodermal swellings (the " PrimitivwiiJste" Hatschek, Fig. 82, pic) as the first rudiment of the ventral chain of ganglia ; these show segmental swellings (rudiments of ganglia), and are separated by the primitive groove {py). In the later stages of development (Fig. 82 B), in the region of the primitive swellings, the ectoderm is seen to be composed of several layers, an outer one, which noAV is changed into the hypodermis {h) of this region, being separated from the inner layers. The latter now, as lateral A ft" Srri 3 Fig. 82. — Development of the ventral chain of ganglia in AstncMn fluviatilis (after Reichenbach). J, cross section through the ttandibular segment of an embrj-o in which the maxillipedes have already appeared. B, cross section through the ganglionic rudiment in a maxillary segment of an embryo in which the maxillipedes have already developed, a", cross section of the second antennae ; bi, connective tissue covering over the inner side of the ganglionic rudiment; cc, ectoderm; en, entoderm; /, bundle of fibres in cross section; gi, ' large ganglion-cells; h, hypodermis; m, invaginated median strand of the ganglionic rudiment; ins, mesoderm ; pr, primitive groove ; pw, primitive swelling of the ganglionic rudiments ; s, lateral strand ; sm, secondary mesoderm. strands (s) represent the rudiment of the ventral cord. Reichenbach, on whose description of the development of the nervous system in Astacus our account chiefly rests, was able to prove that, in the formation of each pair of ganglia of the ventral cord, there enters, besides the corresponding portion of the lateral strands, a median invagination (m) ; this is to be traced back to the primitive groove, and is known as the median strand. This agrees with the discoveries of Hatschek in connection with the origin of the ventral chain of ganglia in the Insecta, M 162 CRUSTACEA. The lateral strands are originally composed of simple embryonic cells. In later stages, however (Fig. 82 B) their structure is found to lie more complicated, a cross section revealing three constituent parts. The first commencements of the formation of nerve-fibres (/) can soon be made out in the innermost (or basal) portion ; these run as two longitudinal bundles below the lateral strands, and are connected with very fine jjrocesses of those cells of the strands which become changed into ganglion cells. Besides these paired bundles of fibres, there is, in the rudiment of every pair of ganglia, an unpaired mass of nerve-fibres which perhaps arises from the median strand and gives rise to the transverse commissures. The lateral strands at an early stage become invested with a layer of mesodermal tissue ; this covering, according to Reichenbach, repre- sents the neurilemma, and penetrates not only into the ganglia, but even into the central mass of fibres. The appearance of masses of fibres at tlie inner or basal side of the lateral strands has probably the significance of an onto- genetic recapitulation of a primitive condition, in wliich the whole nervous system was an epithelial structure, with the masses of fibres developed on its inner or basal side. Even in quite early stages, Reichenbach coiild distinguish, in the rudiments of the ganglia, larger and smaller cells of varying histological character. This distinction is also evident in the fully developed condition. The larger elements (Fig. 82 B, g) give rise to the so-called large ganglion cells in the central nervous system of Astacus. Similar large cells were observed by Nusbaum even in early stages in Mysis. In the later stages, massive accumulations of pigment have occasionally been found in the ventral ganglia ; these are probably deposited in mesoderm-cells. Instances of this are to be found in the ganglion which is connected with the sixth pair of appendages of Crangon (Kingsley), and in the thoracic ganglia of Mysis (Nusbaum). Reichenbach's view as to the participation of a median invagination in the formation of the chain of ganglia has since received onl}" jiartial confirmation. Nusbaum, indeed, observed it in Mysis, and Gkobben thought that it could lie assumed for Moina. Clau.s, however, denied that a median invagination took part in the formation of the ventral cord in BrnncJiipus. On the other hand, Nusbaum recently recognised the presence of tlie median strand in Isopoda {Oniscus, No. 39), in which group Bobeetzky (No. SO) and Bullae (No. 81) described the origin of the ventral cord from an unpaired thickening which only at a later stage divided into symmetrical halves. With regard to the development of the 2^eripheral nerves, Reichenbach (No. 65) and Glaus (No. 9) have shown that it is probable that these do not originate as outgrowths from the rudiments of the central nervous system, but that they appear as distinct ectodermal thickenings at a time when the whole nervous system is still connected with tliat layer. The transverse commissures which are doul)le in each segment in Branchipus arise in a similar way, according to Glaus (No. 9). In tracing the development of tlie hratii or supraoesoplwgeal ganglion, we must first study Eeichenbach's minute descriptions of its origin in Astacus (No. 65). According to this author, the whole central nervous system of the pre-oral part of the body arises in the form of three pairs of ganglia, equivalent to one another and Ijclong- ing to three separate body-segments (Fig. 83). The most anterior NERVOUS SYSTEM. 163 of these ganglia, which develops in the proximal portion of the eye-stalk, yields the optic ganglion (o', o"), while the remaining two belong to the first and second pairs of antennae and enter into the formation of the brain proper or siipraoesophageal ganglion. Of these last, that lying in the segment of the first antenna (antennule) very soon becomes divided by a transverse constriction into a pair of oanglia (a, h), the anterior of which (a) we will distinguish by Packard's name of procerebrum, while the posterior, since it gives off the nerve to the first antenna, has been called the antenmdar fit/L mx"- Fig. S3.— Embryo of Astacusfluviatilis with the rudiments of all the limbs (after Reichenbach). a. anterior, b, middle, c, posterior portion of the rudiment of the brain ; ab, abdomen ; an', first antenna; an", second antenna; d, mandibular ganglion; md, mandible; mx', tirsst maxilla ; mx'', second maxilla ; mf, mf", mf", first, second, and third maxillipedes ; o, ruJi- nieot of the compound eye; o', the part of the optic ganglion which has arisen from the optic invagination ; o", inner part of the optic ganglion ; ol, upper lip ; r, rudiment of the rostrum ; t, telson ; ih, fold of the thoracic shield ; I-F, first live pairs of ambulatory limbs. ). It should be mentioned that Reicbenbach believed that he observed, in late stages, a similar transverse division of the posterior pair of ganglia which give off the nerves to the second antennae, and are therefore called the antennal gaiKjlia {c) ; this latter division, however, is not so striking, and also not of the same significance as the division which takes place in early stages between the procerebrum and the antennular ganglion. It is an important fact that the pairs of ganglia just described 164 CRUSTACEA. show, according to Keiciienbach, in their development, great agree- ment both inter se and with the ganglia of the ventral cord. In each of these pairs of ganglia we can distinguish lateral strands and a median strand ; the lateral strands, in cross section, are seen to be broken up into three parts, as are the ganglia of the ventral cord. The median strand, however, varies in different regions. In the region of the optic ganglion, the two halves of the median strand shift far apart, and enter separately into connection with the corre- sponding ganglia. In the region of the procerebrum and the antennular ganglion, on the other hand, is found that median invagination of this strand which doubtless leads to the develop- ment of the commissural portions of the brain. In the region of the antennal ganglion, again, there is no median invagination. Eeichenbach believes that this invagination has shifted forwards and is represented by that established between the antennular ganglia. If, however, we l^elieve that the transverse commissure between the antennal ganglia was originally post-oral, and is perhaps still to be looked for in such a position (Claus, No. 78), we 'shall not be surprised at the absence of the median invagination between these ganglia. In later stages, the invaginations of the median strand are no longer distinct from one another in the region of the procerebrum and the antennular ganglion, a closer union between the different parts of the brain being then generally apparent. The procerebrum, according to Eeichenbach, gives origin principally to the " anterior cerebral swelling," while the antennular ganglion is connected with the development of the "lateral swellings" (Krieger,. Dietl). These observations of Eeichenbach are to a certain extent in agreement with those of Ivingsley, who found in Craiigon, apart from the optic ganglia, three consecutive pairs of ganglia taking part in the formation of the brain. The most anterior of these (the procerebrum) Kingsley, however, regarded from the first as an independent structure, it alone being originally preoral and homo- logous with the supraoesophageal ganglion of the Annelida. The two pairs of ganglia which folloAV (the antennular and antennal ganglia) are, when first developed, post-oral in position, and must- thus be considered as "anc-lia of the ventral cord drawn into the O o cephalic region. The above oliservations lead us uj) to the question of the primary segmentation of the pre-ora] portion of tlie head iu the Crustacea. Reichenbach, witli whom also Nusbaum (No. 39) is in essential agreement, has been led by his- NERVOUS SYSTEM. 165 ontogenetic researches to assume for tliis region of the body three component parts, homonomous with the other body-segments; these are. the optic, the antennnhir, and the antennal segments. In the optic segment, the optic ganglia would represent the segmental pair of ganglia, while the general position of tlie parts in the Astacus embryo incline Reichenback to return to the old view of Milnk-Edw'akds, according to which the eye-stalks represent tlie limbs of this segment. This last view, which subsequently found supporters in Huxley and Spence Bate,* has, however — and, as it appears to us, rightly — been disproved by Glaus and Fr. Muller, by reference to the ontogeny of the stalked eye of the Phyllopodan larva {Brancliipus) and of the Zoaea of Lucifer, which indicates that the eye-stalks are to be regarded as secondarily abstricted lateral portions of the head which have become indepen- dently movable, while the o}itic-ganglion, as a part of the brain shifted anteriorly, also attains only a secondary independ- ence. These conclusions remove all ground for assuming tlie presence of an independent optic segment. With regard to the segment of the body corresponding to the second 2^(ii>' of antennae, there can hardly be any doubt that we here have to do with an origin- ally post-oral body-segment, i.e., with a true trunk-segment, wliich only second- arily underwent displacement forwards, and thus attained a closer union with the pre-oral parts of the head. This view is supported by the changes brought about during embryonic development in the relative positions of the mouth and the second antennae (p. 157), and above all by the condition ot the nervous system, in which are to be found all transitions between the independent development of the pair of ganglia on this segment and their close fusion with the cephalic mass. It has been known, since the observations of Zaddach, that, in Apus, the origin of the pairs of antennal nerves is to be sought post-orally in the oesophageal com- missures, and later researches (Pelseneer, No. 14) showed the ganglia at this point te-be connected by means of a post-oral transverse commissure, although this has to some extent been otherwise explained. Similar conditions are found in other Phyllopoda. Thus, in Glaus' drawing of the brain of a Cladoceran (Fig. 84), three sections can be distinguished, only the two anterior being pre-oral in position. The most anterior section (c^, corresponding to Packard'.s * [See also C. Heris.st, Uber d. Regeneration v. antennenahnlichen Organen an Stelle v. Augen, Arcliiv. f. Entwlckl. Mcchanik, Bd. ii., 1895-96, p. 643 ; and B. Hofer, Bin Krebs mit eiuer Extremitat statt eines Stielauges, Verh. Dent. Zool. GcsclL, 1894, p. 82.— Ed.] Fig. S4. — Ventral aspect of the brain of Daplimasi)niUs{&itexCh!i.vii). ci, anterior, c-, middle, c3, posterior section of the brain ; go, optic ganglion ; n, nerve of the sensory organ of the neck ; ««', nerve of the first antenna; iia", nerve of the second antenna ; a ', second nerve of the second antenna ; sc, oesophageal com- niissure. 166 ■ CBUSTACBA. procerebium) gives otf the nei'ves to the eyes aiid to the frontal and other sensory organs, the second section the nerves to the first pair of antennae (na'), while the posterior section (c^), Avhicli lies on the conrse of the oesophageal commissure {sc) behind the oesopliagus, supplies nerves to the second antenna {na"). Among other Crustacea, the antennal ganglion undergoes more or less displacement forwards along the commissure and a subsequent fusion Avith the brain. The accejitance of this view of the alterations in position is attended with a certain amount of difficulty owing to the course of the transverse commissures (origin- ally running behind the oesophagus) connecting this pair of ganglia. The following alternative is offered us : either to suppose that a secondary pre-oral transverse connection is developed, or to assume that the transverse fibres, after the ganglia have com})leted their wanderings, retain their original course behind the oeso- phagus. Glaus (No. 78) believes that this primary connective retains its original jiosition in the adult, and he would identify as that structure the transverse post-oral connection between the oesophageal commissures, which is found in many Crustacea {Jpseudcs, Stomatopoda, Decapoda), lying in front of the mandibular ganglion. In other cases, this fibrous connective is perhaps fused with the transverse commissure of the mandibular ganglion. The question now arises whether we are justified in considering the part of the brain lying in front of the antennal ganglion as an originally single complex, or whether, in this also, a separation into (two) consecutive segments can be recognised. We must here mention Ray Lankester's theory (No. 15), which distinguished in the Crustacean l.irain an anterior section connected with the optic ganglion under the name of a rcJdcerebrum , this only being enlarged into a si/nccrchrum when the ganglia of two following segments (the antennular and antennal segments) are drawn into it. This view has been accepted by Packard (No. 86), who distinguishes in the brain of Ascllus as sej)arate sections : (1) the optic ganglia, (2) the procerebrum, (3) the antennular ganglia, (4) the antennal ganglia. In this case, we should have to find the homologue of the Annelidan brain, which develops out of the neural plate, in the pro- cerebrum alone, while the optic ganglia would be a secondary portion of the brain* arising during the later development of the paired lateral eyes, and the antennular and antennal ganglia would be ganglia of the ventral cord. This view stands opposed to that of Claus (No. 78), according to which the antennular ganglia, together witli the procerebrum, formed an originally single complex, the primitive brain. This portion, which is to be deduced from the neural plate of the Annelidan larva, contains the ganglia of the former median sensor}' organs (Xauplius eye, frontal organ) and the anterior antennae, which are morphologically to be homologised with the Annelidan palps already con- nected with the neural plate. Tliis latter view would be supported liy the oljservations of Reichenbach, according to which the rudiment of the corre- sjionding part of the brain is originally to be found as a single complex at the Ijases of the first antennae, and only later breaks up into two pairs of ganglia. A certain amount of support is also afforded by the peculiar structure of the first antennae, which, as carriers of important sensory organs, do not develop in accordance with the fundamental type of the Crustacean limb, a point specially emphasised by Glaus and Boas. Such a heteromorphous structure of the antennule might, indeed, lie secondarily acquired, and might be accounted for liy its jihysiological significance already mentioned, as well as * A view first enunciated liy Hatschek (Beitriige zur Entwickelungsgeschichte der Lepidopteren), and later accepted by Gkobben for the Crustacea. SENSORY ORGANS. 167 Ly its position at the anterior end of the body. The views of Ray Lankester and Packakd are supported most strongly by Kingsley's statements con- cerning Crangon (No. .^5) : not only was the procerebrura observed to originate independently of the antennnlar ganglion, but the antennules and their pair of ganglia were found in a distinctly post-oral position. In accepting this view we should have, indeed, to assume, with Ray Lankester, a backward wandering of the mouth. "VVe must, however, await further researches as to the structure of the Crustacean brain, and above all, as to the development of the whole region now under consideration, before forming a decisive judgment. It is evident that, in a discussion as to the primary segmentation of the ante- rior region of the Crustacean body, the question of the morphological value of the first antenna comes to the front. Two alternatives are presented to us : we nnist either regard it as a true limb, although somewhat modified in shape, or else, Avith Boas, deny that it has this significance, and consider it only as a stalked sensory organ (similar to the stalked eyes). Only in the latter case can we regard it as homologous with the primary cephalic tentacle of the Annelida. We, howcA^er, see many reasons for regarding the first antenna as a true body- appendage. We have only to recall the similarity between its position and development and those of the latter in the embryo, and its use for swimming purposes in the Nauplius stage and in many Entomostraca, in which tlie first antenna is sometimes diverted to other purposes {e.g., climbing and sucking). It is only in the higher Crustacea that this limb is distinctly set apart as a sensory organ. If these considerations incline us to place the first antenna in the series of true trunk-limbs, we then have to ask whether the vestiges of the primarj' cephalic tentacle, so common among the Annelida, are not to be sought in some other structure. It is not difficult to assume this significance for the so-called frontal sensory orgnns (Fig. 123, fs, p. 269) found in the young stages of many Crustacea as paired peg-like or filamentous processes innervated from the procerebrum. This view gains in probability by a comjiarison with PcripaUis, in whose embryos similar blunt processes have been observed, while the antennae of Peripatus, according to their development and their relation to the coelomic sacs, must be considered as modified trunk-limbs. If we held this view, to which, naturally, we can only ascribe a hj^pothetical value, we should be led to distinguish, with Ray Lankester, three sections in the anterior part of the Crustacean body \\hich contains the brain. There would be one actual primary section, originally the only pre-oral section of the body, with the procerebrum, the eyes, and the frontal sensory organs, and two sections following posteriorly, trunk-segments drawn into the head (antennular and antennal segments), for which we must assume an originally jiost-oral position. We must, however, once more emphasise the fact, before accepting any one of these views, that in discussing these questions we are dealing entirely with hypothetical matters. D. Sensory Organs. Of the details of the development of the unpaired triple Nauplius or Entomostracan eye* nothing is as yet known, but mention should * According to Clai'.s {Kaiserl. Akad. JVissensch. Anz. Wien, 1891), the Nauplius eye is composed of three inverted cup-shaped eyes, in which the nerves enter the retinal cells from the side turned away from, while the rods are directed towards, the pigment cups. A certain similarity with the median eye of the Arachnoida is thus brought about. [Bernard (The Apodidae) claims to have discovered paired rudiments for the Nauplius eye oi Apus. — Ed.] 168 CRUSTACEA. be made of the observations of Leydig and Grobben, according to which this eye has a paired rudiment. According to Urbanowicz (No. 23), the eye in Cyclops is formed "of three ectoderm-cells, each of which secretes pigment and becomes a refractive sphere." The development of the paired compound eye has been best studied in the Decapoda (Bobretzky, No. 41 ; Eeichenbach, No. 65 ; KiNGSLEY, No. 52 ; Herrick, Nos. 48 and 49 ; and Parker, No. 62). It has also been observed in Mysis (Nusbadm, No. 39}, Para- podopsis (BouTCHiNSKY, No. 37a), and BrancJiipus (Glaus, Nos. 8 and 9). The description of the development of the compound eye cannot be separated from that of the optic ganglion. The simplest example of the development of a compound eye is that of Branchip7is. The rudiment of the compound eye, as well as of the optic ganglion, can here be traced back to a pad-like growth of hypodermis, the superficial parts of which become trans- formed into the eye, while the deeper parts contain the material for the optic ganglion, which is connected with the brain. The several layers of cells which represent the rudiment of the eye, and which must be regarded as a simple thickening of the hypodermis, soon show an arrangement of the elements into a superficial layer (which yields the corneal cuticle and the crystalline cones) and a deeper pigmented layer for the formation of the retinulae, the latter layer being connected with the rudiment of the optic ganglion by fibrous strands. In the lateral parts of the whole rudiment there soon occurs a histological difi'erentiation of the optic ganglion and of the ommatidia which compose the eye, whereas, in the anterior and more median portions, a proliferating hypodermis of embryonic character (Fig. 85) is retained until a later stage ; this constantly yields new elements for the enlargement of the whole rudiment. Strictly speaking, two budding zones (/.•', A:"), which arc distinct from, but in contact with one another, can be distinguished at this point ; one of these (A:"), by the production of new ommatidia, adds to the size of the eye itself, while the other and more proximal (A;') yields the elements corresponding to the optic ganglion. During these ontogenetic processes, the movable eye-stalks have arisen by simple outgrowth from the lateral parts of the head. The development of the eye in the Schizopoda and the Decapoda is very similar to that in Brancliipus. Here also the compound eye arises from a thickening of the hypodermis, which from the first is in close connection with the rudiment of the optic ganglion. SENSORY ORGANS. 