i^rOC ^ * fss~ /iffteoAZ °f entry* /fY ■ygrr MBL/W HOI 0 0301 00M7fl. y.. 1 -.L-'i-# LONDON: SWAN SONNENSCHEIN AND CO., Ltd. NEW YORK: THE MACMILLAN CO. 1899. PREFACE. The present and third instalment of the translation of the Lehrbuck der Dergleichenden Entwicklungsge&chichte der wirbel- losen TMere contains the Arachnida and appended groups, the Onychophora, the Myriopoda, and the Insecta, thus completing the Arthropodan portion of this work. The remaining volume, which contains the Mollusca, Ascidia, and Cephalochorda, will, I hope, be published at the end of the year. In connection with the present volume, I have to thank Mr. II. I. Pocock for his valuable assistance in the Arachnidan part, and also I have to thank Mr. A. I). Michael, who kindly read through the chapter on the Acarini and corrected many errors in the same. Most of the suggestions made by these gentlemen have been added as editorial footnotes, but some matter, when the original was obviously at fault, has been placed in the text. The very important work by Brauer on the Ontogeny of the Scorpion renders some of the text relating to this genus out of date, and this work should certainly be consulted by those studying the Arachnida. His discovery of an additional segment between the thorax and abdomen is of especial importance in the interpretation of the Arachnidan body. In this connection it is interesting to note that this segment was correctly figured, though misinterpreted, by Metschnikoff in 1871, as a careful comparison of the figures of these two authors will show. Mil PREFACE. Iii the Onychophora, Willey's paper on Peripatus novae - hritanniae is of great importance, especially in connection with the invagination germ -band in the Insecta and the interpretation of the embryonic membranes. Among the numerous additions to the literature on the In- secta, Heymon's works, especially that on Lepisma, are worthy of careful study. Unfortunately, the interpretation of the ontogenetic processes in the Insecta is very difficult, and in consequence we still find a terrible confusion enshrouding the origin of some organs, especially that of the alimentary canal, which a number of recent authors maintain to be entirely ectodermal, a condition which, judging from what occurs in other Arthropods, seems extremely improbable. The germ-cells, as in the two previous volumes, are still treated of as mesodermal, whereas, as has been pointed out in the editorial notes to Vols. i. and ii., these cells are probably handed down from parent to offspring as distinct and con- tinuous structures, their identity being temporarily merged in the egg. In the present volume I have added more notes and literature and made more alterations in the text than in Vol. ii., and I hope that such alterations will tend to bring this volume more up to date. MAETIN F. WOODWARD. Royal College of Science, London. July, 1899. CONTENTS OF VOL. III. The lung-sacs The intestinal canal . The mesodermal derivatives Blood-vascular system and coelom The coxal glands The genital organs . Chapteb XXI. ARACHNIDA. By E. Korschelt I. Scorpiones .... 1. Cleavage aud formation of the germ-layer 2. The origin of the embryonic membranes and the development of the external form of the body . 3. The formation of the organs A. The nervous system and the eyes B. C. D. E. F. G. II. Pedipalpi III. Pseudoscorpiones .... IV. Opiliones V. Solifugae VI. Araneae ..... Oviposition and constitution of the egg 1. Cleavage and formation of germ-layers 2. Development of the external form of the body 3. The development of the organs A. The nervous system . B. The eyes .... C. Survey of the Arachnidan eyes D. The respiratory organs E. The spinning glands and the poison glands F. The intestinal canal and its appendages . The mesodermal structures . G. The blood-vascular system and the body-cavity H, The coxal glands /. The genital organs VII. Acarina ..... Oviposition ..... 1. Embryonic development 2. The formation of the larval integuments and the further course of development VIII. General considerations regarding the Arachlrida Literature .... PAGE 1 1 1 4 12 12 19 19 22 23 24 25 26 27 32 34 37 37 38 45 59 59 63 68 76 79 81 85 87 92 93 93 93 94 97 110 118 3d^f, X CONTENTS. By E. Kohschelt Chapter XXII. PENTASTOMIDAE. 1. Embryonic development 2. The larval development 3. General considerations Literature ..... Chapter XXIII. PANTOPODA. By E. Korschelt Oviposition and care of the brood 1. Cleavage and formation of the germ-layers 2. The further development of the embryo 3. The form of the larva and its transformation into the 4. General coi>siderations . Literature ..... Chapter XXIV. Literature TARDIGRADA. By E. Korschelt adult Chapter XXV. ONYCHOPHORA (Pekipatus). By E. Korschelt Structure of the egg .... 1. Cleavage and formation of the germ-layers Peripatus novae-zealandiae Peripatus capensis The American species 2. The development of the external form of the body 3. The formation of the organs The ectodermal structures The integument The nervous system and the ventral organs The eyes .... The slime- and the crural glands The alimentary canal The mesodermal structures The body-cavity and the blood-vascular system The musculature The nephridia The salivary glands The anal glands The genital organs Another account of the origin of the mesodermal structures General considerations Literature ..... Chapter XXVI. MYRIOPODA. By E. Korschelt Oviposition and the constitution of the egg 1. Cleavage and formation of the germ layers 2. The development of the external form of the body A. Chilopoda .... E. Diplopoda .... The first rudiment of the embryo Flexure of the germ-band The further development of the embryo The interpretation of the mouth-parts of the Myr Post-embryonic development . C. Symphyla and Pauropoda opod PAGE 129 129 130 136 137 139 139 139 144 148 156 160 162 163 164 164 165 166 169 170 175 188 188 188 189 194 195 195 198 201 204 204 206 207 208 209 211 216 218 218 220 223 223 229 229 229 231 232 236 238 CONTENTS. XI Chapter XXVI. MYRIOPODA— conlirvu d. page 3. The formation of the organs . ... 239 The nervous system . . ... 239 The eyes . . . ... 241 The tracheae . . . ... 243 The protective glands . . ... 243 The alimentary canal . . ... 244 The enteron . . . ... 244 The stomodaeum and the proctodaeum . . . 246 The mesodermal structures . ... 246 The body-cavity, the blood- vascular system, the fat-body, and the musculature . . ... 246 The salivary glands . . ... 251 The genital organs . . ... 252 General considerations . .' ... 253 Literature . . . . ... 257 Chapter XXVII. IXSECTA. By K. Heider . . . 260 I. Embryonic development . . ... 260 1. Oviposition and the structure of the ripe egg . . . 260 2. Cleavage and the formation of the blastoderm . . . 263 3. The formation of the embryonic rudiment and the embryonic integuments . . ... 268 A. General view of the germ-band and the germ-envelopes 268 B. The distinction between the superficial and the im- mersed germ-band . . ... 272 C. The distinction between the invaginated germ-band and the germ-band that has been overgrown by the membranes . . ... 274 D. Insects with invaginated germ-band . . . 276 E. Insects in which the germ-band is overgrown by the amniotic fold . . ... 282 F. Transition forms between the two types of development of the germ -band . . . 287 G. General considerations . ... 289 4 Development of the external form of the body . . . 290 A. Segmentation . . ... 290 B. Stomodaeum and proctodaeum. Labrum . . 293 C. Extremities . . ... 295 I). Xervous system and tracheal invaginations . . 300 E. Transition to the definitive form of body . . 302 5. Completion of the dorsal part of the embryo and degenera- tion of the embryonic envelopes . . . 302 A. Involution through the development of a continuous dorsal amnion-serosa sac . ... 304 B. Involution accompanied by dorsal withdrawal of the amnion only . . ... 307 C. Involution accompanied by dorsal withdrawal of the serosa and complete separation of the amnion . . 307 D. Involution accompanied by the amputation of both embryonic envelopes E. General considerations Xll CONTENTS. Chapter XXVII. INSECTA— continued. 6. The formation of the germ-layers . ... 7. Further development of the mesoderm Development of the body-cavity . ... 8. The formation of organs . . ... A. Outer integument . . ... B. Endo-skeleton . . ... C. The nervous system . • . . . . D. The sensory organs . . ... The ocelli . . ... E. The tracheal system . . ... F. The alimentary canal and intestinal glands G. Heart . . . ... H. The musculature, the connective tissue, and the fat-body /. Genital organs . . ... II. Metamorphosis . . . ... 1. The larval forms . . . ... A. Homomorpha . . ... B. Heteromorpha . . ... 2. Development of the imago . . ... A. Development of the external form of the body . B. Development of the internal organs of the imago Hypodermis . . ... Musculature . . ... Intestinal canal . . ... The tracheal system . ... The nervous system . ... The fat-body . . ... The ultimate fate of the phagocytes General considerations regarding the development of the imago in the pupa . ... III. Parthenogenesis, Paedogenesis, Heterogeny . ... IV. General considerations . . . ... Literature . . . . ... PAGE 309 318 318 322 322 323 323 329 329 334 336 338 341 342 355 355 356 359 367 370 378 379 381 382 385 386 386 386 387 388 390 396 Chapter XXVIII. GEXERAL CONSIDERATIONS ARTHROPODA Formation of the germ-layers . Appearance and further development of the organs Nervous system Eyes .... Respiratory organs r . Fore- and hind-guts . Development of the enteron Development of the mesoderm Germ-band Embryonic envelopes, metamorphosis Interpretation and relationships of the Arthropoda Literature .... ON THE Subjects Index Authors Index 411 412 413 413 414 420 421 422 423 424 425 426 431 433 437 CHAPTEE XXL ARACHNIDA. Systematic : — I. SCORPIONES. II. Pedipalpi. III. Palpigradi (Koenenia). IV. PsEUDOSCORPIONES. V. Opiliones. YI. SOLIFUGAE. VII. Araxeae. VIII. Acarina. I. Scorpiones. The Scorpiones are viviparous. The oval or spherical eggs, which are rich in yolk and are each surrounded by a thin membrane, lie in follicles that arise as outgrowths of the walls of the ovarian tubes. Fertilisation takes place either in the ovarian follicles (Euscorpivs and Scorpio, aIetschnikoff, Laurie), or when the egg has left the follicle and passed into the ovarian tube (Androdonus, Kowalevsky and Schulgin). In the former case the embryo remains in the follicle during the greater part of its development (Scorpio, Joh. Muller), or leaves it when the formation of the germ-band commences (Euscorpius italicus). Further development then takes place in the ovarian tubes or oviducts, which thus function as uteri. At birth the young resemble the adult in their general organisation. 1. Cleavage and Formation of the Germ-layer. The cleavage of the egg in Scorpions is discoidal. At the pole of the egg, which is directed from the follicle towards the ovarian tube, in the youngest stage as yet observed, there were found a number of cells which formed a small unilaminar cap on the yolk ; this is the germ-disc (Fig. 1). The blastoderm spreads gradually from this point, advancing very slowly over the yolk (Fig. 2 A and B). Long B ARACHNIDA, before it has grown round the latter, however, the rudiment of the germ-band has appeared, and the first differentiation of the latter takes place at the point where the blastoderm first began to form. A cleavage of the yolk, such as is met with in the eggs of the Araneae, does not occur in the Scorpiones. The discoidal cleavage of the Scorpiones might be compared with the Crusta- cean method of cleavage distinguished as Type IV., and might, like the latter, be traced back to superficial cleavage (Vol. ii., pp. 117 and 118). This would be the more permissible as superficial cleavage is, as a rule, widespread among the Arachnida also. In this respect the Scorpiones, as compared with the Araneae, must be considered as showing a modified condition, although they are in other respects more primitive. The development of the embryo within the body of the mother is a sufficient proof that modification in the primitive method of development has taken place. The Formation of the Germ-layers. The germ-disc does not long retain the character of a single layer of cells. A thickening appears at its centre, which, on the sur- face turned towards the yolk, appears as a swelling. This, according to Kowalevsky and Schulgix, has arisen by a down- sinking of the cells. If we bear in mind, in addition to this, the longitudinal furrow described by Metschxikoff on the surface of the now oval germ-disc (Fig. 4 A, p. 6), we are reminded of the long slit -like blastopore that occurs in Peripatus and in the Insecta, and which constitutes the longitudinal germinal groove. In any case, the differentiation of the inner and middle germ-layers starts from this point. Fig. 1. — Egg of Euscorpius italicus showing the germ-disc (after Metschnikoff, from Balfour's Text-book). Laurie (No. 23), in his recent work, does not actually deny the " down- sinking " of the cells and the presence of the longitudinal furrow, but being unable to convince himself, seems inclined to doubt their existence. This, indeed, cannot be considered as established, especially as the descriptions given of these processes are not very exact. Laurie derives the ento-mesoderm by delamination from the cell-mass of the germ-disc, in which no special regularity of structure is apparent. But he also finds a thickened point at the posterior end of the germ-disc, in which rapid increase of cells takes place (formation of the ento- mesoderm), and which could therefore be compared with an invagination (Fig. 3, CLEAVAGE AND FORMATION OF THE GERM-LAYER. 3 4,OPo2 A, e). The caudal prominence described by Metschnikoff (No. 24) is probably to be identified with this growing point ; the former projects into the yolk, and at a later period shifts into the caudal region of the embryo. Laukie compares the thickened part with the primitive streak in vertebrates, and we are again involuntarily reminded of the conditions found in Peripatw. In the latter the "point of ingrowth" lies at the posterior end of the long blastopore.* When the germ-disc, by the active increase of its elements, lias attained a thickness of several cells, these still appear but slightly differentiated into layers. The inner rf( surface of the germ- disc is now quite irregular, for single ■cells become de- tached from it, and shift into the yolk .C cj$. birth. This latter point was confirmed by Kowalevsky and Schulgin, who observed the gland both in its earlier slightly coiled stage and in its later more compact condition. [See Brauer, No. II.] The structure and position of the coxal glands in the youngest known stage render it highly probable that they are formed from the somatic mesoderm. They are assumed to be nephridia, a view which seems very probable. Considering the primitive character of the coclom in the /fi Scorpiones, we should , expect the nephridia to open into the body- cavity through funnels, and this is actually the case for a time. The further development of the inner terminations of the gland must de- pend essentially on the modifications undergone by the body-cavity, but this point is somewhat obscure. More thorough ontogenetic researches are required before it can be stated with certainty whether, as in Pcripatus and the Crustacea, a part of the body-cavity forms a capsule for the forma- tion of the terminal sac of the gland, or whether the mouth of the funnel is retained for a consider- able time in a wide secondary body - cavity. The most recent writer on this subject, Stxjiiany (Xo. 14) was not able to prove that the coxal glands in the Arachnida opened into the body- cavity, and he inclines to believe in the presence of a closed terminal sac, such as is found in the Crustacea, but here also we must demand actual proofs. G. The Genital Organs. The ontogeny of the genital organs has as yet been little investi- gated. They were first observed by Laurie at a late stage of development shortly before birth, in the first abdominal segment [second, Brauer], as tubular structures at first unconnected with Fig. 14. — Euseorpius italicits. Portions of sections through a newly-hatched Scorpion (A) and an advanced embryo (B) to show the coxal gland and the formation of the genital organs (after Laurie), a, efferent duct of the coxal gland ; ec, ectoderm ; g, efferent duct of the genital organ ; (/.op, genital operculum ; Ih, body-cavity ; in, external opening of the coxal gland ; mes, mesoderm ; n, ventral nerve-cord ; J'.-ii Pti bases of the third and fourth limbs ; so, somatic, sp, splanchnic layer of the mesoderm. 26 ARACHXIDA. the exterior (Fig. 14 B). Kowalevsky and Schulgix, who also noticed them, referred them, though with some hesitation, to the splanchnic layer of the mesoderm. Laurie's account would rather tend to show that they arise from the somatic layer, as do the coxal glands of the Scorpiones and the nephridia of the Annelida (Vol. i., Fig. 137, p. 297). The nephridial character of the efferent genital ducts seems to be confirmed by the fact that they open into the body- cavity in the form of a wide funnel (Kowalevsky and Schulgix [Brauer]). Laurie also believes that at least in part they are nephridial in origin. The ends of the canals which are directed outwards long remain closed, a fact which we do not regard as disproving the nephridial character of the efferent ducts, since even the Annelidan nephridia develop in a similar way. From Laurie's description we might imagine that the mesodermal efferent ducts become directly connected with the ectoderm at the points where the remains of the first pair of abdominal limbs lie in the form of ectodermal thickenings (Fig. 14 B, g.op), as is the case, according to Bergh, with the nephridia of the Annelida. Kowalevsky and Schulgix, however, speak of an ectodermal invagination, towards which the mesodermal efferent duct grows, so as to unite with it. This invagination, as far as can be made out from their short account, is small, and it appears very possible that such an ectodermal invagination might arise at the thickening which indicates the position of the abdominal limbs. An ectodermal termination has also repeatedly been assumed for the nephridia and the genital efferent ducts of the Annelida. It is, however, highly probable that the short unpaired portion is derived from a depression of the ectoderm. In the Pedipalpi this unpaired segment is much larger, and becomes a large cavity (No. 31). The genital glands arise, according to Kowalevsky and Schulgix, as cell- thickenings "apposed to the inner tube." This can only be understood to mean that a part of the peritoneum {i.e., of the secondary body-cavity) is concerned in the formation of the genital organs ; on this point, however, as well as on the differentiation of the mesodermal structures, we await further particulars.* II. Pedipalpi. f According to Bruce, who has made a few statements as to the ontogeny of Phryuas, the embryo here, as in the Scorpiones, lias an embryonic envelope. We may indeed make the general assumption that the course of development in the Pedipalpi resembles that in the Scorpiones. Bruce points out as specially remarkable the existence of a sensory organ at the base of the second ambulatory limb, consisting of columnar cells prolonged externally into filaments. The Pedipalpi are very closely related to the Scorpiones, and, like the latter, show in their organisation many points of agreement with * [See footnote, p. 3. — Ed.] t [The Pedipalpi are oviparous ; the eggs are carried in a gelatinous sac attached to the ventral surface of the mother. For chief ontogenetic features see App. Lit. Pedipalpi, Nos. I.— III., noting presence of reversion of germ-bands and unanimous conclusion that Pedipalpi are more nearly related to the Araneae than to the Scorpiones. — Ed.] PSEUDOSCORPIONES. 27 Limulus (Ray-Lankester, Bruce). Our knowledge of the ontogeny of the Pedipalpi is unfortunately very incomplete, and this may also be said of Koenenia mirabilis, a form discovered by Grassi (under stones in the plains of Catania), which shows great resemblance to the Pedipalpi, but has been placed by him in a separate order, the Microtelyphonidae, the Palpigradi of Thorell.* This form is said to have no special respiratory organs, and Grassi therefore sees in it a transitionary form between the Gigantostraca and the Arachnida, which has " already lost the gills, but has not yet developed respira- tory organs suited to a terrestial existence " ! We can hardly imagine A Fig. 15.— Embryos of Chelifer in their envelopes (after Metschnikoff, from Balfour's Text- book). A, early cleavage stage. B, stage in which the blastoderm (6?) has separated from the yolk-masses within. C, splitting of the blastoderm into two layers. The yolk-masses are seen within the egg. A cell-like albuminous tissue appears between the blastoderm such a transition, and would rather regard the absence of respiratory organs, if it actually occurs, as a degeneration, such as is met with in other air-inhabiting Arthropoda in cases where the body is dis- tinguished from related forms by its specially small size (e.g., in a few Mites, among the Arachnida, and in Pauropus among the Myriopoda). III. Pseudoscorpiones. The little that is as yet known of the ontogeny of the Pseudoscorpiones does not seem sufficiently well established to enable us to form a decisive judgment with regard to the extraordinary development of these forms. Metschnikoff's * [Hansen and Soiiensen (App. Lit. on Palpigradi, No. I.) give a very careful account of Koenenia, and correct many errors in Grassi's description. — Ed.] 28 ARACHXIDA. statements as to the development of Chdifcr np to the time of the formation of the blastoderm are, indeed, confirmed by Stecker with regard to Chthonius, but the description of the latter author is not calculated to inspire confidence. A more recent treatise by J. Barrois* on the ontogeny of Chelifcr is too short to supply many further details. The eggs of Chelifer and of Chthonius are spherical and crowded with yolk-spherules. Each is surrounded by a vitelline membrane, and again by a second envelope probably secreted by the oviduct. These eggs are carried by the mother on the ventral surface of the abdomen, where they pass through their development. The cleavage is at first complete, the egg dividing up into two, four, and eight equal blastomeres (Fig. 15 A). In the latter stage, i.e., when the egg is divided up into eight spheres, clear protoplasmic segments are said to appear on the surface of the yolk-laden spheres. The number of these clear cells soon greatly increases, until they form a layer surrounding a central mass of yolk (Fig. 15 B); this layer may be regarded as the blastoderm. The large yolk-segments with their nuclei can still be clearly seen within the egg.f The whole process must, no doubt, be thus explained : The few nuclei which enabled the yolk to break up into segments, by division, send off nuclei to the periphery, the nuclei which remain within corresponding to the yolk-nuclei of other Arthropod eggs. In the fact that the yolk itself remains segmented these forms are peculiar. As the segmentation of the yolk gradually disappears, the blasto- derm divides into an outer and an inner layer of cells (Metschnikoff, Fig. 15 C). About this time, large clear bodies appear between the blastoderm and the egg - integument ; these contain structures resembling nuclei, and therefore resemble cells (Fig. 15 C). Metschnikoff was reminded by them of an embryonic envelope, but could not convince himself that such a covering was actually present, and regarded these structures as disintegrated masses of albumen, a view also taken by Stecker. These cells recall those found beneath the cuticular envelopes in the Mites (Claparede's haemamoebae, Fig. 53, p. 99). * We have not heard of any more detailed work on this subject by Barrois ; Stecker's preliminary notice also seems not to have been followed by any larger treatise. [See Barrois (App. to Lit. on Pseudoscorpiones, Xo. I.).— Ed.] t [Barrois (App. to Lit. on Pseudoscorpiones, Xo. I.) has recently very fully investigated the development of Chelifcr; he finds that segmentation may be either total or partial, the latter condition predominating and resulting in a core of yolk with peripheral cells, some large, which form the blastoderm, others very small, which become applied to the vitelline membrane. A deep median ventral longitudinal groove appears, from the walls of which mesoderm-cells are pro- liferated off. "Origin of the entoderm obscure, nuclei appear in the yolk.— En.] PSEUDOSCORPIONES. 29 The further differentiation of the embryo is characterised by the early and pronounced development of the future anterior end of the body ; this appears as a great accumulation of cells belonging to the inner layer of the blastoderm. A pair of marked swellings appear on either side of this region, and from each of these a large truncated appendage soon arises (Fig. 16 A). These processes are the rudiments of the pedipalps which are here, as in Scorpio, the first limbs to appear. They are still in a very primitive condition, the inner yolk-mass extending far into them (Fig. 16 A and B). In front of the limbs, towards the ventral surface, there is a swelling which, even at this early stage, is distinguished by its strong muscu- lature, and consequently has a striped appearance (Figs. 16 A and B, r, 17 .4). This is the rudiment of a provisional organ — a kind of sucking proboscis (Fig. 16 C) which serves for attachment and for db Fig. 10. — A and B, larva of Cliclifer ; C, provisional proboscis of an older stage (after Metsch- nikoff). A, ventral aspect ; B and C, from the side, ah, abdomen ; d, yolk ; (j, brain ; p, the four limbs ; pd, pedipalps ; )•, proboscis (provisional larval organ). taking in food. The embryo leaves the egg at this stage, having previously undergone a larval ecdysis. A fine cuticle, which occupies a peculiar position between the bases of the two limbs, becomes detached from the embryo. The larva, when hatched, at the youngest stage shown in Fig. 16 A, has the muscular proboscis, the truncated pedipalps, and the rudiment of the abdomen directed forward. The proboscis, which is regarded as a modified upper lip, already seems to function as a sucker, for the larva attaches itself by means of this organ to the ventral surface of the mother. The proboscis lengthens considerably at a later stage, and becomes applied to the ventral surface of the larva, lying between the limbs (Fig. 16 B). Barrois has described a provisional oral aperture situated between the pedipalps. There are also, according to Barrois, 30 ARACHNIDA. chitinous structures in the proboscis. There is no mention of an external aperture to the proboscis; Metschnikoff could not find one, although he assumes that the larva obtains its nourishment by sucking the blood of the mother. Soon after becoming attached to the body of the mother, it swells considerably, and becomes filled with a clear fluid (cf. Fig. 17 A and B). If this fluid comes from outside, we must certainly assume that an intestinal epithelium has already developed round the inner yolk-mass, although no such differentiation has been recognised. °co on -J.oO°°^ °A< Fig. 17. — Embryo and larvae of Chclifcr (after Metschnikoff, from Balfour). A, embryo in the egg-integument ; B and C, larvae taken from the ventral surface of the mother. «f>, abdomen with the provisional appendages; un.i, anal invagination; ch, chelicerae ; pd, pedipalps ; between the last two {ch and yd) the upper lip is visible in C. Above the pedi- palps are seen, in A the rudiment, in B the base, and in C the last vestige of the proboscis. In l! the rudiment of the oesophageal ganglion can be recognised, lying dorsally to the proboscis. The pedipalps are followed posteriorly by the four limbs, and, in 11, by the rudimentary abdominal appendages. C represents the larva just undergoing ecdysis. The larval integument is partly loosened (noticeably on the ventral side) ; the remains of the proboscis are attached to it. The later stages (Figs. 16 and 17 B) differ from the youngest larvae (Fig. 16 A) in external form chiefly in the swollen nature of the dorsal region, brought about by the presence of the clear fluid mentioned above. Other modifications have also taken place, the rudiments of the first pair of limbs having budded out behind the pedipalps, and these are followed by the three other pairs PSEUDOSCORPIONES. 31 (Fig. 17 B). On the abdomen, which is bent ventrally, four pairs of limb-rudiments appear (Fig. 17 B), which, however, soon com- pletely degenerate. The Pseudoscorpiones agree in this respect with other Arachnida. The most anterior pair of limbs is still wanting, but a paired thickening is found dorsally, above the base of the proboscis ; this has apparently arisen from an invagination, and is the rudiment of the supra-oesophageal ganglion (Fig. 16 B, g). This recalls the cephalic pits of the Scorpiones and Araneae (pp. 12, 53). The larva continues to approach the adult in form, segmentation appearing both in the limbs and in the abdomen, but the cephalo- thorax remains unsegmented. The chelicerae have, in the meantime, appeared in front of the pedipalps. The true upper lip arises between them, some way from and altogether independent of the larval proboscis (Fig. 17 C). The proboscis degenerates, the last vestige of it being lost when the larva moults, at the stage depicted in Fig. 17 C. It is then still found attached by a delicate thread to a point behind the future mouth, until it is cast off with the larval integument (Barrois). A large mass of yolk can still be seen within the body, enclosed in the enteron, which opens exter- nally through the proctodaeum at the posterior end of the body (Fig. 17 C, an.i). The oesophagus is probably also formed by an ectodermal invagination (Metschnikoff). General Considerations. The ontogeny of the Pseudoscorpiones is remarkable on account of the embryo leaving the egg-membrane with a much simpler structure and at a much earlier stage than in other Arachnida. Further, the larvae, in their half parasitic life on the body of the mother, have developed a provisional sucking organ which at first lies in front of the first pair of limbs, but shifts back later, in consecpiience of processes of growth, on to the ventral surface (Figs. 16 and 17); this organ, however, cannot be compared to a pair of limbs. No homologue has so far been discovered among the Arachnida for this proboscis, which must therefore be regarded as an organ acquired by the Pseudoscorpiones through their peculiar method of development. The difference between the ontogeny of the Pseudoscorpiones and that of the Scorpiones, to which they are perhaps most nearly related, is very striking. The cleavage, the formation of the blastoderm, and the first rudiment of the embryo in the two forms can hardly be compared. They also differ in impor- tant points of their organisation. The absence of the taildike abdomen, the disappearance of the abdominal ganglia (Ckoneberg), the position of the genital apertures (in the second abdominal segment), and, not least, their tracheal respiration, remove the Chernetidae from the true Scorpiones so far that the 32 ARACHNIDA. variations in their method of development appear comparatively unimportant. Attempts have been made to connect the Pseudoscorpiones with other divisions of the Arachnida, especially with the Opiliones, but these have not been sufficiently based on the organisation of the two groups. We must therefore, according to a recent investigator of the anatomy of the Chernetidae (Cuonebekg), leave the systematic position of the Pseudoscorpiones undecided, since their ontogeny, so far as it is yet known, throws no light upon the subject. IV. Opiliones. The spherical eggs of the Opiliones are surrounded by two membranes. The inner membrane is secreted by the egg, the outer by the epithelium of the genital duct ; they represent the vitelline membrane and the chorion. The eggs, glued together so as to form a large ball, are deposited in a hole in the ground (Henking). The first ontogenetic processes have been closely studied in Opilio and Leiobunvm, by Henking, but we are unable to accept his view of the origin of" the cleavage-nuclei through free nuclear formation, since it contradicts what is known of other Arthropoda.* Accord- ing to Fatjssek, the egg of Phalangium divides up into a number of large spherical cells filled with yolk-spherules, each cell contain- ing a central nucleus. Cleavage is therefore total. These cells might be compared to the yolk-pyramids in the eggs of the Araneae, but, in the subsequent processes, these cells in the Opiliones seem to differ from those structures. A cleavage-cavity does not appear. The formation of the blastoderm occurs by the separation and more rapid division of some of the peripheral cells. Not all the cells, indeed, not even the majority of them, rise to the surface to form the blastoderm, a large proportion of them remain within the egg as yolk-cells (Henking, Faussek). The formation of the blastoderm takes place more rapidly in one half of the egg than in the other, a condition similar to that observed in the Araneae. Active increase in number of the blastomeres in one region of the blastoderm leads to the formation of a thickening in it; this is the germ-disc. According to Faussek, immigration of cells into the yolk-mass from the disc does not take place ; the entoderm being possibly represented by the cells which remain in the yolk, and from them, at a later stage, the epithelium of the enteron arises. The origin of the entoderm from cells which, from the first, remain behind in the yolk, has been assumed for the Araneae (Schimkewitsch), but the forma- tion of the germ-layers in the Opiliones has not yet been observed sufficiently * [Most cytologists do not believe in the existence of the process termed free nuclear formation ; all modern research tends to prove that every nucleus has originated directly from a pre-existing one. — En.] OPILIONES. 33 closely for us to decide whether this is also the case in them. Faussek found in embryos in which the segmentation of the germ-band is commencing, an accumulation of cells at the posterior end of the band, which strongly resembles the point of ingrowth in the germ-band of the Scorpiones. The statements hitherto made as to the nature of this structure are, however, so contradictory that it is impossible to gain any clear idea of it. Faussek derives these cells, which appear like a thickening of the blastoderm, from a deposit of yolk-cells on the blastoderm. At first he derived the genital glands from this deposit, i.e., from yolk-cells, but he afterwards traced them to a thickening of the blastoderm which appeared at a very early stage. A more exact account of the partly contradictory statements on this subject may be expected in Faussek's larger work [App. to Lit. on Opiliones, Nos. III. and IV.] The mesoderm, so far as we can gather from the few statements on the subject, splits into a somatic and a splanchnic layer, so that in this respect also there is resemblance with the Scorpiones and the Araneae. The enteron seems to form as in the Araneae, apart from the origin of the entoderm, which arises differently according to Faussek. The yolk is directly surrounded by the splanchnic layer of the mesoderm, and the yolk-cells now become applied to this layer, eventually giving rise to the continuous epithelium of the enteron. This process commences in the anterior part of the body. We have only a few isolated statements as to the further develop- ment of the Opiliones. Metschnikoff (No. 34, p. 520) traces the origin of the abdominal limbs, and Balbiani describes a few of the later ontogenetic stages. It appears that the cephalo-thoracic seg- ments to which the four pairs of limbs belong are distinctly marked off from one another in the embryo, but this segmentation disappears during the further course of development, and is not recognisable in the adult. Between the eyes and the bases of the chelicerae lies an unpaired, spine-like structure, which, like similar structures in the Araneae, and especially in the Myriopoda (Chilognatha), we shall call the egg-tooth (p. 58, and cf. the chapter on the Myriopoda). The little that is known of the ontogeny of the Opiliones is in harmony with that of the Arachnida generally. An important feature which is still recognisable in the adult, seems, according to Balbiani, to be very marked in the embryo. This is the occurrence of masticatory ridges on the pedipalps and on the two anterior pairs of limbs. Herein we find a striking resemblance to the Scorpiones. The Opiliones further resemble other Arachnida in the number and position of the limbs, and in the presence of a coxal gland \MacLeod), homologous with the synonymous organ in other Arach- nids. Whereas, however, in other groups, this gland is merely D u ARACHNIDA. provisional, and degenerates in the adult (Scorpiones, Araneae), in the Opiliones it is a well-developed organ, still functional in the adult, and consisting of a large coiled canal, a wide, sac-like reser- voir, and an efferent duct ; the latter opening externally at the base of the third ambulatory limb (Loman, No. 9).* V. Solifugae. f Of the ontogeny of the Solifugae, like that of all the Arachnida already considered, so far as we are aware, very little is known. The little that we do know is in connection with Galeodes araneoides, some of the later ontogenetic stages of which have been described by Croneberg. % B. Km. IS. — A, embryo, and II, newly-hatched young form of Galeodes araneoides (after Crone- berg). a, anus ; ch, chelicerae ; peel, pedipalps ; p, limbs ; r, rostrum. The first embryo discovered by Croneberg was already in an advanced stage, not far from hatching. In Fig. 18, A, it is seen to be very like the embryo of an Araneid. As in the latter, the spherical abdomen, probably well filled with yolk, forms the chief * [Leredinsky (App. to Lit. on Opiliones, No. V.) describes this gland in Phalangium opilio as arising entirely from the mesoderm, the ectoderm only sharing in the formation of the external aperture. He expresses his belief that the coxal glands of Arachnids, the antennae, shell, and coxal glands of Crustacea and Limulus, are all nephridia and thoroughly hemodynamic, but perhaps not thoroughly homologous, some being derived from the primary and others from the secondary coelom. See also, Faussek (App. to Lit. on Opiliones, No. IV). — Ed.] f [See Bernard, App. to Lit. on Solifugae. No. I.] X [BiRULA (App. to Lit. on Solifugae, No. II.) finds that the ova of Galeodes develop within the cavities of the ovaries ; there are no embryonic membranes ; the thoracic and abdominal segments are visible before the appendages. A flexure-reversal occurs as in the Araneae. Hutton states that the Solifugae are oviparous. — Ed. ] SOLIFUGAB. 35 part of the body. The broad and flattened cephalo-thorax seems closely pressed against the ventral surface of the abdomen. The rudiments of the limbs are seen on the cephalo-thorax ; the chelicerae are bent towards the rostrum (Fig. 18, A), the latter being approxi- mated to the slit-like anal aperture. After the embryo is hatched, the abdomen appears longer, and shows a few slight constrictions, which no doubt correspond to seg- ments (Fig. 18, B). It carries two rows of dorsal setae, six in each row. These are the only traces of the hairy covering which is so profuse in the adult. The chitinous integument of the young is thus only provisional. The young probably remain for some time after hatching in a pupa-like condition, resembling in this respect the Araneae (p. 58), which after leaving the egg remain quiescent surrounded by a cuticular envelope, which is not cast off for some time. This fact explains why the limbs (now bent backwards) up to this time show no traces of segmentation (Fig. 18, B, Croneberg), and are also devoid of claws. No abdominal limbs were found in the young animal, nor was their presence to be expected at so late a stage. A very remarkable structure, not occurring in the adult,* is a pair of wing-like appendages, which arise dorsally between the points of insertion of the first and second pairs of limbs. These outgrowths consist of a double layer of cells, invested with a cuticle, and thus represent integumental folds ; no nerves or tracheae extend into them, and they are also devoid of muscles. The significance of these -wing-like appendages is not understood. Croneberg compares them to the paired appendages of the Asellus embryo (Vol. ii., p. 151), which are to be regarded as vestiges of the shell, but lays no special stress on this comparison. f The Solifagae are distinguished from the other Arachnida by a few important features, in which they seem more nearly to approach the Insecta. The most anterior pair of limbs with the segment to which it belongs enters into close relation with the preceding (cephalic) segments, and is marked off from the posterior (thoracic) segments, so that a separate cephalic region with three pairs of limbs ai-ises. This has been compared to the head of the Insecta and the next region, which now consists only of three segments, each with a pair of limbs, to the thorax of the Insecta. The resemblance is increased * Croneberg examined adults of the same species, and found that this structure was altogether wanting in them. t [It is now generally agreed that these structures are embryonic sensory organs, and similar to those found in Phrynus, See Bruce (Lit. on Pedipalpi, 26) and Laurie (App. to Lit. on Pedipalpi, No. I.). — Ed.] 36 ARACHNIDA. by the fact that the abdomen consists of ten segments visible externally. It is a striking fact that the Solifugae, which breathe by means of dendriform tracheae, possess at least three pairs of stigmata; the first opens on the fourth segment of the body, viz., the second thoracic (i.e., the first free thoracic) segment; the second pair opens on the second abdominal, and the third pair, which are closely approximated, open on the third abdominal segment. A fourth opening may be present as a median stigma on the fourth abdominal somite.* We cannot agree with those who find actual relationship to the Insecta implied in the very striking features we have mentioned, and regard the Solifugae as a connecting link between the two stocks of air- breathing Arthropoda. The value of a division of the anterior body into head and thorax, in which the three anterior pairs of limbs would have to be considered as the equivalents of the three pairs of oral limbs in the Insecta, is diminished by the fact that one pair is still wanting, i.e., there is in the Solifugae no homologue for the antennae of the Insecta. The most difficult point to explain is the position of the pair of stigmata on the cephalo-thorax ; we can only assume that it was acquired later. The assumption gains in proba- bility when we find that stigmata appear on the cephalo-thorax in the Acarina also, on the legs in Opiliones (Hansen), and on the head in Scolopendrella and Sminthurus (?). The presence of a spiral filament in the tracheae of the Solifugae is no proof of their relationship to the Insecta, since it occurs also in other Arachnida. In spite of the external division of the body into three parts, the Solifugae agree so closely with the Arachnida in outer and inner organisation, that we are not justified in separating them from that class. The shape of the chelicerae, the possession of a coxal gland, like that which is found in the Arachnida (MacLeod, No. 44), the hepatic tubules derived from the enteron,f the position of the genital aperture on the first abdominal segment, and other less striking features favour the Arachnid character of the Solifugae. We there- fore regard them as a branch of the Arachnid stock developed in a special direction, a view which corresponds to that of Ray Lankester (No. 45) and other writers on this subject. The slight * [Beunakd, op. cit.] t With regard to the liver, it should lie mentioned that a more recent observer (Biiujla, No. 42) has found certain differences in structure between this organ in the Solifugae and the Arachnida in general. He also, however, describes the liver as a well-developed organ filling up the interstices between the other organs, a description which applies to the liver of an Arachnid, but not to that of an Insect. ARANEAE. 37 data that are afforded by ontogeny confirm our view, the embryo of Galeodes closely resembling an Araneid embryo. More accurate data as to the development of the Solifugae are very desirable. „ ,. VI. Araneae.* systematic : A. Tetrapneumones. Avicularia (Mygale), Atypus. B. Dipneumones. Epeira, Theridium, Agalena, Lycosa, and all the other Araneae mentioned. Oviposition and the Constitution of the Egg. The Araneae build nests or prepare cocoons for their eggs, and usually watch over them. In many cases the cocoons are carried about by the mother, held by the chelicerae {e.g., Dolomedes, Pisaura) or attached to the abdomen {e.g., Lycosa, Tarantula). The eggs, which are rich in yolk, are surrounded by a vitelline membrane as well as by an external envelope, probably secreted by the oviduct, the latter being described as the chorion. A thin protoplasmic layer (the periplasm or blastem) covers the yolk, which in turn surrounds a central mass of protoplasm (the centroplasm), within which the nucleus is situated ; from this central mass fine protoplasmic strands extend to the surface, thus breaking up the yolk into columns. Besides the nucleus, a remarkable structure is found in the eggs of Araneae, and called the yolk-nucleus, but this is not yet sufficiently understood. It consists of a compact accumulation of spherules ; occasionally it is quite a complicated structure, composed of several concentric layers. When the egg matures the yolk-nucleus usually disappears, but it appears sometimes to be still retained, and is said to be still found near the nucleus in one of the yolk -complexes in the two- and four-celled stages of cleavage (Kishinouye). According to Ludwig (No. 66), the external envelope is marked out into polygonal areas, but this has recently been referred to the breaking up of the periplasm into polygonal divisions, Sabatier (No. 70) and Locy (No. 64), these writers thus agreeing with older statements made by Balbiani (No. 46). This polygonal marking must not be confounded with blastoderm-formation (which only occurs later) ; the former is said to appear even before cleavage * [Pocock divides the Araneae into two groups — A. Mesothelae, comprising one genus, viz., Liphistius, with a segmented abdomen. ■d r\ ■ ii u i ( Mygalomorphae. B. Opisthothelae J A/a°chn0U10rphae._ED.] 38 ARACHN1DA. takes place. Locy, with whom Kishinouye agrees in the main, explains these markings hy contractions of the egg after it is laid, drawing the periplasm closer to the yolk. The columns of yolk-granules can be separately recognised at the periphery as prominences, and this causes the polygonal markings on the surface. Some of Bai.biani's numerous figures that bear on this point seem to confirm this view, while others contradict it. In these figures, besides the original division of the periplasm into areas, another and true division is shown, caused by the presence of blastoderm cells. Since the egg is said to contract, there might be a regular folding of the vitelline membrane (in the form of polygonal areas), such as is said to occur in Cetochilus (Grobben), but this possibility seems to be excluded, as Locy mentions a perivitelline fluid which appears when the egg contracts, between its surface and the vitelline membrane. 1. Cleavage and Formation of Germ-Layers. Cleavage may here at first be described as total, but passes later into a superficial form. The central nucleus divides, the two daughter nuclei still lying near the centre of the egg (Fig. 21 ^4). Fig. 19. — Three ontogenetic stages of Philodromus limbatus (after H. Ludwig, from Balfour's Text-book). Although there is no furrow dividing the egg into two, a complete division is indicated, at first, however, only in the yolk. The yolk- granules become arranged radially one behind another in the form of cylindrical columns (Ludwig, Figs. 19 and 21 ^4). These columns, radiating from the centre, become divided into two groups by the division of the nucleus into two (Fig. 19 B). Between them lies formative yolk. As nuclear division proceeds, the two groups of columns, which Ludwig described as rosettes, again divide, and yield four rosettes (Fig. 19 C), which then divide further into eight, sixteen, and thirty-two rosettes, following the usual course of total CLEAVAGE AND FORMATION OF GERM-LAYERS. 39 and equal cleavage. Each rosette, which has now become a simple column, has a nucleus. In the further course of cleavage (Fig. 20 A) the nuclei shift to the periphery, accompanied by the formative protoplasm belonging to them. These, together with the periplasm already present, separate from the yolk to form a peripheral layer, which now contains the nuclei, and must thus be described as the blastoderm (Fig. 20 B, bl). The yolk-columns, or rather pyramids, may still be present at this time. Even earlier a cavity appears at the centre, the cleavage-cavity (Fig. 20, B), the central yolk-mass being withdrawn into the blastoineres as they develop, and pressing further towards the periphery. The yolk-rosettes do not seem, as a rule, to be so distinct as Ludwig found them in Philodromus. Yolk-pyramids have also been seeninAgalena, Titer idium, ^ B Epeira, Phol- cus, and other forms, but the groups formed by them (the rosettes of Phi- lodromus) lie closer to one another (Fig. 21 ^4). A stage in which there are eight such groups closely resembles an egg that has undergone total and equal cleavage, and that has a small cleavage-cavity (Fig. 21 B). Each group of yolk-columns with its nucleus corresponds to a blastomere. The blastoineres here also divide further, as in a case of equal cleavage, and when, after repeated division, a large number of blastomeres (about 128) have been formed, the nuclei, which have meantime shifted to the periphery, with their protoplasm, separate from the yolk below them, and thus give rise to the blastoderm (Fig. 21 C and D). The cleavage -cavity, which may be fairly large (Figs. 20 B and 21 G), becomes again filled with yolk, and the regular arrangement of the latter is gradually lost (Fig. 21 D and E). The formation of the blastoderm seems to take place more rapidly in the one half of the egg than in the other (Fig. 21 E), (Salensky, Ludwig, Locy, Morin, Schimkewitsch). The former is Fig. 20.— Superficial aspect and optical section of a later stage in the cleavage of Philodromus limoatus (after Ludwig, from Balfour's Text-book). U, blastoderm ; yk, yolk-pyramids. In the space between the vitelline membrane and the blastoderm the perivitelline fluid is found (11). 40 ARACHXIDA. the region in which the germ-band appears later, and may possibly correspond with the germ-disc from which the blastoderm spreads in Scorpio. The method of cleavage of the Araneid egg agrees closely with that of Crustacean eggs classed under type II. (Vol. ii., p. 109). If, as appears probable to us, a cleavage-cavity does not occur in all Araneid eggs, the centre in some cases remaining filled with an unsegmented mass of yolk, these latter cases would probably be referable to that type which was described, in connection with the Crustacea, as total cleavage with subsequent transition to superficial cleavage. Fio. 21.— Sections through the egg of Theridium mmmlatum in different stages of cleavage and blastoderm-formation (after Morin). bl, blastoderm; d, yolk ; dp, yolk-pyramids ; dz, yolk- cells ; fh, cleavage-cavity ; p, periplasm. [The blastomeres are, for the most part, flattened (Fig. 21 JE), but at one spot the cells become spherical and multiply rapidly, conse- quently a large accumulation of blastoderm-cells, several deep, forms at this spot. By reflected light this spot appears as a round whitish area. Shortly after this a second thickening, and consequently a second white area, appears. These two thickenings herald the formation of the germ-disc] CLEAVAGE AND FORMATION OF GERM-LAYERS. 41 There is little agreement among authors concerning the onto- genetic processes which follow the formation of the blastoderm, some ascribing great significance to the prominence, called by Claparede the primitive cumulus, which appears in the blastoderm by the thickening of the cell-layer * (Figs. 22 B and 23 A and B), others denying its importance. According to Morin, a thickening of the blastoderm arises in the region which corresponds to the later ventral surface, i.e., to the rudiment of the germ-band (Fig. 21 F) ; not only do the cells here increase in size, but some of them separate from the blastoderm, and form definite layers ; the blasto- derm thus becomes multilaminar. At the same time a few cells a. &.■ Fig. 22.— Sections through the egg of Pholeus plialangioides during the formation of the germ- layers (after Morin). c.p. primitive cumulus ; d, yolk ; dz, yolk-cells ; e, point of ingrowth. in this region become entirely disconnected from the rest, and migrate into the yolk (Fig. 21 F, dz). The three germ-layers may be now recognised. An outer layer, which constitutes the greater part of the blastoderm, is the ectoderm. Below this, at one pole of the egg, is the mesoderm, while the cells which have migrated into the yolk represent the entoderm. In the Araneids observed by Morin, the primitive cumulus arose only after the germ-layers had formed, if indeed it arose at all. It was wanting in Theridium, the form in which the origin of the germ-layers has been just * [Unfortunately there is a good deal of confusion surrounding the term "primitive cumulus." As stated above, there are two thickened white areas in the blastoderm ; and according to Kishinotjye, Claparede overlooked the first of these, and applied the term " primitive cumulus " to the second. The former author terms them the primary and secondary thickenings ; but Kingslet, while agreeing that Kishixouye may be right, nevertheless retains Clai'Arede's term " primitive cumulus" for the first thickening. — Ed.] 42 ARACHNIDA. described. It is, however, not impossible that those cases in which it is wanting are not primitive, but are specialised, and that it really is of greater importance than its late appearance in Pholcus and its entire absence in Theridium would lead us to believe. This last view is confirmed by the recently-published work of Kishixouye (p. 44). The primitive cumulus* arises as a thickening of the blastoderm (Fig. 22 B), and may project from it as a prominence of considerable size (e.g., in Tegenaria and Agalena, Fig. 23, A and B, p. 46). It has been found in most of the Araneae as yet examined. A depression is said to appear in front of it (Salensky, No. 71, Schimkewitsch, No. 72). We are tempted to regard the latter as the blastopore, at the posterior edge of which the ingrowth of cells is specially active, as in the Scorpiones (p. 2). Some of the statements as to the relation of the primitive cumulus to the germdayers in the process of formation (e.g., those of Bruce, No. 54, and Lendl, No. 63) must evidently be understood in this way. If we consider the primitive cumulus to lie at the posterior end of the embryo, we find ourselves in the position which was taken up by Balfour (No. 47). Although, since the time of this writer, the ontogeny of the Araneae has been investigated by several zoologists, very little further light has been thrown on this point. According to the above view, the primitive cumulus corresponded more or less to the future caudal end, the depression lying in front of it, and the cephalic lobes again in front of this (Balfour, Schimkewitsch, Lendl) ; according to another view, the caudal end arises at some distance from the primitive cumulus, the cephalic lobes lying nearer it (Balbiani, Locy). In inclining rather to the view that the primitive cumulus corresponds to the posterior end of the embryo, we are actuated chiefly by theoretical considera- tions. The figures given by Morin and Schimkewitsch seem also to support such a conclusion. It is, however, true that there is little convincing evidence for our assumption that the mesoderm arises from the primitive cumulus. There is, indeed, evidence of active proliferation of cells in the primitive cumulus, but in front of it also (in the region of the future germ-band) the blastoderm appears to be multilaminar (Fig. 22 B). It has already been mentioned that Morin entirely denies this significance of the cumulus. According to him, when such a prominence appears, it arises only after the development of the germ-layers. It cannot, however, be denied that Morin himself represents it as of considerable size (Fig. 22 B). It decreases later by giving off isolated mesoderm cells, and, by degrees, shifts dorsally. This displacement is also evident in Claparede'.s figures, if indeed the prominence seen in them actually corresponds to the primitive cumulus (Fig. 25 A and B, p. 48). That the blastopore, or the last * [The authors here use this term in the sense in which it was originally applied by Claparede, i.e., they apply it to the thickening which forms a projection from the surface of the blastoderm, which Kishinouye termed the secondary thickening. Kingsley, on the other hand, terms the primary thickening in Limulus the primitive cumulus. — Er>.] CLEAVAGE AND FORMATION OF GERM-LAYERS. 43 traces of it, should occupy such a position is from previous evidence improbable, unless we may assume that the proliferating area shifted as the posterior end developed, and thus attained a position which is apparently dorsal. Further ■ discussion of this point is unadvisable, as a glance at the figures of Clatakede, Balbiaxi, Salexsky, Balfour, Schimkewitsch, Locy, and Moein shows that they cannot be brought into agreement with one another. It is evident that the difficulty which attends investigation of this point is the cause of our uncertainty with regard to it. Orientation in the almost spherical egg is rendered still more difficult by the appearance of the different parts of the embryo (cephalic lobes and caudal end) simultaneously with the degeneration of the primitive cumulus. On this account one of the more recent investigators of the ontogeny of the Araneae, Kishixouye, was unable satisfactorily to decide the position of the primitive cumulus in the embryo. We must, for the present, accept with some hesitation the view that the depression which appears in the blastoderm of the Araneae and the primitive cumulus corresponds to gastrula- tion, although such an interpretation appears very probable, especially when comparison is made with the Scorpiones. This subject is not exhausted with the question as to whether the germ-layers originate in a region corresponding to the later ventral surface, in which the primitive cumulus represents an area of active cell -proliferation (perhaps a point of ingrowth), for there exists a different interpretation of the origin of the germ-layer. According to the view given above, it is to be assumed that the cleavage-cells shift to the periphery to form the blastoderm, and that the germ-layers originate there by an ingrowth of cells (Figs. 21, F, and 22, A and B). While the mesoderm remains as a compact accumulation on the ventral side, the cells of the entoderm become detached from it and shift into the yolk ; from these the enteron forms later. The origin and the fate of these yolk-cells is otherwise described by Balfour, Schimkewitsch, Locy (?). The most important point in these diverging views is the assumption that some of the cleavage-cells remain in the yolk. These cells, which are not utilised in the formation of the blastoderm, do not represent the entoderm alone, but some of them give rise to mesoderm-elements (Balfour, Schimkewitsch). According to Schimkewitsch, cleavage and the formation of the blastoderm take place in such a way that the egg breaks up into a large number of yolk- pyramids in the manner already described. Each of these pyramids contains a nucleus which at first lies at the centre. The nuclei shift to the periphery later, and there, with the protoplasm which surrounds them, become separated from the yolk. An outer cell-layer, the blastoderm, is thus formed. It appears, however, as if a further division of the nuclei had taken place previously, and a large number of nuclei had remained within the yolk ; at least, this is what we understand from Schimkewitsch's description of the cleavage-process.* During the development of the blastoderm there is a further increase in the number of the nuclei which remained within the yolk. Before following its further fate, we must mention a process which was observed by Schimkewitsch in Araneid eggs, and had been previously noticed by Salexsky. The blastoderm - cells which at first surround the egg, shift towards the ventral side, and there * The statements of Schimkewitsch as to the breaking up of the yolk- pyramids and the formation of uninuclear and multinuclear yolk-cells do not come within our scope, and also require corroboration. As a whole, his figures agree with the descriptions of earlier writers. Schimkeavitsch also found the central cleavage-cavity in a few forms ( Tegenaria, Epeira), and describes it as filled with masses of yolk, in the way described for Theridium (Fig. 21, C and D). •44 ARACHNIDA. form a thickening which, together with the later proliferation of cells at this spot, yields the rudiment of the germ-band. Morin's account also, as far as we can follow it, seems to confirm this, and the figures adopted from him (Fig. 21 D-F) show that an accumulation of blastoderm-cells at first lies on the dorsal side of the egg, while at a later stage only a few cells are perceptible in this region. According to Schimkewitsch, the dorsal side of the egg becomes completely denuded of blastoderm, which only later grows out towards it again. We were at first disposed to attribute the absence of blastoderm on the dorsal side rather to a belated advance of the nuclei out of the yolk, especially as authors state that the formation of the blastoderm progresses from the ventral to the dorsal side. There seemed here to be a distant resemblance to the cleavage and the formation of the blastoderm as observed in the eggs of Scorpio. Further investigation is needed to show whether this conjecture is correct, or whether such a marked redistribution of the blastoderm-cells as is shown in. the figures actually takes place. A similar crowding together of the blasto- derm-cells, though not nearly to such a great extent, has also been observed in other Arthropoda (Astacus, cf. Vol. ii., p. 128). According to Schimkewitsch, who on this point is essentially in accord with Balfoui:, the yolk-cells take part to no inconsiderable extent in the formation of the mesoderm, although the chief mass of them is to be described as entodermic. Schimkewitsch, like Balfouk, assumes a two-fold origin for the mesoderm, inasmuch as it is formed from the thickening of the ventrally situated blastoderm, especially from the primitive cumulus, and also by the addition of yolk-cells to this thickened region. Certain modifications here appear in individual forms (Tegenaria, Epeira, Lycosa) ; upon these, however, we shall not enter, as we are unable to agree with this view. Of the two opposed views, the one assuming the existence of yolk-cells giving origin to the entoderm and the mesoderm to some extent, the other deriving both the entoderm and the mesoderm from the blastoderm by a process comparable to gastrulation, the latter appears to us to be by far the more justifiable. This view is confirmed by Kishinouye's recent work (No. 62). This observer found no nuclei in the yolk after the formation of the blastoderm, but observed cells migrating into the yolk from the blastodermic thickening (Figs. 21 and 22). These cells, which become distributed through the yolk, form the entoderm. Further thickening of the ventral region of the blastoderm gives rise to the mesoderm, as was described above (p. 41). The ventral blastodermic thickening known to us as the primitive cumulus is in any case of significance in connection with the formation of these two germ-layers, for it, like the ventral plate (to be described later), appears before the differentiation of the germ -layers (Kishinouye), and not after it, as Monix assumed (p. 41). When we trace back the formation of the germ-layers to the blastoderm, we thereby imply that the yolk-cells also arise from the blastoderm. These latter, according to the unanimous opinion of authors, contain, in the Araneae, the rudiments of the whole entoderm, giving rise later to the epithelium of the enteron. If these cells were to remain in the yolk when cleavage takes place, the process of blastoderm-formation would have to be regarded as epibolic, but this is contradicted by what occurs in related forms. The germ-layers are moreover formed in the Scorpiones also by the DEVELOPMENT OF THE EXTERNAL FORM OF THE BODY. 45 ingrowth of cells on the ventral side of the blastoderm, and the entoderm, when first appearing, occupies this position in both these divisions of the Arachnida. In the Scbrpiones, it forms a regular epithelium, so that it cannot fail to be recognised as a separate germ-layer, but here also isolated cells enter the yolk. All these facts confirm us in regarding the view of the formation of the germdayers adopted by Morin and Kishinouye (Fig. 21 F) as correct. It cannot, however, be denied that the figures given by Schimkewitsch, and especially those given by Balfour, show yolk- cells in earlier stages and further removed from the thickened part of the blastoderm, which might rather be assumed to have remained behind in the yolk at the time of cleavage, than to have become detached from the thickened part of the blastoderm. If this should be the case, the view here taken is not thereby contradicted; we have then to do merely with single cells which were not utilised in the formation of the blastoderm, and remained behind in the yolk. These cells, as vitellophags, perhaps render the absorption of the yolk possible. In that case we must assume that they do not later enter into the formation of the entoderm, but probably disintegrate during the gradual disappearance of the yolk, as is the case with corresponding (yolk) cells in the Insecta. 2. Development of the External Form of the Body. The development of the external form of the body has been repeatedly investigated more or less thoroughly in the cases of Agalena, Clubiona, Epeira, Theridium, Lycosa, and PJtoIcus, and has been found to follow a very uniform course. In spite of this fact, and although a large number of zoologists, among whom we may mention Herold, Claparede, Salensky, Balfour, Schimkewitsch, Locy, and Kishinouye, have investigated the subject, some points, especially in the earlier ontogenetic stages, still remain obscure. The chief difficulty is connected with the early rudiment of the embryo and the first appearance of segmentation. At a time when the blastoderm is either approaching completion or is fully developed, there appears (probably on the later ventral side) the prominence known to us as the primitive cumulus, the sig- nificance of which has already been discussed (p. 41, etc.) From this there extends forwards a band which is distinguished by its white colour from the rest of the egg, and is caused by a marked thickening of the blastoderm (Fig. 23 A, Claparede, Balfour). Herold mentions a comet-like structure which arises at an early stage on 4G AHACHXIDA. the surface of the Araneid egg, this comparison heing apparently suggested by the band just described, together with the primitive cumulus (Fig. 23). The band soon widens at the end furthest from the primitive cumulus, and it becomes still broader as the thickening of the blastoderm extends out laterally from this region. Such a lateral extension of the blastodermic thickening, starting from the band, implies that we regard the band itself, as well as the primitive cumulus, as thickenings of the blastoderm, which have arisen by active increase of cells at these points. According to Salensky, a depression appears in front of the primitive cumulus ; this soon closes again, and is regarded by him as the blastopore. We are disposed to attribute the same significance to that thickening of the blastoderm which was mentioned above in the description of the formation of the germ-layers. "We thus assume that the primitive ^S. **" A, Fig. 23. — Superficial aspect of three early stages in the development of an Araneid, showing the embryonic rudiment (A and B, Ayalena labyrinthica, after Balfour; C. Theridium, after Morin). c.pr, primitive cumulus; h, posterior; v, anterior. cumulus lies at the future posterior end, and that the band runs out from it anteriorly. Its position therefore indicates the ventral surface. The latter is clearly recognisable as such at a somewhat later stage, the blastodermic thickening extending further, and finally becoming evident on the surface of the egg as a region shaped somewhat like an isosceles triangle (Fig. 23 C). The basal part of this triangle seems to appear first (Fig. 23 B), and then by degrees the parts nearer the apex. The base of the triangle corresponds to the rudiment of the cephalic lobes, the point to the posterior end of the embryo. According to this account, the primitive cumulus would occupy the apex of the triangle, and must be looked for in the posterior region (Fig. 23 B), and the band which developed at first and proceeded from the primitive cumulus (Fig. 23 A) would DEVELOPMENT OF THE EXTERNAL FORM OF THE BODY. 47 then indicate the longitudinal axis of the embryo. The whole of the triangle thus represents the germ-band or the so-called ventral plate. Embryonic envelopes, such as are found in the Scorpiones (p. 4), are wanting in the Araneae. The amniotic fold described by Bruce on the head of the Araneid embryo must without doubt be referred to the infolding which takes place during the formation of the brain. The formation of cuticular larval integuments will be again referred to later (p. 58). At about the time when the rudiment of the germ-band (the so-called ventral plate) first appears (Fig. 23 A-C), the egg is said to be flattened on this side, but at a slightly later stage the ventral surface of the embryo appears much arched (Figs. 24 and 25 A), either because this surface has become secondarily convex, or because this special region has not been affected. In Pholcus it appears to be the dorsal part that is flattened (Fig. 25 A and B), and Claparede mentions that in this way the anterior and posterior ends are approxi- mated. Fig. 24. — Young embryo of Clubiona incompta show- ing the commencement of segmentation of the germ-band (after Salensky). hi, cephalic lobe; si, caudal lobe ; between these are a few segments in the act of forming. The larger cells outside the region of the germ-band are said to repre- sent blastoderm -cells which are here less crowded (Salensky). The segmentation of the germ-band begins with the appearance of a few transverse furrows which mark off a large anterior and a posterior region, as well as several intermediate segments (Fig. 24). These segments at first appear very indistinct, the parts of the body to which they correspond being doubtful. In the youngest segmented stage, three segments were found besides the large anterior and posterior regions (Fig. 24, Salensky, Balfour, Locy, Lexdl). These seem to correspond to the first three thoracic segments. According to Locy, we must, however, assume that the three middle segments represent the second, third, and fourth thoracic segments. He believes that the segments develop in the following order ; fourth, third, second, first thoracic segments, then that bearing the pedipalps, and last of all that carrying the chelicerae. The differentiation of the segments would thus take place from behind forward, an exact reversal of the order usually met with in segmented animals. There is a general resemblance between this view and that adopted by Metschnikoff for the Scorpiones, according to which the embryos at first break up into three regions, the anterior corresponding to the cephalic region, 48 ARACHNIDA. the posterior to the telson, with the as yet undifferentiated segments of the post-abdomen, and the middle part giving rise to the other segments of the body (p. 6). It is well to bear in mind how extremely difficult it is to orient the regions of the body in these early stages of the Araneid embryo (cf. Figs. 23-25). This difficulty throws some doubt upon the correctness of the identification of the body-segments, given above on Salensky's authority, and consequently the whole order of formation of the thoracic somites may be at fault. Fig. 25.— Embryos showing varying degrees of segmentation, but not yet provided with limbs. A and Jl, Phokus opilionoides; C, Club'wna (after Claparede). A and B, lateral aspects ; C, ventral aspect, ch, cheliceral segment ; c.pr, primitive cumulus (?) ; eh, egg- integument; h, posterior, kl, cephalic lobes; ped, pedipalpal segment; I-IV, thoracic segments ; 1, first abdominal segment ; si, caudal lobe ; v, anterior. The segments of the pedipalps and chelicerae are said by almost all authors to appear later than the four thoracic segments. Just as the posterior region of the embryo contains in itself a number of segments, so also does the anterior lobe comprise, besides the cephalic part, the segments of the chelicerae and the pedipalps. A regular separation of segments from before backward, therefore, does not take place. At the stage in which there are six segments interposed between the cephalic and caudal lobes (Fig. 25 A and B) the four posterior segments are much better developed and more distinctly marked off than the two anterior segments. .Balfour, Schimke- DEVELOPMENT OF THE EXTERNAL FORM OF THE BODY. 49 witsch, and Locy give figures of this stage in Agalena, in which the cheliceral segment is still united to the cephalic lobe, or is in the act of separating from it. We are unfortunately not able, from this description, to decide the order in which the thoracic segments become differentiated, but it seems as if the most posterior segment {the fourth) arose after the others. The abdominal segments separate from the caudal lobe in the usual order, i.e., from before backward. As the germ-band segments, it extends further over the egg; and not only do its anterior and posterior ends grow towards the dorsal side, but it extends laterally, and may thus, in a few forms {e.g. Pholcus), cover the greater part of the surface of the egg (Fig. 25 A). Seen from the ventral surface, the germ-band now appears broken up into segments, which extend transversely across the whole surface of the egg (Fig. 25 C). The segments appear somewhat narrow and as if separated by broad transverse furrows. The egg therefore somewhat resembles the dorsal surface of a rolled -up Isopod. This condition, however, is not long retained, a lateral •contraction of the germ-band taking place which causes it to draw back again on to the ventral surface (Fig. 25 B), and to lie there in the form of a segmented band. The cephalic and caudal lobes retain their positions unchanged during this process, and, owing to the dorsal extension of the anterior and posterior extremities of the germ-band, they appear closely approximated (Fig. 25 B). In those forms in which the germ-band does not extend so far over the egg in early stages {e.g. Agalena), the cephalic and caudal ends only approach one another on the dorsal surface at a later period. The shape of the germ-band becomes modified, the cephalic portion widening and assuming a bilateral, bilobed form ; the abdom- inal segments, further, become separated from the caudal lobe, which has also widened. There may be as many as twelve abdominal segments besides the telson (e.g., Pholcus, Schimkewitsch). The abdomen is thus richhj segmented in the Araneid embryo, in direct opposition to its condition in the adult. The complete segmentation of the abdomen does not take place till the later stages, other im- portant modifications in the germ-band preceding it. The first of these to be noted is the appearance of a longitudinal furrow in the ventral middle line (Fig. 28 A), which is caused by the division of the mesoderm lying on the ventral surface into two bands, these subsequently shifting to a more lateral position. The germ-band is in this way divided into two symmetrical halves (Figs. 28 A and E 50 ARACHNIDA. Long before d-~ B, and 26), which may lie so far apart that the yolk protrudes between them (e.g., Agalena, Balfour, Fig. 29, p. 53). Anteriorly, in the cephalic lobes and also at the caudal end, the two halves of the germ-bands remain united (Figs. 28 A and B, 26). the germ -band has divided to such an extent, the rudiments of the limbs- have appeared, the first to be seen being those of the four pairs of ambulatory limbs, as slight prominences a little re- moved from the median groove (Fig. 28 A, 3-6). These are followed by the rudiments of the pedipalps (2), and, a little later by those of the cheli- cerae (1). The rudiments of limbs arise in the same way on the first four abdominal segments (Figs. 28 A, a, 27), so that the abdomen of the embryo is not only much more fully segmented than that of the adult, but even has limbs on some of its segments. In this respect the Araneae re- semble the Scorpiones, which also have limbs on the ante- rior abdominal segments (p. 8). Further similarity is found in the fact that, in the former, the posterior part of the abdomen may be flexed forward ventrally like the post-abdomen of the Scorpion embryo. This is the case in Pholcus, as was pointed out by Claparede, and confirmed by Emerton, Schimkewitsch, and Morin. It is almost universally admitted that the first four abdominal segments carry provisional appendages (Balfour, Locy, etc.). Even Morin's researches, carried out by the help of the latest methods, yielded the same result, although Salensky had mentioned a first limbless segment, and Schimkewitsch had accepted this view. The statements of these two authors are supported by the notes and figures of Bruce, published after his death (No. 54). We were able easily to convince ourselves, by examination of an Araneid embryo Fig. 2ij. — Embryo of Phrfcus opilionoides, ideally unrolled (after Claparede). ch, chelicerae ; d, yolk ; kl, cephalic lobes ; ped, pedipalps ; Pi~Pt, first four pairs of limbs ; 1-3, first three abdominal segments ; pah, the posterior part of the abdomen flexed ventrally. DEVELOPMENT OF THE EXTERNAL FORM OF THE BODY. 51 Ae.__ Fig. 45. — Transverse sections through embryos of Thendium muculatum (after Morin). In A, the embryo, which is curved round the yolk, is cut through twice ; the thoracic limbs and primitive segments can be recognised below, while the abdominal primitive segments are seen above. D, cross-section through the abdomen of an older embryo, in which the primi- tive segments have increased in size, hi, blood-corpuscles; d, yolk; cfr, yolk-cells; ex, limbs ; I, lung invaginations ; n, rudiment of the chain of ganglia ; us, primitive segments. the cephalo-thorax and also in the abdomen, as far as the latter possesses appendages, the primitive segments extend into the limbs ; indeed for the time they withdraw almost entirely into the limbs (Figs. 44 and 45 .4). The mesoderm-bands naturally also take part in the displacement undergone by the two halves of the germ-band in consequence of the pressing forward of the yolk-mass to the ventral side. Whereas they formerly lay near the ventral median line (Fig. 43 C) they now appear removed from it, and divided by the so-called yolk-sac (Fig. 29 A, p. 53). The segmental cavities THE BLOOD-VASCULAR SYSTEM AND THE BODY-CAVITY. 87 increase considerably in size, the primitive segments extending towards the dorsal side (Fig. 45 B). This process exactly corre- sponds to that we have already met with in the formation of the coelom in the Annelida (Vol. i., p. 289). The following are the derivatives of the primitive segments : — 1. The somatic layer gives rise to the body -musculature (as thickenings near the ventral middle line of the abdomen), the two strong longitudinal muscles being specially noteworthy, and also to the subcutaneous connective tissue. According to Schimkewitsch, .the endoskeleton also is derived from the somatic layer, but this statement Ave give with reserve. The covering of the parts arising through invagination of the ectoderm (stomodaeum and proctodaeum, lungs, glands), together with their musculature, thus the strong musculature of the stomodaeum, already mentioned, is also derived from the somatic layer. 2. The splanchnic layer gives rise to the covering of the enteron, the blood-vascular system, and the genital organs. The coxal glands are probably related to the mesoderm and coelom, as in the Scorpiones (pp. 24 and 92), [cf. Brauer and Purcell]. G-. The Blood-vascular System and the Body-cavity. The Blood-vascular System. At a time when the limbs have already developed, there appear, above the primitive segments, between the ectoderm and the yolk, large round cells (Fig. 45 A and B, bl), concerning whose origin there is considerable difference of opinion. Balfour derived these cells from the yolk-cells. To the latter he also traced the origin of the dorsal mesoderm (Fig. 29, p. 53). This last assumption was refuted by Schimkewitsch, Locy, and Morin, who agree in stating that the primitive segments extend to the dorsal middle line. The cells which, at later stages (Fig. 29), are found dorsally, as in Figs. 45 B and 46, therefore belong to the primitive segments. But, besides these, there are the large round cells mentioned above (Fig. 45 A and B, bl), and with regard to their origin, Schimkewitsch and Locv agree with Balfour, deriving them from the yolk-cells. Kishinouye has recently adopted the same view, which seems in accordance with the constitution of these cells. They are much larger than the cells of the primitive segments (Fig. 45 A and B) ; we might, nevertheless, like Morin, derive them from these, and assume that they had separated from the primitive segments in an early stage, when the cells of these segments were themselves larger. Better nourishment near the yolk as the cells increased in number would also determine increase in size. This view is further supported by the fact that they are found in the cavities of the primitive segments (Schimkewitsch). This latter author, indeed, thinks that they reach these cavities from the yolk by breaking through the wall of the segment, but this view seems improbable. 88 ARACHNIDA. SfJ.W. Fig. 46. — Cross-section through the abdomen of an embryo of Phohus phalangioides (after Morin). bg, ventral chain of ganglia ; W, blood-corpuscles; d, yolk ; dz, yolk cells; h, heart; so, somatic, sp, splanchnic mesoblast; sp.iu, spinning mammillae. So long as the origin of the isolated cells lying between the ectoderm and the yolk is not definitely established, we may regard them asmesoderm-cells, and we are especially inclined to consider them as derivatives of the yolk-cells from a comparison with simi- larly related cells found in the Vertebrata, which are there undoubtedly derived from the yolk. These isolated cells eventually become blood-corpuscles. They collect dorsally during the upward growth of the primitive segments (Fig. 45 B), and, as they press somewhat closely against one an- other, they form (especially in the abdomen) a compact strand of cells which prevents the junction of the primitive segments in the dorsal middle line (Fig. 46, U). Subsecptently the mesoderm grows between this strand and the ectoderm, and thus the two primitive segments meet to form a partial dorsal mesentery. At a later period, the walls of the primitive segments grow between the yolk and this strand of cells, and unite with one another below the latter (Figs. 46 and 47 ^4). This strand of cells has consequently become enclosed by a layer of mesoderm having the form of a longitudinally-placed tube, which is at first attached to the somatopleure above and the splanchnopleure below. The tube soon loses its connection with its parent mesoderm (Fig. 47 B), and we now find a continuous layer of mesoderm (somatopleure) lining the ectoderm, while another layer covers the yolk (splanchnopleure) ; between these two layers is the body-cavity, in which the mesodermal tube now lies freely. This tube is the heart, and, so far as can be judged, it is formed directly from the walls of the primitive segments (Schimkbwitsch, Loot, Morin (Fig. 47 A and B)). As a consequence of the development of the heart, the primary continuity of the cell-elements of the primitive segments becomes interrupted at this point (Fig. 47 B). THE BLOOD-VASCULAR SYSTEM AND THE BODY-CAVITY. 89 (Compare with the development of the heart in the Annelida and in the Mollusca.) The isolated cells which had hecome grouped together into a strand become blood-corpuscles. Their crowded condition and their extremely close connec- tion with the walls of the primitive segments suggested the idea that the heart was derived from a solid mesodermal strand extending along the dorsal middle line (Balfour), but this view cannot be verified ; the formation of the heart may be directly compared with the similar process in the Annelida. The cavity of the heart corresponds to apart of the primary body-cavity, enclosed on each side by the primitive segments. Fig. 4". — Transverse sections through the abdomen of embryos of Theridium maculaUua, showing the formation of the heart (after Moris), hi, blood-corpuscles ; c, coelomic cavity ; d, yolk ; dz, yolk-cells ; ec, ectoderm ; h, heart ; so, somatic, sp, splanchnic mesoblast. The heart lies in a depression of the yolk (Fig. 47 B). The latter is covered only by the splanchnic layer of the mesoderm, as the entodermic epithelium is still wanting. From this part of the splanchnopleure, a mesodermal lamella is said to separate and grow round the heart to form the pericardium (Schimkewitsch). The alary muscles of the heart are then formed from the somatic mesoblast. The pulmonary veins arise as outgrowths of the peri- 90 ARACHNIDA. eardrum, while the anterior and posterior aortae, as well as the lateral arteries, originate as prolongations of the heart or as out- growths from it (SchiiMKEWITSCh). While the cavity of the heart appears to be a part of the primary body-cavity, the pericardial space, according to Schimkewitsch, corresponds to a part of the secondary body-cavity. The pericardium in the Arachnida forms a tube, and is not comparable with the synonymous structure in the Insecta. But before we can make any definite statement as to the nature of the pericardium we must have a more exact account of its origin. The Body-cavity. In the Arachnida, as in other Arthropoda, the blood-vascular system is not separated from the body-cavity, but the latter is directly connected with the circulation of the blood. The method of development of the body-cavity in the Arachnida is, however, strikingly different from that in the Crustacea, Myriopoda, and Insecta. While, in these latter, the primitive segments are not large and soon undergo degeneration, in the Arachnida they are almost as largely developed as in the Annelida (Figs. 45 and 46). The primitive segments are also highly developed in Peripatus to begin with (Fig. 100), but this form resembles the Insecta in that the segments very soon cease growing, and after a rich growth of cells undergo early disintegration. The adult body-cavity forms (as a pseudocode) outside the primitive segments. In the Arachnida it forms somewhat differently; it is, however, difficult, from the statements before us, to arrive at a satisfactory judgment, since little stress has until now been laid upon this point. It is certain, however, that the primitive segments are of considerable size even at a someichat advanced stage of development (Bigs. 46 and 47). Between the somatic and splanchnic layers of each primitive segment there is a rather large cavity, and we must assume that when the union of the segmental cavities takes place this passes direct into the adult body-cavity. It is true that here, also, the body-cavity would not retain the coelomic epithelium up to the last, but the wall of the primitive segments would also break up (Figs. 47 A and B, 41, p. 80, 42, p. 82), yielding the muscidar and con- nective tissue elements, so that at last, in the Arachnida, a condition would be reached similar to that attained at a much earlier stage in the development of the Crustacea, Myriopoda, and Insecta. The segmentation of the mesoderm begins to disappear when the primitive segments have grown to a considerable size and the embryo itself is near the stage illustrated in Fig. 27. The segmental cavities unite in the cephalo-thorax and the dividing walls (dissepi- THE BLOOD-VASCULAR SYSTEM AND THE BODY-CAVITY. 91 ments) gradually disintegrate, the cells falling into the body-cavity (Schimkewitsch). These cells probably give rise to blood-corpuscles. The primitive segments of the cephalic lobes seem already to have fused with those of the cheliceral segment, at least Schimkewitsch speaks of a connection between the two which, however, he explains in another way. If we understand Schimkewitsch rightly, he assumes that the pair of primitive segments in the cheliceral segment arise by division from the pair in the head; we should be more inclined to assume the opposite of this, i.e., an extension of the first trunk-segment into the cephalic region. It, however, appears from the accounts and figures before us that the cephalic and cheliceral segments undoubtedly have separate primitive segments. A union of these two pairs of segments, like that described by Kleinenbeisg for Lumbricus, would then take place. The two segmental cavities of the head become united ; such a union of the cavities of the two sides must take place in the trunk also as a result of the processes described in connection with the formation of the heart (Fig. 47). This at least applies to the dorsal side ; on the ventral side, the primitive segments are at first still far apart (Fig. 46), but they shift gradually towards the middle line, so that they finally extend round the whole mass of yolk. In the abdomen the primitive segments remain separate longer, a fact which is in keeping with their later differentiation. Even when they are fused together, the mesoderm represents two extensive layers passing into one another — an outer or somatic layer and an inner or splanchnic layer ; between these is the secondary body- cavity (Schimkewitsch). From the splanchnic layer, the folds already mentioned in con- nection with the formation of the intestine grow into the yolk (Fig. 42, p. 82), in this way cutting off from it isolated masses which correspond to the later hepatic lobes. "We should like here to draw special attention to the important fact that the yolk is so long a time bounded solely by mesoderm (Figs. 46 and 47), and that the epithelium of the enteron develops very late (Fig. 41, p. 80); indeed, the mapping out of a large part of the enteron, that of the liver, seems to be commenced by the mesoderm. Whether the distribution of these folds corresponds to a true segmentation appears doubtful, although this might be indicated by the appearance of four lateral folds in the cephalo-thorax. It appears that these correspond to the thoracic caeca of the enteron (?), for in the abdomen also a number of folds occur, and it is these principally that give rise to the form of the liver (Mof.ix). The folds which penetrate the yolk not only come from the side, but from above and below, and thus represent oblique as well as longitudinal layers 92 ARACHNIDA. (Schimkewitsch), so that it is impossible to trace them back, as Balfour attempted to do, to the partition walls of the somites. A structure resembling the fat-body of the Insecta which is present in the body-cavity (Ray Lankester's lacunar blood-tissue) is, according to Schimkewitsch, formed, like some of the blood-corpuscles, out of the yolk- cells which immigrate into the body-cavity, and these cells also are said to become arranged into a " peritoneal " layer, which envelops the internal organs. In both cases we should, after what has already been said, feel disposed to derive these structures from the mesoderm, i.e., from the primitive segments, although such a derivation would have to be established by further researches. Where a peritoneum is present, it would be of interest to learn its relation to the primitive coelomic epithelium. rrv. Fig. 48. — Portions of transverse se.ctions of Pholcits phalangiddes (A) and Lycosa saccata (B and C), through different regions of the abdomens of embryos (after Schimkewitsch). bl, blood -corpuscles or detached mesoderm-cells ; e, ectodermal covering of the body; g (and gi?), portions of the genital glands descending to the ventral side ; hi, median portion of the rudiment of the ventral cord; ran, muscles; n, rudiment of the ventral cord; so, somatic, sp, splanchnic layer of the mesoderm. H. The Coxal Glands. The coxal glands, which we shall describe as they are found in the already hatched Araneid, show great resemblance to those of Scorpio, and no doubt arise in the same way as in that animal. An actual efferent duct has for the most part only been proved to exist in young Araneids, where it opens at the base of the fifth pair of appendages (Bertkau, No. 51). In the young of Atypus, Bertkau found, on the anterior pairs of limbs, i.e., third and fourth pairs of appendages, slits corresponding in appearance and position to the apertures of the coxal glands on the fifth pair, and this led him to conclude that there were originally several pairs of these glands. THE GENITAL ORGANS — ACARINA. 93 Kishinotjye's derivation of the coxal glands from an ectodermal invagination winch lengthens into a tube is not only incompatible with their origin in the Scorpion from the mesoderm (p. 24), but also with their relation to the body- cavity. According to Kishtnouye's own statement, the tubular rudiment of the coxal gland opens at its inner end in the shape of a funnel into the coelom, so that the accepted view that these glands are nephridia is confirmed, provided the accounts given are correct. If these glands are ectodermal in origin, then they must be regarded, not as coxal, but as crural glands, and we would expect them to end blindly. Sturany (No. 14), the most recent investigator of the coxal glands in the Arachnida, considers them to be nephridia. If his con- jecture that they end in a terminal sac as in the Crustacea proves correct, the latter would no doubt correspond to a part of the body-cavity. We refer the reader to our account of the coxal glands in the Scorpiones (pp. 24 and 87). I. The Genital Organs. According to Schimkewitsch, the genital organs arise in the anterior part of the abdomen, within the two longitudinal folds of the splanchnic layer, which have risen up into the yolk from the ventral side. In the median layer of each of these folds an ovoid thickening appears (Fig. 48 ^4). This consists of large central and flat peripheral cells, the latter representing an enveloping epithelial membrane (Fig. 48 B). The anterior end of the rudiment curves round towards the ventral side, and is said to correspond to the efferent ducts, while the rest represents the germ-glands. "When the young Araneid hatches, there is still no communication between the efferent ducts and the exterior ; this is established later by means of an ectodermal invagination (Schimkewitsch). [Pueceli, (App. Lit. on Araneae, No. VII.) traces the ducts to tubular growths of the abdominal mesodermal segments ; the openings of these ducts into, the coelom become connected with the genital cells which grow forward from the posterior end of the germ-band. Similar structures develop in all the abdominal appendages, but only those on the second segment persist. He regards them as modified nephridia.] VII. Acarina. Oviposition. The majority of the Acarina lay eggs, a few (e.g., Halarachne) are said to be viviparous. Some (e.g., Scutovertex) appear to be viviparous at certain times of the year and ovo- viviparous during the remainder of the season, others are habitually ovo-viviparous. The Acarid egg is surrounded by a more or less strong shell, sometimes covered with prominences ; in many species this protective shell is extremely thick and traversed by fine pore- canals.* The eggs are deposited in various places, according to the * [According to Trox'ssaet (App. to Lit. on Acarina, No. VII.), the female of Syringobiu chelopus among the plumicolous Sarcoptidae at times reproduces parthogenetically ; the eggs thus produced in the absence of males have no shells. — Ed.] 94 ARACHNIDA. habits of the parent. They are found in decaying wood, in damp earth, on the under surfaces of stones, in dung-heaps, on leaves, fruit, etc. Some, but by no means all, parasitic forms lay their eggs on or in the body of the host. The eggs are at times laid in a heap, at other times separately; in the latter case they are often stalked ; those of Myobia musculi have a process at the posterior pole by which they are attached to the fur of the mouse. According to Hallbr, many Oribatidae carry their eggs attached to their backs, others are said to lay them in a part of their cast-off chitinous Fig. 49.— Cleavage ami formation of the blastoderm in the egg of Tetranychus tdarius (after Clapar£de, from Balfour's Text-book). The yolk-granules are represented by clear circles (in .-1 and D). The nuclei, with the clear areas of protoplasm around them, are much larger than the granules. C, an egg in the stage of blastoderm-formation. integument.* The form of the eggs is most commonly elliptical (Fig. 50), sometimes oval, and more rarely globular (Fig. 49), or even discoidal. For their size, they are richly provided with food- yolk. 1. Embryonic Development. The embryonic development in these eggs is difficult to follow on account of their minute size, and is therefore not well known. [Neither of these assertions is quite correct. The carrying of the eggs is almost entirely confined to the genus Damaeus — it is most commonly the immature individuals, not the adults, which pile the eggs on their backs; it is manifest that at this period they cannot have any eggs of their own to carry. What happens is that the larvae are born with soft abdomens, and have the instinct of piling upon their backs dirt, rubbish, etc., as a protection ; they will pick up and carry the eggs and empty egg-shells, from which they may them- selves have emerged, but they will equally pick up and carry the eggs of other Acarina. The statement concerning the eggs being found in the cast integument has never been confirmed, and is very doubtful. — Ed.] EMBRYONIC DEVELOPMENT. 95 Claparede's account (No. 77) is still the most complete.* According to this writer, in Tetranychus telarius, the nucleus, surrounded by formative protoplasm, rises to the surface of the yolk (Fig. 49 A) and soon divides. Eepeated division (Fig. 49 B) gives rise to a large number of nuclei, each surrounded by an area of protoplasm. The nuclei remain lying at the surface of the egg, and by increasing- still further in number, they, with the protoplasm around them, give rise to the blastoderm (Fig. 49 G). pr Fig. 50.— Embryonic development and formation of the first larval integument in Myobia musculi (after Claparede, from Balfour's Text-hook). In D, the egg-integument has split, and the embryo, surrounded by the first larval integument, is in the act of leaving the egg. ch, chelicerae; pd, pedipalps; pHA the first three pairs of limbs; pr, proboscis (which has arisen through the fusion of the chelicerae and the pedipalps) ; sL-s*, four post-oral segments. The yolk is represented by the darker area. According to Robin and Megnin (No. 104), total cleavage occurs in the eggs of the Sarcoptidac. The egg, while still in the oviduct, was seen to hreak up into four cleavage-spheres. This, if correct, reminds us of the condition described in the Araneid egg, which, however, does not there lead to the complete division of the egg into cleavage-spheres. Total cleavage is said to he undergone also by the egg of Cliclifer (p. 2S), and the same has been main- tained, at least at a later stage, of the eggs of the Opiliones also (p. 32). The blastoderm was examined in a number of Acarina, and always consisted of a thin single layer of cells enclosing the yolk. As it develops further, thickenings take place at points corresponding to the future ventral surface, especially in the cephalic and caudal i-egions. The germ-band thus arises here (Fig. 50 B) in the same way as in other Arachnids. At first it is an ecpually thickened * [Henking's account is perhaps more correct.— Ed.] 96 ARACHNIDA. band, but later it breaks up into two symmetrical halves, a ridge of yolk pressing outwards in the median line. Here also there is agreement with the Araneae. The germ-band soon becomes segmented (Fig. 50 .4). The cephalic lobe, which, in Myobia as in the Araneae, curves over to the dorsal side (Fig. 50 B), and the caudal lobe become segmented off from the trunk. The part lying between theni, which corresponds to the cephalo-thorax, is divided up into a number of segments, the truncated rudiments of the mouth-parts and limbs soon appearing on these (Fig. 50 B). This segmentation is less distinct in other Acarina, and, as is well known, eventually disappears. The abdomen is still comparatively large in such an embryo ; in many Acarina it is much reduced, or is united to the cephalo-thorax. Before development has pro- gressed thus far, a delicate struc- tureless integument separates, in At ax, from the embryo, and surrounds it, like a second egg- integument, in the form of a closed envelope (Fig. 51, dm). In other Acarina, this process only takes place later, when the limbs are already present, so that these are found on the envelope in the form of sheaths surround- ing the actual limbs (Fig. 52, dm). This delicate envelope, though separated from the embryo, is thus seen to be a true larval integument. The embryo is now enclosed in a double envelope, and the dorsal surface which, up to this period, showed little signs of development, being covered only by a thin cell-layer, now commences to develop by the growth of the mesodermal elements towards this surface. The yolk for some time longer retains its former appearance (Figs. 50-53), but we must no doubt assume that the formation of entoderm has already begun. Nothing certain is as yet known of the develop- ment of the germ-layers and the rudiments of the organs in the Acarina. The limbs of the embryo lengthen (Figs. 51 and 53 A) and become segmented (Fig. 52). In the stage depicted in Fig. 51, and more especially in the following stages, the embryos of many Fio. 51.— Embryo of Atax Bonzi surrounded by the deutovum and the egg-shell (after Claparede). eh, chelicerae ; d, yolk ; dm, deutovum ; eh, egg-shell ; kl, cephalic lobe ; 2'!-j),,, the three pairs of limbs ; ped, pedipalps ; si, caudal lobe. THE FORMATION OF THE LARVAL INTEGUMENTS. 97 Acarina show great resemblance to those of the Araneae (Figs. 51 and 57 A). The chelicerae and pedipalps unite to form the proboscis (Fig. 53).* The abdomen (in Atax) now decidedly preponderates over the anterior part of the body (Fig. 53). There are only three pairs of limbs when the embryo breaks through its envelopes and begins free life (Figs. 51-53, px-p^). We thus find, in the Acarina, a larval stage with only three pairs of limbs, as distinguished from the four pairs of the nymph and of the adult, which, in other points of both outer and inner organisation, the embryo greatly resembles.t 2. The Formation of the Larval Integuments and the Further Course of Development. It was mentioned that, in many Acarina, e.g., Atax, the embryo casts off a cuticular integument at an early stage when the limbs have not yet developed or are only indicated. Claparede's deut- ovum is thus produced, the embryo within the egg-shell thus becoming enclosed in a second envelope (Fig. 51). The resemblance of the "deutovum" with the embryo enclosed in it to an intact egg is increased by the fact that, after casting off the primary egg-shell (eh), the embryo undergoes further changes in its external form within the deutovum. In Trombidlum and Myobia this cuticular membrane is cast only after the rudiments of the limbs have appeared (Fig. 52). In Trombidium, this membrane is provided with appendages which surround the limbs like sheaths (Hen king), but this is not the case in Myobia. Here the limbs form in the usual way (Fig. 50 B), but when they have grown to a considerable length they become applied to the ventral surface of the body, and gradually become flattened to such a degree as hardly to project from the surface of the body. The whole embryo is once more oval and apparently devoid of appendages (Fig. 50 C). At this stage a cuticular membrane becomes detached from the embryo, bearing near its antero-dorsal extremity (in the nuchal region, according to Claparede) a tooth-like structure, composed of two thin chitinous processes closely applied to one another. This structure is not well depicted in Fig. 50 C and D, * [This is true of the forms described by Henking, but by no means holds good for all the Acarina, in the majority of which the chelicerae remain as perfectly distinct and movable organs. — Ed.] t [In the Phytoptidae the adult has only two pairs of legs. The larvae and nymphs do not always resemble the adults in other respects, for instance, in the Oribatidae, they differ essentially in external appearance, and the adult has a well-developed tracheal system which is entirely wanting in the larva. — Ed.] H 98 ARACHNIDA. /lect.cA where it appears more like a slit (at the left side of the inner envelope). Claparede thought that the tooth served for splitting the envelopes. It thus performs the same func- tion as the egg-tooth of the Araneae (p. 58), but it need hardly be pointed out that the difference in position of the two structures makes it im- possible to homologise them. We might rather compare the structure just described with the egg-tooth of theOpiliones (p. 33). The embryo, sur- rounded by the cuticular membrane, emerges from the egg-shell (Fig 50 D), which, however, con- tinues to surround the greater part of it. This recalls the cuticular membrane in the Araneae, which forms under the egg-shell and encloses the hatched and still motionless embryo. The limbs now grow out again, but are reduced as before, and a second cuticular integument is cast off, so that the greater part of the egg is enclosed by two integuments as well as by the egg-shell. The tritovum of Claparede is thus formed. Within it the embryo attains the six-limbed form in which it finally & ,-/'4, limbs of the nymph (partly broken off) ; r, proboscis (chelicerae and pedi- palps) of the nymph ; zh, intermediate integu- ment. DEVIATIONS IN THE METAMORPHOSIS. 107 number of statements (not indeed always reliable) made as to post-embryonic development, but also because of the number of variations occurring. "We must therefore refer to the literature already quoted for further particulars, and restrict ourselves to the description of a few ontogenetic peculiarities. The formation of the deutovum-membrane in the Acarid egg is apparently very common, and yet it seems to be indisputable that in some forms it does not take place. Claparede, who has given special attention to this point, states that the six-limbed larvae of Tetranychus hatch direct out of the egg-shell, without previously being surrounded by a special chitinous envelope. Tbe appearance of a six-limbed larva also is not universal, although it occurs in most families.* In Phytopta, for example, the larvae are four-limbed, i.e., provided with only two pairs of limbs, and some have been disposed to regard this as a primitive condition. But since, according to Nalepa (Nos. 100 and 101), the adult Phytopta also has only four limbs, this must be considered as a secondary condition both in the larva and in the imago. The great preponderance of the abdomen in the Phytopta and the consequent length of the body must also be regarded as a specialised condition. It is interesting, in this connection, to institute a comparison with the Demodicidae, which also have long abdomens. The six-limbed larva is found in its development and, according to Czokor (No. 78), passes through a course essentially agreeing with that described above. Taking into account the transformation of the six-limbed larva into the eight- limbed nymph, the occurrence of a four-limbed larva has been thought possibly to denote the more primitive character of the four-limbed form, from which the six-limbed form was to be derived. But we have already shown that such a conclusion is unwarrantable. Some light is thrown on the occurrence of the six-limbed larva by "Winkler's observations of Gamasus crassipes. Although the larva of this form has six limbs, four pairs were distinctly developed in the younger embryo (Fig. 57 A and B). Winkler's account is so clear, that all doubt appears to be excluded. t We must assume that one pair degenerates during a moult that takes place within the egg (formation of the deutovum). Shortly before the embryo hatches, when the limbs are already provided with the characteristic setae, there are only three pairs (Fig. 57 C). This statement, which we are hardly justified in doubting, is a strong argument in favour of the secondary origin of the six-limbed larva. The eight-limbed embryos of Gamasus crassipes observed by Winkler appear to be in a lower developmental stage than the six-limbed embryos (Fig. 57 A-C). We therefore assume that, in this form, a stage like that found in Pteroptus vcspcrtilionis is left out, this Acarid having an abbreviated course of development. The embryo of Pteroptus commences free life with eight limbs, i.e., at the nymph stage. It could, however, be shown that the embryo passes through a six-limbed stage in the egg while the latter is still within the mother (Nitzsch). X Limncsw pardina also leaves the egg as a nymph (Neumann). The young of the Phytopta, when they hatch, are very like the sexually mature adult, having * The six-limbed larvae have been observed in the Tetranychidae, Hydrach- nidae, Haiaearidae, Oribatidae, Trombidiidae, Gamasidae, Ixodidae, Tyro- glyphidae, Dermaleichidae, Sarcoptidae, Demodicidae, etc. t [This has since been confirmed by Wagner in Ixodes. — Ed.] J [This observation has not been confirmed, and appears very doubtful ; but cases probably exist in which the whole hexapod stage is passed through in the egg. — Ed.] 108 ARACHNIDA. only two pairs of limbs, and fully-developed mouth-parts. They differ from the adult chiefly in the absence of the external genitalia. These are developed in the course of two moults, and reproduction can now take place (Nalepa, No. 100). The development of Spluierogyna ventricosa appears still more abbreviated. This Acarid, the female of which is distinguished by the greatly swollen abdomen, is ovo-viviparous. The egg, after being laid, yields the sexually mature male and female, and copulation takes place soon after birth (Laboulbkne and Megnin). The course of development may be lengthened by the occurrence of a second nymph-stage following that which proceeds from the larva, and more or less resembling it in form. This is found in Halacarus spinifcr (Lohmann, No. 92), -/ie<£ Fig 57.— Embryos of Gamasus erassipes after the removal of the external egg-envelope, at various stages (after Winkler), abd, abdomen ; ch, chelicerae ; d, yolk ; eh, the cuticular embryonic integument ; kl, cephalic lobe ; ped, pedipalps ; pl-pi, limbs ; si, caudal lobe. and in various Gamasiclae (Reamer, No. 90, "Winkler, No. 106), but it ought to be more definitely ascertained whether these nymphs do not correspond to the pupal stage in other Acarina. It appears, further, that the nymph may be capable of reproduction before it attains the form of the sexually mature animal (Canestrini). This point was established for the Gamasiclae. Berlese distinguishes in this family several ontogenetic series which he describes as normal, and in which the larva, the nymph, and the imago succeed one another in the usual manner, and others which are abnormal, and in which earlier stages, i.e., nymphs, are already capable of reproducing themselves parthcnogcnetically . Such forms do not seem to attain to the complete form of the sexual animal. GENERAL CONSIDERATIONS. 109 It is said that several forms capable of reproduction may occur in this way in one and the same species ; Gamasus tardus, for instance, has no less than five such different forms, each of which might be taken for a different species (Berlese).* These are evidently very complicated conditions, which are far from being sufficiently understood. There is no doubt that early stages of development have repeatedly been regarded as different species, as is now definitely proved in the case of the well-known genus Hypopus (Megnin, Nos. 94 and 95, Michael, Nos. 98 and 99). The members of this genus are minute creatures with a smooth chitinous shell, convex on the dorsal side and flattened on the A-entral side, covering the whole of the body. Acarids with this characteristic appearance are often found on larval and on adult Insects, Myriopoda, etc., and were long regarded as adults. A closer study of the course of their development, however, proved that they merely represent early onto- genetic stages of Tyroglyphus and related genera, which, as a result of hitherto unknown circumstances, have deviated from the usual form of the nymph. These variations only affect isolated individuals, and it has been attempted to trace them back to unfavourable external conditions, which brought about such a modification of the inner organisation (Megnin). This explanation of the origin of the heteromorphic {Hypopus) forms has been disproved by Michael. General Considerations. Attempts have been made to separate the Acarina from the Arachnida, and to give this group the same value as the larger divisions of the Arthropoda (Arachnida, Myriopoda, Hexapoda, Haller, No. 83, A. C. Oudemans, No. 11). The grounds given for this classification appear to us too insufficient to deserve further discussion (p. 100). It rather appears to us that in the organisation and development of the Acarina there is sufficient resemblance to the Arachnida to justify their being classed among these latter, in accordance with the view until now commonly held. The Acarina represent a group of the Arachnida with highly specialised develop- ment, and are thus strongly differentiated in individual points of organisation from other Arachnids. Even the course of development has been influenced, and shows peculiarities which do not occur in other Arachnids. The chief of these are the different consecutive larval and pupal stages, and the free larval form provided with only six limbs. This latter must be considered as a secondary peculiarity. The best proof of this would be afforded by the appearance of a fourth pair of limbs in embryonic stages, which precede the six- limbed larva, if the statements made on this subject by "Winkler (No. 106, cf. p. 107) should be confirmed.! * [According to Michael both these observations are erroneous. — Ed.] t [This has been done for Ixodes by Wagner. — Ed.] 110 ARACHNIDA. VIII. General Considerations regarding the Arachnida. In studying the Arachnida, the point of greatest importance and interest consists in their relationship to those divisions of the Arthropoda classed with them as Tracheata, i.e., the Myriopoda and the Insecta. The Myriopoda, on account of their usually long form of body and the slight differentiation of the different parts of the body, demand less attention in this respect than the Hexapoda, in which the very marked division of the body into three regions calls for comparison with the segmentation of the Arachnida. In such a comparison, however, a serious difficulty at once arises in the different number of segments, and especially of limbs, found in the two groups.* The fusion of segments which often takes place among the Arachnida is of less consequence, since this may also occur to a greater or lesser degree among the Insecta. The fusing of the head and the thorax to form the cephalo-thorax must nevertheless be emphasised as an important Arachnidan character. The Insects, as is well known, carry on the head a pair of antennae, a pair of mandibles, and two pairs of maxillae, which, on account of their structure and ontogeny, are justly regarded as limbs ; further, there are three pairs of limbs on the thorax. The Arachnida have only two pairs of cephalic limbs (the chelicerae and the pedipalps), but four pairs of legs on the thorax. The attempts which have been made to harmonise these differences are too numerous to be treated here in detail. According to what may be described as the prevailing view, there is no homologue in the Arachnida for the antennae of the Insecta, but the chelicerae may be homologisecl with the mandibles, the pedipalps with the first maxillae, and the four ambulatory limbs with the second maxillae and the limbs that follow them. The chelicerae have, however, by some been considered to correspond to the antennae. "We are not disposed to accept either of these views, but, for reasons to be given later, compare the chelicerae to the second antennae of the Crustacea, for which a homologue is wanting in the Insecta. The first antennae of the Crustacea, which correspond to the antennae of the Insecta, are not present in the Arachnida. The pedipalps can at once be homologised with the mandibles of the Insecta (and Crustacea), the four pairs of ambulatory limbs with the two pairs of maxillae and the legs of the Insecta, but in this case one pair of thoracic extremities is wanting in the Arachnida. This, however, * [On this whole discussion compare Editorial note, p. 117.] GENERAL CONSIDERATIONS REGARDING THE ARACHNIDA. Ill does not appear to us important, since we attach no great value to this comparison of the Arachnida with the Insecta, and seek for the relationships of the former not so much in the domain of the "Tracheata" as among the branchiate forms, viz., the Xiphosura, as Ray Lankester and others have also done. We are, therefore, inclined to acrree with those zoologists who consider the Arachnida and the other air-breathing Arthropoda as two distinct series, and also assume a separate origin for the tracheae in these two divisions. The agreement existing between the organisation of the Arachnida and that of the Xiphosura compels us to adopt this view. We have already pointed out the agreement in outer structure between the Scorpiones and Limidus (Vol. ii., p. 357), especially in the numbers of the segments and limbs. In Limidus, as in the Arachnida, we find six pairs of limbs on the cephalo-thorax, so that a homology is suggested. We have just compared the first pair of limbs, the chelicerae, to the second antennae of the Crustacea, chiefly because the ganglia of these limbs, which arise post-orally, become united with the supra-oesophageal ganglion, as is the case also with the second antennae in the Crustacea (Vol ii., p. 164), and this process gains in significance when it is found repeated in the maxillary ganglia of Peripahis (p. 193). Xo such process is to be found in the Insecta, and we conclude that the limb in question is wanting in them. "We must not neglect to record the fact that, in the Opiliones and the Acarina, the chelicerae are said to be innervated from the thoracic ganglionic mass (Leydig, No. 40, b, and Winkler, No. 106). A final elucidation of this point is very desirable. The presence in the Araneae of another pair of cephalic limbs besides the two already mentioned has repeatedly been maintained. Two prominences are said to appear in front of the rudiments of the chelicerae and again to disappear (Croneberg, Jaworowski). It was assumed that these conjectural limbs became united with the rostrum (Croneberg, Lendl*), which, according to other observers, was found to have a paired rudiment (Schimkewitsch). There was a general tendency to seek in the rostrum the rudiment of one, or, indeed, perhaps of several pairs of limbs, and it was thought that this could even be proved in the adult animal (Scorpiones, Solifugae, * According to Lenbl, the vestigial limbs lie between the chelicerae and the pedipalps, and correspond to the mandibles of the Insecta, while the chelicerae, by their position and their manner of moving, show themselves to be true antennae. The shifting forward of the pedipalps pressed the conjectural mandibles against the rudiment of the upper lip, so as to fuse with it. 112 ARACHNIDA. Acarina — Croneberg). It should be noted that, according to Schimkewitsch, the so-called lower lip also arose from a similar paired rudiment, but in this case a pair of limbs seems out of the question. If such a vestigial pair of cephalic limbs is really present, it must be regarded (Croneberg, Jaworowski) as the missing antennae, and would be homologous with the first antennae of the Crustacea. This would necessitate no essential modification of our view. The first antennae, which were present in the ancestors, would still occur in the Araneae as vestiges, the chelicerae, however, corresponding to the second antennae. The pedipalps were compared by us with the mandibles of the Insecta. Each is composed of a masticatory ridge and a many- jointed palp. In the embryonic rudiment, however, both parts are said to consist of a number of joints ; if so, this limb would show a very primitive character, and a certain agreement with the biramose extremities of the Crustacea (Jaworowski). Indications of this biramose character are said to be found in the rudiments of other limbs also (Jaworowski). Further, whichever pair of limbs (chelicerae or pedipalps) is compared with the mandibles of the Insecta, the many-jointed character of the Arachnid limb affords a significant contrast to the Insectan mandible, which always consists of a single joint. Another primitive character is found in the presence of masticatory blades on the third and fourth limbs (in the Scorpiones and Opiliones), these extremities being thus partly utilised as mouth-parts, like the thoracic limbs of Limuhis which surround the mouth. The presence of pincers on the anterior limbs might also well be regarded as primitive, since such pincers are found in Limulus. We do not, however, lay any great stress upon this, as similar structures may arise independently of each other. The condition of the cephalo-thorax and its appendages in the various divisions of the Arachnida shows far more agreement with those of the Insecta than is found in the next section of the body — the abdomen — even if Ave overlook the reduced conditions which are exhibited here in the Acarina. We must here mention that the Solifugae, owing to the fact that the three posterior cephalo-thoracic segments are free, while the anterior region becomes swollen in a manner sucrsjestive of a head, have a certain resemblance to an insect. DO ' In addition to this, the abdomen shows the same number of segments as in the Insecta, and a pair of stigmata appears on the first GENERAL CONSIDERATIONS REGARDING THE ARACHNIDA. 113 "thoracic" or fourth cephalo-thoracic segment. These peculiarities have led to the Solifugae, which breathe through tracheae, being brought into relation with the Insecta ; but we have already shown (p. 36) that Ave cannot regard these characters of the Solifugae as primitive, nor consider the Solifugae themselves as intermediate forms between the Arachnida and the Insecta. In judging of the relationships of the Solifugae, it is important to note that in them also the chelicerae are innervated from the brain (Weissenborn, No. 16), and are thus proved to be homologous with the chelicerae of other Arachnids. An attempt to compare them with the antennae of the Insecta in order to explain their innervation will hardly be made, their whole development being opposed to this. In making a comparison with the Insecta, we should conclude rather that the antennae, which are to be regarded as a pair of cephalic limbs, are here wanting. The abdomen of the Arachnida is characterised chiefly by the great reduction of its segmentation, except in some divisions however, where the segments are very distinct. In the Scorpiones, the posterior part of the body is divided into a pre-abdomen and a post-abdomen, and is of great length. It might, indeed, be con- sidered as doubtful whether the lengthening of this part were not secondary, but for the fact that other Arachnida, during embryonic life, have sometimes this same number of segments, and also show indications of the division into pre- and post-abdomen (Araneae, pp. 50 and 57). In the fossil Xiphosura (Hemiaspis, Belinums), as well as in the Gigantostraca, the number of abdominal segments is larger than in Limulus, this makes it very probable that the abdomen of the latter has arisen through the fusion of a number of post-abdominal segments, and is thus homologous with the post-abdomen of the Scorpiones (Vol. ii., p. 358). The latter thus show, in the retention of the richly-segmented abdomen and in their segmentation generally, a very primitive character. It has been conjectured that the length and mobility of the abdomen are connected Avith the poison-sting which arms its extremity, and which is thus the more easily brought into use (Weissenborn). Great concentration of the organs is evident in the Arachnida, and the further forms are removed from those which we may rightly consider as the most primitive, the greater is the degeneration found in them, this degeneration reaching its highest degree in the Acarina. The derived forms of the Arachnida are thus simpler in their i 114 ARACHNIDA. organisation than the primitive forms, especially as certain systems of organs (circulatory and respiratory systems) may partly, or wholly, degenerate. The abdominal limb-rudiments are of peculiar importance in the comparison of the Arachnida with other Arthropods. Their number in the Scorpiones, as in Limulus, is six. [? cf. Brauer, Kishinodyb.] It is possible that in the Araneae, also, the same number of abdominal appendages was originally present (p. 51). The Arach- nida, like the Insecta, were derived from forms provided with a larger number of limbs. The first pair [second, pp. 10, 25, 57], is related to the genital aperture, while the following pairs show on their posterior surface the invaginations which give rise to the lungs. The lungs of the Arachnida may therefore be homologised with some probability with the gills of the Xiphosura (Vol. ii., p. 358, and Vol. iii., p. 77). This implies an origin for the Arachnidan tubular tracheae different from that in other "Tracheata" (Peripatus, Myrio- poda, Insecta), for there can be no doubt that the tracheae in the Arachnida are in the closest connection with the lungs.* Although the tracheae in a few Arachnids, e.g., the Solifugae, the Opiliones, and some Pseudoscorpiones and Acarina, seem to resemble each other greatly in structure, they must, in the one case, be derived from lungs or gills, and, in the other cases, from simple integumental depressions. Their later similarity of structure must be regarded as a phenomenon of convergence.! The presence of the stigmata in the abdomen only is in accord- ance with the view of the origin of the respiratory organs here adopted, but an exception occurs in the first pair of stigmata of the Solifugae which lies on the first "thoracic," or, rather, fourth cephalo-thoracic, segment. This must for the present be regarded as a secondary acquisition, and we may similarly try to explain the fact that, in the Acarina, stigmata occur in the cephalo-thorax at various points, often very far forward, in the cheliceral region. Similar displacements of the stigmata are also known to occur in Scolopendrella, where they also appear in the head in an unusual manner. * [See Simmons and Purcell (App. to Lit. on Araneae, Nos. VII., VIII.) and footnotes, p. 78.— En.] t [Tubular tracheae are not restricted to these four groups, but are also iouiirt in many Araneae associated with the lungs ; only the Scorpiones and the Pedi- palpi have lungs alone. This has led Bernard (App. to Lit. on Arachnida in gen No. III.) and Jaworowski (App. to Lit. on Araneae, No. II.) to the conclusion that the lung-books are not primitive structures giving rise to the trachea, but rather that both the lung-books and trachea are to be derived from simple sac-tracheae. — Ed.] GENERAL CONSIDERATIONS REGARDING THE ARACHNID A. 115 There are various other points of organisation in which the Arach- nida are removed from the Insecta, but approach the Xiphosura, and perhaps even the Crustacea. In dealing with the eyes, we tried to show that they cannot be classed together with those of the Insecta and the Myriopoda, but have had a different course of development (p. 68). They may, however, well be homologised with the median and lateral eyes of Limulus. In the origin of the Arachnid eyes, inversion plays an important part. Inversion has recently, also, been introduced by Claus as an explanation of the origin of the median eye of the -Crustacea (Xo. 57), and it appears not impossible that a closer connection may be found later between these processes. Further agreement between the Arachnida and the Xiphosura is found in the presence of an endoskeleton, which in the Scorpiones and Limulus is very similar in structure.* Another point which appears to us to be very characteristic, and which also fully applies to the Solifugae, in spite of their apparent deviation from the other Arachnids, is the presence of a large digestive gland (liver), such as does not occur in the Insecta, but is found in Limulus and the Crustacea. Another still more important point of agreement is yielded by the enteron and its appendages, if we grant that the testimony of ontogeny is reliable, viz., the origin of the so-called Malpighian vessels out of the entoderm. If this is the case, it would form an important reason for separating the Arachnida from the Insecta. Tubular appendages occur in the Crustacea at the posterior end of the metenteron ; the Malpighian vessels of the Myriopoda and the Insecta are, however, of ectodermal origin. Another point of resemblance between Limulus and the Arachnida is afforded by the presence of an artery running, in the Scorpiones, above the chain of ganglia, and forming a backward continuation of the oesophageal vascular ring (supra-neural vessel, supra-spinal -artery) ; a condition similar to this is met with in the ontogeny of Limulus. A sub-neural artery, indeed, occurs in the Crustacea, and a supra-neural vessel is also found in the Myriopoda (a fact * [There is considerable disagreement regarding the homology of the endo- sternite. Bernard (App. to Lit. on Scorpiones, Xo. I.), who has made a comparative study of the Arachnidan endosternite, comes to the conclusion that the endosternite of Limulus cannot be homologous with that of the Arachnids, the latter being part of an epidermal endophragmal system, while that of Limulus is mesodermal. On the other hand, Schimkewitsch (App. to Lit. on Scorpiones, No. YI.) maintains that the structure generally termed the endosternite in the Arachnida and Limulus is always mesodermal, and co-exists with, but is independent of, the series of ectodermal apodemes which -are so conspicuous in Galeodes. — Ed.] 116 ARACHNIDA. which makes this point of resemblance appear of less importance), so that this feature may perhaps be inherited from a common ancestral form. A less important agreement with the genital glands of Limulvs is afforded by the corresponding tubular network of genital glands in the Scorpiones. The coxal glands of the Arachnida, derived from the mesoderm, may, according to our present knowledge, be assumed with consider- able certainty to be nephridia, and are comparable with the organs which, in Limulus, occupy a corresponding position. These glands cannot be fully homologised with the antennal and shell-glands of the Crustacea, since these latter differ somewhat from them in position, i.e., belong to other segments. The nephridia that were present in every segment in the ancestral form have undergone great reduction, and the remnants are retained by their descendants in different segments, a feature probably connected with the varying form of the adult body in the different groups. We need hardly point out that the possession of coxal glands (especially strongly developed in youth) is a further distinction between the Arachnida and the Insecta, the latter not possessing any glands which in their development and position could be compared with the nephridia of the ancestral form. The Arachnid coxal glands arise from the mesoderm, the condition of which during embryonic development is a point of special im- portance. While, in the Insecta, the primitive segments are early subjected to change, in the Arachnida, they grow forward dorsally, and only undergo disintegration at a time when the dorsal heart is formed from them. The coelom, which disappears very early in the Insecta, is long retained in the Arachnida. This, which in itself is a primitive condition, further determines a greater simplicity in the rudiment of the heart, perhaps also in that of the coxal glands (nephridia), and probably also of the genital glands. The conditions thus produced recall those in the Annelida more than those in the remaining Arthropoda. It appears open to cuiestion whether much stress should be laid on the agree- ment existing between the cleavage, and the formation of the germinal layers and of the first rudiments of the organs in the Arachnida and the processes described for the Crustacea, or whether these should be explained by a certain similarity prevailing in these processes throughout the Arthropoda. This has already been pointed out in individual cases. It must remain equally doubtful whether the youngest stage of the germ-band in the Scorpiones, which has been compared with a certain ontogenetic stage in the Trilobites (p. 6), is of special importance in this connection. It can hardly be doubted, from all that has GENERAL CONSIDERATIONS REGARDING THE ARACHNIDA. 117 been stated above, that there is a close relationship between the Arachnida and Limulus, and, consequently, points of agreement with the Trilobites might be expected. It is, in this connection, a striking fact that the Scorpiones are of such great age, and that the forms now extant are not very unlike those found in the Silurian strata (PalacopJwnus uuncius, Xo. 15). In conclusion, we must again emphasise the fact that the apparent agreement of the Arachnida with the other Tracheata must he regarded as nothing more than a similarity determined hy their common Arthropodan nature and by a like development as the result of a similar manner of life. We must not assume a nearer connection between these divisions of the Arthropodan stock. We believe, rather, that the Arachnida, together with the Palaeostraca, proceeded from a common ancestral form, and subsequently diverged from one another, while the other Tracheata belong to a distinct stock, the two, however, being connected very far back. The Arachnida, according to our view of them, form a very uniform group. The most primitive forms are those in which the body is distinctly segmented, i.e., the Scorpiones and the Pedipalpi.* The Opiliones and the Pseudoscorpiones are affected by a reduction which goes still further in the Araneae, and reaches its highest degree in the Acarina, in which this far-reaching adaptation is accompanied by essential modifications in development.! Such modifications are also found in the Pseudoscorpiones, probably as the result of similar causes. [In addition to the editorial footnotes inserted here and there referring to Bernard's Arachnidan work, it is necessary to call separate attention to it in some detail, inasmuch as it has a profound bearing upon the question as to whether the Arachnids could be deduced from a Limuloid ancestral form. Arguing that the only scientific method of arriving at the ancestral form of the Arachnida is to compare all the known forms, and to sift out what are obviously the more primitive structural adaptations from the more specialised, this author arrives at the conclusion that the Solifugae come nearest the ancestral form in their segmentation, and in the simplicity of their endosternites. This endo- sternite has no resemblance whatever to the endosternite of Limulus, to which he would assign an entirely different origin (App. to Lit. on Arachnida in gen., No. I., and App. to Lit. on Scorpiones, Nos. I., VI.). He endeavours to show that the typical form of the Arachnidan body is an adaptation to the special manner of feeding. The Arachnids suck the blood of their victims, and, by a force-pump action of the oesophagus, distend the alimentary canal in a manner which would seriously interfere with the rest of the organisation. Their whole inner anatomy, he believes, can be shown to be simply so many adaptations to this serious * [According to Bernard, the Solifugae are in this respect the most primi- tive.— Ed.] t [This statement is a little misleading, for, in the adult Opiliones, only six segments are visible in the abdomen, while, in the Pseudoscorpiones, there are ten to eleven ; further, although the abdominal somites are fused in most adult Araneae (not in Liphistius), yet, in the young, eight to nine segments can be recognised ; these are not lost, but fused together, and, even in the Acarina, one form (Ixodes) exhibits marked segmentation ("Wagner). — Ed.] 118 ARACHNIDA. distention of the intestinal tract — adaptation, that is, of some much less specialised type than Limulus. All the chief organs are dealt with in detail, and, whether the author's conclusions are all of them ultimately confirmed or not, he has succeeded in placing on a new level, not only the controversy regarding the Arachuidan origin, hut also (by his association of physiology with morphology) the science of the whole group. So far his views have not met with much acceptance, and the Scorpiones are still generally regarded as the most primitive Arachnids finding their nearest allies in the Merostomata. — Ed.} LITEKATURE. Arachnida in General. 1. Croneberg, A. Ueber die Mundtheile der Arachniden. Archiv. f. Naturgesch. Jahrg. xlvi. 1880. 2. Eisig, H. Die Capitelliden des Golfs von Neapel. Mono- graphic der Fauna und Flora von Neapel. Berlin, 1887. (On the coxal and spinning glands of the Arachnida.) 3. Fernald, H. T. The Kelationships of Arthropods. Studies Biol. Lab. Johns Hopkins University. Baltimore. Vol. iv. 1890. 4a. Grenacher, H. Untersuchungen iiber das Sehorgan der Arthropoden. Gbttingen, 1879. 4b. Haase, E. Beitrage zur Kenntniss der fossilen Arachniden. Zeitschr. Deutscli. Geologisch. Gesellsch. Jahrg. 1890. 5. Jaworowski, A. Ueber die Extremitaten bei den Embryonen der Arachnoiden und Insecten. Zool. Anz. Bd. xiv. 1891. 6. Lankester, E. Eay. Limulus an Arachnid. Quart. Journ. Micro. Sci. Vol. xxi. 1881. 7. Lankester, E. Bay. On the sceleto-trophic tissues and the coxal glands of Limulus, Scorpio, and Mygale. Quart. Journ. Micro. Sci. Vol. xxiv. 1884. 8. Letjckart, E. Ueber den Bau und die Bedeutung der sog. Lungen bei den Arachniden. Zeitschr. f. Wiss. Zool. Bd. i. 1849. 9. Loman, J. C. C. Altes und Neues iiber das Nephridium (die Coxaldriise) der Arachniden. Bijdragen tot de Dierlcunde. Aflev. xiv. Amsterdam, 1887. 10. MacLeod, J. Becherehes sur la structure et la signification de l'appareil respiratoire des Arachnides. Archiv. Biol. Tom. v. 1884. 11. Oudemans, A. C. Die gegenseitige Venvandtschaft, Abstam- mung und Classification der sog. Arthropoden. Tijdschrift der Nederlandsche Dierlaindige Vereenigung. (2). Deel i. 1885-87. LITERATURE. 119 12a. Saint-Eemy, G. Contribution a l'etude du cerveau chez les Arthropodes tracheites. Archiv. Zool. Exper. (2). Tom. v. Suppl. 1887-90. 12b. Schimkewitsch, \V. Les Arachnoides et leurs amnites. Archiv. Slav. Biol. Tom. i. Paris, 1886. 13. Scudder, S. H. Bearbeitung der Arachniden. Zittel's Hand- bueh der Palaeontologie. Munchen and Leipzig, 1885. 14. Sturany, R. Die Coxaldriisen der Arachnoiden. Arb. Zool. Institut Univ. Wien. Bd. ix. 2. 1891. 15. Thorell, T., and Lindstrom, G. On a Silurian Scorpion from Gotland. K. Svensha Vetenslcaps Akad. Handlingar. Bd. xxi. 1885. Ann. Mag. Nat. History (5). Vol. xv. 1885. 16. "Weissenborn, B. Beitrage zur Phylogenie der Arachniden. Jen. Zeitschr. f. Naturw. Bd. xx. 1885. APPENDIX TO LITERATURE ON ARACHNIDA IN GENERAL. I. Bernard, H. M. Comparative Morphology of the Galeodidae. Trans. Linn. Soc. 1896. II. Bernard, H. M. The Apodidae. London, 1892. III. Bernard, H. M. An endeavour to show that the tracheae of the Arthropoda arose from setiparous sacs. Zool. Jahrb., Bd. v., and Ann. Mag. Nat. Hist. (6). Vol. xi. 1893. IV. Gaubert, P. Recherches sur les organes des senses et sur les systenies tegumentaire, glandulaire et musculaire des appendices des Arachnides. Ann. Sci. Nat. (7). Tom. xiii. V. Hansen, H. J. Orders and characters in different orders of Arachnids. Ent. Meddel. Kjobenhavn. Bd. iv. VI. Jaworowski, A. Homologie der Gliedmassen bei Arachniden und Insekten. Kosmos. Lemberg, 1891. VII. Pocock, Pi. I. On some points in the Morphology of the Arachnida (S. S.), with notes on the Classification of the Group. Ann. Mag. Nat. Hist. (6). Vol. xi. 1893. VIII. Wagner, J. Beitrage zur Phylogenie der Arachniden. Jen. Zeitschr. f. Naturw. Bd. xxix. 1895. IX. Are the Arthropoda a natural group 1 By ten authors. Nat. Sci. Vol. x. I. Scorpiones. 17. Blochmann, F. Ueber directe Kerntheilung in der Embryo- nalhiille der Skorpione. Morph. Jahrb. Bd. x 1885. 1 20 ARACHNIDA. 18. Ganin, M. S. On the Development of the Scorpion. (Kussian, without illustrations.) Supplement to the Protocol Univer. Kharkov. 1867. 19. Kowalevsky, A., and Schulgin, M. Zur Entwicklungsgeschichte des Skorpions (Androctonus ornatus). Biol. Centralbl. Bd. vi. 1886-87. 20. Lankester, E. Ray, and Bourne, A. G. The minute structure of the lateral and central eyes of Scorpio and Lirnulus. Quart. Journ. Micro. Set. Vol. xxiii. 1883. 21. Lankester, E. Eay. On the coxal glands of Scorpio, etc., and the brick-red glands of Lirnulus. Proc. Roy. Soc. London. Vol. xxxiv. 1882-83. p. 95. 22. Lankester, E. Eay. New hypothesis as to the relationship of the lung-book of Scorpio to the gill-book of Lirnulus. Quart. Journ. Micro. Sci. Vol. xxv. 1885. 23. Laurie, M. The Embryology of a Scorpion (Euscorpius italicus). Quart. Journ. Micro. Sci. Vol. xxxi. 1890. 24. Metschnikopf, E. Embryologie des Scorpions. Zeitschr. f. Wiss. Zool. Bd. xxi. 1871. 25. Muller, Joh. Beitriige zur Anatomie des Scorpions. Meckel's Arehiv. f. Anat. u. Phys. 1828. 26. Parker, G. H. The eyes in Scorpions. Bull. Mm. Comp. Zool. Harvard College. Vol. xiii. 1887. 27. Patten, W. On the origin of Vertebrates from Arachnids. Quart. Journ. Micro. Sci. Vol. xxxi. 1890. 28. Rathke, H. Zur Morphologie. Reisebemerkungen aus Taurien. Riga and Leipzig, 1837. APPENDIX TO LITERATURE ON SCORPIONES. I. Bernard, H. M. The Endosternite of Scorpio compared with that of other Arachnids. Ann. Mag. Nat. Hist. (6). Vol. xiii. 1894. II. Brauer, A. Beitriige zur Kenntniss der Entwicklungsgeschichte des Skorpions. Zeitschr. f. Wiss. Zool. Bd. lvii. and lix. III. Laurie, M. On the development of Scorpio fulvipes. Quart. Journ. Micro. Sci. Vol. xxxii. 1891 IV. Laurie, M. On the development of the Lung-books of Scorpio fulvipes. Zool. Am. 1892. V. Marchal, P. La glande coxale du Scorpion et ses rapports mor- phologiques avec les organes excreteures des Crustaces. Compt. rend. Tom. cxv. (Translated in Ann. Nat. Hist. (6). Vol. x.) LITERATURE. 121 "VT. Schimkewitsch, W. Ueber Bau und Entwicklung des Endo- sternites der Araclmiden. Zool. Jahrb. (Anat.). Bd. viii. 1894. II. Pedipalpi. 29. Bruce, A. T. Observations on the Nervous System of Insects and Spiders, and some preliminary observations on Phrynus. Johns Hopkins University Circulars. Baltimore. Vol. vi. 1886-87. No. 54, p. 47. 30. Grassi, B. Intorno ad un nuovo Aracnide Artrogastro etc. I. Progenitori dei Miriapodi e degli Insetti. Memoria V. Bull. Societa Entomol. Ital. Anno xviii. Firenze, 1886. 31. Tarnani, J. Die Genitalorgane der Telyphonus. Biol. Centralbl. Bd. ix. 1889-90. APPENDIX TO LITERATURE ON PEDIPALPI. I. Laurie, M. On the Morphology of the Pedipalpi. Journ. Linn. Soc. Vol. xxv. 1894. II. Pereyaslawzewa, S. (1) Les premiers stades du developpe- ment des Pedipalpi. (2) Les derniers stades du developpe- ment des Pedipalpi. Compt. rend. Tom. cxxv. 1897. III. Strubell, A. Zur Entwicklungsgeschichte der Pedipalpen. Zool. Anz. xv. (Translated in Ann. Nat. Hist. (6). Vol. x. 1892.) III. Palpigrada. 31a. Hansen, H. J., and Sorensen, W. The order Palpigrada Thor. and its relationship to the other Arachnida. Entomologisk Tidskrift. Aug. 18th, 1897. And Grassi, No. 30. IV. Pseudoscorpiones. 32. Barrois, J. Sur le developpement des Chelifers. Compt. rend. Acad. Sci. Paris. Tom. xcix. 1884. p. 1082. 33. Croneberg, A. Beitrag zur Kenntniss des Baues der Pseudo- scorpione. Bull. Soc. imp. Nat. de Moscou. Tom. ii. 1888. 34. Metschnikoff, E. Entwicklungsgeschichte des Chelifer. Zeitschr. f. Wiss. Zool. Bd. xxi. 1871. 35. Stecker, A. The development of the ova of Chthonius in the body of the mother, and the formation of the blastoderm. Ann. Mag. Nat. Hist. (4). Vol. xviii. 1876. p. 197. (Translated from Sitzungsber. bohm. Gesellsch. Wiss. Prag, 1876.) APPENDIX TO LITERATURE ON PSEUDOSCORPIONES. I. Barrois, J. Memoire sur le developpement de Chelifer. Rev. Suisse Zool. 1896. 122 ARACHNIDA. II. Bernard, H. M. Notes on the Chernetidae, with special reference to the vestigial stigmata, and to a new fcrm of Trachea. Journ. Linn. Soc. Zuol. Vol. xxiv. III. Bertkau, Ph. Zur Entwicklungsgeschichte der Pseudoscorpione. C. B. Ver. Rheinl, 1891. IV. Bouvier, E. L. Sur la ponte et le developpement d'un Pseudo- scorpionide, le Garypus saxicola. Bull. Soc. ent. France. 1896. V Vejdowsky, J. F. Sur la question de la segmentation de l'oeuf et la formation du blastoderme des Pseudoscorpiones. Congr. interned. Zool. ii. 1892. V. Opiliones. 36. Balbiani, E. G. Memoire sur le developpement des Phalangides. Ann. Sci. Nat. (5) Zool. Tom. xvi. 1872. 37. Faussek, V. Ueber die embryonale Entwicklung der Gesch- lechtsorgane bei der Afterspinne (Phalangium). Biol. Cen- tralis. Bd. viii. 1888-89. 38. Faussek, V. Zur Embryologie von Phalangium. Zool. Anz. Jahrg. xiv. 1891. 39. Faussek, V. On some ontogenetic stages of the Opiliones. Trud. ross. estestv. Obshchetsvo. St. Petersburg. Zool. Tom. xx., pp. 46-53. 40a. Henking, A. Untersuchungen iiber die Entwicklung der Phalangiden. Zeitschr. f. Wiss. Zool. Bd. xlv. 1887. 40b. Leydig, F. Ueber das Nervensystem der Afterspinnen (Phalan- gium). Arcliiv. f. Anat. u. Phys. 1862. 41. MacLeod, J. Sur l'existence d'une glande coxale chez les Phalangides. Bull. Acad. Roy. Sci. Belgique. (3). Tom. viii. 1884. APPENDIX TO LITERATURE ON OPILIONES. I. Bertkau, P. Die Entwicklung der Coxaldriise bei Phalangium. Zool. Anz. Bd. xv. 1892. II. Faussek, V. Zur Embryologie von Phalangium. Zool. Anz. Bd. xiv. III. Faussek, V. Zur Anatomie und Embryologie der Phalangiden. Biol. Centralblatt. Bd. xii. (Translated in Ann. Mag. Nat. Hist. (6). Vol. ix. 1892.) IV. Faussek, V. On the ontogeny and anatomy of the Phalangidae (Russian). Trud. ross. Obshchestvo estestv. St. Petersburg Zool. Tom. xxii. (No. III. is a resume of this larger work). LITERATURE. 123 V. Lebedinsky, J. Die Entwicklung der Coxal-driise bei Phalan- gium. Zool. Anz. Bd. xv. 1892. VI. Purcell, F. Uber den Bau der Phalangidenaugen. Zeitschr. f. Wiss. Zool. Bd. lv. (1). 1894. VI. Solifugae. 42. Birula, A. Einiges iiber den Mitteldarm der Galeodiden. Biol. Centralbl. Bd. xi. 1891-92. 43. Croneberg, A. Ueber ein Entwicklungsstadium von Galeodes. Zool. Anz. Jahrg. x. 1887. 44. MacLeod, J. Sur la presence d'une glande coxale chez les Galeodes. Bull. Acad. Roy. des Sci. Belgique (3). Tom. viii. 1884. 45. Lankester, E. Bat. Limulus an Arachnid. Quart. Journ. Micro. Sci. Vol. xxi. 1881. p. 644. APPENDIX TO LITERATURE ON SOLIFUGA. I. Bernard, H. M. Comparative Morphology of the Galeodidae. Trans. Linn. Soc. Zool. 1896. II. Birula, A. Beitrage zur Kenntniss der Anatomischen Baues Geschlechtorgane bei dem Galeodiden. Biol. Centralbl. Bd. xii. 1892. VII. Araneae. 46. Balbiani, E. G. Memoire snr le developpement des Araneides. Ann. Sci. Nat. Zool. (5). Tom. xviii. 1873. 47. Balfour, F. M. Notes on the development of the Araneina. Quart. Journ. Micro. Sci. Vol. xx. 1880. 48. Barrois, J. Recherches sur le .developpement des Araignees. Journ. Anat. et Phys. Paris, 1877. 49. Bertkau, Ph. Ueber die Respirationsorgane der Araneen. Archiv. f. Naturgescli. Jahrg. xxxviii. 1872. 50. Bertkau, Ph. Ueber den bau der Augen etc. bei den Spinnen. Verhandl. Naturhist. Ver. Rheinlande und Westfalen. Jahrg. xlii. 1885. p. 218. 51. Bertkau, Ph. Ueber den Verdauungsapparat der Spinnen. Archiv. f. mikr. Anat. Bd. xxiv. 1885. 52. Bertkau, Ph. Beitrage zur Kenntniss der Sinnesorgane der Spinnen. I. Die Augen. Archiv. f. mikr. Anat. Bd. xxvii. 1886. 53. Bruce, A. T. Observations on the Embryology of Spiders. Amer. Nat. Vol. xx. 1886. 124 ARACHNIDA. 54. Bruce, A. T. Observations on the Embryology of Insects and Arachnids. Baltimore, 1887. 55. Carriere, J. von. Kritische Besprechung der neueren Arbeiten iiber Bau und Entwicklung des Auges der zehnfiissigen Crustaceen und Arachnoiden. Biol. Centralbl. Bd. ix. 1889-90. 56. Claparede, E. Recherches sur revolution des Araignees. Naturkundige Verhandl. Provinciaal Utrecht sch Genootshap Kunst. Wiss. Deel i. Stuk i. Utrecht, 1862. 57 Claus, C. Ueber den feineren Bau des Medianauges der Crustaceen. Anz. d. Akad. Wiss. Wien. No. xii. 1891. 58. Emerton, H. Observations on the development of Pholcus. Proc. Boston Soc. Nat. Hist. Vol. xiv. 1870-71. 59. Herold, M. Untersuchungen iiber die Bildungsgeschichte der wirbellosen Thiere im Eie. I. Theil. Yon der Erzeuguns der Spinnen. Marburg, 1824. 60. Kennel, J. von. Die Ableitung der sog. einfachen Augen der Arthropoden, namlich der Stemmata der Insectenlarven, Spinnen, Scorpioniden etc. von den Augen der Anneliden. Sitzungsber, Naturf. Gesellsch. Dorpat. Bd. viii. 1888. 61. Kingsley, J. S. The Embryology of Spiders. Critical notice of ]STo. 72, Schimkewitsch. Amer. Nat. Vol. xxi. 1887. 62. Kishinouye, K. On the development of Araneina. Journ. Coll. Sci. Imp. University, Tokio. Japan. Vol. iv., Part i. 1891. (This journal was only accessible after this work had gone to press. The chief results recorded have, however, been referred to by us.) 63. Lendl, A. Ueber die morphologische Bedeutung der Glied- maassen bei den Spinnen. Mathem. naturw. Berichte aus Ungam. Budapest and Berlin. Bd. iv. 1886. 64. Locy, W. A. Observations on the development of Agalena naevia. Bull. Mies. Comp. Zool. Harvard College. Vol. xii. 1886. 65. Loman, J. C. Ueber die morpholog. Bedeutung der sog. Mal- pighi'schen Gefiisse der echten Spinnen. Tijdschrift der Nederlandsche Dierkundige Vereenigung. Ser. 2., Deel i. 1885-87. 66. Ludwig, H. Ueber die Bildung des Blastoderms bei den Spinnen. Zeitschr. f. Wiss. Zool. Bd. xxvi. 1876. 67. Mark, E. L. Simple eyes in Arthropods. Bull. Mus. Comp. Zool. Harvard Coll. Vol. xiii. 1887. LITERATURE. 125 68. Morin, J. Zur Entwicklungsgesch. der Spinnen. Biol. Centralbl. Bd. vi. 1886-87. 69. Morin, J. On the Development of the Araneae. Zapisk. novoross. Obsh. estestv. Odessa. Tom. xiii. 1888 (Russian). 70. Sabatier, A. Formation du blastoderme chez les Araneides. Compt. rend. Acad. Sri. Paris. Tom. xcii. 1881. {Ann. Mag. Nat. Hist. (5). Vol. vii. 1881.) 71. Salensky, W. On the Development of the Araneae. Zapisk. Kievsk. Obshch. estestv. Tom. ii., Pt. i. 1871. (Russian. Abstracted in Hofmann's and Schwalbe's Jahresb. Anat. Phi/s. Bd. ii. 1875. p. 323.) 72. Schimkewitsch, W. Etude sur le developpement des Araignees. Arcliiv. Biol. Tom. vi. 1887. 73. Watase, S. On the Morphology of Compound Eyes of Arthropods. Studies Biol. Lab. Johns Hopkins University. Baltimore. Vol. iv. 1890. APPENDIX TO LITERATURE ON ARANEAE. I. Damix, X. Ueber Parthenogenesis bei Spinnen. Verh. Ges. Wien. xliii. (Translated Ann. Mag. Nat. Hist. 1894.) II. Jaworowski, A. Die Entwicklung der sogenannten Lungen bei der Arachnoiden und speciell bei Trochosa singoriensis, nebst Anhang iiber die Crustaceenkiemen. Zeitschr. f. Wiss. Zool. Bd. lviii. III. Jaworowski, A. Ueber die Extremitaaten, deren Driisen und Kopfsegmentierung bei Trochosa singoriensis. Zool. Anz. Bd. xv. 1892. IV. Jaworowski, A. Die Entwicklung der Geschlechtsdriisen bei Trochosa singoriensis. Verh. Ges. deutsch. Naturf. 1895. V. Jaworowski, A. Die Entwicklung des Spinnenapparates bei Trochosa singoriensis, mit Beriicksichtigung der Abdominal- anhange und der Fliigel bei der Insekten. Jen. Zeitschr. f. Naturio. Bd. xxx. 1896. VI. Kishinouye, K. Note on the Coelomic cavity of Spiders. Joum. Sci. Coll. Japan. Vol. vi. 1894. VII. Purcell, F. Note on the Development of the Lungs, Enta- pophyses, Tracheae, and genital ducts in Spiders. Zool. Anz. Bd. xviii. VIII. Simmons, 0. L. Development of the Lungs of Spiders. Amer. Joum. Sci. (3). Vol. xlviii. Also in Tuft's Coll. Stud. No. 2. 12G ARACHXIDA. VIII. Acarina. The literature on the ontogeny of the Acarina is so extensive that only a limited number of the works can here be quoted. Fuller references to literature will be found in the works of Furstexberg (No. 80), Henkixg (No. 85), and Lohmanx (No. 92). 74. Bexeden, P. J. vax. Eecherches sur l'histoire naturelle et le developpement de l'Atax ypsilophora. Mem. Acad. Roy. Belgique. Tom. xxiv. 1850. 75. Berlese, M. A. Polymorphisme et Parthenogenese de quelques Acariens (Gamasides) Archiv. Ital. Biol. Tom. ii. Turin, 1882. 76. Caxestrixi, G. Osservazioni intorno al genere Gamasus. Atti del Real. Institute* Veneto d. Sc, Lett. etc. (5). Tom. vii. 1880-81. 77. Claparede, E. Studien an Acariden. Zeitschr. f. Wiss. Zool. Bd. xviii. 1868. 78. Czokor, J. Ueber Haarsackmilben und eine neue Varietat derselben bei Schweinen. Verli. k. k. zool. hot. Gesellsch. Wien. Bd. xxix. 1880. 79. Frauexfeld, G. vox. Zool. Miscellen. Khyncholophus oedi- podarum. Verh. k. k. zool. hot. Gesellsch. Wien. Bd. xviii. 1868. 80. Furstexberg, M. H. F. Die Kratzmilben des Menschen u. d. Thiere. Leipzig, 1861. 81. Gudden, R. Beitrage zu den durch Parasiten bedingten Haut- krankheiten. Archiv. f. physiol. Heilkunde. Stuttgart, 1855. 82. Haller, G. Zur Kenntniss der Tyroglyphen und Verwandten. Zeitschr. f. Wiss. Zool. Bd. xxxiv. 1880. 83. Haller, G. Die Mundtheile und systematische Stellung der Milbeu. Zool. Anz. Jahrg. iv. 1881. 84. Haller, G. Ueber den Bau der vogelbewohnenden Sarcoptiden (Dermaleichiden). Zeitschr. f. Wiss. Zool. Bd. xxxvi. 1882. 85. Hexking, H. Beitrage zur Anatomie, Entwicklungsgeschichte und Biologie von Trombidium fuliginosum. Zeitschr. f. Wiss. Zool. Bd. xxxvii. 1882. 86. Koexike, F. Zur Entwicklung der Hydrachniden. Zool. Anz. Jahrg. xii. 1889. 87. Kramer, P. Zur Naturgeschichte der Milben. Archiv. f. Naturgesch. Jahrg. xlii. 1876. LITERATURE. 127 88. Kramer, P. Ueber Dendroptus. Archie, f. Naturgesch. Jahrg. xlii. 1876. 89. Kramer, P. Ueber die Segmentirung b. d. Milben. Arckiv. f. Naturgesch. Jahrg. xlviii. 1882. 90. Kramer, P. Ueber Gamasiden. Archiv. f. Naturgesch. Jahrg. xlviii. 1882. 91. Laboulbene, A., et Megnin, P. Memoire sur la Sphaerogyna ventricosa. Journ. Anat. et Phys. Annee xxi. Paris, 1885. 92. Lohmann, H. Die Unterfamilie der Halacaridae Murr. u. die Meeresmilben der Ostsee. Zool. Jahrh. Abth. f. Syst. Bd. iv. 1889. 93. MacLeod, J. Communication preliminaire relative a l'anatomie des Acariens. Bull. Acad. Roy. Sci. Belgique. Ser. iii. Tom. vii. 1884. 94. Megnin, P. Memoire sur un nouvel Acarien de la famille des Sarcoptides, le Tyroglyphus rostro-serratus et sur son Hypopus. Journ. Anat. et Phys. Paris, 1873. 95. Megnin, P. Memoire sur le Hypopus. Journ. Anat. et Phys. Paris, 1874. 96. Megnin, P. Sur les Metamorphoses des Acariens de la famille des Sarcoptides et de celles de Gamasides. Compt, rend. Acad. Sci. Paris. Tom. lxxviii. 1874. 97. Michael, A. D. British Oribatidae. Vols. i. and ii. London. Ray Society, 1883. 98. Michael, A. D. The Hypopus question, or the life history of certain Aearina. Journ. Linn. Soc. Zool. Vol. xvii. 1884. 99. Michael, A. D. ^Researches into the life histories of Glyci- phagus domesticus and G. spinipes. Journ. Linn. Soc. Zool. Vol. xx. 1890. 100. Xalepa, A. Anatomie der Phytopten. Sitzungsber. k. Akad. voiss. Wien. Bd. xcv. 1887. 101. Nalepa, A. Beitrage zur Systematik der Phytopten. Sitz- ungsber. k. Akad. tciss. Wien. Bd. xcviii. 1889. 102. Xeuman, C. J. Sur le developpement des Hydrachnides. Entonx. Tijdsskrift. Stockholm. Tom. i. 1880. 103. Xitzsch, C. J. Ueber die Fortpflanzung des Pteroptus vesper- tilionis. Archiv. f. Naturgesch. Jahrg. iii. 1837. 104. Eobin, C, et Megnin, P. Memoire sur les Sarcoptides plumicoles. Journ. Anat. et Phys. Paris, 1S77. 128 ARACHNIDA. 105. Winkler, W. Das Herz der Acarinen nebst vergleichenden Bemerkungen uber das Herz der Phalangiden und Cherne- tiden. Arb. Zool. Inst. Wien. Bd. vii. 1888. 106. Winkler, W. Anatomie der Gamasiden. Arb. Zool. Inst. Wien. Bd. vii. 1888. APPENDIX TO LITERATURE ON ACARINA. I. Jourdain, S. Sur le developpement du Trombidium holoseri- ceum. Compt. rend. Acad. Sci. Paris. Tom. cxxv. 1897. II. Kramer, P. Ueber die Typen der postembryonalen Entwick- lung bei der Acariden. Archiv. f. Naturgesch. Bd. lvii. 1891. III. Kramer, P. Zur Entwicklungsgeschichte und Systematik der Susswassermilben. Zool. Anz. Bd. xv. 1892. IV. Lohmann, H. Die Halacarinen der Plankton-Expedition. Leipzig, 1893. V. Schaub, H. von. Ueber die Anatomie von Hydrodroma. Sitzungsber. k. Akad. loiss. Wien. Bd. xcvii. 1888. VI. Supino, F. Embriologia degli Acari. Atti. Soc. Veneto-Trent (2). Tom. ii. VII. Troussart, E. L. Sur l'existence de la parthenogenese chez les Sarcoptides plumicoles. A?i?i. Soc. ent. France. Tom. lxiii. 1894. And Bull. Soc. ent. Paris, 1894. VIII. Wagner, J. Zur Entwicklungsgeschichte der Milben, Fur- chung des Eies, Entstehung der Keimbliitter und Entwick- lung der Extremitiiten bei Ixodes. Zool. Anz. Bd. xv. 1892. IX. Wagner, J. Die Embryonalentwicklung von Ixodes cal- caratus. Trud. St. Peterb. Obshch. Tom. xxiv. CHAPTEE XXII. PENTASTOMIDAE. Our knowledge of the Pentastomidae still rests principally on Leuckart's observations, supplemented by a few smaller treatises, and recently confirmed and amplified by Stiles. 1. Embryonic Development. The eggs of Pentastomum are surrounded by two envelopes (Fig. 58 A and B, h). The early embryonic development takes place gradually as the ovum passes down the uterus. Cleavage is total (Leuckart, Macalister). The egg breaks up into a number of cells of about equal size, the further fate of which could not be ascer- tained. Macalister describes the formation of a blastoderm and a germ-band, but his statements are not conclusive. According to Leuckart, a germ-band is not formed. The embryo secretes a surface cuticle at an early stage, a disc-like thickening appearing on the dorsal surface of this cuticle. When the cuticle detaches itself from the embryo and forms a third envelope to the latter (Fig. 58 A and B, eh), it remains connected at this thickened disc (dorsal cone, rz). The chitinous integument also, which is now secreted as a covering for the embryo, is correspondingly thickened at this spot, and takes the form of a pit-like depression. The "dorsal cone," which at first connects these two chitinous thickenings, becomes constricted and broken through, but a trace of it is left attached to the embryo ; this, in P. taenioides, is shaped like a raised cross situated in a cup-shaped groove (Fig. 58 B and C, rk). The remainder of the" dorsal cone" is retained on the detached integu- ment as a circular thickening (the so-called facet, Fig. 58 B, /). This structure recalls the micropyle or the dorsal organ of the Crustacea, with which Leuckart has compared it. A certain external similarity in structure is found between the so-called primitive tracheae of the Acarina and the dorsal cone of the Pentastomum ; but these "tracheae" are paired and lie ventrally, so that there is no real agreement between them. K 130 PENTASTOMIDAE. The early shedding of a cuticular integument within the egg, which must be regarded as a moult, recalls the formation of the deutovum-membrane in the Acarina (p. 96) ; similar processes occur also in the Crustacea (Vol. ii., p. 118). Before the dorsal cone is broken through — i.e., before the cuticular envelope is completely detached from the embryo — two pairs of truncated appendages have developed on the ventral side. These are limbs on which claws soon appear. A narrower posterior portion — the so-called tail — has, previous to this, become marked off from the compact trunk (Fig. 58 A and B), to the ventral surface of which it is applied. This caudal appendage is characteristic of the embryos of a few species of Pentastomum. In P. taenioides it is c. Sw Pi p. Fig. 58. Embryos in the egg-integuments and free larva of Pentastomum taenioides (after Leuckart). (1st, stigma of gland ; eh, embryonic integument; /, "facet"; h, egg-integu- ments ; m, oral plate ; px and p2, truncated limbs ; rh, dorsal cross (dorsal organ) ; rz, dorsal cone ; s, caudal appendage. The boring apparatus of the embryo is not shown. somewhat large (Fig. 58 B and C), while in P. proboscideum it is merely a small bifid appendage (Fig. 59, s). The embryo of P. oxycephalum has no caudal appendage, but presents a round posterior extremity. In this form the embryo leaves the egg (Van Beneden, Schubart) ; it is therefore very unlike the parent in shape, and has to pass through radical transformations before attaining the adult form (Leuckart). The Larval Development. The further course of development is marked by the transference of the eggs into the intermediate host and the development of a THE LARVAL DEVELOPMENT. 131 four-limbed larva. The form whose ontogeny was examined by Leuckart, P. taenioides, inhabits, in its sexual condition, the nasal cavity of the dog. The eggs are laid in the nasal mucus, and with this they reach the exterior. For the further development of the embryo an intermediate host is necessary. This, in the case of P. taenioides, is a rabbit, which, by swallowing the eggs, introduces them into its stomach, where the egg-integuments become detached and the larva set free. In P. proboscideum also (Stiles), the early stages are similar to the above. The eggs of this form are found in the lungs of the boa constrictor ; from the lungs they pass into the intestine, where they are found in quantities in the fa?ces, with which they leave the body. They, too, must be swallowed by an intermediate host in order to develop further. Stiles was successful in introducing them into mice. The larva, which has a blunt anterior, and a pointed posterior end, i.e., which is sup- plied with a tail, has two pairs of truncated limbs, provided wuth chitinous claws fur- nished with a sup- porting apparatus (Fig. 58 C, and Fig. 59, st). The two claws are attached to a chitinous ring, and seem to be quite in- dependent of the supporting apparatus. This structure sug- gests that the limb consists of a terminal, and a basal seg- ment, the limb being thus regarded as two- jointed. Stiles, who adopts this view, thought the limb more distinctly marked off from the body than did Leuckart, who regarded it as consisting of one joint only. At the anterior end of the body lies a boring apparatus composed ■of several chitinous spines (Fig. 59, ba), which has been compared Fig. 59. — Quadrupedal larva of Pcntastomum proboscideum, from the ventral side (after Stiles), ba, boring apparatus ; dst, stigma of gland ; dz, gland - cells ; kr, claws ; to, mouth ; ma, stomach ; n, rudiment of the nervous system ; oes, oesophagus ; ^',-j'o, truncated limbs; ro, dorsal organ, seen through the transparent body ; s, caudal appendage ; st, apparatus for supporting the claws ; tp, sensory papillae. 132 PENTASTOMIDAE. with the mouth-parts of the Arthropoda, especially with those of the Acarina, hut such a comparison is hardly permissible on account of the position of this apparatus and its origin in front of the mouth ; it must probably be regarded as a larval organ (Stiles). Near the boring apparatus are two small papillae, which have been regarded as tactile organs (tp). The mouth, in P. proboscideum, lies somewhat far back, about on a level with tbe anterior truncated limb (Fig. 58, m). It is surrounded by a chitinous horsesboe-shaped band, and leads into a narrow oesophagus, which passes into a wider stomach. According to Stiles, there is no anus, although one is to be seen in Jacquart's not very accurate drawings. An accumulation of cells surrounding the oesophagus represents the rudiment of the nervous system (m). Stiles also found within the larva a large accumulation of richly granulated cells distributed in a definite manner, some of which are no doubt glandular cells. Two circular structures lying at the bases of the anterior extremities are regarded as the external apertures of glands (so-called stigmata of the glands, Figs. 58 and 59, dst). The Encysted Larva. The larva which has become free in the intestine of the intermediate host, by the help of the boring apparatus at the anterior pole of tbe body and the limbs, traverses the wall of the intestine and passes into the other organs, e.g.y the liver, where it becomes at- tached and en- closed in a fibrous cyst derived from the tissues of the host. It here passes through a number of moults, during which it throws off the limbs and the boring apparatus. The caudal appendage also disappears, and the larva assumes a compact cylindrical form. Leuckart found, seven weeks after ^ ,ma ces. U-ed. Fig. 60. — Encysted larva of Pentmt^mvm tacnioidcs from the viscera of a rabbit, nine weeks after infection (after Leuckart). a, anus ; ag, efferent duct of the genital gland; dst, glandular stigmata; ed, proctodaeum ; gd, genital gland ; m, mouth ; ma, stomach ; n, rudiment of the nervous system ; oe, genital aperture ; oes, oesophagus. THE LARVAL DEVELOPMENT. 133 infection, in the cysts of P. taenioides, besides the worm-, or, rather, maggot-shaped larva, two cast integuments, on which could be distinguished remains of the embryonic chitinous structures, viz., the dorsal cross and the chitinous oral horseshoe-shaped band, and probably also the remains of the truncated limbs. Several further moults then take place, a long time being occupied by this develop- ment ; five to six months, according to Leuckart, pass before the larva of P. taenioides attains its full development in the intermediate host. The development of P. 2^'oboscideum is somewhat more rapid, but also occupies several months (Stiles). While the larva remains in the cyst, and during the course of several moults, the most important change which takes place is the development of the internal organs; the external form, however, ov. s?iu^ ma- Fig. 61. — Encysted female larva of Pentastomum taenioides from the viscera of a rabbit, about four months after infection (after Leuckart). a, anus ; ed, proctodaeum ; Ih, larval integu- ment (detached cuticle) ; m, mouth ; ma,, stomach ; mu, retractor muscles of the pharynx ; n, nervous system ; od, oviduct : oe, genital aperture ; oes, oesophagus ; or, ovary ; tn, nerve running from the oesophageal ganglion to the tactile papillae ; vwj, vagina. also undergoes a few changes, to be described below. The internal organs of the free larva, as far as could be ascertained, seem to pass direct into those of the encysted larva and of the sexually mature animal. The intestinal canal, which was not extensive in the free larva, widens and becomes differentiated into its separate regions, pharynx, oesophagus, and stomach. The latter soon becomes very large (Fig. 60, ma). It ends blindly posteriorly, and only becomes connected later with the proctodaeum (ed). The accumulation of cells round the oesophagus present in the 134 PENTASTOMIDAE. free larva (Fig. 59, n) during later larval life develops into the sub-oesophageal mass and the oesophageal ring, which represent the central nervous system of the adult. The sub-oesophageal mass, in early larval life, is much larger than in the adult animal, and occupies a considerable part of the ventral surface (Leuckart, Figs. 60 and 61, n). The rudiments of the genital organs can be recognised early, but, according to Leuckart, it is at first impossible to distinguish the two sexes. A long, unpaired strand lying dorsally to the stomach, the germ-gland (Fig. 60, gd), forks anteriorly to form two strands (the rudiments of the efferent ducts, ag). These two strands oir. /7UI. <% ed Fig. 62. — Encysted female larva of Pentastomiim prohoscideum (the so-called P. subcylindrimm) from the viscera of a mouse, six and a half weeks after infection (after Stiles), a, anus ; ed, proctodaeum ; lit, hook-sac ; Ih, larval integument (detached cuticle) ; m, mouth ; ma, stomach ; n, rudiment of the nervous system ; od, oviduct ; oe, genital aperture ; ol, upper lip ; ov, ovary ; rs, receptaculum seminis ; vag, vagina. embrace the anterior part of the stomach, and, after reuniting ventrally, open externally in the region of the ganglionic mass (a). There is very little difference in this respect in the male ; the genital aperture in the adult male retains its primitive position in the anterior part of the body, not far behind the mouth. The genital aperture in the adult female is, however, found at the posterior end of the body, quite near the anus (Fig. 62, oe) ; and Leuckart assumes that it has been thus displaced on account of greater growth of the part between it and the mouth taking place simultaneously Avith arrest of growth in the posterior region. Fig. 61 represents THE LARVAL DEVELOPMENT. 135 a transitionary stage, in which the genital aperture is already shifted further back than in Fi" 60. The differentiation of a vagina from the primitive genital duct has here already taken place. In Fig. 62, the genital aperture has already assumed its final position near the anus. Stiles speaks of a differentiation of the sexes at an early stage ; but the stages described by him in P. proboscideum seem to us to be somewhat more advanced than those observed by Leuckaet in P. tacnioides. According to Hoyle, it appears that the genital glands may originally have been paired. If this were the case, we should have, in the fusion of the germ- glands to form a single organ, a process similar to that in the Acarina (p. 101). The position of the (female) genital aperture at the posterior end of the body, which is in opposition to what is usual in the Arachnida, might, according to Leuckart's explanation, be regarded as secondary. The body of the encysted larva after the first moults looks quite smooth, but later a series of rings make their appearance (Fig. 62). These first arise in the middle of the body, and spread anteriorly and posteriorly. These superficial markings cannot be regarded as equivalent to actual segmentation on account of their late appearance and their development. In some Pentastomidae, e.g., P. protelis (Hoyle), they are somewhat broad, and constrictions form between them, thus increasing the resemblance to a true segmentation. In P. proboscideum also such an appearance can be remarked, and is still more conspicuous in P. taenioides. In other Pentastomidae, raised rings are found like broad hoops round a barrel, separated by inter- spaces (Van Beneden, Jacquart). Small circular apertures in the chitinous integument are found distributed all over the surface of the body, and later, in consequence of the formation of rings, arranged in transverse rows upon it. These resemble the two glandular stigmata of the four-limbed larva (Fig. 62), and were regarded by Leuckart as the apertures of integumental glands. A differentiation of the chitinous covering of the body which arises in later larval stages is found in the so-called circles of spines which appear at the posterior edge of each ring, and are characteristic of the fully-formed larva (Fig. 63, st). The larva of P. taenioides, which was formerly taken for a sexually mature form and called P. denticulatum, has the circles of spines specially well developed. They are probably of advantage to the animal in locomotion. Still more important aids in locomotion and attachment are the hooks — two pairs of claw-like chitinous structures (Fig. 63, h), which develop in two sac-like depressions of the integument in front of the mouth (Fig. 62, Jit). The hooks have no connection 136 PENTASTOMIDAE. with the truncated limbs of the larva, nor can they be regarded as limbs, as might appear from their origin as depressions and in front of the mouth. At a later stage they shift further back towards, or even behind, the mouth (Fig. 63). A further differentiation of the surface is found in the appearance of a large number of papillae arranged in pairs at the anterior end of the body (Fig. 63, tp), which have been considered tactile organs (Leuckart, Stiles). The last larval form and its transference to the final host. While these external and internal ontogenetic processes have been taking place the body has lengthened, and has thus been forced to curl up in the cyst, within which the general form of the adult animal is reached. The larva (Fig. 63) now breaks through the cyst, and wanders away from the part hitherto inhabited by it, the circles of spines assisting it in this process. Should the intermediate host in which it lives at this time fall a victim to a beast of prey, the larva possibly passes direct out of the mouth of the latter into its nasal cavity, there, by renewed ecdysis, throwing off its spiny covering, and finally attaining the complete organisation of the sexually-mature Pentastomum. But if no such favourable opportunity is afforded the larva of reaching its final host, it becomes re-encysted within the body of the intermediate host. Encysted larvae which are swallowed by a beast of prey with the flesh of the host, and thus reach the intestine of the former, if sufficiently mature, break through the intestinal wall, and by active locomotion reach the respiratory tissues and the nasal cavity (Gerlach, Stiles). a- Fig. 63.— Free larva of Pentastomum taenioides (the so-called P. denti- cvlatum), from the liver of the rabbit or the nasal cavity of the dog (after Leuckart). a, anus ; d, intestine ; ft, hooks ; m, mouth ; st, circles of spines ; tp, tactile, papillae. 3. General Considerations. The most important point in the ontogeny of Pentastomum is the appearance of a larva furnished with two pairs of limbs. This larval form distinctly indicates that in Pentastomum we have an Arthropod, a fact which is not so evident from the organisation of the adult. It was this larval form above all that led to the classing of the Pentastomidae with the Acarina. The similarity LITERATURE. 137 would be still greater if a six-limbed larva also appeared in Pentastomum, as was maintained by De Filippi. Unfortunately, but little reliance can be placed upon this otherwise important statement, as may be seen from a glance at his figures. A direct comparison of the Pentastomum larva with an Acarid larva is inadmissible on account of the absence of mouth -parts in the former. In this case degeneration may, indeed, have gone even further than in the Acarina, and it is possible that Pentastomum may be derived from forms resembling these latter animals. Certain Acarids, e.g., the Pliytopta, in which two pairs of limbs disappear, and in which the body is lengthened (pp. 107 and 108) might be regarded as indicating the possible line of origin of a form like Pentastomum (Leuckart). But it must be expressly pointed out that there is no definite ground for this view, and that Pentastomum might, with almost equal justification, be derived from some other group of Arthropoda. Unfortunately, the organi- sation of the adult also fails to afford any definite clue, but only makes it clear that Pentastomum is a form much reduced by parasitism. Important systems of organs, such as the respiratory and excretory systems, which elsewhere, by their characteristic development, help to determine systematic position, are wanting. A distinct blood vascular system also is not developed. In the transversely striated musculature, on the other hand, we have an Arthropodan character, and it has already been pointed out that the genital organs can, perhaps, be interpreted in this sense. The ovary in its structure recalls that of the Arachnida, the eggs bulging out like follicles on its surface, and giving the organ an aciniform appearance. LITERATURE. 1. Bexedekt, P. J. van. Recherches sur l'organisation et le developpement des Linguatules (Pentastoma). Ann. Sci. Nat. (3). Zool. Tom. xi. 1849. 2. Filippi, F. de. Nuova linguatula con embrioni di particolar forma. Archiv. Zool., Anat. e Fisiol. Fasc. i. Tom. i. Genova, 18G1. 3. Gerlach, A. C. Pentastomum denticulatum bei zwei Ziegen. Jahresber: d. 1c. Thierarzneischule zu Hannover, ii. 1869. 4. Hoyle, W. E. On a new species of Pentastomum (P. protelis), from the Mesentery of Proteles cristatus, etc. Trans. Roy. Soc. Edinburgh. Vol. xxxii. 1887. 138 PBNTASTOMIDAE. 5. Jaquart, H. Mecanisnie de la retraction des ongles des Felis et des crochets des Linguatules trouvees dans les poumons des serpents. Journ. Anat. et Phys. Paris. Annee iii. 1866. 6. Leuckart, R. Bau unci Entwicklungsgeschichte der Pentas- tomen. Leipzig and Heidelberg. 1860. 7. Lohrmann, E. Untersuchungen liber den anatomischen Bau der Pentastomen. Archiv. f. Naturgesch. Jahrg.lv. 1889. 8. Macalister, A. On two new species of Pentastoma. Proc. Roy. Irish Acad. Ser. 2. Vol. ii. Dublin, 1875-77. 9. Schubart, T. D. Ueber die Entwicklung des Pentastomum taenioides. Zeitschr./. Wiss. Zool. Bd. iv. 1853. 10. Stiles, Ch. W. Bau und Entwicklungsgeschichte von Pentas- tomum proboscideum u. P. subcylindricum. Zeitschr. f. Wiss. Zool. Bd. Iii. 1891. (This work contains a very complete bibliography of the Pentastomidae.) APPENDIX TO LITERATURE ON PENTASTOMIDAE. I. Spencer, W. B. The Anatomy of Pentastomum teretiusculum. Quart. Journ. Micro. Sci. (2). Vol. xxxiv. CHAPTER XXIII. PANTOPODA. Oviposition and Care of the Brood. The Pantopodan female does not deposit her eggs in the usual manner, but transfers them to the male, who attaches them to his third pair of limbs, the so-called ovigerous limbs (Figs. 74, 3, p. 157), and carries them about until the embryo is mature. The eggs are usually collected into large clumps, containing as many as 100. Several such clumps are found on one male, so that, if Avell laden, he may be found to carry 1000 eggs (Dohrn). Although in such cases, and generally among the Pantopoda, the eggs are very small, they are comparatively large in Pallene (0*25 mm. in diameter), being, for example, 125 times the size of an egg of PhoxichiUdium or Tanystylum (Morgan). Pallene carries only a few glutinous egg-clumps, each containing only two eggs (Dohrn). Nymphon, according to Hobk, has specially large eggs (in N. brevicaudatum, 0*5 to 0*7 mm. in diameter), but yet carries a great number. The large eggs are very rich in yolk, the smaller ones naturally have less yolk. The eggs are spherical, and each is surrounded by a delicate membrane (Fig. 64, D). 1. Cleavage and Formation of the Germ -Layers. The cleavage of these eggs is total (Dohrn, Hoek, Morgan) ; but those genera which have small eggs (e.g. PhoxichiUdium and Tanystylum) show equal cleavage, those with larger eggs (Pallene, Nymphon) unequal cleavage (Morgan). Up to the present time but little has been known concerning the early ontogenetic processes in the Pantopodan egg. Many years ago (1843) Kolliker gave an account of the total cleavage of the egg, and Dohrn has more recently described a few stages in the cleavage of the eggs of Pycnogonum that confirm the above conclusion. Hoek, in examining the Challenger material, found a few ontogenetic stages, which, however, could naturally only give a very incomplete idea of 1 40 PANTOPODA. the development of the embryo. Hoek afterwards tried to complete his account by means of observations made on living specimens (Pdllene, No. 7). Morgan next investigated the cleavage and formation of the germ-layers in these animals (Nos. 10 and 11), and in a more recent work (No. 12) he gives a detailed description of these processes in several Pantopodans. We shall thus have to rely chiefly on his account of these processes. In Pallene, the first line of cleavage divides the egg up into two blastomeres, one of these being large, and the other only about a quarter of its size (Morgan). Each of the two spheres is again divided into two by a cleavage taking place at right angles to the first, so that two micromeres and two macromeres are now formed. The third line of cleavage is perpendicular to the two former lines, and gives rise to four micromeres and four macromeres. This stage is followed by one of eight small and eight large cleavage-spheres. From this point onward the micromeres and the macromeres do not divide at the same rate. At a later stage, sections present an appearance like that given in Fig. 65 A, except that the pole of the micromeres consists of smaller cells .than are there figured. The cells are pyramidal, but their boundaries do not in all cases extend to the centre. We here find an indication of transition to the next im- portant stage. Unequal cleavage seems also to occur in the eggs of Nymphon brevicaudatum, which are rich in yolk, for, according to Hoek's figure (Fig. 2, PI. xix., No. 6), one half of the egg at a late stage is composed of smaller cells than the other half. The nuclei of the pyramidal cells, together with the surrounding protoplasm, shift to the periphery (Fig. 64 A and B), the boundaries of the blastomeres being retained to a certain extent (dp). To some degree, however, they disappear, this being specially noticeable towards the centre of the egg (A and B). The nuclei are surrounded by areas of protoplasm, which send out processes into the yolk. Since these complexes of protoplasm, increasing in number by division and shifting closer together, yield the blastoderm (Fig. 64 C), a stage like that seen in the Araneae is passed through, i.e., the yolk- mass appears first divided up into pyramids which disintegrate later. According to Hoek's description, this breaking up of the yolk can still be distinguished in later stages after the blastoderm has formed (cf. figure of Nymphon brevicaudatum, No. 6, PI. xix., Fig. 5). In the centre of the egg a cavity appears (Fig. 64 A, fh), which must be regarded as the cleavage-cavity. Its occurrence, however, does CLEAVAGE AND FORMATION OF THE GERM-LAYERS. 141 not seem to be constant (Morgan), and in any case it soon disappears again. This also, if proved to be correct, would constitute a certain analogy with a condition described for the Araneae (p. 39). In 35. Fig. 64. — Sections through eggs of PaUene in various stages of blastoderm-formation (after Morgan). In D, an imagination (e) appears in the blastoderm, round which cells (probably the first mesoderm-cells) are given off. bl, blastoderm ; d, yolk ; dp, yolk-pyramids ; e, aperture of the invagination ; eh, external and internal integument ; fh, cleavage-cavity (?) them also there is a transition from total to superficial cleavage. There is also a concentration of the blastoderm towards that pole at which later the first indications of the embryo appear (Fig. 64 C). 