169 Even in early stages, the outer and anterior part can be distinguished as the rudiment of the eye (Fig. 80 A, p. 157, and Fig. 83, o, p. 163), and the inner, posterior part as that of the optic ganglion (Figs. 80, go, and 83, o', o"). The hitter is thus, from the very first, in close contact, not only with the rudiment of the eye, but, proximally, with that of the supraoesophageal ganglion. The rudiment of the eye is thus only a part of the ectoderm (Fig. 86 A) which becomes multilaminar and produces from its superficial layers the cells of the cornea and crystalline cones, while the lower layers give rise to the retinulae and the pigment-cells. A basal membrane secreted on the // /[" inner surface of this hypodermis {Fig. 86, mh) yields the memhrana limi- tans which bounds the eye on the side of the optic gan- glion. Mesoderm- elements are de- posited on this membrane from within, and these, in My sis, yield the pigment of the innermost of the three pigment-layers. In Mysis, according to Nusbaum (No. 39), differentiation occurs at the time when the stalked eye rises from its substratum, the first differ- entiation of the ommatidia appearing in that dorsally curved lamella which, by its curvature, causes the eye to separate from the food-yolk. In this lamella, a very regular arrangement of the cells both into horizontal layers and into vertical columns takes place early. The horizontal lamination separates the corneal cells from those of the crystalline cones, etc. The vertical arrangement gives rise to two alternating columns of cells, which we may distinguish as ommateal and intermediate columns. In each ommateal column, the most superficial transverse layer is formed by two corneal cells (Claus, Ko. 78) destined to secrete the corneal lens ; below this comes the layer of the crystalline cone, also consisting of two cells. In accord- ance with this number of cells, Grenacher was able to prove the origin of the crystalline cone from two separate segments, the Fig. 85. — Left eye of a young Bmnchipus, seen from the ventral side (after Claus). go, optic ganglion ; k', budding zone for the optic ganglion ; k", buddiug zone for the onnnatidia ; m, muscle of the eye. 170 CRUSTACEA. boundaries of wliich can be recognised in the adult crystalline cone. The cells in the lower layers of the ommateal column no doubt give rise to the elements of the retinulae, but Nusbaum is inclined to derive these from the intermediate columns. Claus derives from these columns the anterior and posterior pigment -cells which surround the crystalline cone in Mysls. The observation of these processes is rendered difficult by the early deposition of pigment w^hich is to be found within the rudiment of the eye in two layers, and in a third mesodermal laver below it. A ■G^"^°n°e^,^ce: Fig. 86.— Sections tlirougli the compound eye of the American Lobster {Hotnarus americanxis) in three stages of development (after Parker). A, transverse section through the optic lobe in a young stage. D, older stage when the optic rudiment (r) and the optic ganglion (go) are beginning to be separated by the development of a basal membrane {mh). C, cross section of still older stage, c, rudiment of the brain ; go, optic ganglion ; mh, basal membrane ; r, rudiment of the eye (retiuogen). The account here given of the ontogeny of the eye in Mysis resembles that given by Herrick (No. 48) in connection with Alpheus, and Parker (No. 62) for Homarus. The optic lolies here develop, by proliferation of the ectoderm, into a nuiltilaniiuar rudiment (Fig. 86 A). (Herkick considers that indifterent elements out of the yolk also contribute to their formation). There then takes place a separation into a superficial and a lower part (Fig. 86 B and C). The first layer {retiuogen) becomes the rudiment of the eye (r), while that of the optic ganglion (go) arises from the cell-mass of the lower layer {gangliogen) . In later stages the two layers are separated by a cuticular basal membrane, through which strands of nerve-fibres pass. In the rudiment of the eye itself, which we have called the retiuogen, the separate ommatidia are said to develop (Herrick), the corneal cells forming into groups of two in the most superficial layer, and SENSORY ORGANS. 171 the crystalline cone-cells into groups of four in the subjacent layer, while, in the lowermost layer, the cells of the retinulae form bundles of seven each, these bundles reaching and enclosing the lowermost, pointed ends of the cells of the crystalline cone. The separate rudiments of the ommatidia are divided by numerous undifferentiated ectoderm -cells. In this last point the accounts of Paekeu and Hereick differ. In Hornarv.s it is said (Parkek) that the separate rudiments of the ommatidia lie close to one another, and are not separated by any kind of intermediate pillars. Three layers can be distinguished ; from the outermost arise the corneal hypodermis- cells and the anterior pigment-cells (distal retinulae), from the median layer the cells of the crystalline cone, and from the lowest the actual retinulae.* The paired lateral ej^e of the Isopoda also develops in a similar way. In Bullar's account of Cymothoa (No. 81) the rudiment of the eye stands in close connection with that of the optic ganglion. The two proceed from one and the same ectodermal thickening. While the inner layers of this thickening become detached for the formation of the optic ganglion which is connected with the brain, a suiaerficial hj'podermal thickening becomes marked off by a pigmented basal membrane. This hj'podermal thickening represents the rudiment of the eye, in which the separate ommatidia seemed to be marked off from one another by a strongly pigmented mantle of cells. The details of the development of the ommatidia were not followed in this case. In the cases we have hitherto described, the development of the eye is comparatively simple, but in Astacus (Eeichenbach, ^o. C5) and Crangon (Kingsley, Ko. 52) complication occurs through an invagination Avhich forms on the boundary between the rudiment of the eye and that of the gangHon. This invagination, which, according to Eeichenbach, is replaced at a certain stage of develop- ment by a more solid ingrowth, gives rise to the opiic fold between the rudiment of the eye and the gangliogen ; in this fold an inner and an outer layer are recognisable. Although Eeichenbach did not closely follow the future fate of the optic fold, it appeared probable to him that the outer layer enters into the formation of the eye and yields the retinular layer, while the inner fold enters into tlie formation of the optic ganglion. It was pointed out, especially by Carriere (iSTo. 44), that in such a method of develop- ment of the retinular laj^er the position of its cells is reversed, their basal ends being directed towards the cells of the crystalline cone, and their upper ends towards the rudiments of the ganglion, and that we must thus assume a later re-arrangement in the retinular area, such as takes place among the Araneae, but has not yet been observed among the Crustacea. The suggestion made by Patten appears to us probable, viz. — that the optic fold has nothing to do with the formation of the eye, but merely yields the material *' [Cy. Parker, G. H., Retina and Optic Ganglia in Decapoda, Mitthdl. Stat. Zool. Neapel., xii., 1895.] 172 CRUSTACEA. for the enlargement of the optic ganglion. If so, it -would correspond to the proximal budding point (for the enlargement of the ganglion, Fig. 85, //) in the eye-stalk of Branchipu^: This last view has recently been accepted by Kingsley also, who originally thought that the layer of the crystalline cones and the retinular layer arose from the outer wall of the ojjtic invagination. The above view receives its chief support, as Cakriere pointed out, from the position of a pigmented layer of mesoderm -cells, which, according to Reichenbach, develops between the outer wall of the optic fold and its crystalline cone-layer, and which is, nevertheless, evidently identical with the layer of pigment-cells below the basal membrane of the eye, described above (p. 169) for Mysis, With regard to the significance of the separate parts of the ommatidium, as to which Grenacher and Patten have recently taken opposite views, attention should be drawn to Parker's observations, wliich revealed a con- nection of the retinular cells with fine nerve -fibres, while the crystalline cone-cells, which doubtless reach to the basal membrane, end at that point. This is in agreement with the view of Grenacher, who saw in the retinular cells the percipient elements, while Pattex considered that the crystalline cone-cells known as retinophorae were the elements connected with the nerves. In the develoj^ment of the compound eye of the Cladocera, which was care- fully described by Grobben (No. 11), sjiecial interest is awakened by the formation of an integumental fold which grows over the eye, cutting oflf a hemispherical precorneal space. The movement of the sunken eye is thus assisted. Similar conditions are found in Apus, Esthcria, Limnadia. and Limnetis. The compound eyes, in these forms and perhaps also in the Ostracoda, may be regarded as movable stalked eyes with degenerated stalks which have sunk below the surface. Where, as in the Cladocera, an unp)aired compound eye is found, this must be considered to have arisen by the fusion of paired rudiments. An auditory organ was observed by Reichenbach (Xo. 65) in Asfacus, as a dorsal invagination in the basal joint of the antennule. Even in early stages, the ectodermal sensory epithelium, which probably yields the auditory ridges, is distinguished by the multi- laminar and regular arrangement of its cells. Nusbaum (Xo. 39) similarly was able to observe the origin of the auditory sac in Mysis in the endopodite of the last pair of pleopoda out of an ectodermal invagination. ^o*^ E. GUIs. The branchiae first appear as simple outgrowths of the superficial body- epithelium (ectoderm), within which lacunar blood-spaces traversed by connective tissue-strands develop (Reichenbach). We may with some probability regard all the branchial sacs or tubes which belong to the outer side of the basal joints of the limbs, INTESTINAL (.'ANAL. 173 and can tlius be described as eiipodicd giUs (p. 195), as homologous structures throughout the whole of the class Crustacea. We may also, perhaps, derive them from the branchial tubes of the Annelida, (hi the other hand, it must be pointed out that branchial outgrowths develop at other points also, e.g., on the exopodite of the pleopoda (SquiUa), or on the endopodite of the same limbs (Sii-iella), as dorsal appendages in certain Ostracoda (Astei'ojye), as mantle-folds in the Balanidae, etc. These naturally cannot be homologised with the epipodial gills. The epipodial branchial sacs present in a single row in the Phyllopoda are replaced in the Decapoda by three rows of branched tubes, which, according to their exact points of attach- ment, are distinguished by Huxley as jyodohrandiiae, arthrohranchiae, and pleurohrancliiae. Instead of these, we find in the Euphauaidae and Lojjhogasti'idae only a single row of dendriform branched tubes, so that Claus raises the question whether the three rows of gills of the Decapoda may not be derived from the principal branches of the Schizopodan gill which have shifted apart. F. Intestinal Canal. The intestinal canal arises here, as in most groups of animals, from three separate rudiments, the fore-gut and hind-gut arising as ectodermal invaginations (the stomodaeum and the proctodaeum), while the mid-gut is formed from the cells of the entoderm. "Whereas the two former approach their adult form by a series of comparatively simple changes, the development of the latter is brought about b}"- far more radical ontogenetic processes, on account of disturbances due to the presence of the food-yolk, these processes also varying in the different orders of Crustacea. There is some variation among the sub-groups of tlie Crustacea witli regard to the time of the appearance of the fore- and hind-guts. In the Entomostraca, the rudiment of the fore-giit, as a rule, appears first. This is also the case in Asellus, Gammarus, and Mijsis, while, in Oniscus, the hind-gut develops first. In the Decapoda, the proctodaeal invagination usually appears first, a fact connected with the early development of the abdomen. The position of this invagination with relation to the closed blastopore is of importance. In Moina, according to GnoBiiEN, the point at which the blastopore closes corresponds to the oral aperture, while in the Decapoda, it always lies close to the future anal orifice. According to Reiciienbach, in Astacus, it lies some- what behind the point at which the anal aperture is forming, as is also the case in Atycphyra (Ischikawa). The opposite, according to Lebedin.sky, holds true for Eriijliia. Here the proctodaeal invagination develops behind the blastopore. Kixgsley, on the contrary, believes that, in Crangon, the procto- daeal invagination corresponds exactly to the jioint at which the blastopore 174 CRUSTACEA. closed, this being also maintained by Bobretzky for Astacus and agreeing with P. Mayer's account of Eupagurus. Hoek found that, in the free-living Copeiwda, tlie point at which the blastopore had closed corresponded to the position of the future anal aperture, and Nassonow made the same observation in Balanus, so that, as a rule in the Crustacea, the position of the blastopore may be assumed to be in the neighbourhood of the anal aperture (p. 141). The time at which the three rudiments unite to form a single canal varies according to the stage of development at which the larva leaves the egg. In the free-living Cojjepoda (Cetochilus), the intestinal canal is completed at an early stage, while in the Decapoda, communication is established between the fore- and hind-guts and the mid-gut usually at a later stage. The hind-gut here appears to show its adult structure sooner than the fore-gut. We have as yet very little accurate information as to the manner in which the mid-gut develops in the Entomostraca. In Moina, the entoderm-cells first form a solid strand, a cross section of which reveals a radial arrangement of the cells, hut no lumen (Grobbex). In Cetochilus, on the contrary, the entoderm-sac which is formed by invagination seems to be transformed direct into the mid-gut. In many other Entomostraca, the mid-gut rudiment can be recognised as a central mass of cells filled with food-yolk {Balaniis, Lang, !N"assonow). In later stages, the nuclei of the entoderm-cells, with the protoplasm which surrounds them, migrate to the surface, and as the food-yolk is gradually assimilated, the cavity of the mid-gut appears within. This is also the case in Palaemrm (p. 132). In the parasitic Copepoda also, the mid-gut, according to Van Beneden (No. 17), appears to develop in this way. The mid-gut, filled with food-yolk, is connected at its anterior end with the stomodaeal invagination, and at its posterior end with the proctodaeal invHgi- nation. The development of the mid-gut is best known in the Decapoda. In Astacus, Avhere the cells of the entoderm-vesicle absorb the whole of the food-yolk, without thereby disturbing tlie conformation of the vesicle, the epithelium of the mid-gut arises by the shifting of the nuclei to the surface of the yolk-liearing entoderm-pyraraids, a separation of the cells from the food-yolk there taking place ; as the entoderm-cells increase in number they become arranged into an epithelium which now covers the surface of the disappearing yolk (p. 130). At the same time, the whole rudiment of the mid-gut, by constriction from without, assumes a lobate form. Paired anterior lobes form wliich become connected with the median rudiment of the mid-gut, at whose posterior dorsal portion can be recognised another swelling, the rudiment of the dorsal caecum of the mid-gut. The development of the mid-gut epithelium just described first takes INTESTINAL CANAL. 175 place at the point where the entodermvesicle and the rudiment of the hind-gut are in contact, and where soon an entoderinal epithehal plate can be seen (Fig. 63 B, ep, p. 131). The same has been ■observed at the point from Avhich the formation of the hepatic tubes commences. Separate centres of epithelial formation are to be found at the anterior, lateral and posterior parts of the mid-gut correspond- ing to the anterior, lateral, and posterior hepatic lobes of the adult animal ; at these points, the epithelium soon grows out to form the primary hepatic tubes. The rudiment of the posterior pair of tubes seems from the first to be connected with the entoderm-plate described above. As the epithelium continues to grow over the rest of the entoderm-vesicle, the central part of the alimentary canal arises, receiving the efferent ducts of the mid-giit gland (liver) ; this central part, in the adult, is of no great extent. Its musculature arises by the deposition of mesodermal elements. The stomodaeal invagination soon becomes divided into a narrower oesophageal portion and an inner and more swollen portion, the rudiment of the so-called stomach. In the latter can be recognised the rudiments of the tooth-plates as epithelial thickenings, and those of the gastrolith-sac as two diverticula diverging on the ventral side. The young Astacus hatches with two completely developed gastroliths (Reichenbach). The mid-gut becomes connected with the fore- and hind-guts only at a late stage. The mid-gut develops in a similar manner in those Decapoda in which tlie entoderm-sac does not retain its continuity, but breaks up into single cell- elements, which become distributed in the food-yolk [Palaemon, Eupagurus, ErijJhia, Atyc2}hyra, Crangon, etc.). In these also, the entodermal elements finally rise to the surface, and yield the mid-gut in the way described abo\e. Here also the first appearance of this epithelium was observed in contact with the blind inner end of the proctodaeal invagination (Fig. 64 6', ej)). Three pairs of originally distinct hepatic rudiments, however, appear to be added to it (Crangon, Kingsley). The formation of the mid-frut in the Arthrostraca differs from the type above described as occurring in the Decapoda in that here the mid-gut epithelium does not proceed from elements scattered in the food-yolk, but from a paired lateral mass of cells which lies superficially on the yolk and gradually grows round it (p. 139, etc.), while within the yolk, vitellophags are found only in isolated cases {Oniscus, ^STusbaum) ; in other cases {Porcellio, Amphipoda), cell-elements are here altogether wanting. By the gradual circum- crescence of the yolk from the two sides by the paired entoderm- rudiment, the mid-gut vesicle closes, the very large primary hepatic 176 CRUSTACEA. sacs forming by constriction at its sides. These hepatic sacs give rise later, by longitudinal constriction, to four or six tubes. In some cases (Onisnis, Caprella, Sunamphithoe), the formation of the hepatic tubes precedes the development of the mid-gut vesicle. In most cases, the chief part of the entoderm is used up in the forma- tion of the hepatic rudiment, only a small part entering into the formation of the central portion of the intestinal canal. The latter is almost exclusively composed of the fore and hind -guts, only a short tract in the immediate neighbourhood of the openings of the hepatic tubes being entodermal in origin. As diverticula of the posterior section of the mid-gut, there arise in Gammarus (Pereyas- lawzewa) those paired tubes (uri7iary glands), the entodermal nature of which was recognised even in the anatomical researches of Xebeski. The homologising of these urinary tubes Avith the Mal- pighian vessels of the Insecta is on this account inadmissible, the latter belonging to the hind-gut, and therefore having an ectodermal origin. The observations of Bullae (No. 81) on the development of the mid-gut in C'ymothoa are in fairly close agreement with those of Nusbaum for Oiiiscus. There are, however, no so-called yolk-cells here inside the very considerable mass of food-yolk ; the formation of the mid-gut proceeds from the inner cell-layer of the germ-band. Tlie first trace of a separate entodermal rudiment is found in a paired mass of cells lying somewhat behind the stomodaeal invagination ; this mass of cells gives rise to three pairs of liepatic tubes. At a later stage, a layer of cells connected with tlie epithelium of the hepatic tubes grows over the whole of the food-yolk. The rudiment of the mid-gut now consists of the yolk-vesicle thus formed and the hepatic tubes communi- cating with it. The anterior end of the yolk-vesicle is connected with tlie stomodaeal invagination. The proctodaeal invagination, on tlie contrary, is not in contact with the posterior end of the 3'olk-sac, hut runs forward over its dorsal side so as to pass into it quite near the stomodaeal invagination. The yolk-vesicle appears now as a ventral diverticulum of the alimentary canal, hut, as most of it undergoes absorption, it results that here also only quite a small part of the definitive intestinal canal near the points of entrance of the liepatic tubes belongs to the mid-gut {cf. p. 139 on Ligia). The above description of the formation of the mid-gut in the Arthrostraca, founded on the observations of Bullae, Nusbaum, Perf.yaslawzewa, and RossMSKAYA, appHcs, according to Nusbaum, to Mysis also (Figs. 65 and 66, p. 135). The entoderm here originally lies as a mass of cells in the most posterior section of the germ-band (p. 134). The entoderm-cells soon increase in number, and spread out over the whole ventral surface of the embryo. At a later stage they also reach the lateral and dorsal parts, the food-yolk in this way becoming surrounded by a layer of entoderm-cells "While this circum- crescence is going on in the posterior part of the emliryo, the entoderm in the anterior part (behind the mandibular segment) forms two lateral grooves consisting of large granular cells (Fig. 66, I) ; these are the rudiments of the HEART. 177 hejiatic tubes, which become connected Uxter on the ventral side by entodermal •epithelium. Two longitudinal folds arise and fuse M'ith the inwardly curving •edges of the grooves, and thus separate the hepatic tubes from the middle portion of the intestinal canal. At the same time, the he|)atic tubes become divided through longitudinal folding into four secondary tubes in the same way as in the Arthrostraca. It appears that, in ifysis, when the mid-gut vesicle forms, the whole of the food-yolk is not taken up into it, but part of it comes to lie in the cephalic region outside of the intestine, and thus in the body-cavity. The food-j-olk has a similar position in Moina. Or. Heart. In' describing the development of the heart, we must take as our starting point the observations made by Claus (Nos. 8 and 9) on Brancldpns. The somatic layer of the mesoderm here forms a cell- stratum originating on the ventral surface, and now divided into separate segments ; this mesoblast gradually grows upwards under the lateral parts of the integument. The dorsal edge of these growing mesoderm-segments is formed by a single roAV of large cells {card to blasts, Xusbaum, Figs. 87, c, and 88 A, c), which later •assume a crescentic shape (Fig. 88 B, c), so that a channel is now formed on each side. These semilunar channels, meeting and fusing in the middle line, give rise to a dorsal tube (Fig. 88 C, c). The latter is from the very first divided up into separate segments f( chambers) corresponding with the primitive mesoderm-segments ; the boundaries of these chambers persist as the lateral ostia. This origin of the lieart seems distinctly to show that its lumen must be considered as a remains of the primary cleavage-cavity (Butschli, Schimkewitsch). The ostia develop at the boundaries of the mesoderm-segments. Tlie cardio- lilasts become transformed into the muscle-cells of the wall of the heart. During their development, the latter have their lower ends connected with tlie dorsal portion of the intestinal muscle-layer. From this point, a horizontal septum stretches out towards the Ijody-wall ; this is the pericardial diapliragm •(Fig. 88 C, s), which separates an upper portion of the coelom containing the heart from the rest of the body-cavity. This septum is found in all Crustacea. The heart of Oniscus develops in a similar way by the fusing of two grooves which arise from a single row of cells on each side (Nusbaum). The formation of the lieart in the Amphipoda (Pereyaslawzewa, Rossijskaya) also passes through a similar stage. While, however, in Oniscus, the posterior pai'ts of the body develop iirst, and the formation of the heart proceeds gradually from behind forward, in the Amphipoda, the heart begins to form in the middle region of the body. A vascular trunk arises simultaneously in front of the dorsal organ in the same way. The two rudiments only fuse after the dorsal •organ has degenerated, that organ i-etarding the development of the single dorsal tube. In the Amphipoda the fusing of the two grooves takes place in N 178 CRUSTACEA. such a manner that the ventral precedes the dorsal fusion. This leads us to the condition observed in the Decapoda. In Astacus, the first rudiment of the heart is to be recognised as an accumula- tion of mesoderm-cells in the most posterior part of the embryonic disc (Fig. 63- B, h, p. 131), and thus behind the point from which the ventrally curved thoraco-abdominal region arises. In sections it can be seen that these mesoderm- cells unite to form a transverse plate which becomes opposed to the ectoderm on each side. The cavity found between the ectoderm and the plate of cardioblasts is the future lumen of the heart. This plate already shows slight pulsation, in which the ectoderm passively participates, before it curves round dorsally to> form a tube (Reichenbach). A similar condition is described by Lebedinsky for Eriphia (No. 57). The heart of Mysis develops in a manner analogous with that of Oniscus. The heart here arises as a cavity in a kind of dorsal mesentery formed by the fusion of the lateral edges of the somatic mesoderm. Its formation proceeds from behind forward. As the primitive form of Crustacean heart we must assume a long dorsal vessel provided with numerous segmentally- arranged pairs of ostia ; such a heart has been retained in the Branchiopoda (Fig. 95, h, p. 202), and also occurs in the Stomatopoda. The short, sac-like heart of the Copepoda and Cladocera is a degenerate form of this elongate type. Such degeneration may lead in small Entomostraca to the entire dis- appearance of the heart (many Copepoda and Ostracoda). In the same way the short sac-like heart of the Decapoda is probably to be derived from an elongate type of heart, such as is found in the Stomatopoda and Leptostraca. It seems probable, from Claus's researches on the vascular system of the Stomatopodan larvae with regard to the origin of the arteries, that the heart of the Decapoda corresponds to the most anterior portion of that of the Stomatopoda. FiQ. 87. — Dorsal aspect of a pos- terior segment in the body of a young Drawhi'pvsls.vwa., showing the development of the heart (after Glaus), c, cardioblasts ; h, cardiac cavity ; ms, mesoderm- somites ; os, rudiments of the ostia. H. Glands. Two pairs of glands which occur in the Crustacea must be regarded as modified segmental organs. These are the antennal gland (green gland) and the shell-gland. Eeichenbach (No. 65) and IscHiKAWA (No. 51) have maintained that the former arises as. GLANDS. 179 an ectodermal invagination, while Kingsley (No. 55) has proved that, in Crangon, it arises from a collection of mesoderm-cells, and only secondarily does the vesicle open externally. The mesodermal origin of the shell-gland has been established in the case of the Cladocera by Grobben (No. 11) and Lebedinsky (No. 11a). A ^/^ -SO Fig. 8S. — Three transverse sections through young Branchipus larvae (after Claus). A, transverse sections through a thoracic segment at the Metanauplius stage. B, the same through a thoracic segment at a later stage. C, the same through an abdominal segment at a still later stage, c, cardio- blasts ; d, intestinal canal ; dl, dorsal longitudinal muscle ; g, rudiment of the ventral chain of ganglia ; h, cardiac cavity ; 01', ovary ; s, pericardial septum ; so, somatic, and sp, splanchnic layers of the mesoderm ; li, ventral longitudinal muscle. Our belief that these two glands, which formerly were often confused with one another, are the homologiies of the Annelidan segmental organs, rests partly on the statements of Leydig and Gegenbaur. This interpretation of the antennal gland is supported chiefly by the careful anatomical researches of Grobbex, and by the agreement thereby revealed between the structure of this organ and the nephridia of Peripatus, so thoroughly investigated by Sedgwick. According to Grobben, we must distinguish in the antennal gland (in the Entomostracan larva as well as in the adult Malacostracan) two sections — an "end-sac" and a much-coiled tubular portion which often enlarges before opening on the basal joint of the second antenna into a vesicle (urinary bladder). All these separate parts of the gland may become more complicated by the development of secondary diverticula. In recent researches, in which the indigo -carmine injections recommended by Kowalevsky were used, Weldon (Xo. 68),* working on Palacmon serratus, concluded that the structure * Cf. Maechal, Compt. Rend., Tom. ex., cxi., and Weldon, Quart. Journ. Micro. Sci., Vol. xxxii. 180 CRUSTACEA. of this urinary organ was far more complicated than is generally believed. It had been thought probable, from investigation oi Peripatus, that the "end-sac" of the aatennal gland miglit be considered as the vestige of the coelomic sac of this segment. But "Weldon foimd in Palaemon a large coelomic sac lying in front of the genital glands and not communicating with the rest of the body-cavity, but connected with two nephridial canals running forward to the right and left ; these last widened out to form the urinary bladder. To this canal system, the "end-sac," which may be compared with the Malpighian glomerulus, is a lateral addition connected with the urinary bladder liy five glandular tubes. As the relations of the body-cavity in the Crustacea are still far from clear, further investigation of this remarkaljle condition is very desirable. In the meantime, it is perhaps advisable to adhere to the views put forward by Grobbex. That the above pairs of glands and the parts of the mid-gut participating in the formation of the excretory jiroducts (in Copepoda and Amphipoda) are by no means the only excretory organs of the Crustacea is proved by the experiments of Metschnikoff and Kowalev.sky {Biol. Centralbl., Bd. ix. ) mentioned above. These authors, by means of coloured injections, demon- strated the presence of small tubes in the thoracic limbs of 3Iysis, in which the colouring matter accumulated, and of cell -groups in a corresponding position in Nebalia. Lebedinsky (No. 57) has described the development, in Eriphia spinifrons, of a " segmental organ," arising as a paired outgrowth of the somatopleura. The tube thus produced lengthens out anteriorly and forms a coiled canal, which enters into communication with an ectodermal invagination in the coxal joint of the first pair of maxillipedes. I. Genital Organs. Our knowledge of the development of the genital organs in the Crustacea is as yet very fragmentary. The rudiments of the genital glands belong in all cases to the mesoderm. Grobben (No 21) found that, in CetocMIus, the genital rudiments are paired in the Nauplius stage, and lie ventrally to the intestinal canal. Only later do they shift dorsally to a position above the intestine, where they unite to form a single gland. Each of the genital rudiments consists of a large special genital cell, and of adjoining mesoderm-cells which yield the envelopes and efferent ducts.* In Moina,f probably in connection with its paedo-parthenogenesis, the genital rudiments can be recognised as unpaired genital cells as early as the time when * [Hacker {Archiv. /. mikr. Anat., Bd. xlix., p. 35 ; "Die Keimbahn von Cyclops") traces the rudiment of the genital gland to a single cell, which is completely isolated at the thirty-two-celled stage, and which can be identified from the first cleavage which divides the egg into two cells, one of which is purely somatic, wliile the other (the " Kornchenzelle ") gives rise to a certain number of blastomeres (three), but always retains its individuality, and finally divides into the primitive entoderm and the primitive genital cell. — Ed.] t In Dcqyhnia similis, Lebedinsky (No. 11a) was unable to distinguish the genital cells at as early a stage as is possible in Moina. Even in the Nauplius stage, the genital rudiments were indistinguishable. LITERATURE. 181 the germ-layers separate. A cell-mass arising from these by division shifts within the embryo, and there forms an unpaired plate lying dorsally above the rudiment of the mid-gut ; this plate only secondarily divides into two halves. The above mass of cells is invested later with a mesodermal envelope. The statements of Glaus (No. 9) in connection with the development of the genital organs in Branchipiis are of importance. The rudiment of the genital glands is here a paired strand lying in the three or four anterior segments at the sides of the alimentary canal, and recognisable even in early stages. The first development of the efferent ducts, however, takes place at a later stage of sexual differentiation (p. 200), and proceeds from a transformation of the twelfth and thirteenth post-ce2)halic segment into two pairs of genital swellings. The latter, in the female, join in the median line, while in the male they remain separate. In both sexes the genital swellings of the posterior segment are grown over by those of the anterior segments. The fused swellings then appear, in the female, as a broad median prominence, or in the male as paired lateral prominences. The cell-material found in the genital swellings is utilised in such a M'ay that the mesoderm-cells yield the efferent apparatus (oviduct and uterus — seminal duct and seminal vesicle) as well as the accessoiy glands, while an ectodermal ingrowth on the second pair of swellings becomes, in the female, the short external portion of the uterus (vagina), and, in the male, the long protrusible copulatory organ (cirrhus). In the Decapoda, the genital rudiment has only been observed in the latest stages of embryonic life. Bobeetzky and Reichenbach (No. 65) conjecture that it is represented by two cell-strands which lie above the intestinal canal. According to Bocretzky, these lie in the mid -gut region below the pericardial septum, whereas Reichenbach observed the rudiment in the posterior segments in the hind -gut region. Nusbaum (No. 39) found the genital rudiments, in 3fysis, as a i)aired group of cells lying behind the hepatic rudiment. When the hejiatic tubes develop, this rudiment shifts towards the dorsal side, and later prol.iably fuses to form an impaired rudiment lying between the heart and the intestine. It appears doubtful whether a few large cells observed in the ectoderm at the time when the germinal layers form, and which later lie in the abdomen, are really, as authors maintain, to be identified as the genital rudiment. Although, for general reasons, we feel inclined to attribute the genital rudi- ments to the mesoderm, Pereyaslawzewa and Rossijskawa (Nos. 70 and 73) derive the genital cells in the Amphipoda from the wall of the mid-gut (!). According to these authors, a few entoderm-cells leave the mid-gut epithelium (in Orchestia, the hepatic tubes also), and, surrounded by a mesodermal envelope, become the rudiment of the genital glands. It is well known that such an origin was asserted for these glands in Peripahis by Sedgwick. LITEEATURE REFERRING TO THE EMBRYONIC DEVELOPMENT OF THE CRUSTACEA. Crustacea in General. 1. Beneden, E. van. Recherches sur la composition et signification de I'oeuf. 31em. cour. et Mem. Sav. Etrang. Acad. Roy. Belrjique. Tom. xxxiv. 1870. 182 CRUSTACEA. 2. Beneden, E. van, and Bessels, E. Memoire sur la formation du blastoderme cliez les Amphipodes, les Lerneens et les Copepodes. Mem. cour. et Mem. Sav. Etrang. Acad. Roy. Belgique. Tom. xxxiv. 1870. 3. Gerstaecker, a. Crustacea, Bromi's Classen tmd Ordnungen des Thierreichs. Bd. v. 1. Abtli. 1. Hiilfte, 1866-1879. 2. Hiilfte, 1881-1889. 4. MtJLLER, F. Fiir Darwin. Leipzig, 1864. 5. Weismann, a. Eichtungskorper bei parthenogenetischen Eiern. Zool. Anz. Jahrg. ix. 1886. 6. Weismann, A., and Ischikawa, C. Ueber die Bildung der Richtungskorperchen bei thierischen Eiern. Ber. d. natur- forsch. Oesellsch. Freiburg. Bd. iii. 1887. 7. Weismann, A., and Ischikawa, C. Weitere Untersuchungen zum Zahlengesetz der Richtungskorper. Zool. Jalirh. Bd. iii. Abth. f. Anat. 1889. Phyllopoda. 8. Claus, C. Zur Kenntniss des Baues nnd der Entwicklung von Branchipus und Apus. Abhandl. Akad. Wiss. Gottingen. Bd. xviii. 1873. 9. Claus, C. Untersuchungen liber die Organisation und Entwick- lung von Branchipus und Artemia. Arb. Zool. Inst. Wien. Bd. vi. 1886. X 10. DoHRN, A. Unters. liber Ban und Entwickhnig der Arthropoden. 3. Die Schalendruse und die embryonale Entwicklung der Daphnien. Jen. Zeiischr. f. Naturw. Bd. v. 1870. 11. Grobben, C. Die Entwicklungsgeschichte der Moina rectirostris etc. Arb. Zool. Inst. Wien. Bd. ii. 1879. j 1 la. Lebedinsky, J. Die Entwicklung der Daphnia aus dem Sommereie. Zool. A?iz. Jahrg. xiv. 1891. [Transl. Ann. K H. (6) viii.] 12. MiJLLER, p. E. Bidrag til Cladocerernes Forplantnings-historie. Naturhistor. TidskHft (3). Bd. v. 1868. 13. Nassonow, N. B. On the Ontogeny of Balanus and Artemia (Russian). Izvyest. imp. Ohshch. Lijubit. estestv. Antrop, i. Ethnog. AIoscow. Tom. Iii., 1 (1887). 14. Pelsenebr, p. Observations on the Nervous System of Apus. Quart. Journ. Micro. Sci. Vol. xxv. 1885. 15. Lankester, E. Ray. Appendages and Nervous System of Apus cancriformis. Quart. Journ. Micro. Sci. Vol. xxi. 1881. LITERATURE. 183 16. Weismann, a., and Ischikawa, C. Ueber die Paracopulation im Daphnidenei, sowie liber Reifung und Befruchtung dessel- ben. Zool. Jahrh. Bd. iv. AUh. f. Anat. 1891. APPENDIX TO LITERATURE ON PHYLLOPODA. I. Hacker, Y. Ueber die Entwicklung der Wintereies von Moina paradoxa. Ber. Ges. Freiburg. Bd. vii. and viii. 1893. II. Kerherv]5, L. B. De I'apparition provoquee des ephippies chez les Daphnies (Parthenogenesis of Daphnia). Mem. Soc. Zool. Tom. V. 1892. III. Samassa, p. Die Keimblatterbildung bei den Cladoceren. Arcliiv. f. mih\ Anat. Bd. xli. 1893. Copepoda. 17. Beneden, E. van. Eecherclies sur I'embryogenie des Crustaces. IV. Developpement des genres Anchorella, Lernaeopoda, Brachiella et Hessia. Bidl. Acad. Roy. Belgique (2). Tom. xxix. 1870. 18. Claus, C. Zur Anatomie und Entwicklungsgeschichte der Copepoden. ArcMv. f. Naturgesch. Bd. xxiv. 1858. 19. Claus, C. Die frei lebenden Copepoden etc. Leipzig, 1863. 20. Fritsch, J. A. Note preliminaire sur I'ontogenie de nos Cope- podes d'eau douce. Zool. Anz. Jahrg. v. 1882. 21. Grobben, C. Die Entwicklungsgeschichte von Cetochilus septentrionalis. Arh. Zool. Inst. Wien. Bd. iii. 1881. 22. HoEK, P. P. C. Zur Entwicklungsgeschichte der Entomostraken. II. Zur Embryol. der frei lebenden Copepoden. Niederl. Archiv.f. Zool. Bd. iv. 1877-1878. 23. Urbanowicz, Eel. Zur Entwicklungsgeschichte der Cyclopiden. Zool. Anz. Jahrg. vii. 1884. 24. XJrbanowicz, Fcl. Contributions a I'embryologie des Copepodes. Archiv. Slav. Biol. Tom. i. 1886. APPENDIX TO LITERATURE ON THE COPEPODA. 1. Hacker, Y. Die Keimbahn von Cyclops. Archiv. f. mikr. Anat. Bd. xlix. 1893. II. ScHiMKEWiTSCH, W. Studien uber parasitische Copepoden. Zeitscltr.f.Wiss. Zool. Bd. Ixi. 1896. Cirripedia. 25. Beneden, E. van. Recherches sur I'embryogenie des Crustaces. III. Developpement de I'oeuf et de I'embryon des Sacculines (Sacculina carcini Thomps.). Bull. Acad. Roy. Belgique (2). Tom. xxix. 1870. 184 CRUSTACEA. 26. BovALLius, C. Embryologiska Studier I. Om Ijalanidernas Utveckling. Stodcholm, 1875. 27. HoEK, P. P. C. Zur Entwicklungsgeschiclite der Entomostraken. I. Embryologie von Balanus. Niederl. ArcMv. /. Zool. Bd. iii. 1876-1877. 28. Lang, A. Die Dotterfurchung von Balanus. Jen. ZeifscJir. /. Naturiv. Bd. xii. 1878. 29. Nassonow, j^. Zur embryonalen Entwicklung von Balanus. Zool. Anz. Jahrg. viii. 1885. 30. Nussbaum, M. Vorl Ber. iiber die Ergebnisse einer mit Unter- stiitzung der k. Acad, ausgef. Eeise nach Californien. Sitz- ungsher. Akad. Wiss. Berliji. 1887. 31. Nussbaum, M. Anat. Studien an Calif ornisclien Cirripedien. Bonn, 1890. APPENDIX TO LITERATURE ON CIRRIPEDIA. I. Groom, T. T. Early Development of the Cirripedia. Tr. Roy. Soc. London. Vol. clxxxv. 1894. Leptostraca. 32. Claus, C. TJeber den Organismus der Nebaliden und die systeni- atische Stellung der Leptostraken. Arh. Zool. Inst. Wien. Bd. viii. 1888. 33. Metschnikoff, E. On the Development of Xebalia (Eussian). Zapiski imp. Acad. Naul: St. Petersburg. Tom. xiii. 1868. 34. Metschnikoff, E. Sitzungsher. d. Verhandl. deutsclier Natur- f or seller zu Hannover. 1865. p. 218. Cumacea. 35. Blanc, H. Developpement de I'oeuf et formation des feuillets primitifs chez la Cuma Eathkii Krciy. Recueil Zool. Suisse. Tom. ii. 1885. 36. DoHRN, A. Ueber Bau und Entwicklung der Cumaceen. Jen. Zeitschr.f. Naturw. Bd. v. 1870. APPENDIX TO LITERATURE ON CUMACEA. I. BouTCHiNSKY, P. Zur Embryologie der Cumaceen. Zool. Anz. Bd. xvi. 1893. Schizopoda, 37. Beneden, E. van. Eecherches sur I'embryogunie des Crustaces. II. Developpement des Mysis. Bidl. Acad. Roy. Belgique (2). Tom. xxviii. 1869. LITERATURE. 185 37a. BouTCHiNSKY, p. Observations on the Development of Para- dopsis cornuta Czern. (Russian). Zapiski Novoross. Obshch. Estestv. Odessa. Tom. xiv., 2. 1890. 38. NusBAUM, J. Zur Embryologie der Schizopoden (Mysis cliam- aeleo). Biol. CentmlU, Bd. vi. 1887. 39. NusBAUM, J. L'embryologie de Mysis chamaeleo (Thompson). Archiv. Zool. Exper. (2). Tom. v. 1887. 40. Wagner, Jul. Sur le developpement des Schizopodes. Comni. prel. I. Sur la formation des feuillets embryonnaires de Neomysis vulgaris var. baltica Czern. Rev. Sc. Nat. Soc. Natural. St. Petersbour;/. Ann. i. 1890. II. La formation et la signification du sillon caudal de la Neomysis vulgaris var. baltica Czern. Hev. Sc. Nat. Soc. Natural. St. Peters- hourg. Ann. i. 1890. APPEXDIX TO LITERATURE ON SCHIZOPODA. I. Bbrgh, R. S. Beitrage zur Embryologie der Crustaceen. Zur Bildungsgesehichte des Keimstreifen von Mysis. Zool. Jahrb. (Anat.). Bd. vi. 1893. II. Wagner, J. Zur Entwicklungsgeschichte der Schizopoden. Zool. Anz. Bd. xvii. 1894. Decapoda, 41. Bobretzky, N. On the Embryology of the Arthropoda (Development of Astacus and Palaemon). Zapislii Kievsk. Obshch. Estestv. Tom. iii., vuip. 2. See Hoyer, Jahresber. f. Anat. and Phys. von Hofmann and Sclmalbe. Bd. ii. 1873. 42. Brooks, W. K. The Embryology and Metamorphosis of Sergestidae. Zool. Anz. Jahrg. iii. 1880. 43. Brooks, W. K. Lucifer, a study in Morphology. Tr. Roy. Soc. London. \o\. clxxiii. 1883. 44. Carriere, J. Ban und Entwicklung des Auges der zehnfiissigen Crustaceen und der Arachnoiden. Biol. Ce^itralbl. Bd. ix. 1889. 45. Dohrn, a. Untersuchungen iiber Bau und Entwicklung der Arthropoden. 6. Zur Entwicklungsgeschichte der Panzerkrebse (Decapoda Loricata). Zeitschr. f. Wiss. Zool. Bd. xx. 1870. 46. Faxon, W. The development of Palaemonetes vulgaris. Bull. Mus. Comp). Zool. Vol. v. 1879. 47. Haeckel, E. Studien zur Gastraea-Theorie. II. Die Gastrula und die Eifurchung der Thiere. Jena, 1877. Jen. Zeitschr. f. Naturw. Bd. ix. 1875. 1 86 CRUSTACEA. 48. Herrick, F. H. The development of the compound eye of Alpheus. Zool. Anz. Jahrg. xii. 1889. 49. Herrick, F. H. IS'otes on the Embryology of Alpheus and other Crustacea, and on the development of the compound eye. John Hopkins Univ. Circ. Vol. vi., Ko, 54, 1886 ; and Vol. vii., No. 63, 1887. 50. Herrick, F. H. The development of the American Lobster : Homarus americanus. John Hopkins Univ. Circ. Yo\. ix. 1890. No. 80. 50a. Herrick, F. H. The development of the American Lobster. Zool. Anz. Jahrg. xiv. 1891. 51. IscHiKAWA, Ch. On the development of a freshwater macrurous crustacean Atyephyra compressa, De Haan. Quart. Journ. Micro. Sci. (2). Vol. xxv. 1885. 52. Kingsley, J. S. The development of the compound eye of Crangon. Journ. Morph. Vol. i. 1887. 63. KiXGSLEY, J. S. The development of Crangon vulgaris. 2nd paper. Bull. Essex Inst. Vol. xviii. 1887. 54. Kingsley, J. S. The development of the compound Eye of Crangon. Zool. Anz. Jahrg. ix. 1886. 55. Kingsley, J. S. The development of Crangon vulgaris. 3rd paper. Bull. Essex Inst. Vol. xxi. 1889. 56. Lebedinsky, J. On the development of Eriphia spinifrons (Russian). Zapiski Novoross. Ohshch. Esfestv. Odessa, 1889. 57. Lebedinsky, J. Einige Untersuchungen liber die Entwicklungs- geschichte der Seekrabben. Biol. Centralbl. Bd. x. 1890. 58. Lereboullet, A. Recherches d'embryologie comparee sur le developpement du Brochet, de la Perche et de I'Ecrevisse. Mem. Savans. Etrang. Paris. Vol. xvii. 1862. 59. Mayer, P. Zur Entwicklungsgeschichte der Decapoden. Jen. Zeitschr. f. Naturio. Bd. xi. 1877. 60. Mereschkowski, C. Fine neue Art von Blastodermbildung bei Decapoden. Zool. Anz. Jahrg. v. 1882. 61. MoRiN, J. On the Development of Astacus (Russian). Zapiski Novoross. Ohshch. Estestv. Odessa. Tom. xi. 1886. 62. Parker, G. H. The history and development of the eye in the Lobster. Bull. Mus. Comp. Zool. Cambridge. May, 1890. 63. Ratiike, H. Ueber die Bildung und Entwicklung des Fluss- krebses. Leipzig, 1829. 64. Reichenbach, H. Die Embryonalanlage und erste Entwicklung des Flusskrebses. Zeitschr. f. Wiss. Zool. Bd. xxix. 1877. LITERATURE. 187 65. Reichenbach, H. Studien zur Entwicklungsgescliichte des Flusskrebses. Ahhandl. Senkenberg. Nat. Gesellsch. Frank- furt. Bd. xiv. 1886. 66. ScHiMKEWiTSCH, W. Eiiiige Bemerkungen iiber die Entwick- lungsgescliichte des Flusskrebses. Zooh Anz. Jahrg. viii. 1885. 67. ScHiMKEWiTSCH, W. Some Observations on the Development of Astacus leptodactylus (Russian). Isvyest. imp. Obshcli. Lijuhit. estestv. Antrop. i. Ethnog. Moscow. Tom. 1. 1886. -68. Weldon, W. F. R. Coelom and ISTepliridia of Palaemon serratus. Journ. Alarine Biol. Assoc. Vol. i. 1889. APPENDIX TO LITERATURE ON DECAPODA. I. Allen, J. E. Nephridia and Body-cavity of the Larva of Palaemonetes varians. Proc. Roy. Sac. London. Vol. Hi. 1893. 11. BouviER, E. L. Sur le Developpement embryonaire des Galatheides du genre Diptychus. Compt. Rend. Ac. Sci. Paris. Tom. cxiv. III. Bdmpus, H. C. The Embryology of Homarus americanus. Jour7i. Morph. Vol. v. 1891. IV. Boutchinsky, p. Zur Entwicklungsgeschichte von Gebia litoralis. Zool. Anz. Bd. xvii. 1894. V. Edwards, P. Milne-, and E. L. Bouvier. Sur le Developpe- ment embryonaire des Galatheides abyssaux du genre Diptychus. Compt. Rend. Ac. Sci. Paris. 1891-92. TI. GoRHAM, F. P. The Cleavage of the Egg of Virbius zosteri- cola. Journ. Morph. A^ol. xi. 1895. VII. Parker, G. H. The Retina and Optic Ganglia in Decapoda. Mittheil. Stat. Zool. Neapel. Bd. xii. 1895. VIII. Roule, L. Developpement des Decapodes. 1, Du Palemon serratus. Ann. Sci. Nat. Zool. (viii.). ii. 1896. IX. Weldon, W. F. R. The Formation of the Germ-layers in Crangon vulgaris. Quart. Journ. Micro. Sci. xxxiii. Amphipoda. 69. Bessels, E. Einige Worte iiber die Entwicklungsgeschichte und den morphologischen Werth des kugelformigen Organs der Amphipoden. Jen. Zeitschr. f. Naturtv. Bd. v. 1870. 70. Pereyaslawzbwa, Sophie. Le developpement de Gammarus poecilurus Rathke. Bull. Soc. Natur. Moscou (2). Tom. ii. 1888. 188 CRUSTACEA. 71. Pereyaslawzewa, Sophie. Le developpement de Caprella ferox Chrnw, Bull. Soc. Natur. Moscou (2). Tom. ii. 1888. 72. EossiJSKAYA, Marie. Le developpement d'Orchestia littorea. Bull. Soc. Natur. Moscou (2). Tom. ii. 1888. 73. EossiJSKAYA - Koschewnikowa, M. Le developpement de la Sunampliithoe valida Czerniavski et de I'Amphithoe picta Eathke. Bull. Soc. Natur. Moscou. 1890. 74. Uljanin, V. Blastoderm- und Keimblatterbildung bei Orchestia Montagui und mediterranea. Zool. Anz. Jahrg. iii. 1880. 75. Uljanin, Y. Zur Entwicklungsgeschichte der Amphipoden. Zeitschr. f. Wiss. Zool. Bd. xxxv. 1881. 76. Della Valle, A. Deposizione, fecondazione e segmentazione delle uova del Gammarus pulex. Atti della Sue. dei Natu- ralisti di Modena (3). Vol. viii. 1889. 77. De la Valette St. -George, A. Studien liber die Entwicklung der Amphipoden. Ahh. Naturf. Gesellsch. Halle. Bd. v. 1860. APPENDIX TO LITERATURE ON AMPHIPODA. L Bergh, E. S. Die Drehung des Keimstreifens und die Anlage- des Dorsal organes bei Gammarus pulex. Zool. Jahrh. (Anat.). Bd. vii. 1893. Anisopoda. 78. Glaus, C. Ueber Apseudes Latreillii Edvv. und die Tanaiden. IL Arb. Zool. Inst. Wieii. 1888. Bd. vii. Isopoda. 79. Benedbn, E. van. Eecherches sur I'embryogenie des Crustaces. I. Asellus aquaticus. Bidl. Acad. Roy. Behjique (2). Tom. xxviii. 1869, 80. Bobretzky, N. Zur Embryologie des Oniscus murarius. Zeitschr. f. Wiss. Zool. Bd. xxiv. 1874. 81. BuLLAR, J. F. On the development of the parasitic Isopoda. Tr. Roy. Soc. London. Vol. clxix. 1878. 82. Glaus, G. Ueber die morphologische Bedeutung der lappen- formigen Anhange am Embryo der Wasserassel. Kaiserl. Akad. Wissensch. Anz. Wien. 1887. 83. DoHRN, A. Die embryonale Entwicklung des Asellus aquaticus. Zeitschr. f. Wiss. Zool. Bd. xvii. 1867. LITERATURE. 189 Si. Grosglik, S. Schizocol oder Enterocol 1 Zool. Anz. Jabrg. x. 1887. ■85. NusBAUM, J. L'embryologie d'Oniscus murarius. Zool. Anz. Jahrg. ix. 1886. ■85a. NusBAUM, J. Beitriige zur Embryologie der Isopoden (Ligia oceanica). Biol. Centralhl. Bd. xi. 1891. 