142 PANTOPODA. The peripheral cells, which were also formerly present at the opposite pole (A and B), disappear. At a time when the blastoderm only partly surrounds the egg, a few cells of amoeboid form are seen lying below it (Fig. 64 C). According to Morgan, cells are given off first at the pole of the micromeres, and then at other parts of the periphery. These cells arise by division of the pyramidal blastoderm-cells in a tangential direction, a process which Morgan, comparing it with the result of observations on other Pantopoda, considers to be one of delamination ; a lower cell-layer forms, which is no doubt to be regarded as the entoderm. This view does not appear sufficiently supported by the facts as yet known, and Morgan's observations have made possible another assumption with regard to the formation of the germ-layers. At the pole of the egg that is richer in cells a thickening appears, which has been compared by Morgan to the primitive cumulus of the Araneid egg (p. 42). A depression then appears at this point (Fig. 64 D, e), and from this an active proliferation of cells takes place. Morgan himself regards this as the formation of the meso- derm, and believes that some of the amoeboid cells which grow into the yolk are also of entodermal nature. The two germ-layers are not yet distinct from one another. In any case, the whole process shows great similarity to the formation of the germ-layers in the Araneae. Amoeboid cells are formed which grow into the yolk, and give rise later to the enteron. That some of the cells which originate near the invagination represent the rudiment of the mesoderm cannot be doubted. These cells soon increase greatly in number, and become arranged into two bands, the mesoderm-bands. The invagination which, on account of its relation to the formation of the germ-layers, might be regarded as the blastopore, is held by Morgan to be the stomodaeum. The two genera Tanystylum and Phoxichilidium, possess smaller eggs less richly provided with yolk, and these differ in their develop- ment from the larger eggs just described, inasmuch as they undergo equal cleavage, by means of which the egg breaks up into two, four, eight, and sixteen blastomeres of equal size. In consequence of this, the pyramidal cells of a later stage are also approximately equal in size (Fig. 65 A). The fact that the yolk contained in such an egg is smaller in quantity than in the other egg leads to a difference in the further development. An actual blastoderm is not at first formed, as in Pallene, but forms later by a process of delamination (Fig. 65 B). A cleavage-cavity also seems to arise, as may be seen in Fig. 65 B. CLEAVAGE AND FORMATION OF THE GERM-LAYERS. 143 Each of the pyramidal cells divides tangentially into an inner and an outer cell, both of these cells then continuing to divide. The outer cells form at the periphery a regular layer, the blastoderm (Fig. 65 C, bl), while many of the inner cells lose their regular boundaries. A yolk-mass thus arises in which isolated cells lie (C, d and z). The inner cells, which were evidently the richer in yolk, have now fused to form a common mass. The embryo thus shows a condition similar to that of other Arthropoda, e.g., the Araneae, there being a peripheral layer of cells (the blastoderm) and an inner yolk-mass with cells distributed in it. The latter, indeed, arise in a different way in the Pantopoda, as is shown in Fig. 65 B. The formation of the germ-layers could not be more exactly made out in eggs with equal cleavage, but Morgan assumes that the enteron is formed from the inner cells (the entoderm). In these forms also, Morgan early a. D6. C. Fin. 65.— Sections through eggs of Tanystylum (A and B) and Phoxichilidium (C) in the final stage of cleavage (A) and in the stage of delamination and blastoderm-formation (B and C) (after Morgan), bl. blastoderm ; d, yolk-mass ; r, the cells which become detached from the peripheral cells (blastoderm) and shift inward. noticed a depression of the peripheral cell-layer, which, like the depression already described in Pallene, he regarded as the rudiment of the stomodaeum. This depression is triangular, a fact which has led to its being compared with the triangular stomodaeum of the Araneae. In view of the comparatively slight knowledge which we possess of the first ontogenetic processes in the Pantopoda, it would be too presumptuous to try to form further conclusions. It has already been mentioned that a certain agree- ment with the conditions in the Araneae exists. The splitting of the blastoderm into two layers maintained by Morgan recalls the processes in the Pseudo- scorpiones (p. 28) ; but these also are too little understood to allow of further comparison. The commencement of development and the further differentiation at the one pole might be compared to the formation of the germ-layers in the 144 PANTOPODA. Araclmida over a limited area of the blastoderm. Morgan states very definitely that this budding-off of an inner layer of cells or multipolar delamination takes place in Pallene slowly, while the outer layer of cells is growing round the yolk ; this we might perhaps refer to an ingrowth of cells with a circumcrescence of the yolk, and compare this process with the corresponding one in the Scorpion. It is advisable to direct the attention of future observers to this point. When the depression of the blastoderm described above appears, the entoderm, according to Morgan, is already formed ; the depression could not, therefore, be compared with the blastopore, although in other respects such a comparison is suggested, all the more so that Morgan thinks that the mesoderm arises round this depression. The fact that other processes in the Pantopoda resemble those in the Araclmida is proved by the formation of a germ-band which, however, is much degenerated, but at the same time shows a certain resemblance to that of the Araclmida. Eggs rich in yolk no doubt represent the more primitive condition in the Pantopoda, and the formation of a blastoderm (of the usual Arthropodan consti- tution) and of a germ-band must also be regarded as primitive. The reduction of the yolk probably had a great influence on the ontogenetic processes, which thus attained the condition in which they are now found (p. 154). 2. The Further Development of the Embryo. Our knowledge of the development of the embryo and the origin of its organs is still very incomplete. The following accounts refer chiefly to Pallene, which was made the subject of careful investigation by Morgan. We must, however, point out that Pallene, unlike other Pantopoda, remains within the egg almost up to the time when the adult form is reached (pp. 148 and 153). When the invagination already mentioned has appeared on the thickened side of the blastoderm, other thickenings of the surface take place. Two of these are oval in form (Fig. 66, g), and lie in front of the triangular depression (to). These represent the rudiments of the supra-oesophageal ganglion. Extending posteriorly from the invagination are two rows of thickenings ; these are the rudiment of the ventral chain of ganglia (gu-gjy) ; laterally to these the first indications of the limbs appear as distinct thickenings (Fig. 67, A). These rudiments, taken as a whole, form a band on the ventral surface of the egg, narrow anteriorly, but broader posteriorly, which may with safety be compared with the germ-band of other Arthro- pods. As the yolk-mass is not very large, the germ-band covers a great part of the egg. As the embryo develops further, it extends laterally, covering a still larger part, so that it can no longer be designated as a distinct germ-band, but rather as the embryonic rudiment surrounding the egg. During this process the embryo has also grown somewhat longer (Fig. 67 A). THE FURTHER DEVELOPMENT OF THE EMBRYO. 145 The order of appearance of the limbs varies in the different forms. In Pallene, according to Morgan, the first to develop is the most anterior pair; these limbs lie near the mouth and are chelate, but their first appearance has not been observed with certainty. The next pair to arise is the fourth, and, in the gap which naturally occurs between these limbs, two pairs of ganglia are visible, those of the second and third pairs of limbs (Figs. 66 and 67 A). The fourth pair of limbs is followed by the fifth and sixth. The third pair develops later, but the second pair is alto- gether wanting in Pallene, and the seventh pair, like the third, appears a short time before the embryo leaves the egg-envelope. Pallene is thus seen to possess, as an embryo, all the limbs of the adult. In most other Pantopoda, however, this is not the case, only three pairs of limbs being usually developed within the egg-envelope. Nymphon brevicaudatum resembles Pallene in possess- ing all the limbs of the adult at the time of hatching (Hoek). While the limbs are appearing and gradually developing, the rudi- ment of the nervous system also undergoes further differentiation. Five pairs of large ganglia can be distinctly made out (Fig. 67 A). They belong to the segments carrying the second to the sixth limbs. It would be interesting to discover the relations of the ganglia which innervate the first pair of limbs to the supra-oesophageal ganglion, i.e., whether they represent a post-oral pair of ganglia fused with the supra-oesophageal ganglion. The two anterior of these five pairs of ganglia approximate closely to each other later (Fig. 72 B), and in the adult these two ganglia, belonging to the second and third limbs, are united. The ganglia of the first three pairs of ambulatory limbs (Fig. 67 A), which appear early, are followed at a much later stage by those of the fourth pair (the seventh pair of limbs) and the abdominal ganglia. In each of the ectodermal thickenings which represent the rudi- Fig. 6C— Superficial aspect of an egg of Pallene, showing the anterior part of the embryonic rudiment (after Morgan), g, rudiments of the supra-oesophageal ganglia ; gji-gjy, ventral ganglia belonging to the segment carrying the second, third, and fourth pairs of limbs ; m, mouth ; I, first limb ; IV, rudiment of the fourth limb. 146 PANTOPODA. ments of the ganglia a pit-like depression appears (Fig. 67 A and B, e), round which the cells of the thickening show a regular epitheloid arrangement (Morgan). An ectodermal depression thus takes part in the formation of the ganglion. The invagination closes later, but its cavity can still be recognised after the ganglion has shifted inward, and has lost its connection with the ectoderm (Fig. 68, e). "When the two anterior pairs of ganglia unite they appear as a single pair, in which, however, there are four pits, which proves that this one pair is composed of two. Morgan's statement as to the participation of ectodermal invaginations in the formation of the ventral ganglia is so definite, that we do not seem justified in doubting this fact (cf. Figs. 67 a. and 68). He himself compares these structures with the ventral organs of Pcripatus (p. 189), and there is no doubt a certain similarity between the two ; but it must be pointed out that the ventral organs are by no means in such direct connection with the ganglia as are the depressions in the Pautopoda. A participation of an ectodermal invagination similar to the above in the formation of the brain cannot be established, although it is just here that we should expect it, when we take into account the cerebral pits in the Arachnida. The development of the ex- ternal shape of the body is com- pleted by the addition of the missing appendages, the length- ening of the embryo, and the commencement of segmentation. The first pair of limbs shifts anteriorly and dorsally. At its base, the proboscis or beak appears to arise as an unpaired anterior outgrowth of the body, carrying the mouth at its extremity. At the posterior end of the body, the vestigial abdomen appears as a small pointed appendage, at the end of which the anus forms. The first of the internal organs to claim attention is the enteron. The entoderm has become arranged into an epithelium surrounding the yolk-mass (Fig. 68, ent), and from this, diverticula, also filled Fio. 67.-^4, embryo of Pallene empusa, seen from the ventral side. B, part of a trans- verse section through the same, to show the paired depressions (c) on the ventral surface (after Moroan). I- VI, limbs; bgt ventral chain of ganglia, the depressions (e) being visible in the ganglia ; ect, ecto- derm ; ent, entoderm ; mes, mesoderm. THE FURTHER DEVELOPMENT OF THE EMBRYO. 147 with yolk, grow out at an early stage into the limbs (di). These are the intestinal caeca, which, in the larva (Fig. 72 A), as well as in the adult, run far into the limbs. This arrangement recalls that in Chelifer, where the yolk also extends far into the limbs (p. 29, and Fig. 16). This is also the case in the Acarina, and in the embryos of some Araneae, e.g., Agalena (Locy). This peculiar feature is known to be retained throughout life in the Pantopoda, in which the trunk is much reduced as compared with the limbs. These latter also contain the genital organs in the adult, and this explains the fact that a process of the mesoderm at an early stage runs between the ectoderm and the entoderm into the rudiment of the m«s* Fig. 68. — Transverse section through an embryo of Palknc empusa at a somewhat older stage than in Fig. 07 A. The ventral depressions (c) have closed (after Morgan). lg, ventral nerve-strand showing fibrous structure on the dorsal side ; coe, mesodermal cavity in the limb ; d, yolk ; di, intestinal caeca of the limbs ; e, the closed ectodermal invagination ; ect, ectoderm ; cut, entoderm ; roes, mesoderm ; p, pair of ambulatory limbs. limb. According to Morgan, a cavity bordered by a mesodermal •epithelium lies at the base of each limb, the mesodermal process extending from this point into the limb (Fig. 68, mes). Morgan does not hesitate to speak of the body-cavity of the limbs. In any case we thus have here the primitive segments which, taken together, represent the two already segmented mesoderm-bands. These latter, together with the rudiments of the ganglionic chain and the limbs on each side, form the germ-band (Fig. 66), although this is con- siderably reduced in accordance with the small size of the egg. As these mesoderm-bands develop at the thickened part of the blastoderm, 148 PANTOPODA. the region beneath which the mesoderm extends may be regarded as the germ-band, the Pantopoda, as has already been pointed out, agreeing in this respect with other Arthropods. Should the appearance of primitive segments and their extension into the limbs be confirmed, a strong resemblance to the Arachnida would be established. Pcripatus, indeed, and many of the Insecta, show the same arrangement, but we do not feel confident in laying so much stress either on this or on the similarity to the ventral organ which Morgan specially points out. Trans- verse sections of embryos of Pallene (Morgan) and of Nijmplwn (Hoek) show unmistakable similarity to sections of a spider. The further development of the mesoderm, its relation to the adult body-cavity, and the formation of the heart, have not yet been ascertained with sufficient certainty. The heart appears in the dorsal middle line after the mesoderm has already given rise to a number of schizocoele-like cavities. More accurate accounts of the participa- tion of the primitive segments in these processes (the further differentiation of the mesoderm and the formation of the heart) would be of great interest. The mesodermal tissue with its cavities increases in extent as the yolk-mass degenerates. The latter is absorbed by the surrounding entodermal epithelium. Yolk-cells do not appear to play any special part in this process, and may, indeed, be wanting. The enteron becomes connected with the stomodaeum, which is derived by Morgan from the invagination already mentioned as appearing very early. The proctodaeum does not appear until very late, when the seventh pair of limbs and the abdomen form. The Form of the Larva and its Transformation into the Adult. The Larva. Most of the Pantopoda develop through metamor- phosis. The larvae usually have three pairs of limbs, but some leave the egg in a more advanced condition ; the young Palleney for instance, when hatched is provided with all the limbs of the adult, and this higher stage of development is also attained in the egg by a few species of the genus Nymphon. The various species of this genus differ from each other in this point; in some of them the larva, at hatching, has only four or five pairs of limbs (Hoek). The many Pantopodan larvae that have been described differ only slightly from each other, and are easily derived from a larval form provided with three pairs of limbs. This form, which was first THE FORM OF THE LARVA. 149 carefully examined by Dohrn, has a compact body (Fig. G9), some- times almost square, or else rounded, seldom long or oval. The body is not externally segmented, although it carries three pairs of limbs ; in this respect this larva bears a certain resemblance to the Crustacean Nauplius. It has been compared with the latter, but the resemblance is merely superficial. Fig. 69. — Larva of Achclia lacvis immediately after hatching (after Dohrn). 7-777, limbs; bg, strands of connective tissue ; d\ spine on limb 7 with gland (dr) ; da, enteron ; /, filament of the glandular secretion ; g, brain (with the eyes above it) ; m, muscles ; s, proboscis ; v, vacuoles in the gland. The larva, as already stated, is supplied with three pairs of limbs. The most anterior limb has three joints, and is chelate. At its base it has a movable spine (Fig. 69, d), which, in other genera, is considerably longer than in the larva of Achelia depicted in Fig. 69. This appendage brings about a certain similarity to the biramose limbs of the Crustacea, but we would not lay 1 50 PANTOPODA. any great stress upon this point. A tolerably large spine, comparable with that on the first limb, also occurs on the two following limbs (Fig. 69). That on the first extremity, however, is distinguished from the others by having at its point the aperture of a gland (dr). The fine filaments which can be produced through this aperture serve for attaching those larvae which, after quitting the egg-envelope and undergoing the first moult, fix themselves on the ovigerous limbs of the male. The second and third pairs of limbs possess hooks only (Fig. 69, 77. and 777.). The muscles of all the limbs, especially the first, are well developed. Whereas the first are used for fixation, and especially for prehension, the two posterior pairs are used for crawling and climbing. These larvae live among algae, Hydroids, etc. Another feature of the external organisation of these larvae is the proboscis, or beak, which arises as a ventral conical outgrowth between the bases of the anterior limbs (Fig. 69, s). At its tip lies the oral aperture. It appears as if the proboscis arose near the stomodaeum as an ectodermal outgrowth, although some have been inclined to attribute its origin to fusion of the upper lip with a pair of limbs (Adlehz). It is impossible to decide whether we are justified in comparing it to the provisional proboscis of Clielifery which it cannot fail to recall, on account of the slightness of our knowledge of this latter organ. The intestine is already provided with outgrowths, the anterior pair of which are beginning to extend into the first pair of limbs (Fig. 69). From the intestine, fibres of connective tissue extend to the body-wall. The anus does not yet seem to be present (Dohrn), and no doubt does not appear until later with the rudiment of the abdomen (Fig. 71 B). The nervous system of the larva consists of the supra-oesophageal ganglion and only two pairs of ganglia on the ventral side. Im- mediately above the supra-oesophageal ganglion lie the two eyes in close contact (Fig. 69). The manner in which these arise is of special interest, as it appears to offer a further point of agreement with the Arachnida. The eyes, like the nervous system, attain full development during metamorphosis. The two eyes of the former stage are now joined by another pair. So as to understand how these develop we shall have to explain briefly the structure of the Pantopodan eye, which is as yet very insufficiently understood. These eyes, like those of the Araneae, consist of a corneal lens, a subjacent hypodermis DEVELOPMENT OF THE EYE. 151 (vitreous body), a thick layer of retinal cells, and a layer of pigment behind the whole. In the retina, the cell nuclei lie in front of the rods ; these latter therefore belong to the posterior part of the cells, and thus come into direct contact with the pignient-layer (Fig. 70, st). The nerve-fibres become connected, however, with the outer ends of the visual cells, so that here also the same conditions prevail as are found in the posterior middle eyes and the lateral eyes of the Araneae (Fig. 34, p. 64). This last point, which seems to be implied in the description given by Hoek, has recently been established by Morgan (No. 12). The ontogenetic stages, as well as the adult structure, closely resemble those of the Arachnida, as may be seen by comparing Fig. 70 with the ontogenetic stages of the eyes of the Scorpiones and the Araneae illustrated in Fig. 10, p* 14, and Fig. 35, p. 65. In Fig. 70, an invagination extending from one side below the hypo- dermis is suggested. The retina and the pigment-layer thus arise, and out of the superjacent hypodermis the layer forming the vitreous body, which secretes the lens on its outer side. An inversion thus takes place in the formation of the eye, and its innervation would be from the first explicable in the same way as that of the eye of the Araneae. In earlier stages in the development of the eyes, an invagination is not so distinctly recognisable as in the eyes of the Araneae. The different layers of cells lie somewhat close to one another, and Morgan assumes that no actual (complete) invagination takes place, but rather that new cells are continually being added to the eye from the point of ingrowth, and that thus finally the layers, like those in the Arachnid eye, are formed (Fig. 70). A thicken- ing of the hypodermis, which appears laterally to the eye, perhaps yields the new cell-material. This hypodermal thickening recalls the one found near the Crus- tacean eyes and those of Limulus (Vol. II., pp. 280 and 359). The development and the structure of the Pantopodan eyes suggests through- out a comparison with those of the Arachnida. Morgan's statement that the ettr r Fig. 70. — Longitudinal section through one of the posterior eyes of the larva of Tanystylum (after Morgan), c, cuticle ; cct, hy, ectoderm (hypodermis) ; gl; vitreous body ; pi, pigment- layer ; r, retina ; st, rods. 152 PANTOPODA. rods arise through fusion of the rods of two neighbouring cells, makes the similarity appear still more striking, and leads to the same conclusion in both cases ; viz. , to a derivation of these apparently simple eyes from compound eyes. Our knowledge of the eyes of the Pantopoda is, however, still too slight to allow of any definite conclusions ; Morgan even adopts an altogether opposite view, and explains the inversion which in all cases is present in these eyes, by the degeneration of the posterior part of an optic invagination and the better development of the anterior part. In this way he derives the inverted Pantopodan eyes from such simple eyes (ocelli) as those of the Insecta, being guided in this decision chiefly by a certain bilateral symmetry in the Panto- podan eye. But that method of development as it appears in the ontogeny of the eye, i.e., the growth of the invagination towards one side, is merely a caenogenetic process, and serves for the quicker attainment of the structure now possessed by the adult eye. It has the significance of an abbreviated develop- ment. As a logical consequence of this view, a corresponding assumption must be made for the Arachnid eyes. We cannot here examine Morgan's conclusions Fig. 71. — Larvae of Tanystylum in two different stages, seen from the ventral side (after Morgan), a, anus ; abd, abdomen ; ig, ventral chain of ganglia ; m, mouth ; s, proboscis ; I-IV, first four limbs. more closely, but refer to the original treatise and to our own view of the eyes of the Arachnida given above (p. 68). On the other hand, it must be mentioned that the description recently given by Claus (No. 2) of the origin of the median eye in the Crustacea, involuntarily recalls the condition of the eyes in the Pantopoda. The median eyes of the Crustacea are said by Claus to arise by inversion, and seem to have their elements arranged like those of the Pantopodan eyes. The rods lie on the inner side, directed towards the pigment- cup of the eye. while the nerve-fibres join them from the opposite side, where also lie the nuclei of the retinal cells. The principal change which brings about the transformation of the larva into the adult is the formation of new segments at the posterior part of the body. The limbs already present either pass over directly to the adult, merely growing and developing further, or TRANSFORMATION OF THE LARVA INTO THE ADULT. 153 some of them, usually the second or third, and in many cases both of these or even all the three anterior pairs temporarily degenerate, the corresponding adult limbs growing out at the same points (Dohrn, Hoek). In Pallene the second pair is wanting, and does not even occur as a vestige, while in Tanystylum the first pair is wanting as a functional appendage, but appears ontogenetically as a Avell-developed pineer-carrying limb (Fig. 71 A and B), and only gradually degenerates in the later larval stages ; it is still present in the adult as a small, vestigial two-jointed bud (Morgan). The position of the second and third pairs of degenerated limbs is marked by the appearance of the apertures of what are presumably excretory glands (coxal glands). The tubular spine of the first limb, through which the above gland opens, is thrown off in one of the moults and gives place to an ordinary short spine. It has therefore the significance of a larval organ. The first indication of the formation of new body-segments is, according to Dohrn, found in a paired swelling of the intestine behind the last of the larval limbs, accompanied by a bulging of the body-wall. At the same time, in the posterior part of the ventral surface, a thickening of the ectoderm appears which is the rudiment of a new pair of ganglia. The ectoderm begins to become wrinkled in the posterior part of the body and rises up above the newly- formed lower layer. The larva now moults, after which it is evident that a limb has appeared on the bulging of the body-wall just mentioned ; into this limb an intestinal caecum is continued. It is thus clear that this is a new limb, which soon develops and becomes jointed (Fig. 71 A and B). The other limbs form in the same way. Only when the body thus lengthens do the three anterior pairs of limbs also take part in the transformation (Dohrn). The short abdomen arises as a posterior sac-like swelling, and the anus appears upon it (Fig. 71 B). The transformation of the six-limbed larva just described takes place in some forms, as has already been mentioned, within the egg-envelope ; Nymphon breuicollum, for example, leaves the egg when provided with five well-developed pairs of limbs (Fig. 72 A and B), and the first rudiments of a sixth pair. Other points of its organisation, especially the shape of the limbs with the intestinal caeca extending far into them, can be made out without further assistance from Figs. 72 A and B. The young of Nymphon brevicaudatum possess all the limbs at hatching (Hoek), and the same condition is found in the genus Pallene (Dohrn, Morgan). 154 PANTOPODA. During metamorphosis, the rudiments of the genital organs which were not observed in the six-limbed larva become recognisable. In the larva with four pairs of limbs (Fig. 71 B), a compact mass of cells, the first rudiment of the genital gland, lies in the median line dorsally to the enteron, somewhat near the fourth pair of limbs (or first ambulatory limbs). The anterior end of this mass splits later into two parts, which grow out towards the bases of the limbs just mentioned. The posterior end of the germ-gland then splits in the same way, the genital tubes which run into the limbs thus arising. The wide tubular rudiment of the heart has formed at the anterior part of the body, also from mesoderm cells, dorsally to the genital rudiment, and thus directly beneath the integument. Fig. 72.— Larvae of Nymphon brevicollum soon after hatching. A, dorsal, li, ventral aspect (after IIoek). I-V, the five anterior limbs; bg, ventral chain of ganglia; d, yolk-mass; dl, diverticulum of the yolk-filled enteron in the limb; g, brain ; s, proboscis. The differences observed in Pallene and Nymjihon give rise to the question as to which method of development is to be considered the more primitive among the Pantopoda ; in this respect the appearance of larval organs and the casting of a larval integument, observed by Dohrx in Pallene, suggest that the direct development of this form must be regarded merely as an abbreviation of the indirect method of development, and that the latter is the more primitive. In consequence of the more complete development of the embryos in the egg the latter must have a richer supply of nutritive material. The large amount of yolk in the eggs of Pallene and Nymplwn would, under these circumstances, appear as a later acquisition, and it then seems doubtful whether we ought to ascribe to the first ontogenetic processes of these eggs a truly primitive character, although we feel inclined to do so on account of the greater resemblance of their development to that of other Arthropoda. The course of development in Phoxichilidium differs from that of other Pantopoda in that the form of the larva undergoes THE LARVA OF FHOXICHILIDIUM. 155 considerable degeneration before passing into that of the adult. This is connected with its parasitic manner of life. On leaving the egg, the larva of Phoxichilidium possesses on the whole the organisation of the usual sixdimbed Pantopodan larva, but is distinguished from the latter by the fact that the usually hookdike terminal joints of the two posterior pairs of limbs are Fig. 73. — Various larval stages of Phoxichilidium, (after Dohrn, Semper, and Adlerz). A, free larva with the tendril-like flagellae on the two posterior pairs of limbs (77 and 777). B-D, larval stages found in Hydroid polyps. (A is more highly magnified than the other figures.) 7-777, limbs ; d, intestine with its caeca ; dr, glands of the first limb ; h, larval integument in the act of becoming detached ; n, ventral chain of ganglia ; s, proboscis. much lengthened, and form long flagellae, which can coil up like tendrils (Fig. 73 .4). These flagellae, which may be much longer than those represented in Fig. 73 (e.g., in Phoxichilidium femoratum, (Hoek)), are probably used for attachment, the larvae winding them round the Hydroids (e.g., Hydractinia, Podocoryne, Tuhularia, 156 PANTOPODA. Plumularia), which are chosen by them as hosts. Dohrn assumes that the larvae, after attaching themselves to the Hydroids by the help of the flagellae, throw off during a moult the two posterior pairs of limbs that carry the latter, and pass through the oral aperture of the polyp into its gastral cavity. They are certainly found later in such a position, and here pass through the further stages of their development. The tendril-like flagellae seem not to occur in all Phoxichilidia, for R. von Lendenfeld lias described a larva of Ph. plumulariac not distinguished in any way worth mentioning from the usual Pantopodan larva. This larva further differs from other Phoxichilidia in its manner of life ; it does not penetrate into the polyps, but only attaches itself to them by the help of its pincers and by burying its beak in the host's body at the base of the head. The larva remains in this position until it has almost attained the form of the adult animal. "VVe may gather from v. Lendenfeld's description that the further development of the forms discovered by him takes place as in other Phoxichilidia, for he also mentions a two-limbed stage. It has already been stated that the larvae cast off the flagellae and limbs at ecdysis (Semper, Dohrn). The larva moults several times (Fig. 73 B), the second and third pairs of limbs degenerating completely (Semper) ; but, according to Adlerz, some vestiges of the posterior pairs are retained (Fig. 73 C and D), and it is in place of these that the second and third limbs of the adult arise. The larvae, several of which often occur in one polyp, with their large anterior limbs, have a very peculiar appearance (Fig. 73) in this stage. In the following stages the limbs are found to degenerate still further (this is also evident from the figure given by Adlerz), but, with the bulgings of the intestines, the rudiments of the posterior segments begin to appear. The ganglia of these develop and the outgrowths of the body-wall which yield the limbs soon appear (Semper, Adlerz). These processes seem, on the whole, of the same nature as those before described. When three pairs of ambulatory limbs have formed, and the fourth is present as a rudiment, the young Phoxichilidium leaves the polyp and leads a free life. 4. General Considerations. Although much has been written as to the relationships of the Pantopoda, these are still far from clear. The ontogeny of these animals, as far as it is now known, unfortunately throws little light upon the subject. In comparing the Pantopoda with other divisions of the animal kingdom, attention is turned chiefly to the Crustacea and the Arachnida. The form of the larva is of the greatest GENERAL CONSIDERATIONS. 157 importance in a comparison with the former, while in comparing the Pantopoda with the Arachnida the shape of the adult receives more attention. It cannot be denied that the whole appearance of the Pantopoda suggests a certain similarity to the Araneae. But a nearer comparison at once reveals a difficulty, inasmuch as the Pantopoda possess one pair of limbs more than the adult Arachnid. An attempt has been made to overcome this difficulty by considering the first two pairs of limbs of the Pantopoda (Fig. 74, 1 and 2) as ecpuivalent to the chelicerae and the pedipalps of the Arachnida, and the third to the sixth limb of the former as equal to the ambulatory limbs of the latter (Fig. 74, 3-6). The ovigerous limbs (Fig. 74, 3) would thus represent the first pair of ambulatory limbs of the Arachnida, and the seventh limb would be the homologue of Fig. 74. — Male of Nymphon hispidum seen from the ventral side. The setae are omitted (after Hoek, from Lang's Text-book). 1-7, limbs ; aft, abdomen ; s, proboscis. the first pair of abdominal limbs. In view of the fact that in the Insecta an abdominal segment is separated from the posterior part of the body, and may enter into the closest relation to the thorax, such a view might be defended. Those Avho adopt it consider that the addition of another pair of limbs to those already specialised for locomotion was determined by the withdrawal of the third limb from the ambulatory series for use in the care of the brood. According to this view, the four pairs of ambulatory limbs of the Pantopoda would not be homologous with those of the Arachnida. The last homology must, however, be regarded as possible, and in that case the loss of an anterior limb in the Arachnids would have to be assumed. It has already been pointed out (p. Ill) that the rostrum has been conjectured to represent a pair of limbs. 158 PANTOPODA. Iii carrying further these attempts to homologise the limhs, this last assumption leads to certain difficulties as to the position of those now under consideration. A careful examination of the various views held on this subject, which are all more or less speculative, would lead us too far, but we must draw attention to the fact that the ovigerous limbs have by some been regarded not as independent limbs, but as belonging to the second limb. Schimkewitsch, who adopted this view (Nos. 14 and 15), in defending it laid weight on the fact that the rudiments of the pedipalps in the embryos of the Araneae are biramose (pp. 52 and 112). Each of the branches is said to give rise to a limb. This view is not supported by ontogeny, since, in the Pantopodan larva, the second and third limbs arise quite separately. Just as little does ontogeny support the view that the tripartite proboscis of the Pantopoda arises through the fusion of a pair of limbs with an (unpaired) upper lip. A third pair of limbs would then be added, for it cannot be assumed that the paired pieces are merely parts of one limb. The loss of two pairs of limbs by the Arachnida has even been suggested (Croneberg, p. 111). The ontogeny of the Pantopoda seems to show that the beak is, as Dohrn assumes, only an outgrowth of the lips of the stomodaeum. The number of the ganglia corresponds to that of the limbs, Adlerz, indeed, finds (in the adult), besides the ganglia of the second and third limbs, another pair which innervates the proboscis. A final decision on this point will only be possible when the ontogenetic conditions are clearly established. The first limb is innervated from the brain, while the second and third limbs receive their nerves from the first and second ventral ganglia. It would be of the greatest importance to make certain whether an originally post-oral ganglion unites with the brain, as in the Crustacea and the Arachnida. If this is not the case, the limbs lost in the Arachnida must be considered to be the first limbs of the Pantopoda, and their homologues must be sought in the conjectural rostral limbs of the Arachnida. It does not, however, seem probable that the first chelate limbs should be true antennae, and consequently not comparable to the chelicerae of the Arachnida. We have already several times pointed out various resemblances between the development of the Pantopoda and that of the Arach- nida, but these do not appear to us sufficient to lead to further conclusions as to the relationship of the two groups. Morgan, chiefly on account of his ontogenetic researches, has recently spoken in favour of such a relationship. It appears to us that, in taking up this position, he was largely influenced by the structure of the Pantopodan eyes. But Claus has recently shown (Vol. ii., p. 167, and Vol. iii., p. 115) that the median eyes of the Crustacea also arise by invagination, and that their component parts apparently have the same position as those of the Pantopodan eyes (No. 2), so that in this character there is similarity to the Crustacea just as much as to the Arachnida. In assuming the loss of an anterior limb, we are obliged to shift back any connection between the Pantopoda and the Arachnida to very early times in the history of the Arthropoda, before the THE FORM OF THE LARVA. 159 Arachnid a arose from forms nearly related to the Xiphosura, for the Arachnida agree with the Xiphosura in many more points than with the Pantopoda. If we must remove the union of the two to such a remote period, the few points of comparison again lose their significance, seeing that they refer chiefly to the more highly developed forms and not to the lower forms. To derive the Pantopoda directly from the Arachnida, however, seems im- possible, the latter having attained far too high a grade of organisation to allow of such a derivation. Even if the Pantopoda were originally related to the Arachnida or some other segmented form, they have in their whole organisation become far removed from it, and have become markedly specialised. The decided preponderance of the limbs over the trunk, and the almost complete degeneration of the latter (Fig. 74) determined the displacement of the internal organs (intestinal caeca and genital glands) into the limbs. The opening of the genital organ on the second joint of the limbs is probably a consequence of this change, and thus has a secondary character. In those cases in which the genital apertures are found, not on several limbs, but only on the seventh pair, as in Pycnogonum, we might be inclined to derive this condition from that in Limulus and the Arachnida, in which the genital apertures lie in the first [second] abdominal segment, and to regard it as primitive, but such an assumption is not supported by any convincing evidence. The reduction of the trunk as compared with the limbs becomes still more marked through the degeneration of the abdomen. The latter is merely a short, truncated appendage of the body (Fig. 74), but the presence of two pairs of ganglia (Dohrn) shows that it originally consisted of more than one segment. In Ammothea and Zetes, the abdomen shows externally a division into two parts, and in some other Pantopoda evidence of a larger number of segments (three to seven) is said to be forthcoming (Hoek, Xo. 7, pp. 453 and 454). Should the Pantopoda prove to be connected at the roots with the Arachnid stock, they would thus in a certain way be related to the Crustacea. The latter, however, appear to us to be too far removed in structure to admit of any relation between the Panto- podan larva and the Nauplius. Those recent observers who have most thoroughly studied the ontogeny of the Pantopoda cannot find any close relation between the two. Hoek regards the larva as representing the primitive racial form, just as the Nauplius was 160 PANTOPODA. formerly regarded. Dohrn considers it to be, like the Nauplius, a modified Annelid larva, and derives the Pantopoda from forms resembling the Annelida. Morgan, however, is unable to accept this conclusion, but regards it as a secondary larval form which can no longer be referred to the Annelidan larva. It seems to us that this last view might easily be reconciled with that of Dohrn. Dohrn derives the Pantopoda from the Annelida, without relating them to the Crustacea and the Arachnida. He thus regards them as a distinct, independent group, and this is also Hoek's view (No. 7). Morgan, on the contrary, favours the relationship to the Arachnida, a view towards which Schimkewitsch also inclines (No. 15). He attributes the Pantopoda to the same racial form as the Arachnida, but believes that they branched off early and developed in a different direction. The most recent investigator of the Pantopoda, G. 0. Sars (No. 13), does not connect them with either the Crustacea or the Arachnida, but wishes to make them into a separate class. In consequence of all these varying opinions we are unable to define with any degree of certainty the systematic position of the Pantopoda. On the whole, according to the present state of our knowledge, we shall do best to follow the conclusions of Dohrn (No. 4). If, notwithstanding this last decision, we have appeared to place the Pantopoda as next in order to the Arachnida, and to dwell on the possibility of the relationship of these two forms, this was done for practical purposes, since we should otherwise have been obliged to classify them in a less satis- factory manner, because they seem to show some slight similarity in their development, and a convergence in some anatomical points, to the Arachnida.* LITERATURE. Only a few of the many treatises describing the ontogenetic stages of the Pantopoda are here enumerated. The following literature, however, will afford further references. 1. Adlerz, G. Bidrag till Pantopodernas Morfologi och Utveck- lings historia. Bihang till k. Svenska Vetenskap. Akad. Handlingar. Ed. xiii. Afd. iv. No. 11. Stockholm, 1888. 2. Claus, C. Ueber den feineren Bau des Medianauges der Crustaceen. Anz. k. k. Akad. Wiss, Wien, Mai, 1891. No. 12. * [Ihle (App. to Lit. on Pantopoda, No. I.) holds that it is impossible to ally the Pantopoda with the Arachnida or with the Crustacea, but thinks that the Mvriopoda may he regarded as the ancestral stock. He lays special stress on the presence of a pair of abdominal appendages. — Ed.] LITERATURE. 161 3. Dohrn, A. Untersuchungen iiber Bau unci Entwicklung der Arthropoden. 2. Pycnogoniden. Jen. Zeitsclir. f. Natunv. Bd. v. 1870. 4. Dohrn, A. Die Pantopoden des Golfes von Xeapel. Fauna und Flora des Golfes von Neapel. Monographie iii. Leipzig, 1881. 5. Faxon, "W. Bibliography to accompany "Selections from Embryological Monographs." Pycnogonida. Bull. Mus. Comp. Zool. Harvard College. Vol. ix. 1882. p. 247. (Contains the older bibliography.) 6. Hoek, P. P. C. Keport on the Pycnogonida. Voyage of H.M.S. Challenger. Zoology. Vol. iii. 1881. 7. Hoek, P. P. C Nouvelles Etudes sur les Pycnogonides. Archiv. Zool. exper. Tom. ix. Paris, 1881. 8. Hodge, G. Observations on a species of Pycnogon (Phoxichili- dium coccineum) with an attempt to explain the order of its development. Ann. Mag. Nat. Hist. (3). Vol. ix. 1862. 9. Lendenfeld, Pv. vox. Die Larvenentwicklung von Phoxichili- dium plumulariae. Zeitsclir. f. Wiss. Zool. Bd. xxxviii. 1883. 10. Morgan, T. H. Preliminary Note on the Embryology of the Pycnogonids. Johns Hopkins Univ. Circidars. Baltimore. Vol. ix. No. 80. 1890. 1 1 . Morgan, T. H. The relationships of the Sea-spiders. Biological lectures delivered at the Marine Biologiccd Laboratory of Woods Holl. Boston, 1891. 12. Morgan, T. H. A contribution to the Embryology and Phylogeny of the Pycnogonids. Studies Biol. Lab. Johns Hopkins Uvniersity. Baltimore, Vol. v. 1891. 13. Sars, G. 0. Pycnogonidea. Den Norske Nordhavs-Expedition 1876-7S. Bd. xx. Christiania, 1891. 14. Schimkewitsch, W. Etude sur l'anatomie de l'Epeire. Ann. Sci. Nat. (6). Zool. Tom. xvii. 1884. 15. Schimkewitsch, W. Les Arachnides et leurs affinites. Archiv. Slaves Biol. Tom. i. Paris, 1886. 16. Semper, C. XJeber Pycnogoniden und ihre in Hydroiden schmarotzenden Larvenformen. Arb. Zool. Inst. Univ. Wiirzburg. Bd. i. 1874. APPENDIX TO LITERATURE ON PANTOPODA. I. Ihle, J. E. W. Ueber die Phylogenie und Systematische Stellung der Pantopoden. Biol. Centralbl. Bd. xviii. 1898. M CHAPTEE XXIV. TARDIGRADA: The eggs of the Tarcligrada are laid either singly {Macrobiotus Hufelandi) or several together, and are left in the cast-off skin of the mother. In the case of eggs laid singly, the egg-integument is thickly studded with small prominences which render its examination very difficult. When several eggs are laid together the egg-envelope is smooth and transparent. The species investigated by Kaufmann seems to have been comparatively easy to study, and he was able to establish the fact that its cleavage is total and equal, as v. Siebold had already stated. Kaufmann followed the process of cleavage up to the formation of a morula stage composed of cells of about equal size. He then observed the separation of a peripheral cell-layer from the central mass, and the flexion of the embryo which supervenes. The concave side of the embryo seems to correspond to the ventral surface, for the limb-rudiments here arise on the two sides. Two pairs of prominences appear first ; these are apparently the two anterior pairs of limbs, which are followed by the third and fourth pair. When the young leave the egg, they possess the full number of limbs and have the general form of the mother. v. Siebold's statement (No. 4, p. 553), that the Emydiac have only three pairs of limbs when they leave the egg, may be traced to a misunderstanding of Doyere's account (No. 1). This author states that the limbs are here not fully developed, not that one pair is wanting. It does not appear from v. Siebold's account that he himself investigated this point, which is of interest in con- nection with the comparison that has repeatedly been made between the Tardigrada and the Acarina. The accounts of the ontogeny of the Tardigrada are unfortunately so scanty that we can hardly gain anything from them applicable to the whole group. We cannot even tell for certain if a blastoderm and germ-band develop, although this is probable. The armature of the mouth is evidently a product of the stomodaeum, as may be gathered from the study of the adult anatomy ; mouth-parts (in the sense in which the term is used of the Arthropoda) apparently do not appear TARDIGRADA. 163 even as rudiments.* For this reason alone, the association of the Tardigrada with the Arachnida and especially -with the Acarina, which has repeatedly heen attempted, chiefly on account of the number of limbs, cannot be maintained. "With regard to the number of their limbs, the Tardigrada cannot be compared with any other division of the Arthropoda, and the form of their limbs is so simple as to distinguish them in this respect also from all other Arthropoda. The segmentation of the body in the Tardigrada is peculiar, inasmuch as the abdomen is wanting and the anus lies in front of the last pair of limbs. In other points also the organisation of the Tardigrada shows peculiarities which distinguish it from that of other Arthropods; we may mention, by way of example, the unstriped muscle-fibres and the absence of special respiratory organs, and of a vascular system. For these and other reasons (cf. Plate, !No. 3) we are led to consider the Tardigrada as a lateral branch of the Arthropod stock which separated from it near its root, and developed in a special and unique direction. LITERATUEE. 1. Doyere, M, Memoire sur les Tardigrades. Ann. Sci. Nat. (2). Tom. xiv. Zool. 1840. 2. Kaufmann, J. Ueber die Entwicklung und systematische Stel- lung der Tardigraden. Zeitsclir. f. Wiss. Zool. Bd. iii. 1851. 3. Plate, L. Beitrage zur Naturgeschichte der Tardigraden. Zool. Jahrb. Abth. f. Anat. Bd. iii. 1888. 4. Siebold, C. Th. vox. Lehrbuch der vergleichenden Anatomie der wirbellosen Thiere. Berlin, 1848. pp. 552 and 553. APPENDIX TO LITERATURE ON TARDIGRADA. I. Erlanger, R. von. Beitrage zur Morphologie der Tardigraden. Zur Embryologie eines Tardigraden, Macrobiotus macronyx. Morph. Jahrb. 1895. And Biol. Centralbl. Bd. xiv. and xv. II. Kennel, J. von. Die Verwandschaftsbeziehungen und die Abstammung der Tardigraden. Sitzmigsb. Ges. Dorp. 1891. * [Erlanger (App. to Lit. on Tardigrada, Xo. I.) has made an exhaustive investigation into the development of Macrobiotus macronyx. He finds that cleavage is total and equal and results in the formation of a long, oval blastula with a eleavage-eavity ; an invagination-gastrula arises whose blastopore, which occupies the position of the future anus, soon closes ; a very short proctodaeum is formed, and the archenteron divides into a pharynx, gullet, and stomach ; four pairs of enterocoeles form, and the embryo becomes divided into a head and four thoracic segments. The head-coelom enters into connection with the first pair of appendages ; the coelomic pouches of the second and fourth thoracic segments enter into relation with the remaining appendages, while that of the third segment gives rise to the gonads and (?) to a pair of enteric glands. The oral papillae, the salivary glands, the nerve-ganglia, and the eyes all arise from the ectoderm. Erlanger regards the head and first two thoracic segments as the cephalo-thorax ; the third and fourth segments as the abdomen, behind which is a transitory post-abdomen. — Ed.] CHAPTER XXV. ONYCHOPHORA (Peripatus). Structure of the Eggs and Nourishment of the Young by the Mother. The eggs of Peripatus pass through their development in the uterus, but there is considerable variation in this respect in the different geographical species. This point has been carefully investi- gated up to the present time in P. novae-zealandiae (Australia), P. capensis and P. Balfouri (Africa), and P. Edwardsii, torquatus, and Imthurni (South America). These species differ even in the size of the egg and of the mature embryo. The oval eggs of P. novae- zealandiae are 1/5 mm. long and 1 mm. broad, and the young which hatch from them are not much larger than the eggs themselves ; the eggs of P. cajyensis and P. Balfouri are 0"4-0'6 mm. long, but the newly-hatched young of the former has a length of 10-15 mm., and that of the latter is about half as long. In P. Edioardsii the mature embryo attains the length of 22 mm., i.e., a third of that of the mother, while the egg is here only O04 in diameter. The species in which the young are largest at birth have thus the smallest eggs,. and vice versa. The explanation of this striking fact is to be found in the circumstance that in the South American species the egg or embryo remains in close connection with the mother, and is nourished by means of a "placenta and an umbilical cord" (Fig. 88, p. 179).* This accounts for the extraordinarily small size of the eggs in this case, and for their being devoid of nutritive material. In the African species the eggs are larger, but the newly-hatched young are smaller than in P. Edwardsii, there being no correlation between the size of the egg and that of the embryo, the latter, although not * We are here following the definite statements of v. Kennel (No. 4), which rest upon his own observations, although we are aware of the statements made by Hutton (No. 3) as to the size of the newly-hatched young of P. novae- zealandiae. These, according to this latter author, measure from 8 to 10 mm. Since v. Kexnel's statements were not contradicted by the more recent observers of the Peripatus of New Zealand, we must assume that the difference is only apparent, and that the large size of the embryo as compared with that of the egg must be traced not to its greater mass, hut rather to its increase in length and to its extension after leaving the egg. CLEAVAGE AND FORMATION OF THE GERM-LAYERS. 165 attached to the wall of the uterus, is nevertheless nourished by fluid yielded by that organ. In the New Zealand species, such nourishment from the mother is not needed, since the embryo is not essentially larger than the egg. In this case, therefore, the material for the development of the embryo must be contained in the egg itself. It is actually found that the egg of P. novae- zeal andiae is very rich in yolk, as are those of most Arthropods. The general course of its cleavage also agrees with what is found, for instance, in the Insecta. Considering the close relationship of Peripatus to the Arthropoda, which can hardly be disputed, it seems likely that the condition of the egg of P. novae-zealandiae is the primitive condition. It is probable that Peripatus, like the terrestrial Arthropoda generally, origin- ally produced eggs rich in yolk which it laid. This state of things is recalled by the presence of a firmer egg-envelope in P. novae-zealandiae, already pointed out by Sedgwick (No. 11) ; the laying of eggs not fully developed also in this same species points in this same direction, even though we find that eggs laid thus early do not attain full development (Hutton, No. 3). The capacity for developing the eggs within the body must have been secondarily acquired. The ■egg of the New Zealand species, which is rich in yolk and develops within the uterus, represents the first step in this newly-acquired course of development. An accumulation of nutritive material in an egg which develops within the uterus is unnecessary, and this is opposed to the assumption that in P. novae- zealandiae we have a specialised form in which the egg has been secondarily supplied with yolk. A further step in adaptation would be represented by P. capensis. The eggs here show a spongy structure as if penetrated by fluid yolk, and this, as well as the method of their development, seems to indicate that they to a certain extent represent a degenerate condition of eggs originally rich in yolk. Isolated granules of yolk also appear in these eggs, and in P. Balfouri the egg is still somewhat rich in definite yolk-masses. In the species found in the West Indies, the nourishment of the embryo by the mother has become so complete that no trace of the former rich supply of yolk remains in the eggs which have become extraordinarily small. These biological conditions naturally find expression also in the method of development of the different species. * 1. Cleavage and Formation of the Germ-Layers. Although the early development, the cleavage and the formation of the germ- layers, has repeatedly been investigated in different species, our knowledge of these processes still remains very incomplete. This fact is accounted for by the * [Willey (App. to Lit. on Onychophora, No. II.), from the study of the «gg of P. novae-britanniae, has come to conclusions which are exactly the reverse of those given above ; and here he is in agreement with v. Kenxel, who believes that the ancestral Peripatus discharged a small yolkless egg into the water, and that the intra-uterine development was concomitant with the adaptation of the parent to a terrestrial existence. These authors then conclude that the development of yolk in the eggs of P. novae-zealandiae is quite a secondary condition, which, Willey believes, culminates in the return to the oviparous condition observed by Dexdy (App. to Lit. on Onychophora, No. I.) in P. oviparus, which Willey regards as a secondarily acquired habit. — Ed.] 166 OXYCHOPHORA. difficulty of obtaining material, for the eggs, taken from living animals brought to Europe, were sometimes in such a bad condition (Sedgwick, No. 10, Pt. I., Figs. 7-13), that the researches made on them could not be of any great value. Some of the observations also are incomplete, or, as in the case of the South American species examined by v. Kennel, important differences of opinion arise between observers (v. Kennel and Sclatek) which can only be finally settled by further research. A connected description of the first ontogenetic processes and their inter-relationships in the various species is as yet impossible. We shall first consider the development of P. novae-zealandiae which, for the reasons given above, we regard as showing the most primitive condition, and then deal with the African species. The South American species, from what we as yet know of them, seem to claim a position distinct from the others. Peripatus novae-zealandiae. Cleavage is here superficial. The eggs are rich in yolk, and the cleavage-nucleus appears to have a peripheral position. Its division gives rise to nuclei surrounded by islands of protoplasm ; these for the most part also lie peripherally, but single nuclei shift towards the centre of the egg, as may be seen in the figures given by Lilian Sheldon (Fig. 75, Xo. 12) It is no doubt due to the distribution of the nuclei in the yolk that this latter breaks up to some extent into distinct rounded areas (Fig. 76 A), although Lilian Sheldon was not always able to prove that these were regular yolk-pyramids either in origin or form. This break- ing up of the yolk led former observers (Hutton, No. 3, v. Kennel, No. 4, Pt. I.), who could only make observations on insufficient material, to the conclusion that the egg of P. novae-zealandiae underwent total cleavage. According to the descrip- tion given by L. Sheldon, the cleavage - nucleus and the nuclei which first arise seem to lie on the later dorsal side and opposite to the point at which the blastoderm forms. These nuclei increase in number and form a peripheral accumulation (protoplasmic or polar area, Fig. 76 A, a), starting from which, circumcrescence of the yolk takes place (formation of the blastoderm). The active increase in number 75.— Section through the egg of P. novae- zealandiae (after L. Sheldon). In the yolk are I lie nuclei surrounded by areas of protoplasm. TERIPATUS NOVAE-ZEALANDIAE. 167 of the cells and their constant shifting towards the periphery, leads to the almost complete circumcrescence of the yolk as far as a point lying almost opposite the original accumulation of nuclei, where the yolk remains uncovered. Here an ingrowth of cells then takes place, the appearance of an invagination being thus produced (Fig. 77 A and B). The point of invagination is the blastopore (bl), and the base of the depression is formed of yolk in which nuclei can be recognised (Fig. 77 B). The germ-layers do not yet appear to be differentiated from the cell-mass surrounding the blastopore, which represents the rudiment of the germ-band. Miss Sheldon seems to assume that the part of the cell-mass underlying the superficial cell- layer (or ectoderm) yields chiefly mesoderm, while the entoderm arises from the cells lying in the yolk, and which, according to Miss Fig. 76. — Portions of sections through the egg of P. novae-zealandiae, showing the blastoderm- formation (after L. Sheldon). A shows the " polar area " and the cleavage of the yolk. B, the commencement of the circumcrescence of the egg. a, " polar area " ; ds, "yolk segments." Sheldon, arise and multiply by a process of free nuclear forma- tion (I), as she was unable to observe any karyokinesis.* The blastopore lengthens later and then resembles a small groove, the base of which is formed by the nucleated yolk. We here have a resemblance to the condition in P. capentis illustrated in Fig. 84 A. As far as we can gather from the description of L. Sheldon, the process of circumcrescence is regarded by her as an epibolic gastrulation. The yolk-mass, with the nuclei contained in it, would correspond to the entoderm. A study of the figures, however, has compelled us to form another conclusion, which gains in probability from the fact that we are here dealing, as in the case of many * [It is extremely doubtful if there is such a process as free nuclear formation. All recent research on nuclei tends to prove that every nucleus originates from a pre-existing nucleus either by mitotic or amitotic division. — Ed.] 168 ONYCHOPHOIU. Arthropods, with an egg very rich in yolk. Whether the blastoderm is really formed by circumcrescence of the egg starting from one pole, or whether the nuclei contained in the yolk, by shifting to the surface, help to form it, the peripheral accumulation of cells which recurs in the same way in various stages claims identification with the cell-accumulation in the neighbourhood of the blastopore {cf. Figs. 76 and 77). We should then not be obliged to assume gastrulation through epibole, which is unusual in eggs so rich in yolk, but should rather assume that at the point where this cell-accumulation is found a depression (invagination) occurs (Fig. 77 B). Whether the base of this depression is formed of yolk (containing nuclei), or whether a closed archenteron is present, would in this case still have to be decided. If the blastopore lengthened later [cf. also P. capensis) there would be a resemblance to the gastru- lation of the Insects. In these latter, as in Peri- patus, the mouth and anus show a connection with the two terminal points of the long blastopore. In this conception of the cleavage and formation of the germ -layers, it may be noticed that the invagina- tion apparently takes place at the animal pole of the egg. But if it is remem- bered that in P. capensis the brain arises in the im- B mediate neighbourhood of . ^-^^ M- the blastopore, it will be seen that we must rather regard this as a shifting of the vegetative pole, or the region of entoderm-forma- tion, towards the animal pole, than as a gastrulation at the animal pole. The same is the case in the Insecta and in many Crus- tacea (Vol. ii. , pp. 141 and 142). The view here adopted receives a general support from the conditions in the Crustacea, in which the cir- cumcrescence of the yolk (or the formation of the blasto- derm) takes place from one point, gastrulation after- wards occurring in that region (Vol. ii., p. 115). We are unable in the present state of knowledge to obtain any light upon these processes from the ontogeny of P. capensis. &«°?&^>&& ■^mm?o®m? , 'o •sOO ^mmMmmmm L ~-CC i 0°°isr Ocn o % qo oQnO oOOOOor°n Cfpn Qo ??J \S.o00 " Fig. 77.— Sections through the egg of P. novae-zealandiae showing the formation of the blastoderm and invagina tion (after L. Sheldon), bl, blastopore. PERIPATUS CAI>ENSIS. 169 Peripatus capensis. In consequence of the eggs of Peripatus capensis being poor in yolk, their cleavage is apparently total. According to Sedgwick, an animal pole (corresponding to the later dorsal side) can be dis- tinguished from a vegetative pole. Two meridional furrows divide the egg into four blastomeres of equal size, each of which contains a portion of the animal and a portion of the vegetative protoplasm. These cleavage-planes are said not to cut through the whole egg, the blastomeres being united centrally. At a later stage, an equatorial furrow separates the smaller ectoderm-cells (animal pole) from the larger entodermal blastomeres. At the close of segmentation, the cells are very loosely connected, the smaller ectoderm-cells are closely applied together, while the larger entoderm-cells are amoeboid and scattered irregularly within the egg-membrane. The stage which follows may roughly be compared with the blastula. The entoderm- cells draw together and lie directly beneath the smaller ectodermal cells, which then grow round the entodermal .--" elements, a solid (and vJ''(' > ' therefore epibolic) gastrula being thus formed in the course of further develop- ment. The archenteric cavity is said to arise in — E. the entoderm through the formation of "vacuoles"! It opens externally at the point which has remained unaffected by the cir- cumcrescence, and thus corresponds to the blasto- pore. Behind this, an increase in number of the cells of the superficial layer takes place, which leads to a thickening of this layer and then to a separation of the lower layers, that have thus arisen, as the mesoderm. During the lengthening of the blastopore, which soon occurs, and the simultaneous increase in length of the whole embryo (Fig. 84 A), the mesoderm grows forward on both sides of the blastopore and thus yields the mesoderm-bands. The rudiment of the germ-band is thus produced (Sedgwick). Fio. TS.— Section through a 16-celled embryo of ]'. Edwardsii lying in the uterus (after J. v. Kennel). E, embryo; i.Uv\ inner wall of the uterus; Ue, uterine epithelium. 170 ONYCHOPHORA. In the better preserved eggs figured by Sedgwick, large cavities can be seen in the protoplasm, and this leads us to conjecture that, in the normal condition, the eggs might be filled with a more or less fluid mass of yolk. These spaces in the body of the egg are very large, occupying a large part of its interior, so that, when the very unsatisfactory condition of the material investigated is taken into account, we are led to the conclusion that the cleavage may in this case also be superficial. The cavities in the blastula-stage just described would then be filled with yolk, and the gastrula would perhaps be formed by invagination, as was conjectured in the case of P. novac-zealandiac. As we have not personally examined these eggs, wc do not feel justified in giving definite expression to this view, but we cannot refrain from making a conjecture which appears to us so- probable. There would in this case be a certain similarity between the African, and the New Zealand species, especially as it may with probability be assumed that eggs poor in yolk are to be derived from eggs rich in yolk. This last view is held by Sedgwick himself, and in a later treatise (No. 10, Pt. iii.) he calls- the egg of P. capensis meroblastic, because of the central connection mentioned above as existing between the blastomeres. The American species. On account of the small size of their t»W«»flBM coo° Us. f~. E E. ire..._L' ' " r^ -_-> fM "+~- ■"'&•'■&*• w t.Uro. - ' 3P o - Pig. 70.— Sections through embryos of P. Edwardsii together with the uterine wall (after J. v. Kennel). E, embryo; am, amnion; a.Uw, outer wall of the uterus; i.Uvj, inner wall of the uterus; Ue, uterine epithelium [embryonic derivative, Sclater and Willey]. and the connection between these and the wall of the uterus, the American species differ entirely in their develop- ment from the forms we have so far considered. The small eggs, poor in yolk, undergo a total and apparently fairly regular (equal) course of cleavage. The embryo, even at this early stage, appears to obtain nourishment from the uterus, for it increases in size in a marked manner (v. Kennel). When it has reached the 32-cell stage, it forms, according to v. Kennel, a solid cell- mass, completely filling the narrow lumen of the uterus, and thus in close contact with the inner surface of the uterine epithelium (Fig. 78). This THE AMERICAN SPECIES. 171 latter at first consists of very deep cells which, however, under the influence of the growing embryo, seem to flatten. The embryo then, according to v. Kennel, enters into direct connection with this epithelium, this change being accompanied by a peculiar alteration in the shape of the former. The embryo, which is said to give off fluid and to decrease in size, becomes applied to the epithelium as a lenticular cell-mass (Fig. 79 A). The figures show the close nature of the connection between the embryo and the epithelium, the latter may, indeed, occasionally become detached from the wall of the uterus, and may surround the embryo as a special layer (Fig. 79 B, Ue). The central part of the embryo now rises from the surface of the uterus, while the edges, which still remain in contact with the latter, become somewhat approximated through these changes; the em- bryo thus assumes the form of a cap open towards the surface of the uterus (Fig. 79 B). From the surface of the embryo a few amoeboid cells become detached ; some of these be- come applied to the uterine epithelium and, finally, these amoeboid cells unite and give rise to an envelope which sur- rounds the whole embryo, and which has been termed the Fig. SO. — Median section of a pear-shaped embryo of P. Ed- wardsii, with amnion and uterine wall (after J. v. Kennel). om, amnion ; n, umbilical cord ; j).c, embryonic, p.u, uterine placenta ; Ue, uterine epithelium ; I'w, wall of the uterus ; w, point of ingrowth. amnion or serosa (v. Kennel, Fm\ 80, am). The margins of the cap-shaped embryo now become approximated and fuse together, so that the embryo becomes a closed vesicle. The embryo then grows out from the wall into the cavity of the uterus ; its point of attachment narrows and thus forms a stalk (Fig. 80, ?i). A proliferation of cells then takes place at the base of the stalk, this growth being called by v. Kennel the "embryonic placenta." Corresponding to this is a 172 ONYCHOPHORA. circular thickening of the uterine epithelium, which, as the " uterine placenta," enters into close connection with the former (Fig. 80, p.e and p.u). The stalk connecting the embryo with the placenta continues to narrow, and is described by v. Kennel as the "umbilical cord." According to this account the embryo becomes closely con- nected with the wall of the uterus, and a thickening of the connective tissue layer of the latter takes place, causing a constriction of the uterine lumen in front of and behind the region which contains the embryo, thus forming a closed brood-cavity (Fig. 88, p. 179). The amnion and the uterine epithelium are now separated from the embryo by a considerable cavity (Fig. 80). The germ-layers begin to form by an active increase and a con- sequent ingrowth of the cells which commences opposite the point of attachment of the embryo (Fig. 80, w). In comparing the development of P. Edicardsii with that of other species of Peripatus, the point at which this ingrowth takes place will recall the accumula- tion of cells in the blastoderm in P. novae-zealandiae, in which invagination eventually occurs, and which at the same time represents the first indications of the germ-band. In the South American species this point of ingrowth, which in position corresponds to the ventral side of the embryo (the latter is attached by its dorsal surface), must be regarded as the blastopore. From this point the ingrowth proceeds continuously, and fills the whole inner space of the embryo down to the " umbilical cord " (Fig. 80). The cells of the latter have shifted apart, leaving a central lumen, and have become arranged into an epithelium such as is also found all round the embryo, except at the point of ingrowth (Fig. 81). This outer epithelium corresponds to the ectoderm. The further differentiation of the germ-layers is said by v. Kennel to take place through the appearance of a cavity in the more dorsal part of the central cell- mass and the regular arrangement of the cells in its neighbourhood (Fig. 81, ent). The cell-layer thus differentiated, the entoderm, is in this way distinguished from the ventral cell-mass lying at the blastopore, which represents the mesoderm. This latter remains connected with the ectoderm for a long time, even during the later changes of shape of the embryo, and at this point (w) new cell- material is continually produced (v. Kennel, Sclater). In the above description of the first ontogenetic processes in P. Edicardsii, we have followed the account given by v. Kennel liecause the material at his disposal, with regard to quantity and state of preservation, seems to guarantee the reliability of his statements, but it should be mentioned that these processes THE AMERICAN SPECIES. 173 have received another interpretation. Although this latter has been opposed by v. Kennel for very important reasons (No. 5), it has been adopted by Sclater, and seems to have a certain value in so far as it affords some explanation of the peculiar early developmental stages. According to Sclater (No. 9), cleavage gives rise to a blastula formed of large cells, and containing a small cavity (Fig. 82 A). An invagination then takes place in this (pseudogastrula, Fig. 82 B). The invaginated part alone yields the embryo (Sclater)) while the outer layer, by the peculiar growth of its cells, sejiarates from the embryo and becomes very thin, thus forming a membrane which envelops the embryo (Fig. 82 C, a). From the embryo itself another envelope arises, by the splitting off of single cells, this latter corresponding to the amnion described by v. Kexnel. The figures given by Sclater agree on the whole with those of v. Kennel, but they are interpreted by the two authors in an entirely different way. What v. Kennel regards as uterine epithelium is considered by Sclater as an embryonic envelope, for this no doubt is the meaning of his pseudogastrula. Fig. 79 B (v. Kennel) must therefore be regarded as the stage of invagination corresponding to Fig. 82 B (Sclater' s pseudogastrula), and Fig. SO must be interpreted in a similar way. Fig. 79 A, according to this view, should be regarded as an older stage, similar to that represented in Fig. 82 C. Further, the two stages in which the conjectural vesicle has either thin or thicker walls ought not to be unhesita- tingly derived one from the other, as is done by Sclater. Indeed, far stronger proofs must be brought forward for the view adopted by Sclater before it can be finally accepted ; it nevertheless appears to us worth mentioning because it seems best to account for the origin of the embryonic envelopes which are attributed to Pcripatus.* In any case, the two en- velopes which are said to surround the embryo suggest the double embryonic envelopes (amnion and serosa) of the Insecta, all the more that this double embryonic integument may have arisen here as there by the formation of folds in the blastoderm. The position of the embryo in relation to the folds might even correspond to that of the Insectan germ in relation to the embryonic integuments, but we know too little of the * [Willey's observations on the development of P. novac-britanniac (App. to Lit. on Onychophora, No. II. ), in which he finds that the egg gives rise to a large, thin-walled vesicle (trophoblast) with a thickened invaginated embryonic area, tend to support Sclater's views regarding the relations of embryonic envelopes in P. Edwardsii, and are opposed to those of v. Kennel.^To us they appear conclusive on this point. — Ed.] Fig. SI.— Median section through a pear- shaped embryo of P. Edwardsii (after v. Kennel), ent, entoderm ; n, um- bilical cord ; w, point of growth. 174 ONYCHOPHORA. ontogeny of Peripatus to be able to make further comparisons. We must, however, add the description given by L. Sheldon (No. 12) of the earlier stages of P. novae-zealandiae, according to which the embryo proper still within the egg-shell is surrounded by a layer of yolk (the ectodermal yolk of Miss Sheldon). Unfortunately no details as to the significance and origin of this " external yolk " are known, but we might in this also see an embryonic envelope, especially as structures resembling nuclei are found in this outer layer. We are led to adopt this assumption all the more on account of the condition of those Insecta (or Myriopoda) in which the germ-band sinks into the yolk, a condition which finally leads to formation of the embryonic envelopes. In consequence of the Insectan embryonic envelopes arising in this way the embryo here also may be apparently surrounded by an outer layer of yolk, which in reality lies between the embryonic envelopes. The presence of such envelopes derived from folds is not confirmed by what is found iu P. capensis. In this form nothing of the kind has been observed, nor can we assume that such a feature has been overlooked. The ectoderm of P. capensis is only so far peculiar in that it is in the younger stage extraordinarily rich in vacuoles and of a spongy texture, and, in consecpience of its structure, is aide to take in nourishment endosmotically (Sedgwick, No. 10). L. Sheldon connects this structure with the so-called ectodermal yolk of the New Zealand species, but we can- not consider this a happy com- parison. On the other hand, this condition of the ectoderm helps to explain the formation of the external organs of nutrition in the embryos of the American species, whether these are formed direct from the ectoderm of the embryo itself or represent a specially differ- entiated portion of the embryonic envelopes. If, in dealing with the early ontogeny of the different species of Peripatus, we have appeared to dwell almost entirely on the rela- tive probabilities of the processes described, we can only again point out how very little is known with certainty of the earliest develop- ment of this animal. The great importance of this form forces us to take into account statements that are not sufficiently confirmed. We have therefore tried to gather together the facts as yet known into a connected whole, but do not for one moment assume that the conclusions arrived at are final. Fig. S2.— Sections through [embryos of P. 2m- thnrni at various stages (after Sclater). E, embryo ; a, outer, i, inner cell-layer of the embryo ; m, (cuticular) membrane bounding the uterus internally ; v, placenta-like growth of cells. THE DEVELOPMENT OF THE EXTERNAL FORM OF THE BODY. 175 2. The Development of the External Form of the Body. In spite of the variations found in the first ontogenetic stages of the several species of Peripatus, the latter differ very little from one another in the development of the external form of the body. In the following descriptions we shall first deal chiefly with P. capensis, which was very carefully examined first by Moseley (No. 6), then by Balfour (No. 1), and later by Sedgwick (No. 10, Pt. i.). P. capensis. It has already been shown, in describing the forma- tion of the germ-bands, that a thickening of the blastoderm arises behind the lengthening blastopore by the proliferation of cells, this spot being recognisable externally as an oval area (Fig. 83). "We saw that at this point the mesoderm originates, and extends forwards in the form of two bands, to the right and left of the blastopore. In each mesoderm-band segmentation takes place, a segmentation which may in all respects be compared to that of the Annelida. For instance, at the anterior ends of the two bands, cell-complexes are cut off and commence, by the formation of cavities, to form the primitive segments (Fig. 84 A and Z>), fresh rudiments being continually added posteriorly. At the posterior end of the blasto- pore the mesoderm-bands pass over into an undifferentiated cell-mass. During the differentiation of the mesoderm-bands another im- portant change takes place in the embryo. The lips in the middle region of the elongated blastopore approach one another and fuse, so that the only remains of the blasto- pore are an anterior and a posterior --'"" ^^v. aperture (Fig. S4 A and B). These /' "\ two apertures are henceforward retained / ' •- , .©, '-.'■'• '■% (C and D), giving origin (in connection j: B w~~. with ectodermal invaginations) to the mouth and the anus. \ ■ / ; .'*7 The next changes in the embryo \ • j &•. take place as follows : as the differ- \ entiation of the primitive segments Continues, the first Of these Shift Fig. S3.— Embryo of P. capensis (after further forward, and, in addition to ^IZZ W'blastopore; w' zone this internal segmentation of the em- bryo, an outer segmentation now appears (Fig. 84). At the anterior end the cephalic lobes begin to appear, and it is to be specially noted that, as rudiments, they show great resemblance to the body-segments. The posterior end of the hitherto straight embryo curves round 176 ONYCHOPHORA. K. B. M. ventrally, thus covering the posterior aperture derived from the constriction of the blastopore (Figs. 84 and 85). Before describing the further development of the embryo, we must glance at the corresponding processes in other species of Peripatus. The observations recorded above on the development of the external form have dealt chiefly with the shaping of the ventral surface, this being first developed as two symmetrical halves. We are here reminded of the development of the eggs of the Oligochaeta and Hiru- dinea that are rich in yolk, and still more of that of the eggs of the Myriopoda, Insecta, and Arachnida. In Peripatus, as in these, a germ-band forms. Its composition out of two halves is still more distinct in P. novae-zea- landiae. In this form, in consequence of the large size of the egg, caused by the abundance of its yolk, the two halves of the germ-band lie somewhat far apart, separated by a ventral protrusion of the yolk-mass covered with ectoderm and entoderm (Fig. 86 A and B), so that a kind of ventral yolk-sac arises resembling the one met with in the Araneae (p. 54). While already well developed at the anterior end, the two halves of the germ-band become less and less differentiated posteriorly, and end near the blastopore in the as yet undifferentiated cell-mass (primitive streak of English authors). At first sight there appears to he a fundamental difference between P. capcnsis with P. Edwardsii and P. novcte-zealandiae regarding the position, relative to us. - ra- il. m.- Fio -Embryos of P. capensis to illustrate the closing of the blastopore, the segmentation of the mesoderm, and the flexure of the embryo (after Balfour and Schimkewitsch). a, anus ; U, blasto- pore ; m, mouth ; us, primitive segments ; w, zone of growth. THE DEVELOPMENT OF THE EXTERNAL FORM OF THE BODY. 177 the anus (? = blastopore), of this undifferentiated cell-mass (primitive streak). In the two first species this zone of growth is undoubtedly situated behind the anus (i.e., behind the blastopore = anus in P. capensis, Fig. 84 A, w, and in the region of the blastopore and behind the anus in P. Edwardsii, Fig. 89 A), whereas in P. novae-zealandiae this zone of growth appears, from a superficial examination, to be situated in front of the anus (Fig. 86 A), at least the two halves of the germ-band bend forward and unite in front of that aperture. According to L. Sheldon this is only an apparent difference, for in sections of P. novae-zealandiae through this region the zone of growth of the mesoderm (primitive streak) is found to be situated, as in P. capensis and P. Edwardsii, behind, not in front of, the anus. The possibility of the zone of growth being situated in front of the anus is of interest when we make a comparison between the embryos of Peripatus at this stage and those of the Annelida. Such a comparison made between the embryo of P. novae-zcalandiae (Fig. 86 A) and that of Clepsine among the Hirudinea (Vol. i., Fig. 152, p. 322) reveals a striking similarity between the two, especially in the configuration of the mesoderm-bands which, in the Hirudinea, however, unite in front of the anus, a condition which, it is true, is suggested from a superficial examination of the embryo of P. novae-zealandiae, but which is not substantiated by the investiga- tion of sections, and one which would, moreover, seem improbable from a comparison with the two other species of Peripatus mentioned above. Further, in the Hirudinea, there appears to be no connection between the blastopore and the anus, which makes a comparison with Peripatus more difficult than would otherwise be the case. In other Annelida, however, the primitive mesoderm- cells are met with at the posterior edge of the blastopore (Vol. i., pp. 264 and 283), but the condition is different from that in Peripatus in so far as the blastopore does not pass direct into the anus. In view of these possible differ- ences between the species of Peripatus inter se and of the possible resemblances to some Hirudinea, a renewed investigation of the relations of the growing zone in P. novae-zealandiae, which must be regarded as the most primitive form [one of the most specialised according to Willey and v. Kennel], is much to be desired. The South American species, in conse- quence of the small size of their eggs and of the connection of the latter with the wall of the uterus, have a different form in the early stages. Our description of the embryo ceased at a stage in which it was somewhat pear-shaped (Figs. 80 and 81). From this stage it passes into the mushroom stage (Fig. 87), the embryo proper gaining in size as compared with the umbilical cord, through extending in both directions at right angles to the axis of the cord (Fig. 87). These two directions correspond to the length and the breadth of the embryo. Growth takes place at first chiefly in the first of these two directions, Fig. 85. — Embryo of P. capensis (after Balfour and Sedgwick), a, anus; m, mouth. 178 ONYCHOPHORA. af„ with the result that the embryo becomes elongate (Fig. 88). Dorsally, it is attached by the umbilical cord, while the ventral surface is free. The blunt end becomes the head, and the pointed end the posterior extremity. The blastopore, which is probably secondarily displaced, lies quite near the latter (Fig. 89 A, bl). The space between the blastopore and the umbilical cord is much longer than that between the latter and the anterior end, since, starting from the blastopore, new cell-material is continually being produced posteriorly. Two mesoderm-bands, divided up into the primitive segments, are present as in P. capensis, but, in conse- quence of the smaller size of the egg, the paired nature of the germ-band is not so distinct, although it can be recognised here also (Fig. 89). The mouth arises in a position corresponding to that in P. capensis, but quite independent of the blastopore, the latter, as a small and shallow depression, n J9f„ having come to lie ^ai. at an early stage at the posterior end (Fig. 89 A, bl). The anus also is said by v. Kennel to arise independently of the blastopore. It arises in front of the latter as a slit-like depres- j sion (Fig. 89 A, a). J If the accounts given of P. capensis by Sedg- wick prove correct, we shall have to conclude that, in P. Edwardsii also, the oral and anal apertures were originally connected with the blastopore, since the position of the two apertures is similar to that in P. capensis. With regard to the position of the growing zone, P. Ediuardsii, according to v. Kennel, agrees entirely with P. capensis and P. novae-zealandiae ; for since this zone proceeds from the blastopore, and the latter lies behind the amis (Fig. S9 A), the undifferentiated cell-mass is also found behind it. The connection of the embryo with the mother must here again be referred to. According to v. Kennel, the embryo is connected with the mother by means of the umbilical cord, as well as by the Fio. S6. — Embryos of P. novae-zealandiae. A, ventral, and B, lateral aspect (after L. Sheldon), a, anus ; at, antennae ; ex, limbs ; m, mouth. THE DEVELOPMENT OF THE EXTERNAL FORM OF THE BODY. 179 Fig. 87. — Mushroom-shaped embryo of P. Edwardsii in the brood-cavity (after v. Kennel), b, brood-cavity ; e, embryo ; n, umbilical cord ; p, placenta ; it, lumen of the uterus. embryonic and uterine placenta (Fig. 80, p. 171, and Figs. 87-89) The great development of these organs shows that, in the younger stages, they contribute to its nourishment. They degenerate later, and the embryo is then, like that of P. capensis, nourished by the uterine secretion. In conse- quence of the close organic connection of the embryo with the uterus, the former is unable to shift from its position. The embryo, firmly enclosed in its brood-sac (Fig. 88), can only move on into the vagina by the growth of the parts lying between the ovary and the brood-cavity itself, and by the gradual absorption of the posterior parts. When the embryo which lies nearest the vagina passes over into the latter, its brood-cavity must be com- pletely absorbed before the next embryo can reach the vagina. The extrusion of the embryos in the South American species of Penpatus closely resembles the passage of Insectan eggs from the oviduct into the efferent apparatus. There also the empty follicle left after the expulsion of the -egg is completely absorbed before the next ■egg is able to pass out. The further development of the external form of the body consists ■essentially in the lengthening of the body, the marking off of the head and trunk, and the appearance of the limbs •and sensory organs. It agrees on the whole in the different species, so that separate accounts are here unnecessary. An important change in the form of the young embryo is brought about by the great development and marking off FlG. ss.— Embryo of p. Edwardsii in Of the cephalic segment from the trunk the brood-cavity (after J .v. Kennel, r ° _ from Langs Text-book of Comp. (Figs. 86 and 89). This change, which <.). e, embryo; ep, placenta. 180 ONYCHOPHORA. occurs early, is introduced by the shifting forward of the first pair of primitive segments to the extreme anterior end of the body, where they become considerably enlarged. A pair of large swellings (cephalic lobes) thus arise at the anterior end ; these soon become marked off from the body by a transverse furrow, and thus constitute the cephalic segment. On the ventral side of these lobes is the oral aperture ; on the dorsal side a pair of prominences appear (Fig. 86 A and B) which soon increase in size and become recognisable as the rudiments of the antennae. In P. capensis these are said to appear before the limbs (Sedgwick), but this distinction seems to be of no great significance ; in P. Edwardsii the antennae are said to appear simultaneously with the rudiments of the truncated legs, which they closely resemble. They are, however, distinguished from the latter by their more dorsal and pre-oral position (Figs. 86, 90, and 91). In front of the rudiments of the antennae, and lying more medianly, there are, at an earlier stage, two small prominences (Fig. 90, x), which shift later towards the anterior margin of the head (Fig. 94 A and B). These prominences, which were observed by v. Kennel in P. Edivardsii, and the nature of which is as yet unknown, can still be recognised at a later stage than that depicted in Fig. 94 B, and disappear R. B. --m. from view only when folds begin to form in the cephalic integument. Wc shall refer to them again at the end of this section (p. 187). The limbs arise as latero- ventral outgrowths of the segments consecutively from before backward (Figs. 86, 90, and 91). The segmentation of the body is brought about chiefly by the outgrowth later- ally of the primitive segments. The embryo, especially in its lateral parts, thus appears notched (Figs. 86 and 90). The paired nature of the germ-band is still indicated by the presence of a median ventral furrow (Fig. 89). This especially applies to P. Edivardsii, in which also the limbs appear later than in the African and Australian species. This retardation is no doubt due — U. Fig. 89.