86. Packard, A, S. On the structure of the brain of the sessile- eyed Crustacea. Mem. National Acad. Sci. Washington, 1884. Bd. iii. 57. Rathke, H. Untersuchungen liber die Bildung und Entwick- lung der Wasserassel. Leipzig, 18.32. 58. Eathke, H. Recherches sur la formation et le developpement de I'Aselle d'eau douce. Ann. Sci. Nat. Zool. (2). Tom. ii. 1834. 89. Rathke, H. Zur Morphologic. Eeisebemerkungen aus Taurien. Riga and LeijJzig, 1837. 90. Eeinhard, W. Sur le developpement de Porcellio scaber. Trav. Soc. Natur. Kharkoiv. Tom. xx. 1887. 91. Reinhard, W. Zur Ontogenie des Porcellio scaber. Zool. Anz. Jahrg. x. 1887. 92. RouLE, L. Sur revolution initiale des feuillets blastodermiques chez les Crustaces isopodes (Asellus aquaticus L. et Porcellio scaber Latr.). Comp. Rend. Ac, Sci. Paris. Tom cix. 1889. APPENDIX TO LITERATURE ON ISOPODA. I. McKwoBn/j'. P. The Embryology of the Isopod Crustacea. Journ. Morpli. Vol. xi. II. NusBAUM, J. Matergaly do enibryogenii i histogenii rovvnonogoii (Isopoda). Ahli. Krahauer Akad. Bd. xxv. 1893. And Biol. Centralhl. Bd. xiii. 1893. III. RouLE, L. Etudes sur le developpement des Crustaces. La segmentation ovulaires et le fagonnement du corps cliez I'Asellus aquaticus. Ann. Sci. Nat. Zool. (8). 1. 1896. IV. RouLE, L. Etudes sur le developpement des Crustaces. Le Developpement du PorceUio scaber. Ann. Sci. Nat. Zool. Tom. xviii. 1894. 190 CRUSTACEA. II. Metamorphosis. 1. The Nauplius Stage. The most typical feature of the metamorphosis in the Crustacea, and therefore, perhaps, the most convenient one with which to commence our study, is tlie so-called Nauplius stage. In those cases in which the young animal is hatched at a later stage of develop- ment {e.g., in the Cladocera, the Arthrostraca, and most Decapoda), the NaiqMus stage is thrown back among the series of embryonic stages ; nevertheless, we commonly find that this stage is indicated by a period of rest, such as an ecdysis of the Nauplius integument (p. 118). The Kauplius body (Fig. 89) is as a rule oval, the anterior end being the more rounded, and the posterior the narrower end of the body. In other respects, there is great variety in the shape of the body. Dorsally compressed, laterally compressed, long, and even broad NaiipJii are to be found. The possession of three pairs of linibs — the future first antennae (a'), second antennae (a"), and mandibles (^md) — points to the fact that, in the Nauplius, we have an already segmented larval form. This segmentation is, however, not externally recognisable in the body of the Naujilius, although the limits of the segments may be indicated in the ontogenetic stages which lead up to the Nauplius (free-living Copepoda, Cirri- pedia). Typically, the Nauplius has no shield-like reduplication of the dorsal integument, but such a fold (which usually only appears at a later stage) may be distinctly recognisable in individual cases (Cirripedia, Fig. 102 A), or at least may be indicated by a slight fold in the integument of the dorsal surface. The posterior end of the body is still without the paired furcal processes, but is armed with paired setae (furcal setae). The most anterior pair of limbs, the first antennae {a'), are uniramose, and consist of few joints or are still unjointed. They serve for locomotion, and are also of importance as bearing sensory organs. The two pairs of limbs which follow these are developed in the form of biramose swim- ming limbs. The first of these, the second antennae (a"), lie at the sides of the mouth, and are distinguished by a strong, hook-like, masticatory process springing from the basal joint. The third pair of limbs, the mandibles (md), also function principally as locomotory organs. A masticatory blade is not, as a rule, developed on the basal joint, indeed, this structure, so characteristic of the appendage in the adult, is hardly indicated, although it may in some cases THE NAUPLIUS STAGE. 191 occur here. The U-shaped antennal gland {at) opens externally on the basal joint of the second antenna. The oral aperture which lies between the two antennae is overhung by a large upper lip (labruni), and leads into the alimentary canal, in which can be distinguished a short oesophagus, a widened mid-gut, and a hind- gut. The anal aperture may be wanting in the first Nmiplms stages (Cttochilus, Cyclops). It has been observed in many cases that this aperture origin- ally opens on the dorsal surface (Cirripedia, Fig. 103, p. 212; CetocMlus, Fig. Ill B, p. 233; em- bryonic stage in the Cladocera, Fig. 72, p. 147), the aperture shift- ing later to the posterior end of the body between the furcal processes. The nervous system retains its original connection with the ectoderm, and consists of the supraoeso- phageal ganglion, circum- oesophageal commissures, and the first pair of gan- glia of the ventral cord. The second antennae are innervated from a pair of ganglia lying behind the mouth (Claus, Dohbn), an arrangement which is retained by the Phyllopoda in the adult. As a sensory organ, there is the unpaired J^auplius eye, which lies in the frontal region, and is composed of three parts. The most developed muscles are those of the limbs, Avhich are attached to a point in the mid-dorsal region. The heart is not yet developed. Although there is no external segmentation in the Nauplius body, we can nevertheless recognise in it the following segments: — An anterior pre-oral or primary cephalic segment, a posterior terminal or anal segment at the jiosterior end of the- body, and the true trunk-segments which lie between these. As members of the last group, we recognise the section of the body belonging to the second antenna and the mandibular segment. We have, however, already pointed out (p. 165) that the region of the first antenna possibly also corresponds Fig. 89. — Naxlii of the Lepadidae, when hatched, seem closely to resemble tlie Naiqilii of the Bcdanidae in shape, differing only in this one respect, that the long frontal horns are bent backwards. Soon after hatching they undergo a first moult, after which the spine-like thoraco-abdominal appendage (ventral spine) and the dorsal caudal spine first attain their full length. During later moults, the NaupUus grows to a great size (12 mm. long), and is then distinguished by a number of accessory structures. In Lepas fascicularis, seven moults take place before the Metanaupilius stage is reached. These later stages (Fig. 103) show the thoraco-abdominal appendage (t), and the dorsal caudal spine (ds) drawn out as long, pointed processes beset with small hooks. A similar process (d) has arisen at the central point of the dorsal shell ; this process, which runs obliquely backwards, is known as the dorsal spine. At the margin of the dorsal shield also, besides the frontal horns, eight large spine-like processes can be recognised, the first pair lying anteriorly between the frontal horns, and the others being distributed laterally and posteriorly. Both the dorsal shield and the larger processes are further beset with a series of minute spines, which are wanting only on the frontal horns. Many of these minute spines, as well as the marginal spines of the dorsal shield, are found to be jierforated and connected with the ducts of unicellular integumental glands. On the ventral side of the thoraco - abdominal appendage, two more posteriorly-placed, immovable thorns can be distinguished,, and, anteriorly, six pairs of larger movable spines (x), which do not appear simultaneously but successively, each new moult develop- ing a fresh pair, until, after the last moult, the full number of six is attained. The conjecture, made with some hesitation by Fig. 103. — Larva of Ltims cmstralis (Archizoaea gigas) (after Dohrn). a, anal aperture ; d, dorsal spine; ds, dorsal caudal spine; h, frontal horns ; ol, upper lip ; t, thoraco- abdominal appendage (yentral spine); x, paired movable spines. THORACICA. 213 DoHRN (No. 42) and Willemoes-Suhm (No, 62), that these pairs of spines correspond to the rudiments of the six pairs of thoracic limbs in the Cypris-like larva appears to us somewhat plausible. This last stage would then have to be assumed to be the Meta- naupUus stage. For more recent observations on the larvae of the Lepadidae we are above all indebted to Willemoes-Suhm (No. 62). Dohrn (No. 42) has minutely described a large Nauplius larva — probably to be referred to Lepas australis under the name of Archizoaea gigas. These stages have, however, absolutely nothing to do with the Zoaea larva of the Malacostraca. The development of long spine-like processes is a protective ada^itation often occurring in pelagic animals {cf. the skeletons of the Radiolaria and the spines of the Plutcus larvae). In the whole series of Nauplius stages, except the last, there is no essential alteration apart from the increase in size of the body and the above-mentioned development of thorn- and spine-like processes. In the last stage which immediately precedes the CijprU stage {Metanauplius, Fig. 102 B), important new rudiments of organs appear. With respect to the general form of the body, it should be emphasised that the lateral parts of the dorsal shield now already begin to bend downwards, and, covering the body laterally, fore- shadow in position the valves of the shell in the Cypris stage. The three anterior pairs of limbs still bear a general resemblance to the typical Nauplius limbs, but already, within the first pair (first antennae), the rudiment of the sucking disc so important in the Cypris stage can be made out (Krohn, No. 50 ; Willemoes-Suhm, No. 62 ; Claus, No. 8). Tliis fact aftbrds a direct proof that the adhesive antennae of the later stages are actually derived from the most anterior pair of Nauplius limbs (Krohn, No. 50 ; Fr. Muller). Behind the mandible is to be seen the slightly developed rudiment of the fourth pair of limbs ; it is, however, probable, as we shall see further later, that we have here the rudiments of two con- secutive pairs of limbs {mx) (Metschnikoff, No. 53 ; Claus, No. 8). The thoraco-abdominal region of the larva has increased considerably in size ; in it, beneath the Nauphus cuticle, can be recognised the rudiments of the six pairs of swimming limbs {f^-f^^) of the larva (thoracic limbs), as well as the caudal processes of the short abdomen (furcal processes) (Krohn, No. 50 ; Claus, No. 8). Another im- portant point is that in this stage, near the Nauplius eye, the paired, movable compound eyes (Fig. 102 B) already appear as rudiments. In this Metanauplius stage, therefore, all the more important organs and limbs of the Gypris-Wko, larva are formed, and the larva itself. 214 CRUSTACEA. by the casting of the Metanauplius integument, emerges from the preceding stage. The free-swimming Cypris stage (Fig. 104), which is only of short duration^ derives its name from the bivalve shell that envelops the whole body. The actual segmentation of the body and the inner organisation have nothing in common with the Ostracoda. In respect to these the larva at the Cypris stage already very nearly approaches the adult. The two shell-valves which can be brought together by an adductor muscle, are directly continuous dorsally. The dorsal border appears arched, whereas the ventral edge is flattened. Anteriorly they are rounded, but run out to a point posteriorly. In the anterior section of the valves, near the ventral side, a small projection may be remarked on which can be recognised the frontal horn of the Nauplius. The ventral margins of the shell are closely approximated in the middle of their length. Posteriorly, a fissure opens between them for the passage of the swimming limbs of the larva (?/), and there is also an anterior aperture through which the first antennae, the adhesive organs (Z), are protruded. The latter limbs, in the free Cypris stage, are used for the occasional attachment of tlie larva, which precedes and prepares the way for the final attachment. Tliey consist at this stage of four joints, the basal joint joining the body by a broad base, and carrying various chitinous ingrowths (apodemes) for the attachment of muscles. The second joint, which is lengthened like an arm, is bent at an angle to the basal joint. The short third joint carries at its outer side the adhesive disc, at the centre of which the duct of the cement-gland opens, while the truncated fourth joint seems to be provided with one ordinary and one large olfactory seta (Claus, No. 39 ; Willemoes-Suhm, No. 62). The cement-gland (Fig. 105, cd) shows various degrees of develop- ment in the Cypris larva. In the pupal stage which follows, it Fio. 104. — Ci/p''is-like stage of Lepas fascicularis (after Claus, from Lang's Text-book), ua, Nauplius eye ; jm, paired eye ; rf, thoracic (swimming) limbs ; ab, abdomen ; I, first antenna. ■- THORACICA. 215 shifts to the base of the eye-stalk formed from part of the head of the larva, and here takes the form either of a closely-packed coiled gland (Lepas pectincda), or else of widely scattered cells connected only by the efferent duct. On account of its structure, the cement- gland must be regarded as an accumulation of unicellular glands on a much-branched common duct, the single glandular cells being sessile on this duct, like grapes. We are still very nnich in the dark as to the genesis of this gland and its morphological derivation. Attempts have been made to derive it from one of the two typical pairs of Crustacean glands (antennal and shell-glands), although such a homology is attended by considerable difficulty on account of the aperture of the gland occurring on another limb. Glaus (No. 8) tinds, within the shell- fold of the free-swimming Cypris stage of Lepas aicstralis, a coiled gland-like cell strand, and is disposed to homologise this with the shell-gland of other Entomostraca, while he conjectures that the cement-gland which is recognisable in later stages is derived from this cell-strand. Willemoes-Suhm, on the contrary (No. 62), found, even in the Nauplius stage, a paired glandular mass lying at the sides of the upper lip, out of which, according to him, the cement- gland develops. While authorities thus hold conflicting views, we shall do well for the present to regard the cement-gland as a peculiar structure found in the Cirripedia, and to abstain from attempting to homologise it with the glands of other Crustacea. A great change occurs in the structure of the limbs which surround the oral aperture of the Nauplius. The second antennae seem altogether to disappear (if they are not to a certain extent, as Pagenstecher (j^o, 58) conjectured, retained in the palp-like ajDpendages of the upper lip). The actual mouth-parts, together with the upper lip, are already shifted on to a slightly projecting oral cone, and appear in the form of three pairs of truncated rudi- ments, in which we recognise the future mandibles, the first maxillae, and the lower lip which results from the fusion of the second maxillae. In what way these mouth -parts are derived from the limbs of the MetanaupUus stage is still far from clear. The most probable view appears to us to be that of Glaus (No. 8), according to which the mandibles are derived from the basal joints of the third pair of Nauplius limbs, while the outer segment of the rudimentary limbs following these in the MetanaupUus (Fig. 102 B, mx) yield the first maxillae. The under lip, on the contrary, is said to rise from a rudiment on the inner side of this limb. We should thus have to assume that, in this imperfectly developed limb, we have the rudi- ments of the first and second pairs of maxillae crowded together. Metschnikoff (No. 53) also derives two pairs of limbs from this rudiment, but identifies them with the mandible and maxilla of 216 CRUSTACEA. the adult Cirripede, the third Nauplius limb as well as the second being supposed to vanish completely. In the more posterior section of the body, we find six pairs of swimming limbs (rf), which, in their structure, strikingly recall the thoracic limbs of the Copepoda, and are also similarly used as swimming feet. Each of these consists of a two-jointed basal segment provided with an exopodite and endo- podite, each of these, again, being composed of two joints and clothed with long swimming setae. Posteriorly, the body terminates in a short abdomen {ah) composed of four segments, the last of which ends in two furcal appendages bearing long setae. With regard to the internal organs, it should be mentioned that, at this stage, sac-like outgrowths of that enlargement of the intestinal canal which is called the stomach begin to appear : these develop into the so-called liver ; further, that the rudiment of tlie ovaries is to be seen in the shape of paired tubes situated ventrally in the cephalic region. Fig. 105.— Pupa of Lepas pectitiata (after Claus, from Lako's Text-hook), iia, Naiiplius eye ; pa, paired eye ; r/, thoracic limbs (with the tendril-like— cirriforiii— feet beginning to form inside) ; o, mouth ; d, intestine ; L, liver; sm, adductor muscle ; sc, scutum; f, tergum ; ca, carina ; cd, cement-gland ; I, first antenna with the sucking disc. The free-swimming Ci/j^ris-like larva, after becoming finally attached, gives rise to the Jixed Ctjp^-is-Wie larva (Fig. 105) of the Cirripedia. As, at this stage, the larva ceases to feed and loses all power of locomotion, while important external and internal changes are going on in the body beneath the cuticle of the Ci/jyris stage (as if beneath a pupal integument), this stage has been well called the pupal stage (Cyjms-like pupa). From this, by the casting of the cuticle of the (Tyjjm-like stage, proceeds the young Cirripede. For details as to the processes through which these transformations are brought about, we are indebted chiefly to the THORACICA. 217 observations of Darwin (No. 40), Pagbnstecher (No. 58), and Claus (Nos. 39 and 8). The larva attaches itself firmly with the sucking discs found on the first antennae and by means of a sticky fluid secreted by the ■cement-gland. At first the whole of the ventral surface is in contact with and parallel to the surface of attachment (Fig. 106 ^ and B). The changes Avhich first attract attention are certain processes of growth by means of which different parts of the body approach the final shape. For example, the mouth-parts at the top of the buccal cone, which until now were very rudimentary, become distinctly -K- i-f ^'o Vdy / Fig. 106.— Diagrammatic illustration of the metamorphoses of Lepas. A, Cypris-VCiia stage. B, attached pupa. C, young Lepas stage, still surrounded by the loosened Cypris shell (s). a', first antenna; ah, abdomen; c, carina; d, intestine; to, mouth; o, unpaired eye; pa, paired eye ; rf, thoracic limbs ; s, Cypris shell ; sc, scutum ; t, tcrgum ; x, reflection of the dorsal integument ; y, ventral fold. developed ; in the Lepadidae the principal feature in this process is the great increase in size of the upper lip. The thoracic limbs also elongate and become (Fig. 106 B) the rudiments of the tendril-like feet ()/). All these parts develop under the sonaewhat loose Cypris cuticle. The long tendril-like feet naturally have not the room necessary for development in the short envelope of the thoracic limbs of the Cypris stage, and therefore become much curved, and even press back into the thorax (Fig. 105), The abdomen degenerates almost completely, while from its base (genital segment) the unpaired penis projects as an outgrowth. 218 CRUSTACEA. While these processes of growth are taking place, the whole thorax undergoes a significant change of position (Fig. 106 yl and B). Whereas, in the earlier stages, the thorax lay almost parallel to the ventral surface, it now rises into a more perpendicular position with regard to that surface, so that the anal aperture no longer lies behind the oral aperture, hut above it. Simultaneously with this change of position, a sharper distinction arises between the thorax and the cephalic region, the point {x) at which the wall of the thorax unites with the surface of the mantle shifting ventrally (c/. the position of X in Fig. 106 A, with its position in Fig. 106 B). Meanwhile many changes take place in the anterior cephalic region of the pupa, these being preparatory to its transformation into the stalk of the adult form. The broad basal joint of the adhering antenna first completely fuses with the head and is taken up into the latter, so that the adhesive antennae of the adult Cirripede each consist of only three joints. Further, somewhat behind this point, a deep infolding of the surface of the body occurs (Fig. 106 B, y), so that that portion of the head which forms the stalk is at this stage sharply bent on itself. This fold arises by the withdrawal of the stalk-integument from the cuticle of the Cypris larva. Some of the important organs of the larva, however, which are not to be taken over into the adult, remain attached to that cuticle. The chief of these are the paired eyes (the Kauplius eye, however, passes over into the adult Cirripede and is retained throughout life), and the chitinous processes called by Darwin the apodemes, which served for the attachment of the antennal muscles and are cast otf in the moult which follows. These details are not represented in the figure. The most important change which now follows is in the position of the larva and the lengthening of the stalk which is connected with it. By an opening out of the fold above mentioned, the larva now rises from the surface of attachment, its ventral surface standing at right angles to the latter. At the same time the stalk passes out from between the shell- valves of the Cypris stage (Fig. 106 C), and lengthens to form the adult peduncle. Darwin (No. 40) pointed out the fact that the part by means of which the attachment of the larva at first takes place does not correspond to the frontal extremity of the body, but to the most anterior portion of the ventral surface. Only after the upright position has been assumed does the frontal extremity come into contact Math the surface of attachment, to which it is glued by a secretion. In Crytcyphialus, on the contrary, and also in AlcipjK, Lithotrya, and Anelasma, this part is not firmly attached, but grows out further. This further growth can only take place when the substratum can give way correspondingly. THORACICA. 219 In Anelasma, the substratum is affected by simple pressure, but in the other forms just mentioned by a boring action of the stalk. The passage of the stalk out of the Cypris shell is rendered possible by the fact that the latter at this stage surrounds the body very loosely. Whereas, in the Cypris stage, the Avhole of the ceplialic region is contained within the bivalve shell, in the adult, the anterior part of that section (the stalk) is not covered by the shell, here, as in the Cladocera, only that part of the head which carries the mouth-parts being included within the shell. We may imagine this change to have arisen at the time when the stalk grew out by a secondary flattening out of that anterior part of the mantle-fold which, in the Cirripede pupa, covered the most anterior part of the head. The adult Cirripede shell now appears more distinctly beneath the Cypris shell, and in it can be made out the first rudiments (primary valves, Darwin) of the five calcareous plates (scuta, terga, and carina, Fig. 106 C, sc, f, c). These primary valves are distinguished by their sieve-like sculpturing, which is caused by the boundaries of the matrix-cells remaining evident in the calcareous secretion. The surfaces of the valves are covered by a thin cuticle. The primary valves do not directly increase in size, but new calcareous layers are continually being secreted below them, and these attain a size greater than that of the original valves. In a superficial view of the primary valves, they are now seen to be surrounded by concentric lines representing the subjacent calcareous layers. By this method of increase of the valves, the non-calcareous parts of the shell which spread out between the valves become more and more circumscribed, but in some cases these intermediate spaces are retained to a considerable extent (^Conchoderma). It should be mentioned that in those forms which have a large number of valves, only the five primary valves are at first formed. The metamorphosis of the Balanidae, in the first stages, resembles that of the Lepadidae. Here also the Cypris pupa gives rise to a young form attached by a short fleshy stalk. Only later does the broad base characteristic of the Balanidae develop and form that external secondary mantle-fold, within which the upper part of the mantle which carries the scuta and terga appears like an operculum. The first rudiments of the shell are here membranous, and the sculpturing mentioned above, which is characteristic of the Lepadidae, is wanting. While the external form of the adult is reached in this manner, the inner organs are also undergoing important changes, some of which are as yet little iinderstood. Some organs are cast off at ecdysis (paired eyes, antennal apodemes), others simply degenerate (antennal muscles). Meantime the hepatic outgrowths appear in the intestinal canal. The cement-gland increases considerably in size, and the genital organs show special development. The ovaries undergo their characteristic displacement, shifting to a position within the stalk. The moult which follows, and in which the cuticle of the Cypris stage is cast, closes this period of development. In the moult, the outer cuticle of the two 220 CRUSTACEA. shell-valves is cast first, the thorax and the inner mantle-cavity losing their cuticle later. In most of the Tlioracica, metamorphosis seems to run the course above described. Only in individual cases is it more abbreviated. Thus, according to KoREN and Danielssex (No. 48), the larvae of Anclasma squalicola jmss througli the greater part of their metamorphosis within the mantle-cavity. KossMANN, however, describes the Nauplii of these forms whose larvae, according to the above authors, when hatched, are provided with six pairs of limbs. Pagenstecher has rightly connected this feature with the attachment of Anelasma to sliarks. Still more ablireviated is the metamorphosis of Scalpellum Stroma, the Cypris stage of wliich, surrounded by the Nauplius cuticle, was found by Hoek (No. 45) even within the egg-envelope. B. Abdominalia. The metamorphosis of Alcippe, a form which bores into the columella of the shells of Fusus and Buccijium (especially when these are inhabited by Pagurids) seems to agree in essentials with that of the Thoracica. The Nauplii were first described by Hancock. The Cypris stage described by Darwin (No. 40) is distinguished by the possession of six pairs of thoracic limbs, a fact worth mentioning in contrast to the reduced number and shape of these limbs in the adult ; only four pairs of these are retained, the first as a palp, the second biramose, and the third uniramose. In Kocldorine, which bores into the shell of Haliotis, Noll (No. 56) found two kinds of larval forms, a small form i^rovided with adhering antennae, but without a mantle, and a larger form having a bivalve shell and resembling the Cypris stage. It is probable that the metamorphosis here is closely related to that of the nearly allied genus Cryptophialus. The metamorphosis of Cryptophialus, first described by Darwin (No. 40), is greatly abbreviated. The egg here passes almost direct into the Cypris-\\kQ larval form. There first arises from the oval egg a larva in which two processes can be recognised as the rudiments of the adhering antennae. A third jjrocess indicates the posterior end of the body. At a later stage, the adhering antennae are ap- proximated, while the body assumes generally a more pointed egg-like shape. From this stage the Cypris form emerges, in which can be distinguished the mantle-fold, the paired eyes, and the well-developed adhering antennae. The rudiments of the thoracic limbs are here wanting, but three pairs of setae are found on the abdomen. These larvae creep about by means of the adhering antennae in the mantle- cavity of the mother, and finally change into the adult form. For the development of the stalk in these forms, see p. 218. RHIZOCEPHALA. 221 C. Rhizocephala. The free-swimming stages of the Rhizocephala (NaiqjUus or Cypris stage) were early observed by Fr. Muller (Xos. 54 and 55), KossMANN (No. 49), and others. The later transformations, on the contrary, which bring about the transition from the Cypris stage to the parasitic form, were first observed by Delage (No. 41) in the case of Saccuiina carcini. We shall, therefore, in our description, follow the statemezits of Delage. The Nauplius of Saccuiina carcini leaves the egg with a more or less compact form, but by a moult which follows soon after hatching attains a greater length (Fig. 107 A). Otherwise it shows the normal Cirripede type. The two frontal horns with their glands {gl) are well developed, as also are the filamentous frontal organ (/s) and the Nauplius eye {ua), lying close to the brain. As a remains of the labium a projection is found which is known as the rostrum ; on the other hand, the oral aperture, the intestinal canal, and the anal aperture are wanting. The intestinal canal is here replaced by a large accumulation of food-yolk. The Nauplius limbs (/, 2, 3) are developed in the typical manner, but the protopodites of the two posterior pairs are without the masticatory hooks usually found on them. A median mass of cells lying within the body between the brain and the rostrum is assumed by Delage to be the ovary [ov). Fritz Muller asserted the presence of a broad, oval, dorsal shield in the Nauplius of Saccuiina. Kossmann, however, has pointed out that Muller's figures are of larvae about to undergo ecdysis, in which the Nauplius cuticle is no longer in contact with the body. The most important change found after a third moult is the further growth of the thoraco-abdominal region of the Nauplius, in which the six thoracic segments, with the rudiments of their limbs, soon appear distinct from the abdomen. At the same time the most anterior limbs be^in to assume the character of the adhering antennae [Metanauplius stage). By the fourth moult the free Cypris stage (Fig. 107 B) is reached. In this moult the biramose limbs of the Nauplius stage are entirely cast off, and not only do the chitinous envelopes of these limbs adhere to the cast skin, but it appears that even some of the soft parts are thrown off in the moult. The Cypris stage in which, as in the former stages, there is no trace of an alimentary canal, closely resembles in form the similarly-named stage in the Thoracica. The bivalve shell has almost the same shape ; the segmentation of the thorax, the form of the swimming limbs (I-VI) are in agreement. 222 CRUSTACEA. The abdomen {ah) is very rudimentary, consisting of a single joint bearing two furcal appendages. The first antenna (i) is witliout sucking disc and cement-gland. It consists of three joints : a basal joint which is broadened and connected with the chitinous apodemes, a long middle joint, and a short, terminal joint provided with three Fig. 107. — Consecutive larval stages of Sacculina carcini (after Delage, from Lang's Text-book). A, Nauplius after tlie first moult. 11, free-swimuiing Cypris stage. C, Cyj^ris stage after the larva has become attached to a seta (hb) of the host. D, formation of the Kcntrogon larva. E, the Kentrognn larva after the Cypris shell has been thrown off and the arrow has formed. F, the arrow has bored through the chitinous cuticle of the host. 1, 2, 3, the three pairs of Nauplius limbs ; I-VI, thoracic limbs ; ah, abdomen ; bh, seta of the host ; /, fat globules ; fs, frontal sensory organ ; rjl, glands of the frontal horns ; ov, rudiment of the ovary ; ff, arrow; va, Nauplius eye. hooked setae. The Nauplius eye (wa) persists, but the paired eyes are absent. Among the internal organs present at this stage, we note the very distinct ovarian cell-mass {ov), the strongly-developed musculature, the persistence of the glands of the frontal horns (yZ), and the accumulations of jDigment and of food-yolk (/), RHIZOCEPHALA. 223 The next stage which we may compare with the pupa of the Thoracica is called by Delage the Kentrogon stage. In this stage the attachment of the larva to the body of the host {Carclnus maenns), and its passage into the body-cavity of the latter are accomplished. After the free Ci/2)ris-li]<.e larva has swum about for three or four days, it seeks out a host, a young shore-crab (from 3 to 12 mm. broad), and attaches itself to one of the integumental setae (l>b), one of the adhering antennae (1) of the Cypris larva surrounding such a seta near its point of insertion (Fig. 107 C). The point at which the larva attaches itself is not, as we should expect, a priori, on the ventral surface of the abdomen (of the host), but seems to be in- discriminately chosen. Fixation often takes place on the back of the host, or on one of the legs. The next change to take place may be described as a moult in which many important parts of the body are cast off (amputated). First the soft contents of the adhering antennae are drawn in, the apodemes (chitinous tendons of the antennal muscles) being at the same time expelled from within the body. These latter remain attached for a long time to the envelopes of the adhering antennae, which are also retained for some time (Fig. 107 D), as they are of importance in bringing about the attachment of the larva to the host. While the soft parts are everywhere withdrawn from the chitinous envelope, the thorax is protruded far beyond the shell -valves and amputated in toto (Fig. 107 C). This can only be accomplished by a somewhat extensive rupture of the body-wall, and, through this, remnants of internal organs are thrust out into the larval shell and lost. Thus a great part of the pigment found in the larva, as well as remains of food-yolk, are thrown out, the frontal glands and the whole body musculature undergoes degeneration, and the masses of detritus thus produced, together with the Nauplius eye, are now eliminated. The remainder of the body left after the separation of all these organs draws together to form a solid oval sac (Fig. 107 D), which soon becomes surrounded with a chitinous envelope. This latter is closely contiguous to the sac; only at its most anterior end, that turned towards the adhering antennae, it can be noticed that the soft body seems to lie naked against the inner sides of those organs. The newly-formed envelope is in this region probably exceedingly delicate, and very closely apposed to the inner surface of the antennae. The layers into which the contents of the sac break up are at this stage not at all clear. A superficial ectodermal cell-layer can, however, be distinguished from the inner mass which, in all probability, is meso- 224 CRUSTACEA, dermal, and the chief constituent of which is the cell-mass of the ovary. Besides remains of pigment and food-yolk, other mesodermal elements no doubt enter into the formation of this second layer, from which are to he derived the rudiments of the testes as well as the musculature and other organs of the adult Sacculina. It is important to bear in mind that the encysted sac thus produced has been derived, after the expulsion of the whole thorax, exclusively from the cephalic section of the Ci/pris larva. The soft body of the sac-like larva now begins to develop a small fine point at its anterior end (Fig. 107 D), which protrudes into the inner cavity of the antenna used for at- tachment, and this is followed by the secret- ing of a new chitinous envelope at the surface of the soft body (second moult of the Kentrogon stage, Fig. 107 ^). As the new cuticular layer thickens considerably over this anterior point, it forms the arrow-like tube which character- ises the Kentrogon stage. This tube, in- creasing in length and becoming somewhat bent, invaginates the anterior surface of the sac (Fig. 107 E). At this stage the cast ofT Cypris shell is either only very loosely at- tached to the sac, or is even completely severed from it. The protrusion of the arrow now takes place (Fig. 107 F, pf), the invagination just described being re-evaginated. The arrow first enters the inner cavity of the adhering antenna, and thence, con- ducted by the latter, passes into the soft articular membrane of the am Fig. lOS. — Two sections through the nucleus of a Sacculina interim (after Del age). A, younger stage. B, older stage, am, outer mantle-layer ; im, inner mantle-layer ; m, meSoclerm-cells ; o, aperture of the perivisceral cavity; ov, rudiment of the ovary ; p, perivisceral cavity. RHIZOCEPHALA. 225 seta to which the Cyj^^'^s-like larva attached itself. Through this membrane the arrow bores, thus establishing a communication between the inner cavity of the sac and the body-cavity of the host. During these processes the soft inner body becomes surrounded with another very delicate cuticle (third moult of the Kentrogon stage). There is at this point a gap in our knowledge of the development of Saccnlina. There can, however, be no doubt that the soft body Fig. 109.— Longitudinal sections through two stages of development ot Saccnlina carcini (after Delage). .-1, Saccnlina interna. B, Saccnlina externa, a, atrium (widening of the oviduct); a»)i, outer mantle-layer ; &, brood-ca\'ity (mantle-cavity) ; B, basal membrane ; C, central tumour ; cl, cloacal opening ; D, intestinal wall of the host ; dr, glands of the ovarian sac ; /, aperture of the perivisceral cavity ; g, ganglion ; im, inner mantle-layer ; L, body- wall of the host ; ov, ovary ; p, perivisceral cavity ; pe, perivisceral ectodermal layer ; R, root- processes (some in cross section) ; t, rudiment of testes. of the larva passes through the canal of the arrow, so as, in this way, to reach the body-cavity of the host. The Sacculina thus becomes an endoparasite {^Sacculina interna). Q 226 CRUSTACEA. Sacculina interna. The endoparasitic larva now Avanders from the point at which the attachment of the Cijpris hirva occurred, further into the host, until it reaches the ventral side of the intestinal canal where its final fixation takes place. At the same time it sends out an exceedingly wide-spread network of rootlets which penetrate throughout the whole body of the host, covering superficially all the organs and only leaving the heart and the gills unaffected. At the point where the actual Sacculina lies, all the rootlets gather together to form a plate {hasal memhrane, Fig. 109, B), in the middle of which can be seen a swelling {central tumour, Fig. 109, C). The body rudiment of the Sacculina lies sunk in this central tumour as its so-called Jiucleus. The rootlets, the basal membrane, and the central tumour show essentially the same histological structure. They consist of a superficial epithelium (ectoderm) and an inner cavernous tissue composed of star-like anastomosing cells of connective tissue. The nucleus (Fig. 108) is completely sunk in the central tumour, and therefore lies in a cavity which is called by Delage the perivisceral cavity (^)), and which opens externally through one small aperture. Even this aperture closes (Fig. 108, B), to re-appear later in the form of a transverse fissure (Fig. 109, /, fente de sortie). The point at which the nucleus is in contact with the wall of the perivisceral cavity is now known as the stalk (peduncle) of the nucleus. In the nucleus itself, a superficial ectodermal layer can be dis- tin^'uished (Fig. 108 ^) which, near the stalk, passes into the wall of the perivisceral cavity. The inner mass of the nucleus at this stage consists almost exclusively of the rudiment of the ovary (ov) ; but in the stalk there are some mesoderm-cells which are of importance in connection with the development of the testes, the musculature, the connective tissue, &c. Delamination next takes place in the ectodermal layer of the nucleus, which divides into two layers that shift apart (Fig. 108 B). Into the space between the two layers, some of the mesoderm-cells just mentioned wander to yield the musculature of the mantle. The two layers of ectoderm which have thus arisen are known as the Older {am) and inner {im) mantle-layers, on account of their relation to the future mantle of the Sacculina. A second similar process of delamination now takes place in the inner mantle-layer, an inner ectodermal layer surrounding the central part of the nucleus being thus split off. The latter, as surrounding the future visceral sac, is distinguished as the perivisceral ectodermal layer {pe). Between it and the inner mantle-layer there now appears a cavity lined with chitin (Fig. 109, h); this is the so-called brood-cavity {cavite incubatrice). EHIZOCEPHALA. 227 We thus see that, by changes in the nucleus, the most important parts of the body of the adult SaccuUna begin to be formed ; viz., the inner visceral sac, the brood-cavity, and the mantle-fold, which surrounds it. The visceral sac is not completely encircled by the brood-cavity, for, at the point of attachment of the sac, the inner mantle-layer passes over into the perivisceral ectodermal layer ; this transition point is the so-called mesentery, which Delage assumes to liave a ventral position in the body. Having now obtained a general idea of the developnaent of the form of the body in Sacculina, vre must add a few words as to the rise of its most important organs. In the mantle region the changes are not great. In the late stages, the mantle-cavity breaks through into the perivisceral cavity, and thiis arises the cloacal aperture (Fig. 109, d), wdiich lies almost opposite the stalk of the Saccidina, but somewhat to the left side of the body. While the ectodermal ■cells of the mantle lengthen to form transverse connective fibres (Fig. 109 B), the mesoderm-eells change into longitudinal muscle-strands and the sphincter of the cloaca. More important alterations occur in the region of the visceral sac. The ■ganglion (g) here forms by an immigration of ectoderm-cells, in which not only the perivisceral ectodermal layer, but also the inner mantle-layer (by means of the mesenterial margin ?) are said to take part. Whereas, in earlier stages, the whole inner space of the visceral sac was almost exclusively occupied by the ovarian rudiment, numerous mesenchyme-cclls now^ wander from the stalk into the perivisceral sac, surround the ovary, forming a peritoneal envelope around it, and fill the space between the body-wall, the ganglion, and the ovary. The ovarian rudiment simultaneously breaks up into two lateral lobes connected by a commissure. The manner in which the short oviduct arises is not quite •clear, but Delage believes that he can trace it back to a paired lateral ecto- dermal invagination. This latter, widening inwardly, gives rise to the so-called atria (a), on whose walls the glands of the ovarian sac (cement glands, dr) •appear as lateral outgrowths. The vasa deferentia arise in the same way through ectodermal invaginations near the stalk of the visceral sac, while the actual testes [t) are derived from mesoderm-cells which become attached to the ■ends of these ducts. After the Sacculina, completely enclosed within the central tumour ■(in the perivisceral cavity), has in almost every respect attained the grade of development of the adult, it rises to the surface of the central tumour, passing out through the widened opening of the perivisceral •cavity (Fig. 109 A, /). The fold by Avhich this cavity was formed now draws back to the base of the stalk and soon entirely disappears. An increase in size takes place in the Sacculina after leaving the central tumour, and this causes constant pressure on the ventral wall of the abdomen of the host (Fig. 109 B, L) leading to gangrene of the part thus affected, and to the consequent formation of an opening through which the body of the SaccuUna passes out freely, its stalk 228 CRUSTACEA. still connecting it with the basal plate and the network of rootlets within the host. The Sacculina is, by these processes, changed into a SaccuUna externa (Fig. 109 B). The parts that lie outside of the host now become strongly chitinised, there is further increase in size, and the stage of sexual maturity is reached. The metamorphosis of Sacculina above described is without doubt one of the most remarkable processes of transformation in the animal kingdom. The intercalation of a temporary endoparasitic condition must no doubt be referred to the protected position thus obtained^ and, indeed, the whole process of development of this form has undergone marked coenogenetic changes. Although, considering the unusual simplification in structure of the Kentrorjon larva, we may not be ontogenetically justified in tracing back the various parts of the adult body to those of the GyxJvis-VikB larva, the consideration of other forms {e.g., AneJasma) leaves us not a moment in doubt as to how the body of the adult SaccuUna is to be interpreted. Such consideration would lead us to compare the peduncle of the Sacculina, running out at its base into root-like processes, to the stalk of the Lepadidae, the mantle of the SaccuUna to the shell of the latter, and the brood-cavity of Sacculina to the mantle-cavity of the Lepadidae. The cloacal aperture would then correspond to the shell-cleft in the latter family. This interpretation is supported most of all by the similar position of the ovarian sacs in these cavities. It becomes probable, on comparison with Anelasma, that Delage's definition of the mesenterial margin as the ventral side of the Sacculina is actually correct. Frequent attempts have recently been made to oppose the Khizo- cephala to all other Cirripedia as an independent group (sub-order). On the other hand, it must be pointed out that they, in the NaupUus and the Cypris stage, show such complete agreement with other Cirripedes that too great stress must not be laid on changes of structure which have evidently arisen secondarily as the result of parasitic life. D. Ascothoracida. This group comprises a few forms {Laura Gerardiae, Synagoga mira, Petrarca hathyacUdis) which live parasitically in Anthozoa, and Dendrogaster astericola, an endoparasite in the Asteroidea {Solaster, Echinaster). In the most important features of their organisation,, these forms are true Cirripedes, although they claim a special position within that group. Laura Gerardiae, thanks to the re- THE MORPHOLOGICAL DERIVATION OF THE COMPLEMENTAL MALE. 229 searches of Lacaze-Duthiers (No. 51), is the member of this group which we know best. The animal is surrounded by a large mantle which here shows direct relation to the shell-valves of the Cypris stage ; each half contains between its two lamellae not only the hepatic outgrowths of the intestinal canal, but also the ovaries. The body proper appears very much reduced, but is still distinctly segmented, the mouth-parts are adapted for sucking, the six (or five) pairs of thoracic limbs have degenerated, and the abdomen is short. It should be mentioned as a characteristic of the group that the first antennae are here never used, as in other Cirripedes, for the ■attachment of the body. For the general morphological elucidation of these forms indeed, we must compare them not with the adult Lepas, but rather with the free-swimming Cijpris larva. Very little has been made known, up to the present time, of the ontogeny of these forms. The cleavage of the egg in Laura seems to resemble that in Balanus. The Nauplii oi Laura show very little similarity with ^ the typical Cirripede Nauplius, on account of the absence of the very characteristic frontal horns. A small form observed by Lacaze-Duthiees, and no doubt representing a stage in the course of development of Laura, may possibly be regarded as a complemental male. The Cypris stages of Dendrogaster (Fig. 110), however, are known (Knipowitsch, No. 47), the metamorphosis in this form appearing to be abbreviated, no free Na uplius stage being passed through. The larva, in which an anal aperture, as in the adult, is wanting, bears a general re- semblance to the Cypris stage of the Cirripedia. Both the single and the compound eyes are, however, wanting. A very large olfactory filament («') is develojied on the first antenna. Tliere are five biramose pairs of thoracic limbs ; the first abdominal segment carries the rudiment of a penis {p). The abdomen, [ah) which is distinguished by its length, consists of five joints and tlie furcal appendages. Fig. 110. — Free-swimming Cypris stage of Dendogaster astericola (after Knipowitsch). a, first antenna ; ab, abdomen ; d, intestine ; m, buccal cone ; n, nervous system ; p, rudiment of penis. E. The Morphological Derivation of the Complemental Male. The sexual differentiation of the Cirripedia is very complicated and difficult to explain. As a rule the Cirripedia are hermaphrodite. We shall not err if, considering the almost universal separation of 230 CRUSTACEA. the sexes in all other Crustacea, ^^-e regard this hermaphroditism as secondarily acquired in consequence of the attached manner of life. We must assume that the free-swimming ancestors of the Cirripedia were of separate sexes, and that hermaphroditism was only gradually acquired after the fixed sedentary mode of life had become fully established as a characteristic of this group. While, in the Balanidae and Khizocephala, hermaphroditism has become the exclusively prevailing condition, there is a tendency in many groups of the Lepadidae to retrogression in the direction of the separation of the sexes. Male forms here appear either side by side with herma- phrodite individuals, being then called complemental males, or else, in cases of the complete separation of the sexes, in company with true females. These complemental males are always smaller tlian the hermaphrodite individuals or the females ; they are found attached parasitically to the bodies of the hermaphrodites or females. In isolated cases, the shape of the body in the male deviates only slightly from that in the hermaphrodite (Scalpellum villosum and S. Peronii), but in other cases there is striking sexual dimorphism in this respect, the male undergoing a process of degeneration in which the calcareous portions of the skeleton, the limbs, the mouth, and the alimentary canal have been lost, thus sinking to a very low grade of organisation, and becoming actually a dwarf or complemental male. The following degrees of degeneration of the male are found in the genus Scaliiellum (Hoek, No. 46). I. True hermaphrodite forms [Scal2)ellum balanoides, Hoek). II. Large hermaphrodite forms with small complemental males, (a) The males resemble the hermaphrodites in structure. The division into capitulum and peduncle is recognisable, mouth and alimentary canal are present [Scalpellum villosum, S. Peronii). (h) The males have degenerated ; they are without mouth and intestine ; without a shell, or have only a vestigial shell ; without a peduncle [Scalpellum vidgare, S. rostratum). III. Separate sexes. The female is large and resembles the hermaphrodite individuals of the other species. The male is very small [Scalpellum ornatum, S. regium, Hoek, etc.). The Abdominalia {Alcippe, Cryptophudus) show a differentiation akin to that of this last groiqi. Here also we find separation of the sexes with highly developed sexual dimorphism. The complemental males appear very much reduced. They have no tendril-like feet, no mouth, and no alimentary canal. In other respects, if we take into account the reduction that has commenced, their structure can be traced back to that of the female. A condition resembling that above described for Scalpellum is found in the COPEPODA. 