— Embryo of P. Edwardsii. A, ventral, and B, lateral aspect (after v. Kennel), a, anus; hi, blastopore; m, mouth; n, umbilical cord. THE DEVELOPMENT OF THE EXTERNAL FORM OF THE BODY. 181 to the modified method of development within the uterus, the close connection of the embryo with the wall of the uterus leading to the later development of its external form. In P. novae-zeal andiae the limbs are to be found while the two halves of the germ-band are still far apart (Fig. 86), and in P. capensis also they appear early. The embryo, at an early stage, becomes curved, and, as the body ■elongates, its posterior end becomes rolled up ventrally (Fig. 85, p. 177), this being determined either by its position within the egg- shell or (secondarily as in P. Edwardsii) within the brood-cavity. In P. Edwardsii the posterior end forms several coils. The posterior extremity of the embryo of P. capensis is also at first bent in towards the ventral surface of the body (Fig. 85), but subsequently this pos- terior region grows parallel with that surface, the bend being retained at the middle of the body, and the embryo lies in the egg-envelope in such a way that the anterior and posterior halves of the body are almost parallel to one another, the head touching the posterior end. In P. novae-zcalandiac, in a stage earlier than that illustrated in Fig. 86 A and B, a ventral flexion apparently occurs in the embryo, the latter consequently assuming a curved form, but it soon straightens again to some extent, and retains the form shown in Fig. 86 A and B (L. Sheldon, No. 12, Pt. i. ). Here also the two halves of the germ-band are at first very far apart, as may be seen from Fig. 86 A and B. In keeping with the unspecialised external form of the adult Peripatus, the further development of the embryo is very simple, and, apart from the anterior region of the body, presents no specially noteworthy features. The formation of the limbs continues in the manner above described (Fig. 91), until the final number is reached. Where the two halves of the germ-band lie far apart, as in P. novae-zeal andiae, they eventually shift together to form the ventral surface, a process which is assisted by the gradual absorption of the yolk. The dorsal surface at the Fig. 90. — Anterior part of an embryo of P. Edicardsii, dorsal aspect (after v. Kf.nxel). at, antenna ; k, max- illary segment ; op, segment of the oral papillae ; p, first adult trunk segment ; x, prominence in front of the antennal rudiment (c/. pp. ISO and 187). same time assumes its final shape. The annulations of the body, and the papillae which are seen on its surface in the adult condition, appear in the form of folds and slight elevations of the epidermis. 182 ONYCHOPHORA. The terminal region of the body, up to the time when the adult form is assumed, is almost button-shaped. At its lower side, either in a depression (as in P. Edicardsii) or on a papilla, as in P. capensis, lies the anus. Two slight outgrowths, the anal papillae, which apparently belong to the terminal section, must be regarded as rudi- ments of limbs, and thus indicate a true segment. The limbs themselves have assumed their adult form, being better marked off' from the body, and exhibiting a ringed appearance not unlike segmentation At their free ends the two cuticular chitinous claws arise. The limbs have shifted from their former more ventral position to their final position between the dorsal and the ventral surface. With regard to the position of the anus it must be mentioned further that, in consequence of its being found in front of the growing zone, it must be related to a true segment. In various drawings made by v. Kennel and Sedgwick of sections cut through the anal aperture, well developed primitive segments are seen round the terminal region of the intestine. We must then in any case- assume a shifting forward of the anus which originally belonged to the terminal region of the body. The relation of the anus to the segmentation of the body in the adult does not seem satisfactorily settled, nor is it clear whether it subsequently shifts out of the segmented region to the extreme end. The development of the anterior region of the body is less simple than that of the trunk. Complications arise in the former through two other segments besides the actual cephalic segment being drawn into the formation of the adult head, and through the corresponding modification of the append- ages of these segments. We thus find in Peripetias a state of things already met with in the Crustacea, and still more closely re- sembling conditions found in the Arachnida, Myrio- poda, and Insecta. In the cephalic segment, the rudiments of the antennae have undergone alteration; they have lengthened considerably, and rings like those on the limbs have appeared on them (Fig. 91, at). The Fio. 91.— Embryos of P. capensis of different ages (after Sedgwick), at, antenna; aw, eye; /, fold, contributing to the formation of the buccal cavity; k, jaw ; op, oral papilla ; p,-pti,, first three pairs of limbs. THE DEVELOPMENT OF THE EXTERNAL FORM OF THE BODY. 183 eyes (azi), as slight ectodermal depressions, are already present as rudiments (P. capensis), situated somewhat ventrally to the antennae. In P. Edwardsii they arise immediately behind the bases of the antennae. The further development of the mouth and of the two pairs of limbs connected with it is of special importance. The anterior aperture remaining after the partial closing of the blastopore does not directly give rise to the mouth, but becomes carried inwards by an invagination of the ectoderm, the stomodaeum, and thus forms the aperture between the latter and the enteron.* Neither does this second aperture represent the mouth of the adult, for it becomes covered by various outgrowths of the ectoderm, which form above it a secondary buccal cavity. This process commences by the appearance of a fold on the outer side of the limb next in order to the antenna (Figs. 91 and 92, 7c); this fold is closely applied to the limb, and is continued posteriorly along the ventral surface of the embryo (Figs. 91 and 93, /, and Fig. 92, p). It appears notched, and, in P. Edwardsii, is repre- sented by a series of papillae lying one close to the other (Figs. 92 and 94). In later stages these two folds shift closer towards the oral aperture, and thus press the limb-rudiments that lie on the inner side of them towards the mouth. As the folds grow still higher, these limbs, to- gether with the stomodaeal aperture, come to lie in a cavity, the adult buccal cavity (Fig. 94), the limbs themselves becoming the jaws of the adult. The distal part of each of the limbs, at the time when the formation of the buccal cavity just described begins, appears deeply notched, and the two strong chitinous teeth arise at this part (Fig. 94 A and B). These terminal teeth, which are to be compared with the double claws on each of the legs, prove, even in the adult, that these jaws are true limbs. Several other structures contribute to the complete development of the buccal cavity. Between the cephalic lobes, and ventrally to them, a somewhat long prominence arises (Fig. 94, ol), which * Cf. below, p. 196. Fig. 92.— Anterior portion of an. embryo of P. Edwardsii, seen from the ventral side (after v. Kennel, from Lang's Text- book of Comp. Anat.). Tc, jaw ; no, aperture of the nephridium belonging to the segment of the oral papillae (op) ; p, pa pillae of the folds which sur- round the jaws laterally. 184 ONYCHOPHORA. lies directly in front of the sharp edge of the stomodaeal aperture, and thus, when that aperture is walled in by the lateral folds, shifts with it into the cavity thus formed (Fig. 94 B). The folds then unite in front of this unpaired papilla, the upper lip of v. Kennel (Fig. 95). The posterior unnotched continuations of the lateral folds form the posterior boundary of the buccal cavity, on the floor of which the primitive aperture of the stomodaeum now lies surrounded by the jaws and the upper lip. From the above it will be seen that, in Peripatus, there are three distinct apertures, each of which in turn must be regarded as the oral aperture: (1) the primitive blastopore mouth (Fig. 84 D, m), which persists in the adult as the opening between the oesophagus and the stomach-intes- tine; (2) the stomodaeal mouth (Fig. 93, m), which in the adult puts the buccal cavity into communication with the pharynx; and (3) the external opening of the buccal cavity, which functions as the mouth in the adult, and is formed by the concrescence of two ectodermal folds. The shifting forward of the lateral folds towards the oral aperture has also caused the ventral organs of the first two segments to shift into the buccal cavity (Fig. 94 A, vol and vo2). We shall refer to these again later. Another pair of folds exactly like those which have walled in the oral aperture are sometimes present, according to v. Kennel, on the outer side of the lateral folds, but these do not seem to be of constant occurrence. They, however, seem further to support the view, which appears very probable, that the folds found near the mouth of Peripatus do not represent limb-rudiments, as has been conjectured by Moseley. The third pair of limbs are less closely connected with the mouth than are the jaws, for while they also shift towards the oral aperture, they remain outside the lips of the buccal cavity (Fig. 95, op). Apart from the fact that no chitinous hooks develop on them, they retain to a greater degree the character of limbs. They are early distin- guished from the other limbs by their greater development (Fig. 91 B, op). These limbs are known as the oral papillae, at the tips of which the slime-glands open. In the adult, these papillae lie as far forward as the jaws (Fig. 95), and the segment to which they belong must therefore be reckoned as a cephalic segment. Three Fig. 93.— Cephalic part of an embryo of P. capensis (after Sedgwick). at, antenna ; /, oral fold ; k, jaw ; in, stomodaeal aperture ; op, oral papillae ; sp, aperture of the sali- vary gland ; vo, aperture of the ventral organ. THE DEVELOPMENT OF THE EXTERNAL FORM OF THE BODY. 185 segments beside the primary cephalic segment, viz., those of the antennae, of the jaws, and of the oral papillae thus take part in the formation of the head. The head of the embryo, in consecpience of the great development of the cephalic lobes, at first appears very large in proportion to the rest of the body (Figs. 86 and 89). In the course of development, however, it decreases in size, the mouth shifts more to its anterior end, and the form of the adult is thus practically attained. R. B. g+no Fio. 94. — Anterior parts of embryos of P. Edwardsii, ventral aspect (after v. Kennel). at, antenna ; /,, the folds broken up into papillae, surrounding the mouth; /,, the folds lying outside of/ ; g+vo, ganglion and ventral organ of the cephalic segment with the slit-like depression of the ventral organ ; k, jaws ; ol, upper lip ; op, oral papillae ; j\ , p,/, first and second pairs of legs ; vox~vo4, ventral organs of the jaws, oral papillae, and of the first two trunk-segments; vo., is divided into an anterior and a posterior part; x, prominences in front of the antennal rudiments (pp. ISO and 1S7). The young are born provided with the complete number of limbs. Their development, according to authors, lasts unusually long (Sedgwick, No. 11). P. novae-zecdandiae is said to require eight to nine months for its development, and P. capensis thirteen months (?). The umbilical cord, which in the South American species connects the embryo with the placenta, at the time when the embryo lengthens and coils up its posterior end, changes and finally degenerates ; its lumen closes first near the embryo. The embryo is now nourished by swallowing the surrounding albumen, 186 ONYCHOPHORA. a method of feeding which also occurs in the embryos of P. capensis ; in addition to this there is a kind of endosmotic inception of nutritive fluid. Interpretation of the cephalic appendages of Peripatus. The nature of the two posterior pairs of cephalic appendages of Peripatus cannot be doubted. They correspond to the limbs of the trunk, and might Avithout further question be assumed to be limbs which, when two (primary) trunk -segments were fused with the head, were transformed into jaws and oral papillae. This is not the case with the antennae, which are distinguished from the limbs of the trunk by their dorsal and pre-oral position. In this respect they entirely agree with the antennae of the Myriopoda and the Insecta, with which we a consider them homo- logous. The antennae of Peripatus seem un- doubtedly homologous with those of all the other Tracheata, but not with those of the Annelida. The an- tennae, not only of Peripatus, but of the Myriopoda and In- secta, have been com- pared with regard to their position to the cephalic tentacles of the Annelida, which are found (pre-orally) in the cephalic seg- ment, occupying the same position with relation to the neural plate as do the antennae of the higher forms with relation to the brain. The manner in which the antennae of Peripatus originate, however, seems to us to tell against such a comparison. The antennae, both as rudiments and when developing, show great agreement with the trunk-limbs (Figs. 91-94), a fact which is strikingly evident in the figures given by Sedgwick and v. Kennel. Like the limbs, they are externally ringed, and a process of the primitive segment runs into each of them, so that they too are hollow cones. Indeed, a canal is said to run from the primitive segment of the antenna Fig. 05.— Head of P. Edwardsii, ventral aspect (after Sedgwick, from Lang's Text-book of Comp. A net.), a, antenna (the greater part of which is removed) ; op, oral papilla. In the buccal cavity are the double jaws. The cavity itself is surrounded by the folds cut up into papillae. THE DEVELOPMENT OF THE EXTERNAL FORM OF THE BODY. 187 to the exterior, and this would correspond to the nephridial canal of the trunk -segments. In fact, there is such close agreement between the antennae and limbs as rudiments, that it is difficult to believe that they are essentially different structures. We should feel inclined rather to attribute to them the same origin, and to assume merely that the antennae had shifted further forward. This conjecture is supported by a comparison of the antennae of the Insecta with those of Peripatus. The former also as rudiments show not only in form, which would there be less remarkable, but also in position the closest agreement with the (primary) trunk- appendages, indeed, to begin with they even lie post-orally (Fig. 147). We might conclude from this that the antennae of Peripatus and those of the Insecta were homologous structures, but that they could not be compared with the cephalic tentacles of the Annelida, in other words, that they were originally appendages of the primary trunk and not of the primary cephalic region. If we accept this view we shall have to assume that the primary cephalic region has greatly degenerated, and that a primary trunk-segment (the first) has to a certain extent taken its })lace. An indication as to how and why this happened is to be looked for in the fusion of the other and undoubted trunk- segments in the adult head. The utilisation of the anterior limbs as mouth- parts was accompanied by their partial transformation into sensory organs (palps of the Insecta), and the final preponderance of one pair as feelers. Again the brain would in this case have to be reckoned as belonging to the first (primary) trunk-segment, and could not be derived from the neural plate alone. This view, however, presents no difficulty when we see how, in Peripatus, the ganglia of the maxillary segment passes from a post-oral to a pre-oral position and is absorbed into the brain (p. 193). In the Crustacea the ganglia of the second antennae undergo corresponding changes. The entire filling up of the so-called cephalic segment in Peripatus by a regular pair of primitive segments with unbroken epithelial walls agrees with what is found in a trunk-segment, but not with the condition of the cephalic region in the Annelida. If the primary cephalic segment which, in the Annelidan Trqchophore, carries the cephalic tentacles, has really undergone degeneration, we might expect to find traces of this fact. The two small prominences which appear in front of the antennal rudiments, the significance of which is still obscure, might be regarded as vestiges of this kind (Figs. 90 and 94, x). We might conjecture that they are possibly vestiges of the primary Annelidan tentacles. This interpretation of them, which appears to us very plausible, also leads to a striking agreement with the Crustacea. In homologising the cephalic appendages of the latter (Vol. ii., p. 166), a similar view was adopted, the same significance being ascribed to the frontal sensory organs as is now given to the small prominences (x) in front of the antennal rudiments in Peripatus. It cannot be regarded as altogether improbable that the adult Peripatus should still retain vestiges of this organ, the agreement of which with the frontal sensory organs still functioning in many Crustacean larvae would be stil 188 ONYCHOPHORA. more striking. The prominences now under consideration are, according to v. Kennel, retained for a long time, and have perhaps escaped observation in later stages owing to the development of papilla-like prominences of the integu- ment, such as occur in great numbers in the adult. St. Remy (No. 8) describes and illustrates a paired ganglionic swelling on the brain of the adult Peripatus ("formation de nature inconnue"), which in position corresponds to the two prominences on the head of the embryo, and which might well be regarded as the lobe of a primary tentacle-nerve. In the posthumous works of Balfour also, similar structures are described as pairs of nerves (running to various points of the dorsal surface), and of these one might belong to such a sensory organ. We cannot refrain, in this connection, from calling attention to the sensory organs found in the cephalic region in many Myriopoda (e.g. , Lithobius, Polyxenus, and Glomeris), the innervation of which is said to take place from the "optic thalamus" (Tomosvaky, No. 22, p. 760). We must however impress upon the reader that the actual material required for a successful comparison of this peculiar sensory organ with the frontal organ of the Crustacea, or with the still insufficiently investigated prominences of Peripatus has not yet been obtained. In the younger embryos of Peripatus (such as that illustrated in Fig. 91 A), the change of position of the antennae, if these are considered as proceeding from limbs, is not very marked, especially as compared with the corresponding change that takes place in the Insecta. The position of the eyes in Peripatus is less easily reconciled with this view. The eyes lie further back than the antennae, close to that part of the brain which must be derived from the first (primary) trunk-segment. The eyes, however, may well be ascribed to the primary cephalic segment, especially as, in Peripatus, they agree with the eyes of the Annelida rather than with those of the Arthropoda. This can only be explained as having been brought about by the shifting of the various parts which participate in the formation of the head. 3. The Formation of the Organs. The Ectodermal Structures. The Integument. The ectoderm forms a single layer of cubical cells over the greater part of the body of the embryo. In P. capensis these cells, especially on the dorsal surface, are said by Sedgwick not to be sharply demarcated externally, and to exhibit a spongy structure. Sedgwick on this account ascribes to them the capacity for absorbing fluid nourishment, and believes that the placenta described by v. Kennel might arise as a more specialised ectodermal organ for taking in nourishment. The changes undergone by the ectoderm when trans- formed into the adult integument are not very important. The delicate cuticle which occurs in the adult is secreted externally. At some points, e.g., on the ventral side of the limbs, the ectoderm becomes multilaminar and here secretes a thicker layer of chitin, and this is also the case at the distal extremities of the limbs where the claws are formed. THE NERVOUS SYSTEM AND THE VENTRAL ORGANS. 189 The Nervous System and the Ventral Organs. The nervous system and the ventral organs arise from two massive thickenings of the ectoderm formed by the active increase in number of the cells on the ventral side of the cephalic and primary trunk- segments. The two longitudinal swellings thus produced appear at the time when the limbs become sharply marked off from the body, and develop from before backward. According to Sedgwick, each of these swellings passes directly into a corre- sponding thickening of the cephalic (antennal) segment, but this, according to v. Kennel, is not the case, the swellings ending bluntly where the cephalic section commences (Fig. 92), so that the part of the central nervous system pertaining to the cephalic segment arises separately from the rest.* A much slighter thickening of the ectoderm does, however, occur, according to v. Kennel, between the cephalic and the trunk portions of the longitudinal swellings at the time when the latter appear. And this, since it denotes the formation of a commissure, might be regarded as indicative of a continuity between the cephalic and the trunk portions of the longitudinal swellings. This question as to the continuous origin of the cephalic and the trunk portions of the central nervous system has already been discussed in connection with the Annelida (Vol. i., p. 287). It was not indeed there finally settled, but it is in connection with them that a decision of the question can best be expected. The paired thickening on the ventral surface just described gives origin not only to the nervous system, but also to the ventral organs (v. Kennel, No. 4). Transverse sections of the embryo show that the thickening projects both externally and internally (Figs. 100, p. 200, and 101, p. 202). In the middle of the cell-mass Avhich forms it, a horizontal fissure then arises extending from before backwards and separating the mass into an outer and an inner portion (Fig. 100 B). The inner mass of cells represents the rudiment of the nervous system (»), the outer, remaining in connection with the epidermis, represents that of the so-called ventral organs (vo), the development and significance of which must now be discussed. The ventral organs. As the cleft, which in each segment divides the rudiment of the nervous system from that of the corresponding ventral organ, is interrupted by intersegmental cellular strands con- necting its two walls (Fig. 101 B, p. 202), a segmentation of the ventral organs takes place which is visible even externally. The connecting strands between the ventral organ and the nerve-cord, which occur between the consecutive pairs of limbs, are retained * An entirely distinct origin for the brain and for the ventral chain of ganglia cannot here be asserted, inasmuch as the ganglion of the maxillary segment is also drawn into the brain, as will be shown presently. On this account and also because of the relation above pointed out, of the antennae to the limbs, there is room for doubt as to the true cephalic nature of the brain in Peripatus. 190 ONYCHOPHORA. until the embryo is mature, and are even found in the adult (v. Kennel). As development proceeds, the ventral organs shift together and finally unite in the middle line ; they become flattened and a slight depression is seen on their outer surface. Whereas at first they were very massive (Figs. 100 and 101 B), they now appear much smaller as compared with the size of the embryo (Fig. 102, p. 205). As the embryo develops, they become less and less con- spicuous, and, in the adult, are represented merely by a small unpaired follicular depression of the epidermis situated medianly between the bases of the limbs on each segment (v. Kennel), and until recently overlooked. an. - Fig. 96.— Transverse sections through the head of an embryo of P. Edwardsii (after v. Kennel). In A, only half the section is drawn, an, antennal nerve ; n, brain (consisting of cell- and iibre-substance) ; us, primitive segment of the head ; co, ventral organ. The ventral organs of the anterior segments differ from the rest. Those belonging to the segment of the oral papillae, as well as those of the maxillary segment, are drawn into the buccal cavity, and can still for a time be recognised on its floor (Fig. 94 A, vo1 and vo2). Of these, the two posterior unite to form the ventral wall of the oesophagus, while the anterior organs remain distinct. Consequently each of these latter develops further independently, and in both the external depressions are more marked than the other trunk ventral organs (thus retaining, according to v. Kennel, the more primitive character). This is still more the case with those structures which must no doubt be regarded as ventral organs of the cephalic segment. These are two deep epidermal depressions lying near one THE NERVOUS SYSTEM AND THE YEXTRAL ORGAN-. 191 another on the ventral side of the cephalic segment (Fig. 96 B, which have arisen, like the ventral organs of the trunk, by the splitting up of the ectodermal layer into an outer and an inner portion, and the subsequent invagination of the former. These depressions, -which at first are quite open, but later close almost completely, can be recognised even in surface view, first as pits, and later as irregular slits on the ventral surface of the cephalic lobes (Figs. 93, p. 18-1, and 94 B, p. 185). At a later stage the ventral organs close completely and lose their connection with the epidermis. As the two vesicles sink down deep into the mass of the brain (Fig. 96 B) and thus become closely connected with this latter, it is clear that, when the brain becomes small in comparison to the head and shifts to the dorsal side of the latter, the vesicles follow the brain, and remain connected with it in the form of a thick-walled vesicle, the so-called brain-appendage of Peripcdu*. The ventral organ of the cephalic segment, if, indeed, this vesicle is to be considered as such, would be distinguished from those of the trunk by the complete loss of its connection with the epidermis. The significance of the ventral organs has until now remained obscure. Their great development in the early part of embryonic life, and their reduction in the adult, indicates that they are organs which were more highly developed in the ancestors of Peripatus. From their position it might be concluded that perhaps the greater part of the ventral surface, by means of its strong ciliation, functioned for locomotion, like the ventral ciliated area of the Annelida. The connection of the ventral organs with the nervous system is not surprising, considering the origin of the latter out of these ectodermal masses. It is possible that during ontogeny the ventral organs may be concerned in supply- ing the cell-material for the development of the ventral chain of ganglia, v. Kexxel's statement that the gradually diminishing cell-mass of the ventral organs is used in the further development of the epidermis seems in keeping with the original connection of these organs with the ectoderm, especially as, with the exception of the ventral organ of the cephalic segment, the greater j>art of each organ retains this connection. The similarity between these cephalic ventral organs and the "cephalic pits" of the Arachnida, which are in the closest connection with the nervous system (pp. 12 and 53), is very striking. Fig. 96 B shows how closely the "ventral organs" of the head of Peripatus become applied to the rudiment of the brain, and comparison of Figs. 93 and 94 B, with Fig. i C, p. 6, Fig. 7, p. 10, and Fig. 28, p. 52, shows that a marked agreement exists even in the external position of the depressions in the two groups. In the present state of our knowledge, however, we are not justified in carrying this comparison further.* The Nervous System. When the rudiments of the two longi- tudinal nerve-trunks first separated from those of the ventral organs, * [Willey (App. to Lit. on Onychophora, Xo. II.) finds what he believes to be ventral organs persisting in the adult on the anal segment. — En.] 192 ONYCHOPHORA. a thin layer of fibres had already appeared on the dorsal side of the former. As development proceeds this gradually thickens (Figs. 101, p. 202, and 102, p. 205). The position of this fibrous layer on the mass of the ganglion-cells is practically retained in the adult, for even there the fibrous mass lies dorsally to the ganglionic cells (Balfour, No. 1), and only a very few of the latter attain a position dorsally to the fibrous mass. This feature must be regarded as a primitive one. In more highly differentiated forms, e.g., the Crustacea and the Arachnida, the fibrous mass is indeed peripheral when it first appears, but is soon covered by ganglionic cells, and comes to lie within the mass of the ganglion. It has already been pointed out, in connection with the Crustacea (Vol. ii., p. 162), that the appearance of the fibrous substance on the inner periphery of (i.e., dorsally to) the ventral strands might represent a primitive condition. The transverse commissures which are, in Peripatus, found in large numhers connecting the longitudinal nerve-trunks, grow out from the latter like the peripheral nerves, which are said to be formed by the outgrowth of nerve- fibres (v. Kennel). The brain arises in a manner agreeing with the origin of the rest of the nervous system, but certain complications are caused by the fact that it is formed by the fusion of the ganglia of two distinct segments. The separation of the ganglionic rudiment of the cephalic segment from the epidermal thickening (ventral organ) takes place somewhat as in the trunk-segments, but the fibrous tissue here lies much deeper in the mass of ganglionic cells, and is partly covered dorsally by the latter (Fig. 96 B). From this dorsal cell- mass a strand of cells is continued into the antennal rudiment, and forms the rudiment of the antennal nerve (v. Kennel, No. 4, Sedgwick, Xo. 10, Pts. iii. and iv.). The latter therefore appears as a direct continuation of the cerebral ganglion, and is in this way distinguished from all the other peripheral nerves, which are merely outgrowths of nerve-fibres (without participation of ganglionic cells). The nerve-mass yielded by the cephalic segment soon grows to such a size as to occupy the greater part of the head; the two masses of ganglion cells, from which the antennal nerves proceed, shift towards the middle dorsal line, where they form a pair of large egg-shaped swellings (Fig. 97, g). The pair of ganglia composing the brain are at first separated by a deep slit. This becomes bridged over later, the fibrous mass of the two halves of the brain uniting THE NERVOUS SYSTEM. 193 to form a commissure (the so-called supra-oesophageal commissure, v. Kennel). This commissure is thus of secondary origin, and seems also to involve parts of the brain lying further back. These parts, however, do not belong to the cephalic segment, but are formed by the (maxillary) segment that follows it. When the jaws become enclosed in the developing buccal cavity, the ganglia of the maxillary segment also sink below the surface, and pass toward the dorsal surface, so that they can soon be recognised in viewing the embryo from the dorsal side (v. Kennel, Fig. 97, gm). It must be assumed that this upward displacement takes place along the oesophageal commissures which are already present. The maxillary ganglia are thus approximated to those of the cephalic segment, with which they subsequently fuse. This fusion is very intimate, and the maxillary ganglia can be recognised as two moderate prominences behind the antennal ganglia (Fig. 97, gm). The fibrous masses of the maxillary ganglia from the two sides unite to form a com- missure, the sub - oesophageal commissure, this being facilitated by the downward slope of the posterior ends of these ganglia. This method of formation of the sub-oesophageal commissure renders it very improbable that it is a primitive structure. A commissure lying further back and consisting of cells (Fig. 97, c) might rather be considered as such. This latter commissure connects two ganglionic swellings, which may perhaps be attributed to the segment of the oral papillae. All the commissures which follow this are, as already mentioned, said to be derivatives of the fibrous substance. The point that must be regarded as of most importance in the formation of the brain in Peripatus is the fusion of the maxillary ganglia with the ganglia of the cephalic segment, for this feature distinguishes Peripatus from the Myriopoda and the Insecta so far as is at present known, and connects it rather with the Crustacea, in which the ganglia of the segment of the second antennae unite with the brain (Vol. ii., p. 165). It therefore seems likely that o Fig. 97. — Anterior part of the central nervous system of an embryo of P. Edwardsii at a somewhat earlier stage than that depicted in Fig. 94 B, dorsal aspect (after v. Kennel), at, antenna ; au, eye ; c, first commissure after the sub-oesophageal com- missure ; gj and g/j, cephalic portion of the brain ; gm, portion belonging to the maxillary segment ; gIV, the next follow- ing ganglion ; op, oral papilla ; p/t first foot ; s, passage for the oesophagus ; sd, slime-gland. 194 ONYCHOPHORA. the jaws of Peripatus are to be homologised, not with the mandibles of the Insecta, but rather with the second antennae of the Crustacea. The question which naturally arises as to whether the corresponding segment has been lost in the Insecta, or, in other words, as to the relation to it of the mandibular segment, can hardly, in the present state of our knowledge, be profitably discussed. The close connection brought about between the maxillary segment and the cephalic segment increases the probability of the view expressed above, that the antennal segment also (now known as the cephalic segment) may have been united with a cephalic section formerly present, and now to a great extent degenerated. We were led to this assumption by the presence of the two prominences in front of the antennal rudiments (Fig. 94, x), and by the close agreement in manner of formation between the antennae and the feet. It is, indeed, difficult to reconcile with this view the statement that the antennal nerve forms in a manner essentially different from the peripheral nerves, but this point has as yet received too little attention to be considered as of decisive importance. The Eyes. The rudiments of the eyes have already appeared before the separation of the nervous system from the ventral organs. On the dorsal boundary of the ectodermal thickening in the cephalic segment, a small pit is formed on each side, behind and somewhat ventrally to the rudiments of the antennae ; the floor of this pit is at first connected with the ectodermal thickening, but soon becomes detached from it. The pit closes to form a vesicle, which becomes constricted off from the ectoderm. Outwardly, i.e., towards the epidermis, this vesicle is unilaminar, but on the inner side it is multilaminar. Pigment appears on the inner boundaries of its cells, and in its cavity the lens is secreted. The cells of the inner and lateral walls yield the rods of the retina. A differentiation into cell- and fibre-substance has already taken place in the thickened inner wall of the optic vesicle, and a connection which occurs between this part and a process sent out by the brain gives rise to the optic nerve, which is thus a secondary formation (v. Kennel). Sedgwick's account of the origin of the eyes in Peripatus is somewhat different. According to him, the region in which they originate still belongs to the brain, and they do not lose their connection with the latter, the inner wall of the optic vesicles remaining united with the cell-mass of the brain. The optic nerve arises at this point later by simple constriction. The eyes thus originate chiefly from the brain, and are covered merely by the ectoderm of the surface ; they are "cerebral eyes," according to Sedgwick, in opposition to v. Kennel, who believes, as stated above, that they arise independently of the brain. It is possible that the observations made on the origin of the eyes in Perijmtus can be harmonised with those of the development of the eyes in the Annelida. The eyes of Peripatus agree closely with more highly organised Annelidan eyes, THE SLIME- AND THE CRURAL GLANDS. 195 such as those of the Alciopinac. According to Kleinenbeeg (Arol. i., p. 289, Annelidan Lit., No. 26), the eyes in this family arise independently of the cephalic ganglion as two ectodermal invaginations, but the inner wall of the optic vesicle is said to become closely connected with the brain, giving off cell- material direct to the latter. The elements of the two organs, in any case, seem for a time to be closely united, at the very point where the optic nerve forms later. If Kleinenbeeg's observations are confirmed, a similar condition might be thought to prevail in Pcripatus also, and the opposing views of v. Kennel and Sedgwick might thus be explained. The Slime- and the Crural Glands. The slime-glands are of ectodermal origin, arising as depressions on the tips of the oral papillae (Fig. 93). At first the pits are shallow, but they gradually deepen and their blind ends grow inwards and backwards. The pit has thus, at the stage depicted in Fig. 94 B, become a conical tube (Fig. 97, sd), which has grown back to the intestine. This growth continues in the following stages, so that the glands eventually attain a considerable length. They retain their simple tubular form; the branches which occur in them in the adult appear as outgrowths shortly before the embryo is mature and ready for birth (v. Kennel). The slime-glands are no doubt to be regarded as modifications of the crural glands, which, as sac-like structures, lie in the lateral divisions of the body-cavity and open on the ventral side of the feet. In the different species of Peripatus they differ in number and in distribution. These glands first appear at a late stage of embryonic development as ectodermal invaginations lying at the bases of the feet distally to the apertures of the nephridia (Fig. 102, c, Sedgwick). In the male (P. capensis) the crural glands of the last pair of feet are transformed into long glandular tubes (Balfour). The Alimentary Canal. With the exception of the short stomodaeum and proctodaeum, which are ectodermal derivatives, the alimentary canal is of ento- dermal origin. The following account is derived principally from the description given by Sedgwick of P. capensis, this form being chosen because we must regard it as more primitive than the American species examined by v. Kennel. The two forms vary principally in the first stages of the development of the intestine, the later stages showing great similarity. In order to understand the formation of the intestine, we must revert to the gastrula stage of P. capensis. The blastopore there leads into a cavity, which is lined by a thick protoplasmic 196 OXYCHOPHOEA. syncytium containing nuclei and rich in vacuoles. This voluminous nucleated mass must no doubt be regarded as corresponding to the yolk with its nuclei found in P. novae-zealandlae. In the latter form, the nucleated yolk forms part of the boundary of the arch- enteric cavity. In both forms the blastopore lengthens (Fig. 99 .4), and is constricted in the middle of its length, where its edges become approximated and fused ; thus the original single blastopore becomes divided into two apertures (Fig. 99 A-C). During this process, the vacuolated entodermal syncytium becomes arranged into a regular epithelium which, where the blastopore is still patent, passes over into the ectoderm, but in the region of the closed blastopore forms a tube that is said at first to be connected with the mesoderm-bands lying in that region, but to become isolated later, thenceforth form- ing a distinct entodermal tube. In P. novae-zealandiae, in consequence of the large amount of yolk present, this process is somewhat different. The entoderm-cells are here said to become arranged at the periphery of the yolk into an epithelium which thus surrounds the yolk. The latter would then he gradually absorbed during the further development of the intestine. The mouth and anus form as in P. capensis (Sheldon). The two apertures derived from the elongation and constriction of the blastopore (Fig. 99 D) are the primitive mouth and anus. They do not, however, persist as those organs in the adult, owing to the appearance of a depression of the ectoderm at each of the openings, so that the point of union between the ectoderm and the entoderm is shifted inward, and an ectodermal stomodaeum and proctodaeum are formed. The changes of form undergone by the embryo have their influence on the rudiment of the intestine. As a consequence of the curvature of the embryo, the entoderm extends anteriorly and posteriorly above the mouth and the anus (Fig. 98 ^4). The anterior wall of the stomodaeum thus runs forward. During the further development of the embryo, however, the course changes. When the mouth is shifted more to the anterior end, the anterior entodermal sac degenerates, and the stomodaeum now appears directed posteriorly (Fig. 98 B). The dorsal wall of the anterior portion of the intestine up to this point was closely apposed to the body-wall (Fig. 98 A and B), but the latter now separates from the gut and forms the swollen anterior end of the embryo (Fig. 98 C). It is followed in this course by a diverticulum of the entoderm, while the stomodaeum retains its former position. This diverticulum is THE ALIMENTARY CANAL. 197 also obliterated in the further course of development, and the intestine then runs straight back. The stomodaeum gives rise in Peripatus to the muscular oesophageal swelling (pharynx), meso- dermal tissue also contributing to its formation. The external changes in the mouth have already been described in connection with the external form of the body (Fig. 94, p. 185). The growth of the embryo produces similar changes at the posterior end of the intestine. A B. C. Fig. OS. — Median longitudinal sections through embryos of P. capensis at various ages (after Sedgwick), are, anus; di, anterior entodennal diverticulum; ■rut, entoderm; m, mouth; st, stomodaeum. In the American species of Peripatus, the intestine even at its first appearance differs from that of P. capensis, as no elongation of the blastopore occurs in these forms (v. Kenuel). The rudiment of the enteron, which is completely closed to the exterior and has been produced by the ingrowth of cells (Figs. SO, p. 171, and SI, p. 173), is here sac-shaped. As the embryo lengthens, the enteron also extends in the form of a tube. Its connection with the ectoderm is brought about through the fusion of the entoderm with the ectoderm, an invagination of the latter at this point forming the oral aperture. The mouth arises ventrally ou the boundary between the head and the trunk, and the anus in front of the blastopore (Fig. 89 A). It has already been pointed out (p. 178) that these two apertures occupy the same positions as in P. cwpensis, and that they perhaps 198 OXYCHOPHORA. B. were originally related to the blastopore. v. Kennel, however, does not believe this, and, further, seems little inclined to attribute much value to the observations on this point made in other species of Peripatus. He attributes an altogether different significance to the groove in the blastoderm observed by himself and described by us iu accordance with the views of English authors as the blastopore. The further development of the pharynx takes a course similar to that above described, the primary oral aperture shifting inwards, while an anterior ento- dermal diverticulum appears. The anal aperture, on the contrary, which arose in front of the blastopore through the formation of a slit (Fig. 89 A), is said not to coincide with the adult anus. The former closes by the approximation of its edges, and an ectodermal invagination arises a little distance in front of it, grows inward towards the entoderm and fuses with it. The rectum and anus are thus formed, the latter then shifting more to the posterior end of the embryo in consequence of the unequal growth of the latter (v. Kennel). The Mesodermal Structures. The formation of the chief mass of the meso- derm proceeds from a zone of growth lying at the posterior end of the blastopore, and extends forward from this point in the form of two bands (mesoderm - bands) lying symmetrically to the ven- tral median line. Where a slit -like blastopore is present, as in the African and Australian species, the mesoderm -bands lie in close contact with it, and are thus situated in the region where the ectoderm passes into the entoderm. After the blastopore has partially closed, the posterior (anal) aperture lies in front of the growing zone, and its position is the same in Fig. 99.— Ventral aspect of embryos of P. capensis, to ,, . . . illustrate the segmentation of the mesoderm (after the American species, 111 Balfour and Sedgwick), a, anus ; hi, blastopore ; which the blastopore is to, mouth.; its, primitive segments; w, zone of . growth. not slit-like. THE MESODERMAL STRUCTURES. 199 English authors, in accordance with the terminology used in the Vertebrata, have called the growing zone the primitive streak, and the groove-like depression that occurs in it the primitive groove. If such a groove occurs, it must no doubt be regarded as a continuation of the blastopore, and we must assume that it is not the most posterior part which is retained as the anus. The growing point itself must be considered as lying on the posterior margin of the blastopore. At this point, a great accumulation of cells takes place, and the germ-layers are here still fused and undifferentiated. In so far as the mesoderm-bands extend forward from this undifferentiated cell -mass, the condition here to a certain degree resembles that in the Annelida. Sedgwick even speaks of polar cells of the mesoderm, but of these nothing definite is known. There can be no doubt that the mesoderm is chiefly produced from behind, i.e., from the growing zone, but in consequence of the close apposition of the mesoderm-bands to the edges of the blastopore, the participation of the latter in their growth cannot be excluded (Sedgwick). In the American species, it appears certain that no such partici- pation occurs. The forward growth of the mesoderm-bands takes place from the point of ingrowth, which must be regarded as the blastopore, and their growth determines the lengthening of the whole embryo. The mesoderm-mass here separates from the sac-like rudiment of the enteron (Figs. 100 and 101), but not so sharply as to exclude a connection of the former with the ectoderm on the one side and with the entoderm on the other, which can be proved to exist even at later stages, when the mesodei'm has become far more highly differentiated. The mesoderm may thus be regarded even in this case as arising on the boundary between the ectoderm and the entoderm. The further development of the mesoderm-bands takes place in a manner very similar to that in which they develop in the Annelida. Before they have reached the anterior end of the blastopore, they break up into paired, regularly arranged segments (Fig. 99 A-C, us). Cavities then appear in these, and, as these gradually widen, the cell-material of the separate segments becomes arranged into a regular epithelium. The paired primitive segments thus arise. As they extend further, the outer wall becomes applied to the ectoderm and the inner to the entoderm (Fig. 100), like the somatic and splanchnic layers of the Annelida. A pair of primitive segments belongs to each body-segment. The differentiation of the primitive segments commences in the most anterior part of the mesoderm-bands and extends backward, their number increasing with the growth of the body ; the first pair of primitive segments to develop is thus that belonging to the cephalic segment, and this is also much larger than any other pair. It extends almost to the ventral and dorsal middle lines ; the two halves, however, do not come into contact, and con- sequently no mesentery is formed. Transverse sections through the body of an embryo at the stage when the primitive segments are being differentiated closely resemble, especially in the anterior and posterior regions, similar sections through an Annelidan embryo ; 200 ONYCHOPHORA. as. they show the ectoderm with its ventral thickenings, and the two niesoderm-segments containing the primitive body-cavity, bounded by the epithelial walls, applied to the ectoderm and the entoderm (Fig. 100). Such anatomical and histological differentiation is present in the embryo represented in Fig. 88, and no further essential change appears until twelve to fifteen segments are visible externally, together with the full (adult) number of internal segments (v. Kennel). When the mesoderm-bands have broken up into the series of consecutive primitive segments, the resemblance with the Annelida is very striking, but the further course of development differs, inasmuch as it is not the segmental cavities which yield the A. 1. nus. us. vor.tn Fig. 100. — Transverse sections through embryos of P. capensis (A) and P. Edwardsii (P.) (after Sedgwick and v. Kennel). A, transverse section through the region of the oral papillae in an embryo at about the stage depicted in Fig. 91 A. B, transverse section through a trunk- segment of a young embryo, d, intestine (entoderm) ; Ih, dorsal and ventral spaces between ectoderm and entoderm (parts of the primary and adult body-cavity) ; I, lateral, m, median portions of the segmental cavities ; mes, portions of mesoderm detached from the primitive segments; n, rudiment of the ventral cord; us, primitive segment; vo, ventral organ ; vo+n, common thickening of the ventral organ and the ventral nerve-cord. body-cavity of the adult, for, in Pervpatus, the latter arises as a pseudocoele independent of the primitive segments. All that is retained of these segments enters into the formation of the nephridia and the genital organs (v. Kennel, Sedgwick). The formation of the future body-cavitj' and of the nephridia is commenced by a thickening of the ventral Avail of the primitive segments ; and subsequently, by an ingrowth of the cells of this thickening, a separation of the segmental cavity into two spaces is brought about, one dorso-median and the other lateral (Fig. 100 B, m and 1) ; these are at first connected, but become completely separated later (Fig. 104 A, p. 210). The dorsal portion shifts THE BODY-CAVITY AND THE BLOOD-VASCULAR SYSTEM. 201 towards the dorsal median line, while the greater part of the lateral portion is withdrawn into the rudiments of the limbs (Fig. 104, v. Kennel, Sedgwick). Even before this separation has commenced, while the primitive segments still retain their sac-like shape, the antero-dorsal portion of each grows forward over a part of the preceding primitive segment, and thus extends into the preceding body-segment. This explains the fact that in transverse sections we not only see the segmental cavity of the segment through which the section passes, but also a portion of that belonging to the next segment, and that this latter lies above the ventral portion of the segmental cavity of the preceding segment. The lateral portions of the primitive segments yield the nephridia, and the dorso-median the genital glands in the segments which contain these organs ; in the other segments these portions disappear, their cell-elements being used in the formation of the blood vascular system and the musculature, and for the further development of the pseudocoele, which now comes under consideration. The Body-cavity and the Blood-vascular System. Even before the division of the primitive segments into two portions, the ectoderm had separated from the entoderm with which it was until then in close contiguity, thus giving rise to a free space dorsally and ventrally to the intestine. These spaces are the first indication of the body-cavity of the adult (Fig. 100 A and B, 111), and into them the mesoderm-cells which become detached from the primitive segments wander. As these cells become applied to the entoderm and ectoderm, the cavity which is at first bounded merely by these two germ-layers, and is therefore to be regarded as the primary body-cavity, becomes lined with mesodermal elements (Fig. 101 A, Hi). These spaces, in consequence of their origin, are not segmented, but the other and lateral portions of the future body-cavity, which arise by separation of the cell-elements in the inner thickened somatic wall of the lateral portions of the primitive segments, exhibit a segmental arrangement (Fig. 101 A, l.lh). These cavities, at first distinct from one another, fuse together later, and give rise to the two spaces, the lateral sinus of Sedgwick, which later run continuously through the body, and in which the nerve- strands lie in the adult. Another space on each side of the body agreeing in origin with these latter spaces, develops still more peripherally in the limb-rudiments and surrounds the nephridia (Figs. 101 and 102, p.lli). This last part of the body-cavity, which may best be described as the pedal body-cavity, unites later in some 202 OXYCHOPHORA. places with the lateral spaces, so that, where this is the case, the nephridia and the longitudinal nerves come to lie in one common cavity. Several cavities unite to form the central space which, in the adult, contains the intestine and the genital organs. According to n. us n 3* ***«' Wi>v '^$m?m& Fig. 101. — Transverse sections through embryos of P. capensis at different ages, A being taken through the segment of the oral papillae (somewhat diagrammatic, after Sedgwick), d, intestine; lh, dorsal and ventral median portions of the body-cavity; m.lh, lateral portion of the median body-cavity; n, rudiment of the ventral cord; ne, nephridia (in A, rudi- ment of the salivary glands) ; oc, external aperture of the same ; p, limb ; pe, pericardial cavity; p.lh, pedal body-cavity; sh, segmental cavity; its, dorsal portion of the primitive segments ; vo, ventral organs. Sedgwick, two new spaces (Fig. 101 B, pe and m.lh) appear on the outer side of each of the dorsal portions of the primitive segments (sh), the wall of which to some extent forms their inner boundary. THE BODY-CAVITY AND THE BLOOD-VASCULAR SYSTEM. 203 The lower of these spaces (m.lh) at a later stage grows above the intestine, and unites with its fellow and with the space which has already appeared beneath the intestine (Ih) to form the greater part of the permanent median cavity, the so-called central compartment of the body-cavity, while the upper one represents (pe) the rudiment of the pericardial cavity. The pericardial spaces on each side extend towards the median line, the remains of the primitive segments being thus displaced downwards. The cavity (lit) that arose early above the intestine thus appears confined, together with the dorso-median portions of the primitive segments (s7i), between the two pericardial spaces (pe) ; the latter grow above the pseudocoele (Ih), and also between the latter and the dorsal portions (sh) of the primitive segments, and unite with one another in the middle line. Thus the common pericardial cavity is formed, surrounding the dorsal pseudocoele (Ih) ; the latter now assumes a tubular form and becomes the definitive heart (Fig. 102, h). According to Sedgwick, the primitive segments take no part in the formation of the heart. The ostia of the heart, the formation of which has not been closely observed, do not arise until later, when the embryo is ready for birth. Detached mesoderm-cells, which become applied to the outer wall of the heart, give rise to the cell-mass within the pericardial cavity, which has been compared to the fat-body of the Insecta. It involuntarily reminds us of the cell-growth on the dorsal vessel of the Annelida, which is probably homologous with the pericardial gland of the Mollusca ; but we are prevented from homologising the two structures because the pericardial gland, as an outgrowth of the peritoneal epithelium, lies within the secondary body-cavity, while the cell- mass in Peripatus lies outside the latter. The pericardial space in Peripatus, like that of other Arthropods, does not correspond to the pericardium of the Mollusca or the coelom of the Annelida. Only its ventral wall (the pericardial septum, Fig. 101 B, and 102, ps) is perhaps in part formed by the somatic wall of the primitive segments, as is also the case in the Insecta. In Peripatus, as in the Arthropoda, the dorsal vessel is in direct communication, in the adult, with the body-cavity, and this fact is explained by the similarity in the development of this system of organs in the two divisions. In the two anterior (cephalic and maxillary) segments, the trans- formation of the primitive somites undergoes certain modifications determined by the special form of these parts. In the maxillary segment, the inner or dorsal part of the primitive somite is not extensive, and fuses with the corresponding part of the succeeding segment which projects into this segment. The different spaces of the permanent body-cavity are here less distinctly developed. The lateral parts of the primitive segments which occupy the rudi- 204 ONYCHOPHORA. merits of the jaws undergo considerable thickening of their outer walls, the formation of the strong musculature of the jaws being thus brought about, the inner wall supplies the cells which form the muscles of the pharynx and stomodaeum. The primitive somites in the cephalic segment are at first very large and occupy the greater part of the segment. As the ventral organs and the brain increase in size, the primitive segments are, however, pressed towards the dorsal surface, and thus become less extensive. Parts of the primitive segments pass into the antennae (as elsewhere into the feet), so that these latter at first appear to be hollow, though the cavity degenerates later (Fig. 96 A, us). The wall of the first primitive segment gives off cells for the formation of the musculature of the oesophagus. According to Sedgwick, the anterior primitive segment, like the rest, is divided into a dorsal and a lateral portion, the significance of which will be discussed below (cf. the Nephridia). The Musculature. Even in early stages, before any differentiation had taken place in the primitive segments, cells became detached from them and became applied to the ectoderm. These cells, and others which follow them during the further development of the mesoderm, give rise, immediately below the ectoderm, to a layer of circular muscle-fibres, which at first is thin, but in later stages becomes much thicker (Sedgwick). The longitudinal muscles arise later than the circular fibres, their fibres appearing in the cell-layer that covers the latter internally. According to Sedgwick, they are distributed into various com- plexes, one ventral, two ventro-lateral, two lateral and two dorsal, corresponding to the longitudinal muscle-bundles of the adult. The musculature of the intestine and of the inner organs generally is derived from the wandering cells which become detached from the primitive segments and applied to these organs. The Nephridia. The nephridia arise in the following way from the lateral portions of the primitive segments, the greater part of which occupy the bases of the limbs. Each primitive segment has a conical outgrowth directed towards the ventral side, which lengthens and, at the base of the foot, fuses with the ectoderm, which becomes perforated at this point, and thus the cavity of the primitive segment opens on to the exterior (Fig. 101 A); this aperture persists as the external opening of the nephridium (Sedgwick). The nephridium is now essentially complete (Fig. 101), for it does not possess a funnel opening into the adult body-cavity, as was formerly believed by Balfour and Gaffron, but, according to the latest observations of Sedgwick, throughout life ends blindly in this direction, the canal THE NEPHRIDIA. 205 of the nephridium being continued into tlie terminal saccular enlargement (Fig. 102, es). "We must thus assume that the nephrostome of the Annelida is represented by the opening of the nephridium into the terminal sac. The terminal sac, therefore, corresponds to the coelom (secondary body-cavity of the Annelida), a view which is confirmed by the manner in which the nephridia arise. A part of the coelom has thus come into direct relation to the kidney, and a state of things is found very similar to what has already been met with in the Crustacea (Vol. ii. , p. ISO), and, with certain modifications, will be found to recur in the Mollusca. jnes Fig. 102.— Transverse section through the posterior region of the body of an advanced embryo of P. capensis (after Sedgwick, somewhat diagrammatic), c, rudiments of the crural glands ; d, intestine ; e.?, end-sacs of the nephridia; h, heart ; l.lh, lateral, m.lh, median, p.lh, pedal portions of the adult body-cavity ; mes, mesodermal tissue ; n, ventral nerve-trunk ; ne, nephridial canal ; oc, external aperture of the nephridium ; p, foot ; pe, pericardial cavity ; ps, pericardial septum ; sb, collecting vesicle (urinary bladder) of the nephridium ; sd, slime- gland ; so, sole of the foot (thickening of the ectoderm) ; st, transverse commissure connecting the nerve-trunks (n) and the ventral organ (vo) ; g, gonad. The above description of the simple formation of the nephridia applies specially to those of the segments carrying the first to third limbs (of P. capensis). Those of the following limbs are distinguished by the fact that the canal becomes much coiled in later stages and widens towards its outer end (Sedgwick, Fig. 102, sb), like the urinary bladder in the nephridia (antennal glands) of the Malacostraca. Apart from the transformation which we shall find in the nephridia during the formation of the salivary glands and the genital organs, there are im- 206 ONYCHOPHORA. portant changes to be observed in the cephalic and maxillary segments. In the latter the nephridia have degenerated ; traces of them only are said to be found (v. Kennel). In the cephalic segment, on the contrary, the two segmental cavities (in early stages) are said still to open outward through canals (Sheldon, No. 12, Pt. ii.). v. Kennel and Sedgwick describe a (canal-like) continuation of the primitive cephalic cavity which descends on the outer side of the ectodermal thickening (rudiment of the nervous system) and fuses with the ectoderm, immediately in front of the jaws (Sedgwick) ; according to L. Sheldon, indeed, it even opens outward at this point. This canal has been considered homologous to the canal of the nephridia. According to Sedgwick, it therefore belongs to the lateral portion of the first primitive segment. We cannot clearly make out, from the figures given, the relation of this lateral portion to the coelomic cavity of the antennae. We therefore refrain from discussing the position of this efferent canal as conrpared with those of the other nephridia, and merely point out that a remarkable change in the position of the nephridium towards the limb must have taken place if this canal is really the nephridial canal of the so-called cephalic segment, and if our former assumption that the antennae of Per (pectus are transformed limbs is correct (c/. p. 186). The Salivary Glands. According to v. Kennel and Sedgwtick, who agree on this point, there can be no doubt that the paired gland -which opens into the buccal cavity through a short, common duct arises from the nephridia of the segment carrying the oral papillae. These ducts develop in the same way as the undoubted nephridia. They originate, after the separation of the dorso-median part, from the lateral portions of the primi- tive segments which develop an external aperture (Fig. 92, no, and 93, sp, p. 184). Their furtherdevelop- ment is peculiar only in so far that the canal, at the point where it passes into the terminal sac, begins to lengthen posteriorly (Fig. 103 .4), so that a long, blind tube arises at this point (Fig. 103 D, k). This tube gives origin to the principal part of the salivary gland ; it, however, retains through- out life the vesicular portion of the rudiment (s) corresponding to the terminal sac (v. Kennel, Sedgwick). The connection of the Fig. 103.— Formation of the salivary glands of 7'. capensis (after Sedgwick), k, canal of the gland; n, nephridial canal ; s, terminal sac, the walls of which in A appear thinner than in 1'.. THE ANAL GLANDS. 207 latter with the glandular tube becomes drawn out into a short canal (Fig. 103 B), which enters the latter dorsally (Sedgwick). The two external apertures of the nephridia (Fig. 103, sp) are displaced into the buccal cavity by the fold which encloses the mouth. They here come to lie in a transverse groove, which, as the buccal cavity develops further, becomes deeper and shorter. This groove eventually becomes a short canal with a slit-like lumen, into which the two nephridial canals (salivary glands) open. This is the common efferent duct of the salivary glands opening into the buccal cavity (v. Kennel). The Anal Glands. The so-called anal glands, a pair of glandular tubes which, in the male of P. Edwardsii, open ventrally on either side of the anus, and, in P. capensis, open through a short, common efferent duct at the genital aperture,* and are evidently related to the genital apparatus, are shown by their development to be modified nephridia (v. Kennel). They arise in P. Edwardsii from the primitive segments of the last (limbless) segment upon which the anus opens ventrally. The anal glands occur as rudiments in both sexes ; in the male only, however, do they attain the functional tubular form ; in the female they degenerate. In P. capensis, at the male genital aperture, a pair of glands open which are apparently the homologue of the anal glands of the American species. But since the nephridia of the segment which carries the genital aperture give rise to the efferent ducts of the genital apparatus (see below, p. 209), these glands must have a different origin. It appears probable that they are derived from one of the two additional pairs of primitive segments found by Sedgwick in P. capensis behind the primitive segments of the anal papillae. In this form, the genital aperture has shifted to a position quite near the anus, lying in front of it on the segment carrying the anal papillae. In P. Edwardsii, on the contrary, the genital aperture is found two segments further forward, on the penultimate limb-bearing segment. Since, according to Sedgwick, there are still two segments which remain in an undeveloped condition behind the last fully formed primitive segment (that of the efferent genital ducts), it might be assumed that these corresponded to the last limb-bearing segment and to the so-called anal segment of the American species. The latter would thus have two well-developed segments (the genital segment and that following it) in a region where in the African and New Zealand species a degeneration occurred, which led to the genital and anal apertures coming to lie on apparently one and the same segment. This would also explain the approximation of one of the last pairs of nephridia (the anal glands) to the antepenultimate pair (the efferent genital ducts). This assumption seems to be confirmed by the fact recently made * [In P. novac-britanniae (Willet), the pygidial (anal) glands open by a median aperture situated immediately above the anal orifice. — Ed.] 208 ONYCHOPHORA. known by L. Sheldon (No. 13) that, in P. novae -zealandiae, in the so-called anal segment, there are two coiled glandular tubes, each of which opens independently at the side of the body and laterally to the nerve-trunks, i.e., at a spot where normally the nephridial apertures open. These two glands are the equivalents of the anal glands (Sedgwick, Sheldon), and are more correctly called accessory glands of the male genital apparatus ; from their position, they may safely be regarded as modified nephridia. It should be mentioned further that the American species, which thus shows the more primitive condition in the segmentation of the posterior end of the body, shows on the other hand a less primitive method of reproduction. The shifting of the anus forward from the terminal segment must, indeed, in any case be regarded as secondary. The Genital Organs. In the fifteen anterior segments of the embryo of P. capensis, the dorso-median portions of the primitive segments are concerned in the formation of the pericardium and heart, but in the following segments their fate is quite different. After their separation from the lateral or nephridial portions, they shift towards the dorsal median line, and, decreasing in size, come to lie as small triangular sacs between the wall of the intestine and the pericardium (Fig. 102, g). It is these, according to Sedgwick, Avhich yield the genital glands. Cells appear in them at a very early stage ; these, which are distinguished by their size and specially large nuclei, are the primitive genital cells. We might assume with v. Kennel, that these arise in the Avail of the primitive segments themselves, or in the mesoderm-mass, before it breaks up into primitive segments, as will be described later in connection with the Insecta. On the other hand, Sedgwick ascribes an entodermal origin to the genital cells. [See Editorial Preface, Vol. II.] By the fusion of the dorsal portions of the primitive somites pertaining to consecutive body-segments and the breaking through of their transverse walls, two tubes are formed, and these come to lie in the middle division of the body-cavity. Up to this point, the rudiments of the genital organs are alike in the two sexes, but a histological differentiation now takes place, inasmuch as the genital cells increase more rapidly in the male, and become smaller, whereas in the female the germ-cells retain their large size. There is also an anatomical differentiation, the genital rudiments in the female fusing at the anterior end, while in the male they remain distinct in corre- spondence with the form of the genital apparatus in the adult. We must assume that the median portions of these posterior primitive segments yield the genital glands, while the efferent ducts are derived from the lateral portions of that primitive segment ORIGIN" OF THE MESODERMAL STRUCTURES. 209 which develops the genital aperture (in P. capensis this is the segment of the anal papillae, and in the American species the antepenultimate segment).* An actual separation into a lateral and a dorso-median section, such as takes place in the other primitive segments, does not, however, occur in the genital segment; here, indeed, the primitive somite extends dorsally, but its widened dorsal portion remains connected with the lateral portion. After this primitive segment, like the rudiments of the nephridia, has acquired an external aperture, its dorsal part fuses with the posterior ends of each of the tubular genital glands, and the rudiment of the genital organs is thus essentially completed. The two external apertures shift towards the middle line so as to lie beside one another. An invagination of the ectoderm then yields the unpaired terminal region (ductus ejaculatorius, vagina) of the genital apparatus. The development of the genital organs makes it evident that the cavity of the genital glands is homologous with the secondary body-cavity (or coelom). Its cellular lining thus corresponds to the peritoneal epithelium of the Annelida ; in both cases the genital products become detached from this, fall into the secondary body-cavity, i.e., in Peripatus the cavity of the genital glands, and pass out of the body through the nephridia. That the efferent ducts of the genital apparatus in Peripatus are homologous with the nephridia cannot be doubted. This is not only proved by their manner of developing, but is also confirmed by the fact that in the American species the nephridia are wanting in the antepenultimate segment, which carries the genital apertures, while they are regularly developed in the preceding segments and in the segment that follows (Gaffkon). The transformation of nephridia into efferent genital ducts such as are found in Peripatus is of special interest, because of the continuity of the transformed nephridia with the genital glands, and because this demonstrates a similar morphological constitution of the whole of the genital apparatus, such as is met with in other Arthropoda. This continuity tends to mask the nature of the efferent ducts ; their true character as nephridia can only be ascertained with certainty from the study of their development. Another account of the origin of the Mesodermal Structures. The description given by v. Kennel of the transformation of the primitive segments ditl'ers in some essential points from those of the English authors. Since those points are of great importance, they must be separately discussed. According to v. Kennel, besides the cell-growth which, in the form of a fold projecting from below, divides the segmental cavity into two spaces (Fig. 100 B, p. 200), a second fold grows in from the inner wall of the cavity (Fig. 104 A), so that the cavity is divided into three spaces which for a time are in communication (Fig. 104 A, I, II, III). The dorso-median portion (III) then becomes partitioned off, and this as well as the greater part of the lateral portion, which lies principally in the foot, is used up in providing elements * Cf. on this point the remarks on the anal glands, p. 207. P 210 ONYCHOPHORA. for the formation of the muscles and the connective tissue. The boundaries of the pseudocode are thus formed, the latter arising on the whole in the way already described by the separation from one another of the primary germ- layers and the formation of cavities in the massive mesodermal tissue, the remaining spaces belonging to the primitive segments naturally being added as their boundary walls break up. The continuity of the lining epithelium is only retained in the middle (ventral) portion (II) of the primitive segment. This yields the funnel of the nephridium (Fig. 104 A-C, II), which, according to this account, as also according to the statements of Balfour and Gaffron, is open towards the adult body- cavity. This funnel, of mesodermal origin, is joined by a ventral invagination of the ectoderm, which proceeds from the base of the foot and grows out like a tube (Fig. 104 A-C, nc). Thus while Sedgwick derives the whole of the nephridium from the mesoderm, v. Kennel traces the origin and the greater part of the nephridium to the ectoderm. In our previous description we followed the accounts of Balfour and Gaffron, because, from researches made by Bergh * in connection with the Annelida, it has become highly probable that the nephridia in those animals are formed entirely from the mesoderm. The derivation of the nephridial canal from the ectoderm must also affect v. Kennel's view as to the formation of the genital organs, since he too acknowledged the efferent genital ducts as transformed nephridia. Not only the terminal portion, but the whole of the efferent ducts is therefore of ectodermal origin ; only a short piece, connecting the ectodermal uteri and vasa deferentia with the genital glands is yielded by the mesodermal nephridial funnel (Fig. 105 A-C, ml). From this, in the female, are produced the acces- sory structures of the uterus. A glance at Fig. 105 A-C makes this clear. The female in P. novae-zcalandiae has a paired receptaculum seminis, and in the American species a paired receptaculum ovorum as well, which opens between the former and the ovary, and close to the latter, into the uterus. These accessory structures are not found in the female in F. capensis. Fig. 104. — Portions of sections through embryos of P. Ed wardsii at various stages (dia grammatic, after v. Kennel from Lang's Text-book of Comp, Anut.). d, intestine; I, foot Ih, body-cavity ; m, mesodermal tissue ; n, ventral nerve-trunk nc, nephridial canal ; I, II, III the three spaces of the primi tive segment-cavity, II repre- senting the rudiment of the funnel. The paired receptaculum seminis arises as follows : — Each of the two uteri makes a sharp bend behind the ovary, so that * R. S. Bergh, Neue Beitriige zur Embryologie der Anneliden, Theil i., Zcilschr.f. wiss. Zool., Bd. 1., 1890 ; cf. also Vol. i., p. 297. GENERAL CONSIDERATIONS. 211 at one point the ascending and descending portions approach each other. As the bent part widens at the same time, fusion and perforation of the wall of the uterus takes place, so that the parts of the uterus in front of and behind the bend enter into direct com- munication. From this point, however, two canals lead into the widened portion of the uterus, which has thus become vesicular, and forms the receptaculum seminis. The receptaculum ovorum arises between the receptaculum seminis and the ovary as a hernia-like outgrowth of the uterus (or oviduct). When this has attained a certain size, the epithelium at its point is said to rupture (v. Kennel), the so-called ovarial funnel of Gaffron thus arising. This funnel, however, is not, as this author assumes, open towards the body- cavity, but remains covered by the connective tissue which invests the uterus (v. Kennel). These points seem to require re-investi- gation. In the meantime, Sedgwick's conjecture that the receptaculum ovorum corresponds to the terminal sac of the nephridium of the genital segment seems to merit attention.* v. Kennel agrees with Sedgwick as to the formation of the genital glands in so far as he also derives them from the dorso- median portion of the primitive segments, hut, if we understand him rightly, he thinks that the dorsal portions of only two primitive segments, viz., those belonging to the genital segments, take part in the process ; these remain united with the lateral portions (Fig. 105 A-C), as Sedgwick .also showed. General Considerations. The possession of tracheae and seg- mentally arranged nephridia brings Peripaius into relation with the Artliropoda on the one hand and the Annelida on the other. In addition to these two principal characters, Peripatus has a number of other Fig. 105. — Diagrammatic sections through the genital segment of female embryos of P. Edwardsii at various ages (after v. Kennel, from Lang's Text-book of Comp. Anat.). d, intestine ; ee, nephridial canal (arising through an invagination of the ectoderm); ml, mesodermal por- tion (funnel) of the nephridium ; n, ventral nerve-trunk ; ov, ovary (dorso- median portion of the primitive seg- ment) ; va, vagina (unpaired ecto- dermal invagination). * [Willey (App. to Lit. on Onychophora, No. II.) supports Sedgwick's views. — Ed.] 212 OXYCHOPHORA features in common with these two divisions, some of these occurring in its ontogeny. Although the eggs of some species of Peripatus have little, or even no yolk, it is highly probable that they are to be traced back to eggs ricli in yolk, like those of P. novae-zealandiae* These undergo superficial cleavage and become covered with a blastoderm, and thus resemble the eggs of the majority of Arthropods. The long, slit-like blastopore which, in closing, leaves the oral and anal apertures, finds its homologue among the Insecta. The formation of the germ-layers can also be compared with processes found among the Insecta or the Crustacea, but in the development of the mesoderm -bands from the edge of the blastopore and their gradual shifting forwards, and in the nature of the segmentation of these bands, we find an agree- ment with the Annelida. The same is true with regard to the form of the germ -bands, which is, of course, largely determined by the nature of the mesoderm-bands. Balfour pointed out the great similarity between the germ-bands of Peripatus and those of the Myriopoda and the Arachnida (e.g., Geophilus, Scorpio, Agalena), which is shown in the form of the limb-rudiments, and especially in that of the cephalic lobes. But we must remember that the formation of such germ-bands was first introduced in the Annelida ( Oligoeliaeta, Himdiuea). The great development of the brain and the possession of limbs produces again a greater resemblance to the Arthropoda. This resemblance finds further expression in the union of several segments to form the head, and in the transformation of their appendages (limbs) into mouth-parts. The inclusion of one or more trunk- segments in the head has, indeed, been stated for the Annelida, but these segments are never so radically transformed as in the Arthro- poda and in Peripatus. Such agreement suggests the question whether the cephalic segments of the Onychophora and those of the Arthropoda may not be homologous, but we are here met with difficulties, the number of the segments involved in the formation of the head differing in the two groups, and the relation of the segments to each other also varying. The latter finds its expression in the composition of the brain. In Peripatus the ganglia of the maxillary segment are included in the formation of the brain, which is not the case in the Myriopoda and Insecta. The jaws of Peripatus cannot, therefore, be homologised with the * [For "Willey's conclusions (App. to Lit. on Onychophora, No. II.) see footnotes, pp. 105 and 216. — Ed.] GENERAL CONSIDERATIONS. 213 mandibles of the Insecta, but this condition in Peripatus recalls a similar feature in the Crustacea, in which the ganglia of the second antennae are incorporated in the brain. The second antennae of the Crustacea may thus better be compared with the jaws of Peripatus. Then, however, the question arises, whether a segment has not been lost in the Insecta. This point can only be fully discussed later on (ch. xxviii.). Our view of the antenna-bearing segment, and of its relation to that of the Crustacea or of the air- breathing Arthropoda on the one hand, and to the cephalic segment of the Annelida on the other, has already been stated (p. 186), and to this we must refer the reader. We must, however, point out that, in the transformation of the anterior limbs into mouth-parts, Peripatus approaches the Arthropoda and is removed from the Annelida, in which the jaws are mere cuticular developments of the stomodaeum. There can be no doubt that, by the development of limbs armed with claws, Peripatus is in advance of the Annelida ; on the other hand, the limbs are without the segmentation characteristic of the Arthropoda ; the lateral position of the feet also seems a more primitive character inclining towards the Annelida, this character, together with the homonomous segmentation of the body, giving Peripatus a worm-like appearance. Another point of resemblance to the Annelida is found in the crural glands, which have no doubt rightly been traced back to the glandular sacs (setiparous glands) of the Annelidan parapodia (Balfour). The crural glands are found again in the higher Tracheata, although in these transformed nephridia (coxal glands) have repeatedly been held to be the homo- logues of the crural glands of Peripatus. The passage of the primitive segments into the limbs, which is so characteristic of Peripatus, recurs, though not to such a great extent, in the Myriopoda, the lower Insects, and the Arachnida. When the mesoderm-bands first appear and break up into segments, their wide extension brings about a great resemblance between the embryo of Peripahis and the Annelida, though it must not be forgotten that this similarity is greatest just in the species in which the yolk is most reduced, and which we must therefore assume show a derived condition (African and American species).* With regard to the development of the mesoderm, Peripatus shows, on the whole, a greater resemblance to the Arthropoda, if the musculature and the segmental repetition of * [See footnote, p. 165. — Ed.] 214 ONYCHOPHORA. the nephridia are left out of account. The muscles, with the exception of those of the jaw, show no transverse striation. They form a dermo-muscular tube composed of several layers of diagonal and longitudinal fibres arranged in symmetrically distributed bands. Here Ave have features which recall the Annelida far more than the Arthropoda, in which the dermo-muscular tube breaks up into separate groups of muscles distributed in a definite manner. The body-cavity, on the contrary, in its origin (as a pseudocoele), as- well as in its final development, is altogether Arthropodan in character. This is also the case with the dorsal vessel, which is connected by ostia with the pericardium, and thus with the pseudo- coele, for the jjericardial space is, as in the Arthropoda, a part of the pseudocoele, and is formed on the whole in the same manner as that of the Arthropoda. The development of the body-cavity, and its division into separate spaces in the embryo, may be compared with what is found in the ontogeny of the Myriopoda and the Insecta, and may thus be regarded as an important point of agreement between Peripatus and these forms {i.e., the Arthropoda generally). The nephridia seemed to be a specially strong bond of union between Peripatus and the Annelida as long as we had to assume that they opened, as in the latter, through a wide funnel into the body-cavity (Balfour, Gaffron). But since it has become known that they are closed by means of a vesicle towards the adult body- cavity (Sedgwick), although their segmental repetition still offers an important point of comparison with the Annelida, a still greater inclination towards the Arthropoda is shown, the nephridia (antenna! and shell-glands) in the Crustacea having the same form.* This similarity of structure renders it probable that the nephridia of Peripatus are no longer ciliated f; if, however, the statements that have been made as to the presence of a ciliated epithelium in the nephridia that are transformed in the efferent genital ducts J should 11 [This vesicle or end-sac is a thin-walled remnant of the coeloni, homologous with that of an Annelid, with which the nephridinm communicates by a thick- walled funnel. The body-cavity of the adult Peripatus is a pseudocoele like that of other Arthropoda, with which we should certainly not expect to find the nephridia communicating. The nephridia of Peripatus are specially interesting since they appear to combine certain Arthropodan features (the coelomic end-sac) with others only met with in the Annelida (segmental repetition and marked funnel). — El).] t We have not been able to find any definite statements as to the presence or absence of cilia in the nephridia of Peripatus. J Gaffkon describes and figures a thickly ciliated epithelium lining the vasa deferentia. GENERAL CONSIDERATIONS. 215 prove correct, this fact would furnish another Annelidan characteristic for Peripatus. The nephridia in Peripatus, as in the Annelida, are utilised as efferent ducts of the genital organs. The genital products here, as there, arise in the wall of the secondary body-cavity (which, in Peripatus, is very much circumscribed), and pass from it into the funnel of the nephridium. At this point, however, the develop- ment of Peripatus seems once more to bring it near to the Arthropoda. The ducts unite with the genital glands to form one whole, a condition which, indeed, has already been met with in various divisions of the Annelida. The structure of the eyes in Peripatus and their mode of origin show no connection with the organisation of the Arthropoda, but, on the contrary, agree very closely with that of the Annelida. The eyes closely resemble the eyes of the Alciopinae. Further, in comparing Peripatus with the Myriopoda, the absence of Malpighian vessels, or, indeed, of any trace of such organs, is a striking peculiarity. In forming a comprehensive judgment of the anatomical and ontogenetic relationships of Peripatus, we have to admit that it unites Annelidan with Arthropodan characters, but that the lattei preponderate ; not only in its external form, but in its inner organisation, does Peripatus appear far more like an Arthropod than an Annelid. Phylogenetically, Peripatus may well be con- sidered as an intermediate form in a series beginning with the Annelida and ending with the Insecta, although this does not, of course, imply that Peripatus is actually to be regarded as the ancestor of the Myriopoda and the Insecta. Another peculiarity, worthy of note from an ontogenetic point of view, is the late appearance of the tracheae, the origin of which has not so far been observed even in the oldest embryos, and the interpretation of which is rendered appreciably more difficult by our ignorance of their formation. We can hardly err in tracing them back to ectodermal invaginations, and it therefore seems probable that they are to be derived from modified integumental glands, or, still better, from respiratory portions of the body-covering. Whether we may, from the absence of observations on this point, conclude that the tracheae actually appear very late, or whether we must consider that they have been overlooked, does not appear very certain, but Ave are inclined to adopt the first of these views, and to explain the late ontogenetic appearance of the tracheae by 216 ONTCHOPHORA. their late phylogenetic acquisition. The irregular distribution of the tracheae in Peripatus, as contrasted with their regular arrange- ment in the higher Tracheata, seems to indicate a lower condi- tion of the tracheal system, and thus confirms the view that it is newly acquired, and found, in Peripatus, to a certain extent in its initial stage.* LITERATURE. 1. Balfour, F. M. The anatomy and development of P. capensis. Quart. Journ. Micro. Sci. Vol. xxiii. 1883. 2. Gaffron, E. Beitrage zur Anatomie und Histologie von Peripatus. Theil. i. u. ii. Schneider's Zool. Beitrage Bd. i. Breslau, 1885. 3. Hutton, F. W. On P. novae-zealandiae. Ann. Mag. Nat. Hist. (4). Vol. xviii. 1876. 4. Kennel, J. von. Entwicklungsgeschichte von P. Edwardsii und P. torquatus. Theil. i. u. ii. Arb. zool. Inst. Univ. Wiirzburg. Bd. vii. u. viii. 1885 u. 1886. * [A most important paper dealing with Peripatus lias recently been published by Willey (App. to Lit. on Onychophora, Xo. II.). This author, in describing the anatomy and development of a new Peripatus [P. novae-britanniae), advances such a large series of new facts and fresh interpretations of old ones, that the above account becomes very incomplete without a summary of his results, consequently all students of the Onychophora should refer to this monograph. Unfortunately, the earliest stages obtained were not well enough preserved to make out the process of cleavage. The egg is small (roinm. ), without yolk, and enclosed in a remarkably thick egg-membrane. Willey's most important discovery is that at an early stage the embryo has the form of an oval and, for the. most part, thin-walled vesicle, only a portion of which gives rise to the embryo, the latter forming a thickened postero- ventral area which becomes, as it were, invaginated into the vesicle. The thin-walled vesicle, which is com- posed of both ectoderm and entoderm, serves to absorb nourishment and to protect the embryo. Willey regards it as physiologically analogous to the blastodermic vesicle of the Mammalia, and adopts for it Htjbrecht's term, trophoblast ; it is possibly homologous with the embryonic envelopes of the Insecta. A comparison of "Willey's figures with those of v. Kennel and Sclateu will show that this trophic vesicle is undoubtedly the homologue of the thin-walled vesicle which encloses the embryo of P. Edwardsii, and which v. Kennel derived from the uterine epithelium. Willey's observations thus support those of Sclater on this point. The trophic cavity eventually becomes the gastral cavity, but the embryonic entoderm which was largely used up in the forming of wandering trophocytes ( = vitellophags) has to be reconstructed ; during this process the entoderm undergoes histolysis, and yolk -granules appear to be formed. Willey regards the yolk of P. novae-zealandiae as a secondarily acquired structure, consequently he differs from our author's conclusions regard- ing the primitive nature of that form, and considers it as one of the more specialised species of Peripatus. He does not consider that the large empty trophic vesicle of P. novae-britanniae indicates the former existence of yolk, but refers to Hubrecht's conclusions regarding the similar condition in the higher Mammalia which he thinks justify his view, that the presence of yolk in the eggs of Peripatus is a very recent secondary condition. — Ed.] LITERATURE. 217 5. Kennel, J. von. Ueber die friihesten Entwicklungsstadien der siidamerikanischen Peripatusarten. Sitzungsber. Naturf. Gesellsch. Dorpat. Bd. viii. 1888. 6. aIoseley, H. N. On the structure and development of P. capensis. Phil. Trans. Roy. Soc. London. Vol. clxiv. 1874. 7. Moselet, H. N". Kemarks on observations by Capt. Hutton on P. novae-zealandiae, etc. Ann. Mag. Nat. Hist. (4). Vol. xix. 1877. 8. Saint Kemv, G. Contribution a l'etude du cerveau chez les Arthropodes Tracheites. Archiv. Zool. Exp. (ii.). Tom. v. Suppl. 1887-90. 9. Sclater, W. L. On the early stages of the development of a South American Species of Peripatus. Quart. Jonrn. Micro. Sci. Vol. xxviii. 1888. 10. Sedgwick, A. Tbe development of the Cape Species of Peripatus. Parts i.-iv. Quart. Journ. Micro. Sci. Vols, xxv.- xx viii. 1885-88. 11. Sedgwick, A. A Monograph of the Species and Distribution of the Genus Peripatus. Quart. Journ. Micro. Sci. Vol. xxviii. 