231 gemis Ihla. In Ibla qitadrivalvis, side by side with the hermaphrodite form, there is a small compleraental male with a large peduncle, but a very reduced capitulum and a diminished number of thoracic limbs, while in Ibla Cumingii a similar male is found side by side with a true female, complete separation of the sexes being here obtained. It follows from the above that, in accordance with H"Oek (No. 46), in tracing the sexual condition of the Cirripedes, we start from the hermaphrodite form, proceeding to others with dwarf males in company with hermaphrodite individuals, and reach finally a complete separation of the sexes in which marked sexual dimorpliism has secondarily developed. The series of com- pleraental males, as well as the dwarf males, would thus be derived from the hermaphrodite form by the degeneration of the female genital rudiment. It is perhaps difficult to imagine how the want of complemental males can in the first instance have made itself felt in hermaphroditic forms. But we must bear in mind that, according to F. Muller, cross-fertilisation appears to be the rule even among the true hermaphroditic Lepadidac, This cross- fertilisation is of the greatest importance in securing the vitality of the race ; but owing to the fixed mode of life of the Cirripedia there is some danger that it may not take place. All such danger is precluded by the development of these dwarfed forms living semi-parasitically on the larger hermaphrodites. The dwarfs, by the atrophy of their ovaries, have become males, and by the attachment to the hermaphrodites ensure cross-fertilisation. We must regard this condition as a partial retrogression in the direction of separation of the sexes, which has again been reached in individual cases in the further course of this development. The view sketched above is in opposition to that of Claus (No. 8), according to which the sexually distinct forms {e.g., Alcippe and Cryptophialus) have re- tained the primitive condition. From this originally dioecious condition, by a transformation of the females into the large hermaphrodite form, the condition found in most Lepadidac was developed, while the males were only retained in isolated species as complemental males. Consequently the males would be a vestigial remainder from those times when hermaphroditism had not yet become the rule among the Cirripedia. The occurrence of dwarf males side by side with hermaphrodite individuals in a number of forms would be easily explicable by this ingenious theory (not, however, in Scalpellum villosum and ^S". Peronii). HOEK, however, has pointed out in opposition to it, that, in this case, the dwarf males ought to show a greater resemblance in structirre to the Cypris form than they actually do show. The dwarf males in reality appear to be connected by gradual transitions with the complemental males of Scalpellum villosum and ^S*. Peronii, which latter forms are evidently to be derived from the herma^jhrodite form. The view held by Claus would apply to the Rhizocephala if the statements of Fe. Muller and Delage were to prove true that complemental males occur in these forms which, throughout life, do not develop further than the Cypris pupa. But this last view, which rests upon the discovery in the cloacal aperture of a young Saceulina externa of an attached dead Cypris envelope, must still be considered doubtful, and has actually been denied by Giard. 6. Copepoda. The Copepoda are very numerous and rich in varieties of form, in spite of the simplicity of their body segmentation ; they never- theless show morphological characters which, in relation to those 232 CRUSTACEA. of the hypothetical ancestor of the Entomostraca, must be regarded as decidedly degenerate. Among these we must reckon the small size of the body and the comparatively small number of its segments, the reduced form or entire absence of the heart, the want of separate respiratory organs (branchiae), the loss of the paired lateral eyes, which are retained only in the families Conj- caeidae and Pontellidae, and perhaps also the slight development of the dorsal shield. On the other hand, there are certain indications that the Copepoda ought to be counted among the most primitive of existing Crustacea. Among these characters we should specially mention the use of the two pairs of antennae as locomotory and clasping organs, the very primitive structure of the mouth-parts in the free-living forms (occurrence of a biramose mandibular palp, the segmentation of the first maxilla, Fig. 91 A, p. 194), and the meta- morphosis which, in the free-living forms, shows very primitive features. "With regard to the segmentation of the body, we must distinguish as the most anterior region of the body a simple cephalic region carrying the antennae and the mouth-parts. The latter consist of three pairs of appendages (mandibles, first and second maxillae), the last pair separating into two appendages (the exopodite which shifts forward is called the first maxillipede, tlie endopodite which shifts backwards yields the second maxillipede). The thoracic region con- sists of five segments provided with biramose swimming limbs (Fig. 90 A, p. 193); the last of these segments may be vestigial, whilst the anterior segment often fuses Avith the cephalic segments, this union giving rise to an anterior region known as the cephalo-thorax. The abdomen consists of five segments, the most anterior of Avhich alone carries the rudiment of a limb (genital prominence). A fusion of the two anterior abdominal segments usually gives rise, in the female, to a double genital segment which bears the genital aperture. In a few Pontellidae, the ceplialo-thoracic region, by tlie apjwarance of demarcations between the segments, becomes subdivided into regions (each consisting of two segments). This peculiarity stands almost alone in the whole series of Crustacea. Such a segmentation can only be regarded as a secondary re-appearance of the long-lost independence of the cephalic segment. It has, nevertheless, a certain interest. A. Gnathostomata. The metamorphosis of the free-living Copepoda is accomplished as a very gradual transition from the Nauplius to the adult form through many moults. There is, hoAvever, at a certain period, a more sudden GNATHOSTOiMATA. 233 change of shape, and this enables us to separate the course of metamorphosis into two periods, the first comprising the series of NaiqMus and Metanauplius forms, while the second is distinguished by a name taken from the metamorphosis of the Cyclopidae as the series of Cyclops-Vike larval forms. In the first series, the Xauplius limbs show a general resemblance to the primitive form, the abdomen is not yet distinctly marked off, and the furcal processes have not yet Fig. 111.— Larval stages of Cetochilus septentrionalis (after Grobben, from Lang's Text-hook). A, Naiiplius. B, Metanauplius. C, older Metanauplius. I and 3, first and second antennae ; 3, mandible; A, maxilla; 5, 6, exopodite and endopodite of the second maxilla ( = first and .second maxillipedes) ; /, II, first and second pairs of thoracic limbs ; an, anus ; g, brain ; gz, genital cells ; in, mouth ; me, primitive mesoderm-cells ; ol, upjier lip. developed. In the second series, these last advances in development are made, while the antennae and mouth-parts approach the adult form. The development of the free-living Copepoda has been specially investigated by Claus (Nos. 64 and 67). We take, as the foundation of our description, the metamorphosis of Cetochilus, which has been minutely described by Grobben (No. 73). The free-living Copepod leaves the egg as a strikingly unspecialised Nauplius (Fig. Ill A). 234 CRUSTACEA. The body, which is usually oval (only in individual cases long, trans- versely broadened or barrel-shaped), shows no traces of external segmentation, and carries on its ventral side the large upper lip (oZ), as well as the three typical pairs of Nauplius limbs (i, S, 3). The anterior limb (first antenna) is uniramose, the next (second antenna)- biramose and provided with a masticatory hook on the protopodite. The mandible in CetocMlus is without such a masticatory process, and is a simple biramose swimming limb. The limbs are beset at their ends with long setae. In the alimentary canal we can distinguish a stomodaeum, a long enteron, and a proctodaeum. The latter, in CetocMlus, in the first NmqyUus stage, is still found as a solid ingrowth of the ectoderm, the anal aperture not yet having broken through. The nervous system is still connected for its whole course with the ectoderm. As a sensory organ, we find the Nauplius eye. The coiled antennal gland, which is probably developed at this stage, functions as^ excretory organ, besides which the cells of the enteron seem to have undertaken an excretory function ; urinary concretions- have, at least, been proved to exist in certain projecting cells- in the enteron of the Nauplius of Cyclops (Leydig, Fig. 89, ds, p. 191). The terminal region of the Nauplius is sharply bent ventrally,. and provided with two setae. We here find, internally on each side, a large mesoderm-cell {ine) ; these are assumed by Grobben to be primitive mesoderm-cells. Later Nauplius stages are distinguished by the greater length of the body and by the outgrowth of its posterior region. During this latter process, the more strongly chitinised integument of the dorsal parts becomes marked off as the cephalo-thoracic shield by the de- velopment of a fold at its margin. The proctodaeum has now become perforated, a distinct dorsally-placed anal aperture being apparent. The brain is connected behind the Nauplius eye with a paired ectodermal growth, in which can be recognised the rudiments of paired lateral eyes and their optic ganglia (secondary brain) ; these,, at a later period, become vestigial. The rudiment of the genital organs is to be recognised in large mesoderm-cells lying one on each side of the alimentary canal (Fig. Ill B). There now appears, behind the mandible, a small biramose limb^ the rudiment of the first maxilla (^), and the larva passes into the first Metana^iplius stage (Fig. Ill B). A later Metanauplius stage (Fig. Ill (7) reveals three more rudi- GNATHOSTOMATA. 235- ments of limbs, viz., the second maxillae (-5), from the two branches of which the so-called maxillipedes of the Copepoda are derived, and the first two pairs of thoracic limbs (/, //). This stage still has a distinct ISTauplius appearance. The body has grown in length, but still, when seen from the side, shows the characteristic ventral curvature. The posterior end of the body is still without the furcal processes. The two pairs of antennae have not essentially changed from their condition in the former stage, except in the increased number of their setae. The masticatory hooks are still present on the basal joints of the second antennae. In the mandible (3) a large masticatory blade can be seen projecting from the basal joint. The first maxilla {4) is like a small lobed plate, while, in the second maxilla (o), indications of the separation of the exopodite (so-called anterior or outer maxillipede) from the endopodite (so-called posterior or inner maxillipede) are to be seen. The rudiments of the two anterior pairs of thoracic limbs are found as bilobed swellings (/, //). At this period, several moults take place, no essential change occur- ring in the shape of the larva, beyond the appearance behind the second thoracic segment of the rudiment of a third. The whole series ends by a moult through which the larva passes on into the series of Cf/dojjs-like stages. The first of these stages, which, in accordance with the accepted terminology of the Copepodan metamorphoses, we shall call the Jii'st Cyclops stage (more correctly named Cetochilus stage by Grobben), reveals essential changes of form. The body is no longer flexed ventrally, but is straight. Its most posterior region is sharply marked off from the anterior portion of the body by a constriction ; the furcal appendages and the rudiment of a fourth thoracic segment have developed. The limbs approximate in shape to those of the adult, although they have not so many joints. The first antenna has passed from the short leaf-shape into the long, cylindrical, oar-like form ; it stands out laterally from the body, and consists of many joints. The second antenna has remained biramose, but has lost its masticatory process ; on the mandible, the masticatory blade has greatly increased in size. The maxilla is larger and more richly jointed, the maxillipedes have become transformed into large prehensile organs. The two anterior pairs of thoracic limbs have developed as swimming limbs ; the basal segment of each already consists of two joints, but the two rami are still unjointed ; the third pair of thoracic limbs, on the contrary, can only be recognised as a pair of bilobed rudiments. 236 CRUSTACEA. Among the transformations which take place in the internal organs during the Cxjdops stages, we must mention the degeneration of the paired optic rudiment and of the secondary brain. Now for the first time the nerves which mn to the paired frontal organs are distinctly recognisable. The antennal gland degenerates, while the shell-gland, which opens externally at the base of the anterior maxillipede, becomes functional in its stead. The anal aperture no longer lies dorsally, but shifts to the posterior end of the body between the two furcal appendages. The genital organs show an advance by an increase in number of genital cells and by the development of the efferent ducts. The paired genital rudiments now meet above the intestine and fuse to form a single gland. The heart develops between the first and second thoracic segments out of a paired rudiment of mesoderm-cells. Fig. 112.— Two stages of development of Canthommptus staphylinus (after Claus). A, Meta- nauplius stage. B, Cyclops stage with three pairs of swimming limbs, a', first antenna ; a", second antenna; md, mandible ; mx, maxilla; mf, second maxilla (rudiment of the two so-called maxillipedes) ; mfi, first so-called maxillipede ; tnfn, second so-called maxillipede ; pi, pii, pill, first, second, and third jiairs of thoracic limbs. In the seco7id Cyclops stage (Fig. 112 B), there first appear tlie rudiments of the limbs of the fourth thoracic segment and the delimitation of the fifth thoracic segment. Tlius we find, at this stage, behind the cephalo-thorax, four free thoracic segments, and, following these, the as yet unsegmented abdomen. Of the thoracic limbs, the three anterior pairs are well developed. In the third Ci/dops stage, the fourth pair of thoracic limbs has attained full development and tlie first abdominal segment has formed ; in the following Cgclops stages, the gradual segmentation of the abdomen takes place, as well as the complete transition of the limbs to their PARASITA. 237 final shape. Besides the formation of a more richly segmented body, in the famiHes of the Cijdopidae and Corycaeidae, the (inner) accessory branch of the second antenna is lost as early as the first Cyclops stage, and the mandibular palp degenerates. The metamorphosis of the Calanidae (Cetochilus), which has here been considered as the type, is distinguished by the regular development of t!ie limbs from before backwards, but in the Harpactidac (Fig. 112) and Cijdopidae an exception to this regular order is found in the second maxilla {mf), which in the later Nauplius stages has, indeed, begun to form, but is still in an exceedingly rudimentary condition, so that the pair of limbs which follow it precede it in development. ^Ye have here a parallel to the condition of the maxilla in the Phyllopoda. B. Parasita. The free-living Copepoda are connected with the more specialised parasitic genera by many transitional forms ■\vlnch mark the various degrees of parasitism, and we consequently also find various stages in the transformation and degeneration of the segmentation of the body. It may be laid down as a general rule that the female, on account of the part j^layed by her in reproduction, shows a greater tendency to adopt the parasitic mode of life, and consequently to undergo degeneration of the locomotory organs, obliteration of the boundaries between the segments and deformation of the body. Thus, even in the little modified SajDphirina, we find that the females become parasitic in the respiratory cavity of the Tunicata, or in the umbrellar cavity of Diphyes, while the males are always found swimming freely about. An extreme example is found in Lernaea (Fig. 114 J. and B), the metamorphosis of which ends in a Cydops-Wke free-swimming stage, during which copulation takes place ; the female afterwards becomes attached to a fish (often one of the Gadidae), and undergoes great deformation of the body (Fig. 114 C and D). Heteromorphism of the sexes develops in this group, in so far as the male is only slightly removed in the segmentation of its body from the later Cyclops stages, while the female is greatly transformed in accordance with her parasitic habit. The same is the case in the Philichthyidae and the Chondracanthidae. In some forms, on the contrary, the male also departs from the Cyclops-\\ke shape of the later larval stages by a secondary transformation. Whereas, in the Caligidae and Dichelestiidae, the two sexes are not strikingly diff"erent in shape of body or size, in the Lernaeopoda (Fig. 115 Z> and E), the heteromorphous development of the two sexes takes place in a way different from that shown in Lernaea. The males here 238 ■ CBUSTACBA. participate in the degeneration of the body -segments as a result of their parasitic life, and they further undergo arrest of growth, so that an enormously large female is contrasted with a dwarf male. This kind of heteromorphous development of the two sexes must he regarded as an excessive adaptation to the different sexual functions, which is rendered possible by parasitic life. The parasitic forms can be deduced from the free-living forms by imagining that the latter have, in consequence of parasitism, under- gone certain changes. Thus the parasitic Copepoda first passed through Metanauplius and Cyclops stages to a stage approaching the shape of the free-living form, and then, through a series of further stages, attained parasitic deformation. The metamorphosis of the parasitic Copepoda has thus been lengthened by the addition of final parasitic stages. The first two series of stages, however, seem to be correspondinglj^ shortened. The larvae of the parasitic Copepoda frequently do not hatch in the Naiqjiius form, but in -an advanced Metanauplius stage, or even in the Cyclops stage (Fig. 73, p. 148; Cltondracantlms, Tracheliastes, Achtheres, AnrJiorella, Brac.hiella, etc.). On the other hand, the metamorphosis may be shortened by the suppression of the later Cyclops stages, since, in cases of the most specialised parasitic forms, the very first Cyclops stage passes at once into the parasitic form {Chondracanthus, Lernaeop)ocUdae). A further distinction between the metamorphosis of the parasitic Copepoda and that of the free-living forms arises from the circum- stance that, even in the larval condition, a sedentary manner of life (on the gills of a host) is adopted, and that, in keeping with this, there is a development of a peculiar attaching organ (the frontal band of the larvae in the Caligidae, Lernaea, and Lernaeopoda) and of resting stages with reduced limbs (so-called pupal stages). It would take us lieyond the limits we have assigned to ourselves to give a complete enumeration of the very scattered notices of individual larval forms among the parasitic Copepoda, especially as there are still many gaps in the observations made on the development of these forms. We must content ourselves with selecting a few of the more important forms, of whose meta- morphosis a more accurate knowledge has been obtained. We must here, in the first place, separate those larvae which apparently are not provided with the larval adhering apparatus (frontal band) from those in which such an organ has been observed. In those families in which the adult retains more or less of the body seg- mentation of the free-living Copepoda (e.g., Corycaeidae and the Notodelphyidae, which latter is placed among the Gnathostomata), the metamorphosis appears not to differ essentially from that above described for free-living forms. In the PARASITA. 