1888. 12. Sheldon, L. On the development of P. novae-zealandiae. Parts i. and ii. Quart. Journ. Micro. Sci. A'ols. xxviii. and xxix. 1888-89 13. Sheldon, L. Notes on the anatomy of P. capensis and P. novae-zealandiae. Quart. Journ. Micro. Sci. Vol. xxviii. 1888. 14. Sheldon, L. The maturation of the ovum in the Cape and New Zealand Species of Peripatus. Quart. Journ. Micro. Sci. Vol. xxx. 1890. APPENDIX TO LITERATURE ON ONYCHOPHORA. I. Dendy, A. Description of Peripatus oviparus. Proc. Linn. Soc. N.S.W. Vol. x. 1895. II. Willet, A. The Anatomy and Development of Peripatus novae-britanniae. Willey Zoological Results. Cambridge, 1898. For recent views on the systematic position of Peripatus, see Nat. Sci., Vol. x., pp. 97 and 264, 1897. CHAPTEE XXVI. MYRIOPODA. Systematic : — I. Chilopoda, with a dorso-ventrally compressed body, two pairs of maxillae, and one pair of maxillipedes ; with one pair of limbs on each body-segment ; genital aperture on the penultimate segment. Geophilus, Lithobius, Scolopendra, Scutigera. II. Symphyla, small delicate forms with only twelve segments, to each of which is added an intermediate segment ; with twelve pairs of limbs on the principal segments ; with one pair of maxillae, without maxillipedes ; genital aperture on the fourth segment ; a single pair of branching tracheae opening on the head ; at the posterior end, two stylet-like processes (cerci). Scolopendrella. III. Pauropoda, small delicate forms with twelve body-segments, nine of which only carry limbs ; with one pair of maxillae ; without maxillipedes ; cbaracterised by the possession of three long flagellae on the antennae ; paired genital apertures situated at the base of the second pair of legs; tracheae not known. Pauropus. IV. Diplopoda (Chilognatha),* with cylindrical body ; with one pair of maxillae (gnathochilarium), without maxillipedes ; the fifth and subsequent segments are double, each carrying two pairs of limbs ; genital aperture between the second and third pairs of limbs. Polyxenus, Glomeris, Polydesmus, Strongylosoma, Julus. Oviposition and the Constitution of the Egg. The eggs of the Diplopoda are usually laid in large numbers, enclosed in an earthen nest formed by the female, by whom they are watched for a long time, often until the young are hatched. * [POCOCK subdivides the Diplopoda into the Pselaphognatha, containing the single family Polyxenidae, and the Chilognatha. For recent views on relation- ship of the Chilopoda and the Diplopoda, see footnote at the end of this chapter. —Ed.] 0V1P0SITI0X AND THE CONSTITUTION OF THE EGG. 219 The nests are found in the damp earth, under stones, the bark of trees, etc. The eggs within them may be glued together into large clumps (Jirfus). The Polyxenidae surround the heaps of eggs with a thick envelope formed from their own hair. Glomeris lays its eggs singly and at wide intervals ; each egg is surrounded by the female with a special capsule formed out of earth moistened by a glandular secretion.* It has been stated that viviparous forms occur among the Scolo- pendridae / but, on the other hand, some of these animals have also been observed to lay eggs in large clusters. In these latter cases it Avas found that the eggs are taken care of by the female, Fig. 106. — Sections through eggs of Gcophilus ferrugineus at two early stages, illustrating the formation of the blastoderm (after Sogkaff). d, yolk ; dp, yolk-pyramids ; /.-, nuclei, each surrounded with a protoplasmic area. which rolls itself spirally round them and remains immovable until the young are hatched. Llthobius lays its sticky eggs one at a time and rolls them in the soil until they become coated with earth. The eggs, which are usually spherical, seldom oval, are very rich in yolk. They are surrounded by a vitelline membrane and another structureless but firmer envelope, the chorion, which is apparently secreted by the genital ducts. * Statements of a more detailed nature as to the time and mode of laying the eggs in the Diplopoda are to be found in the older works of Newport and Fabre, and more especially in two treatises by 0. v. Rath (Nos. 16 and 17). Further accounts of the laying of the eggs and the care of the brood are given by L.vtzel (No. 10) in his description of different species of Myriopoda. 220 MYRIOPODA. 1. Cleavage and Formation of the Germ-layers. The cleavage of the Myriopodan egg has repeatedly been described as total, but deserves this designation even less than, for example, the eggs of the Araneae. The egg, it is true, shows a number of segments, which at first is not large, but increases later, and these produce the appearance of division into more or less sharply marked blastomeres, but this appearance is not the expression of total cleavage in the strict meaning of the term, and only arises some time after the division of the cleavage-nucleus and its descendants inside the egg. Fig. 107.— Sections through eggs of Geophilus ferrutjincus showing the blastoderm-formation (after Sografk). hi, blastoderm ; dp, yolk-pyramids ; gr, groups of blastoderm-cells on tin- future dorsal surface ; k, nuclei with the protoplasmic areas surrounding them. The cleavage of the external surface of the egg does not appear to take place in all Myriopodan eggs. Heathcote, for instance, points out specially, in connection with the Julus tcrrestris, Leach, that no outward segmentation is to be observed in this form, although Metschnikoff, in another species of Julus (J. Morelletti, Lucas), carefully described and figured the segmentation of the surface of the egg. The absence of this external cleavage in other species is perhaps to be accounted for by the great abundance of yolk in these forms. The cleavage-nucleus lies, surrounded by a mass of protoplasm, in the centre of the egg. It here divides first into two, and these soon increase by further division, so that many nuclei, each surrounded with an area of protoplasm, are found at this stage in the centre of CLEAVAGE AND FORMATION OF THE GERM-LAYERS. 221 VI. the egg (Fig. 106 A). Only after this has taken place do the separate cell-areas become marked off from one another (Fig. 106 B), and thus produce the appearance of total cleavage mentioned above. We may assume that some of the central nuclei had already shifted towards the periphery, and that the cleavage of the egg is a con- sequence of this. It is probable that the yolk -pyramids which are bounded by the furrows are provided with nuclei, although it has been impossible up to the present time to prove this with certaintj'. Sograff assumes that the nuclei belonging to the pyra- mids lie at their tips, and are therefore not far removed from the central nuclei. The yolk-pyramids are not completely marked off from one another, but are connected at the centre of the egg where the central nuclei lie (Fig. 106 B), After the cleavage of the yolk has taken place, a migration of the central nuclei towards the periphery occurs. The nuclei force their way into the yolk-pyramids, the number of which has increased, and shift towards the periphery of the egg (Fig. 107). Judging from Sograff's figures, this migra- tion seems to take place chiefly along the boundaries of the yolk-pyramids (Fig. 107 .4). Eeachino- the surface of the egg, the nuclei at first are not evenly distributed, but arranged in groups (Metsch- nikoff, Sograff, Heathcote), but later they form a con- tinuous layer of cells, the blastoderm. The latter de- velops first on the ventral surface (Fig. 107 B), where the cells divided more quickly, and consequently become smaller, and proceeds from this to the dorsal surface, where, until now, the cells were still arranged in groups (Fig. 107 B, gr). The yolk-pyramids remain distinct until some time after the formation of the blastoderm. Most authors agree in assuming that in the formation of the blastoderm a large proportion of the nuclei remain within the egg, perhaps in the yolk-pyramids. This cell-material represents for the Fig. 108.— Section through an embryo of Julus tcrrcstris on the sixteenth clay of development, somewhat diagrammatic (after Heathcote). hi, blastoderm ; d, yolk ; do, dorsal surface ; dz, yolk-cells ; k, the keel-shaped accumulation of cells on the ventral side (ue). 90 0 MYRI0P0DA. most part the entoderm, but is also said to take part in the formation of the mesoderm (Sograff, Heathcote). The latter arises partly as a cell-growth which takes place in the ventro- median portion of the blastoderm. There arises in this region a thickening of the blastoderm projecting inward in the shape of a keel (Fig. 108), in the formation of which cells are said to take part that migrate from the inner part of the egg and become applied to the thickening (Heathcote, Sograff). This thickening is at first median, but divides later into two lateral bands (the mesoderm-bands), which then break up into segments containing cavities. — mXi Fig. 109.— Surface views of three early stages in the development of Geophilus ferrugineus, showing the rudiment of the germ-band (after Sograff). A and B, lateral views. C, ventral view of the anterior part of the germ-band, a, anus ; at, antenna ; d, yolk ; Id, cephalic lobes; md, mandibular segment; mxlt segment of the first pair of maxillae ; r, median longitudinal groove. AsJ a longitudinal furrow appeared early in the ventral middle line, the formation of the germ-bands in the Myriopoda might be interpreted in the same way as in Peripatus, i.e., the furrow might be regarded as an indication of gastrulation, especially as the mouth and the anus appear at its anterior and posterior ends. This view, indeed, is in opposition to the statements regarding the origin of the entoderm from cells which have remained in the yolk, but we must remember THE DEVELOPMENT OF THE EXTERNAL FORM OF THE BODY. 223 that similar statements have been made in connection with the Insecta, which later research has not confirmed, and, moreover, the ontogeny of the Myriopoda has not yet been exhaustively investigated, so that nothing definite is known as to the relation of this furrow to the formation of the germ-layers.* The cleavage and the formation of the germ-layers in the Myriopoda has been investigated by Metschnikoff in various Diplopoda (Strongylosoma, Polydcsmus, Polyxenus, and Julus, No. 11), as well as in the Chilopoda (Geophilus, No. 12) ; the last-named form has also been investigated by Sograff (Nos. 19 and 20), while Heathcote has re-investigated Julus (No. 7). The earliest observations are those of Metschnikoff, but these unfortunately were made without the assistance of sections. Sograff's treatise being in Russian has remained to a certain extent inaccessible to us, and the descriptions of Heathcote are not very satisfactory. Besides these there is a treatise by Stecker on the early development in the Myriopoda (No. 21). The results obtained by this author are, as Balfour pointed out, in absolute contradiction to what is otherwise known of the ontogeny of the Myriopoda. This author found no free yolk, and observed a blastula giving rise to an invagination-gastrula with a wide archenteron. It has been conjectured by Sograff (No. 19), that Stecker mistook Gastropod eggs for 'the eggs of Myriopoda, but this hardly appears possible when we consider Stecker's definite statements as to the species examined by him and his description of their later stages. Further investiga- tions of the genera examined by Stecker have led to totally different results, which justify us in regarding Stecker's account as unreliable and inadmissible in our account of the ontogeny of the Myriopoda. 2. The Development of the External Form of the Body. A. Chilopoda. As has already been mentioned, the development of the blastoderm is first completed on the ventral side of the egg, from whence it spreads later to the dorsal side (Fig. 107 B). This ventral blastoderm consists of small cells, and here the rndiment of the germ-band appears, its cephalic lobes being first visible (Fig. 1 09, M). Posteriorly no differentiation can be observed, the rudiment of the germ-band there fading away into the undifferentiated blastoderm, which has not yet completely surrounded the egg. The rudiments of the antennae appear early at the posterior boundary of this first recognisable segment of the germ-band, the cephalic lobes (Fig. 109, B and C, at). The next segment to appear belongs to the mandibles, and then follow the segments of the two pairs of maxillae and the maxillipedes (Fig. 110, md-mp). The limbs and segments generally develop in regular order from before backward, as can be clearly made out from Figs. 109 C, 110 and 111. According to Sograff, the anus forms very early in Geophilus * [Heyjions (App. to Lit. on Myriopoda, No. III.) states that there is no gastrula-groove in Scolcqjcndra ; he finds that the entoderm forms both from the yolk-cells and by a budding oil" of cells from all parts of the blastoderm. — Ed.] 224 MYRIOPODA. (Fig. 109 B, a). The mouth, on the contrary, is said to appear only at a later stage. It can be seen as an invagination between the cephalic lobes when five segments have developed. It is difficult to ascertain from the descriptions given whether the oral invagination lies in front of the antennae ; this is certainly indicated in Fig. Ill of a Geophilus embryo, while in other figures (e.g., Fig. 110) it is not Oir— b£—\ Fios. 110 and 111.— Two stages in the development of the germ-bands of Geophilus femtgineus, unrolled (alter Sograff). o, anus ; at, antenna ; Id, cephalic lobes ; m, mouth ; md, mandi- bles ; m.g, Malpighian vessels ; mp, maxillipedes (or segment of the maxillipedes) ; mxlt m.i.., lirst and second maxillae (or segments of the same); nr, neural groove; ol, upper lip ; p, legs; /),, p2, segments of the first two pairs of legs; r, ventral longitudinal groove; s, the lateral parts of the germ-bands which already show segmentation ; si, caudal lobe ; ul, paired rudiment of the lower lip. evident. In this latter figure, which represents an earlier stage than Fig. Ill, the antennae rather lie in front of the mouth, but another circumstance still to be described in connection with the formation of the organs indicates that the condition here resembles that in Peripalus, and that the antennae have CHILOPODA DEVELOPMENT OF THE EXTERNAL FORM. 225 here, perhaps, as there, the character of post-oral limbs. The difference in the position of the antennae in the two embryos may, perhaps, be explained by imagining a change both of the time and the place of their appearance, such as is occasionally found in other embryonic rudiments. In the embryo of Geophilus depicted in Fig. Ill, behind the oral aperture, two somewhat large prominences {ul) can be seen, resembling a pair of limbs, R. Fig. 112. — Two embryos of Geophilus, lateral aspect. The germ-band surrounds a large part of the yolk and still shows the dorsal curvature. The two lateral swellings found in the embryo at the stage depicted in Bare omitted for the sake of clearness (after Metschnikoff). a, anus ; at, antenna ; d, yolk ; Jcl, cephalic lobe ; md, mandible ; mxlt mx„, first and second maxillae ; rap, maxillipedes ; m\v, mouth-parts ; p (p,, p„), legs ; si, caudal lobe. but lying in front of the mandibles. Sograff calls these structures the lower lip, but it is not clear whether he actually considers them to be limbs, and to what part of the adult he refers them. The mouth-parts known to us in the Chilopoda only develop behind these, as already seen (Figs. 109-111). Similar structures met with in the Insecta are not regarded as limbs, but as a lower lip. 226 MYRIOPODA. In the middle of the germ-band a shallow furrow appears (Fig. 109 G) extending from the mouth to the anus, the two apertures marking the ends of the furrow (Sograff). This furrow, which soon disappears, naturally invites comparison with the long slit-like blastopore of Peripatus on account of its corresponding position and relation to the mouth and anus (c/. pp. 175 and 222). Somewhat later, on each side of the middle line, the rudiments of the ganglionic chain appear as thickenings of the ectoderm also separated by a median furrow (Fig. Ill, nr). This furrow, however, must not be confounded with the furrow just mentioned, which has disappeared before the formation of the groove now under consideration. So far the germ-band of Geopliilus, with its numerous segments, has developed from before backward, and extends with a dorsal curvature round the greater part of the yolk-mass (Fig. 112 A). The posterior end at a some- what later stage grows even further up on to the dorsal side, so as to approach still more nearly the cephalic lobes, as may be seen in Fig. 112 B. Now, however, a change takes place, a transverse ventral fur- row appearing in the region of the twentieth segment. This furrow deepens considerably (Fig. 113), and finally causes the curvature of the embryo to change completely from a dorsal to a ventral flexure (Fig. 114), This causes the posterior end of the body to separate from the cephalic lobes, and consequently the dorsal surface, which was formerly much shortened, now undergoes extension (Figs. 112 B, 113 and 114). The ventral surface of the posterior half of the body now lies exactly opposite to, and parallel with, the anterior half (Fig. 114), so that the tips of the extremities of the one half touch those of the other, and this flexure of the embryo again leads to the approximation of the caudal and cephalic lobes (Fig. 114, M and si). Up to this point the actual embryo, as the germ-band, has remained Fig. 113. — Embryo of Geophilus ferrugineits at the stage when the germ-band begins to flex ventrally ; seen obliquely from the ventral side (after Soqraff). at, antenna ; d, yolk ; el , point at which the band bends ; ex, limbs ; lest, germ-band. CHILOPODA — THE EMBRYONIC ENVELOPE. 227 distinct from the yolk-mass upon which it lay, as is evident from a glance at Figs. 109-113. The germ-band now, however, extends laterally and grows round the sides of the yolk-mass, so that the dorsal surface of the embryo begins to develop, and its segmentation commences (Fig. 115). At the same time the two halves of the body that lie parallel to one another lengthen, and approach more and more the final shape, although still showing the ventral curvature (Fig. 115). A cuticle was secreted at the surface of the embryo at an earlier stage. When the embryo assumed the ventral curve, the cuticle did not follow that curvature, but bridged it over, and thus remained somewhat separate from the body. In later stages, the body, as well as the anterior limbs, is found sheathed in this cuticle. The mature embryo is still enveloped in it, and it is only cast off after the egg-shell (Fig. 115, eh) has split as the first lar- val integument. In the Geophilus inves- tigated by Metsch- nikoff, a tooth is found on the cuticle covering the second maxilla (Figs. 114 and 115, ez), this, ac- cording to Metsch- nikoff, is used for splitting the egg- envelope, and is cast off with the cuticle. "We thus have here a recurrence of the structure known as the egg-tooth in the Araneae (p. 58). The provisional cuticle in any case corresponds to the envelope formed in other Myriopoda at a still earlier stage which surrounds the embryo in the same way as do the blastodermic cuticle, or the deutovum-membrane of the Acarina (cf. pp. 97 and 234). The embryo splits the egg-shell (Fig. 115, eh) at an early stage of development. It still retains the ventral curvature and is sur- rounded by the provisional cuticle. It continues to grow in length Fig. 114.— Embryo of Geophilus after completion of the ventral flexure. The ventral surface of the anterior part of the body is turned towards that of the posterior part, and lies almost parallel to it (after Metschnikoff). at, antenna ; d, yolk ; ek, point at which the germ-band bends ; ez, egg-tooth (on the second maxilla) ; kl, cephalic lobe ; p, legs ; si, caudal lobe. 228 MYRIOPODA. at the expense of the yolk now accumulated in the enteron, this growth in length proceeding as before, by the formation of new segments from the still undifferentiated caudal lobe (rf. Figs. 112- 115). The antennae are now distinctly segmented (Fig. 115, at), and the mouth-parts are approaching the adult form, but the other limbs are still simple and truncated. At a stage somewhat later than that depicted in Fig. 115, when the embryo has thrown off the provisional cuticle, Metschnikoff observed the first movements, which consisted of slow extension and flexion of the body. Metschnikoff points out that, in these movements, the extremities might better be compared with the ventral cirri of many Annelids than with the rapidly moving limbs of the Myriopoda. rd "mro Fio. 115.— Embryo of Geophihis after the splitting of the egg-shell (eh). The ventral curva- ture is still retained (after Metschnikoff). a, anus ; at, antenna ; d, yolk ; ch, egg-shell ; ez, egg-tooth on the second maxilla ; g, brain ; miu, mouth-parts ; p, legs ; si, caudal lobe ; vd, stomodaeum. Throughout the course of development that we have followed, the body of the embryo has been cylindrical, and it retains this shape for a time after it has hatched. It thus resembles in form a Diplopod, until the dorso-ventral flattening of the body character- istic of the Chilopoda takes place. At the stage of hatching, when the " larval integument " is cast, Geophilus is said to possess all its limbs, although these are still truncated (Fig. 115) and do not enable it to move with freedom. The young Chilopod probably passes through several moults before attaining the complete form and size of the mother, although when hatched it bears a strong DirLOPODA THE FIRST RUDIMENT OP THE EMBRYO. 229 general resemblance to the latter. This is also the case in the Scolopendridae, whereas, in the Scutigeridae and the Lithobiidae, the embryos leave the egg with only seven pairs of legs (not taking into account the maxillipedes). The number of legs is completed during post-embryonic development. But since the young animals in those cases also possess essentially the form of the mother, their post-embryonic development, accomplished through several moults, is fairly simple. The Chilopoda have been divided according to this distinction in the manner of their development into Chilopoda epimorpha (Scolopendridae, Geophilidae) and Chilopoda anamorpha {Scutigeridae, Lithobiidae). B. Diplopoda. The embryos of those Diplopoda whose ontogeny has as yet been investigated (Polyxenus, Glomeris, Pohjdesmus, Strongylosoma, Jidus), leave the egg-envelope at a stage when only comparatively few segments are developed and with only three well-developed pairs of legs (Figs. 121 B, p. 235, and 122, p. 237). As contrasted with the Chilopoda, which possess a large number of segments when hatched, the young Diplopoda are thus comparatively far removed from the adult form. They have been distinguished as larvae, but it should be pointed out that, in those parts of the body that are developed, they already show the organisation of the adult. The First Rudiment of the Embryo. Flexure of the Germ-band. Julus. The formation of the germ-band and the first rudiment of the embryo seem to appear in the same way as in Geophilus, but the germ-band does not in this case extend so far over the egg, and does not, therefore, assume a marked dorsal flexure. When the cephalic lobes have appeared as rudiments, the stomodaeum lies between them and the proctodaeum, almost at the posterior end of the germ-band, and when, further, the post-cephalic segments have become marked off and show the rudiments of limbs, a transverse furrow appears between the sixth and seventh segments and soon deepens. This is the same process as that which, in Geophilus, results in the transition from the dorsal to the ventral curvature (cf. Fig. 113). Since, however, the germ-band in Julus is small as compared with the mass of the whole egg, the former sinks into the yolk during this process (Fig. 116 A). The posterior and still undifferentiated portion of the germ-band now lies bent parallel 230 MYRIOPODA. to the anterior portion, the ventral surfaces of the one half turned towards that of the other (Fig. 116 A and B). In the Chilopoda, we were able to trace the inward flexure of the germ-band to the fact that, during the original dorsal curvature of the long germ-band which extends almost entirely round the egg, the development of the dorsal surface is not possible, and consequently the change to the ventral curvature takes place. In consequence of the length of the embryo, the latter is obliged to assume a bent position within the egg. The germ-band of the Diplopoda, however, is short, and the dorsal surface might very well develop without the intervention of the ventral flexure. We nevertheless find the formative processes met with in the Chilopoda recurring in the Diplopoda, and the conditions which, in the former w/v Fio. 110.— Two embryos of Jnlus Moreletti, illustrating the ventral flexure and the sinking of the germ-band into the yolk (after Metschnikokf). at, antenna ; d, yolk ; Id, cephalic lobe ; md, mandible ; Pi~p:i, first three pairs of legs ; si, caudal lobe ; uk, maxilla. case, were mechanically necessary for the development of the long embryo, are here perhaps rather adapted to the support of the embryo. It is also possible that the larger extent of surface thus brought into contact with the yolk facilitates the nutrition of the embryo. The ventral flexure was thus retained, although its primitive significance is lost. These processes are of special interest when compared with the sinking of the germ-band into the yolk in Peripatus (footnote, p. 216) and in the Insecta. In Julus the invagination of the germ-band takes place only after the rudi- ments of the antennae, the mouth-parts and three pairs of legs have appeared (Fig. 116 A), but in the Strongyhsoma, Polydcsmus, and Polyxenus it occurs at a DIPLOPODA — THE FURTHER DEVELOPMENT OF THE EMBRYO. 231 very early stage, the change from the condition in the Chilopoda, which is regarded by us as the primitive condition, having advanced still further than in Julus. Of these genera, Strongylosoma has been the best investigated, and Metsch- nikoff has shown that the first indication of the germ-band appears almost in the way that has been described. A transverse furrow, however, seems to appear very early, even before a trace of limb-rudiments is visible. The furrow deepens as in Geophilus and Julus (Fig. 117 A), and here also leads to the ventral curvature of the germ-band (Fig. 117 B). In Strongylosoma the germ-band does not sink in deep, but in Polyxenus it projects somewhat further into the yolk. We can also make out from Fig. 117 B, that the whole of the germ-band does not sink into the yolk, as the most anterior and posterior parts of it (cephalic and caudal ends) still remain on the surface. This seems indicated in Julus also (Fig. 116). Besides the transverse furrow, Metschxikoff also observed a longitudinal groove extending far forward (and no doubt back- ward as well) over the germ - band. This groove, which is here somewhat deep, corresponds to that described by Sograff in Geo- philus as appear- ing in early stages, but it seems to be much more dis- tinct in Strongy- losoma than in Geophilus (p. 222). The antennae and mouth-parts very soon develop on the anterior, sunk portion of the germ-band, and the first pair of legs develop posteriorly. In this way a stage is reached similar to that attained by the bending of the germ-band of Julus when already provided with limb-rudiments. Fig. 117. — Embryos of Strongylosoma Guerinii (A) and Polyxenus lagurus (£), to show the early invagination of the germ-band before the appearance of appendages. The dorsal part of the yolk has been omitted. In B the lamination of the germ-band due to the formation of the germ-layers can be seen (after Metschnikoff). d, yolk ; eh, egg-shell ; ek, point at which the germ-band (kst) is flexed. The further Development of the Embryo. Various factors co-operate to bring about the transformation of the ventrally flexed embryo, sunk in the yolk, into the adult form. By the growth of the germ-band towards the dorsal surface, and its simultaneous extension anteriorly and posteriorly, the yolk is taken up into the embryo. The ventral surfaces of the anterior and posterior portions of the body, till now approximated, move apart, and the whole embryo, the dorsal surface of which has also now 232 MYRIOPODA. developed, lengthens somewhat, so that a stage resembling that depicted in Fig. 118 is reached. The development of the limbs is the next point of importance. It is as difficult here, as in the Chilopoda, to ascertain with certainty the position of the antennae with relation to the mouth. The antennae, as in Geophilus (Figs. 114 and 115), are specially highly developed (Fig. 118, at). The mandibles also (md) are very large. The rudiments of the maxillae are of special importance. According to Metschnikoff's observations, which are confirmed by v. Kath, they arise out of a pair of limb-rudiments following the mandibles (Fig. 116, uJi), and are themselves immediately followed by the rudiments of the legs (p^p^). The Interpretation of the Mouth-parts of the Myriopoda. Whereas the Chilopoda have two pairs of maxillae and one pair of maxilli- pedes, the Diplopoda possess only one pair of maxillae, which have united to form a lower lip, the gnathochilarium (Fig. 122, gch, p. 237), and possess no maxillipedes. The structure of the adult gnathochilarium seems to indicate that it has resulted from the fusion of two pairs of maxillae (Fig. 120, viXi and mx2), and this view has repeatedly been adopted. It is rendered all the more probable by the fact that the Chilopoda have two pairs of maxillae (Fig. 119). These two would be homo- logous with the first maxillae and the lower lip of the Insecta. Although such an assumption seems both likely and attractive, it has not so far been sup- ported by ontogeny, accord- ing to which, as mentioned above, the gnathochilarium is derived from one jiair of limbs only (Metschni- koff, No. 11, v. Rath, No. 15). Further investi- gation of these points is, indeed, very desirable. The structure of the adult mouth- parts, however, seems to lead to a conclusion directly opposed to the above view of the composition of the gnathochilarium out of two pairs of jaws. The gnathochilarium of the Diplopoda (Fig. 120), like the first maxillae of the Chilopoda (Fig. 119, stm + me + mi), is composed of several paired pieces, and there is therefore a certain agreement between them. We may thus perhaps conclude that the whole of the gnathochilarium is homologous with the first Fi<;. 118. — Embryo of Polydesmus complanatus in a late stage of development. The egg -envelope has been removed (after Metschnikoff). at, antenna ; ch, em- bryonic cuticle ; d, yolk ; g, brain ; md, mandible ; p,-p:i, first three legs ; si, caudal lobe ; uk, maxilla. INTERPRETATION OF THE MOUTH-PARTS OF THE MYRIOPODA. 233 pair of maxillae of the Chilopoda. Another pair of limbs will then have been drawn into the formation of the mouth-parts in these forms, and will have yielded the second maxillae. That such an inclusion of a pair of legs among the mouth-parts is not inadmissible in the Myriopoda is shown by the change of the first pair of legs in the Chilopoda into maxillipedes. The second maxillae of the Chilopoda themselves do not essentially differ from legs (Fig. 119, pi), and the fact that, in the Diplopoda, the first pair of legs may shift to a position very near the head (Fig. 122, b, p. 237), does not seem without significance. According to this last view, a comparison of the mouth-parts of the Myriopoda with those of the Insecta would suggest that the gnathochilarium of the Diplopoda and the first maxilla of the Chilopoda should be homologised only with the first maxilla of the Insecta. The second maxilla of the Chilopoda and the first pair of legs of the Diplopoda, on the other hand, would correspond to the lower lip of the Insecta. The external similarity of the plate-like gnathochilarium to the lower lip of many Insects would be explicable not through a direct homology between the two structures, but merely through the similarity of their functions. The mouth-parts are fully developed even during embryonic life, and therefore have attained their definite form when the young animal hatches (v. Eath, Fig. 122). Three pairs of legs are at first developed in the larva (Fig. 121 B), but these do not always appear to belong to three consecutive segments. Thus, in the larvae of Sirongylosoma and Poly- desm/us (Figs. 121 B, and 122), the segment coming next but one after the head carries no limbs ; in the larva of Jul us, this is the case with the third segment (Xewport), in keeping with the absence of limbs on that segment in the adult. The third pair of legs is followed by the rudi- ments of several other limbs, which, however, vary in number in the different forms. These limbs are at first truncated and are hidden under the integument, only appearing as free legs during post-embryonic life. The number of segments has increased posteriorly, so that when hatched, the larva usually has from seven to nine trunk-segments, but the number of these also seems to vary Fig. 119. — The head of Lithdbius validus, seen from below (after Latzel, from Lang's Text-book of Comp. Anat.). a, antenna ; me, outer, mi, inner blade of the first pair of maxillae ; pi, palp of the second pair of maxillae ; oc, ocelli ; sk, ventral portion of the cephalic shield ; M, basal plates of the second maxillae ; stm, basal plates of the first maxilla. 234 MYRIOPODA. slightly iii different forms. The segmentation is distinguishable not merely, as in earlier stages, on the ventral surface, but is now continued towards the dorsal side (Fig. 118). "When the larva is hatched, its whole body is, as a rule, distinctly segmented (Fig. 121 B), though variations occur in this respect. The embryonic envelope of the Diplopoda (and of the Myriopoda generally*) is a structureless cuticle secreted by the superficial ectoderm of the embryo. In Julus, its secretion has already taken place before the germ-band shows any signs of segmenta- tion. This membrane which, in its origin, closely resembles the blastodermic cuticle of the Crustacea, forms a sac round the embryo and soon separates somewhat from the surface of the latter. When the germ-band attains its ventral flexure, a corresponding infolding appears in this envelope, which remains somewhat closely apposed to the ventral surface. The cuticle is retained during the course of further development, and still surrounds the embryo like a sac when the latter emerges from the egg-shell. The newly - hatched embryo of Julus in consequence somewhat resembles a maggot, as is evident from Newport's descriptions and figures. This larva, surrounded by the embryonic envelope, me. Fi<:. 120. — The gnathochilarium of Lysiopctahon cari- natum (after O. v. Rath, from Lang's Text-book of is in a lower stage of ridges (me and mi) ; mx2, the so-called tongue-plate which carries anteriorly a toothed blade (hi); mx1 and mx„, have been regarded as corresponding to a lirst and second pair of maxillae. development than other newly-hatched Diplopods. The head is not distinctly marked off from the body, nor are the segments fully developed, the germ-band not having fully extended over the dorsal surface. At this stage the larva is still incapable of movement, and may be described as a pupa. Beneath the pupal integument a 1 The embryonic integument of the Chilopoda, according to observation made on Gcophilus, shows the same characters as that of the Diplopoda (p. 227). DIPLOrODA — THE EMBRYONIC ENVELOPES. 235 second cuticular envelope is said to rise from the body. This envelope is secreted after the ventral flexure has taken place (Heathcote). Since the limbs were then developed, it no doubt shows outgrowths corresponding to the latter. The fact that two cuticular envelopes (besides the later cuticular covering of the body) form in Julus seems to be evident from Metschnikoffs description, since confirmed by Heathcote, but he does not throw much lmht on their nature. The larva of Julus passes through a resting stage within the pupal integument, and finally reaches the stage at which other Diplopoda leave the egg. The embryonic or pupal envelope, which has for some time become quite separated from the body, splits, and the larva now for the first time becomes capable of free movement. A B rg^ Fig. 121. — Two larval stages of Strongylosoma Guerinii (after Metschnikoff, from Balfour's Text-book). In A the larva is surrounded by the cuticular envelope which is provided with the egg-tooth ; in B it has shed this envelope, and has entered upon free life, at, antenna, above and posterior to which, in Fig. A, the egg-tooth can be seen; 3, A, 5, the three pairs of legs of the embryo. While, in Julus, the secretion of the cuticular embryonic envelope takes place specially early, in other forms it occurs later, when the limbs have already appeared as rudiments, so that these become separately ensheathed by it (Fig. US, cli). Its character as a larval envelope is thus more evident. In Strongylosoma a special larval organ appears as a cephalic thickening of this envelope in the form of a chitinous cone (Fig. 121 .4). According to Metschnikoff, this serves for splitting the egg-shell, and may thus be described as an egg-tooth, like the corresponding organ in Geophilus. This latter, indeed, belongs to a pair of limbs (Figs. 114 and 115, ez), and is thus not homologous with the unpaired egg-tooth of the Diplopoda, 236 MYRIOPODA. which, in form and position, more nearly resembles the unpaired egg-tooth met with in the Opiliones (p. 33), while the egg-tooth of Geophilus corresponds in position to the paired structure described in connection with the Araneae (p. 58). Structures functioning in this same way thus occur in the Arthropod a in very different positions. The egg-tooth is, in any case, cast off later with the larval cuticle. In the embryos of Polydesmus, which otherwise closely resemble those of Strongylosoma, the egg-tooth is wanting (Fig. 118), no doubt because the egg-integument in this form is much thinner, and a special organ for splitting it is thus rendered unnecessary (Metsch- nikoff). Julus also has no egg-tooth. The larvae of Polydesmus and Strongylosoma in a similar way remain (though for a very short time) in the embryonic envelope (Figs. 118 and 121), and are thus also at first incapable of free movement. The pupal stage of Polyxenus, on the contrary, could not be discovered, and Glomeris also does not seem to pass through such a stage (vom Rath). It need hardly be pointed out how strongly the blastodermic cuticle of the Crustacea, and still more the deutovum-membrane of the Acarina, are recalled by the cuticular integument of the Myriopod embryo. The resemblance to the deutovum-membrane is increased by the discovery in Polyxenus of free amoeboid cells, like the haemamoebae of the Acarina (Fig. 53, p. 99), outside the embryo and between it and the egg-integument (or the cuticle, where this is present) (Metschxikoff). "We must, however, regard this merely as an analogous condition. Post-embryonic Development. Stages of post-embryonic development are represented even while the embryo is still enclosed within the cuticular envelope, for the embryo in many cases leaves the egg surrounded by this integument, and must therefore already be regarded as a larva, and the envelope as a larval integument (Fig. 121, p. 235). It has already been mentioned that the so-called larva of the Diplopoda, apart from the small number of its segments, does not differ greatly in form from the adult. The possession of three pairs of legs brings about a striking resemblance to an Insect larva ; vom Rath points out especially its resemblance to the young Podurid. This is of course merely an external resemblance, for, in the first place, the homology of the cephalic regions of the Insecta and the Myriopoda (in respect DIPLOPODA POST-EMBRYONIC DEVELOPMENT. !37 of the number of segments utilised in the formation of the head) is still very doubtful (p. 232), and further, in the latter, one of the anterior trunk-segments, usually the second, is, as a rule, devoid of extremities (Figs. 121 and 122),* so that the first three pairs of legs are distributed on four segments, whereas the thorax of the Insecta, as is well known, consists of three segments, each possessing a pair of limbs. The transformation of the larva into the adult, the so-called anamorphosis of the Diplopoda, has been a frequent subject of investigation (Newport, Fabre, Bode, Latzel, vom Eath, and others). Variations occur in different forms, but these are not of such importance as to require special attention from us. The most important features of the post- embryonic development are the addition of new segments and the manner in which the double segment characteristic of the Diplopoda originates. The for- mation of new somites always takes place between the anal segment and that last developed (Latzel), and the formation of double segments is now proved to be due to the fusion of two of the originally distinct primi- tive segments (Heathcote). As already mentioned, the six- limbed larva has several other pairs of legs as rudiments beneath the integument. The number of these varies in different forms. The larva of Glomeris when hatched, behind the three anterior well - developed pairs of legs, has five more pairs of truncated, freely projecting limbs (vom Rath). Thus the Glomeris larva, which is said not to pass through a pupal stage, corresponds in this respect to a stage of development attained by other Diplopoda only after several moults. The first three pairs of legs in the larva of Polydesmus (Latzel, vom Rath) * [In Polydesmus the second trunk-segment is devoid of limbs both in the larval and adult condition. — Ed.] Fig. 122. — Newly-hatched larva of Polydesmus complanatus (after O. vom Rath, from Lang's Text-book of Comp. Anot.). a, antenna ; an, anus ; frj-6.,, the three pairs of legs of the larva; bt, cheeks ; gch, gnathochilarium ; Ibr, labrum ; sd, the stink-glands. Stage I. 7 segments II. 9 !) III. 12 >) IV. 15 )» V. 17 )) VI. 18 )> VII. 19 )5 VIII. 20 M 238 MYRIOPODA. are found on the first, third, and fourth trunk-segments (Fig. 122).* Two trun- cated pairs of legs lying below the integument belong to the fifth segment, and another pair of the same kind to the sixtli segment. After ecdysis these limbs project freely, and the stage with seven segments and three pairs of legs is followed by one with nine segments and six pairs of legs. The next (third) stage has twelve segments and ten ( g ) or eleven ( ) pairs of legs. The sixth segment now has one instead of two pairs, and the seventh segment of the female has two pairs, while in the male it carries only one. The copulatory limbs of the male which lie on the seventh segment develop only in the adult. The further development of segments and limbs is set forth in the following table : — 3 pairs of legs. 6 10 „ (, Malpighian vessels. THE HEART AXD THE SALIVARY GLANDS. 251 the cavities of the pseudocoele. Ventrally to the heart a peri- cardial septum forms, which has the same origin as the heart itself (Heathcote). According to Sograff, the heart of Geophilus arises from a series of paired cell-accumulations lying on the already formed intestine (Fig. 128 A). When it is said that these belong to the splanchnic layer, the splanchnic layer proper of the primitive segment can hardly he intended, but rather one of the coverings of the intestine derived from the parenchymatous tissue. This is the sense in which Heathcote also speaks of a splanchnic and a somatic layer (Fig. 126, sp and so). In the cell-accumulations, each pair of which corresponds to one of the chambers of the adult heart and to a body-segment, cavities appear (Fig. 128 B). The two sacs belonging to each pair fuse together, uniting in the middle line, and thus form a chamber of the heart, and the consecutive chambers, uniting together, form the whole dorsal vessel (Fig. 124). The heart in its development closely resembles the dorsal vessel of the Annelida which is derived from paired rudiments (Vol. i., p. 291), and, if this account of its origin prove correct, the Chilopoda would appear to show in this respect a still more primitive condition than is found even in Peripatus. The body-musculature arises out of the mesodermal elements which become applied to the ectodermal body-wall, but, as we said before, nothing certain is known of the derivation of these elements. e> The Salivary Glands. Although one cannot but be predisposed to homologise the salivary glands of the Myriopoda with those of the Insecta, and therefore to regard them as ectodermal structures, we are compelled by Heath- cote's statements to consider them as mesodermal. The salivary glands are said by him to be formed as tubular outgrowths of the somatic portion of the primitive somite of the maxillary segment, i.e., the inner division of the primitive segment (p. 248). When the tube has grown to some length it opens externally by fusing with the ectoderm on each side at the base of the maxillary plate. Should the origin of the salivary glands out of the mesoderm actually be con- firmed, we should have to regard them as transformed nephridia, whereas the salivary glands of the Insecta, in consequence of their ectodermal origin, must no doubt be considered as crural glands. The salivary glands of the Myriopoda would then be formed in the same way as those of Perijxdits (p. 206), although in the latter it is the lateral not the inner portion of the primitive segment that gives rise to them. The question as to the direct homology of the salivary 252 MYRIOPODA. glands iii the Myriopoda and Peripatus, is identified with that of the homology of the mouth-parts in the two groups. These glands, in Peripatus, belong to the segment of the oral papillae, and the question of the co-relation of these appendages to the mouth-parts of the Myriopoda will be discussed at the end of this volume. In itself, the origin of the salivary glands out of the mesoderm does not appear to us very probable. The Myriopoda, moreover, possess several pairs of these glands (Hekbst, No. 9) occurring in the individual cephalic segments, just as the pairs of salivary glands in the Insecta are distributed on the segments of the mandibles, the maxillae and the lower lip. Nothing appears more likely than that the structures in the Insecta and in the Myriopoda are homologous. It is indeed possible that, as in Peripatus, glands of mesodermal origin may occur side by side with those of ectodermal origin. It would therefore be of primary importance to establish accurately the manner of development of these glands. If the salivary or cephalic glands (Herbst applies these two names indifferently to them since spinning glands also are found among them) are of ectodermal nature, they would have to be regarded as crural glands. Such glands (probably ectodermal) are also often found on the trunk in the Myriopoda, and have been compared to the crural glands of Peripatus, and further, to the parapodial glands of the Annelida. In the Myriopoda these glands vary greatly in character. They are described in detail by Eisig (No. 2), who is disposed to regard the protective glands which were described above, according to the accounts of Metschnikoff and Heathcote, as ectodermal, as transformed nephridia. The Genital Organs. The little that is as yet known as to the formation of the genital organs relates to the genital glands of the Diplopoda. These, as in Peripatus, proceed from the somatic or inner part of the primitive segments, which, however, is not, as in Peripatus, shifted towards the dorsal side, but remains in a ventral position. A large number of primitive segments is used in the formation of the genital glands. The somatic part of these segments shifts towards the median line and comes to lie above the ventral chain of ganglia. The coelomic sacs of the right and left side of each segment come into contact in the middle line. About the time when the embryo hatches, the two fuse so that a single cavity results, and as the consecutive coelomic sacs also unite, a long tube is formed lying between the ventral chain of ganglia and the intestine. This is the genital tube. We have no altogether reliable accounts of the relation of the genital tube to the efferent ducts. In Peripatus, the efferent ducts are known to correspond to a pair of nephridia and open out posteriorly. In the Myriopoda, a special interest attaches to this question, because the genital organs in the Chilopoda open posteriorly (on the penultimate segment), in the Diplopoda, on the other hand, somewhat far forward (behind the second pair of legs). "We are inclined to regard the arrangement in the Chilopoda as the more primitive, and to assume a secondary displacement of the efferent ducts in the Diplopoda, a process which can best be explained through the utilisation of another pair GENERAL CONSIDERATIONS. 253 of nephridia. A displacement of the genital aperture by the occurrence of an additional segment behind this, the penultimate segment, and by the dis- appearance of anterior segments seems altogether improbable. The position of the genital glands themselves appears to us more primitive in the Chilopoda than in the Diplopoda, and like that of their efferent ducts, more nearly to correspond to the condition in Peripatus. The glands lie dorsally to the intestine, and in the embryo appear as two accumulations of cells near the dorsal vessel (Fig. 128 B, cj). Their origin is, unfortunately, as yet obscure. [An important contribution to the ontogeny of the Chilopoda has recently been made by Hetmons (No. III.). This author has entirely reinvestigated the development of Scolqpevdra, but unfortunately so far has only published a short summary of his results. He finds that the egg possesses a central unsegmented yolk containing segmentation-nuclei ; some of the latter migrate to the surface and form the blastoderm, while others remain within the yolk and give rise to some yolk-cells. The yolk-cells also arise as immigrants from all parts of the blastoderm, as also do the entodenn-cells. There is no gastrula-groove. The body is found to consist of the following segments : — a primary cephalic plate and a primary anal piece (telson), between which are found (1) an antennular, (2) an antenna], (3) an intercalary, (4) a mandibular, (5, 6) two maxillary, (7) a maxillipedal, (8-28) the body-, and (29, 30) the genital segments. The salivary glands are purely ectodermal and not modified nephridia. A coelomic cavity is present in each segment, cephalic plate and telson excepted, in all thirty pairs. The unpaired gonad and genital duct of the adult are paired in the embryo, and arise in connection with the coelom ; as in the Insecta, the ducts have ectodermal terminations with accessory glands. For the brain, see foot- note, p. 241. Heymoxs concludes that the Chilopoda show close relationship to the Hexapoda and are very remote from the Diplopoda, the Myriopoda not forming a natural group. — Ed.] General Considerations. In considering the ontogeny of the Myriopoda, two important questions arise — (1) whether the developmental history of the Myriopoda testifies to their near relationship to Peripatus, and (2) in what way their ontogeny is related to that of the Insects. Since the Myriopoda appear in a certain sense as intermediate forms between the Insecta and the Onychophora, these questions naturally suggest themselves. It must at once be stated that up to the present time the ontogeny of the Myriopoda is too little known to enable us to answer these questions in a manner as satisfactory as might be desired. Even with regard to the first ontogenetic processes in the Myrio- podan egg, we must hesitate in instituting a comparison with Peripatus. A superficial cleavage accompanied by segmentation of the yolk takes place in the eggs of the Myriopoda, and the same kind of cleavage has been affirmed of the egg of Peripatus novae- zealandiae, which is rich in yolk (Fig. 76 A, p. 167). The eggs of other species of Peripatus undergo total cleavage as has been seen, but this method of cleavage was regarded as probably 254 MYRIOPODA. secondary. Further, it is of interest in this connection that the eggs of the lowest Insects (Podura) are also said to undergo total cleavage, although we are still without certainty on this point. The formation of the germ-layers in the Myriopoda is still too little understood for us to draw definite conclusions from it, but the external form of the body offers a few points of comparison, though these are perhaps not very certain. It was shown that the Myriopodan egg undergoes a decided ventral folding at an early stage of development, and that this may lead to the sinking of the whole germ-band into the yolk (Figs. 113-116, pp. 226-230). This also appears to be the case in PerijJatus, judging from the statements and drawings of L. Sheldon, and it seems not impossible that these developmental processes which occur in the Myriopoda and the Insecta are foreshadowed in Peripatus.* Indications of a low grade of development, and at the same time of resemblance to Peripahis, are afforded by the suggestion of ventral organs (of head and trunk), the continuation of the primitive segments into the extremities, especially into the antennae, the condition of the ectodermal (crural) glands and of the salivary glands (which perhaps arise from the mesoderm), also by the double rudiment of the heart and by the formation of the genital glands ; but unfortunately our knowledge of the ontogenetic processes in these cases is not sufficient to raise conjecture to the level of certainty. The adult animal is better understood, and in it the constitution of the mouth-parts, the segmentation of the nervous system, the structure of the eyes, the presence of the Malpighian vessels, as well as the condition of the blood-vascular system ami the body-cavity prove without doubt the near relationship of the Myriopoda to the Insecta. By far the most striking point of agree- ment is afforded by the tracheae, which are constituted exactly like those of the Insecta. If we now ascribe great importance to a point which was not considered applicable in a comparison with the Arachnida, it is because a derivation of the long and fairly homonomously segmented Myriopoda from forms like Peripat/t* already provided with tracheae is naturally suggested, while such a^derivation of the Arachnida is met with great difficulty, as has already been shown more in detail (p. 110). In spite of the great agreement existing between the tracheal system of the Myriopoda and that of the Insecta in the adult .stage, one fact in connection with the former seems to suggest * [See footnote, p. 216.— Ed.] GENERAL CONSIDERATIONS. 255 the condition found in Peripatus. The tracheae in the Myriopoda, as in Peripatus, appear very late ; they are said to form only during post-embryonic life, whereas in the Insecta they appear as rudiments in an earlier stage of development. At first sight, the occurrence of larval forms provided with com- paratively few segments and still fewer pairs of limbs might seem to be a fact of great significance, all the more that these closely resemble in appearance the young forms of the lowest Insects, viz., the Thysanura. This brings us to the question whether the rich segmentation of the Myriopodan body represents on the whole a primitive condition, or whether it is a secondarily acquired char- acter. We might answer that the racial form of the Myriopoda was a homonomously segmented form, consisting like Peripatus of a large number of segments, or we might, with Haask (No. 5) assume that the large number of segments such as are now found in the Myriopoda represents a later acquisition by these forms. The con- tinuous lengthening of the body has been explained by the manner of life of the Myriopoda, which is accompanied by such a develop- ment of the body in the same way as in the Serpents among Vertebrates. It is interesting to see how this lengthening of the body leads to a modification of its morphological characters. In those Chilopoda whose bodies consist of many segments, unpaired •chitinous plates appear in the soft intersegmental ventral integument, which in the shorter Chilopoda is only slightly developed, and as the length and the number of the segments increase, these become broad ventral plates, the unpaired scuta (Haase, No. -6). The common primitive form of the Insecta and the Myriopoda has repeatedly been sought in some form approaching the Symphyla : but Scolopendrella, to which, on account of its striking resemblance to the Thysanura (Figs. 192 and 193), this great significance was attached, shows as well as the latter certain peculiarities of organisa- tion which prevent it as much as the Thysanura from being regarded as a primitive form. We do not indeed doubt that the iSymphyla as well as the Thysanura are very ancient forms, but we would assume a racial form for the Myriopoda with still more primitive organisation, the Symphyla being somewhat removed from that form and the Thysanura still further. The differentiation of a thorax, which is an important character of the latter, but which is merely indicated in the Myriopoda, will be discussed in dealing with the Insecta. The Diplopoda resemble most insect larvae in leaving the egg 256 MYRIOPODA. at a stage with few segments and with only three well developed pairs of legs. The Chilopoda when hatched always have a large number of segments and pairs of legs, some even possessing the full adult number. We should feel inclined to consider this as the more primitive condition, especially as Peripatus also possesses the full number of segments when hatched, did not the whole organisation make it appear doubtful which of the two divisions, the Chilopoda or the Diplopoda, is the more primitive. The invagination of the embryo, or ventral flexure of the germ- band of the Chilopoda, as well as their further development, seems to take place in a primitive way, as it appears to be merely a consequence of increase in length, while the early bending of the germ-band of the Diplopoda does not admit of such a natural explanation, but must rather be regarded as a derived condition. On the other hand, the cylindrical form of the Diplopodan body seems to represent a more primitive condition, since the Chilopodan embryo also is cylindrical and becomes flattened dorso-ventrally only after hatching. Whereas, in the Chilopoda, each body-segment carries a pair of limbs, in the Diplopoda we see every two segments fusing together to form one, which is then provided with two pairs of limbs. Ontogeny has shown that for every segment of the Diplopoda, two primitive segments and two ganglia appear as rudiments ; the double nature of these segments can thus no longer be questioned. In this we certainly have a secondary character in the Diplopoda. The mouth- parts of this division, nevertheless, are far simpler than those of the Chilopoda, in that the former probably possess only one pair of maxillae, while in the Chilopoda two more pairs of extremities are drawn in to assist this pair in the work of mastication. The tracheal system is simpler in the Diplopoda and more complicated in the Chilopoda, but, on the other hand, a more primitive condition of the genital organs is found in the latter, the genital glands first appearing dorsally to the intestine (as in Peripatus) and retaining this position, while in the Diplopoda they are found ventrally to the intestine. In the former, the genital aperture belongs to the penultimate body -segment, whereas, in the latter, it lies near the anterior end of the body, between the second and third trunk- segments. It can hardly be doubted that the position of the genital aperture at the posterior end of the body represents the primitive condition, and that in other cases that condition has been modified. When it is further added that the Diplopoda appear palaeonto- LITERATURE. 257 logically as the older, and the Chilopoda as the younger forms, it only remains to be said that the latter also became separated from the racial form early, and each branch, developing further inde- pendently, while retaining ancient features, acquired characters which, in consequence of their separate development, were not the same in the two groups. The most important feature in the organisation of the Myriopoda is the uniform development of the trunk-segments and the possession of limbs on all or nearly all these segments. This feature gives the Myriopoda a specially primitive character, and brings them near to those other forms which show a homonomous segmentation of the body, viz., Peripatus and the Annelida. [In the above account the Myriopoda are treated as a natural group, and as such they are probably regarded by the majority of zoologists. This, however, is not the opinion of those who have specialised in this group, and a separation of the Diplopoda from the Chilopoda was suggested as early as 1887 by Pocock, a view which he afterwards amplified (No. VI.). Pocock regards the Chilopoda with the Symphyla as much more nearly related to the Hexapoda than to the Diplopoda with the Pauropoda, and he proposes to divide the Tracheata into two groups— (1) the Opisthogoneata, including the Hexapoda, Chilopoda, and Symphyla ; and (2) the Progoneata, embracing the Diplopoda and the Pauro- poda, the latter, according to Kexyox, being modified Diplopods. He regards the Symphyla as standing nearest the ancestral form of the whole group. Precisely similar conclusions have been arrived at by Silvestri (No. IX.), who also has made a special study of the Myriopoda. These views have been accepted by Ray Lankester (No. V.) and others. Kingsley (No. IV.), while agreeing that the Myriopoda form a purely artificial group and that the Chilopoda are closely related to the Hexapoda, differs from Pocock in concluding that the Diplopoda have probably no relation at all to the Chilopoda. The view that these two subdivisions of the Myriopoda are quite distinct from one another, or at the most but slightly related, is amply confirmed by the study of their ontogeny, as may be seen from the numerous points of difference mentioned above. — Ed.] LITERATURE. 1. Bode, J. Polyxenus lagurus. Ein Beitrag zur Anatomie, Morphologie u. Entwicklungsgeschichte der Chilognathen. Zeitschr. f. d. ges. Naturwiss. Bd. xlix. Halle, 1877. 2. Eisig, H. Die Capitelliden. Monographie der Fauna, u. Flora des Gol/es von Neapel. xvi. Berlin, 1887. 3. Fabre, L. Becherches sur l'anatomie des organes reproduc- teurs et sur le developpement des Myriopodes. Ann. Sci. Nat. (4). Tom. iii. Paris, 1855. 4. Grenacher, H. Ueber die Augen einiger Myriapoden. Archiv. f. Mih-o. Anat. Bd. xviii. 1880. 5. Haase, E. Beitrag zur Ontogenie u. Phylogenie der Chilopoden. Zeitschr. f. Entomologie. N. F. Heft viii. Breslau, 1881. s 258 MYRIOPODA. 6. Haase, E. Die Abdominalanhange der Insekten mit Beriick- sichtigung der Myriopoden. Morph. Jahrb. Bd. xv. 1889. 7. Heathcote, F. G. The early development of Julus terrestris. Quart. Journ. Micro. Sci. Vol. xxvi. 1886. 8. Heathcote, F. G. The post-embryonic development of Julus terrestris. Phil. Trans. Roy. Soc. London. Vol. clxxix. 1888. 9. Herbst, C. Anatomische Untersuchungen an Scutigera coleop- trata. Dissertation. Jena, 1889. 10. Latzel, R. Die Myriopoden der Oesterreich-Ungarischen Monarchic Wien, 1880 and 1884. 11. Metschnikoff, E. Embryologie der doppeltfiissigen Myrio- poden (Chilognatha). Zeitschr. f. Wiss. Zool. Bd. xxiv. 1874. 12. Metschnikoff, E. Embryologisches iiber Geophilus. Zeitschr. f. Wiss. Zool. Bd. xxv. 1875. 13. Newport, G. On the organs of reproduction and the develop- ment of the Myriopoda. Phil. Trans. Roy. Soc. London. 1841. 14. Packard, A. S. On the Morphology of the Myriopoda. Proc. Amer. Phil. Soc. Philadelphia. Vol. xxi. 1884. Also in Ann. Mag. Nat. Hist. (5). Vol. xii. 15. Rath, O. vom. Beitriige zur Kenntniss der Chilognathen. Inaug. -Dissertation. Bonn, 1886. 16. Rath, 0. vom. Ueber die Fortpflanzung der Diplopoden (Chilognathen). Ber. d. Naturforsch.-Gesellsch. Freiburg i. Br. Bd. v. 1891. 17. Rath, 0. vom. Zur Biologie der Diplopoden. Ber. d. Natur- forsch.-Gesellsch. Freiburg i. Br. Bd. v. 1891. 18. Saint Remy, G. Contributions a l'etude du cerveau chez les Arthropodes tracheites. Archiv. Zool. Exper. (2). Tom. v. Suppl. 1887-1890. 19. Sograff, 1ST. Zur Embryologie der Chilopoden. Zool. Anz. Jahrg. v. 1882. 20. Sograff, N. On the embryonic development of Geophilus ferrugineus and G. proximus. (Russian). Lzvyest. Imp. Obshch. Lyub. Estestv. Antrop. i Ethnog. Moscow. Tom. xliii. 1883. (Abstracted in Zool. Jahresb. d. Stat. Neapel. 1883.) 21. Stecker, A. Die Anlage der Keimblatter bei den Diplopoden. Arch. f. Mihro. Anat. Bd. xiv. 1877. LITERATURE. 259 22. Tomosvary, E. Eigenthiimliche Sinnesorgane der Myriopoden. Math. Xaturwiss. Berichte aus Ungarn. Bd. i. Pest and Berlin, 1882-83. APPENDIX TO LITERATURE OX MYRIOPODA. I. Adensamer, T. Zur Kenntniss der Anatomie und Histologie von Scutigera coleoptrata. Verh. Z. But. Ges. Wien. Ed. xliii. Ueber das Auge von Scutigera coleoptrata. Sitzungsber. Z. Bot. Ges. Wien. Bd. xliii. 1893. II. Cholodkovsky, N. A. Zur Embryologie der Diplopoden. C. R. Soc. St. Petersburgh. 1895. (Printed in Russian. Resume in German.) III. Heymons, R. Zur Entwicklungsgescbicbte der Chilopoden. Sitzungsber. k. preuss. Almd. Wiss. 1898. IV. Kingsley, J. S. The classification of the Arthropoda. Amer. Nat. 1894. V. Lankester, E. Ray. Are the Arthropoda a Natural Group 1 Nat. Sci. 1897. VI. Pocock, R. I. On the classification of the Tracheate Arthro- poda. Zool. Anz. Jahrg. xvi. 1893. VII. Rosenstadt, B. Zur Morphologischen Eeurtheilung der Augen von Scutigera. Zool. Anz. Jahrg. xix. 1896. VIII. Schmidt, P. Beitrage zur Kenntniss der niederen Myriapoden. Zeitschr. f. Wiss. Zool. Bd. lix. 1895. IX. Silvestri, F. I. Diplopodi. Part i. Sistematica. Ann. Mus. Civ. Genova (2a). xvi. 1896. X. Verhoeff, C. Zur Phylogenie der Myriapoden-Ordnungen. (Two papers.) Zool. Anz. Jahrg. xix. 1896. XI. "Willem, V. Les ocelles de Lithobius et Polyxenus. Bull. Soc. Mai. Belg. Tom. xxvii. CHAPTER XXVII. INSECTA. Systematic (after Brauer, No. 146) : — A. Apterygogenea. «. Thysanura {Campodea, Japyx, Machilis, Lepisma). b. Colleinbola (Podura, Sminthurus). B. Pterygogenea. a. Dermaptera (Forficula). b. Ephemeridae. c. Odonata {Libcllulidae). d. Plecoptera (Pcrlidae). c. Orthoptera genuina {Blattidae, Phasmidac, Mantidae, Saltatoria). f. Corrodentia (Termitidae, Psocidae, Mallophaga), g. Thysanoptera (Physapoda, Thrips). h. Rhynchota. Neuroptera (Sialidae, Megaloptera). Panorpatae. Trichoptera (Phryganea). m. Lepidoptera. n. Diptera. o. Siphonaptera. p. Coleoptera. q. Hynienoptera. I: I. Homomorpha. - Heteromoi-pha. I. Embryonic Development. 1. Oviposition and the structure of the ripe egg. Most Insects are oviparous, only a few forms bringing forth their young alive, e.g., the parthenogenetic generations of the Aphidaey many Diptera (Sarcophaga, Tachina, Oestridae, Pupipara, Cecidomyia larvae), the Stylopidae, and a few Coleoptera (many Staphylinidae). The eggs when laid are protected from external injuries in many different ways, either by being glued to some surface or by being deposited in water, below ground, or within the tissues of plants. In the last case the laying of the eggs often gives rise to excrescences on the plants (galls). Insects whose larvae live as parasites in the OVIPOSITION AND THE STRUCTURE OF THE RIPE EGG. 2G1 body-cavity of other Insects (Ichneumonidae) lay their eggs in the body of the future host within which the embryonic and post- embryonic development take place. Many Insects surround their eggs with a web, others envelop them in a secretion which, in the ■case of eggs laid in water, swells up as a jelly (spawn of the Culicidae and Phryganeidae), but in other cases hardens into a firm capsule by exposure to the air (so-called egg-case or cocoon of Periplaneta and Mantis), etc. The eggs of Insects are usually distinguished for their large size. They show great variety of form, the most prevalent being an elongate oval, the long axis of the egg corresponding with the future long axis of the larva. In such eggs a distinction between the future dorsal and ventral surfaces is indicated by a difference in curvature (Fig. 129, c? and v). The mature egg is enclosed within two envelopes, an inner vitelline membrane (Fig. 129, dh) secreted by the egg itself, and an outer chorion (ch) secreted by the epithelium of the ovarian tube. The latter occasionally breaks up into two layers, the endochorion and the exochorion. The vitelline membrane is usually a homogeneous delicate structureless membrane, but the chorion is seldom thus con- stituted. In most cases it is ornamented by the presence of a network of thickened ridges and markings, which vary greatly in the different genera and species. The chorion is pierced at one or more points (micropyles, Fig. 129, m) to allow of the passage of the spermatozoa, and the modification of the chorion that takes place round these micropyles often results in a very complicated micropylar apparatus, round which the vitelline membrane appears attached to the chorion (Fig. 129), so that both membranes appear perforated at this point. In Insect eggs there is always a distinction between an anterior and a posterior pole. The anterior pole is that which, in the body of the mother, lay directed towards the head, and thus corresponded to the upper end of the ovarian tube. In later stages of embryonic development, the head end of the embryo always lies at this pole, while the posterior end of the embryo is directed towards the posterior pole of the egg. The micropylar apparatus usually lies at the anterior pole of the egg. A cap of glutinous matter often covers the micropyle-area (Fig. 129, g), and may extend as an envelope over the greater part, or even the whole of the egg. In the egg itself there is usually a separation of a superficial layer consisting of formative protoplasm (Fig. 129, A') from an inner mass 262 INSECTA. T ~ composecl principally of food-yolk (do). The peripheral layer of protoplasm, the periplasm or perivitellus ("Keimhautblasteni" of Weismann, No. 87), has only been found wanting in a few cases,. but it is usually quite thin and incon- siderable, and, as compared with the amount of the central food-yolk, it might almost be thought to be dis- appearing. Only the eggs of certain small Insects are found to be com- paratively poor in yolk. Some of these are forms in which the larva which emerges from the egg is dis- tinguished by its small size (oviparous Aphidae), or in which the nourishment of the embryo is provided for in some other way, either by its development within the body of the mother (vivi- parous Ajriiidae), or by the embryo passing through its development endo- parasitically in the coelomic fluid of some other Insect (Ichneumonidae). In all these forms the poverty in yolk has a determining influence on the course of embryonic development as will be seen later. These modifications are probably secondary, and an ovum well provided with food-yolk is no doubt to be regarded as the primitive type of Insect egg. The central yolk-mass in the Insect egg (do) consists of a fine network of formative protoplasm, within the meshes of which the numerous particles of food-yolk and spherical fat-drops are contained. The elements of the food- yolk appear as strongly refractive bodies which are spherical or polygon ally flat- tened by mutual pressure, and appar- ently structureless and homogeneous. The germ-vesicle of the maturing insect egg lies in the central part of the yolk, and appears as a large vesicular nucleus provided Fni. 129. — Diagrammatic median sec- tion through the egg of Musca at the stage of fertilisation (taken from drawings by Henking and Blochmann). ch, chorion ; d, dor- sal side of the egg ; dh, vitelline membrane ; do, food-yolk ; g, glu- i mi 'us cap over the micropyle; k, peripheral protoplasm (periplasm or perivitellus); m, micropyle; p, male and female pronucleus before fusion ; r, polar bodies ; v, ventral side of the egg. CLEAVAGE AND THE FORMATION OF THE BLASTODERM. 263 with a delicate chromatine network. During the process of maturation it shifts to the surface of the egg, there a spindle is formed, and the nucleus undergoes division. In this way the first, and afterwards the second, polar body is formed (Blochmann, No. 5). The position of the polar spindle varies in the different groups of the Insecta. In some (Picris) it lies directly at the anterior pole of the egg, but in most Insects it is formed nearer the middle of the long axis of the egg. Blochmann (No. 5) found it in Blatta in the middle of the dorsal surface, while in Musca it occurs at about one-third to one-fourth of the whole length behind the anterior pole on the concave (dorsal) side of the egg (Fig. 129, r). In the Formicidae it also lies laterally, but near the anterior pole, while in the Aphidac it is situated at the middle of one of the lateral surfaces of the egg. In Hydrophilus (Heider) it lies somewhat behind the mid-lateral region. 2. Cleavage and the Formation of the Blastoderm. The type of cleavage universally prevalent among the Insecta* is the pure superficial type so common among the Arthropoda (Vol. ii., (p. 113). The first cleavage-nucleus (produced in the fertilised egg by the union of the male and female pronuclei) shifts inward and increases by indirect nuclear division (Figs, 131 A and 175 A, B, C, /). The formation by division of the numerous cleavage-nuclei from the first has only been directly observed in those eggs that are poor in yolk (Aphidae, Cecidomyia, Cynipidae). But it can hardly be doubted that in the larger yolk-bearing eggs of other Insects the numerous cleavage-nuclei which are found distributed throughout the egg soon after it has been laid are actually pro- duced from the first cleavage-nucleus by nuclear division. These numerous nuclei, with the star-like areas of protoplasm which sur- round them, represent the formative elements of the blastoderm. Tichomiroff has, however, conjectured in the case of the egg of Bombyx, and Henking (No. 39) has more definitely maintained in the case of Mtisca that these nuclei in their protoplasmic islands distributed throughout the yolk-masses have been produced by free formation of nuclei.t But this view appears to us altogether un- tenable. It is contradicted by the observations of Blochmann (Xo. 5), according to which all the cleavage-nuclei in Musca undergo 1 Ul.tanin (Xo. 83) believes that total and equal cleavage occurs in the Poduridae. It appears, however, from the researches of Lemoine that even here cleavage is superficial, and the same conclusion is arrived at by Grassi (No. 33) from the condition of the food-yolk in the later stages in Jcqnjx. [According to Henneguy (No. XII.), cleavage is total in the eggs of Smicra, one of the Chcdcididae. This is obviously a derived condition, as these are parasitic forms and the egg is nourished at the expense of the host. — Ed.] t [Henking (No. XL) has since modified this statement. See also footnote p. 167.— Ed.] 264 INSECTA. division simultaneously (Fig. 130 A A jXx, OR . ■ oH^o^L " -eta, - *««*:* Fig. 130. — Stages of blastoderm-formation in Musca (Callipkora) vomitoria (after Blochmann). The drawings represent parts of sections through four eggs. A, the nuclei of the cleavage-cells have become arranged parallel to the surface of the egg. ]:, the cleavage-cells fuse with the peripheral protoplasm. C, the surface becomes in- dented by furrows ; all the nuclei of the blastoderm-cells are shown in the act of division. D, the blastoderm-cells in the form of long cylindrical epitheloid cells. 6, peripheral protoplasm ; bz, blastoderm- cells; d, food-yolk; dz, yolk-cells; fz, so- called cleavage - cells ; i, inner peripheral protoplasm. O), which indicates that they represent a generation of descen- dants of the first cleavage-nucleus, all being of the same age ; and it is further disproved by the direct observations above mentioned as being made on some small eggs. According to Weismanx (No 89) in lihodites and Biorhiza aptera (Ci/ni- pidae) the first cleavage-nucleus divides at first into two nuclei which shift apart in the direction of the longi- tudinal axis of the egg, and, according to their positions, are known as the anterior and posterior "pole nuclei."* While the anterior nucleus remains in- active for some time, the posterior, by a kind of budding (?), gives rise to numerous nuclei, which take part in the formation of the blastoderm. The anterior nucleus, on the contrary, after the completion of the blastoderm, is said to produce by division the nuclei of the so-called inner germ -cells or yolk-cells. The process of the formation of the blastoderm in larger eggs O Oct rich in yolk was first followed in detail by Bobretzky (No. 6) and Graber (No. 149), with the help of sections. More recently Blochmann (No. 5) has made investigations on the Muscidae with which the statements of Heider (No. 38) concerning Hij- drophilus agree. The cleavage- nuclei at first lie at the centre of the egg, more or less in the longi- tudinal axis (Fig. 131 A). Each of these nuclei (/) is surrounded by a star-like mass of protoplasm, and the whole is therefore not * [This term is used only to describe the position of these two first cleavage- nuclei in the elongate egg, and has no connection with the nuclei of the polar bodies, or with the "pole-cells," p. 352.— Ed.] CLEAVAGE AND THE FORMATION OF THE BLASTODERM. 265 unlike a wandering amoeboid cell. Since, however, all these proto- plasmic islands are connected by a fine network of processes, the whole egg constitutes a syncytium, the yolk being contained in the meshes. Nevertheless, although these formative centres are not yet distinctly marked off they are already called (though inaccurately) cleavage-cells. In later stages these " cleavage-cells " shift somewhat nearer the surface of the egg, and become arranged to form a sphere (Fig. 130 A, 131 B) almost parallel to the latter. In sections of the egg they therefore appear arranged as a circle (Fig. 130 ^4). Gradually, during further processes of division, they reach the surface of the egg and fuse with the peripheral protoplasm found there (Figs. 130 B and Fig. 131.— The formation of the blastoderm in HydropMlus (after Heider). b, developed blas- toderm ; d, food-yolk ; /, so-called cleavage-cells ; fc, peripheral protoplasm; z, yolk-cells. 131 C). Division into separate cell- territories corresponding to the cleavage-nuclei now takes place (Figs. 130 C and 131 D) by the appearance of furrows that press in from the surface, and gradually traverse the whole of the peripheral protoplasm (Fig. 130 D). After the surface of the egg has been covered in this way with an epithelium (blastoderm), there follows, in many Insects (Chironomus, Musca, H>jdro])hilus), the separation of the so-called inner peripheral protoplasm (Fig. 130 D, i), i.e., of a layer of protoplasm containing coarse granules which develops between the blastoderm and the central yolk-mass. By taking up this layer of protoplasm, the blastoderm -cells increase in height, and now form a cubical or "266 INSECT A. cylindrical epithelium which continuously covers the surface of the egg. The point at which the cleavage-cells first reach the surface varies in the different groups of Insects. In the Muscidae, according to Graber, the formation of the blastoderm first commences at the posterior pole of the egg, while in Apis (Kowalevsky), Pieris (Bobretzky, No. 6), and Chironomus (Weismann, No. 89), the first blastoderm-cells were noticed at the anterior pole. In Hydro- philus (Heidek, No. 38) the blastoderm first forms round the middle of the egg as a transverse girdle, somewhat nearer the posterior pole of the egg, and develops last at the poles. In Blatta (Wheeler) and Gryllotalpa (Korotneff) the first cells forming the blastoderm appear on the future ventral surface. As it is at this side that the rudiment of the germ-band arises, the early appearance of the blastoderm-cells at this part recalls the premature development of the blastoderm often occurring in the Crustacea in the region of the embryonic germ-zone (Vol, ii., p. 115). A similar development is found in Occanthus (Ayers, No. 1). A method of blastoderm-formation differing somewhat from the above and more normal type has been observed in some Orthoptera (Blatta and Gryllotalpa). As a rule, the "cleavage-cells" increase within the food-yolk so rapidly that when they reach the surface of the egg they are closely crowded together, and here at once constitute a continuous epithelium, but this is not the case in Gryllotalpa (Weismann, No. 89, and Korotneff, No. 47) and in Blatta (Wheeler, No. 95). In these forms the first "cleavage- cells," which are comparatively few in number,* migrate to the ventral surface of the egg and there multiply so that separate cell- islands are temporarily formed. Only in later stages do the cleavage- cells, greatly increased in number by division, become distributed equally over the whole surface of the egg. It was maintained by Wheeler that, in Blatta, when the amoeboid cleavage-cells had reached the surface of the yolk, their nuclei no longer showed mitotic division, but here (as well as later in the serosa) multiplied by direct or amitotic division. The question as to the origin of the so-called yolk-cells or vitello- phags is of importance. It has been observed that, as a rule, not all the " cleavage-cells " shift to the surface to take part in the formation of the blastoderm, but that a few remain behind within the yolk (Fig. 130 D, cte, and 131 C, D, z), where they increase in number, obtain equal distribution throughout the yolk, and become the so-called yolk-cells, whose function is to liquefy the mass of food-yolk and to bring about its assimilation. The origin of the yolk-cells from cleavage-cells which have remained in the yolk has * [In Blatta they are numerous, sixty to eighty cells being found scattered in the yolk before migration commences. — Ed.] CLEAVAGE AND THE FORMATION OF THE BLASTODERM. 267 recently been definitely maintained among others by Kowalevsky, Blochmann, F. Schmidt, and Graber (No. 28) for Muscidae, by \Vheeler for Doryphora, and by Heider for Hydrophilus. Patten, on the contrary, has proved in connection with the egg of a Phry- ganid (Neophylax), and Wheeler in connection with that of Blatta, that in these forms all the "cleavage-cells" migrate to the surface and take part in the formation of the blastoderm, so that there is a stage at which the surface of the egg is covered by the blastoderm, while the centre of the egg is devoid of nuclei. In these cases the Pig. 132. — Two diagrammatic sagittal sections through an insect-embryo to illustrate the development of the embryonic envelopes. In A, the germ-band (/.-, k') is not completely grown over by the amniotic fold. In B, the amniotic folds have united and completely cover the germ-band, a, anterior, b, posterior pole of the egg ; v, ventral ; d, dorsal ; af, amniotic fold; ah, amniotic cavity; am. amnion; do, food-yolk; ec, ectoderm; /;, cephalic end of the germ-band ; V, posterior end of the germ-band ; s, part of the serosa derived from the amniotic fold; .?', part of the serosa derived from the undifferentiated blastoderm; u, lower layer. so-called yolk-cells only appear later, single blastoderm-cells wander- in g again into the interior. As we shall see later, even in the DO ' forms first described, a secondary increase of yolk-cells takes place by immigration from the blastoderm (or from the germ -band), these forms, in which all the cleavage-nuclei reach the surface and in which the immigration of the yolk-cells only takes place later, 268 IXSECTA. perhaps represent the primitive condition, while in most Insects there is a kind of abbreviation of development which causes some of the cells to remain within the yolk from the first. Cf. in this connection the formation of yolk-cells in the Crustacea (Vol. ii., p. 144), Arachnida (pp. 43-45), and Myriopoda (p. 221). In the case of the Aphidae also, Will (No. 97) has maintained that the yolk- cells arise exclusively through the immigration of cells from the blastoderm during its formation. As a rule, all the blastoderm-cells are at first of the same shape and size. An exception is afforded by the eggs of the Diptera, in which the so-called pole- cells, to be described later (p. 352), which represent the early differentiation of the genital rudiment, present us with elements which for a moment, indeed, are incorporated in the blastoderm, but are distinguished by their size and their contents from the blastoderm-cells {cf. Fig. 174 C, })z, p. 353, and Fig. 175 B, p, p. 354). 3. The Formation of the Embryonic Rudiment and the Embryonic Integuments. A. General view of the Germ-band and the Germ-envelopes. The embryonic rudiment in the Insecta, as is often the case in the Arthropoda, takes the form of a long band-like thickening, usually extending along the ventral side of the egg, this being known as the germ-band, embryonic band (Fig. 134 E). In most cases the boundaries of the future body-segments are already indicated on this band by consecutive transverse furrows. A cross section through the germ-band of an insect (Fig. 133 B and C) shows it to be multi- laminar. It consists* of an outer layer of cells, the ectoderm (ec), * We shall here therefore give the name "germ-band " to the whole embryonic rudiment in contradistinction to the transitory portion of the egg, which com- prises the food-yolk with its vitellophags and the embryonic envelopes. Such a use of the term " germ-band " is universal in connection with the Arthropoda. It should, however, be pointed out that in the Hirudinea (Vol. i., p. 321) this term is used in another sense, and onl}7 embraces a part of the embryonic rudiment. Indeed, the expression "germ-band" is occasionally used as the equivalent of "mesoderm-bands." t According to the published statements, we must assume that cellular embryonic envelopes are not present in the Apterygogenea. They are said to be wanting in the Poduridac (Ul.iaxix, No. 83). A cuticular larval integument, such as is repeatedly found in other groups of Arthropoda (Arachnida and Myriopoda, pp. 9, 58, 97), is said to form in this case. This may be provided with prominences to assist in splitting the egg-integuments, and its presence has been definitely proved by the observations of SoMMER (No. 76) and Lemoine (No. 51). Indeed, it appears that the Podurid embryo passes through several moults before batching. From this fact the absence of the amnion might be concluded. GRASSI (No. 33) who observed a dorsal organ in Japyx which occurs in the same way in the Poduridac, sees in this a proof of the presence of the amnion. Since, however, in the Poduridac, this organ develops in the earliest stages of the formation of the germ-band, it seems doubtful whether we may compare it with the dorsal organs developing by the involution of the serosa in the higher Insects (p. 304). We must therefore await further investi- gations of these points. (See p. 304, and Hevmoks, No. XVI). THE GERM-BAND AND THE GERM-ENVELOPE and an inner layer which comprises n the entoderm and the mesoderm, and as long as these two germ- layers are not sharply distinguished one from the other, this is known as the inner or lower layer (u). It is characteristic of the Insecta, but only rarely occurs in other Arthropoda (e.g., in the Scorpiones, Fig. 3, p. 5), that the germ-band does not remain freely exposed on the surface of the egg, but is grown over by an amniotic fold rising from its edges (Figs. 132 A, of, and 133 B, af), so that the former appears somewhat sunk below the ventral surface. As the amniotic fold extends from all sides over the germ-band, a cavity is enclosed between the two. This is the amniotic cavity (ah), which, when the amniotic folds have completely covered the germ-band and have united over it, appears as an en- tirely closed cavity (Figs. 132 B and 133 C). (See footnote t, p. 268.) The germ-band, after its develop- ment, thus appears covered by a double cellular envelope derived from the amniotic fold. The outer of these two envelopes is distin- guished as the serosa (s). This passes without any break of con- tinuity into that undifferentiated portion of the blastoderm which takes no part in the formation of the germ-band (Fig. 132, s), and Fig. 133. — Transverse sections through three consecutive stages in the formation of the germ-band and the embryonic envelopes of an insect-embryo. A, formation of the ventral plate (bp) and the gastrula-invagination (g). B, rise of the amniotic fold (of). C, complete overgrowth of the germ-band by the amniotic folds, v, ventral, d, dorsal surface ; of, amniotic fold; ah, amniotic cavity; am, amnion; bl, blastoderm; bp, ventral plate; do, food-yolk; ec, ectoderm ; g, gastrula-invagination ; s, serosa ; u, lower layer. 270 INSECTA. which covers the surface of the food-yolk. This part of the blasto- derm, after the completion of the envelopes, is usually reckoned as part of the serosa, so that in this sense we can say that the serosa forms a closed sac covering the whole surface of the egg (Fig. 133 C, s), one part of it enveloping the surface of the food-yolk and the other that of the germ-band.*1 The inner of the two envelopes covering the germ-band which was derived from the inner layer of the amniotic fold is known as the amnion (Figs. 132 and 133, am). This, at the edges of the germ- band, is continued into the ectoderm of the latter, the transition being in most cases quite gradual. The amnion and the ectoderm (ec) of the germ-band thus together form an epithelial sac closed on all sides, the lumen of which is represented by the amniotic cavity. Hi!-/ Pio. 134.— Ventral aspect of five stages in the development of Eydrophihis (after Heideb, from Lang's Text-book). The anterior end is directed upwards, a and b, points at which the blastopore closes ; a/, edge of the amnion-fold ; ft/', caudal fold ; a/", paired cephalic fold of the amnion; an, antenna; es, terminal segment; g, pit-like invagination (rudiment of the amniotic cavity); k, cephalic lobes; r, groove-like invagination; s, part of the germ-band covered by the amnion. The origin of the germ-band is to be sought in a thickening of the blastoderm on the ventral side of the egg (Fig. 133 A, bp). "While, as was mentioned above (p. 268), the blastodermic cells originally exhibited the same shape and size over the whole surface of the egg, they soon become differentiated in such a way that the cells of the dorsal side flatten to form a thin pavement epithelium, * The fact that Grabee (No. 27) observed in Melolontha the secretion of a cuticle from the outer surface of the serosa, after the completion of the development of the embryonic envelopes, deserves mention. A certain parallel may perhaps exist between this process and the development of the blastodermic cuticle in the Crustacea and other Arthropoda. THE GERM-BAND AND THE GERM-ENVELOPE. 271 while the cells belonging to the ventral side, owing to their more rapid division, become crowded close together, assume a prismatic form, and thus constitute a columnar epithelium. The ventral thickening of the blastoderm that has thus arisen, and which, in its extent, represents the first rudiment of the germ-band, was named by Balfour the ventral plate (Fig. 133 A, bp). The invagination of that part of the ventral plate which lies in the median line (g) gives rise to the formation of the lower layer. This invagination, which at a certain stage represents a groove running along the median line for the whole length of the germ-band (Fig. 134 A and B), must be regarded as the gastrula-invagination of the Insecta (for details, see pp. 309 et seq.). The lower layer yielded by it (Fig. 133, B and C, u) then extends beneath the whole of the ventral plate up to the edges of the amniotic fold (Figs. 133 B and 134 C). It should be mentioned that the ventral plate from its commencement is not in all cases a uniform structure, but sometimes proceeds from several distinct rudiments. Thus it has been pointed out by F. Schmidt, in connection with Musca, and by Heider in connection with Hydrophilus, that the anterior and posterior ends of the germ-baud appear first, the middle part only developing later. Another originally independent element of the germ-band is afforded in Hyclrojyhilus by the rudiments of the cephalic lobes (Fig. 134 A, k), the inde- pendent origin of which was also observed by Will (No. 97) in the Aphidae. These originally distinct formative centres only secondarily unite to form the common rudiment of the germ-band. The lateral delimitation of the germ-band seems determined by the rise of the amniotic fold, and since, when the amnion is complete, it consists of somewhat columnar cells and, even at later stages, owing to its histological character, more nearly resembles the ectoderm of the germ-band than the serosa, some investigators have assumed a closer connection between the amnion and the germ-band. "Will regards the amnion directly as a part of the germ-band, and Graber (No. 30) also derives the amnion from the thickened epithelium of the ventral plate. "We have used the term "germ-band" in the usual manner, understanding by it the segmented and already multilaminar embryonic rudiment (consisting of the ectoderm and the lower layer). It is, however, certain that this term may also be applied in a wider sense, as Graber (No. 30) has recently insisted, to the embryonic rudiment of earlier stages, in which segmentation and the forma- tion of the germ-layers has not yet commenced, presupposing that the embryonic rudiment as such is distinctly marked off from the rest of the egg. From the time of its origin onwards, the germ-band grows con- tinually in length (Fig. 134-4 to E), and in many cases extends in such a way that it no longer covers only the ventral side of the egg, but its anterior and posterior ends bend round to the dorsal side of the egg. This extension of the germ-band to the dorsal side may, in some cases (Phryganeidae, Chironomus), go so far that the anterior 272 INSECTA. and posterior ends almost come into contact (Fig. 141, p. 283). The germ-band thus appears at these early stages, i.e., roughly speaking, during the first half of its embryonic development, dorsally flexed. In the later stages, in consequence of the development of organs, and the more complicated shape thus brought about, the band shortens by contraction, so that finally the oral aperture comes to lie at the anterior pole of the egg, and the anal aperture near the posterior pole (Fig. 143, m and an). These positions of the apertures are very typical of Insect embryos at later stages. The embryonic rudiment now no longer appears curved dorsally, but is straight. Indeed, curvature in an opposite direction often takes place, the most posterior segment of the embryo appearing ventrally curved (Phryganeidae, Lepidoptera, Hydrophilus, Blatta, etc., Fig. 142 C, p. 285, and Fig. 143 B, p. 286). Graber (No. 30) lias recently pointed out that the Insects may be divided into two groups, according to the extension and the increase in the length of the germ-band. In the first group the conditions of growth of the embryo described above prevail, while, in the other forms (e.g., Blatta, Stcnobothrus) the embryonic rudiment from the very first extends over only a quite short area of the periphery of the egg, and during the whole of the later development never grows in the same way as forms belonging to the first group. In forms in which the germ-band is short, the dorsal curvature is naturally not noticeable in the earlier stages, and the germ-band appears to be straight. The growth of the embryonic rudiment in length also progresses more equally during the whole development. Shortening is not perceptible in the later stages. Insects might therefore be divided into two groups as having, on the one hand, germ-bands which are long at first and shorten later, and, on the other hand, germ-bands which remain from the first comparatively short. This distinction, however, does not appear to us to be based upon differences of any importance. B. The distinction between the superficial and the immersed Germ-band. The general description of the position and the origin of the germ-band and the embryonic envelopes, given above (p. 268, etc.), only holds good for some of the Insecta. These conditions are to be found in many Orthoptera (Blatta), the Phnjganeidae, Lepidoptera, Hymenoptera, many Diptera (Chironomus), and to some extent in the Coleoptera. In details, however, an abundance of variations which will be mentioned later are found to occur; these may be traced back to the shape of the egg, the amount and distribution of the food-yolk, and also to some extent to the vestigial condition of the embryonic envelopes. In other groups of Insects (Pseudo- neuroptera, Hemiptera), on the contrary, we find that the phenomena manifested in the formation of the germ-band and the embryonic THE SUPERFICIAL AND THE IMMERSED GERM-BAND. 273 envelopes, as well as the position of the former, differ still further from those which we described above, and which we took as a starting-point for the sake of clearness. "We must now deal with all these conditions in detail. From the above description of the rise of the amniotic fold (Figs. 132 and 133, of), it becomes evident that a cavity, con- tinuous with the space containing the food-yolk, extends in between the serosa and the true amnion. It is thus possible for spherules of food-yolk to pass into this cavity and entirely to fill it (Fig. 135). In this case the amnion and the serosa are separated from each other by a somewhat wide space filled with food-yolk, whereas, in other cases, where the food-yolk does not penetrate this cavity, -n Fig. 135.— Section through the germ-band of a Lepidopterovts Insect (combined from drawings by Bobretzky and Hatschrk). ah, amniotic cavity ; am, amnion ; c, coelomic cavity ; do, food-yolk (divided up into separate masses, each containing a nucleus) ; cc, ectoderm ; m, mesoderm ; pr, thickenings of the ectoderm, representing the rudiments of the ventral nerve-cords ; s, serosa. the amnion and the serosa are in direct contact (Fig. 158 B-F). "We may thus divide the eggs of Insects into two groups, according to the presence or absence of this space between these two embryonic envelopes. 1. Eggs in which the germ-band is superficial, i.e., in which the elements of the food-yolk have not penetrated into the space between the amnion and the serosa. The germ-band is here comparatively superficial (Figs. 132, p. 267; 133, p. 269; 140 A, p. 281; 143, p. 286). 2. Eggs in which the germ-band is sunk or Immersed, the space between the amnion and serosa being filled by particles of food-yolk. In such cases the germ-band appears, as compared with type 1, more deeply sunk within the egg (Figs. 135, 136 C-E, and 142, p. 285). 274 INSECTA. The germ-band is superficial in many Orthoptera (Oecanthus, Gryllotalpa, Blatta, Mantis), in many Hemiptera (Corixa), in the Phryganeidae, the Diptera, and the Hymenoptera. In the Coleoptera also, the greater part of the germ- band appears superficial, but its posterior end is in the first stages immersed. An immersed germ-band occurs in the Libellulidae, many Hemiptera (Pyrrho- coris), many Orthoptera (Stenobothrus), and in the Lepidoptera. C. The distinction between the invaginated Germ-band and the Germ-band that has been overgrown by the Membranes.* With regard to the manner in which the germ-band arises and to its position, there are two opposite types among the Insecta, these, however, being connected by means of transitional forms. In the one type the ventral plate is invaginated into the inner part of the egg, and in the other the amniotic folds rising from its edges grow over it. 1. When the germ-band during its formation is invaginated, e.g., Libellulidae, Brandt (No. 7), its first rudiment appears in the form of a small thickening of the blastoderm lying ventrally in the posterior half of the egg (ventral plate, Figs. 136 A, bp and 137 ^4), in the posterior region of which invagination soon takes place (Fig. 136 A, Jch). The lumen of this invagination is the first rudiment of the amniotic cavity (Fig. 1 36 B, ah) ; the thickened ventral portion of its wall (k) forms the germ-band, while the thinner dorsal portion gives rise to the amnion (Fig. 136 B, C, am). The blind end of the invagination denotes the later anal end of the germ-band (&'). Since, however, the invagination grows from behind forward in the egg, it results that the primitive position of the germ-band appears to be the exact reverse of its later position, its posterior end becoming forwardly directed, while its cephalic end lies near the posterior pole of the egg. In a similar way, that surface of the germ-band which was primarily on the ventral surface of the egg becomes secondarily turned towards the dorsal side of the egg, so that the ventral surface of the developing embryo is now seen through the dorsal wall of the egg. To bring the germ -band into its definitive position the process described as reversal, rotation, eversion, or revolution, which will be described below, is needed. It should be mentioned that in eggs of this type the anterior end of the germ- band, which is distinguished by the extension of the cephalic lobes (Fig. 136 * On this distinction rests the division of Insects into those with inner and those with outer germ-bands (inner and outer germs, or entoblastic and ecto- blastic forms, Grabek). Gkaber has recently suggested the terms entoptic and ectoptic germ-formation to describe these categories. We have not adopted these terms, because they are liable to confusion with the superficial and immersed germ-bands, given above (p. 273). THE INVAGINATED AND OVERGROWN GERM-BANDS. 275 C, J>, k), does not take part m the process of invagination. It remains on the true ventral surface of the egg, and only becomes covered by the growth of the embryonic envelopes (af) taking place in the usual manner. Thus the relation of the anterior portion of the germ-band in this type answers to the description given below for the second type. 2. When the formation of the germ-band is accompanied by the growing over of an amniotic fold, the ventral plate, and the germ- band which develops out of it retain throughout the course of development the position which is typical of the later stages in Fig. 136. — Diagrammatic median sections, to illustrate the development of the Libellulid egg (after Brandt). A-C, development of the germ-band (Jv, A'') accompanied by invagination. D, development of the amniotic folds («/) which grow over the cephalic end of the germ- band. E, the aperture of the amniotic cavity is closed, v, ventral side of the egg ; d, dorsal side ; a, anterior, o, posterior pole of the egg ; af, amniotic fold ; ah, amniotic cavity ; it, amnion ; U, blastoderm ; bp, ventral plate ; do, food-yolk ; k, cephalic end of the germ-band ; k', anal end of the germ-band ; kh, germ-prominence or commencing invagina- tion ; s, serosa. all Insect eggs. This type of development which is exemplified in the Diptera (Chironomus, Simulia, Cecidomyia) is the one already described (p. 268, etc.). The germ -band in this case essentially belongs to the ventral side of the egg. Its anterior end corresponds to the anterior pole of the egg, and its posterior end to the posterior pole (if we do not take into account the dorsal extension mentioned above, p. 271). There is therefore no reversal or rotation in this case. The embryonic envelopes are formed by simple folds which 276 INSECTA. rise from the edges of the germ-band (Figs. 132 and 133, pp. 267 and 269). If we take into consideration the position of the germ-band at the time of its formation, the two types here distinguished might also be defined as a type with the germ-band inversely placed, and one with the germ-band normally placed. It would be still simpler to define them as types either with or without the reversal or rotation. It might indeed be objected that, in the germ-band that is overgrown,, changes of position have also occasionally been observed, and these are often very difficult to distinguish from true reversal or rotation. In the order Coleoptera, we shall find forms in which the forma- tion of the germ-band affords a direct transition from one of the types above distinguished to the other. D. Insects with Invaginated Germ-band. Libellulidae. We shall first consider the egg of the Libellulidae (A. Brandt, ]S~o. 7) as the best representative of this type of develop- ment. This family, as we shall see below (p. 288), seems to exhibit conditions which might be direct modifications of those found in the Myriopoda, and which must therefore be regarded as the more- primitive. In Calopteryx, the first rudiment of the germ-band is found in a thickening of the blastoderm (ventral plate) lying in the posterior ventral half of the egg. The most posterior portion of the germ-band soon becomes pressed into the egg (Fig. 137 A, g). While this invagination, which by many authors is called the germ-prominence> continually deepens, it becomes directed forward and grows out towards the anterior pole of the egg (Fig. 137 B and C). The lumen of the invagination is the first rudiment of the amniotic cavity. A difference in the thickness of the two walls of the invagination is very soon perceptible. The dorsal wall which represents the amnion (am) becomes gradually thinner and its cells flatten, while the other wall thickens and represents the actual germ-band (ps). Almost the whole of the germ-band is here invaginated into the egg, its posterior end pointing forward. Only a small part of the band, its primitive anterior end, retains for a time the superficial ventral position of the original thickening of the blastoderm (Fig. 137 C) ; this soon broadens out to form the cephalic lobes. This part now becomes completely grown over by a circular amniotic fold derived from the surrounding blastoderm. When this circular fold closes over the cephalic lobes, the amniotic cavity is cut off from the INSECTS WITH INVAGINATED GERM-BAND. 277 exterior (Fig. 138 A). The thin layer of the blastoderm surrounding the egg, which has retained its superficial position, now represent* the serosa. In the subsequent stage, which is characterised by the possession of limb -rudiments, the remarkable position of the germ -band can be distinctly recognised (Fig. 138 A). We see that its cephalic end (v) is directed towards the posterior pole of the egg, while its hook- like posterior end (ab) is directed towards the anterior pole. We can also see, from comparison with other stages (Fig 138 C), that the ventral side of the germ -band on which the limb-rudi- ments form is turned to the dorsal side of the egg. The defi- nitive position of the •embryo is brought about by a process of reversal or rota- tion of the germ- band, the embryo undergoing rotation round its transverse axis, and being at the same time evagi- nated from the am- niotic cavity (Fig. 138 B). This pro- cess is commenced by the fusion and subsequent rupture of the amnion and serosa near the cephalic region. This rent gives rise to an opening into the amniotic cavity at the very point where the original aperture of the invagination was situated, and through this aperture first the head and then the consecutive segments of the germ-band emerge and become applied to the ventral portion of the egg-shell, the head shifting towards the anterior pole (Fig. 138 C). In proportion as the embryo emerges from the amniotic cavity, the latter diminishes in size, and finally completely disappears, the •embryo being now only surrounded by the egg-shell. As the germ-band now lies on the surface of the egg, the area Fig. 137. — Three stages of development of the embryo of Calopteryx (after Brandt, from Balfour's Text-book). The embryo is represented inside the egg-shell, am, amnion ; g, lateral edge of the ventral plate ; ps, rudiment of the germ-band ; se, serosa. 278 IKSECTA. occupied by the serosa has become considerably diminished (Fig. 138 C). This envelope now contracts towards the anterior pole of the egg, thickening considerably at the same time (Fig. 138 C, se). In consequence of this contraction, the edge of the rent where the serosa and the amnion coalesced and finally the amnion itself are drawn anteriorly over the food-yolk (Fig. 138 C, se and am), so that the two envelopes together finally form a sac lying dorsally to the embryo : this sac is filled with food- yolk, and may be defined as a kind of (dorsal) yolk-sac. As the lateral and dorsal parts of the embryo now develop further, the contents of the yolk-sac are taken up more and more into the intestinal cavity, which communicates with the sac, and are used up, so that finally, by a process to be described later, the serosa itself is (appar- ently) drawn into the ernbryo and assimilated (p. 304). Fig. 13S. — Three stages in the development of Calopteryz (after Brandt, from Balfocr's Text-book). The embryo is represented inside the egg-shell, a, secondary opening of the amniotic cavity, through which the embryo emerges ; ab, abdomen ; am, amnion : at, antenna : mrf, mandible ; nu.1, uu-, first and second maxillae ; oe, oeso- phagus ; j'l, p", i'3, the three pairs of thoracic legs ; »-•:, serosa ; v, anterior end of the germ-band. Since the germ-band of the Libellulidae arises by an invagination which grows into the interior of the egg, it is seen that the amnion and serosa are here separated by a wide space filled with food-yolk. The germ-band of the Libellulidae is therefore immersed. Its cephalic end, however, is excepted from this immersion, and so belongs to the superficial type. Rhynchota. The type of development of the germ-band described above for the Libellulidae occurs also, as far as is at present known, in all Rhynchota. Metschxikoff (Xo. 55) and Braxdt (Xo. 7) thus found in Hydrometra, and Graber (Xo. 27) in Pyrrhocoris, conditions of development of the egg which in all important points agreed with those observed in the Libellulidae, [NSBCTS WITH INVAGINATED GERM-BAND. 279 A modification of the type described is met with in Corixa (Metschnikoff, No. 55 ; Brandt, No. 7). In this form the germ-prominence, which becomes invaginated at the posterior pole, is indeed also at first surrounded by food-yolk, but very soon becomes closely applied to the dorsal side of the egg, so that the serosa and the amnion are here in close contact. The germ is consequently not immersed, but superficial. In other respects, the process of rotation and the acquisition of its definitive position by the embryo occur in exactly the same way as in the Libellulidae. The germ-bands of the Pediculina and the Mallophaga also, according to Melnikoff (No. 53), agree with regard to position with that of the Libclhilidae. But the condition in these cases is to some extent simpler, as the aperture of invagination into the amniotic cavity remains permanently open. An invaginated germ-band is also found in the Pliysapoda (Dohun, No. 21 ; Jordan, No. 44). The processes of development found in the eggs of the Phytophthires form a direct sequence to those described for the Libellulidae. The descriptions given by Metschnikoff (No. 55) and Brandt (No. 7) of the development of the viviparous Coccidae (Aspidiotus, Lecanium) show almost complete agree- ment with the Libellulidae ; and the Psyllidae also, according to Metschnikoff, seem to follow the same course. Certain peculiarities, on the other hand, are shown by the summer eggs of the viviparous Aphidae, which pass through their development within the egg - follicle. The eggs of these forms that develop parthenogenetically do not, as Will (No. 97) has pointed out, undergo the full process of maturation found in other insect eggs. The former exhibit a pre- cocious embiyonic development, which commences at the stage when only the very first phenomena of maturation are to be found in the appearance of small drops of deutoplasm. After the development of the blastoderm, the egg contains only a small amount of food-yolk, which soon disappears (primary food-yolk, Fig. 139 A, do), and in which single yolk-cells are found. In the stages that follow, however, the embryo is provided with a fresh mass of yolk (secondary yolk, pseudovitellus, sd) through the development of a kind of placental formation from the follicular epithelium of the parent (Fig. 139 A, sd). At the posterior pole of the embryo, at which the formation of the blastoderm is not fully completed, and where consequently there is a gap in the blastoderm,* a fusion occurs with the corresponding part (x) of the follicular epithelium (/). A mass of cells here develops by repeated division as an outgrowth of the follicular epithelium. This mass, by the degeneration and complete disintegration of the cells composing it, becomes transformed into an accumulation of yolk-spherules (secondary yolk), and the yolk -material thus produced is projected into the interior of the embryo through the gap in the blastoderm (Fig. 139 A, sd). The secondary yolk-mass, which thus comes to lie in the primary body-cavity, and into which yolk-cells (dz) soon wander from the embryo, remains for some time connected by means of a strand of yolk with that part of the follicular epithelium from which it originated. The development of the germ -band takes place in the Aphidae by an invagination at the posterior pole in exactly the same way as in the Libellu- lidae. This invagination develops round the gap in the blastoderm already mentioned (Fig. 139 A). It is consequently not closed at its inner end, this being the aperture through which the secondary yolk enters into the interior * This gap has been called the blastopore by "Will, and the immigration of yolk-cells proceeding from this point has been assumed to be gastrulation, a view which we cannot share. 280 IXSECTA. of the egg. Only after the secondary yolk has withdrawn into the primary body-cavity, and the connective strand has been absorbed, does this aperture or gap close (Fig. 139 B), and the invagination then assumes a shape exactly recalling the corresponding stage in the LibclluUdac. As the rudiment of the germ-band continues to grow, it develops a hook-like curvature (Fig. 139 C), Fig. 139. — Diagrammatic median sections through five stages in the development of the egg of a viviparous Aphis (adapted from Will). The orientation agrees with that in Fig 136. The genital rudiment is omitted. A, invagination of the germ-band (fc') and growing in of the secondary yolk (sd). B, closing of the pore through which the secondary yolk was taken in. C, hook-like flexure of the posterior end of the germ-band (k"). D, rise of the amnion-folds («/). E, development of the cephalic serosa (s'). of, amniotic folds; ah, amniotic cavity; am, amnion; do, remains of the primary food-yolk with its yolk-cells; dz, yolk-cells; /, follicle-epithelium; k, cephalic end of the germ-band (cephalic lobe3) ; k', posterior section of the germ-band ; k", posterior end of the germ-band bent in like a hook; I, primary body -cavity; s, serosa; s', cephalic serosa; sd, secondary yolk; x, point at which the secondary yolk forms. which may later become double (Witlaczil, No 98). Certain changes in position also take place. The curved rudiment, which is at first symmetrical with regard to the median plane, is soon too long to retain its position, and certain deviations in a lateral direction occur. The outer aperture of the INSECTS WITH INYAGINATED GERM-BAND. 281 amniotic cavity, which originally belonged to the posterior pole, in the course -of further development shifts more to the dorsal side. At the same time, the rudiments of the cephalic lobes (k), which have arisen as blastodermic thicken- ings and which formerly lay at the anterior pole of the egg, shift backward over the ventral side so that finally they extend over the posterior pole (Fig. 139 D). The whole of this blastodermic thickening is not, as in the Libellulidae, included in the invagination of the germ-band, and its true anterior end is therefore not at first covered by the embryonic envelopes. Soon, however, a circular amniotic fold appears surrounding the primitive anterior end of the germ-band and the aperture of invagination (Fig. 139 D, of). This circular fold at the time of its rise consists, like every other amnion-fold, of two layers (amnion and serosa). In the course of further growth, however, the serosa outstrips the amnion, so that the cephalic lobes appear covered by only one epithelial cell-layer, the so-called cephalic serosa (Fig. 139 E, s').* d~* v Fig. 140. — Rotation of the embryo of Oecanthus (diagrams after Ayers). a, anterior pole of the egg ; am, amnion ; b, posterior pole of the egg ; d, dorsal side of the egg ; k, germ- band ; r, dorsal plate (caused by the contraction of the serosa); s, serosa; v, ventral side The other, later processes of development — the rupture of the cephalic embryonic envelope, the evagination of the embryo through the aperture thus produced, and the rotation of the germ-band occur in just the same way as in the Libellulidae, The ontogeny of the Aphidae has been described chiefly by Brandt (No. 7), Metschnikoff (No. 55), Witlaczil (No. 98), and Will (No. 97). In the above account we have principally followed "Will. * This is a case of the imperfect development of the amnion, such as has been asserted for certain Hymenoptera. It should be mentioned that the account given by Brandt makes it appear possible that the embryonic envelopes in the cephalic region in the Coccidac (and perhaps even also in the Libellulidae) develop in the way described by Will in connection with the Aphidae. In this case, in these forms also, the envelope covering the cephalic region would consist of a single layer of cells. 262 INSECT A. In this type of development "\ve must also include one of the Gryllidae, Oecanthus, although by so doing we place this form in opposition to the other Orthoptera. In this genus the first rudiment of the germ-band, indeed, does not arise by invagination as has been shown by Ayers (No. 1), but forms as a short ventral plate which is overgrown by the amniotic fold. The inner layer of the fold (the amnion), as in the cephalic fold of the Aphidae, does not at first grow over it as rapidly as the outer layer. The serosa therefore at first forms the only complete covering of the germ- band. The amnion, however, grows out later under the serosa and becomes closed, so that the embryo is finally covered by a double cellular envelope. The germ-band is therefore in this case to be classed among those which are grown over by a fold; and it is also superficial. It lies originally (and this is the point which has determined us in the above classification) on the dorsal side of the egg, with its cephalic end directed posteriorly (Fig. 140 A), and thus in position entirely agrees with Gorixa (p. 279). It is therefore obliged, when the rent has taken place in the embryonic envelopes, to undergo a process of rotation (Fig. 140 B, C, D) so as to attain its definitive position. This process and the later degeneration of the serosa through the formation of an invagination (dorsal tube) show such complete agreement with the processes in other examples of this type that we feel justified in classing Oecanthus among them. E. Insects in which the Germ-band is overgrown by the Amniotic Fold. Orthoptera genuina. In all the forms belonging to this group as yet investigated, with the exception of Oecanthus, the germ-band lies from the first on the ventral side of the egg, with its cephalic end pointing anteriorly. In these therefore there is no rotation of the germ -band. The embryonic envelopes arise through the formation of folds. The germ-band is in most cases comparatively short (Blatta, Cholodkowskv, No. 19, and Wheeler, No. 95; Stenobothrus and Mantis, Graber, Nos. 26 and 30). Only in Gryllotalpa (Korotneff, No. 47) does the germ-band attain a con- siderable length, and consequently appears with its anterior and posterior ends bent over towards the dorsal side. In all these forms the posterior (abdominal) end of the germ-band in later stages appears flexed ventrally, as is also the case in the Libelhdidae, Rhynchota, Oecanthus, Phryganeidae, in many Coleoptera, and to a INSECTS IN WHICH THE GERM-BAND IS OVERGROWN. 283 still greater degree in the Lepidoptera and certain Hymenoptera. This curvature becomes lost and the abdomen straightens as a rule even before hatching. It should be mentioned that in Stenobothrus the amniotic fold forms at a very early period in the development of the germ-band. At a time when gastrulation is commencing and the ventral plate is still round and shield-like, the latter is already grown over by the amniotic fold (Graber, No. 26). In GriillotaJpa, the embryonic envelopes develop in the form of two folds rising laterally. In Blatta, a caudal fold and paired cephalic folds corresponding to the two cephalic lobes appear first (as in the Coleoptera). The germ-band in the Orthoptera is, as a rule, superficial. Only in Stenobothrus does it become immersed by the appearance of particles of food-yolk between the two enveloping layers, a condition which we shall meet with again in the Lepidoptera. Fig. 141. — Lateral aspect of the egg of Chironomus in the stage at which the embryonic envelopes develop (diagram adapted from Weismann and Kupffer). A, showing the commencement of the cephalic and caudal folds (kf and .«/). B, union of the two folds laterally and their continuous advance over the germ-band, v, ventral side ; d, dorsal side ; ft, uncovered portion of the germ-band ; am, amnion ; do, food-yolk ; k) cephalic end of the germ-band ; k', ventral portion of the germ-band ; k", portion of the germ-band bent over dorsally ; k"', hook-like bend of the posterior end ; kf, cephalic fold of the amnion ; kl, cephalic lobe ; r, dorsal umbilical passage ; s, serosa ; sf, caudal fold of the amnion x, x', union of amnion with the ectoderm. Diptera. In all Diptera as yet examined, the germ-band attains a considerable length ; it thus not only covers the ventral side, but its two ends bend over dorsally, the posterior end extending very far forward (Fig. 141), so that on the dorsal side of the egg they 284 INSECTA. are seen lying somewhat near each other (x, %'). The later develop- ment, in which the posterior end of the germ-band draws back to the posterior pole of the egg, is therefore attended by great shorten- ing of the embryonic rudiment. The amniotic fold does not here arise simultaneously along the whole edge of the germ-band, but a fold first rises round the cephalic end (hf) and a second at the posterior end (sf) of the germ-band {cephalic and caudal folds of the amnion). Only later do folds form at the sides of the germ-band and connect the cephalic with the caudal fold (Chironomus, Weismann, No. 87, and Kupffer ; Simulia, Metschnikoff, No. 55). As the ends of the germ-band are bent over dorsally and lie very near each other, the edges of the amnion also, which arise from the ectoderm in front of and behind the ends of the germ-band (at x and x' in Fig. 141 A), lie near each other. It follows that the region in which the serosa lies directly on the surface of the food-yolk (>•) is in these forms very limited. Similar conditions will be met with in the Lepidoptera and the Phryganeidae. The germ-band in the Diptera is throughout superficial ; only its most posterior end, in Chironomus and Simulia, and possibly also in the Muscidae, appears bent in the shape of a hook and sunk into the yolk (Fig. 141 B, &'"). We have here an approach to conditions to be described in the Coleoptera. The fact that in a few Diptera the amniotic folds remain imperfect and never completely grow over the germ-band deserves mention. This is the case, according to Metschnikoff (No. 55), in the embryos of the viviparous Cecidomyia larvae, in which the cephalic and caudal folds appear as rudiments, but do not develop further. It is also the case, according to Kowalevsky and Geabeu (Nos. 27 and 28) in the Muscidae, in which the cephalic fold remains extremely small, and only the caudal fold develops somewhat more distinctly. In the later development of the embryo, these folds simply flatten out again and then take a certain part, as it appears, in the develop- ment of the dorsal integument. Trichoptera. The conditions to be observed in the rounded egg of the Phryganeidae, as Patten found in Neophylax (No. 65), approximate very closely to the normal type of the Diptera (Chiro- nomus). The very long, superficial germ-band here also covers the greater part of the periphery of the egg, so that its anterior and posterior ends almost touch dorsally. We shall see that even the phenomena of degeneration of the germ-envelopes in the two groups belong essentially, to the same type (Graber, No. 27). Lepidoptera. The Lepidoptera also, in the general conditions of the germ-band and the embryonic envelopes, stand very near the two last groups. A remarkable point in their development is that the amniotic fold forms at a very early stage of the development of the germ-band (as in Stenobothrus, p. 282), at a time when the INSECTS IN WHICH THE GERM-BAND IS OVERGROWN. i'sr, rudiment of the band or ventral plate is only a round, shield-like thickening of the blastoderm (Fig. 142 A), from the edge of which the amniotic fold rises. Only later does the germ - band begin to lengthen, and very soon, by the passage of food -yolk masses be- tween the amnion and the serosa, becomes immersed (Fig. 142 B). Since, as in the Diptera, increasing length leads to the sharp dorsal curvature of the germ-band, and since the amniotic cavity follows this curvature, that dorsal portion which represents the connection between the embryo and the germ- bands appears to become more and more limited (Fig. 142 C, at x). There thus develops a dorsal um- bilical passage which is here of significance in so far as it repre- sents the passage through which the food-yolk mass taken into the interior of the embryo communi- cates with that lying between the amnion and the serosa. Taking into account this feature, it might be said that, in the Lepidoptera, the embryo is surrounded by a yolk-sac connected with it through the dorsal umbilical passage. Hymenoptera. In the Hymen- optera, conditions are found which agree in essentials with those described for the Diptera. The germ -band here also is always superficial, and is covered by a double cellular envelope (amnion and serosa) formed by the growth ventralwards of a cephalic and a am a/7i Fig. 142. — Diagram of the formation of the embryonic envelopes in the Lepidoptera (A after Kowalevsky, B ami C after Tichomiroff). k, germ-band ; am, am- nion ; se, serosa; do, food -yolk; vd, invagination of the stomodaeum ; ed, invagination of the proctodaeum ; m, mouth ; an, anal aperture ; x, dorsal umbilical passage. 286 INSECTA. caudal amniotic fold (Fig. 143 A). This process has been described by Kowalevsky for Apis, and still more clearly by Graber for Polistes gallica and Formica, and more recently for Hylotoma berberidis (Xos. 27 and 30). In Apis, at least, the cephalic fold seems to take a greater share in the overgrowth of the germ-band than the caudal fold. A vd * d ., mi x-~. jr— - _ — mx -/* ~/i' s -* s — - &nt—- an Fig. 143. — Diagrammatic median sections of two stages of development of Hylotoma berbi ridis (after Graber). ai-«io, first ten abdominal segments ; am, amnion ; on, anus ; at, antenna ; bg, ventral chain of ganglia ; do, food-yolk; ed, proctodaeum; m, month; md, mandible; dub1, first maxilla ; mx-, second maxilla ; og, supra-oesophageal ganglion ; ol, labrum ; P\ P~, P8! the three thoracic limbs; s, serosa; sp, salivary glands; vd, stomodaeum x, <•, point at which the amnion passes into the ectoderm. The germ-band in the Hymenoptera remains as a rule comparatively short. It is not longer than the egg (Fig. 143 A), and thus remains restricted to the ventral side. On the other hand, the amniotic cavity itself continues to extend over the anterior and posterior ends of the germ-hand towards the dorsal surface, thus causing the points where the amnion unites with the ectoderm of the germ-hand (x and x') gradually to approach one another, a rare condition TRANSITION FORMS BETWEEN THE TWO T\TES. 287 among the Insecta, but one which apparently also occurs in the Lepidoptera. The dorsal umbilical passage is in this way more and more circumscribed, until, by the fusing of these inner folds and the absorption or rupture of this solid cord, it is completely obliterated (Fig. 143 B). The embryo, whose dorsal wall is now completely formed, lies henceforth entirely free within two cellular sacs, the outer one corresponding to the serosa and the inner to the amnion (s, am in Fig. 143 B). Although the presence of a double cellular envelope (amnion and serosa) in the Hymenoptera can hardly, according to Graber's recent observations, be doubted, we must here mention that other authors expressly point out that only a single embryonic envelope is present, which then must be assumed to be the serosa. Although some confusion on this point may arise from tin- fact that the inner envelope (amnion) becomes closely applied to the germ- band (Gkaf.ek) and is indistinguishable from the latter, we cannot deny the possibility that the true amnion at first remains stationary, as was described above (p. 2S1), in the case of the cephalic fold of the Aphidae and Oecanthus (p. 282). There would then be a separation of the amnion from the serosa at the edge of the amniotic fold, and the latter would grow out independently by a process of overgrowth (cf. the description given of the formation of the amnion in the Scorpions, p. 5, Fig. 3). The same conditions were found in Apis by Butschli (No. 11) and Gkassi (No. 32), also in Polistcs gallica and Chalieodoma muraria by CarriEre (No. 13).* AVe are still altogether in doubt as to the presence and constitution of the embryonic envelopes in the Pteromalina (cf. on this point the account of Platygastcr), in which the endoparasitic life of the embryo and larva has essen- tially modified the course of the development. F. Transition forms between the two types of development of the Germ-band. Coleoptera. The germ-band of the Coleoptera, which, like that of the Hymenoptera, does not attain to any great length, shows in its anterior and principal portion (Fig. 144, /,-) the characters of a germ-band grown over by the embryonic envelopes. It is superficial and is grown over by the forward extension of a caudal fold (af) and paired cephalic folds (af) (Fig. 134 C, af", p. 270), which soon fuse, to which are added, in Lina (Graber, Xo. 30), lateral folds that arise independently. The posterior end of the germ-band, on the contrary, develops entirely according to the invaginating type described in connection with the Libel! uh'dae. In HydrophUus (Kowalevsky, No. 48, and Heider, Xo. 38), at the posterior end of the rudiment of the germ-band, there is a pit (Fig. 134 A, g, p. 270) which exactly corresponds to the invagination known by authors as the germ- prominence (p. 276). As this pit deepens, the most posterior end * [Burger (No. II.) has published a full account of the embryology of this bee, based upon CarriERE's notes. He finds only one envelope arising from the peripheral portion of the blastoderm and persisting for a short time. — En.] 2S8 INSECTA. of the germ-band develops (Fig. 144, k'), bends round dorsally, and sinks into the yolk. The most posterior part of the germ-band is thus here immersed ; the anal end is directed forward and applied to the dorsal side of the egg ; in short, it shows all the characters of the invaginated germ-band (Fig. 144, 7c). The germ-band, in the Coleoptera, is thus originally bent round dorsally over the posterior pole of the egg. The cephalic end of the germ-band accord- ingly lies at first some distance from the anterior pole (Fig. 134 D, p. 270). It, however, gradually moves towards the anterior pole (Fig. 134 E), while the pos- terior end moves back from its dorsal situation to the posterior pole. This shifting causes the posterior invaginated portion of the germ-band to be, as it were, drawn out of the yolk, so that, finally, the germ-band throughout its whole length is superficial. This shifting of the germ- band corresponds exactly to the process of rotation. In Hydrophilus, however, the rupture in the embryonic envelopes takes place only at a later stage. Conditions like those just described for Hydrojihilus are found in the other Coleop- tera, as may be gathered with special dis- tinctness from the observations of Gkaeer (No. 30) on Lina and of "Wheeler (No. 95) on Dory2)hora. Here also the posterior end of the germ-band is bent in dorsally and sunk into the yolk. The principal difference between these cases and that of Hydrophilus is found in the fact that the cephalic end of the band appears from the first near the anterior pole of the egg, and conserpiently the movement accompanying rotation is not here to be observed. "We have already pointed out (p. 284) that the posterior end of the germ-band in the Diptera is sunk into the yolk, in the same way if not to the same extent as in the Coleoptera. "We have here also the last indications of the formation of a germ- band by invagination. The presence of these vestigial conditions, and above all the condition of Hydrophilus (and Oecanthus) seem to indicate that the formation of the germ-band by invagina- tion is the primitive method in the group of Insecta, while the growth over it of the amniotic fold represents a secondary condition (Will, No. 97)., The movement of rotation which can be observed in Hydrophilus and Oecanthus is only comprehensible on this assumption. Fig. 144.— Diagram of a median longi- tuilinal section through a Hydrophilus embryo in the stage depicted in Fig. 134 1), p. 270 (after Heider). of, anterior amniotic fold ; af, posterior amniotic fold; o.h, amniotic cavity; am, amnion ; do, food-yolk ; k, the segmented germ-band, which is already trilaminar; h', posterior end of the germ-band bent round dor- sally and sunk into the yolk ; v, ventral side of the egg; d, dorsal side of the same. GENERAL CONSIDERATIONS. 289 Gr. General Considerations. "We have seen above (Figs. 113, p. 226, and 114, p. 227) that, in the Myriopoda, the germ-band, as it increases in length, is flexed ventrally and sinks into the interior of the egg. In this invagination, which we must imagine to have come about at first through the difficulty of accommodating the long germ-band within the spherical egg, we shall have to seek (as Graber, No. 149, indicated, and Will, No. 97, more recently proved in greater detail) the starting-point for the development of the invaginated germ-band of the Libellulidae. We shall therefore consider the invaginated form of germ-band as the most primitive in the Insecta. A careful comparison between the condition of the Myriopoda and that of the Libellulidae, indeed, reveals certain differences. In the Myriopoda, the germ-band alone is drawn into the invagination. In the Libellulidae, on the contrary, in which the germ-band is comparatively short, it occupies only one side of the depression, while the opposite side seems to be occupied by a part of the blastoderm which has been drawn into the depres- sion with the band, and is then known as the amnion. The part of the blastoderm not concerned in the formation of the germ-band in this case therefore is more extensive, and this marks the first commencement of the formation of the embryonic envelopes. In the Myriopoda, the parts of the germ-band not drawn into the depression remain simply uncovered. In the Libellulidae, on the contrary, they are grown over by a fold (amniotic fold) which arises secondarity. This formation of folds is a new acquisition in the Insecta by which the system of embryonic envelopes is com- pleted. It is contrasted by Will (No. 97), as the secondary part of the embryonic envelopes, to the primary part which arises by invagination. We should, however, hesitate to lay much stress upon this distinction. In the more highly developed and specialised Insect types, the secondary formation of folds becomes more prominent, while the development of the germ -band through invagination sinks into the background. The germ -band grown over by the amnion is thus derived from the invaginated band, and the development of the former marks an ontogenetic advance, as the complicated process of rotation is now eliminated. The cases of vestigial development of the embryonic envelopes which have been observed in endoparasitic eggs (Pteromalina, Tachinidae), in the eggs of the viviparous Cecidomyiidae, and u 290 INSECT A. in the Muscidae must be regarded as secondarily acquired, when we take into consideration the condition of other nearly related forms. "We have as yet no certain data to help us in discussing the question of the physiological significance of the germ -envelopes. Although the increase of the yolk-absorbing surface may have been of importance for the development of the invaginated germ-band, this consideration does not help to explain the development of the amniotic folds that have grown over the germ-band. In the latter we seem to see the influence of an ontogenetic tendency which led to the germ-band being separated from direct contact with the inner surface of the chorion (or vitelline membrane). This may have afforded greater protection against certain mechanical injuries, perhaps also against the danger of desiccation or adherence. The latter hypothesis seems to receive special support from the fact that eggs with degenerated embryonic envelopes (Cecidomyia, Tachina, Muscidae) are, in consequence of the nature of their surroundings, less exposed to this danger. All these conjectures, however, afford little satisfaction. 4. Development of the external form of the Body. A. Segmentation. The first traces of segmentation are found very early in the germ-band of the Insecta, which becomes divided up by superficial transverse furrows into a number of somites. This segmentation, in the form of consecutive metameres, may appear as early as the very beginning of gastrulation (Hydrophilus, Kowalevsky and Heider, Fig. 134 A and B, p. 270, and Chalicodoma muraria, Carriers, No. 13, Fig. 156, p. 315). The transverse boundaries of the segments then extend not only over the middle plate (p. 310), from the invagination of which the lower germ-layer arises, but laterally over the lateral plates (Fig. 156, s), which become the ectoderm of the germ-band. These transverse furrows owe their origin to the alternate thickening and thinning of the epithelium, which at this stage forms the embryonic rudiment, the furrows corresponding to the thin areas. It follows that, in the forms just enumerated, after gastrulation has taken its course, not only the ectoderm, but the lower layer also, is segmented. Heider (No. 38) maintained in the case of Hydrophilus that the first indications of segmentation even precede gastrulation. Similar transverse zones of the blastoderm have been observed by Wheeler £ Q&). 95) in Doryphora and by Graber (No. 30) in Lina, but these i ,-4 ■ SEGMENTATION. 291 authors interpret them in another way not connected with the later segmentation. Such an early appearance of segmentation as that found in Hydrophilus and Clialicodoma must be regarded as a modification of the ontogenetic processes founded on heterochrony. We shall have to regard as primitive the condition found in other forms (e.g., Lina and Stenobothrus, Graber, Xo. 30), in which the gastru- lation and the separation of the lower layer take place in the unsegmented germ-band, and the division into segments only occurs Fig. 14.j. — Three stages in the development of the germ-band of Lina (after Graber). a, unsegmented germ-band ; in D and C, the segmentation is recognisable in the lower layer. B, with the rudiments of the mandibular and two maxillary segments, to which, in C, the three thoracic segments and the two anterior abdominal segments are added, ft', a", tirst and second abdominal segments ; af, amniotic fold ; hi, blastopore ; k', mandibular segment ; k", k'", the segments of the two maxillae ; M, cephalic lobes ; m, mouth ; V, t", V", first, second, and third thoracic segments ; th, extension of the germ-band in the thoracic region ; v., lower layer. at a later stage (Fig. 145). In these forms the segmentation is princi- pally noticeable in the invaginated lower layer, although probably, in all cases, the ectoderm also is affected by it at an early stage. In the completely segmented germ-band of the Insect (Fig. 134 E, p. 270, and Fig. 146 A, p. 295) we distinguish two peculiarly-shaped regions, one corresponding to the anterior end and another corre- sponding to the posterior end. The anterior or primary cephalic ^92 INSECTA. region carries the oral aperture, and is characterised by its great lateral extensions, the cephalic lobes (Figs. 134, Jc and 145, M), while the posterior terminal section, the so-called anal segment or telson, carries the anal aperture (Fig. 146 A, a). Between these two regions lies the segmented primary trunk-region, which, in the Insecta, seems without exception to consist of sixteen segments. The three anterior of these segments represent the mandibular and the two maxillary segments, which are later drawn into the forma- tion of the head (Fig. 146, md, mxv mx2), while the three following develop into the permanent thoracic segments (pv p.2, 2>3), so tnat ten segments (besides the telson) must be reckoned as belonging to the posterior or abdominal region of the body. Ten abdominal segments together with a telson seem typical throughout the group of the Insecta. This number has been observed recently in the germ- band of Hydrophilus by Heider, and in various forms (Lina, Stenobothrus, various Lepidoptera, and Hylotoma) by Graber (No. 30). Wheeler (No. 95), Cholodkowsky (No. 19), and Carriere (No. 13) have all made similar observations. In the later stages of embryonic development, this number is- apparently in a few forms decreased to nine, the tenth abdominal segment fusing with the telson. This appears to be the case in Hydrophilus and Lina -r in the Lepidoptera, according to Graber (No. 30), a fusion cf the ninth and tenth abdominal segments takes place, the telson remaining independent. With regard to the primary cephalic region, it should be mentioned that,, taking into account the segmentation of the brain recently observed by Patten (No. 67) and confirmed by several other authors, it has to be assumed that this region is composed of several (three) fused segments (cf. pp. 325-328 on the development of the brain). Another point to be noted is that, according to the statements of various authors, among whom Wheeler and Carriere deserve special mention (the former in connection with Doryphora, No. 95, and the latter with Chalicodoma, No. 13), a slightly developed and transitory segment, the so-called pre-maxillary segment, is intercalated between the primary cephalic region and the first body- segment proper (which represents the mandibular segment). According to Carkikre, this structure represents a vestigial pair of limbs and a corre- sponding pair of ventral ganglia. The latter is said to be concerned in the formation of the circum-oesophageal commissure. The cephalic lobes usually appear very early (Fig. 145, kl). Even when the germ-band is still altogether devoid of segmentation, the primary cephalic region is already characterised by the extensions of the cephalic lobes. A slight broadening can also often be observed in that part of the still unseg- mented germ-band which corresponds to the later thoracic segments (Fig. 145- A and B, th). Indeed, Ayers (No. 1) was able to distinguish in the still unsegtnented germ-band of Oecanthus a primary cephalic region, a maxillary,, a thoracic, and an abdominal region, these later regions of the body being indicated by variations in the bulk and breadth of the germ-band. It is on these first rudiments of the body-regions, which are only recognisable as wavy swellings of the lateral contour of the germ-band, that Graber (Nos. 26 and 30) founds his view of the primary segmentation of the Insectan germ-band. STOMODAEUM AND TROCTODAEUM. LABRUM. 293 According to Graber, the law of the development of the body-segments from before backward, which has been accepted on the whole for the Arthropoda and was specially insisted upon by Balfour, does not apply to the Insecta. In this group the germ-band is said to break up at first into macrosomites, i.e., the slightly indicated swellings of the germ-band recognised by Ayers and corresponding to the permanent regions of the body. The macrosomites are said, by means of a secondary segmentation, to break up into microsomites (the later body-segments). This peculiar type of segmentation, which deviates from that of the other Arthropoda, is to be regarded as inherited from a hypo- thetical racial form. AVe, however, are not able to accept this view. Apart from the fact that in Hydrophilus (Heider), Chalicodoma (Carriers, No. 13), Mantis (Viallanes, No. 84), and Xiphidium, one of the Locustidae (Wheeler, No. 94), there is no sign of any breaking up into macrosomites preceding definitive segmentation, it appears to us that the broadening of the germ- baud at the part where later the thoracic region develops may be traced back merely to an accumulation of plastic material, and that it should not therefore be regarded as the expression of a true segmentation. If the lower layer were also affected by this apparent breaking up into macrosomites, the case would be different. Such a condition was actually stated by Graber to exist in Stenobothrus (No. 26). From his more recent publication (No. 30), however, it appears that the formation of macrosomites in the lower layer in Stenobothrus is not cpiite distinct. We have therefore only the statements of Nusbaum (No. 59) in connection with Meloe and, as no division into macrosomites is found in Hydrophilus and Lina, the point seems to require reinvestigation. As a rule, the development of the body-segments in the germ- hand of the Insecta takes place from before backward. This has recently been observed, especially by Graber (No. 30), in various forms (Stenobothrus, Hylotoma, Lina). In Lina, for instance, the mandibular and the maxillary segments (Fig. 145 B, k'-k'") develop first, and in the next stage the three thoracic segments and the two anterior abdominal segments are added (Fig. 145 G), while the other abdominal segments only develop later. In other cases, the development of the segments seems to proceed more equally along the whole length of the germ-band. Our knowledge is, however, very incomplete on this point. An exception to the rule is afforded by Hydrophilus, in which the development of the seg- ments of a middle region is somewhat retarded, while the anterior and posterior parts of the germ-band develop more rapidly. In Pieris, according to Graber (No. 30), the thoracic segments precede all the others in development. The maxillary segments soon follow, and finally the abdominal segments form. B. Stomodaeum and Proctodaeum. Labrum. After the segmentation of the germ-band is completed, the next ontogenetic changes to be remarked are the development of the 294 INSECTA. stomodaeum and the proctodaeum and the rudiments of the limbs. The fore-gut and the hind-gut appear as ectodermal invaginations, the stomodaeum and the proctodaeum, in the primary cephalic region and on the telson (Figs. 145 C, m, and 146 A, m and a). As a rule, the stomodaeum begins to develop a little earlier than the proctodaeum in the Insecta (Fig. 145 0, m). To this rule, however, the Muscidae form an exception, if the observations of Voeltzkow (No. 85) and Graber (No. 28) as to the early appearance of the proctodaeal invagination in these forms are confirmed. About the time when the stomodaeal invagination appears, and anterior to it in position, a forward swelling of the anterior edge of the primary cephalic region is to be remarked. This is the so-called procephalon (Fig. 146, vie), which represents the common rudiment of the labrum and the clypeus. In many cases, this rudiment first appears in the form of a small paired prominence (Fig. 160, I, p. 326), which gives rise later, by fusion in the median line, to an unpaired swelling which is still somewhat indented at the middle. This is the case in the Coleoptera (Hydrqphilusy Kowalevsky, Graber, No. 25, and Heider; in Lina, Graber, No. 30 ; in Meloe, Nusbaum, No. 63 ; in Acilius, Patten, No. 67), in the Lepidoptera (Tichomirofp, No. 79, and Graber, No. 30), in Clialicodoma (Carriere, No. 13), and in other forms. The rudiment is, on the contrary, originally single in Apis (Grassi, No. 32), in Blatta (Cholodkowsky, No. 19), and in Mantis (Viallanes, No. 84). The rise of the procephalon which, by many authors, is called simply the labrum, from a paired rudiment has repeatedly led to its being compared with a pair of pre-oral appen- dages, but the grounds for such a comparison are, as we think, insufficient. This view has been adopted recently by Patten (No. 67), who described the procephalon simply as the first pair of antennae, and also by Carriere (No. 13). The labrum of the Insecta seems to us to find its homologue in the structures called* by the same name in other Arthropoda (especially in the Crustacea), to which the interpretation just mentioned would be inapplicable. It should be mentioned that, in the early embryonic stages of many Insects, a provisional lower lip, arising from a paired rudiment, is found just behind the mouth. This is not to be confounded with the permanent lower lip of the Insecta, which arises by the fusion of the second pair of maxillae. The provisional lower lip was first observed by Butschli (No. 11) in Apis, and called by him the inner antennae ; it was found later by Tichomiroff in the Lepidoptera. Heider described it as the "lateral oral lips" in Hydrophilus> EXTREMITIES. 295 and it has recently been observed by Nusbavjm (No. 63) in Meloe. This structure may best be compared with the paragnatha of the Crustacea, although we are apparently precluded from homologising it with this latter. C. Extremities. The limbs appear as sac-like swellings of the surface of the segments, which, as a rule, are directed backward. The antennal rudiments must be regarded as the most anterior pair of true limbs ; this belongs to the cephalic region, and arises near the posterior edge of the cephalic lobes, at the point where these pass into the mandibular segment (Figs. 146, a?i, and 147, at). It should be specially pointed out that the antennal rudiment, even when it first appears, is, as Weismann (No. 87) has shown, post-oral in position (Fig. 147, at) and shifts towards the mouth only later, finally coming to lie in front of or above it. The an- tennal rudiment, in its external appearance, de- velopment, and position closely re- sembles the other limb-rudiments. s£- ff— 9 — tf Weismann's im- portant discovery that the antennal rudiment is origin- ally post-oral in position has recently been confirmed by various observers (Graber, No. 25, and Heider, No. 38, for Hydrophilus ; Patten, No. 67, for Acilius; Graber, No. 30, for Stenobothrus, Lepidoptera, Hylotoma; Nusbaum, No. 63, for Meloe ; Wheeler, No. 95, for Doryphora; Carriere, No. 13, for Chalkodoma, etc.). This position, as well as the agreement in form between the antennal rudiment and the other limbs, lends important support to the view we have already expressed in connection with Peripatus (p. 186, etc.), and which also applies to the Insecta, that the antennae are structures secondarily shifted to a position in Fig. 146.— Embryos of Hyclrophilus with limb-rudiments (after Heider, from Lang's Text-book), a, anal aperture ; an, antenna ; g, rudiment of the ventral chain of ganglia ; m, oral aperture ; Bid, mandible ; mxlt first, mx„, second maxilla ; pi, p2( p3( the three pairs of thoracic limbs ; pi, p$, pT, ps, rudiments of the first six abdominal limbs ; st, stigmata ; vk, procepbalon. 296 INSECTA. front of the mouth, and that they are entirely homonomous with the other trunk-limbs, and cannot therefore be traced back to the primary cephalic tentacles of the Annelida.* Carriere (No. 13) has asserted the presence in Chalicodoma of a pre-antennal limb-rudiment. According to him, the rudiment of the procephalon represents a first pair of limbs, the pre-antennal rudiment the second, the antennae the third, the transitory limb of the hypothetical pre-maxillary segment (p. 292) the fourth, and the mandible the fifth pair of the series. These statements require confirmation before we can accept them as exjfiaining the true relation- ship of the series of the limbs. Of the limb-rudiments following the antennae, the three next pairs are transformed into the jaws (mandibles, first and second maxillae, Figs. 146 and 147, ind, mxv mx.^). These rudiments develop early and with a complicated form in keeping with their later specialisation, the mandibles appearing toothed and the maxillae lobed. The second maxillae fuse together in later stages to form the lower lip. The three pairs of limbs which follow these (the thoracic legs, Figs. 146 and 147, p1, p>2, pB,) exhibit a massive development, the first traces of the future segmentation soon becoming apparent on them. In the Libcllulidac., the rudiment of the second maxilla appears very large in the embryo (Fig. 138, mx2, p. 278), so that it looks more like that of a thoracic limb than like those of the other jaws. Its special development is probably connected with the size attained by the lower lip (mask) of the larva which proceeds from it (p. 359). With regard to the order of appearance of the different limbs, our knowledge is as yet somewhat incomplete. Here also we find repeatedly that the general order of development is, according to the ontogenetic law, from before back- ward. In many forms the antennal rudiments seem to be the first to appear, while the maxillary rudiments and those of the legs all develop simultaneously, but somewhat later than the antennae. This is the case in Hydrophilus, Melolontha, and Stenobothrus. In Lina, according to Graber (No. 30), the mandibles precede the antennae. Among the Libcllulidac, according to Brandt (No. 7), the rudiments of the thoracic limbs appear first, then those of the maxillae, and only later those of the antennae. In those Insects whose larvae are limbless, on the contrary, the rudiments of the thoracic limbs appear late and in an arrested condition (Apis and Chalicodoma), or are altogether sup- pressed (Muscidac). In the first case, the limb-rudiments degenerate before the larva hatches. It would be interesting to trace the relation of these degenerating rudiments to the imaginal discs of the thoracic limbs that develop later, concerning which, as far as we know, no statements have been published. t * From this point of view, the malformation observed by Kriechbaumer in Bombus (Entomol. Nachr., Jg. xv.) is not without interest ; an antenna was by this author found deformed so that it resembled a leg, and at its end carried two well-developed claws. See Bateson, Materials for the Study of Variation, p. 146. t [BtJRGER (No. II.) finds in Chalicodoma that the thoracic appendages of the embryo flatten out and their hypodermal cell-layer thickens and becomes the imaginal discs of the thoracic limbs of the adult. — Ed.] EXTREMITIES. 297 Soon after the appearance of the thoracic limbs, rudimentary appendages can be seen on the abdominal segments also (Figs. 136 Pa P§> and 137 A, flj-ffg). These, in most cases, exactly correspond in position and in the manner of their development to the limb- rudiments of the preceding segments, so that we may consider them as fully equivalent to the latter. The first statements as to the presence of limb-rudiments on the first abdominal segments were made by Rathke (for Gryllotalpa), and the first mention of the presence of limb-rudiments on all the abdominal segments by J — a Fig. 14".— Two stages in the development of the germ-band of Melolontlta (after Graber). A, stage with eight pairs of abdominal limb-rudiments (al-a8). B, older stage ; the germ- band is very much broadened, a1, limb belonging to the first abdomiDal segment (in B, widened out into a sac) ; ofi, limb belonging to the eighth abdominal segment ; an, anus ; at, antenna; bg, ventral chain of ganglia ; fir, brain; 1, labrum ; m, mouth; md, mandible; rax', first, m:u", second maxilla ; p\ p-, p3 first, second, and third thoracic limb ; s, lateral strand of the ventral nerve-cord ; st, stigma ; x, point of attachment of the sac-like first abdominal limb. Butschli (No. 11, Apis). These statements have recently repeatedly been verified in numerous Insects (for the literature on this point see especially Graber, Nos. 25 and 30, Wheeler, No. 91, and Carriere, No. 15). The first point to be noted is, as Graber has shown, that, in the Orthoptera and Coleoptera, as well as in some Hemiptera, the appendages of the first abdominal segment, as compared with those of the subsequent segments, are more massive and in later stages develop characters peculiar to themselves, while, in the Lepi- 298 INSECT A. doptera and Hymenoptera, the limb-rudiments of the first abdominal segment are, in some cases, less developed than those of the other segments, and in no case do they attain a greater development. In the Orthoptera and Coleoptera, the limb-rudiments of the first abdominal segment show, as is often the ease with vestigial organs, considerable variability in their later development. They are most leg-like in Mantis, according to Grabeb, and in this genus, as well as in some other forms, they even show signs of segmentation, the finger-like process appearing divided into two by a transverse constriction. The limb-rudiments of this segment in Melolontha attain an altogether excessive development (Fig. 147 B, a1, Gbabek), being transformed into large vascular sacs, the walls of which seem to be composed of massive coarsely-granular elements. In many other cases a glandular sig- nificance is suggested for these appendages, the walls at their distal parts being formed of very large coarsely-granular glandular cells which are often pigmented. In such cases, the appendages are mushroom-shaped (Gryllotalpa, HydropMlus) or, when the distal glandular surface sinks in, they assume the form of stalked cups {Mcloe, Nu.sb.vum). Finally, they may be represented by a sac sunk below the surface of the body (Tenebrio, Carriers), or a similarly-shaped solid structure (Cicada and Ncpa, Wheeler). The different shapes assumed by this structure are connected by means of many transition forms. The secretion which has been observed may be gelatinous (Mcloe, Nusbaum, and Cicada, Wheeler) or filamentous (Ncpa, Wheeler). The physiological sig- nificance of these organs still seems very obscure, in spite of the observations- which have been published ; they have been claimed as embryonic respiratory organs (gills) or glands. It should be pointed out that the character of the cells here regarded as glandular agrees closely with that of the elements of the dorsal organ (invaginated serosa) before the latter begins to disintegrate. The appendages we have just been discussing invariably degenerate completely before the larva hatches. The same is, as a rule, the case with the appendages of the posterior abdominal segments, which are usually considerably smaller. It is possible that when the latter disappear they take a certain part in the formation of the lateral parts of the ventral plate, as was conjectured by Haase (No. 153) when reviewing the condition of Machilis and Blatta, and as was more recently rendered probable by Grabeb (No. 30) for Melolontha. With regard to the development of the abdominal extremities (pedes spurii or prolegs) of the caterpillars of the Lepidoptera and the caterpillar-like larvae of the Tenthrcdinidac, it appears from the researches of Kowalevsky (Sphinx), Tichomiroff (Bombyx), and Grabeb (No. 30, Bombyx and Hylotoma) that limb-rudiments first form on all or most of the abdominal segments, but that they very soon disappear on those segments which, in the larva, have no limbs, while on the other segments they are transformed into the functional prolegs. To this view the observations of Goossens and Knatz, according to which single pairs of these prolegs first develop during larval life, are apparently unfavourable. We should here have to suppose, as Gbaber (No. 30) also has pointed out, an embryonic rudiment remaining for a considerable time in a latent condition. On the whole, the embryological data seem to support the view of Balfour, which Cholodkowsky has recently adopted, and to which Grabeb (No. 30) is inclined, that the abdominal appendages of the caterpillars of the Lepidoptera and Hymenoptera are to be regarded as true limbs. We have already had several examples in the Crustacea of the disappearance and EXTREMITIES. 299 re-development of a limb out of a rudiment which has meantime been latent (mandibular palp of the Decapod larva, ATol. ii. , p. 312, maxillipedes of the Stomatopoda, Vol. ii., p. 300). A similar example is afforded among the Insecfa by the thoracic limbs of man}' Hymenoptera ; these appeal" as rudiments in the embryo, disappear later, and reappear in the imago.* The same process will be found to explain the phylogenetic appearance of the abdominal limbs of the caterpillars and Tenthredinid larva ; for it can hardly be doubted that the Lepidoptera and the Hymenoptera, as well as all Heteromorpha, are to be derived from homomorphous ancestral forms which, in the larval condition, were devoid of abdominal limbs. The larval form of the caterpillars, in spite of its apparent resemblance to Perijmtus, must be accepted as a secondary ontogenetic condition acquired in adaptation to certain conditions of life (p. 366). Special mention should be made of the appendages of the last abdominal segment (anal or terminal segment, which in many orders of Insects, especially in the lower orders (Oithoptera genuhia, Eplwmcridaz, Odonata, Plecoptera), persist throughout life as the so-called cercopoda (cerci). It must still be considered doubtful, on account of the nature of the terminal segment, whether we may consider these appendages as the equivalents of the other true limbs. According to the observations of Cholodkowsky (No. 19), their development in Blatta seems to support such a view. They here appear not only in a form resembling that of the other abdominal appendages, but a process of the coelomic sac which develops in the terminal segment extends into them as into the other limb-rudiments. The homologue of the cerci is perhaps found in the posterior extremities of the Lepidopteran caterpillar, which lie beneath or near the anus, the so-called anal prologs which, according to Graber (No. 30), develop on the terminal segment. The three-jointed anal cerci of the Tenthredinid genus Lyda and the structure known as anal spikes in other forms {Nemaius, Zaddach, and Hylotoma, Graber, No. 30) correspond to them. The so-called anal prolegs of the larvae of many Tenthredinidae are, on the contrary, appendages belonging to the tenth or penultimate abdominal segment. There is a certain relation also between the typical abdominal limb- rudiments and the unjointed appendages of the ventral plate of the ninth abdominal segment, known as the styli ; these are found in many Oithoptera, and persist throughout life in the males. According to Cholodkowsky (No. 19), they are derived in Blatta from the embryonic limb -rudiment of this segment. Haase (No. 153), on the contrary, will not allow that either the appendages under consideration or those small movable processes found on the abdominal segments of the Thysanura (ventral stylets) have the morphological significance of true limbs, but regards them merely as the equivalent of the coxal spurs of Scolopendrella. We are here led to ask to what extent the external genital appendages, the so-called gonapophyses, are to be traced back to limb-rudiments. The researches of Kraepelin and Dewitz (No. 103) have revealed that the ovipositors of the Hymenoptera and the Locustidae, and the corresponding genital appendages of the male in these forms, are derived from imaginal discs of the eighth and ninth abdominal segments, which, when they first appear in the larva, closely resemble those imaginal discs of the larva of Corethra, which yield the thoracic limbs (p. 371). Butschli (No. 11) and others have therefore attempted to refer the gonapophyses of these forms to true abdominal limb-rudiments. In support of this assumption, we might point out that these imaginal discs * [See footnote, p. 296.— Ei>.] 300 INSECTA. develop from the abdominal limb-rudiments present in the embryo. It should, however, be mentioned that Haase (No. 153), following Uljaxix, has recently opposed this view, although, as it appears to us. with insufficient reason, maintaining that the gonapophyses should be regarded merely as secondarily- acquired external appendages.* We cannot deny that a certain phylogenetic significance attaches to the presence in the Insect embryo of abdominal limb-rudiments that degenerate later. Considering the near relationship that exists between the Insecta, the Myriopoda, and Peripatus, Ave must see in the appearance of these rudiments the ontogenetic recapitulation of the conditions belonging to an ancestral form of the Insecta, in which all the body-segments were still provided with well-developed pairs of limbs resembling the present thoracic limbs. We should have to attach a certain importance to the fact that, in the Orthop- tera, the embryonic limb-rudiments of the first abdominal segment are always more developed than those of the following segments and, in Mantis, exactly resemble legs. Since, in Campodea (Haase, No. 153), a true rudiment of a leg is retained on this segment, we are justified in raising the question Avhether, in the degeneration of the abdominal extremities in the series of ancestors of the Insecta, the hexapod condition was not preceded by an octopod condition. This would explain the fact that the segment in question, in many points of its development, resembles the thoracic rather than the abdominal segments. The limb-rudiments which are found as sac-like bnlgings of the surface of the germ -band are from the commencement of their development filled with mesoderm. In most Insects there is at first no arrangement in the mesodermal cells that enter the limb-rudiments, but the Orthoptera seem more nearly to follow the Myriopoda and Pcripaius, in so far as, in them, diverticula of the coelom extend into the rudiments (Cholodkowsky. No. 19 ; Graber, Nos. 26 and 30). D. Nervous System and Tracheal Invaginations. The rudiments of these two systems of organs help essentially to determine the external form of the Insectan germ-band. The rudi- ment of the nervous system usually appears very early, before the limb-rudiments are recognisable. We find, as rudiments of the ventral * [Heymons (Nos. XVI. and XXII.) has investigated the development of the cerci, gonapophyses, and stylets in Lcpisma and other Insects, and he concludes that the cerci are true appendages, that the styles appear to be dermal processes replacing true appendages and intimately related to them, and that the gonapo- physes have no relation to appendages. Wheeler (No. XLI V.), on the other hand, is strongly in favour of regarding the gonapophyses as modified abdominal appendages. Uzel (No. XL.) regards the ventral stylets of Campodea as direct derivatives of abdominal appendages. — En.] NERVOUS SYSTEM AND TRACHEAL INVAGINATIONS. 301 chain of ganglia, two swellings running longitudinally along the germ- band near the median line (primitive swellings, Fig. 147 A, s) and a channel lying between them (primitive groove, neural groove). Segmentation takes place early in the primitive swellings, broader paits (rudiments of the ventral ganglia) alternating with constricted parts (longitudinal commissures) in regular segmental order (Fig. 146 .4, g). Anteriorly, the primitive swellings diverge from one another, as the circum-oesophageal commissures, and pass directly into the cephalic lobes. Here each passes into the brain-rudiment, a some- what large ectodermal thickening, the shape of which will be described more in detail below (p. 326). The rudiment of the brain and that of the ventral chain of ganglia are thus, in the Insecta, connected from their first appearance. The tracheae arise as ectodermal invaginations recurring in each segment (Fig. 146 and 147, st). The apertures of the invaginations afterwards become the stigmata. The tracheal invaginations occur regularly on the first to eighth abdominal segment. In the thorax, in which the presence of a pair of such invaginations in each segment may no doubt be assumed as the primitive condition, there is variation in this respect in the different groups. In the Lepidoptera, one tracheal invagination appears in the pro-thorax, while none is found in either the meso- or the meta-thorax. The embryos of most Coleoptera and Hymenoptera (Apis, Butschli, Hylotoma, Graber, No. 30), on the contrary, have no tracheal rudiment in the pro- thorax, but possess such a structure on both the meso- and meta- thorax. The same is the case in the embryo of Mantis (Graber, No. 30). The tracheal invaginations as a rule develop only after the appearance of the limb -rudiments. An exception to this rule is afforded by Apis, in which the tracheal invaginations appear in the thoracic region before the belated limb- rudiments. As a rule the invaginations appear almost simultaneously, only rarely is there any indication of the order of development from before backward. In H>/dro2)hihis, for instance, the meso-thoracic stigma appears somewhat earlier than the stigmata of the other segments (Gkaber, No. 25). In the Coleoptera, structures conjectured by Heider (No. 38) and Wheeler (No 95) to be the vestiges of tracheal invaginations have been observed on the ninth and tenth abdominal segments. It should be mentioned here that certain ectodermal invaginations appearing in the head have been regarded as tracheal formations which have lost their primitive function and become secondarily modified. Carriers (No. 13) follow- ing Moseley and Palmen (No. 161) has thus regarded the salivary glands and the tentorial invaginations as modified tracheae. Others (Butschli, Geassi) have considered the Malpighian vessels to be of the same type as the tracheal invaginations. We shall further on give our reasons for not adopting this view. 302 INSECTA. E. Transition to the Definitive Form of Body. The development of the definitive shape of the body is accomplished through the circumcrescence of the Avhole of the nutritive yolk by the germ-band. We have seen above (p. 272) that, in the later stages of development, the germ-band as a rule lies in such a way that its anterior end corresponds to the anterior pole of the egg, and its posterior end to the posterior pole. As the germ-band grows considerably in breadth, its lateral edges shift up dorsally over the surface of the food-yolk (Figs. 150 A-F, 169, 170, 171, and 172). In this way the lateral parts, and later the dorsal parts, of the larval body are formed. By means of this circumcrescence, the food-yolk comes to lie entirely within the embryo, and finally fills the lumen of the archenteron (Fig. 150 F). The closing of the larval body dorsally through the circumcrescence of the food-yolk by the germ- band is so intimately connected with the degeneration of the embryonic envelopes that we shall have to return to these processes later on. The dorsal parts of the embryo in the cephalic region develop independently of the broadening of the germ-band described above. The segments of this region, i.e., the maxillary, only take part to a small extent in the development of the dorsal portion, the latter being mainly formed by the bending over dorsally and the backward extension of the cephalic lobes as well as of the procephalon. The anterior end of the germ-band is therefore here bent over dorsally. An actual dorsal flexure of the cephalic region develops, as was first pointed out by Weismann and later by Hatschek and H eider (No. 38). During this flexure of the anterior end of the body, the part, of the procephalon lying near the mouth appears as a transverse swelling (labrum). The former most anterior part of the procephalon now becomes the clypeus and assumes a more backward position. The cephalic lobes in this process of shifting pass towards the dorsal side, and the antennal rudiments consequently shift in front of or above the mouth. 5. Completion of the dorsal part of the Embryo and degeneration of the Embryonic Envelopes. Tn most of the Arthropoda that have so far come under review (Crustacea, Arachnida, Myriopoda, etc.), development takes place through the formation of a so-called germ-band, but without the formation of actual embryonic envelopes. The surface of the whole COMPLETION OF THE DORSAL TART OF THE EMBRYO. 303 G/KI egg is then covered partly by the hand-like embryonic rudiment and partly by the unmodified blastoderm. The dorsal part of the embryo is there formed by the continuous broadening of the germ-band which by its growth, extends over the greater part of the surface of the egg, the region covered by unmodified blastoderm becoming more and more circumscribed. It is as a rule assumed that the latter takes part in the closing of this dorsal region by being transformed histo- logically to form the ectoderm of the germ-band. It is possible that in these forms also part of this blastoderm gradually degenerates. "We have (Vol. ii., p. 150) conjecturally referred the formation of the so-called dorsal organ of certain Crus- tacea to such a process of degeneration. A similar method of development of the dorsal part of the embryo perhaps also occurs in the Poduridae, in which a dorsal organ is found which develops in the early embryonic stages, and is connected with a larval cuticle that envelops the embryo (Lemoine, No. 51), but in other respects its significance is somewhat obscure (p. 268). In most insects the process is more complicated, in so far as an amniotic fold arises at the junction of the germ-band with the undifferentiated part of the blasto- derm, the degeneration of this fold being intimately connected with the completion of the dorsal surface of the embryo. A very simple case of the formation of the dorsal region in the embryo which, however, we can certainly not regard as primitive, is found in the Muscidae and a few other Diptera which the amniotic fold is incomplete am- m Fig. 148. — Diagram of the development of the dor- sal tube through invagi- nation of the dorsal plate (transformed serosa). Succeeding the stage de- picted in Figs. 138 C and 140 D. am, amniotic fold (now forming the provisional dorsal in- tegument) ; r, dorsal tube, which is already commencing to disinte- grate. (p. 284). Here (according to Kowalevsky, No. 49, and Graber, No. 28) the amnion is simply flattened out again. The amnion and the serosa then together form a simple epithelium which corresponds to the unmodified part of the blasto- derm in the Crustacea, Arachnida, and Myriopoda, and here also seems to take the same part in the development of the dorsal ectoderm. More complicated and very varied methods of formation of the dorsal region and of the involution of the embryonic envelopes are found in the other Insecta, the four following types being distinguishable. 304 INSECTA. [In Leplsma, one of the Thysanura (Heymoxs, No. XVI), the germ-band attains a ventral flexure, and is invaginated into the yolk, in a manner suggestive of the Diplopoda (p. 229), at a very early period. Here, however, a slight amnion forms, and by a narrowing of the cavity of invagination an amniotic cavity arises ; the amniotic folds, with the serosa, which latter sur- rounds the greater part of the egg, however, never unite, so that the amniotic cavity is never closed, and rotation takes place without rupture of the embryonic membranes. The germ-band, commencing at its anterior end, simply emerges from the amniotic cavity through tire persistent amnion-pore. The serosa contracts dorsally, becomes invaginated, and forms the dorsal organ or sac, which then disintegrates. This condition is distinctly more primitive than that seen in the Libellulidac, and recalls the condition of the germ-band in the Myriopoda. In this connection an important and highly suggestive paper by "Willey (No. XLV.) should be studied. Willey believes that the amniotic cavity of insect embryos was originally a product of invagination of the germ- band, and that this invagination was primarily derived from and associated with a ventral flexure of the embryo. In this respect he differs from Heymoxs, who considers that the dorsal flexure of the Chilopoda and Poduridae is primi- tive, whereas Willey would rather regard the ventral flexure of Lcpisvia and the Diplopoda in this light. Willey further regards the dorsal organ of the Poduridae and the indusium of the Locustidue as vestiges of a trophoblast such as occurs in Perijudns novae-britanniae (p. 216). — Ed.] A. Involution through the development of a continuous dorsal amnion-serosa sac. In describing the development of the Libellulidae (Fig. 138 C, p. 278) we saw that, after rotation had taken place, the embryonic envelopes (the amnion and the serosa) which had grown together, represent a membrane which envelops the dorsal yolk-sac (am + se). The condition then somewhat resembles that seen in the Mtiscidae after the flattening out of the amniotic fold. In this membrane, the part yielded by the amnion is clearly distinguishable from that yielded by the serosa, for while the serous portion has greatly thickened by continuous contraction to form a dorsal plate* the amnion has retained its character as a delicate flattened epithelium (Fig. 140, 0 and D, am, r, p. 281). The further fate of the embryonic envelopes in the Libellulidae has not been observed. We can, however, complete our description by reference to other forms which show the same type of develop- ment. As development advances, the food-yolk becomes more and more restricted to the interior of the embryo, or more strictly speaking, of the developing enteron. The yolk-sac consequently diminishes in size, and the absorption of the food-yolk into- the enteron produces a collapse of the dorsal plate, this latter sinking in and forming a thick-walled sac, the so-called dorsal tube (dorsal organ, Fig. 148, r). The walls of this sac soon undergo disintegra- * The dorsal organ of the Podurid embryo seems to be quite peculiar in its formation, and cannot be referred to the dorsal plate here mentioned, as is shown by its early appearance (Lemoine, No. 15). INVOLUTION OF THE EiMBRYONIC ENVELOPES. 305 tion ; the degenerating serosa-cells lose their epithelial connection, and in this disintegrated condition are absorbed into the intestinal canal with the rest of the food-yolk. Simultaneously with this disintegration, which leads to the complete degeneration of the dorsal organ, the outer aperture of invagination completely closes. In this way the serous part of the wall of the yolk-sac becomes disintegrated. There now only remains the amniotic portion of this wall, which, standing in direct communication with the ectoderm of the embryonic rudiment, represents a provisional dorsal integument. It still appears doubtful to what extent this pro- visional integument passes over into the permanent wall, i e., to what extent the amnion is transformed into the definitive hypodermis (a view which Graber <(Xo. 27) and others have been disposed to -adopt). It would ap- pear very strange if A ar Fio. 149. — Three embryos of Hydrophilus from the dorsal side (after Kowalevskv, from Balfour's Text-book). A, the serosa has retracted to the dorsal side and has thickened to form the dorsal plate {do). B, the dorsal plate (do) is partly invaginated and covered by the amnion (Fig. 150 D). C, the dorsal tube is completely developed and opens externally only through an anterior pore (c/. Fig. 150 E). at, antenna; do, dorsal organ in various stages of development. the permanent dorsal integument were to be utilised in earlier embryonic stages as a provisional ventral embryonic envelope (amnion), and as, on the other hand, as we shall show (p. 307), the degeneration of the amnion was directly observed by "Wheeler in Doryphora, we must leave the question open whether, as a rule, in the Insecta, the germ- hand alone forms the whole of the embryonic rudiment, and also brings about, by its dorsal extension and subsequent union, the completion of the permanent dorsal integument, while the amnion serves as a provisional integument, which later undergoes gradual absorption. The above-described process of the completion of the dorsal body-wall by the development of a dorsal organ and provisional completion by means of the amnion, probably applies to the Libcllulidac. It is also found in all Rhyncota ec / / gastrula-furrow, and therefore ■::M: <&$• ■&*: Fio. 155. — Diagram illustrating the separation of the germ -layers in the most anterior region of the germ-band of Hydrophilus, transverse section (after Heider). dz, yolk-cells ; en, ectoderm ; en, ento- derm ; ms, mesoderm. belong, not to the ectoderm, but to the lower layer. The anterior and posterior entoderm- rudiments are said to arise by the proliferation of cells from the blind ends of these two invaginations. Grabeii (No. 28), indeed, has confirmed Kowalevsky's statements for the anterior entoderm - rudi- ment, and also assumes the ectodermal origin for the stomo- daeum. As to the proctodaeum, on the contrary, and the posterior entoderm-rudiment, Graber entirely agrees with Voeltzkow, with the single exception that, for the growth of the pos- terior entoderm-rudiment, he claims not only the blind end, but a long band of the ventral side of the proctodaeum. "We may here object to this view of Voeltzkow and Graber that if, in reality, in the Muscidac, a posterior section of the intestine arose by invagination from the lower layer, we should not be able to call it the proctodaeum, for in that case we should not be able to regard it as homologous with the similarly-named section of the intestine of other Insects, in which it forms, as in all other animals, from the ectoderm. It, however, appears to us that the sections of the posterior end of the germ-band of the Muscidac, which are in any case difficult to understand, can be more Theorie des Mesoderms. Morph. Jahrb. 1S89. THE FORMATION OF THE GERM-LAYERS. 