239 Chondracanthidae, on the contrary, we find the above-mentioned abbreviations ■of metamorphosis. The young hirvae which hatch from the egg of Chondra- ■canthus gihhosus already show behind the ISTauplius limbs the rudiments ■of two other pairs of limbs, and must therefore be described as Mdanaiqjlii (Claus, No. 71). The youngest parasitic females remain essentially at the level of development of the fii'st Cyclops stage. Of the four distinct thoracic segments seen at this stage, only the two anterior ones carry bilobate rudiments of limbs without setae, while the posterior region of the body (abdomen) is small and divided into two parts. No further Cyclops forms follow these first stages, but while the small male remains during life at this grade of develop- ment, the female undergoes a secondary transformation, the region of the third and fourth tlioracic segments increases in size and forms the main portion of tlie body. The large thoracic region now becomes transformed in an extraordinary manner, dorsal and ventral swellings and lateral projections appearing as secondary outgrowths on each of the thoracic segments (except the first). Somewhat similar phenomena are found in the family of the Philidhyidae. Here the larval form which hatches from the egg in tlie genus Lernacascus is a A^'upJiics with a large provision of food-yolk, whose second antennae are without masticatory hooks, the adhering apparatus (frontal band) being absent, as in the Metanauplius of Chondracanthus. The parasitic form arises from a Cyclops stage in wliich the thorax and abdomen are distinctly segmented, but only the two anterior thoracic segments show a well-developed rudiment of a swimming limb, while the third thoracic segment has only the vestigial remains of a limb. The male retains the shape of one of these developmental stages, while the female is parasitically transform.ed, the thoracic region lengthening, while a peculiar development of asymmetrical rows of chitinous scales makes its appearance (Claus, No. 69). In the family just described, the ascending series of larval forms does not appreciably rise above the level of the first Cyclops stage, but in the Dichelestiidac, where the body of the adult undergoes less modification, the later Cyclops stages are also passed through. The young which hatch from the egg are true NaiqMi. The frontal band appears to be wanting (?) in the larvae of this group. An adhering apparatus of this kind is found in the larvae of the Caligidae, which, in their earlier stages, strikingly resemble the Cyclops stage (pupae) of Lcrnaca (Claus, No. 70, see below). The later larvae, which more nearly Fig. 113. — Two larval stages of Lernaea hrancMalis (after Claus). A, first Cydo%is stage. B, so- called pupal stage, a', first antenna ; a", second antenna; //, first, ///, second pair of thoracic limbs ; Ic, adhesive mass ; oc, eye. 240 CRUSTACEA. resemble the adult in form, but are still distinguished by the possession of the frontal band, were described by Burmeistee as distinct forms under the name of Chcdimus. Later, however, F. Muller jiroved that these forms belong to the ontogeny of Caligus, as had already been conjectured by Kroyer. The Lcmaeidac are very interesting with respect not only to the parasitism of the larval forms, but to the deformation that occurs in the body of the female after copulation. The metamorphosis of Lernaea hrunchialis has been made known by Metzger and Claxts (No. 70). It is probable that the larva which hatches from the egg is a Nauplius resembling that of Aclitheres, in which the segmentation of the body of the first Cyclops stage appears to be commencing beneath the cuticle. In these stages there is a free-swimming period, during which a search is made for the first host (Platessa flesus). The youngest forms found attached to the gills still show in all respects the segmentation of the first Cyclops stage (Fig. 113 A). They correspond to the first stage of development of Achtlieres. The cephalo-thoracic region is followed by three free thoracic segments and a posterior unsegraented region carrying the furcal processes. Two pairs of well-developed swimming limbs {fl, f^^) can be recognised, one attached to the cephalo-thorax and the other to the first free thoracic segment, as well as a third pair of truncated limbs (on the second free segment). The mouth-parts are already of the true Siphonostomatous type. The upper and lower lips (labruni and paragnatha ?) have fused to form a sucking tube which contains the stylet- shaped mandibles, while the pointed 2)alp-like maxillae are attached to its sides. The first antennae [a') are beset with setae, the second antennae («"), as well as the anterior maxillipedes, are changed into hooks for attachment. The posterior maxillipedes have completely degenerated, a point in which this larva differs from that of Aclitheres. The later Cyclops stages which succeed each other (Fig. 113 B) show decided adaptation to parasitic life. A hardened mass of secretion (k), projecting from the head and comparable to the frontal band of the larva of Caligus, brings about the attachment of the larva to the gills of the host ; this permanent attachment precedes the degeneration of the locomotory organs. Almost all the limbs, and especially the swimming limbs {fl, f^^), are now unjointed and truncated ; they have no setae and are immovable. These stages in which independent movement is lost have also received the name of pupal stages. In this pupal condition the segments of the body and the pairs of limbs which were still wanting are developed. A stage can be distinguished with three pairs of swimming limits and four free thoracic segments ; at this period, in the male, the posterior niaxillipede, until now suppressed, becomes distinct; then a further stage is developed with four pairs of swimming limbs ; this last stage leads through another moult to the free-swimming stage in which copulation takes place (Fig. 114 A and B). Apart from the slight segmentation of the abdomen, the body shows the full development attained by free-living Copepoda. The first antennae {a') are now jointed and beset with setae and sensory filaments, the four pairs of swimming limbs {fl, /^''), witli their clothing of setae, are adapted for active swimming, while in the structure of the second antennae {a") and of the mouth-parts, the Siphonostomatous type is marked. The female (Fig. 114 B) is distinguished by a great lengthening of the genital segments, which gives the whole abdomen the appearance of a long, vermiform appendage. The female genital organs are not yet sufficiently developed for the production of eggs capable of fertilisation ; the receptaculum seminis, on the other hand, with the two pores (g) for the reception of the seminal masses from the PARASITA. 241 spennatoplioi'es, has attained full development. Tins free-swimming copulatory Cyclops stage is the last stage of the male, the fertilised female, however, in which she undergoes marked trans- seeks a new host (one of the Gudidae), formation of the body (Fig. 114 C and D). The genital segment, which has enlarged for the development of the eggs, is now a large doubly -curved portion of tlie body, the small abdomen, with its truncated furcal appendages, forming its termina- tion. The cephalo- thorax is changed by the addition of three horns which function as barbed hooks, carrying at their points fork -like out- growths. During these transforma- tions, the limbs are all retained, but are to a certain extent trans- formed 1 ly strong chitinisation. The remarkable form Sjjliacronella Lcuckartu, which is parasitic in the brood- cavity of Amphitlioe, is connected with Lernaea by its metamorphosis. Salkxsky (No. 80) found in Sphaeronella an extremely degene- rate pupal stage fol- lowing the first free- swimming Cyclops stage. Neither seg- mentation nor limbs were recognisable in the sac-like body which was attached by a larval adhering apparatus to the epimeral plates of tlie host, led, through gradual transitionary stages, to the adult form. The metamorphosis of the Lernaeopodidac is best known through the works Fig. 114. — Sexually mature stage of Lernaea hranehioMs (after Claus). a, male. B, female at the copulatory stage. C and D, later conditiou of the female transformed by parasiti.sm, slightly magnified, o', first, a", second antenna ; }l-S"^t first four pairs of thoracic limbs ; g, opening of the receptaculum seminis ; mxj, maxillipede ; oc, eye ; sp, spermatophoral sac ; t, left testis. This stage R 242 CRUSTACEA. of KoLLAK (No. 77), V. NoRDMANN (No. 79), Claus (No. 66), Vejdowsky (No. 81), and others. Its course in the various forms seems to show great agreement, so that Achthcres, described by Glaus, may be selected as a type. The young animal which hatches from the egg (Fig. 115 A) exactly resembles a Nauplius, swimming about with difficulty by means of its two anterior pairs of limbs (iirst and second antennae). Closer examination, however, reveals the fact that the body which is hidden within the Nauplius cuticle already shows that degree of organisation characteristic of the first Cijclops stage. Not only the mouth-parts, but two pairs of thoracic swimming feet (^j^, p^), lie hidden within the Nauplius integument. The mandibles [md) and first maxillae {mx) lie as small stumps at the sides of the upper lip, which enters into the formation of A z Fig. 115.— Metamorphosis of Achthcres percnriini (after Claus, from Balfour's Text-hook). A, so-called Nauplius stage. B, first Cyclops stage. C, older stage of the male larva. J>, sexually mature female. E, sexually mature male. ««i, at-, tirst and. second antennae; md, mandible ; mx, maxilla ; pmi, pvi'^, first and second niaxillipedes ; pi, j)2, first and second swimming limbs ; z, frontal organ ; i, intestinal canal ; o, Nauplius eye ; b, glandular body ; t, organ of touch ; ov, ovary ; /, spine derived from the fused niaxillipedes ; g, cement gland ; rs, receptaculum seminis ; n, nervous system ; tc, testis ; v, vas deferens. the adult rostrum. The position of the two niaxillipedes {jmi'^, ■pm") is of interest, in so far as it proves distinctly that they develop as the exopodite and endopodite of one and the same limb (second maxilla). Besides the above organs, we recognise the future adhering organ in the form of a s))irally-coiled filament ending in a spherical swelling which grows out from a highly-refractive frontal process (~). Glaus considers that this apparently homogeneous filament is a tube filled with a fluid secretion, and is the duct of a glandular mass which secretes a cementing medium. This first stage which seems to cover the whole series of Nauplius stages, undergoes ccdysis after a few hours, the larva which follows possessing the organisation of the first Cyclops stage (Fig. 115 B). It PARASITA. 243 has a long ce})lialo-tliorax, followed by three free thoracic segments and an nnsegmented abdomen. In the thoracic region, two pairs of Avell-developed swimming limbs (^j^, J'J-) and a third rudimentary pair (|j'^) can be recognised. The first antennae {at^) are cylindrical three-jointed limbs beset with setae. The second antennae are still biraraose (at-), but are already changed into the adliering organs of the larva, the longer brancli ending in a hook bent like a claw, ^Yhile the shorter branch is beset with papillae. The upper lip has united with a channelled lower lip (derived from the paragnatha ?) to form a conical sucking proboscis, from the outer sides of which project the short, conical mandibles, representing the transition from the masticatory jaws of the Cyclo- pidae to the piercing stylets of the Parasita ; here also the palp-like first maxillae are found. These are followed by the two pairs of maxillipedes transformed into adhering hooks {2rin^, pvi^), the outer one already assuming a more forward position, and the inner one lying more posteriorly. Among the internal organs, the intestine, the Nauplius eye shifted far back, and two bean-shaped bodies {b, glands ?) at the sides of the latter can be recognised. It is probable that the larva, after a short life of free activity in the water, becomes attached as early as this stage to the mucous membrane of the jaw of the percli (v. Noudmann). The peculiar adhering organ, however, appears to become free, and to be used for its sj^ecial pjurpose only after a furtlier moult. In this and two other attached stages, which probably follow through further moults, the mandibles shift to a position within the sucking proboscis, while a reduction in the setae on the swimming feet probably takes place. These stages, however, did not come under observation, but in a somewhat later stage which was observed (Fig. 115 C), there was already considerable approxi- mation in the structure of the body to the adult Achthcres. The body has now become almost vermiform, the first thoracic segment having separated from the head and having united with the four following segments to form a sac-like region, at the end of which the small, pointed, furcal appendage can be recognised. The antennae and the mouth-parts already essentially resemble the corresponding parts of the adult. The frontal adhering organ, with the exception of a vestige of its basal section (::), has vanished ; on the other hand, a new provisional adhering organ in the shape of a stiff filament (/) has arisen on the outer (anterior) maxillipede (pim}) ; this filament starts from the tips of the fused outer maxillipedes. It is an interesting fact that at this stage sexual differentiation begins to be noticeable. Smaller individuals (yoiing males) have remarkabl}' strong outer maxillipedes [pvi^), which are only united at the point of insertion of the adhering filament, and carry a strong terminal hook. When this filament is thrown off in the next moult, these structures give rise to the anterior maxilli- pedes of the male, which remain distinct and function as hooks for attachment (Fig. 115 E). Tlie posterior maxillipedes {pni') are also very large and each carries a small anchoring hook. In the female, on the contrary, the anterior maxillipedes (;;»i^) are rather long and retain the fused condition, ending in a sucking disc (Fig. 115 JD) ; the posterior maxillipedes can in the same way be distinguished from those of the male by a large, hook-like, terminal joint With regard to the inner anatomical peculiarities of this larval form, we must first mention the degeneration of the Nauplius eye, which, indeed, is not universal among the parasitic Crustacea. The Nauplius eye is retained in the dwarf males of Choiidracaiithus and of Lernaeopoda, as well as in many females {e.g., Chondracanthus cornutus). In the posterior cephalic region of the larvae under consideration, at the sides of the intestine, there are two pairs 244 CRUSTACEA. of glands derived from the bean-shaped bodies above mentioned ; these glands, whose efferent ducts open out at the base of the maxillipedcs, yield a stiff secretion. Between these glandular bodies a dorsally-placed, pulsating organ can be noticed ; this is probably a short, sac-like heart. A similar organ has been observed in Tracheliastes by Vejdowsky, and in the larva of Lcrnaea, by Hesse. Tlie rudiments of the genital organs are already clearly recognisable. At the next moult the animal becomes sexually mature. The male does not increase in size, but, in the female, the posterior region of the body becomes very much enlarged. C. Branchiura. The Branchiura are usually considered to be nearly related to tlie Copepoda. This supposed relationship is based principally on the similarity existing between the swimming limbs of Argulus and those of the free-living Copepoda, and on the structure of the mouth-parts, which, according to the investigations of Claus, are very like those of the parasitic Copepoda (Siphonostomata). Hook- shaped mandibles and stylet -like maxillae can be distinguished. These appear to be enclosed in a proboscis formed by the fused upper and lower lips, together with lateral parts which must be regarded as derived from the mandibles. These are followed by two pairs of maxillipedcs which serve as adhering organs. The view that these latter are the two rami of the second maxilla which have become independent receives special support from the position of the ducts of the shell-gland, discovered by Claus on the basal portion of the second maxillipede. Such a derivation of the maxilli- ])edes would constitute a fresh link between these forms and the Copepoda. Argulus, indeed, shoAvs in the structure of the genital organs, as well as in other points of its organisation, some remark- able peculiarities, and by the possession of paired, movable, lateral eyes and of branched, hepatic tubes approaches the Phyllopoda, sa that we must probably consider the Branchiura as an offshoot from the common ancestor of the Copepoda. The eggs, which are rich in food-yolk, and in which the germ-band attains a ventral curvature, are attached by the female in rows to stones, etc. The young, when hatched (Claus), already greatly resemble the adult in shape (Fig. 116), having the same body- segmentation as well as the same number of limbs (apart from the maxillae). The shield-like anterior portion of the body (cephalo-thorax) consists of the cephalic segments fused with the most anterior limb- bearing segment which carries a pair of swimming limbs. Three free thoracic segments follow, each having a pair of swimming BRANCHIURA. 245 limbs, and these again are followed by an unsegmented abdomen with terminal furcal appendages (in the young form). In the adult, these latter have shifted dorsally. The young are further distinguished from the adult by the small extent of the cephalo-thoracic shield (Fig. 116), whicli does not yet cover the thoracic segments dorsally. The further metamorphosis principally concerns the transformation Fig. 116. — Newly-hatclied larva of Argulus fuliaccus (after Clacs). a', first antfiiiiia ; a", second antenna ; md, mandible ; mdt, mandibular palp ; mfi, first, m/", second maxilli- pede ; pi, :2)', tirst four swimming limbs. of the different limbs. In the newly-hatched larva, the first antenna (a) is short, three-jointed, with a large hookdike appendage on its basal segment. The second antenna (a") is distinctly larger and biramose, an endopodite ending in a hook, and an exopodite provided Avith setae being distinguishable in it. In the mandible (md) are found a basal segment (masticatory blade), which becomes included 246 CRUSTACEA. in the proboscis, a middle portion forming tlie lateral wall of the proboscis, and a long, palp-like appendage provided with setae and projecting freely. This latter, the outer branch of the second antenna and the anterior pair of swimming appendages, are the most important locomotory organs of the larva. jS'o trace could be found in tlie larva of the stylet-like maxillae which are contained within the adult proboscis. The two pairs of maxillipedes terminating in hooks (?»/', mf") serve for adhering organs. Of the four pairs of swimming limbs {p^-p^), only the anterior pair is free and movable, and somewhat resembles in shape a biramose Copepodan limb. The biramose rudiments of the three posterior swimming limbs are still unjointed and immovable. During the course of metamorphosis, after several moults, the basal hook-like processes of the two pairs of antennae grow stronger ; the exopodite of the second antenna and tlie palp of the mandible disappear, while the endopodite of the second antenna, losing its terminal liook, is transformed into a simple palp. The large sucking disc develops on the first maxillipede. The swimming limbs become biramose and are provided with setae, thus resembling the limbs of the Copepoda. In the two anterior pairs of these limbs the rudi- ments of tlie inner branches, known as the flagella, appear, while sexual differentiation is evident in the two posterior pairs, a char- acteristic transformation of tlie protopodite taking place in the male. 7. General Consideration regarding the Segmentation of the Body and the Metamorphoses of the Malacostraca. In the IMalacostraca the body consists of three primary regions, each of which, in all the divisions of this sub-class, contains a definite number of segments. In the anterior ceplialic region, to which belong five pairs of limbs (the two pairs of antennae and the three pairs of jaws), there is, with few exceptions, no trace of a separation into distinct segments. In the thoracic region, Avhich follows posteriorly and consists of eight limb-bearing segments, while the boundaries of the different segments are still more or less marked, only a slight power of movement between the individual segments is retained. But in the abdominal region, which consists of six limb-bearing segments and a terminal portion (telson), the full mobility of the separate segments is, as a rule, preserved, a fact connected with the development of the posterior end of the body into a swimming tail of great importance for the locomotion and the steering of the body. GENERAL CONSIDERATION. 247 The terga of tlie cephalic segments are thickened to form the carapace (dorsal shield) ; this, at its lateral and posterior margins, passes into an integinnental fold which runs back over the thoracic region, and thus covers some or all of the thoracic segments dorsally. Only rarely do the terga of the thoracic segments covered by the dorsal shield retain any degree of independence (Stomatopoda, Fig. 141, p. 298, a few Schizopoda and Nehalia) ; in most cases the terga of the thoracic segments fuse closely with the integumental fold of the dorsal shield lying above them. A fusion of the cephalic and thoracic segments thus takes place, forming a common region of the body {cephalo-thorax). In one series of Malacostracan forms, however, the marginal fold of the dorsal shield has undergone degeneration (Arthrostraca) ; here, as a rule, only the anterior thoracic segment fuses with the cephalic region forming the small cephalo-thorax of this group, which is followed by seven free and movable thoracic segments. The obliteration of the boundaries and consequent loss of move- ment between the segments of the cephalic and thoracic regions just mentioned has its effect on the condition of the limbs. Only in rare cases {Nehalia, Euj)hausiidae) do all the eight pairs of thoracic limbs agree more or less in structure. As a rule, one or more pairs belonging to the anterior part of the thorax enter into close relation with the mouth and become modified for the purpose of mastication, 'i'hese are tlien distinguished as maxillipedes, while the succeeding thoracic limbs which serve for locomotion receive, in many groups, the name of ambulatory limbs. In the Arthrostraca, only the anterior pair of thoracic limbs is changed into a pair of maxillipedes, but in the Decapoda there are three pairs of maxillipedes, and in the Stomatopoila as many as five anterior pairs of thoracic limbs are thus transformed. We must assume as the fundamental form of Malacostracan limb a biramose swimming limb with basal epipodial appendage, such as is retained as a thoracic limb in the Schizopoda. The shape of the thoracic limb in Nehalia (Fig. 91 -0, p. 194) suggests to us that this form may perhaps have been developed from a more lamellate type of limb resembling that of the Phyllopoda. A two-jointed protopodite passes into a five-jointed endopodite, while the exopodite (flagellate branch), which often undergoes degeneration, frequently exhibits a large nuuiber of closely-crowded joints beset with setae. If we compare the metamorphosis of most Entomostraca (especially that of the Phyllopoda) with that of the Malacostraca (p. 193), 248 CRUSTACEA. we find, in the first group, a more gradual transition through many moults from the NaupUus to the adult form, while in the Malacostraca the metamorphosis lias attained a higher degree of specialisation, inasmuch as the separate stages appear more distinct from one another, and larval stages are intercalated which are not on the direct path of transition from the young to the adult form. These latter, by the development of secondary peculiarities, attain a certain independence and lead on to the adult condition only through further important changes in form. The metamorphosis of the lower Crustacea thus bears the same relationship to that of the Malacostraca as does incomplete to complete metamorphosis amono- the Insecta. As examples of these newly-introduced stages in the ontogenetic process, we must mention especially the Zoaea of the Decapoda and the Zoaea-like stage of the Schizopoda {CaJyptopis) and the Stomatopoda, which are distinguished by the fact that although they possess the full nu)nber of body-segments, those of the middle region of the body are in a rudimentary condition. In this larva the posterior (five to seven) thoracic segments seem to be unusually backward in developing, and to have either no limbs or only very rudimentary ones, while the segments of the abdomen are already highly developed. The Zoaea is evidently a larval form secondarily changed by marked adaptation to a pelagic manner of life. From this standpoint it seems appropriate enough that the body retains a compressed form, tliat the most important locomotory organs (the maxillipedes and in some cases the antennae) are developed in the anterior region of the body, and that a posterior movable region (abdomen) develops early for swimming and steering purposes. The rudimentary condition of the middle region of the body thus ap- pears to some extent explicable. In regarding the Zoaea as a larval form secondarily intercalated in tlie course of develojjment, which has attained a certain independent value and significance, special interest attaches to the development of the heart. Bearing in mind the condition of the heart in Ncbalia and the Schizopoda, we should expect to find, in the larval forms of the Decapoda, a long tubular heart. We should also presuppose tliat the three pairs of ostia occurring in the Decapodan heart would already be found in the heart of the Zoaea. This, however, is not the case. The lieart of the Zoaea is a short sac, recalling to some extent that of the Copepoda. It has only two pairs of lateral ostia (in individual cases, such as Fenacus, Euphausia, only one). The missing pairs of ostia only appear later. This proves clearly that the heart has undergone secondary modification corre- sponding to the requirements of the Zoaea. The phylogenetic stages in the development of the heart have been modified to suit the altered conditions of organisation of the Zoaea stage. GENERAL CONSIDEBATION. 