315 satisfactorily explained by interpreting the parts differently, as Graber for- merly did (No. 27). We may perhaps assume that, in the Muscidae, as in Chironomus, the posterior end of the germ-hand not only sinks into the yolk, but also makes a hook-like bend inwards, so that the germ-band in transverse sections of this region is cut through three times. In this way, the posterior end of the germ-band, sunk into the yolk, and the part lying in reality anterior to it, but in transverse section appearing on the dorsal side of the egg, are, by means of the still open gastrula-groove, in communication in such a way that, in a series of transverse sections, the lumina of portions of the gastrula- furrow belonging to these two parts flow together, thus yielding the peculiar dumb-bell-shaped figure. By this assumption, the invagination which Yoeltzkow and Graber (No. 28) erroneously held to be the proctodaeum would more correctly appear as the so-called germ-prominence (p. 276), and the lumen of this invagination would then have to be considered as the amniotic cavity, and the aperture at its dorsal side, not as the anus, but as the aperture of that cavity. The proctodaeum seems to appear only later in the form of an invagi- nation from this cavity. This view is supported throughout by Ritter's obser- vations of the development of the procto- daeum (No. 71). We must here mention Grader's view of the presence of a lateral gastrulation in the Muscidae. Graber finds, in the germ- band of the Muscidae, near the median or principal gastrula - furrow, lateral folds which are specially marked in the most anterior and posterior parts of the germ- band, and which are said to give off elements to the lower layer. These paired folds, which were already known to BtJTSCHLi (No. 12) and Yoeltzkow (No. 85), and which mark the lateral edges of the germ-band, are, according to Graber, supplementary gastrula - furrows which serve the purpose of supporting the gas- trula-furrow in its plastic activity in the formation of the lower layer. Graber has, however, not proved that elements are given off from these lateral folds to the lower layer. Since it was already known to Voeltzkow that, in the stage under consideration, the portion of the blasto- derm not taking part in the formation of the germ-band shows a great tendency to the formation of folds, these folds probably come under this category, and ought not to be regarded as connected with the further development of the embryo. —Ae Fio. 156. — Flask-shape gastrula-stage of Chnlieodoma (after Carri£re). /, folds which bound the middle plate laterally (lips of the blasto- pore) ; m, the partly segmented middle plate (here mesoderm-rudi- ment) ; s, the segmented lateral plates (later ectoderm of the germ- band); ve, anterior entoderm -rudi- ment; he, posterior entoderm -rudi- ment. 316 INSECTA. The formation of the germ-layers in the Hymenoptera seems to deviate somewhat from the common type. Kowalevsky and Grassi (No. 32),* indeed, agree that here also the entoderm originally forms a part of the lower layer. But the separation of the entoderm from the mesoderm in Apis takes place in such a way that the two ends of the lower layer bend over the dorsal side of the egg, and the anterior and posterior entoderm-rudiments which have thus come to lie on the dorsal side grow towards one another. When the two rudiments, which here also are horseshoe -shaped, have met and fused, the circumcrescence of the food-yolk begins ; in this case the process thus starts from the dorsal side and is completed on the ventral side. It results from this that the layer of entoderm- cells in Apis at first does not lie below the germ-band, but on the dorsal side of the egg below that flattened epithelium which, arising from the amniotic fold, provisionally completes the dorsal surface (p. 287). The condition of the entoderm-rudiment in Chulicodoma is somewhat similar (Carriere, No. 13). Here also the entoderm -bands do not lie below the germ- band, but extend beyond the latter towards the dorsal side of the egg. As to the first separation of the germ-bands, Carriere arrived at views approaching those just described, but still revealing in the case of Chalicodoma a peculiar type. The middle plate («?), which becomes invaginated by the formation of the gastrula-furrow, and which, like the lateral plates, shows signs of segmentation at an early stage, is here said to yield the mesoderm exclusively, while the anterior and posterior entoderm-rudiments (ve and lie) arise from a growing zone closely succeeding the middle plate, in the region of which the separation of the mass of entoderm-cells by delamination from the superficial cell -layer which remains in continuity with the ectoderm takes place. We have still to mention the yolk-cells and the secondary cleavage of the yolk. The yolk-cells are elements scattered in the food-yolk, some being cells which remained in the yolk at the time when the blastoderm formed (Fig. 131 O and D, z, p. 265), and seme having reached the yolk by subsequent immigration from the blastoderm and its derivatives. Graber first pointed out the immigration of cells from the lower layer into the yolk, and his observations have been confirmed by other authors. In individual cases, indeed (e.g., * GRASSl's researches mark a turning-point in the conception of the formation of the germ-layers in the Insecta. It must be recorded to his credit that he was the first to oppose the universal opinion of the time that the yolk-cells represented the actual entoderm of the Insecta, and to prove that the entoderm is a part of the lower layer. The presence of an anterior and a posterior entoderm-rudiment was also correctly made out by him. His views were adopted only later by Kowalevsky (No. 49) and Heider (No. 37), though it should be pointed out that the views put forward by Kowalevsky in his first treatise nearly coincided with what is now known to be the actual condition. THE FORMATION OF THE GERM-LAYERS. 317 in Melolontha), these later immigrated cells are said to be clearly distinguishable by their histological character from the cells origin- ally found in the yolk. The yolk-cells are distributed in a regular manner through the food-yolk. Their principal function appears to be digestive, particles of food-yolk being taken up by these cells and changed in such a way as to render the yolk assimilable by the growing cells of the embryo. This leads, after the development of the germ-band is completed, to the marking-off of the territories belonging to the individual yolk-cells, and this process has been described as secondary yolk-cleavage (Fig. 158 C-F, p. 321; Fig. 135, p. 273). In indi- vidual cases (Apis, Mused) such cleavage, however, seems not to occur. The yolk -cells can still be recognised in the completely developed enteron in the remains of food-yolk which fill it, and here they gradually disintegrate. It was long considered by followers of Doiirn, Balfour, and Hertwig that the yolk-cells represented the actual entoderm of the Insecta, as it was thought that these cells finally became arranged at the surface of the food -yolk to form the enteric epithelium. This view has to be relinquished in face of the more recent researches, on which the account of the formation of the germ-la3Ters given above is founded. It appears that the yolk-cells do not in any way take part in the formation of the embryo. It was indeed suggested in several quarters that they gave rise finally to blood-corpuscles or parts of the fat-body (Dohrn, No. 21, Tichomiroff, No. 79, and especially Will, No. 97). A number of more recent authors, however, oppose this vieAV, and maintain that the yolk-cells, after having fulfilled their function as vitellophags, simply disintegrate. This last view seems to us the most probable, since another origin has been proved for the fat-body and the blood-corpuscles (p. 341). Bearing in mind the statements made above in connection with the Crustacea (Vol. ii., p. Hi), we may probably regard the yolk-cells as an abortive portion of the entoderm. [Recent observations have once more rendered uncertain the origin of the mesoderm, the nature of the epithelium lining the alimentary canal, and the true significance of the primitive groove. Thus Heymoxs (Nos. XV. and XX.), states that in the Orthoptera, the ento-mesoderm of other authors is to be regarded as consisting of mesoderm only, the lining of the definitive alimentary canal arising from the ectodermal epithelium of the stomodaeum and procto- daeum. He further states that the primitive groove (blastopore of authors) may be completely wanting, and even when present is not to be regarded as connected with gastrulation. Lecaillon (No. XXIX.) finds that in the Chryso- mclidac the whole alimentary canal is ectodermal. These two authors think that the higher Insects exhibit no entoderm in the alimentary canal of the adult, while in the lower forms (Heymoxs, No. XVI., Lepisma) the enteron arises from the yolk-cells. On the other hand, Burger and Careiere (No.. II.), with whom Wheeler agrees, are fully convinced that a true enteron exists in Chalicodoma, and entirely dissent from Heymoxs' views. They show that the entoderm arises from the undifferentiated blastoderm, and that the stomodaeal and proctodaeal invaginations arise from the superficial layer of blastoderm- cells, the only layer that can properly be called ectoderm. See also Heiher (No. X.).-En.] 318 INSECTA. 7. Further development of the Mesoderm. Development of the Body-cavity. We have seen (p. 271) that an invagination running along the whole length of the germ-band gives rise to a layer of cells, which soon extends on the inner side of the germ-band and so forms a second, lower layer (Fig. 158 C). From this layer, at the anterior and posterior extremities of the germ-band, the entoderm becomes separated and becomes closely applied to the stomodaeal and procto- daeal invaginations which have meantime arisen. The remaining and by far the largest part of the lower layer may, from this stage onward, be considered as mesoderm. An arrangement of the latter into two lateral bands (mesoderm- bands) now takes place, its cells withdrawing more and more from the median line (Fig. 158). This withdrawal from the median line is, however, not complete. Into the space between the two mesoderm-bands the yolk often thrusts itself, giving rise to the so- called median ridge. Segmental cavities [cavities of the primitive segments, us) now appear in the lateral parts of the mesoderm, and the mesoderm-cells become arranged as an epithelium round these cavities and form the wall of the primitive segments or coelom-sacs. The cavities of the primitive segments arise, as a rule, by a splitting of the mesoderm. Heiber (No. 38) thought that in the case of Hydrophilus he had convinced himself that they arose merely by the widening of a slit, which was already recognisable at an earlier stage between the two layers of the mesoderm, and which could be traced back to the lumen of the archenteron compressed dorso-ventrally. Graber (No. 30), however, in his more recent investigations on this point, was not able to satisfy himself of the persistence of these slits. On the other hand, Heider's view has been confirmed by Carriere (No. 13) in the case of Chalicodoma. These observations afford support to the view first adopted by 0. and R. Hertwig that the cavities of the primitive segments in the Insecta represent paired diverticula of the archenteron. Tlie large primitive segments of Phyllodromia arise in a different manner from those of Hydrophilus. The mesoderm of the germ-band is here at first only a single layer of cells. This simple layer, as the limb-rudiments develop, separates with the ectoderm from the surface of the food-yolk, and cavities thus arise in every segment, these cavities, surrounded by mesoderm -elements, becoming the closed coelomic sacs (Heymons, No. 43). The parts of the mesoderm lying laterally in the germ-band are used in the formation of the primitive segments (Fig. 158 D and E). Not all the meso- dermal elements, however, enter into their formation. Some of the mesoderm- cells which lie nearer the median line always remain distinct (cf. Fig. 157 A, m). The greater the size of the primitive segments, the smaller is this remainder, and vice vcrsd. These elements are irregularly arranged and represent a kind of mesenchyme. It was pointed out by Heider (No. 3S), and recently by Graber (No. 30), FURTHER DEVELOPMENT OF THE MESODERM. 319 Fig. 15". — Transverse sections through the abdominal region of three consecutive stages in development of Phyllodromia germanica (after Heymons). am, amnion ; [& O s * 03 so c c; © p. to p '3 S3 O 44> 03 4-> CO 3 C3 4-3 03 73 Pi 09 a CO 03 A 03 -*— ■a 43 GO 4-3 - •- 03 _^ o ca o > VI 43 o c <0 o 3 ci -^ — CO 7t Q O 03 o: '-3 03 09 7? 03 O 03 03 a '5c ea a Fh H £ P &C o 03 /*N rt a K] > 0 t- 03 43 a 93 0 is <; 03 tc — a - = 03 -* „ w o r4 CO t*. 03 C3 43 S «H o 43 b3 tab .- 03 o a p E > *3 CM O Ed 4J a O 33 03 03 >> 03 05 Cm O u '£ C3 O 3 ■M 03 Q3 — if 43 C9 33 03 Gfl ca 03 .r DO o> bC c3 43 43 CO £ 03 tc - +3 CQ +3 3 M 03 03 03 Ka ,~ *~ 03 > ,a ~ ■ VI rt ■ •> 43 +3 H - B 43 s o 03 EC o o 43 43 u H -r CO 03 43 - 03 O 43 g o aa 93 to a 03 I 3 DO Cj eg X £ 03 X > d - 09 -5 03 03 i — o & ~ 03 -J o o •- 60 O [-< 43 OS 09 o o 3 a c s o o 43 C3 0) o o 4^ 00 II t3 a .2 O Q3 03 c g* 43 a 03 r^ £3 03 CO >> B3 H ri to 03 cO 03 pd 4-3 CO 03 U) 03 00 Bd 43 03 > 23 5 — 43 o >> Ti c o 3 — o o 3 a 03 09 03 —j o -r tt o Tc - ^ a tc* -4-3 £ *3 o OS to «* GO o o 03 O 03 +3 o 03 (4 - CO "el — a 3 C3 03 SO 03 OS p 03 Q B J3 09 -^ O 4-3 » o "~ £ .2 o I C3 CO - ° 3 to ^3 „ S a "2 ° s" £ ° 322 INSECTA. which appear at a comparatively late stage and run out from the definitive body-cavity. The mesoderm -bands, which become separated, unite again later, the mesen- chyme-cells from the one side becoming closely applied to those from the other in the median line. After the median yolk-ridge degenerates, a cell-accumula- tion often forms here (Fig. 158 E) ; this extends below the rudiment of the ventral chain of ganglia and owes its origin to the mesenchyme-cells. It is this cell-strand which has been called by Nusbaum (No. 57) the chorda of the Insecta. It is finally used up in the formation of connective and other mesodermal tissues. The permanent body-cavity of the Insecta arises quite independ- ently of the coelomic cavities, by a separation of the germ-band from the yolk (Butschli, No. 11, Fig. 158 F, I). It appears to be bounded on the one hand by the surface of the food-yolk and on the other by the irregularly arranged mesenchyme-cells. We can at first distinguish in section three distinct subdivisions of the body-cavity (Hydrophilus, Heidbr) : a median space and two large paired, lateral spaces which unite later with one another, and with other lacunae which have arisen by the shifting apart of the mesenchyme-cells {e.g., in the limbs). We may trace back the spaces of the permanent body-cavity, as in Peripatus (p. 201), to the primary body-cavity or cleavage-cavity. It becomes apparent as a series of lacunae in the mesenchyme and everywhere shows the character of a pseudocoele (cf. Introduction, Vol. i., p. 11). In later stages of embryonic development, the coelomic sac's and the permanent body-cavity enter into communication (Fig. 167 A, us, Ih). The consecutive coelomic sacs first fuse together through the degeneration of the transverse dissepiments that separated them ; a slit then opens in the median wall of the coelomic sacs and connects their lumina with the permanent body-cavity. In the later transformations undergone by the wall of the coelomic sacs, the latter can no longer be recognised as separate sections of the whole body-cavity. 8. The Formation of Organs. A. Outer Integument. The hypodermis arises by direct transformation from the cells of the ectoderm. In later embryonic stages, the cuticle of the youngest larval stage is secreted at the surface of the hypodermis. The accessory structures, such as hairs, setae, etc, rise from specially large hypodermal cells (setal mother-cells, Tichomiroff, No. 78). Similar cells (scale mother-cells) give rise in the pupae of the Lepidoptera to the scales of the wings (Semper, No. 126). ENDO-SKELETON THE NERVOUS SYSTEM. 323 B. Endo-skeleton. The endo-skeleton of the head {tentorium) develops out of two pairs of ectodermal invaginations ; the anterior invagination develops on the inner side of and somewhat in front of the mandihle, the posterior within and somewhat in front of the second maxilla. The anterior pair becomes connected with the posterior and gives off two supporting columns which ascend on the inner surface of the clypeus towards the dorsal side. The median fusion of the posterior pair leads to the bridging over of the sub-oesophageal ganglion and, in many Insects, a transverse trabecula is thus formed in the cavity of the posterior part of the head (Tichomiroff, Grassi, Patten, Heider, Carriere). Similar ectodermal invaginations bring about the development of a chitinous tendon for the flexor mandibulae and a similar smaller tendon for the antagonistic muscle. Hatschek (No. 36), who was unaware of the relation of these invaginations to the hard structures of the head, thought them to be tracheal invaginations. They have been regarded in the same way recently by Carp-jerk (No. 13). Since this kind of endo-skeleton is found in other groups {e.g., the Crustacea, Vol. ii., p. 160), and the hypothetical transformation of a trachea into an endo- skeletal structure of this kind involves the idea of a considerable change of function, we do not consider the homology between the invaginations under consideration and tracheal invaginations sufficiently well established. We are inclined to regard the former as structures of a distinct character, all the more so that they do not by any means everywhere agree so closely in position with tracheal stigmata of the following segments as they do in Chalicodoma* C. The Nervous System. All the parts of the nervous system are derivatives of the ectoderm and appear in the embryo as ectodermal thickenings. The rudiment of the ventral chain of ganglia is found, as was first shown by Hatschek, soon after the gastrula-invagination closes, in the form of two longitudinal ectodermal thickenings running on either side of the median line. These are the so-called primitive swellings (Fig. 147 A, p. 297), which extend from the cephalic re&ion to the terminal segment and show between them a median depression, the primitive groove (Fig. 158 C, pr and pw). Soon after the appearance of the primitive swellings, the first signs of segmentation can be seen on] them, the swellings being thicker near the middle of the body-segments than at their boundaries. The primitive swellings pass anteriorly at the sides of the oesophageal * [Cf. footnote, p. 78.] 324 INSECTA. invagination (rudiment of the oesophageal commissures) on to the cephalic segment, and from the first are directly connected with the rudiment of the brain which develops from a thickening of the cephalic lobes. This character has recently been specially emphasised by Patten (No. 67), and has also been maintained by Heider (No. 38) and Graber (No. 30) against Will (No. 97), who considers that the brain-rudiment of the Aplridae (neural plate) arises inde- pendently and becomes connected with the rudiment of the ventral chain of ganglia through the oesophageal commissures which arise only secondarily as in the Crustacea (cf. Vol. ii., p. 160, footnote). The widenings of the primitive swellings in the segments give rise to the ganglia of the ventral chain, and the inter-segmental constrictions to the paired longitudinal commissures. Transverse sections (Fig. 158 C, p. 321, and Fig. 135, p. 273) show that the ectoderm becomes multilaminar in the region of the primitive swellings (pw). At a later stage, the lower layers separate by delamination from the superficial layers (Fig. 158 D-F, s) and form the so-called lateral cords, i.e., the rudiments of the longitudinal strands of the ventral chain of ganglia. The primitive groove (pr) deepens meantime and forms an invagination extending between the lateral cords. The cells at the base of this invagination represent the so-called middle cord and give rise in the middle of the segments to the transverse nerve-commissures of the different pairs of ganglia (Hatschek). With regard to the condition of the middle cord in the inter-ganglionic region, opinions are still divided. Hatschek's view that the primitive groove flattens out in this region, its wall being used up entirely for the formation of hypodermis, has been generally accepted, but Graber maintains (No. 30) that in this region also a median cord splits off which degenerates at a later stage. The nerve -fibrillae arise first on the inner or basal surface of the lateral strands and the middle strand. Secondarily, by shifting their position, they become enveloped by ganglionic cells (cf. on a similar condition in the Crustacea, Vol. ii., pp. 160 and 161). Leydig stated that, in many Insects, there is a double transverse commissure in each of the ganglia of the ventral cord, and a corresponding double rudiment has been repeatedly shown to exist in the embryo (Patten, No. 66 ; Ayers, No. 1 ; Heider, No. 38 ; Wheeler, No. 95 ; Graber, No. 30). No detailed accounts have as yet been given of the manner in which the nerves given off peripherally by the ganglia of the ventral cord arise. The ventral diaphragm bridging over the ventral ganglionic chain (Fig. 167 A, dv, p. 340), which has been observed in many Insects, is derived by Korotneff (No. 47) from the mesoderm, but Heider (No. 38) believes that he was able, in Hydropltilus, to trace its origin back to ectoderm-cells lying laterally to the rudiments of the ganglia. THE NERVOUS SYSTEM. 325 An unusually regular arrangement of the cells is shown in sections through the rudiments of the ventral cord in many Orthoptera (Fig. 159). Wheeler (No. 94) has recently recognised as "neuroblasts" four large cells (n-n) lying on each side on the surface of the lateral cords in Xiphidium; these cells, by repeated tangential division, give rise to the ganglionic cells, which are consequently arranged in vertical columns (.:). Graber (No. 30) and Viallastes (No. 84) have observed similar phenom- ni n, Fio. 159. — Transverse section through the rudi- ment of the ventral cord of Xiphidium (after Wheeler). /, fibrous tissue in transverse sec- tion ; m, neuroblast-cells of the median cord ; n1-7ii, neuroblasts of the lateral cords; z, column of ganglion-cells proceeding from the neuroblasts. ena, the former in Stenobothrus and the latter in Mantis.* The middle cord, according to Wheeler, has neuroblasts (m) only in the interganglionic region ; these, however, soon shift to the posterior side of the transverse commissures. In any case, as Wheeler has pointed out, the presence of eight longitudinal rows of neuroblasts points back to a similar condition in the Anne- lida, where only two such rows, produced from neuroblasts, are found (Vol. i., p. 294). The rudiments of one pair of ganglia of the ventral chain originally appear in each of the sixteen segments of the primary trunk. Fusion may occur between these rudiments later, and may bring about an apparent reduction in their number. The ganglia of the three maxillary segments, for instance, unite to form the sub-oesophageal ganglion, and the last pairs of abdominal ganglia fuse in varying numbers, shifting further forward at the same time. In individual cases (e.g., many Diptera), a considerable concentration of the ventral cord is brought about by the fusion of consecutive pairs of ganglia. The brain (supra-oesophageal ganglion) develops in the anterior region of the expanded cephalic lobes. We can, at an early stage, distinguish in the brain-rudiment the following sections : — 1. Paired thickenings of the ectoderm running forward at the * [Burger (No. II.) finds that the ganglionic cells in Chalicodoma arise similarly from neuroblasts, but the arrangement is not so regular as that observed by "Wheeler in Xiphidium and Doryphora. — Ed.] 326 INSECTA. sides of the oral aperture into the anterior cephalic region, which represent direct prolongations of the primitive swellings (Fig. 160, b\ b2, bd), and from which are derived those parts of the brain known as the primary ganglia, and named by Vi all anes the proto- cerebrum, deutocerebrum, and tritocerebrum. These swellings become early separated as three consecutive brain-segments. Patten (No. 67) has the merit of having first drawn attention to this segmentation. 2. A large ectodermal thickening lying laterally to the swellings in the cephalic lobes just mentioned (Fig. 160 A, og). This is the rudiment of the optic ganglion, which exhibits in its postero-external J A it m. f\ / f &1 B v s* y "*" met Fig. 160. — Diagram of the development of the brain in Acilius (after Patten). .4, anterior end of the germ-band of an Acilius embryo. B, the same in three-quarter profile. at, antenna ; 61, first, 62, second, IS, third segments of the brain ; i, invagination of the optic ganglion ; i\ anterior, i", posterior portions of the invagination ; /, paired rudiment of the upper lip ; m, mouth ; mil. mandible ; mx', first, mx", second maxilla ; og, optic ganglion ; og\ first, og~, second, og3, third segment of the optic ganglion ; op, optic plate ; 1-G, rudiments of the six larval eyes ; I-IV, the four anterior segments of the ventral chain of ganglia ; I, that belonging to the pre-mandibular segment (?) ; //, that belonging to the mandibular segment ; III and IV, those belonging to the first and second maxillae. region a semi-circular ectodermal invagination (i, Patten's ganglionic invagination), which yields further elements for increasing the optic ganglion, and corresponds in position with a similar invagination found in the Decapoda (Vol. ii., p. 171) and the Arachnida (pp. 12 and 53). The part of the ectoderm lying externally to this invagination (Fig. 160 A, op} also becomes thickened, increases considerably in extent, and yields at a later period a large part of the cephalic integument and the rudiments of the eyes ; it is therefore known as the optic plate. THE NERVOUS SYSTEM. 327 The separation of the brain-rudiment from the ectoderm, like that of the lateral cords, takes place by a process of delamination. An exception to this rule is found in the composition of the optic ganglion, which is formed from the invagination described above. The fibrous tissue of the brain develops here as in the lateral cords, first on the inner surface of the brain-rudiment, and only later sinks into the interior of the brain and becomes surrounded with a layer of ganglionic cells. The rudiments of the two halves of the brain are originally distinct from one another. Later, when the dorsal part of the head has formed, the two halves of the brain shift on to the dorsal side, approaching one another until, finally, with the assistance of a median invagination (like the transverse commissures of the ganglia of the ventral cord), a commissural connection is established between them (Grassi, No. 32 ; Heider, No. 38 ; Graber, Nos. 28 and 30). The most important recent details of the development of the brain in the Insecta have been given by Patten (No. 67) for Acillus, and by Viallanes (No. 84) for Mantis. According to Patten, the whole head-rudiment shows signs of being composed of three segments (Fig. 160), this segmentation affecting not only the primary parts of the brain-rudiment, but also the rudiments of the optic ganglion and of the optic plates. On the three consecutive segments into which the optic plate (op) is thus divided, the rudiments of the six ocelli of the larva are distributed in Acilius, two ocelli occurring on each segment (Fig. 160 A, 1-6). In the shifting of the separate parts of the rudiment of the head, which takes place at a later stage in con- nection with the development of the cephalic terga, as above men- tioned (p. 302), changes occur in the position of the ocelli with regard to each other, but these we cannot here enter upon. The invagination above described also, which participates in the formation of the optic ganglion, is broken up, according to Patten, into three sections corresponding to the segmentation of the brain (Fig. 160 B, i1, ir) ; in Acilius, only the two anterior sections can be recognised as distinct invaginations, while the third is replaced by a solid ingrowth. Patten's statements have been almost entirely confirmed by Wheeler (No. 95) in the case of Doryp>hora. Cauriere's (No. 13) observations also - qo Fig. 161. — Anterior (ventral) aspect of the developed brain of Oedipodtt (after Vial- lanes). c, circum- oesophageal commis- sure ; c', transverse commissure behind the oesophagus ; do, deutocerebrum ; go, optic ganglion; If, labro-frontal nerve; na', an- tennal nerve ; na", accessory antennal nerve ; no, nerves of the three ocelli ; pc, protocerebrum ; r, root of the paired stomato-gastric ganglion ; tc, tritocerebrum. 328 1NSECTA. seem to confirm Patten's views. Heider (No. 38) and Graber (No. 30), however, although convinced of the presence of a primary segmentation of the brain (bl-b3) in Hydrophilus, were unable clearly to recognise it in the optic ganglion and the optic plate. A comparison with the condition in other Arthropoda, especially in the Crustacea (Vol. ii., p. 162 ct seq.), also supports the view that the optic ganglion is a secondary section of the brain belonging exclusively to the most anterior part. This view is in agreement also with the recent statements of Viallanes (No. 84) with regard to Mantis. According to this author, the primary part of the brain breaks up into three sections, corre- sponding to the protoccrebrum (pc), the deutocerebrum (dc), and the Iritocercbrum (tc) of the adult. Of these, the protocerebrum is connected with the optic ganglion (go) and also yields the nerves to the ocelli (no), as well as the dorsal integumentary nerves ; the deutocerebrum yields the antennal nerves (na' and na"), while the tritocerebrum gives off the labro-frontal nerves (If) which are connected with the frontal ganglion. In the rudiment of the optic ganglion, Viallanes could only recognise a division into an outer and an inner part (premier lobe protocerebral and deuxieme lobe 'protocerebral). Cholodkowsky also (No. 20) observed the segmentation in the brain of Phyllodromia. He, however, considers the optic ganglia as belonging to the third segment of the brain. The above considerations incline us to regard the primary cephalic region as being derived from three fused segments. Of these the most anterior would have to be called the true primary cephalic segment. The segment of the brain belonging to it (the protocerebrum) would be the homologue of the Annelidan brain derived from the neural plate. The second cephalic segment which we should have to identify with the antennal segment* would have to be regarded as a post-oral trunk-segment which has shifted forward secondarily (p. 295), and the third cephalic segment would also have to be regarded in a similar manner, being followed eventually by the hypothetical pre-mandibular segment and then by the mandibular segment. Taking into account what has just been said, it must appear remarkable that, so far, observers have been able to find only one pair of coelomic sacs in the primary cephalic region (p. 320). This pair, according to Cholodkowsky, belongs to the antennal segment into the appendages of which it is prolonged. We should have to assume that the pair of primitive segments between these coelomic sacs and those of the mandibular segment have been secondarily suppressed. It should be mentioned that the frontal ganglion and the unpaired oesophageal nerve connected with it are independent structures which only secondarily enter into connection with the brain. They owe their origin to an ectodermal invagination which belongs to the anterior wall of the oesophageal depression. This invagination yields the material for the formation of the frontal ganglion and the oesophageal nerves (Heider, No. 38 ; Carriere, No. 13). * It should be mentioned that Patten (No. 67) and Carriere (No. 13) reckon the antennae as belonging to the third brain-segment. [Burger, in his work on Chalicodoina (No. II.), based largely on Carriere's notes, claims the antennae as belonging to the deutocerebral segment. A pair of minute evanescent appendages were found by Carriere on the protocerebral and another pair on the tritocerebral segment. It is thus evident, from his post- humous work, that Carriere had ceased to reckon the antennae as belonging to the tritocerebrum. — Ed.] THE SENSORY ORGANS. 329 D. The Sensory Organs. The Ocelli. Detailed accounts have recently been given of the development of the ocelli by Patten (No. 67). There are, on each side, six ocelli which, according to Patten, are distributed in three pairs on what he assumes to be the three most anterior cephalic segments. The Fig. 162.— Two stages in the development of the fifth ocellus of an Acilius larva (after Patten), c, cuticular rods ; cl, rudiment of the chitinous lens ; h, hypodermis ; I, lentigen layer (vitreous body) ; n, nerve ; ?•, rudiment of the retina ; sp, vertical slit in the retina ; x, the retinal cells bordering this slit laterally. individual ocelli of these three pairs differ considerably from one another in structure and development, although a certain uniformity of type can be recognised. The fifth ocellus (the ventral ocellus of the third pair which, however, has shifted far forward in the larva) 330 INSECT A. approaches this common type the most nearly, and we shall therefore content ourselves by describing its development alone. The rudiment of this ocellus (Fig. 162 .4), at a certain stage of its development, strikingly recalls the simple optic pits or cup-shaped Fig. 103.— Two later stages in the development of the fifth ocellus of the Acilius larva (after Patten), cl, chitinous lens; i, so-called pigmented iris ; 1, lentigen layer (vitreous body); m, middle inverted layer of the eye ; r, retina ; $p, vertical slit in the retina ; st, rods ; x, cells bordering the vertical slit. THE OCELLI. 331 eyes found in certain Molluscs (Patella). It is a simple pit-like depression of a thickened part of the hypodermis. The elongate cells which compose the wall of this depression are arranged in a simple layer and, at their free ends, which are turned to the optic pit, carry a striated cuticular margin (c), while their inner or basal ends give off the nerve-fibres which unite to form the common optic nerve. Wi 13 ' T" ; ' According to Patten, this apparently simple rudiment has arisen by the fusion of at least four distinct pits which represent primary em- bryonic organs, and in structure recall the eye- pits on the margin of the mantle in Area. The nerve correspondingly shows its composition out of four originally separate bundles. In later stages, the eye-pit closes towards the exterior (Fig. 162 B), the marginal parts pushing inward until they meet over the deeper parts. In this way the pit-like rudi- ment gives rise to an eye-cup which has by this process become bilaminar. The cen- tral part of the outer or superficial layer (I) becomes the lentigen layer (vitreous body), while the peripheral parts become the pig- mented iris. The cuticular margin of these cells gradually gives rise to the cuticular chitinous lens (cl) of the ocellus. Laterally, the superficial layer of the eye passes direct into the hypodermis (/*). The deeper layer of the eye, which still retains its cup-like curve, m, ri Fig. 164. — A, rudiment of the eye in a Hydrophilus larva just hatched. B, a somewhat older larva (after Patten), cl, chitinous lens ; h, hypodermis ; I, lentigen layer; m, middle layer of the optic rudiment ; n, nerve ; o, aperture of the optic invagination ; r, retinal layer ; rb, rods (in A arranged in a single row). 332 INSECT A. must be considered as the rudiment of the retina (r). From its cuticular margin are derived the optic rods. Certain peculiarities characteristic of the eye of Acilius now develop. The chief of these is a slit traversing the retina perpendicularly (sp), which is bordered by the horizontally-placed rods of the large retinal cells (x) which lie next to it. In the further course of development (Fig. 163) a flattening of the cup-like cavity occurs, owing to the growth of the cells forming the base of the optic cup, and the rods belonging to these cells Ps ffk l consequently assume a more vertical position. The retinal cells at the edge of the cup, on the contrary, curve inward and form an inverted marginal layer (m) with its rods directed towards the base of the retinal cup ; these cells may be regarded as the rudiment of a third layer intercalated between the two principal layers of the eye. The above would justify us in deriving the trilaminar Insectan eye from a three-layered eye by the atrophy or incomplete development of the middle layer. The original presence of three layers in the ocellus is, according to Patten (No. 66), still more distinctly recognisable in the eye of the youngest larvae of Hydropliilus (Fig. 164), in which the optic invagination presses into the optic rudiment not from the middle, but from the edge and from the dorsal side (Fig. 164 A). Even in later stages a vestige of the middle cell-layer (Fig. 164 B, m) is retained. According to Grenacher's observations (No. 151), thc- ocelli of certain Insects appear to remain throughout life in a much more primitive condition than would be expected from Patten's statements, the optic vesicle in them never closing completely, and the layers of retinal cells and of cells forming the lens remaining in direct continuity with the hypo- dermis (e.g., Fig. 165). The statements of Patten do not agree with those of Carriere (No. 14). If we rightly understand the latter author, in the development of the ocelli of pupae of Chrysididae and Ichncumonidac, the separation of the retinal layer from the lentigen layer takes place by delamination, while the optic invagination which forms later develops according to the type of the cup-eye, and at the same time stands in a certain relation to the development of the corneal lens. The larvae of the holometabolic Insecta are, as a rule, devoid of compound lateral eyes (facet-eyes). These develop only in the gradual transition to the imaginal stage. The larvae, on the contrary, Fig. 165.— Section through the eye of a Coleopteran larva (Dytiscus) (after Grenacher, from Hatschek's Text-book), c, chitinous cuticle ; I, corneal lens ; h, hypodermis; pz, pigment - cells ; gk, vitreous body ; r, retina ; b, basal membrane. THE OCELLI AND THE COMPOUND EYES. 333 possess a number of laterally placed ocelli (very often six). The question now arises as to the relation of the compound lateral eyes of the imago to the ocelli of the larva. It is certain that the latter degenerate and are not taken over into the imago. In the pupa of the Lepidoptera, the degenerating ocelli can be seen detached from the hypodermis, and drawn back into the interior of the larva on the optic nerve as on a stalk (Carriere, No. 147). Since, at this time, the rudiment of the compound eye can be seen as a hypodermal thickening, it might be thought that the latter was altogether a new acquisition. But, according to Patten's observa- tions on Acilius, there seems to be a certain relation between the larval and imaginal eyes. In Acilhis, the highly developed and complicated larval eye (the first) has a peculiar dorsal appendage, which perhaps represents the vestige of an ocellus. The hypodermal thickening, which leads to the development of the imaginal lateral eye, develops first in the neighbourhood of this appendage. In later stages, this rudiment forms a thickened band which almost com- pletely surrounds the six ocelli. This position perhaps favours a view which regards the complex of the six larval eyes and the compound eye that develops later merely as differently developed parts of one and the same optical area. We should here recall Grenacher's view, according to which the omma- tidia of the compound eye on the one hand, and the ocelli on the other, represent merely different ontogenetic forms and grades of development of one and the same type of eye (chap, xxviii.). The frontal ocelli of the imagines of many Insecta have nothing to do with the larval ocelli. Against Patten's view that they may perhaps stand in nearer relation to the compound eyes we might adduce the independent condition of their innervation. These ocelli are often three in number. Patten (No. 66) observed in Vespa that the median unpaired ocellus is derived by fusion from a paired rudiment. The details of the development of the compound lateral eyes (fan-shaped or facet-eyes) are so far chiefly known in connection Fig. 1C6.— Section through the rudi- ment of the compound eye oUVcspa (after Carriere, from Hatschek's Text-bool). 1, cylindrical cells (later accessory pigment-cells); 2, crystal- line cone-cells ; 3, principal pigment- cells ; U, retinulae ; 5, nerve, which gives off branches to the different ommatidia. 334 INSECT A. with the pupae of the holometabolic Insecta (Diptera, Weismann, No. 129; Lepidoptera and Hymenoptera, Carriere, No. 147). In many cases, the first rudiment of the facet-eyes, as we saw above, is a paired lateral ectoderm -thickening, while in others, only a crowding together of the individual ectoderm -cells can be seen. The separation of the single ommatidia (single eyes) here takes place exclusively through histological differentiation (Fig. 166). At an early stage, as described in connection with Mysis (Vol. ii., p. 169), the single ommateal pillars and the undifferentiated tissue between them, which in Vespa form very massive intermediate inllars, can be distinguished. In the region of the ommateal pillars the cells become arranged to form two layers, the' outer yielding the cells of the crystalline cone {2) and the principal pigment-cells (3), while the inner yields the retinulae (4), which are connected with the nerve-fibres (J). The cells of the crystalline cone secrete outwardly the cuticular corneal lens, while the crystal- line cone develops within them in eyes of the eucone type. The cells of the intermediate pillars (2) give rise to the so-called accessory pigment-cells. In the course of further development the optic rudiment thickens considerably, the single ommatidia thus becoming taller and narrower, and also shifting closer together. The retinula-cells especially gain greatly in height. Pigment becomes deposited both in the retinula-cells and in the various pigment-cells which cover the outer side of each ommatidium. The development of the most essential features of the ommatidia seems to be thus completed (Carriere). E. The Tracheal System. The tracheae arise as paired, segmentally arranged ectodermal invaginations lying laterally to the limb-rudiments (Fig. 146 A, st, p. 295; Fig. 147 A st, p. 297; and Fig. 158 E, tr, p. 321). The tracheal invaginations are usually found developing at a somewhat early stage soon after the appearance of the limb-rudiments. It, however, appears from one of Grassi's observations (No. 33) that, in Japyx, the tracheal system does not develop until a late embryonic stage. This would have to be regarded as the more primitive con- dition, recalling the Myriopoda (pp. 243 and 254). For since the tracheal system, phylogenetically considered, represents one of the latest acquisitions of the racial form, its early development in most Insects must be considered as being secondarily shifted back to early stages on account of its importance. THE TRACHEAL SYSTEM. 335 The originally simple tracheal invaginations (Fig. 158 E, tr) first widen at their bases, and then soon give off the tracheal branches as diverticula, while the narrowed aperture of the invagination is retained as the stigmatic branch and stigmatic aperture. The two longitudinal trunks of the tracheal system arise from the consecutive tracheal invaginations; these give rise to horizontally placed out- growths, which grow out longitudinally until they meet and finally fuse with those of the next segment (Butschli, No. 11). Only in late embryonic stages is the cuticular tracheal intima secreted. The tracheae become filled with air, according to "Weismaxn (No. 87), to a certain extent even before the embryo hatches, the air being exuded, as it appears, from the tissues and the body-fluid. The further development of the tracheal ramifications is brought about, as far as has yet been observed, by the continuous formation of diverticula. The branches which thus arise are therefore intercellular structures. On the other hand, it should be mentioned that the finest tracheal branches are intracellular canals. Although Schaffer (No. 124a)* has correctly pointed out that the difference between these two methods of formation is not of any great impor- tance, since in both cases there is merely an increase of surface (of a cell-plate in intracellular origin, or of a few cells in intercellular formation), the distinc- tion has a certain interest when we compare the condition in Peripatus. The tracheae of Peripatus consist of numerous very fine tubes which, united to form a tuft, arise from a short funnel connected with the stigma. We may perhaps consider the fine tubes of Peripatus as equivalent to the intracellular and the funnel to the intercellular portion of the tracheal system of the Insecta. At certain stages of development the tracheae bear a strong resemblance to the rudiments of the salivary glands f and the Malpighian vessels. This circum- stance, as well as the position and number of these invaginations in the Hymenoptera, gave support to the view (held by Butschli, No. 11, Grassi, No. 32, and to some extent also by Carriere, No. 13) that we have in the tracheae and these glands homologous organs. If we consider the anatomy of Peripatus, we shall find that there are objections to this view. The apparently irregular distribution of the tracheae in Peripatus, and the facts that glands (salivary and excretory) similar to those of the Insecta and perhaps homologous with them are also present, show that the agreement in position and in number is of no consequence. Above all, however, the tracheae of Peripatus in structure differ greatly from the glands under consideration. Even Moseley's view that the tracheae are transformed integumental glands, a view also held by Palmen, offers many difficulties. Apart from the circumstance that, in forms that stand nearer to the conjectural racial form of the Tracheata, integumental glands of this kind are not known to exist, the transformation of a secreting organ into an air-filled respiratory organ presupposes a change of function difficult to imagine. It is therefore most probable that we must regard the tracheal * Here also the literature on this point will be found. t In the efferent portion of many spinning glands also a spiral thread altogether similar to that found in tracheae develops. But the fact that a similar spiral thread occurs also, for instance, in the vas deferens of the Ci/theridac shows that no weight can be laid on this circumstance (KAUFMANN). 336 INSECT A. invagination as a structure mi generis. We may here point to the condition of the terrestrial Isopoda, in the branchial lamellae of which air-containing spaces altogether analogous to tracheae develop {Tylus). F. The Alimentary Canal and Intestinal Glands. Of the three sections of the alimentary canal, the stomodaeum, enteron, and proctodaeum, the first and third arise as ectodermal invaginations. In most cases the rudiment of the stomodaeum appears in the germ-band somewhat earlier than that of the procto- daeum (p. 294 and Fig. 145 C, m, p. 291). The musculature of these sections is yielded by the surrounding mesoderm. In the stoinodaeal invagination an unpaired dorsal depression soon appears, from which are derived the frontal ganglion and the oesophageal nerve (p. 328). The actual ectodermal character of the fore- and hind-guts has been established with considerable certainty by the unanimous testimony of observers, and by comparison with the conditions in other groups of Arthropoda. Voeltzkow (No. 85), indeed, has recently derived both structures from the lower layers, and Graber has adopted this view for the proctodaeum of Musca. With regard to this, we must refer to what has already been said (p. 315) as to the condition of Musca. The connection between the cavities of the stomodaeum and proctodaeum with that of the enteron is usually established at a somewhat early embryonic stage. In certain larval forms, however (many Hymenoptera, e.g., Apis and Myrmeleon), no communication is established between the enteron and the proctodaeum, the latter then having an exclusively excretory function. The enteron develops from two originally distinct rudiments, the anterior and posterior entoderm-rudiments (p. 313), which from the very first bear a close relation to the invaginations of the stomodaeum and proctodaeum. Although originally applied as simple cell-accumu- lations to the inner ends of these invaginations, so closely, indeed, that Voeltzkow (No. 85, 86), Patten (No. 68), and Graber (Nos. 28 and 30) derived them directly from the epithelium of the latter, they soon extend, through continuous cell-proliferation, and assume the U-shape (Fig. 153 A, en' and ere").* The ends of the U-shaped rudiment in the anterior entoderm-mass are directed posteriorly, but in the posterior mass anteriorly. These ends grow out towards each other, meet and fuse, and thus form two paired entoderm-bands running dorsal to the germ-band along its whole length. * [As already stated (footnote, p. 317), Heymons (No. XX.) and Lecaillon (No. XXIX.) consider that the entoderm is quite wanting in the adults of the higher Insects, the mid-gut originating, according to these observers, as ecto- dermal ingrowths from the stomodaeum and proctodaeum. — Ed.] THE ALIMENTARY CANAL AND INTESTINAL GLANDS. 337 The paired entoderm-bands belong to the lateral parts of the germ-band. They lie, as a rule, immediately below the series of consecutive coelomic sacs (Fig. 158 F, p. 321), where the entoderm- bands (en) are seen cut through below (really dorsal to) the coelomic sacs (us). The dorsal wall of the primitive segments is thus in immediate contact with the entoderm-bands. Active proliferation of cells now takes place from this wall of the primitive segments, and the cell-material thus produced, which splits off from the dorsal wall of the primitive segments, forms the outer layer of the enteron- rudiment, the splanchnic layer or intestinal fibrous layer (Fig. 158 F, spm; Fig. 170, sp, p. 344). The remainder of the dorsal wall of the coelomic sac left after this separation enters into relation with the genital rudiment and yields the terminal filament (cf. p. 345, etc., and Fig. 170, ef). The entoderm-bands with the immediately contiguous splanchnic layer may now be termed the mid-gut rudiment (Fig. 150, m, p. 306, and Figs. 170, 171, 172, sp + en, pp. 344-346). In the following stages the enteron is distinguished for its great lateral growth, which causes it to spread over the surface of the food-yolk, which it finally completely surrounds (Fig. 150 C-F, p. 306, and Figs. 170-172, pp. 344-346). This circumcrescence, as a rule, occurs in such a way that the two entoderm-bands unite first in the region of the ventral middle line (Figs. 150 F, and 171). Only later do they unite on the dorsal side (Figs. 150 F, and 152). The food-yolk in this way comes to lie entirely within the enteron, and with it is included the remains of the dorsal tube or dorsal organ (Fig. 150 F, s) where such is present. The description just given of the development of the enteron, which rests principally on the observations made on Hydrophilus and Pltyllodromia, seems to be directly applicable to most Insects. In individual cases, indeed, we find certain deviations, as, for instance, in Musca, where the coelomic sacs do not attain distinct development (p. 320), and where the whole of the food-yolk is not taken into the enteron, a portion of it remaining (as in other Diptera) in the body-cavity where it is gradually absorbed (Kowalevsky, Voeltzkow, Graber). The conditions also differ to a certain extent in the Hymenoptera (Apis, C'haHcodoina, Kowalevsky, Grassi, Carriere), where the entoderm originally occupies a dorsal position (p. 316) and is only gradually grown over by the germ-band. The circumcrescence of the food-yolk by the entoderm here proceeds from the dorsal to the venti-al side. The salivary glands which open into the buccal cavity and may consist of several pairs* (1-3), arise as ectodermal invaginations * According to Schiemenz (No. 125), the various cephalic glands of Apis (imago) are distributed in such a way that, originally, a pair occurs on each of the three maxillary segments. 338 INSECTA. which originally open, not into the stomodaeum, but on the surface of the body. They may therefore be regarded as integumental glands, the apertures of which have been drawn into the buccal cavity (?). In the Trichoptera and Lepidoptera, an anterior pair of these glands develops in the anterior and inner angle of the mandibular rudiment (Hatschek, No. 36 ; Patten, No. 65). A second pair which here, as in the Hymenopteran larva, is transformed into spinning glands, belongs to the segment of the second maxillae (Fig. 143 A, sp, p. 286). Carriere, however, reckons it as belonging to the first thoracic segment. When the second maxillae fuse to form the lower lip, the apertures of the paired invaginations are approximated, and a short unpaired efferent duct forms, opening into the buccal cavity (Butschli, No. 11, etc.). We should be predisposed to homologise the salivary gland of the Insecta with the glands which open into the mouth in the Myriopoda. This view is opposed by the consideration that the latter, as transformed nephridia, are said to arise from the mesoderm (p. 251), while the salivary glands of the Insecta are purely ectodermal structures. We must therefore leave the question of the homology of these organs and of their relation to similar glands in PeHpdfacs to be decided by further research. The Malpighian vessels develop as paired outgrowths of the proctodaeum, which from the very first have a lumen. They are thus ectodermal structures. Two or three pairs usually make their appearance (Lepidoptera, Phryganeidae, Hydrophilus). In those forms which, at a later stage, have a larger number of these vessels, these secondary tubules develop as diverticula of the primary ones (Gryllotalpa, Rathke). The Malpighian vessels usually appear only after the development of the proctodaeal invagination, as diverticula of this latter, but in the Hymenoptera (Apis and ChaHcodoma), they form before the proctodaeum develops, as invagina- tions of the ectoderm, and consequently at first open on the surface of the germ-band. They then somewhat resemble in appearance tracheal invaginations, and this perhaps led to their being homologised with the latter, a view which we are unable to share, and which Carriere also (No. 13) did not adopt. Only later do they shift with the developing proctodaeal invagination into the interior of the embryo. G. Heart. The earliest recognisable rudiment of the dorsal vessel or heart in the Insecta appears as a longitudinal strand of cells (cardioblasts) running along the upper and external border of the dorsal sub- division of the primitive segments (Fig. 170, li, p. 344, and Fig. 171, h). During the continuous circumcrescence of the yolk by HEART. 339 the germ-band, this rudiment shifts more and more towards the dorsal side. It is directly connected with the wall of the primitive segments (Figs. 170 and 171), and indicates the junction of the dorsal with the lateral wall of the coelomic sac. According to Korotneff (No. 47), whom we have to thank for the first detailed account of the development of the heart in the Insecta, the cardio- blasts are derived from the wall of the primitive segments. In GrylJotalpa, the form described by Korotneff, the condition is in many respects peculiar. The formation of the dorsal organ is here introduced, in the way described above (p. 304, etc.), by the rupture of the embryonic envelopes. The serosa contracts to form a thickened plate (Fig. 167 A, rp), of which the very degenerate amniotic folds appear as a lateral appendage (am), the whole being far removed dorsally from the edges of the germ-band (*x-y*) (cf. Fig. 150 C, p. 306). The interval between the rudiment of the amniotic fold and the lateral edge of the germ-band (*x-y*) is occupied by an epithelial lamella (I) in which we recognise the former amnion. This lamella is not closely applied to the yolk, but is separated from it by a spacious blood-lacuna (bs), in which can be seen numerous blood-corpuscles that have immigrated from the mesoderm. The cardioblasts which are derived from the wall of the primitive segment (us) have become arranged to form a channel (gr) on each side, and thus surround the lower part of the blood- sinus. As the circumcrescence of the food-yolk by the germ-band pro- gresses, after the invagination and degeneration of the dorsal plate has taken place, the two blood-lacunae fuse together dorsally to form one lacuna (Fig. 167 B, bs). This now represents the rudiment of the lumen of the heart. The two vascular channels, moving towards -each other until they come into contact, form, by fusing together, the wall of the heart (Fig. 167 C, r, and Fig. 172, h, p. 346). The venous ostia arise, according to Butschli (Xo. 11), as paired invagina- tions of the lateral walls, in the base of each of which a slit develops. The rudiment of the heart, as we have seen, is intimately connected with the primitive segments. The lateral Avail of the primitive segments, after giving off the elements of the somatic mesoderm, gives rise to an epithelial plate which represents the first rudiment of the pericardial septum or dorsal diaphragm (Fig. 167 A-C, del; Fig. 170, p. 344; Fig. 171, p. 345, and Fig. 172, ps, p. 346). As soon as the two halves of the rudiment of the heart have united in the middle line, the two halves of the pericardial septum also become 340 INSECT A.. & connected and bound the* pericardial space, which is closed towards the rest of the body-cavity (Fig. 172, ps). For a time the pericardial sep- tum remains connected with the wall of the heart, but it becomes- separated from it later (Fig. 167 C, dd). The relations of the rudi- ment of the heart and the pericardial septum to the terminal filament of the genital rudiment will be discussed below (p. 345 et seq.). The statements as to the development of the heart in the Insecta that have been made by other authors (Giiassi, Patten, Ticho- mihoff, Ayees, Heidee, Caeeieke, Heymons, etc.)' can" easily be traced back to the type described for Gryllotalpa. The principal difference in the formation consists in the absence or slighter development of the large blood-lacunae above described. The rudiment of the heart in the first stages is consequently of small extent, and can often hardly be recognised. Fig. 167. — Diagrammatic transverse section through three consecutive stages of Gryllotalpa to illustrate the formation of the dorsal vessel (after Korotnefk). (The rudiment of the fibrous layer of the intestine is omitted in these diagrams). A, youngest stage. The germ-band exiends from *x to y*. The embryonic envelopes are rent and retracted dorsally. am, edge of the rent ; rp, dorsal plate (serosa); I, lamella connected with the ectoderm of the germ- band (amnion reflected back). B, second stage. The germ-band has almost completely grown over the food-yolk. The dorsal organ is absorbed. » C, third stage, dorsal region. The formation of the heart is completed, am, vestige of the amnion-fold ; bs, blood-sinus ; dd, rudiment of the dorsal diaphragm ; dv, veutral diaphragm ; do, food-yolk ; dz, yolk-cells ; ec, ectoderm; gr, vascular channel (rudiment of the heart); I, lamella of the reflected amnion; lh, permanent body-cavity; m, transverse muscle; n, ventral cord; r, heart; rp, dorsal plate ; sp, splanchnic, so, somatic layer of the mesoderm ; its, cavity of the primitive segment; *x-y', lateral termination of the germ-band. THE MUSCULATURE, CONNECTIVE TISSUE, AND FAT-BODY. 341 111 Gryllotalpa and Oceanikus (Ayers) the posterior portion of the heart develops first. The order of development of the heart is here from behind forward. This is an unusual condition, which is due to the fact that the closing of the dorsal region is retarded by accumulated masses of yolk in the anterior part of the body. [As in Gryllotalpa, the heart in Agelastica (Peteunkewitsch, No. XXXVI.) and Bomhyx (Tichomiroff, No. 78) is formed by the eircumcrescence of the yolk by the mesoderm-bands. When these reach the mid-dorsal line, the two layers fuse immediately below the ectoderm, while they remain distinct below ; in this way a mesodermal groove arises which, owing to the fact that the ento- dermal epithelium is still incomplete dorsally, is in open communication with the yolk. This constitutes the gastro-vascular canal of Tichomiroff, which, in transverse section, has the form of a figure 8. At this stage the dorsal and lateral walls of the heart are formed by mesoderm, while the incomplete ventral wall is entodermal. The entodermal epithelium now unites in the middle line, and thus completely separates the heart from the enteron and, soon after, the mesoderm grows in from either side towards the middle line below the cavity of the heart and above the entoderm, and finally, by fusion of the two ingrowths, the mesodermal walls of the heart are completed. The closure of the heart takes place earlier at the anterior and posterior ends than in the middle. — Ed.] The blood-corpuscles are traced back by Korotneff to cells of the somatic mesoderm which have lost their connection with the rest of the mesoderm and have passed into the body-cavity. Our own researches incline us to agree with this statement. Other authors, however (Dohrn, and recently Will also, No. 97), have derived the blood-corpuscles from yolk-cells. Ayers (No. 1) even claims for their formation the cells set free by the disintegration of the dorsal plate. It should here be pointed out that Schaffer {No. 124a) recently maintained that certain cell-complexes connected with the fat-body in caterpillars are formative centres for the blood- corpuscles (p. 372). H. The Musculature, the Connective Tissue, and the Fat-body. The groups of muscles, as well as the connective tissue, are derived through histological differentiation from the somatic layer of the meso- derm (Fig. 167, so). Our own researches, and those of Kowalevsky, <3rassi, and Carriere, show that the fat-body also arises from the mesoderm. In Hydrophilus, a dorsal band-like fat-body, running over the intestine, arises by direct transformation from the wall of the coelomic sac. For the rest of the fat-body also, for instance, for the lobes accompanying the tracheal system, a mesodermal origin can be indisputably established. The observations made by Heyjions on Phyllodromia are in harmony with this. Certain cells of the wall of the coelomic sac early undergo a transformation, which results in their being recognisable as the rudiment of the future fat-body (Fig. 169 B and C, /). 342 INSECT A. Many authors, however, differ greatly as to the origin of the fat-body. Doiirn, Tichomiroff (No. 79), and, recently, Will have derived it from the yolk-cells, while other authors have claimed an ectodermal origin for it. Among these latter are Korotxeff (No. 47) and Schaffer (No. 124a), who, in confirming former statements made by Weismann, trace back the fat-body of Musca to growths from the tracheal matrix and partly from the hypodermis. Graber also (No. 31) has recently maintained the ecto- dermal origin of the fat-body in Hydrophilm and Stenobothrus. As regards Hydrophilus, we are unable to agree in this view. I. Genital Organs. The published accounts of the develop- ment of the genital organs are, with the exception of those relating to the pecu- liar and specialised conditions in the Aphidde and Diptera, much scattered and for the most part fragmentary and unsatisfactory. We must refer the student to the works of Balbiani (No. 3), and Witlaczil —om, (No. 98), and especi- ally to those of Hey- mons (No. 43). We are able, however, to gather (chiefly from the writings of Grassi, No. 32, Heider, No. 38, and Wheeler, No. 95) that the genital glands are mesodermal in origin, and develop from the wall of the coelomic sac. The development of the efferent ducts has been best described by Nusbaum (No. 61) and Palmen (No. 162). Heymons (No. 43) has recently published accounts of the rise of the genital organs in Pie. 10S. — Lateral sagittal section through the abdom- inal part of a germ-band of Phgllodromia gerw.av.icd after the primitive segments have completely formed (after Heymons). 1-7, tirst seven abdominal segments, from the eighth (8) to the terminal segment (es) the abdominal germ-band is flexed ventrally ; am, amnion; c, coelomic sac ; d, food-yolk ; es, terminal segment ; gz, genital cells, lying partly in the dissepiments and partly in the wall or the cavities of the primitive segments. GENITAL ORGANS. 343 —•jo Fig. ICO. — Transverse section through the abdomen in three consecutive stages of develop- ment of Phyllodromia germanica (after Heymons). am, amnion ; bg, rudiment of the ventral chain of ganglia ; c, coelomic cavity ; c', dorsal, and c", ventral sections of the coelomic sac ; cz, cells of the wall of the primitive segments, applied to the ventral side of the genital rudiment ; d, food-yolk ; dw, dorsal wall of the coelomic sac ; ec, ectoderm ; ep, epithelial cells ; ex, abdominal limb-rudiments ; /, rudiment of the fat-body ; gz, genital cells ; Iw, lateral wall of the coelomic sac ; m, mesoderm-cells, which do not take part in the formation of the coelomic sacs; miv, median wall of the coelomic sac; so, somatic mesoderm-layer ; vm, ventral longitudinal muscle. 344 INSECTA. Phyllodromia germanica, on which, as the most detailed, we shall found our description.* In Phyllodromia, distinct genital cells can be distinguished at an early stage of embryonic development by their different histological character. They are larger than the other cells, and show a slightly stainable nucleus with a distinct nucleolus. Tbese genital cells, which have developed by the transformation of the embryonic mesoderm -eel Is, originally lie in the splanchnic layer or on the surface of this layer, which is turned towards the food-yolk at V f* - w Fig. 170. — Transverse section through the abdomen in a somewhat older germ-band of Phyllodromia germanica (after Hevmons). bg, rudiment of the ventral chain of ganglia ; c, remains of the coelomic cavity ; cz, rudiment of the efferent genital duct ; ec, ectoderm ; ef, terminal lilament; en, entoderm; fk, fat-body; gz, genital cells; 7<, rudiment of the heart; jj, rudiment of the pericardial cavity; ps, rudiment of the pericardial septum; so, somatic mesoderm; sjj, splanchnic mesoderm. the boundaries of the segments. After the coelomic sacs are com- pletely formed (Fig. 168, gz), they are found in the dissepiments which divide the consecutive sacs from one another. New genital cells continually develop here from mesoderm -cells. These cells develop in the second to the seventh abdominal segment.! * The following description and figures were taken in advance from a treatise since published in the Zcitschr. f. tciss. Zool., Bd. liii., kindly placed at our disposal by the author. f [According to recent researches on the origin of the germ-cells in Crustacea and other Invertebrate, we should have to look for the first origin of these cells at an early cleavage-stage, at any rate, they should be visible as soon as the mesoderm-bands are formed. Further, we should not expect them to arise or to increase from the ordinary mesoderm-cells, but from germ-teloblasts. — Ed.] GENITAL ORGANS. 315 The crenital cells shift later into the interior of the coelomic sacs, soon reaching their dorsal Avail (Fig. 169 A, gz) and passing in between its cells. The coelomic sacs (c), in transverse section, at this stage are approximately triangular in outline, so that we may distinguish a dorsal (dw), a lateral (Iw), and a median wall (mw). The dorsal wall is in contact with the surface of the yolk, and later, by delamination, yields the splanchnic mesoderm (Fig. 170, sp), while from its remainder is formed the terminal filament of the Fig. 171.— Transverse section through the abdomen in an older germ-band of Phyllodromia gernianiea, in the stage when the circumcrescenee of the yolk commences (after Hetmons). bg, ventral chain of ganglia ; c, remains of the coelomic cavity ; ez, rudiment of the efferent genital duct; ef, terminal filament; en, entoderm; fk, fat-body; gz, genital cells; h, rudi- ment of the heart ; ps, pericardial septum ; sp, splanchnic mesoderm ; vm, ventral longitudinal muscle. genital gland (ef). The lateral wall, which is parallel to the germ- band, takes an active part in the formation of the somatic layer (Fig. 169 C, so) of the mesoderm, the pericardial septum being eventually derived from what remains of it (Fig. 170, ps). When the genital cells have entered the dorsal wall of the primitive segments, they are already so numerous as to form a continuous strand running from before backward. The genital 34G INSECTA. abdominal segment into the eighth rudiment then consists of a cell-strand lying on each side in the dorsal wall of the primitive segments, and extending from the second Not only genital cells take part in the formation of this strand, but undifferentiated mesoderm-cells (Fig. 169 B, G) are added ; these originate in the dorsal wall of the coelomic sacs, and become closely applied to and in part surround the genital cells. These mesoderm-cells form the epithelial cells (ep) of the genital rudiment, while others compose a cell-strand (cz) lying medianly and ventrally to the genital cells. The genital cells in the female give rise solely to the egg-cells (and to the nutritive cells, in those forms in which these occur). The follicular epithelium of the oviducts, on the contrary, as well as a corresponding cell-layer of the terminal chamber are yielded by the mesodermal epithelial cells.* Phyllodro- 7nia, to which the above description refers, and the Orthoptera generally, show in this respect a somewhat simple condition, the germ- or ter- minal chamber of the ovary consisting in them of com- paratively few cells. In most other Insecta, and especially in those which have a great * [Each ovarian tube in the adult Insect consists of three parts : (1) the ter- minal filament, (2) a terminal chamber, often serving as a nutritive chamber, and (3) the actual ovarian tube divided into chambers, each containing an ovum ; this last segment is much the longest. — Ed.] Fig. 172.— Transverse section through the ab- dominal region of an embryo of 1'hyllodromia germaniea, after the circumcrescence of the yolk and the formation of the dorsal surface are com- pleted, bg, ventral ganglionic chain ; cz, rudi- ment of the efferent genital duct ; d, food-yolk ; r/, terminal filament ; en, entoderm ; fk, tissue of the fat-body; gz, genital cells; h, heart; ps, pericardial septum ; s, tracheal stigma ; sp, splanchnic mesoderm ; vm, ventral longitudinal muscle. GENITAL ORGANS. 347 number of nutritive cells in the ovary, the germ- or terminal chamber is extremely large. In this connection it has been maintained that the various cell-elements of the Insectan ovary originate from indif- ferent cells (Korschelt, Xo. 155, Wielowiejski, and others). The ventral cell-strand (cz) becomes transformed in the proximal part of the oviduct, which widens into a cup and receives the separate ovarian tubes. The transformations which take place in the male will be described later. The extent of the coelomic sacs, during the further course of development, becomes much restricted through the degeneration of Fig. 173. — Longitudinal section through the female genital rudiment of Phyllodiomia gerrncm ica (after Heymons). In A, the development of the ovarian tubes is beginning. In B, they have advanced further. c~, rudimeDt of the efferent genital ducts; cf, terminal filaments; ep, nuclei of the epithelial cells ; gz, genital cells. those parts of them which extend into the limbs (p. 320), by the development of the fat-body (Figs. 169,/, and 170, /A), and through the separation of the somatic and the splanchnic mesoderm-layers (Fig. 170, so, sp). Only a small portion of them (c) finally remains, bounded laterally by the rudiment of the pericardial septum (ps), and internally by the terminal filament of the genital gland (ef). The dorsal region, where these two lamellae pass into each other, seems to be intimately connected with the cardioblasts (h). The strand-like genital rudiment now appears attached to the terminal filament as to a mesentery (Fig. 170, gz). 348 IXSECTA. During the degeneration which takes place through the lateral circumcrescence of the yolk by the germ-band (Figs. 171, 172), the paired rudiment of the heart shifts more and more towards the dorsal median line, and the genital rudiment connected with it by means of the terminal filament follows it. This rudiment thus comes to lie on the dorsal side of the developing enteron (Fig. 172, gz). The terminal filament (ef) originally represents a simple longi- tudinally-placed epithelial plate. A rearrangement of its cells soon, however, takes place, these becoming arranged in vertical rows, each of which corresponds to a developing ovarian tube. In this way the original plate-like terminal filament is broken up into the separate terminal threads of the ovarian tubes (Fig. 173, ef). In this process, however, the uppermost dorsal margin of the plate-like terminal filament does not participate, but persists as an undivided filament in the adult, where it is prolonged anteriorly and connects the different ovarian tubes; this is the so-called Mutter's thread. The latter is originally connected with the pericardial septum, but at a later stage it appears to lose this connection. As the separate ovarian tubes, which in Phyllodromia number about twenty, develop, they bend continuously inward from the dorsal towards the ventral side of the ovarian rudiment (Fig. 173). At the same time the epithelial cells (ep), some of which originally lay between the genital cells, become arranged so as to form an epithelium covering the surface of the ovarian tubes, this epithelium further secretes on its outer surface a structureless cuticular tunica propria. The outer peritoneal envelope of the ovaries is formed from the cells of the surrounding fat-body. The genital rudiment originally, as Ave have mentioned, extended from the second into the seventh abdominal segment. In the latter, however, the genital cells are from the first few in number and completely disappear later, so that the genital strand here seems composed solely of epithelial cells. This part is the rudiment of the oviduct proper, and forms a direct continuation of the cell-strand (cz) mentioned above as lying ventrally to the genital cells, out of which, as we have seen, the proximal, cup-shaped portion of the oviduct is formed. The posterior section of the oviduct bends round towards the ventral side, and becomes connected with the hypodermis at the boundary between the seventh and eighth abdominal segments. The rudiment of the oviduct is originally a solid strand, in which a lumen forms later through the shifting apart of the cells. GENITAL ORGANS. 349 Iii later stages the genital rudiment shortens considerably, so that it is then restricted to a smaller number of abdominal segments than at first. The separate ovarian tubes at the same time pass from their original vertical position and become more horizontal. The paired insertion of the rudiment of the oviduct into the hypodermis of the intersegmental furrow between the seventh and eighth abdominal segments recalls conditions observed by Palm£n in certain Ephemeridae, in which the paired aperture of the genital efferent ducts is retained throughout life. This is the original condition in the Insecta. In the female of Phyllorfromia, the un- paired terminal section of the genital passage develops during larval life from an ectodermal invagination, a genital pouch forming in which the egg-cocoon is carried. This genital pouch is formed, as Haase has proved, through the invagination into the interior of the body of the chitinous sternal plates of the eighth and ninth ab- dominal segments. With regard to the development of the efferent genital ducts in the Insecta, we must refer to the results obtained by Nusbaum (No. 61) and Palmen (Xo. 162), which fully agree with those here mentioned as obtained by Heymons for Phyllodromia. Nusbaum studied the development of the efferent ducts in Pediculina and Periplaneta. He found that only the vasa deferentia or the oviduct is derived from the posterior portions of the genital rudiment, i.e., from mesodermal structures, while the rest of the efferent apparatus (uterus, vagina, receptaculum seminis, ductus ejaculatorius, penis, and all the accessory glands) develop from the integumental epithelium, and are therefore of ectodermal origin. The unpaired portions (the uterus, penis, receptaculum seminis, and unpaired glands) develop out of paired hypodermal rudiments. The posterior strands of the genital rudiments become applied to the hypodermal growths just mentioned and fuse with them. A fusion in the middle line of the paired hypodermal growths gives rise to the rudiment of the unpaired organs. These observations agree entirely with the results obtained by Palmen from the standpoint of com- parative anatomy. Palmen found the most primitive type of efferent ducts in Heptagenia (an Ephemerid), an unpaired section being here altogether wanting. The oviducts open separately into the fold between the seventh and eighth abdominal segments, while the vasa deferentia open into a paired penis on the posterior margin of the ninth sternite. An unpaired section may develop from this paired rudiment in individual cases ( $ Forficula, Meinert) through defective 350 INSECTA. formation, one side atrophying after the paired terminal hypodermal growth has fused. In most cases, however, the unpaired terminal section must be regarded as a secondary integumental invagination. This point has not yet been investigated in all groups of the Insecta. The agreement thus found to exist in the position of the genital apertures in Phyllodromia (Heymoxs) and in the Ephemeridae (which, according to Palmen, the Perlidae also resemble) may perhaps justify us in concluding that the opening of the genital glands on the boundary between the seventh and eighth segments corresponds to the primitive condition for all groups of the Insecta, and that the more posterior position of the apertures found in many forms has arisen secondarily by a backward displacement. If this is the case, we must assume that the condition in the Thysanura, in which the genital aperture is paired and opens between the eighth and ninth abdominal segments, or on the last, is a secondary modification (cf. Haase, No. 153). The external genital appendages arise in most Orthoptera (as Dewitz has proved for the Locustidae) out of two pairs of cone-like projections belonging to the eighth and ninth abdominal segments, the posterior pair of which very soon become double. The six parts of the ovipositor of the female arise in this way, while, in the male, corresponding shorter projections are found. The ovipositor of the female Ichneumonidae and Cynipidae come under the same category, as well as the sting of Apis (Kraepelix, Dewitz, No. 103). Since the first rudiments of these paired appendages closely resemble the imaginal discs of the Dipterous larvae, they have repeatedly been regarded as abdominal limbs (p. 299, and footnote p. 300). The ovipositors of many Diptera and Coleoptera, on the contrary, as well as the penis of the Coleoptera, are to be derived from the most posterior abdominal segment, which is invaginated and telescopic. The male genital gland, in Phyllodromia, at first develops in a way similar to that described above in connection with the female organs. Only in later embryonic stages can sexual differentiation be recognised. It is then found that, in the male, four accumula- tions of genital cells become surrounded with an epithelium. These accumulations, which represent the four testicular follicles of Phyllo- dromia, are closely connected with the rudiment of the efferent genital ducts (vasa deferentia), and, in later stages, shift with the latter somewhat backward and away from the original genital rudi- ment. A remnant of the genital rudiment still remains attached to the terminal filament, and this, according to Haase, is the female GENITAL ORGANS. 351 part of the originally hermaphrodite genital rudiment, and may, in individual cases, even develop imperfect ovarian tubes and eggs. It has yet to be proved that the vestigial organs found in the adult male have originated from this remnant of the genital rudiment. Heymoxs concludes from the above observations that hermaphroditism was the original condition in the ancestors of the Insecta. This view, if correct, would account for the frequent occurrence of hermaphroditic conditions in adult Insects. In the female, the whole rudiment of the primitive efferent ducts is directly transformed into the oviduct. In the male, on the contrary, the whole length of this rudiment does not become trans- formed into the vas deferens, but its distal terminal portion degenerates and is replaced by a secondary terminal section, which then becomes connected with the ectodermal ductus ejaculatorius. If we pass in review the origin of the genital organs, as shown in the above description* of Phyllodromia, we have first of all to point out that, in the derivation of the genital cells from the epithelium of the coelomic sacs, we find direct agreement with the Annelida, f In the later development of the paired genital gland and of an efferent duct directly connected with it, there is a certain resemblance to the condition found in the Onychophora (p. 208). The dorsal position of the glands in these two Arthropodan groups is a special point of agreement. On the other hand, it should be pointed out that the genital gland of Peripatus, according to Sedgwick, arises by the direct fusion of the consecutive coelomic sacs (the Myriopoda have been placed in the same category by Heathcote, p. 252), and that therefore, in Peripatus, the genital cavity comes from the coelomic sacs. In the Insecta, on the contrary, the genital rudiment lies, indeed, in the wall of the coelomic sacs, but the lumen of the efferent duct here arises independently of the coelomic sacs, the cavities of the latter forming a small part of the permanent body- cavity. From this point of view, we shall have to regard the * The fact that this description not only applies to Phyllodromia, but is approximately applicable to many other Insects, perhaps to all, seems to be shown by the harmonious though fragmentary statements of Heider and AVheeleil with regard to the Coleoptera. t [It would seem highly probable, from recent researches on the origin of the germ-cells in various Invertebrata, that these cells, although not always to be distinguished by our limited vision from the general mesoderm, and especially from the coelomic epithelium, are in reality quite distinct from an early cleavage-period, and differ essentially from the mesoderm-cells in their nuclear organisation. This is rendered probable from the researches of Hertwig on Sagiita, Haecker on Cyclops, Boveri and others on Ascaris, WeismAnn on Ch ironomus. — Ed. ] 352 INSECTA. condition in Peripatus and the Myriopoda as the more primitive, following directly on that of the Annelida, while the condition of the Insecta is, on the contrary, modified. If we are to homologise the efferent genital ducts of the Insecta with those of Peripatus, we should have to trace them back to a pair of transformed nephridia. Their origin from the mesoderm in the Insecta would be in harmony with this ; but, in other respects, we find no features retained in the development of the efferent genital ducts in the Insecta which can be considered as supporting such a view. Mention should be made of Heymons' observations that, in the genital rudiment of Phyllodromia, the genital cells can from the first be distinguished from the epithelial cells. This statement is not favourable to the view, until now universally held, that the follicle-cells and egg-cells are derived from one- and the same sort of cell by later differentiation. Regarding their first origin, however, in Phyllodromia also, the two kinds of cells are to be traced to the same source. Special attention should be called to the fact that, in the Diptera and Aphidae, the genital rudiments can be recognised in a very early stage of embryonic life. This is certainly to some degree connected with the parthenogenetic and paedogenetic manner of reproduction, which is common in these two groups, and which (as in Moina, Vol. ii., pp. 123 and 180) leads to an early separation of the sexual rudiments. In the Diptera, the first rudiment of the genital glands is repre- sented by the "pole-cells." These cells (the "globules polaires " of Robin, also described by Weismanx in Chironomus and Musca),* which become separated at the posterior pole of the egg even before the formation of the blastoderm, were discovered by Leuckart and Metschnikoff (No. 55) in the asexually developing egg of the viviparous larva of Ceeidomyia (Fig. 174, pz). A rather large, highly granular cell (pz) here becomes severed from the posterior pole of the egg (Fig. 174 D) even before the blastoderm forms, and soon breaks up, first into two and then into four "pole-cells" (Fig. 174 F). When the blastoderm is completely formed, these "pole- * [It would be safer to discard Robin's term "globules polaires" for these cells, since that term was also applied by him and is still applied to the polar bodies or directive corpuscles, which latter, in all probability, have nothing whatever to do with the "pole-cells" of Chironomus, etc., which are simply the precociously separated germ -cells. This point, however, requires renewed investigation, and search should be made for the probable occurrence of true polar bodies and for the relationship of the " pole-cells " to the cleavage-nuclei. —Ed.] GENITAL ORGANS. 353 cells" first shift between the blastoderm-cells (Fig. 174 67), and then into the interior of the embryo, where, in later stages, they become arranged symmetrically into two groups, and, surrounded by cells of the neighbouring tissue, become transformed into the genital rudi- ment (Metschnikoff). In Chironomus (Fig. 175), two "pole-cells" become constricted from the posterior pole of the egg almost simultaneously (Balbiani), these, by division, yielding groups of four and eight cells. Just as in Cecidomyia, these cells are taken into the interior of the embryo (Fig. 175 C), and break up into two groups lying at Fig. 174. — First ontogenetic stages in the parthenogenetic egg of the Cecidomyia larva (after Metschnikoff). b, peripheral protoplasm ; 6/, blastoderm ; d, central food-yolk ; /, cleavage-nucleus; n, degenerating nutritive cells (the so-called corpus luteum); pz, "pole-cells." the sides of the proctodaeal invagination. In quite young larvae that have hatched from the egg, these two spindle-shaped groups, the cells of which soon increase in number, can be recognised at the sides of the dorsal vessel, surrounded by a distinct membrane which, anteriorly and posteriorly, is continued into a ligamentous structure. The anterior corresponds to a terminal filament, the so-called Midler's filament. It is inserted into the dorsal vessel, and was held by Schneider (Xo. 74) to be muscular, this author consequently tracing back the genital rudiment of the Diptera to a transformed fibre of the alary muscle of the heart, an assumption which was refuted by 2 a 351 IN SECT A. Balbiani. The posterior terminal filament is the rudiment of the paired efferent ducts of the genital gland. The division of the inner cells of the ovarian rudiment gives rise to rosette-like groups of cells, each of which corresponds to an ovarian tube. Eichter's more recent accounts (No. 71) agree with the above statements of Balbiani. In the Aphidae, just as in the Diptera, the first rudiment of the genital organs becomes distinct at a very early stage. Even in those early stages in which the first rudiment of the amniotic cavity forms by invagination at the posterior pole (p. 279), and before the formation of the lower layer, a cell-group (the genital rudiment) severs itself from the wall of this invagination, and, as a paired rounded mass, comes to B fe^D^^o^j lie in tne inside of tne embryo. This cell-group, according to Balbiani and "Witlaczil, is de- rived by division from a single cell. It becomes horseshoe - shaped, and breaks up into a number of rounded cell-accumu- lations, which become arranged in equal num- bers on each side of the median plane of the body and represent the rudi- ments of the terminal chambers of the ovarian tubes. These cell-masses are enveloped in an epithelial cover which, anteriorly, passes into the terminal filaments and, posteriorly, into the efferent ducts. The origin of this epithelial covering is doubtful. The efferent ducts of the different ovarian tubes fuse on each side to form a common oviduct, and this is continued into an unpaired ectodermal invagi- nation lying beneath the proctodaeum ; this invagination gives rise to the accessory genital organs (Metschnikoff, Witlaczil, Will). 00O2°o°0( )o°° a . VAfro ° * °OoO g ' Fir.. 175. — Three longitudinal sections through Chironomus embryos (after Ritter). In A, the "pole-cells" (pz) lie outside the developing blastoderm. In B, the pole-cells have pressed in between the blastoderm-cells. In C, they lie within the embryo, b, peripheral proto- plasm ; bl, blastoderm ; d, food-yolk ; k, nuclei of the blastomeres ; p, "pole-cells.'' METAMORPHOSIS. 355 II. Metamorphosis. 1. The Larval Forms.* The Thysanura and the Collembola emerge from the egg in a form resembling that of the adult, so that there can here be no question as to the occurrence of a metamorphosis ; they may consequently be described as true Ametabola (Insects -without metamorphosis). All other Insects, on the contrary, pass through a true metamor- phosis. When they leave the egg they are distinguished from the adult not only by their smaller size, but also by the absence of the wings. Many Insect larvae differ further from the adult {imago) in a number of ways. If we compare the young forms (larvae) of Insects when hatched with those of many Crustacea which leave the egg as Nauplii, we rind a marked distinction between them. In the Insects, the typical number of segments is developed in the embryo ; the limbs also, and the rudiments of the most important organs are already present. Only the wings are wanting. In other respects, the young emerging from the egg has the characteristics of a well-formed Insect. There is no doubt that the Thysanura, and among these the Gampodea •especially (Fig. 193), stand very near the fundamental form of this wingless larva. We have in the Thysanura undoubtedly the most primitive living representatives of the class of the Insecta. Lut we must not lose sight of the fact that, even amongst these forms, many systems of organs {e.g., the tracheal system) have undergone reduction, possibly on account of the small size of the body. The orders of the Insecta may be grouped in two divisions, according to the manner of their metamorphosis. To one of these groups belong those orders which Ave are accustomed to regard as the more primitive on account of their organisation ; among these we find some with an invaginated germ -band, thus suggesting a connection with the Myriopoda. The larva here passes gradually, by a series of stages each marked by an ecdysis, into the imaginal form. During these stages the rudiments of the wings grow out, increasing in size gradually. Metamorphosis thus here takes the form of onward growth within the limits of the segmentation and rudimentary organisation already present. Such development is distinguished as incomplete metamorphosis, and the Insects belonging to this type are known as the Homomorpha. * In the following account we have chiefly followed Lubbock (No. 156) and Brauer (No. 146). 356 IXSECTA. The metamorphosis of the second group, to which the higher orders of the Insecta belong, is more complicated. The larva here leaves the egg in a condition which often differs considerably from that of the imago, not only in form, but also in manner of life. This larval stage, which is characterised by the large amount of nutrition taken, and repeated moults, attains a considerable size, and finally changes into a resting or pupal stage. The faculty of loco- motion is now suppressed ; the pupa hardly moves and takes no nourishment ; all the animal processes step into the background, while the vegetative processes bring about the further (chiefly internal) changes in the body. The larval stage is thus followed by one which in many respects resembles the embryonic stages ; the pupal stage might be defined as a recurrence of embryonic development. A certain distinction between the two is, however, evident to the careful observer. In the embryonic stages, the organs develop chiefly from uniform rudiments, whereas they are here often built up by the concrescence of a number of disconnected formative centres, the so-called imagined discs. These imaginal discs must be regarded as persistent embryonic structures which have lasted throughout larval life in a latent condition, and in which the regenerative capacity of the embryonic rudiment has been retained. Those organs, on the contrary, which functioned in the larva undergo disintegration (p. 368). The pupal stage gives place, after another moult, to the stage of sexual maturity, the winged imago-stage, during which there is no further growth of the body. The Insects that develop according to this type are known as the Heteromorpha, and their metamorphosis is a complete metamorphosis* i.e., they belong to the Metabola or Holo-metaboJa. A. Homomorpha. The post-embryonic development of the Insects belonging to this type is in most cases a true metamorphosis, in so far as the young animal that emerges from the egg, although similar in other respects, is distinguished from the adult by the absence of wings (and of those abdominal appendages which are transformed into the external genital organs). In some cases also, alteration in the manner of life may be accompanied by changes in the form of the extremities (Cicada). The transformation into the perfect Insect is gradual. The last larval stage, with the rudiments of the wings already developed, is known as the nymph or pupa. In the Pediculina HOMOMORFHA. 