249 In the complete series of developmental stages of the Decapoda, which however is only retained in ftdl in a very few cases, the following stages can usually be distinguished passing one into the other by a series of moults. 1. Tlie Nauplius stage (Fig. 122 A, p. 267). This stage shows great agreement with the Nauplius of the Entomostraca in form and in the possession of the three typical pairs of limbs, the anterior limb (first antenna) being uniramose, but the two posterior limbs (second antenna and mandible) biramose. A free Nauplius stage is foiznd in Penaeus and in Euphausia, among the Schizopoda. 2. Tlie Metanauplius stage (Fig 117, p. 254, and Fig. 118, p. 258), in the ioxm of the V)ody, closely resembles the last stage, but shows, behind the Nauplius limbs which have shifted somewhat forward, the rudiments of four more pairs (in Eupjhausia only three). An integumental fold, which arises laterally and posteriorly, is the first rudiment of the dorsal shield. The posterior end of the body is marked by two short prominences bearing setae (furcal processes). The Metanauplius stage is the starting-point in the metamorphosis of L%icifer. 3. The Protozoaea stage (Fig. 119 A, D, p. 2G0, and Fig. 121 A, p. 264). The pairs of limbs which appeared as rudiments in the Metanauplius (first and second maxillae and second maxillipede) have now developed fully. The anterior region of the body is covered by the dorsal shield ; posteriorly, the body passes into a narrower region combining the rudiments of the thorax and the abdomen, and already showing segmentation anteriorly, Avhile the posterior (abdominal) region is not yet fully segmented. The antennae still possess the Xauplius character and function as oars, as also do the biramose maxillipedes. The mandibles have greatly altered ; the basal joint is retained as a masticatory blade, while the distal part (palp) has disappeared. The Protozoaea stage occurs in the metamorphoses of the Penaeidae and Sergestidae. It is distinguished by the presence of distinct furcal processes. In individual cases (Sergestes) the third pair of maxillipedes may also be developed at this stage. 4. The Zoaea stage {¥ig. 119 ^, p. 260; Fig. 123 C, p. 269 ; F% 124, 23. 272, and F^ig. 136, p. 291). This stage agrees in all important characters with the preceding stage, from which it is distinguished by the distinct segmentation of the posterior or abdominal region. The sixth abdominal segment, it is true, often remains united with the telson for some time. The limbs of the Zoaea stage are the same as those of the preceding stage. In the more primitive Decapoda, 250 CRUSTACEA. the antennae also serve as oars, wliile, in tlie Zoaea of the Brachyura, these limbs are kept in the background, and locomotion is carried on exclusively by the two pairs of biramose maxillipedes in con- junction with the movable abdomen. In many cases, the third pair of maxillipedes has already begun to function. The limbs which follow these may be present as sac-like, unjointed rudiments in close contact with the body, but they never function in the Zoaea. The pleopoda are still altogether wanting, except the sixth pair (uropoda), which may, in individual cases, develop as early as this stage. The spine-like process, which rises from the cephalo-thorax and occurs tyjncally in the Brachyuran Zoaea, was formerly considered as a characteristic of this stage, and too great stress was laid upon this point. A really important characteristic of this stage, however, is the fact that the six posterior thoracic segments (commencing with that carrying the third pair of maxillipedes) are generally rudimentary and often difficult to recognise, while the abdominal segments are very apparent on account of their large size and distinct boundaries. The Zoaea stage indicates, in many Decapoda, the beginning of metamorphosis, nuxny leaving the egg in this condition. The Protozoaea and Zoa,ea, in contrast to the Metanauiilius, are dis- tinguished by the gradual development of the paired stalked eye which, as in Branchipus, first arises as a lateral outgrowth of the cephalic region [c.f. the Zoaea of Lncifer, Fig. 119 E, p. 260), the eye-stalk developing very gradually. 5. The Mysis stage (Fig. 120 A, p. 262, and Fig. 123 D, p. 269) and Metazoaea stage (Fig. 133 B, p. 287). By the development of the posterior thoracic limbs, the Zoaea passes into the Mysis- or Schtzojyoda-Uke stage. These limbs, which now begin to function, resemble the maxillipedes in being biramose swimming appendages provided with setae ; they assist the maxillipedes in propelling the body, and recall the limbs of the Schizopoda. At this stage the pleopoda develop. In the Brachyura and Anomura, the process of development seems to be simplified so far as the rudiments of the ambulatory limbs are concerned ; these limbs, which are present in the late Zoaea stage, never resemble the Schizopodan limbs, but from the first are uniramose and pass direct into the adult form. An exopodite never develops on the rudiments of these limbs. Consequently, the Zoaea stage in these animals is followed by a stage in which the general form of the body resembles that of the Zoaea, but, in addition, the rudiments of the five pairs of ambulatory limbs are GENERAL CONSIDERATION. 251 more or less developed, though still closely pressed against the body. This stage, AAdiich replaces the Mysis stage in the Anomiira and Brachynra, has been called by Claus (No. 7) the Metazoaea. G. Last stages of metamorphosis. These stages are now only distinguished from the adult form to which they lead by unim- portant characteristics. In the Serr/estidae, the loss of the exopodite on the thoracic limbs and the enlargement of the abdomen leads from the Mysis stage to the Mastigopus stage (Fig. 120 C, p. 262). In the Penaeidae and Caridea the corresponding stage is called the first shrimp stage. The last stages of metamorphosis in the Anomura and Brachyura are known as the Megalopa (Fig. 133 J. and B, p. 287), the transition from this stage to the adult form involving considerably greater changes in the Brachyura than in the Anomura, because the latter remain throughout life at a stage of development nearer that of the Megalopa. If we review the series of larval stages here described, we see that the order of development of the segments and limbs from before backward is as a rule retained. Only in details are there certain characteristic deviations from this order. For instance, the develop- ment of the thoracic segments in the Zoaea stage is usually retarded as compared with the development of the abdominal segments, and, among the limbs, an exception occurs in the early appearance of the sixth pair of pleopoda. These exceptions to the rule can be shown to result from the adaptation of the larva to a pelagic existence. The whole series of ontogenetic stages described above is only passed through in very few cases in the Decapoda. Penaeus and Lucifer may serve as examples of such complete metamorphosis. As a rule, metamorphosis is more or less abbreviated, the early stages being obscured or hurried through within the egg. The Sergestidae, for instance, hatch at the Protozoaea stage, most of the Caridea at the Zoaea stage, the marine Astacidea at the Mysis stage. The most extreme case of abbreviated metamorphosis is found in many forms living in fresh water and on land {Astacus, Telp>husa, and Gecarciims). The abbreviation of metamorphosis is also attained in another way, viz., by the tendency towards the obliteration of the charac- teristics of the special ontogenetic stages. Thus we shall see that, in the Caridea, the Zoaea stage appears altered by anticipating certain of the peculiarities of the Mysis stage. The complete dis- appearance of the Mysis stage in the metamorphosis of the Brachyura aod Anomura can be explained in a similar way. 252 CRUSTACEA. The series of ontogenetic stages described above refers to the Dccapoda, but the Schizopoda {Eupliausiidae) and the Stomatopoda are in this respect closely related to this order. The Cahjptopis stage of the Eupliausiidae might be assumed to be a Protozoaea or a Zoaea stage if it did not ditfer from the latter in the absence of the second pair of maxillipedes. In the metamorphosis of the Stomatopoda, on the other hand, we see that a suppression of the thoracic segments and the limbs belonging to them, similar to that in the Decapodan Zoaea, leads to larval forms which have been called the Psewiuzoaeae of the Stomatopoda. A second series of forms among the Malacostraca, however, including the Cumacea and the Arthrostraca, shows, in connection with the care of the brood which there prevails, an entire dis- appearance of metamorphosis. In these groups we have conditions related to those found in the Mijsidae and Leptostraca. We do, indeed, find in the belated appearance of the last pairs of thoracic limbs in the Isopoda a last remains of that ontogenetic tendency which, in the Decapoda, led to the development of the Zoaea stage. 8. Leptostraca. The Leptostraca (Nebalia), like the Mi/sidae, have no free- swimming larval stage. When the young animals leave the shell- cavity of the mother, which is used as a brood-chamber, they have in all essential points (Metschnikoff, jS'o. 82) attained the final shape. Metamorphosis is therefore here, as in the Mysidae, confined to those stages which were passed through in the brood-cavity after the shedding of the egg-integument. With regard to the appearance of the difterent limbs, the order from before backward is retained. In this respect, and in the absence of a distinctly marked Zoaea stage, the Leptostraca nearly approach the Phyllopoda. The three pairs of Nauplius limbs appear first. There then follows a stage in which these three pairs of limbs have advanced in development, while behind them can be recognised four other pairs (two pairs of maxillae and the two anterior pairs of thoracic limbs). This stage thus shows, in the number of limb-ri;diments, a certain agree- ment Avith the Zoaea form. At a later stage, the rudiment of a third pair of thoracic limbs can be made out. The embryo lies in the egg, curved in such a Avay that the ventral surface of the caudal region is in contact with the ventral surface of the anterior part of the body. Tlie egg-integument then splits, and the larva, which is thus freed, but is still enveloped in the larval cuticle, and which SCHIZOPODA. 253 shows the rudiments of all the thoracic limbs, not merely straiglitens out, but becomes even somewhat dorsally curved. We thus have, in Nehalia, a repetition of the changes with regard to the relative position of the body regions which we saw (p. 154) commencing in Mysis, but with the difference that, in the latter, the rupture of the egg-integument and the straightening of the body take place as early as the Ncmplius stage, whereas here they occur at a later stage. The pleopoda now gradually appear in regular order from before backward, the body approaches the adult form, and the young leave the brood-cavity (Metschnikoff, No. 82). 9. Schizopoda. A very primitive form of metamorphosis through many moults is retained among the Schizopoda by the Euphaudidae. The Mysidae, on the other hand, emerge from the Nauplius cuticle in a condition closely resembling the adult in form, and at this stage they leave the brood-pouch of the mother and swim about freely (p. 153). The different larval stages of the Eupliausiidae, none of which can be exactly identified with the Protozoaea and Zoaea of the Decapoda, were described by Dana as separate genera under the names of Calyjjtopis, Furdlia, and Cyrtopia, Glaus (jSTo. 97) being the first to prove that these forms belonged to--the ontogeny of the Euphau- siidae. The youngest stages were made known by Metschnikoff (Nos. 93 and 94), who established the important fact that the larva of Euphausia leaves the egg as a true Nauplius. The most important points in the later ontogenetic stages were made known to us chiefly by Glaus (Nos. 91 and 8) for Euphausia, and more recently the process of development in various forms has been traced in greater detail by G. 0. Sars (No. 95), Brook and Hotle (No. 90). The Nauplius of Euphausia, on leaving the egg, has an oval unsegmented body still without a shell-fold, and carrying on its anterior half the three pairs of typical Nauplius limbs. The anterior pair of these is uniramose, the two others biramose ; their distal ends are beset with setae. A segmentation of the limbs is not yet distinctly evident. Only a very small oral aperture can be seen. Later stages (Fig. 117 .4) are distinguished by the development of three more limb-rudiments (two pairs of maxillae (4. and 5) and the first maxillipedes), and must therefore be regarded as Metanauplius stages. The three anterior pairs of limbs {1, 2, S) have still the Nauplius character. We can recognise the rudiments of the Nauplius eye, the upper lip (o), the paired paragnatha (u), and a shield-like 254 CRUSTACEA. fold covering the lateral parts of the posterior limb-rudiments. The posterior end of the body, below the anal aperture which is now becoming more distinct, is lengthened into two rounded f ureal processes fringed with setae. Later MetanaupUus stages show a change in the third limb, which has entirely lost the form of Fig. 117.— Three stages of developineiit of Eiiphmida (from Lano's Text-hool). A, Meta- naupUus (after Metschnikoff). B, Calyptopis (after Claus). C, older Cahjptopis (after Glaus). 1 and 2, first and second antennae ; 3, mandible ; i, 5, first and second maxillae ; /, first pair of niaxillipedes ; ab, abdomen ; (rtj-as) first five abdominal segments ; a^, sixth pair of pleopoda; an, anus; fs, frontal sensory organ; o, upper lip; «, paragnatha ; til , thoracic segments. a swimming appendage and has developed into a masticatory blade (mandible) with a greatly reduced palp. The shield-like fold has now developed in the anterior region of the body, and thus surrounds the cephalic portion of the larva (rf. the later stages, I'ig. 117 ^ and C). In the posterior limbs, the first indication of SCHIZOPODA. 255 the fonnation of lobes is evident, while, near the Nauplius eye, a paired frontal organ (fs) has developed, similar to that already described in the Phyllopod larva. The MetanaupUus gives rise tlirongh further moults to the series of Calyptopis stages (Fig. 117 5 and C), which are characterised by the development of the six most anterior pairs of limbs, and of the long posterior (thoraco-abdominal) region of the body. The tAvo pairs of antennae still retain essentially the Nauplius character, although they are now jointed. The first antenna now exhibits a three-jointed shaft, at the end of which two short processes (tlie rudiments of the future flagella) are inserted. In the second antenna, at the end of the exopodite which is covered with setae, a distinct segmentation into closely-crowded rings is evident. The two maxillae (-^, 5) and the first maxillipede (/) appear as richly-lobed appendages, showing in their form considerable agreement with the Phyllopodan limbs. The first maxilla (-^), besides its two masticatory blades and its endopodite, shows a short truncated portion bearing setae (exopodite), the rudiment of the future fan-like plate. In the second maxilla {5) the exopodite is in quite a rudimentary condition, while, on the inner side of the protopodite, four masticatory processes have developed. The first maxillipede (7) has the character of a biramose swimming limb (especially in Ntjdiphanes). At the beginning of the thoraco- abdoraen, which lies behind these appendages, the closely-crowded rudiments of the other thoracic segments can already be distinctly seen (Fig. 117 Z?, th), while the abdomen {ah) still appears un- segmented. The posterior end of the abdomen has already been transformed into the middle plate of the caudal fin, and is beset with strong spines along its posterior edge. In front of the anal aperture the first rudiments of the lateral parts of the caudal fin (sixth pair of pleopoda, a^) can be recognised. The cephalo-thoracic shield, which covers the anterior part of the body, has undergone o'reat development. In Euphauda it is distinguished by the presence of an unpaired dorsal spine directed backwards, by the delicate dentation of its edges, and by a notch in its ventral margin which recalls the excavation in the edges of the shell in Cijpridina and Halorypris. In other genera {Nydiplianes) the shell has no dorsal spine and the margins are not toothed ; there are also only very indistinct lateral notches in the margin of the shell. The internal organs, which should be noticed at this stage, are the gradually developing paired rudiments of the eye, the hepatic outgrowths of the alimentary canal, and the short sac-like heart which is continuous 256 CRUSTACEA. with a well-developed arterial system, and is provided with one pair of venous ostia. Later CaJi/ptojns stages (Fig. 117 G) are distinguished from those just described by the more distinct development of the rudiments of the eyes which are still hidden under the dorsal carapace, and by the full segmentation of the body. ISTot only is the thorax (th) divided into seven (short) segments, but the abdomen also (a^-o^) appears fully segmented In the last of these stages, the sixth pair of pleopoda (a^) are developed as freely projecting lateral wings of the caudal fin. If we compare the series of Cahjptopis stages with tlie other stages of Malacostracan larvae, we must class the younger Cahjptopis stage with the Protozoaea, the older with the Zoaea. The Cahjptopis stages, however, differ from these in the absence of the second pair of maxillipedes which have not yet developed. The later stages, known as the Ftirc'dia stages, are characterised above all by the complete development of the stalked eye which is now movable, and which from this time onwards is not covered by the forward extension of the dorsal shield, but projects freely through an indentation at its edge. A corresponding change is found in the part of the cephalic shield lying between the eyes, which is gradually transformed into a frontal plate running out into a point to form a rostrum. While the six anterior pairs of limbs still retain for the time the shape seen in the Cahjptopis stage, the missing posterior limbs are developed, the first FurdUa stage showing the rudiments of the second pair of maxillipedes and of the first pair of abdominal limbs. The other abdominal limbs develop very soon, while the third maxillipede and the ambulatory thoracic limbs, as well as the branchial rudiments belonging to them, appear more gradually in order from before backward. At the same time, in Enphausia, the rudiments of the eye-like luminous organs at the bases of the limbs develop. The most characteristic feature of the Gtjrtopia stage is the change in the shape of the antennae, which from this time are no longer used as oars, but approach the adult form. The two flagellate branches of the first antenna have lengthened considerably and have become segmented into many rings. In the second antenna, the transformation of the endopodite into the flagellum, and of the exopodite into a scale, are noticeable. Through the completion of the number of limb-rudiments and the development of the last thoracic limbs, the Gtjrtopia larva passes gradually into the adult form. DECAPOD A. 257 The Mysidac, ike Nehalia, only undergo metamorphosis within the brood- cavity of the mother, the young which leave that cavity already showing the form of tlie adult. We have mentioned above (p. 154) that the egg-integument in Musis is shed at the Nauplius stage. The larva, which is then enveloped in the Nauplius cuticle alone, is essentially embryonic in character. It is maggot- shaped and can move but little, the limbs are soft and have no fringe of setae. The larval cuticle, beneath which the remaining limbs develop, is distinguished, in Mysis vulgaris and M. fiexuosa, by the fact that, at its posterior end, it runs out into two setose furcal processes. The next ten pairs of limbs (two pairs of maxillae and eight pairs of thoracic limbs) appear simultaneously. The first of the pleopoda to appear is the sixth pair which enters into the formation of the caudal fin (Fig. 77 E, p. 153) ; the five anterior pairs grow out only after the shedding of the Nauplius cuticle (P. J. and E. Van Beneden, Nusbaum). The coui'se of development in the LopJiogastridae seems to be in comjjlete agreement with that in the Mysidac. Sars, at least, has figured an ontogenetic stage of LopJiogasfer, which completely resembles a late larval stage of Mysis, the only distinction being that, in the former, all tlie pairs of pleopoda are already present ; these limbs perhaps appearing somewhat earlier in Lopho- gaster than in Mysis. The development of the Mysidae and the Lophogastridae may in a general way be described as essentially abbreviated in comparison with that of the Euphau- siidae. In this respect, as well as in the inner ontogenetic processes, the Mysidae and Lophogastridae approach the Cumacea and the Arthrostraca. 10. Decapoda. A. Sergestidae,* Among the Decapoda, the Sergestidae and the Penaeidae are distinguished by the primitive conditions recalled by their meta- morphoses, Avhich begin with a very simple stage [Nauj?Hus or MetanaiipHus), and also by the regular order of appearance (from before backward) retained by the body-segments. In the family of the Sergestidae, the metamorphosis of the genus Lucifer has been specially accurately observed. The Protozoaea of this ontogenetic series was called by Dana Eridhina demissa ; Claus (No. 8) afterwards found the Zoaea stage belonging to it, but WiLLEMOES-SuHM (No. 157) first established its connection with the development of Lucifer, while Brooks (No. 109) observed the complete course of metamorphosis, from the hatching of the egg to the attainment of the adult form. His observations have been found to agree with those made on the Challenger material by Spencb Bate (No. 100) and Willemoes-Suhm. The actual Naiqjlius stage is passed through in the ≫ the young Lucifer larva hatches at a stage which we must describe as a Meta- nauplius (Fig. 118 ^). On the short oval body we can recognise the * [The Sergestidae and Penaeidae form the tribe Penaeidea comparable to the Caridea, etc. — Ed,] 8 258 CRUSTACEA. Naiiplius eye, the projecting upper lip (oZ), and a few furcal setae, indicating its posterior end. There is as yet no trace of a shield-like fold. In the anterior region of the body, the three pairs of jSTauplius limbs (a', a", md) are inserted. Of these, the first {a) is uniramose ; it consists of five joints and carries swimming setae at its end. The second antenna (a") has a two-jointed protopodite and two swimming rami ; when this appendage is compared with that of other forms, we are inclined, in opposition to Brooks, to regard the ramus, which has several joints and is provided with setae, as the exopodite and the simpler, unjointed branch as the endopodite. The third pair of limbs (mandibles) resembles the second in structure, but is smaller and less jointed. It has an unsegmented protopodite, a single-jointed A £ Fig. lis.— Two Metanauplius stages of Luci/er (after Brooks). A, larva just hatched. B, somewhat older stage, a', first antenna ; ft", second antenna ;