357 and the Mallophaga alone, in consequence of parasitic life and the loss of the wings, metamorphosis is lost (acquired ametahole, Lang). The Insects belonging to this type may be divided into two groups, according to the manner of their metamorphosis. 1. Paurometabola. The post-embryonic development is accom- plished, through a series of moults, by the gradual growth of the body, of the wing-rudiments and the outer genital apparatus. This growth, though probably continuous, appears to be intermittent, being only visible at each moult. In Insects belonging to this group, the young stages resemble the adult, not only in form, but also in manner of life. The Dermaptera, Orthoptera genuina, Corrodentia, Thysanoptera, and most Rhynchota conform to this type. Fio. 176. — A, larva. B, pupa. C, imago of Cicada septemdicem (after Packard). The young forms of most Rhynchota resemble the imagines in the structure of the mouth-parts and in the shape of the body, and change gradually into the adult form. The genus Aleurodes is an exception ; its shield-shaped larva differs in appearance from the winged imago and passes into a resting, pupal stage, which is covered by the larval integument. There is here therefore complete metamorphosis. The same is the case in the male of the Coccidae, which changes into a resting pupa enclosed either in a protective larval integument or in a spun cocoon. The Cicadidae also (Fig. 176) attain a higher degree of metamorphosis. The larvae (A) live beneath the ground on the roots of trees, and are provided with hook-shaped fore-limbs adapted for digging. The nymph (B) is here capable of movement. Only shortly before the imago hatches does it remain quiet, while waiting for the bursting of the integument. 358 IXSECTA. 2. Hemimetabola. The young stage differs from the imago, not only in the absence of wings, but also in the presence of provisional (larval) organs. The larvae live in water and differ from the imago in the condition of their respiratory organs, the former possessing tracheal gills, either external or internal (intestinal respiration) in position. To this group belong the Ephemeri- dae, the Odonata, and the Plecoptera. The Ephemeridae represent a very primitive group. In them alone have the paired efferent ducts of the genital organs heen retained in their original form. The larvae (Fig. 177) live in water, ami leave the egg in a Campodei- form stage.* In the later stages, they are distinguished by the possession of external tracheal gills (k), which may be leaf-shaped or tufted, etc., and are usually attached to the postero-external margin of the terga of the seven anterior abdominal segments. Within these integumental outgrowths, richly-branched tracheal trunks extend, and here the exchange of gases with the sur- rounding medium takes place. In keeping with the aquatic manner of life, the stig- mata are closed, and the stigmatic tracheae found in the meso- and meta- thorax and in the eight anterior abdominal segments are merely thin strands cpiite devoid of air (dosed tracheal system). Only at the moment of hatching do the stigmatic tracheae and the stigmata open, so as to allow of the passage of the tracheal intima which is shed with the body-cuticle (Palmen, No. 161). The num- ber of moults which mark the successive stages through which the larva gradually approaches the imaginal form is very con- siderable (in Chloe over twenty, Lubbock). The last moult but one {i.e., the last larval or nymph-stage) in which imperfect * [There are two types of Insect larvae: the Campodeiform, so-called from its resemblance to the Thysanuran Campodea, and the eruciform. In the first we find the three typical regions of the body clearly defined, biting mouth-parts, ambulatory thoracic limbs, and sometimes terminal abdominal appendages. This larva is characteristic of the Ametabola and the Hemimetabola. In the second type the head is usually well-defined, but the body-segments are simple and cylindrical, and the animal has a vermiform aspect, the mouth-parts are usually adapted for biting, but may be much reduced, the thoracic limbs are usually present together with functional abdominal appendages, the "prolegs." The cruciform larva is very generally found among the Heteromorpha ; it attains its most characteristic development in the Lepidoptera, while in the Diptera it is much modified and degenerate, being apodal. — Ed.] Fig. 177 — Ephemeriil larva, k, tra- cheal gills ; t, principal trunks of the tracheal system. HETEROMORPHA. 359 wings are present, passes into a stage closely resembling the imago (sub-imago) ; this is distinguished from the preceding stages by the fact that, during it, no nourishment is taken. In this moult, the stigmata and stigmatic branches are definitely opened, and the tracheal gills, becoming constricted at their points of insertion, are cast off and remain in the empty skin (exuviae) of the last nymph-stage. One more moult on dry land leads from the sub-imago-stage to the form of the imago. The larvae of the Odonata are in some cases elongate and very like the imago, in others they are distinguished from it by their more compact form. AH Odonatan larvae are characterised by the remarkable modification of the lower lip which forms the protrusible " mask " (seizing pincer). The respiratory organs are variously modified in the different genera. The closable stigmata on the thorax and abdomen of the larva (Hagen) appear to be used chiefly for giving off air, but the older Libellulid larvae also breathe in air through the thoracic stigmata (Dewitz). The tracheal gills are internal in Aeschna and Libellula, being situated as folds or outgrowths on the walls of the rectum {intestinal gills) ; in the Agrionidac, three branchial leaves are found on the last abdominal segment. These, as well as the intestinal gills of the Libellulirfae (Hagen), are cast off when the larva passes over into the imaginal form. In Euphaea, which is distinguished by the presence of abdominal appendages, long, conical gills are found on either side of the body near the stigmata on the second abdominal segment to the eighth (Hagen). Plecoptera. In so far as, in the Perlidae, the tracheal tufts (Fig. 178 A, k) are retained in the imago, and the actual metamorphosis consists only of the gradual growing out of the wings, this family, strictly speaking, should be classed among the Paurometabola. The larvae are Campodeiform (Fig. 178 B) and their respiratory organs appear in various positions, as lateral gill-tufts (Fig. 17S A, k) at the sides of the thorax, as pro-sternal gills on the first ventral shield, at the sides of the anal aperture, or on the lateral margins of the abdomen. We are justified in classing the Perlidae among the Hemimetabola, by the fact that the branchial tracheal gills in the imagines do not function as such, but are retained in a shrivelled and vestigial condition. B. Heteromorpha. The larvae of the Insects that belong to this group are very different in appearance from the imagines. Some of them recall the Campodeiform larva, but they are often modified in adapta- tion to a definite manner of life, and frequently degenerate through the preponderance of the vegetative functions and the partial suppression of free locomotion in consequence of more or less parasitic life. The climax of degeneration is reached in the limb- less and eyeless "maggot" with reduced masticatory organs. In most cases the larva differs entirely from the adult in its manner of life. "We must regard complete metamorphosis as a more specialised form of the process of development, an acquired differ- entiation in larval life distinguishing the highly-developed, but probably, according to their origin, younger orders of Insects as contrasted with, the Homomorpha. 360 INSECTA. The last stage of larval life is always the pupal stage, which, in the form of the body, the development of the limbs, and the structure of the mouth-parts, resembles the imago. In this stage the Insects cease to take nourishment, and also, as a rule, the capacity of locomotion is lost (quiescent or resting pupa). The pupa is often enclosed in a cocoon spun by the larva. If the limbs of the pupa stand out freely from the surface of the body it is known as free- limbed (pupa libera, the exarate, incomplete, or sculptured pupa). In other cases the limbs which, in the resting pupa, are held closely pressed against the ventral side, become free imme- diately after the casting of the larval integument, but soon become glued to the sur- face of the body by the hardening of a tough secre- tion, so that their outlines are less distinct (Lepidop- tera and many Diptera). Such forms are known as mummy pupae (obtected, larcate, or signate pupa, chrysalis). Among the Diptera it often happens that the pupa remains surrounded by the last larval integument (barrel-like pupa, pupa coarctata). The number of moults undergone by an insect with complete metamorphosis is limited, and never attains that found among the Homomorpha (Ephemeridae). Neuroptera. The larvae of the Sialidac, which in appearance resemble many Coleopteran larvae, have mouth-parts adapted for biting, like those of the imago. The larvae of the Megaloptera, on the contrary, have their mouth-parts singularly transformed for sucking the juices of their prey ; the mandibles have a furrow on the lower side, and, with the maxillae, form on each side a sucking tuue. Some of these larvae are elongate and recall the Coleopteran larvae Fm. 17S. — .4, Perlid larva, lateral aspect (after Geaber). k, ^ill-tufts ; st, stig- mata. B, larva of Perla bicaudata (after Westwood). HETEROMORPHA. 361 {Mantispa, the remarkable form Sisyra which is parasitic on Spongilla), while others have a short and more compact body {Myrmclcon). We find here a rare phenomenon, the substitution in the imago of biting mouth-parts for the sucking mouth-parts of the larva, which also occurs in the Dytiscidae among the Coleoptera. The pupa is essentially quiescent and free-limbed ; in the Megaloptera it lies enclosed in a coarse-meshed, rounded cocoon. In some forms, however, immedi- ately before changing into the imaginal form, it becomes capable of locomotion and wanders about before casting the pupal skin. In this condition we find a transition to the metamorphosis of the Paurometabola with nymphs capable of movement. Panorpatae. The larvae are eruciform and live beneath moss or underground. The head is heart-shaped and the mouth-parts strong and adapted for biting. There may be eight pairs of ventral feet (on the first eight abdominal segments). At the end of the abdomen there is the rudiment of a pair of anal forceps which recalls that of the Forficulidae. These larvae are distinguished from similarly-shaped Lepidopteran and Hymenopteran larvae by the possession of compound eyes composed of closely crowded ocelli. Trichoptera. The Phryganeid larvae live chiefly in water in cases constructed by them out of foreign bodies (stones, parts of plants, snail-shells) ; these are in some cases attached to stones. In appearance (Fig. 179) they resemble the Coleopteran larvae. They have three pairs of long thoracic limbs, and at the end of the abdomen a pair of processes beset with hooks (h). At the sides of the abdomen (and of the meso- and meta-thorax) tracheal gills (k) are found in the form of tubes or tufts. The pupa is free-limbed ; the pupal stage is passed through within the larval case after another envelope has been spun within it. Before the imago emerges, the pupa becomes capable of locomotion, leaves the pupal envelope and creeps on to dry land, where the transformation takes place. Lepidoptera. The larvae here all agree in appearance and take the form of caterpillars (eruciform). Most of them live on land ; only a few Pyralidac spend their larval life in water. In these latter, tubular tracheal gills may develop (Paraponyx; Acentropus, Hydrocampa, and Cataclysta, on the contrary, are devoid of true tracheal gills). The most anterior of the thirteen externally recognisable rings of the body represents the cephalic complex of segments. It carries the usually three-jointed, short antennae and the biting mouth-parts. A line running in the median plane, and known as the fork-line, corresponds to the growth-suture of the cephalic lobes. On the two sides of the head are found six (less frequently five) ocelli arranged in a semicircle. The three thoracic rings which follow after the head resemble in form the abdominal rings. The first pair of stigmata belongs to the pro-thorax, the eight subsequent ones to the first eight abdominal segments. The limbs are rarely altogether wanting (Microptcryx among the Tincidae) ; in other cases they are vestigial Fig. 179.— Larva of Phry- ganea fusca (after Pictet). h, grasping hooks ; k, tra- cheal gills. 362 INSECT A. (a few burrowing caterpillars). Usually, three pairs of short, jointed thoracic limbs and five pairs of abdominal limbs are present. The latter are found on the third to the sixth abdominal segment, and on the terminal segment as the so-called prologs. They are truncated, and have a bilobate or circular sole beset with minute hooks. In Ncpticula, there are in all eighteen limbs. In other cases, the number is diminished by the reduction of the abdominal pairs, in some caterpillars, only two or three abdominal pairs of limbs being found (on the sixth to the ninth abdominal segment) besides the three thoracic pairs. The pupa is a mummy pupa (pupa obtccta), and is frequently enclosed in a cocoon. In a few Tincidae (especially in Microptcryx), the limbs are said to be partly free. The mouth-parts in their structure essentially resemble those of the imago. Diptera. The larvae of the Diptera must in general be regarded as essentially degenerate forms. The variation found in the different sub-groups is all the greater on this account. "We have here the best examples of the type of limbless, soft-skinned "maggots," whose body consists of a number of similar rings. Functional thoracic limbs are always wanting, and vestiges are only found on the first thoracic segment. In the same way, truncated ventral feet occasionally develop on the abdominal segments. The mouth-parts also are often quite vestigial. In most cases the integument is soft, but it may be of a firmer character (Slratiomys, in the larval skin of which, according -to Leydig, lime salts are deposited). The soft constitution of the integument may also extend to the cephalic segments (headless larvae) ; at this part, as a rule, however, a chitinous oesophageal framework develops, or a more or less marked maxillary capsule. But only in those cases in which this chitinous capsule contains the cephalic ganglia is it designated as the actual " head " (Brauer, encephalic larvae). The pupa is not always quiescent. In individual cases (Culicidae) it moves about in the water by the contraction of the abdomen. The resting pupa is often enveloped in the larval integument, and is then known as the barrel- shaped pupa. It is either free-limbed (x>upa libera), or, like the Lepidopteran pupae, is provided with limbs glued to the body (pupa obtccta, mummy pupa). The forms assumed by the Diptera larvae are utilised by Brauer (No. 100) for systematic purposes. He distinguishes two principal types, according to the manner in which the larval integument splits before the pupal stage is entered upon (or, in cases in which a barrel-like pupa is formed, when the imago hatches): (1) the Orthorhapha, in which, as a rule, a longitudinal slit opens on the back and a transverse slit at right angles to it ; (2) Cyclorliapha, in which the slit is a circular one, transverse to the long axis of the pupa, so that one or two caps are pushed off at the anterior pole. To the first type belong the encephalic larvae of the Culicidae and Chironomidac, and further the larvae of the Tipulidac, Cecidomyiidac, Stratiomyidae, etc., while the Muscidac, Syrphidae, and Pupipara belong to the second type. Great variety prevails with regard to the condition of the respiratory organs. Many larvae breathe only through the last pair of stigmata at the posterior end of the body which have remained open (mctapneustic respiration), in others a pair of anterior pro-thoracic stigmata remain open in addition to the posterior pair, while the others are closed (umphipncustic respiration); again, in other cases, some of the intermediate stigmata are also partly open (pcripneustic respiration). The pupae of many forms, on the other hand, breathe only through the most anterior pair of stigmata which occur on the pro-thorax (propncustic respiration). HETBROMORPHA. 3G3 Siphonaptera. The larva is limbless, has biting mouth-parts, and consists of a head and twelve more or less similar segments. It has ten pairs of stigmata on the three thoracic and the seven anterior abdominal segments. The pupa is free-limbed. The mouth-parts and the form of the body resemble those of the imago ; it lies in a cocoon. Coleoptera. Many Coleopteran larvae are Campodeiform. They have three well-developed pairs of limbs on the thoracic segments and, at the end of the abdomen, in many cases, there is a pair of filamentous or stylet-shaped appen- dages. More frequently a pair of truncated prolegs is found at the posterior end of the body. The head, which is always well developed, shows the fork- line mentioned in connection with the Lepidoptera, and carries antennae, which are usually short, and a variable number of ocelli (six or fewer) on each side ; these, however, are often wanting. The mouth-parts are adapted for biting, the mandibles are, in individual cases (Dyliscidac), changed into sucking organs. There are generally nine pairs of stigmata, the first of which occurs on the first or second thoracic segment, or on the boundary between the two, while the others belong to the first eight ab- dominal segments. The aquatic larvae {Dytiscus, Hydrophilus) are meta- pneustic, some having tracheal gills (Gyrinus). The body may be elon- gated as in the thread- like larva of the Elater- idae ; in other cases, it broadens out into a shape resembling the Isopoda (Pamidae). TheLamelli- corn larvae are eyeless, soft-skinned, and whitish in colour, and are further distinguished by the sac- like enlargement of the last ring of the body (Rhizotrogus). In forms which bore their way into wood or under the bark of trees, the limbs are vestigial or are altogether wanting {Buprcstidac, Cerambycidac). Such degenerate larval forms may finally become maggot-like (Curculionidae, Bostrychidae). The pupa is free-limbed, and resembles the imago in the form of the body and in the structure of the mouth-parts. A complicated metamorphosis, named by Fab re (No. 105) hyper-metamor- phosis, is undergone by the Meloidae in adaptation to the peculiar manner of life of the larva. The young is at first an active Campodeiform larva (Fig. 180 A), which attaches itself at the first opportunity to the male of Anthophora, and, during copulation, passes over to the female. As soon as the host has laid its eggs in the cell prepared for them in the earth and filled with honey, the Sitaris larva takes possession of the cell, devours the egg, and subsequently lives upon the honey. Here it moults and passes into a stage in which it can only move slightly, and is maggot -like with educed limbs (Fig. 180 B). It then changes into a pseudo-chrysalis (Fig. 180 C), a quiescent, pupa-like stage. From the pseudo-chrysalis there emerges first a larva resembling the Fig. 180. — Metamorphosis of Sitaris (after Fabre, from Lubbock). A, first larval stage. B, second larval stage. C, third larval stage (so-called pseudo-chrysalis) D, fourth larval stage. E, pupa. 364 INSECTA. second stage (Fig. ISO D), and then the actual pupa (Fig. 180 E), which changes into the imago. The freely moving as well as the resting stages are thus here multiplied. Hymenoptera. The larvae of the Hymenoptera belong to various types. The larvae of the Tcnthredinidae, which feed upon leaves, in appearance and colouring resemble the Lepidopteran larvae, and are therefore called false caterpillars (Fig. 181). They are distinguished from true caterpillars by the possession of a single ocellus on each side of the head, and by the unusually large number of abdominal limbs, the anterior pair belonging to the second, and not, as in the true caterpillars, to the third abdominal segment. There are generally six to eight pairs of abdominal appendages. An exception is afforded by the genus Lyda, in which, besides the thoracic limbs, there is only a pair of jointed appendages (cerci) at the posterior end of the body. These false caterpillars resemble the larvae of the Uroccridae, which bore into wood, but the latter are distinguished by the absence of eyes and of abdominal limbs. Most of the other Hymenoptera have degenerate larval forms in consequence of their peculiar and often parasitic or semi-parasitic manner of life. Whether the larva develops in vegetable outgrowths (galls), like many Cynipidae, or parasitically in other Insect larvae, like some Cynipidae, the Pieromalidae, Ichitcumonidae, etc., or whether it finds nutritive material in the cells con- structed and stored with food by the parent, or is fed during growth (Fossoria. Vespidae, Apidae, Formicidae) the passivity connected with its manner of life always brings about a reduc- tion of the limbs and of the mouth - parts, and an approximation to tin- general appearance of the maggot. In the larvae of Bees and AVasps, the enteron remains closed posteriorly, and does not communicate with the proctodaeum which receives the Malpighian vessels. The pupal stage is generally passed through within a spun cocoon. The pupa is free-limbed, and resembles the imago in structure, since, when the larva passes into the pupa, the limb-rudiments are only gradual!}' protruded from the imaginal disc (pp. 371-374), the pupal stage is preceded by a form showing the limbs only half protruded (Dewitz, No. 102), and it is this form that is known as the semi-pupa, sub-nymph, or pro-nymph. The eggs of the Ichncumonidac, Braconidae, and Ttcromalidac develop in the eggs or larvae of other Insects. The larvae of the Ichneumonidac are, as a rule, maggot-like. They may, however, possess at the posterior end of the body caudal appendages (Anomalon) or caudal vesicles (Microgastcr), which are lost on entering the pupal stage. The Pteromalidae, on the contrary, undergo a very remarkable metamorphosis. The ontogeny of these forms which has been described by Filippi, Metschnikoff, Ganix, Ayehs, and Lemoine, is char- acterised by the absence of nutritive yolk from the egg, by the absence or imperfect development of the embryonic envelopes, by the early hatching of the larva, and by the strange shapes of the larval forms. "We are still very much in the dark as to the first stages of development. In Plaiygastcr, a continuous process of division gives rise to numerous cells, some of which soon become arranged to form a superficial layer which surrounds the embryo in the form of an envelope (corresponding to the serosa). The other cells form the Fio. 181, — Eruciform larva of a Tenthridinid (Trichiosoma Iticorum, after Westwood). HETEROMORPHA. 365 rounded embryonic rudiment, in which an outer ectodermal layer and an inner (lower) layer can soon be distinguished. The embryo now lengthens and is soon divided by means of a groove which sinks in ventrally into an anterior widened cephalic section and a posterior narrowed section. The stomodaeum appears as an ectodermal invagination in the cephalic section, and soon becomes connected with the enteron which has developed from the inner cells. The proctodaeum arises considerably later, and does not communicate with the enteron until a very late stage. In the cephalic section (Fig. 182) a pair of grasping hooks (kf) develop at the sides of the mouth, and behind these a lower lip [ill). Another Fig. 1S2. — Stages in the development of Platygaster (after Ganin, from Lubbock). 23, 2h, 25, so-called Ctyctops-like larvae of three species of Platygaster ; 26, second larval stage ; 27, third larval stage, a, antenna ; ag, salivary duct ; ao, anus ; bsm, ventral ectodermal thickening ; ed, intestine ; ew, rectum ; /, furcal appendage ; fk, fat-body ; ga, genital organs ; gh, proctodaeum ; gsae, supra-oesophageal ganglion ; kf, hook-like feet ; Ifg, lateral limbs ; tin, sp, salivary glands; md, mandibles; mo, mouth ; ms7, stomach; dk, slkf, oesophagus; tr, tracheae ; id, lower lip. pair of limbs {Ifg) arise later at the posterior boundary of the anterior section, and a pair of short antennae (a) develop anteriorly. The posterior section of the embryo becomes divided up into several segments and runs out into a fork- shaped appendage (/) recalling the furca of the Copepoda. On this account the first larval stage that hatched from the embryonic integument after the harden- ing of the chitinous cuticle, was known as the Cyclops-like stage (Figs. 182 and 183). It appears that, in this stage, only the intestinal canal and the limb- muscles have been differentiated, while the other still undifferentiated organs 366 IXSECTA. lie as rudiments in a ventrally placed germ-hand, and only attain development in the next stage. Into this latter the Cyc/ ops-stage passes through a moult, and the larva is now an oval limbless body without segments (Fig. 182, 26). The nervous system, the salivary glands, and the proctodaeum now form as ectodermal invaginations, and the groups of muscles, by the arrangement of which the segmentation is recognisable, gradually develop. The last (third) larval stage, which follows the preceding after a moult, has the form of a segmented maggot devoid of limbs {27). The larval forms in other related genera seem to vary greatly. In Teleas there is also a Cyclops-stage, but it is preceded by a spindle-shaped larva which is more equally segmented and has small stump-like mouth-parts, while still devoid of grasping hooks (Ayers). Development here begins with the formation of a coeloblastula (Metschnikoff, Ayers), in the inner cavity of which a lower layer forms by the immigration of cells. The rise of a median groove marks the bilateral symmetry of the embryo and an anterior thickening dis- tinguishes the cephalic end. All these larval forms of the Pteromalidac must be regarded as highly specialised, but we are not in a position to determine in individual cases the ontogenetic significance of the development of these remarkable forms.* The larval forms of the Insecta are very varied. A comparative review of them shows most clearly that the manner of life of the larva is the chief factor in determining its outward appearance. We thus have, in the phytophagous larvae that feed on leaves, the eruciform type or poly pod caterpillar, in the forms that bore into wood, a similar type with powerful mouth-parts and strong cephalic capsule, but with degenerate limbs ; where the life of the larva is more or less parasitic, the form is that of a maggot, etc. In other groups (Orthoptera genuina), the larvae of which agree in their manner of life with the adult, the outward appearance of the imagines is to a large extent already foitnd in the larval forms. It is evident from these considerations that the metamorphosis of the Insecta can only to a limited extent be utilised for phylogenetic purposes. Above all, we must bear in mind that the larva which comes from the egg already shows the typical segmentation of the body, and that, therefore, in no single case are the ancestral forms which preceded the oldest Insect forms reproduced in the larvae. All that we can learn from the larvae of the Insecta is of value merely within the limits of this class. * [Kulagix (No. XXVIII.) has recently reinvestigated the development of Pla- Jygastcr with special regard to the origin of the germ-layers ; there is no yolk, and the total cleavage which occurs is regarded by this author as a modification of superficial cleavage. Henneguy (No. XII.) similarly finds total cleavage and one embryonic envelope in the nearly allied Chalcididae. A curious condition is found by Marchal (No. XXXIII.) in Encyrtus fuscicollis, a form closely related to I'latygastcr ; here the ovum gives rise, not to one egg, but to a legion of small morulae, which form chains of 50-100. — Ed.] DEVELOPMENT OF THE IMAGO. 367 The absence of wings in all insect larvae points back to the primitive nature of the Thysanura, and many insect larvae actually agree closely in appearance with the members of this genus. The Campocleiform larvae, the importance of which was specially pointed out by Brauer (No. 145), would thus represent the larval type which has most nearly retained the primitive character. As the chief characteristics of this type we would name : biting mouth- parts, jointed antennae, thoracic segments more or less resembling the abdominal segments, well developed thoracic limbs, a long, slender, ventrally compressed form of body, and two jointed pro- cesses (cerci) at the end of the abdomen. This type is fairly accurately adhered to by the larvae of the Ephemeridae, Perlidae, many Xeuroptera, and many Coleoptera. The metamorphosis of the Insecta, as a rule, is more sharply marked in the higher orders, the separate stages being more unlike one another, and the transition between them not being gradual. "We must therefore regard incomplete metamorphosis as the more primitive condition, and complete metamorphosis as a higher grade of individual development acquired in the Insecta. Consequently, the larvae of the Metabola must all be considered as derived forms. But in the Hemimetabola also, certain characters, phyletically con- sidered, must "also be regarded as new acquisitions, e.g., the presence of a so-called closed tracheal system and of tracheal gills in many aquatic larvae, since, in all probability, this manner of life must be considered as only recently adopted. While, therefore, little importance attaches, phyletically, to the larval forms of Insects, certain features are perhaps of some value, in so far as the acquired larval forms also show a tendency to repro- duce the morphological characters of the ancestral forms. Among the features which have thus come to the surface again, we have : (1) the softer nature of the integument of the body-surface; (2) the less marked separation of the thorax from the abdomen ; (3) the more uniform segmentation of the extremities; (4) the absence of the facet-eyes ; (5) the frequent occurrence of abdominal limbs. 2. Development of the Imago. We have already (p. 355) pointed out the characteristic distinction existing between the homomorphous orders of Insects on the one hand, and the holometabolic forms on the other with regard to the manner of development of the sexually mature (imaginal) condition. In the first group, the development of the adult is accomplished 368 INSECTA. through a series of gradual internal and external changes, not differ- ing essentially from the ontogenetic processes which occur in the metamorphosis of most other animals. We can here trace back the development of the wing-rudiments, the external genital apparatus, and all other alterations of form to simple growth of the larval body. The transformation of the internal organs, chief among which are the genital organs, is accomplished in an equally simple manner. We may perhaps assume, although this point has not yet been thoroughly investigated, that, simultaneously with the growth of the internal organs, a gradual regeneration takes place in them, as, indeed, is frequently the case with functional organs. We may assume that some of the cells or cell-groups, exhausted through the performance of their vital functions, are absorbed and replaced by fresh elements, so that a constant gradual regeneration of these organs is in progress. In the holometabolic orders of Insects, on the contrary, the tran- sition from the last larval stage to the adult form is accomplished through the intercalation of a resting stage {pupal stage), in which the acts of feeding, and usually also of locomotioD, are suppressed, while the whole life-activities of the organism seem directed towards the important and complicated ontogenetic processes, which involve a complete destruction of many of the organs of the larva and their renewal from certain rudiments {imaginal discs) already present in the latter. Only a few of the organs found in the larva pass directly over into those of the pupa and the imago. Among these we must reckon the rudiments of the genital system. The heart also and the central part of the nervous system undergo only slight changes. Most of the other organs of the larva, on the contrary (the hypo- dermis, most of the muscles, the whole of the intestinal canal, and the salivary glands) are completely destroyed. Their cells, under the influence of the blood-corpuscles (leucocytes), which here act as phagocytes, break up into pieces which are taken in and digested by the latter, while, simultaneously with these processes of disinte- gration, the reconstruction of the organs from the formative centres (imaginal discs) already present in the embryo is accomplished in such a way as, in most cases, to preserve the continuity of the organ. We shall only be able to understand these processes by regarding them as an extreme case of the regenerative processes, which we assumed must occur in the Homomorpha. We shall then have to assume that at first only a part of the rudiment of an organ develops and functions for the use of the larva ; this part becomes exhausted DEVELOPMENT OP THE IMAGO. 369 during larval life, so that it is no longer capable of performing its functions and therefore disintegrates, while another part of the rudiment remains from the first in an undeveloped condition, persisting as the imaginal disc, in order, during the pupal stage, to undertake the regeneration of the organ. It should here be pointed out that this remarkable method of development of the organs of the imago, although most marked in the Insecta, is also found indicated in other animal groups. We find repeatedly that, instead of the gradual transformation of a larval organ into the corresponding adult organ, another course is entered upon, the larval organ being destroyed or degenerating, and the corresponding organ in the adult appearing anew as a rudiment. We refer here to Vol. ii., p. 312, where the disappearance and reappearance of limbs during the metamorphosis of the Crustacea is described. A similar phenomenon was mentioned in connection with the Acarina (pp. 104 and 105), in which a partial destruction and a reconstruction of the internal organs occurs. Where the distinction between the larval and the imaginal form of an organ is very marked, the latter mode of development may even appear as a simplification of the ontogenetic process. Although Swammerdam had already shown that the limb-rudiments can be recognised under the integument of the larvae of holometa- bolic insects, our more detailed knowledge of the changes connected with the pupal stage is due to the researches of Weismann (No. 129) in connection with the ontogeny of the Diptera. The fact that later students of this subject, among whom should be named Ganin, Viallanes, Kunckel d'Herculais, Kowalevsky (No. 112), and Van Kees (Xo. 121), restricted their investigations to the same order, accounts for our being most familiar with the processes of meta- morphosis in the pupa of the Dipteran family Muscidae. Our description will therefore chiefly refer to this family. But since, as is easily seen, we have in the Muscidae the most complicated and the most modified ontogenetic conditions, we shall often have to take as starting-points the simpler formative processes found in other Holometabola, such as the Nematocera (Corethra), the Hymenoptera, and the Lepidoptera (Weismann, Ganin, Dewitz, etc.). It should be mentioned that our knowledge of the subject is still very incomplete, and only the principal features can be regarded as established. We have, especially, no knowledge as to how far the conditions of the inner metamorphosis ascertained as prevailing in the Muscidae occur also in the other groups of Insecta, although it must be regarded as probable that similar processes take place in the pupae of the Lepidoptera, Hymenoptera, and perhaps also of the Coleoptera. We shall consider these ontogenetic processes under two heads, 2 B 370 INSECTA. discussing first the development of the external form of the body, and then the rise of the internal organs of the imago. A. Development of the external form of the Body. The external form of the imaginal body is already complete in rudiment in the pupa, so that the passage of the pupa into the imago takes place merely by an unfolding and hardening of parts already present. The form of the body of the imago must therefore be prepared in the last larval stages, and attains complete develop- ment at the pupal moult (the transition to the pupal stage). In most cases the transition from the larval shape to that of the imago consists principally of a modification of parts already present, new rudiments participating in it only to a limited extent. In the Lepidopteran caterpillar, for example, the head, together with the antennae and mouth-parts, and, further, the thoracic limbs (though in an essentially modified form) pass over direct from the larva to the pupa. The compound eyes and the wing-rudiments arise as new rudiments. The latter appear on the meso- and meta-thorax of the larva in the form of imaginal discs (icing-discs). The same is the case in very many other Holometabola, in which the transition from the larva to the pupa rests essentially upon a transformation of parts already present. We ought here further to mention modifications which occur in the abdomen, and which consist partly of the growth of the abdominal rudiments (extremities'?) into external genital organs (ovipositors, poisonous stings, pp. 299 and 300), and partly of an apparent diminution in the number of segments. The latter may be brought about by a fusion of distinct segments, or by a union of the first abdominal segment with the meta-thorax (Hymenoptera), or else may be traced to a transformation of the most posterior segments into a telescopic genital appendage (ovipositor, penis). In those cases in which the larva is limbless (Diptera, many Hymenoptera, and Coleoptera), the limbs of the imago also arise as new formations in the form of imaginal discs (limb-discs). The metamorphosis of Corethra (Weismann, No. 130) may serve as an example of the simpler type of metamorphosis. The larva belongs to the eucephalic type of Dipteran larvae, and consequently the head of the adult is present as a rudiment in this stage; this larval head, through certain modifications of its parts, passes directly over into the pupa. Even the compound eye is already found in the larva, a rare and exceptional occurrence among the Holometabola. The thoracic limbs, the wings and the halteres, on the contrary, are DEVELOPMENT OF THE EXTERNAL FORM OF THE IMAGO. 371 developed as entirely new rudiments. We consequently find, in the last larval stage that precedes the pupal stage, correspondingly arranged imagined discs. Each thoracic segment shows four such discs, two ventral and two dorsal (Fig. 183, ba and fa). The ventral discs (ba) become the limb-rudiments. Of the dorsal pairs of discs {fa), that occurring in the meso-thorax becomes changed into wings, that in the meta-thorax into the halteres, while the corresponding rudiment in the pro-thorax yields, in Corethra, the stigma-bearing dorsal process and, in Simidia, a tuft of tracheal gills. If we examine such a rudiment (imaginal disc) of a limb more closely, we see that the limb itself as elsewhere (e.g., in the Hemimetabola) arises as an outgrowth of the surface of the body. The only dis- tinction here found is that the limb- rudiment, as a whole, appears sunk below the level of the surface of the body. It arises at the base of an invagination, in the same way as the head- and trunk-discs in the Pilidium larva of the Nemertini {Vol. i , p. 221), and the rudiment of the lower surface of the Echinoid body in the Pluteus (Vol. i., p. 439). Such instances of the occurrence of rudiments of important parts of the adult body in an invaginated condition might easily be multiplied. For instance, the body-wall of the primary zooecium of the ecto- proctous Bryozoa is found invaginated in the larva (as the sucking-disc and mantle-cavity). The lumen of the invagination in which the limbs of Corethra (and of other Holometabola) appear as rudiments was called by Van Rees the jperipodal cavity, and the sheath which bounds it externally, and which naturally is continuous with the hypodermis, was named the peripodal membrane. We must assume that, from the very first, an ectodermal and a mesodermal -portion derived from the corresponding layers of the germ-band, take part in the rudiments of the limbs. The ectoderm of these rudiments is in continuous connection with the peripodal membrane, and through it with the hypodermis •covering the larval body. Weismann was inclined to derive the organs that develop within the limb-rudiments (tracheae, muscles, etc.) from growths of the neurilemma of a nerve joining the imaginal disc. For nerves and tracheal ramifications appear early on the inner surface of the imaginal discs. Fig. 183.— Diagrammatic trans- verse section through a thoracic segment of a larva of Corethra (from Lang's Text-book), ba, limb-rudiment ; fa, wing-rudi- ment ; be and fe, peripodal depressions ; Ihy, larval hypo- dermis ; Ih, chitinous cuticle of the larva. 372 INSECTA. When the limb-rudiments increase in size, the peripodal membrane is correspondingly stretched, while the appendage within it assumes a more or less bent position. In consequence of this, the wing- rudiments seem folded, and the rudiments of the legs in Corethra are spirally coiled. The unfolding of the limb-rudiments is brought about by their protrusion from the invagination in which they were at first hidden. As they become more and more protruded, the peripodal depression becomes continually shallower, and finally the peripodal membrane becomes completely evaginated and forms a part of the general hypodermis. The internal organs of Corethra, as compared with those of other Holometa- bola, seem during metamorphosis to undergo only unimportant alterations. Of the striking processes of disintegration and of subsequent regeneration, which have been so well established for the Muscidae, nothing is to be observed in Corethra. It deserves to be mentioned, however, that, according to Kowalevsky (No. 112), a disintegration of the larval and the development of the imaginal epithelium of the enteron of Corethra takes place in the same way as in Musca (see below, p. 383). Most of the larval organs pass directly over into the pupal and imaginal stages ; the general musculature also remains unaltered, but the muscles of the limbs and of the wings oome from new rudiments. The latter, according to Weismann, arise in the last larval stage from cell-strands which appeared as rudiments in the embryo. When we consider the unimportant character of the internal changes which occur during the metamorphosis of the Tipulidae, of which Corethra serves as- an example, we shall hardly doubt that the conditions here found represent a transition from the incomplete to the complete method of metamorphosis. This is confirmed inter alia by the short duration of the pupal stage and its- capacity for free movement, as also by the early appearance of the compound eye, a character which Corethra has in common with the Hemimetabola. We must now describe more in detail the development of the wings, which has been best ascertained in the Lepidoptera by Semper (No. 126), Landois- (No. 114), Pankeitius (No. 120), and Schaffer (No. 124a). The wings, like the other rudiments of extremities, arise as simple outgrowths of the hypo- dermis within a peripodal depression. They thus at first represent a simple fold of the hypodermis, the point of insertion of this fold being connected internally witli peculiar modifications of the fat-body and of the tracheal system. The fat-body at this point shows accumulations of small cells, which have been regarded by Schaffer as formative centres. The tracheae which join the wing-rudiments form a close network of very fine tracheal tubes, which develop as intracellular structures within single large matrix-cells (Landois, Schaffer). These networks of tracheae degenerate after the pupal stage has been entered upon. On the other hand, large tracheal ramifications develop which run into the wings and lead to the development of their venation. When the caterpillar passes into the pupa, the wing-rudiments are evaginated from the peripodal cavity by the action of increased blood-pressure. The wing-rudiments thus become vesicles filled with blood which contain tracheal ramifications. At a later stage, however, the layers corresponding to the future upper and lower surfaces of the wings become closely applied and fuse, except along those lines DEVELOPMENT OF THE EXTERNAL FORM OF THE IMAGO. .'573 that are occupied by the tracheal ramifications ; here also the blood-thud circulates, and these lines become transformed later into the network of veins in the wing. In later stages, tracheae are no longer to be found within the veins ; they have either degenerated or, as in Musca, according to WeismANN, they have been withdrawn out of the veins into the thorax. There, however, remains in the veins a strand, which was discovered by Semper in the Lepi- doptera, and which in early stages accompanies the tracheae ; this we may call the rib-strand ( Semper' s wing-ribs). It resembles a tracheal tube, and consists of an outer matrix and an inner intima which gives off projecting dendriform processes into the lumen. The centre of the lumen is occupied by a longitudi- nally striated strand (a secretion ?). Semper was able to prove the connection of these rib-strands, which, in the adult, are only to be found in the basal half of the wing, which they serve to support, with the tracheal system. These strands must therefore be transformed tracheae. Nerve-trunks are also found in the wing-veins. The cuticle of the wing, which does not develop until somewhat late, is considerably thickened on the surface of the veins. The manner in which the two hypodermal lamellae of the wings fuse is of some interest. A " basal membrane " develops on the inner surface of the hypodermis on each side, while the hypodermal cells themselves become pillar-like. The two basal mem- branes become closely applied to one another, fuse, and finally disintegrate, so that, in the adult wing, the hypodermal pillars extend continuously through- out the whole thickness of the wing. It should here be mentioned that the facts of ontogeny are not favourable to Adolph's theory of wing-venation. According to this theory, the veins of the fully formed wing are to be divided into convex and concave veins, which differ in their origin, the concave veins being derived from tracheae, while the convex veins develop out of cell-strands into which tracheae can extend only secondarily. The system of convex veins and that of concave veins are origin- ally altogether distinct. But it has been proved by Brauer and Redtenbacher (No. 101) for the wings of the Odonata, and by Grassi for those of the Termites, and more recently by Haase (No. 108) for those of the Lepidoptera, that the branches of one and the same tracheal trunk may be changed partly into convex and partly into concave veins, so that the postulate on which the theory rests is negatived. This theory is also opposed by Van Bemmelen (No. 99), who confirms the observation made by Fr. Muller in connection with the Nymphalidae, that the system of veins in the Lepidoptera immediately after entering the pupal stage differs in details from that of the adult form. The observation of the development of the venation in the wings has thus a certain phylogenetic significance. The hairs and scales of the Lepidopteran wing arise as outgrowths of single hypodermis-cells (mother-cells of hairs and scales, Semper). The characteristic definitive markings develop only after the differentiation of the scales. It must, however, be mentioned that, according to Van Bemmelen, the permanent markings are preceded by transitory markings, which are essentially distinct from the former, but have a few features in common with them. Much more complicated ontogenetic processes are met with in the Muscidae. The limb- and wing-rudiments here arise in the same way as in Corethra. But, in the Muscidae, the whole of the imaginal rudiment is shifted much further into the interior of 374 IXSECTA. the body, the peripodal cavity appears closed and the peripodal membrane is connected with the hypodermis merely by a delicate thread-like stalk (Fig. 184 A, is; Fig. 185 A, st). These connective stalks, which were recognised by Dewitz* (No. 102), who correctly grasped their significance, have a fine lumen, as was shown by Van Rees (No. 121), who has recently studied these structures more closely. Although the first development of the imaginal discs in the embryo of the Muscidae is still unknown, Ave shall not err in tracing them back, like those of Corethra, to hypodermal invaginations. We must then regard the stalk-like connection as the loner drawn out neck of this invagination. In other respects the development of the ex- tremities (Fig. 184) takes just the same course as in Corethra. The rudiments of the legs increase in size and early show traces of the later segmentation. They appear packed into the peripodal cavity in such a way that the differ- ent joints of the limbs telescope into each other " like the rings of a traveller's drinking cup " (as Van Rees appropriately expresses it). The evagination of the developed limb-rudiments, which occurs on the first day after the commence- ment of the pupal condition, takes place by the shortening of the stalk of the imaginal disc (Figs. 184 B, and 185 B) and the widen- ing of its lumen, which allows the extremity, as in Corethra, to emerge finally through the widening aperture of the peripodal invagination (Figs. 184 C, and 186 A). While the latter at the same time gradually disappears, the peripodial membrane is utilised * Kunckel d'Herculais (No. 113) also recognised these strands. Fio. 1S4. — Diagrammatic transverse section through the larva and pupa of Musca, to illustrate the development of the wings, the legs, and the imaginal hypodermis (from Lang's Text-book), b, limb- rudiments ; ft., wing-rudiments ; ihy, imaginal hypo- dermis, in D extending from the base of the imaginal discs ; iid, imaginal discs of the wings ; H-v, imaginal discs of the legs; is, strand connecting the rudiment with the hypodermis; IK, chitinous integument of the larva ; Ihy, larval hypodermis (indicated by two thin parallel outlines, while the imaginal hypodermis is represented by thick black lines). DEVELOPMENT OF THE EXTERNAL FORM OF THE BODY 375 for the formation of a hypodermal thickening in proximity to the point of insertion of the limb, and this thickened part of the hypodermis gives rise, as Ave shall see below (p. 379), to the forma- tion of the whole imaginal thorax, while the hypodermis of the larva disintegrates. We must here touch upon the question of the first appearance of the meso- dermal portion of the rudiments of the extremities. In the imaginal discs of developing larvae of Musca, a separation into an outer ectodermal and an inner mesodermal portion is always found. Gaxin (No. 107) derives the mesodermal portion from a differentiation and delamination of the innermost layers of the - as Fig. 1S5. — Diagrams illustrating the position of the imaginal discs in the larva (.-I) and pupa (B)of Musca (taken from Van Rees). The wing-rudiments are omitted, as, optic disc ; at, antennal rudiment; 61, &'■*, W, rudiments of the three thoracic limbs; bg, ventral chain of ganglia; g, brain; h, so-called "brain-appendage"; m, peripodal membrane; o, aperture of the brain-appendage into the pharynx; oe, oesophagus ; p, so-called "pharynx" ; r, rudiment of the proboscis ; ss, frontal disc ; st, stalk-like connection between the peripodal membrane and the hypodermis ; /, II, III, the three thoracic segments. ectodermal portion, and Van Rees has supported this view. Kowalevsky (No. 107), on the contrary, tends to the view that the mesodermal part of the imaginal discs is to he derived from the cells of the embryonic mesoderm. He finds, scattered in the mesoderm, beneath the hypodermis of the larva, so-called wandering cells (Fig. 190 A, vj, p. 383), which differ in appearance from the leucocytes, and which represent the elements from which the formation of the mesodermal part of the imaginal rudiments proceeds. Kowalevsky is disposed to assume an imaginal rudiment for every segment, this rudiment being, however, so delicate and undifferentiated as not to be discoverable in the first stages. From these imaginal rudiments of the mesoderm, the above-mentioned wandering cells would be derived, and would only secondarily become connected with the imaginal discs. 376 INSECTA. The development of the cephalic region in the Mwcidae is more complicated, and still, in spite of the descriptions of Weismann (No. 129), Van Rees (No. 121), and Iyowalevsky (No. 112), difficult to understand. In this connection we must recall the fact that the cephalic region, in the larva of the Muscidae, occurs in an extremely- reduced condition, the head-region being represented only by the most anterior and smallest of the twelve segments of which the body of the conical larva is composed. Its small size is partly to be ascribed to the fact that a considerable portion of the head is here present in an invaginated condition. For, as has been shown by the researches of Weismann, the anterior part of the head, the mandibles, and the whole of the region surrounding the mouth are invaginated in the last embryonic stages, and in the fully-formed maggot are represented by that depression (Fig. 185, p) in which the hook- apparatus, characteristic of the larvae of the Muscidae, develops. This invaginated part of the head into the base of which the oesophagus now opens, has been named, not very happily, the oesophageal bulb or pharynx, and it must for the present be held that the cavity thus named does not belong to the alimentary canal. It is an invaginated section of the head, and the formation of the imaginal head consists for the greater part merely in the evagination of this region. The first rudiments of the most important parts of the head (the eyes, the antennae, and the frontal region) are found in the youngest larvae as a pair of cell-masses lying in the thorax, closely applied to the halves of the brain (and therefore called by Weismann brain- ap>pendages). These are, probably from their first origin, connected anteriorly with the pharynx, and might be described as the imaginal discs of the head. In the later stages these assume the form of a pair of long sacs expanding posteriorly (Fig. 185 A and B, h), and may no doubt, according to their origin, be regarded as outgrowths of the larval pharynx (see above). Epithelial thickenings soon appear in the walls of these sac-like "brain-appendages," and in these can be recognised the rudiments of definite parts of the head. A disc-Jike thickening appears in the posterior widened portion of each of the appendages, this represents the rudiment of the com- pound eye, and is consequently called the optic disc (as). On the basal surface of the optic disc there is a cellular expansion connected with the supra-oesophageal ganglion by a nerve. This nerve becomes the optic nerve of the adult, while the optic ganglion becomes more distinctly separated from the brain. In the anterior, more cylindrical DEVELOPMENT OF THE EXTERNAL FORM OF THE IMAGO. 377 or tubular portion of the " brain appendages " we find the frontal disc (ss), from which the antennal rudiment soon grows out in just the same way as do the limb-rudiments from the base of the imaginal disc that gives origin to them. Originally (Fig. 185 ^4) the "brain appendages" lie somewhat far back, in the thorax of the larva, so that they connect the posterior part of the wall of the pharynx with the most anterior segment of the brain. Later, however, after the pupal stage has been entered upon, they, together with the central nervous system, shift further forward (Fig. 185 B), so that their anterior extremities, which are Fig. 186.— Diagram illustrating the transformations that take place in the pupa of Mvsca before it hatches (adapted from Kowalevsky and Van Rees). The wing-rudiments are not drawn, as, optic disc ; at, antennal rudiment ; b1, b-, b3, rudiments of the three thoracic limbs ; bg, ventral chain of ganglia ; g, brain ; k, cephalic vesicle (formed by the union of the pharynx with the brain appendage) ; oe, oesophagus ; r, rudiment of proboscis ; ss, frontal disc ; /, II, III, the three thoracic segments. now bent ventrally, embrace the pharynx laterally (if we have rightly understood the descriptions of Weismann and Van Eees). At the same time the communication between the "brain-appendages" and the pharynx (Fig. 185 B, o) becomes wider and wider, and soon extends in the form of lateral oesophageal slits along the whole length of the brain-appendages. By this means, the lumina of the brain-appendages and of the pharynx flow together so completely that the two soon represent only one single vesicle, the cephalic vesicle (Fig. 186, k). The walls of the cephalic vesicle are nothing 378 INSECTA. more than the external surface of the permanent head, and the most important parts of the latter (antennae, eyes, rudiments of the proboscis) can already he recognised on them. It only remains for the cephalic vesicle to he evaginated through the aperture of the pharynx ( + to + ), so as to produce the completed head of the pupa. By this evagination of the parts formerly invaginated, that which was before the aperture of the pharynx becomes the neck- region (Fig. 186 B, + to + ), which now connects the head with the thorax (Van Rees). The protrusion of the " cephalic vesicle," which was directly observed by Weismann, seems to be caused by the increase of pressure from within, brought about by the con- traction of the posterior parts of the body. In accordance with the conformation of the imaginal head thus produced, the anterior end of the oesophagus becomes ventrally flexed. We have pointed out above that the so-called pharynx is nothing more than an invaginated portion of the external surface of the larval head. The " brain-appendages " must be regarded as diverticula of this invagination, in which the separate parts of the head appear as rudiments in an invaginated condition. They may thus be compared throughout with the rudiments of the thoracic limbs. All these "imaginal discs," according to their origin, have to be derived from invaginated portions of the external surface of the body. It is difficult to reconcile with this hypothesis the accounts of Graber (No. 28), who in a later embryonic stage of Galliphora [Musca of most authors) observed the rudiments of the imaginal discs lying as simple epithelial plates in the interior of the body. Since Graber left uninvestigated the preceding and the subsequent ontogenetic stages, we can only record his statement, and must leave the problem to be solved by future researches. B. Development of the Internal Organs of the Imago. It has already been mentioned that most of the organs of the larvae of the Muscidae (and of most Diptera, Lepidoptera, Coleoptera, a and Hymenoptera) undergo disintegration through the action of the blood-corpuscles (leucocytes), and that their reconstruction proceeds from certain embryonic cell-groups, the imaginal discs. Disin- tegration and reconstruction take place during the pupal stage so gradually that in Fig. 187.— Diagram illustrating the formation of """a*-- & j the imaginal hypodermis in the abdomen of the many cases the continuity Muscidue (from Lang's Text-book), hi, imaginal . . . ,. , , , discs of the hypodermis ; Ui, larval hypodermis. 01 the Organ IS not disturbed HYPODEBMIS. 379 during the course of these processes. Such transformation affects especially the hypodermis, the intestinal canal, the muscles, the fat- body, and the salivary glands. The modification of the tracheal system can only to some extent be classed in this category ; other- wise it appears to be due to simple regeneration through division of the cells. The heart, the central nervous system, and the genital rudiments undergo slighter alteration. The changes that take place in the different organs must now be discussed separately. Hypodermis. The hypodermis of the imago arises through an extension of the ectodermal portion of the imaginal discs. This has already been stated in connection with the thorax (p. 375). While the limbs of the thorax gradually attain development in the pupa, a hypo- dermal layer, consisting of numerous small cells, extends from their points of insertion ; this layer, which apparently arises from the peripodal membrane, spreads more and more over the surface of the pupal thorax, while at the same time the area of the large-celled larval hypodermis is correspondingly more and more circumscribed. The flat edges of the newly -formed hypodermis (Figs. 187, hi and 188, i) grow into the slit between the superficial cuticle and the larval hypodermis (Fig. 188, h), so that, at these points, the old hypodermis which is undergoing disintegration comes to lie on the inner side of the newly-formed epithelial layer (Fig. 188 B). It is thus evident that, during the substitution of the new for the old hypodermis, the continuity of the superficial epithelium does not anywhere appear interrupted. Since the edges of these two hypo- dermal envelopes overlap, there is not anywhere a point of the body-surface devoid of epithelium. The disintegration of the larval hypodermis is accomplished by the action of the leucocytes (Fig. 188, h), which become massed in the neighbourhood of the dis- integrating hypodermal cells and ingest the latter. Since the assimilated fragments assume the shape of rounded granules, the leucocytes may now be distinguished by the name of the granular spheres (Weismaxn). The granular spheres, which abound in the future body-cavity of the pupal stage, are therefore nothing more than leucocytes (blood-corpuscles) which have assimilated the dis- organised tissue of the disintegrating larval body. It should here be noted that the breaking-up of the larval tissues is not preceded by the death of the cells, but is the result of the action of leucocytes on the still living tissues which have lost their active functions. While 380 IXSECTA. tissues fully capable of vital activity, e.g., those of the imaginal discs, resist the attack of the leucocytes, the larval tissues, less capable of vital activity, are broken up into fragments by the attack of the leucocytes, and are simply devoured and digested by them. These processes may best be followed in the disintegration of the larval musculature. The destruction of most larval organs is thus due to the capacity for the taking in of nourishment and for intra- cellular digestion possessed by the amoeboid blood-corpuscles. This capacity has been specially emphasised by Metschnikoff (Nos. 116 and 117), who with reference to this significance of the blood- corpuscles has called them phagocytes. Fig. 188. — Sections through a hypodermal imaginal disc in the abdomen of Musea (after Kowalevsky). A, through the larva. B and C, through the pupa, h, larval hypodermis ; V, detached portions of the same attacked by phagocytes ; i, imaginal disc ; k, phagocytes containing disorganised hypodermal cells (so-called granular cells); k', phagocytes enclosing hypodermal nuclei ; m, mesodermal rudiment of th>' imaginal disc ; w, wandering cells. The re-formation of the hypodermis is accomplished in the head and in the abdomen in the same way as in the thorax. In each of the eight segments of which the abdomen of the larva consists, there are, according to Ganin (No. 107), four islands of small cells; these are the imaginal discs (Figs. 187, hi, and 188, i) from which the re-formation of the hypodermis proceeds. Van Rees has recently discovered another pair of small imaginal discs on the abdominal segments. The four discs that occur on the last body-segment, MUSCULATURE. 381 closely crowded together, encircle the anal aperture (Fig. 189, tins) and take part in the formation of the proctodaeum, yielding the rudiments of the rectal sac and the rectal papillae. To this segment apparently also belong the two pairs of imaginal genital rudiments (rudiments of the external genital organs) which were demonstrated by Kunckel d'Herculais (No. 113) in Volucella. It should be mentioned that a cell-accumulation representing the permanent mesoderm (Fig. 188 C, m) is found on the inner surface of the abdominal imaginal discs as on that of the thoracic discs, this accumulation being the starting-point for the development of the mesodermal structures of the abdomen. Kowalevsky, as above mentioned (p. 375), has traced back the origin of this mesoderm-accumulation to the so-called wandering cells (Fig. 188 A, w), while earlier authors were inclined to derive them through delamination from the ectoderm of the imaginal discs. The newly-formed hypodermis extends very rapidly over the surface of the body, so that the areas of bypodermis originating from the different imaginal discs soon flow together. While this perfecting of the permanent hypodermis is taking place, that of the larva is finally destroyed by the phagocytes. Musculature. The greater part (or the whole mass) of the larval musculature undergoes a process of disintegration by means of phagocytes precisely like that described above in connection with the larval hypodermis ; the disintegration of the muscles is, indeed, the first process to take place in the pupa. The muscles of the most anterior segments of the body disintegrate first ; and the superficial layers are affected before the deeper ones. The disintegration of the larval muscles is brought about in the following way. A large number of amoeboid blood-corpuscles, which have collected on the surface of the muscle-bundle, penetrate through the sarcolemma and wander into the interior of the muscle-substance, pressing into fissures which develop in it. It often appears as if the muscle-substance is actually cut out in the parts corresponding to the processes of the phagocytes which extend into it. The muscle, in this way, breaks up into a number of particles which soon become rounded and are immediately swallowed by the phagocytes. The muscles are thus transformed into a great accumulation of granidar spheres, which finally shift apart and become scattered in the body- cavity of the pupa. The muscle-nuclei are digested and assimilated by the phagocytes in the same way as the muscle-substance. 382 INSECTA. WIS Fio. ISO.— Larval digestive tract of one of the Mvscidue with the imaginal discs depicted (after Kowalevsky). bd, caecal tubes of the chylific stomach ; eft, chylific stomach ; /, fat-cells at the apex of the salivary glands ; h, proctodaeal imaginal Tring ; ht, proctodaeum ; ie, imaginal cells of the mid-gut epithelium ; im, Imaginal cells of the muscles of the mid-gut ; ims, pos- terior abdominal imaginal disc; is, imaginal rings of the salivary glands ; ma, Malpighian vessel ; pr, proventriculus ; s, sucking stomach ; sp, salivary glands; v, stomodaeal imaginal ring. Van Rees and Kowalevsky are. in entire agreement with regard to the details of the disintegration of the larval muscles by phagocytes, which had already been the subject of conjecture to Metschnikoff and Ganin. According to Van Rees, not all the muscles of the larva undergo such disintegration. Certain dorsal groups of the ex- ternal oblique muscle of the second thoracic segment are retained and pass over into the wing-muscles of the adult after radical internal modifications consisting of an in- crease in the number of the nuclei and a rearrangement of the muscle- substance. This manner of tran- sition from larval to imaginal musculature appears very remark- able, but the descriptions of Van Rees leave hardly any doubt as to the accuracy of these observations.* As a rule, the formation of the imaginal muscle-groups takes place from the permanent mesoderm, which is derived from that of the imaginal discs (Fig. 188 C, m). We have already stated (pp. 375- 381) all that is as yet known as to the origin of the former. Intestinal Canal. The disintegration of the larval intestine and the de- velopment of the permanent * [According to recent investiga- tions by Kawawaiew(No. XXIV.), phagocytosis plays a very unim- portant role in the metamorphosis of Lasius, this being especially noticeable in the disintegration of the larval muscles ; the phagocytes here do not cause the liquefaction of the fibres, but are only concerned in the absorption of the liquid mass. — Ed.] INTESTINAL CANAL. 383 organ from distinct imaginal discs take place, as with the hypo- dermis, side by side, in such a way that the continuity is not anywhere interrupted. "We owe our knowledge of the imaginal discs of the intestinal canal to Ganin (No. 107). More recently, Kowalevsky (No. 112) and Van Kees (No. 121) have described the development of the intestinal canal in detail. The imaginal discs of the intestine, which in the pupa is very short, are found in the enteron in the form of numer- HI j- ous scattered cell-islands (Fig. 189, ie); in both the stomodaeum and proctodaeum these ap- pear as a ring (v and h) of imaginal tissue capable of great increase. The imaginal ring of the fore-gut (v) lies in the region of the so-called proventriculus (pr, cf. Fig. 191, im), while that of the hind -gut is to be sought immedi- ately below the aperture of the Malpighian ves- sels. m -^. — rru The regeneration Fig. 190.— Transverse section through the pupal mid-gut of one of the Muscidac (after Kowalevskv). e, degen- erating larval epithelium ; /, the newly-formed cell- layer round the same ; m, muscular coat ; m', imaginal cells of the muscular coat; o, imaginal discs of the mid-gut epithelium ; t, tracheal trunks. of these two parts of the intestinal canal is not exclusively brought about by these two rings, but the imaginal rudiments of the neighbouring parts of the body- surface participate in it. It thus appears that the most anterior part of the oesophagus is yielded by the imaginal discs round the mouth, while the discs of the eighth abdominal segment that surround the anus (Fig. 189, ims) produce, by invaginating, the rectal sac and the rectal papillae. The development of the permanent mid-gut (chylific stomach) proceeds in such a way that the island-like imaginal discs, increasing considerably in number, extend over the external or basal surface of the epithelium of the larval enteron (Fig. 190, o) until they come into contact and fuse, the wall of the imaginal intestine being thus formed. The whole of the larval enteric epithelium (e) is at the same time cast off into the interior of the gut and, surrounded by 384 INSECT A. a layer of small cells (/), perhaps derived from the imaginal discs, as well as by a gelatinous envelope, forms the so-called yellow body which, until its disintegration, remains lying in the pupal intestine. The larval muscular coat (hn) remains intact so long as the imaginal mid-gut is not completely developed, but is afterwards attacked by phagocytes and destroyed. The permanent muscular layer develops from single cells lying on the outer surface of the imaginal discs (Figs. 189, im and 190, m), which must be described as special imaginal cells of the intestinal muscles. The metamorphosis of the fore-gut is commenced by the degeneration of the proventriculus and of the sucking stomach. The proventriculus or gizzard (Fig. 191, pr), which seems to be formed by a process of infolding or intussusception of the fore -gut, degenerates through the flattening out of this fold. The sucking stomach also degenerates in a similar way, retreating more and more into the oesophagus, so that, in place of the original diverticulum, there now only remains a widening of the lumen of the oesophagus. At the same time, this part of the gut is attacked and disintegrated by phagocytes, while the disorganised parts are replaced by the gradually extending imaginal portions of the wall. The imaginal ring of the fore-gut (Fig. 191, im), which, according to Kowalevsky, undertakes the formation of a great part of the permanent oesophagus, closes posteriorly, so that the communication with the mid-gut seems to be interrupted. The transformation of the hind-gut takes place in a similar manner. Here also the imaginal ring extends so as to form a tube which, growing round the opening of the Malpighian vessels, closes towards the mid -gut, while posteriorly it is connected with the disintegrating proctodaeum. In a similar way, the territory of the proctodaeum is circumscribed by an imaginal tube growing from behind, formed by one of the imaginal discs found in the neighbourhood of the anal aperture, till finally, when the whole of Fig. 191.— Longitudinal section through the larval proventri- culus of one of the Mttscidae (after Kowalevsky). im, stomodaeal imaginal ring ; oe, oesophagus ; pn, proventri- culus. THE TRACHEAL SYSTEM. 385 the larval hind-gut is changed into granular cells, the two imaginal sections of the tube seem to be approximated until they come into contact. In the above description, we have mainly followed the accounts of Kowalevsky. According to Van Rees, the reconstruction of the fore- and hind-gats is brought about, not merely by the imaginal discs already mentioned, but a simultaneous regeneration of the larval epithelium takes place, only some of the cells of the larval epithelium undergoing disintegration, while others, on the contrary, undergo repeated division and form a portion of the imaginal oesophagus. The salivary glands of the larva (Fig. 189, sp) are entirely destroyed by phagocytes. They are reconstructed from the imaginal discs which, according to Kowalevsky, form a ring at the anterior end of the glandular tube (c/. the statements of Schiemenz, No. 125). From the accounts hitherto given, it is difficult to make out what kind of transformation is undergone by the Malpighian vessels. According to Van Rees, a regeneration of the larval cells through division may take place, or these elements may disintegrate. The method of transformation of the intestinal canal described above seems very widespread among the holonietabolic Insecta. It has been observed not only in Diptera, but also in Lepidoptera (Kowalevsky, Fkenzel), Coleoptera (Ganin), and Hymenoptera (Ganin). The casting of the mid-gut epithelium was also found by Kowalevsky in Corethra, Culex, and Chironomus. [In this connection the more recent works of Rengel (No. XXXVII. ), Mobtjsz (No. XXXIV.), and Kakawaiew (No. XXIV.) should be con- sulted.— Ed.] The Tracheal System. The fact that the tracheal system undergoes important transforma- tion during metamorphosis is demonstrated by the entirely different form assumed by it in the larva, the pupa, and the imago. We have only to recall that the larva of the Muscidae breathe through a pair of stigmata at the posterior end of the body, the pupa through one occurring in the pro-thorax, while the imago possesses six pairs of stigmata (situated on the meso-thorax, meta-thorax, and four abdominal segments). There is no doubt that in the larva and pupa the other stigmata are closed. The tracheal strands connected with the latter, as well as some other parts of the tracheal system, as pointed out by Wbismann, seem to function as imaginal discs for the regeneration of the tracheal matrix (A7ax Rees), and a regeneration of this epithelium can also frequently be seen to proceed from simple division of the cells. The disintegration of the degenerating parts of the tracheal system is accomplished under the influence of phagocytes in the way already described. 2c 386 INSECTA. The Nervous System. The central parts of the nervous system pass directly over from the larva to the imago, although they undergo considerable modifica- tions of form and position. At the same time certain histological changes, known as histolysis, are said by Weismann to take place in them, e.g., a disintegration and reconstruction of the tissues within the organs without disturbing their continuity. Recently, however, the term hystolysis has been applied to the disintegration of the tissues of the pupa generally. "We have as yet little light on the question of the transformation of the peripheral nervous system. Although it must be considered probable that the destruction of the larval muscles is accompanied to some extent by a degeneration of the motor nerves, this is not the case with the nerves that run to the extremities. These can be recognised in the larva in the form of nerve-strands connect- ing the imaginal discs with the central nervous system. These strands, according to Van Rees, pass over from the larva to the pupa and imago, so that, Avhen the limb-rudiments develop further, only the distal parts of the nerves belonging to them appear as new formations. The Fat-body. The fat-body of the larva also is destroyed through the action of the phagocytes in the way described in connection with other tissues. Its reconstruction appears to proceed from the mesoderm of the imaginal discs. It is possible also that the accumulations of embryonic cells, assumed by Schaffer to be formative centres, have to do with the regeneration of the fat-body. In any case it is to be derived from mesodermal tissue. Even though Wielowiejsky observed the origin of the fat-body in Corethra from a cell-layer lying beneath the hypodermis of the larva, such an observation does not necessarily support the view of Schaffer, who thought he had convinced himself that, in Muse a, the fat-body of the larva is derived partly from the hypodermis and partly from the tracheal matrix, and thus from ectodermal tissue. The ultimate fate of the Phagocytes. We have seen that the development of the imaginal organs, in cases where these are not taken over direct into the pupa, always proceeded from the imaginal discs. The phagocytes, the number of which increases greatly in the pupa, do not (as was formerly GENERAL CONSIDERATIONS. 387 thought) take any direct share in the building up of the tissue. Their significance seems to be that of destroyers of the larval organs ■which are doomed to destruction ; the constituent parts of these organs are taken in and digested by them, and, through their capacity of locomotion, they conduct particles of nourishment to the organs that are in process of reconstruction. But what is the fate of these elements after the ontogenetic processes in the pupa are completed 1 There can be no doubt that some of the so-called granular cells develop into ordinary blood-corpuscles. The majority of them apparently undergo degeneration. The phagocytes them- selves are finally used as food for the newly-formed tissues. Interest attaches here to the observation of Van Rees that many phagocytes finally wander into the newly-formed hypodermis, and there, in the spaces between the hypodermal cells, undergo degeneration. General considerations regarding the development of the Imago in the Pupa. "We have seen that the development of the body of the imago proceeds from distinct formative centres (imaginal discs) already present in the larva, having appeared during embryonic life. We have met with such imaginal discs in connection with the different parts of the head, the limbs, the hypodermis, and the various parts of the intestinal canal. We have seen that the development of the mesodermal organs of the imago (muscles, connective tissue, fat-body) proceeds from a mesodermal part of the imaginal discs, the first origin of which, however, is still somewhat obscure. Simultaneously with the building up of the imaginal organs, we have the destruction of the larval organs through the action of the phagocytes. These two processes (disintegration and regeneration) go on side by side in such a way that the continuity of the organ is in most cases perfectly preserved, complete disintegration of the larval tissue only occurring after the development of the permanent organ. The musculature of the larva here forms an exception, as it undergoes disintegration very -early. We must, in conclusion, once more point out that the sharp distinction between the larval, the pupal, and the imaginal stages seems to be founded only upon the appearance of the external surface of the body, as resulting from the consecutive moults. The phenomena of internal development, on the contrary, represent a ■complete, continuous series of transformations, which do not show 388 INSECTA. any such abrupt changes. We can, however, in the main, distinguish, according to the vital functions belonging to them, the forms of the larval, the pupal, and the imaginal stages. III. Parthenogenesis, Paedogenesis, Heterogeny. A capacity for developing unfertilised eggs in a parthenogenetic manner has repeatedly been observed in the Insecta. Partheno- genesis may here be either occasional {e.g., many Lepidoptera, Bombyx, Liparis) or may be of normal occurrence, often recurring at fixed intervals in the ontogenetic cycle.* The males of the social Wasps and Bees, for instance, are produced from eggs that develop parthenogenetically. This is also the case in the Ants, and in Nematus and other Tenthredinidae, while, in the Gynipidae, only females are produced from the parthenogenetic eggs. In the Lepidoptera it seems to be the rule that females come from the parthenogenetic eggs. In Psyche and Soleiiobia, for example, a succession of many parthenogenetic generations was observed, while males were only seldom met with. The same is the case in Apatania among the Trichoptera (Klapalek). In certain Cynipidae there is- a cyclic alternation of parthenogenetic females and male and female sexual forms of a different shape (true heterogeny). There thus develops, in the galls produced by a form known as Spathegaster baccarum, a gall-wasp of different shape called Neuroterus ventricularisy of which only parthenogenetic females are known. The unfertilised eggs laid by Neuroterus, which develop in peculiarly shaped galls,, give rise again to the sexual generation of Spathegaster. With the possibility of attaining reproduction by means of un- fertilised eggs is connected the shifting back of this process to an early stage of development (paedogenesis). Thus, according to Grimm, in one species of Chironomus, the pupa lays eggs, while other Diptera (Cecidomyia), even as larvae, are capable of reproducing themselves parthenogenetically and viviparously. The partheno genetic reproduction of the Aphidae must also to some extent be regarded as paedogenesis ; in these Insects it may happen that the embryo produced parthenogenetically may in its turn reproduce itself. The heterogeny of the Phytophthires seems to be founded on the * [In this connection Nussbaum (No. XXXV.) lias recently made a series of care- ful experiments on certain Lepidoptera, viz. , Bombyx, Porthcsic, and Liparis. He only succeeded in demonstrating the parthenogenetic condition in Bo7iibyx, in which form two per cent, of the unfertilised eggs (1100) showed segmentation. but never hatched. — Ed.] PARTHENOGENESIS, PAEDOGEXES1S, IIETEROGENY. 389 •definite alternation of a generation of parthenogenetic females with a normal generation of males and females, the latter generation reproducing the former hy means of a fertilised egg, these genera- tions being distinguished from each other by certain features in the structure of the body. In the Aphidae, the hibernating fertilised winter-eggs yield in spring a generation that reproduces itself parthenogenetically and viviparously, and which is followed during the spring and summer by a series of generations reproducing themselves in the same way, the individuals of which are often winged, but may also be wingless. This series is closed towards autumn by a generation known as the sexupara, the parthenogenetic and viviparous descendants of which are, as a rule, winged males and wingless females. After copulation has taken place, the female lays the fertilised winter -eggs, from which, in the next spring, the first generation capable of parthenogenetic reproduction hatches. Under certain circumstances it, however, appears that single indi- viduals of the parthenogenetic generations are able to hibernate, and to give origin in the spring to a new parthenogenetic series. In the same way, among other Phytophthires, there are often parallel series of cycles of generations (Dreyfuss, No. 137). A further complication in the cycle of development of the Aphidae is brought about in connection with frequent migration from one plant to another. A winged parthenogenetic generation frequently appears, and then may migrate to a different plant, there to reproduce itself, and in a later generation returns to the original host. These wandering generations, the occurrence of which was often pointed out by Lichtenstein, have been distinguished as emigrantes, alienicolae, and remigrantcs by Blochmann (No. 135). In Pemphigus terebinthi, for example, according to Derbes, the fertilised egg gives rise to a wingless parthenogenetic generation (I.), which produces another winged generation (II., emigrantes). This generation leaves the place occupied up to this time and produces a third generation (III., remigrantes = sexupara), which, after hibernating, returns to the original host and produces the small, mouthless, wingless sexual animals without intestine (IV., sexuales). The cycle of generations in Pemphigus terebinthi is interesting because the sexual generation does not here occur, as it usually does, in the autumn, but in the spring, being produced by hibernating parthenogenetic forms. Conditions similar to those in the Aphidae are found in the Chermetidae, which have recently been much investigated. The chief distinction between the two is that here the parthenogenetic, like the sexual generation, is also oviparous. In Phylloxera, quercus, according to Lichtenstein, the winter-eggs that are laid on Quercus cocci/era give rise to a mother animal (fundatrix), which produces parthenogenetically a winged generation capable of partheno- genetic reproduction (emigrantes) ; these wander over to the leaves of Quercus pedunculate, and Q. pubescens. Several wingless generations (alienicolae) now follow, which reproduce parthenogenetically, the return to Quercus cocci/era being finally made possible by the production of the winged sexupara. The 390 INSECT A. eggs laid by the sexupara here give rise to the wingless sexual generation devoid of proboscis and intestinal canal, which lays the winter-eggs. In Phylloxera vastalrix, the young animals that develop out of the winter-eggs laid beneath the bark of the trunk wander to the root, there to give origin partheno- genetically to several generations of wingless Phylloxera, which cause the swellings on the root. The series of these generations closes by the production of winged sexupara, which wander up the trunk and swarm. These forms also are parthenogenetic. Their eggs, which vary in size according to the sex of the developing embryo, yield sexual animals devoid of proboscis, intestine, and wings, which produce the winter -eggs. Parallel series are introduced into this cycle of generations also, e.g., the wingless Tctrancura, living on leaves which run parallel with the generations of Phizobia. In the cycle of generations of the genus Chermes recently investigated by Blochmann (Nos. 134 and 135), Dueyfuss (No. 137), and Cholodkovsky, similar conditions are found, but these are in some respects still very obscure. In Chermes abietis, the fertilised egg gives rise to a wingless parthenogenetic female (fundatrix, I.), which hibernates at the base of the buds of the fir-tree and, by piercing the buds, deforms them into galls. From this generation is produced a second (II.) consisting of winged parthenogenetic forms, some of which migrate to- the larch and there give rise to a wingless generation (III.) which feeds on the needles and hibernates beneath the bark. These parthenogenetic alienicolae, in the following spring (the second year of the cycle), produce the winged remigrantes (IV.) or sexupara, which return to the fir-tree and there produce the wingless female and male, the fertilised eggs of which give rise once more to a fundatrix (I.). This cycle also is accompanied by a parallel series of forms that do not emigrate to the larch, but remain on the fir-tree. IV. General Considerations. It can hardly be doubted that the Insects and the Myriopoda are very intimately related. If it is considered that the anatomical features possessed in common and the similarity in development (although, indeed, the ontogeny of the Myriopoda is only partly known) are not sufficient to establish this relationship, great stress can be laid on the presence of transition-types, such as the Symphyla (Scolopendrel/a, Fig. 192) and Thysanura (Campodea, Fig. 193), which connect the two groups. It has only to be pointed out here that in the Thysanura, which are most intimately connected with the Orthoptera, we have, in the absence of wings and in the presence of the sac-like protrusible ventral sac, a recurrence of morphological characters which, while they are wanting in the higher Insects, are nevertheless found in the Myriopoda. On the other hand, the Myriopoda are closely related to Peripatas, so that we are justified in regarding the Onychophora, the Myriopoda, and the Insecta as belonging to a single phyletic ontogenetic series, which, through Peripatus, is linked on to the hypothetical racial form of the GENERAL CONSIDERATIONS. 391 Arthropoda (Protostraca) and, through the latter, to the Annelida (cf. Vol. ii., p. 315, and Vol. iii., p. 427). The Insecta represent the highest grade of development of this phyletic series. That they are more highly specialised than the Myriopoda can he seen in the sharper demarcation of the different regions of the hody, the fixation of the number of body-segments, and the development of a new locomotory system, the wings. The marking-off of the three regions which can be distinguished in the body of the Insect (head, thorax, and abdomen) seems to be foreshadowed in the Myriopoda. Here also we find an anterior region, the head, sharply distinguished from the rest of the body. Further, of the trunk-segments that follow this region, the anterior (thoracic) segments may be dis- tinctly differentiated from those which follow (the abdominal region) ; thus, by way of example, we may recall the fact that in the Diplopoda the thoracic segments do not unite to form double segments, as is the case with the other trunk-segments. "We have, however, already pointed out (p. 236) that the region here distinguished as the thorax cannot be entirely identified with the thorax of the Insecta, since, in the Diplopoda, a limbless segment is inter- calated between the three limb-bearing segments of the thorax (Fig. 121 B, p. 235, and Fig. 122, p. 237), a modification not found in the Insecta. Although the division of the body into regions can also be recognised as indicated in the Myriopoda, it is much more distinctly marked in the Insecta. The boundary between the thorax and the abdomen especially is much more distinct. This is connected with the division of labour between the two regions. In the Insecta, the most important loco- motory organs are concentrated in the thoracic region. This has led to the greater rigidity of the thorax and the development of large masses of muscle, while the softer, more extensible abdominal region is the receptacle for almost all the vegetative organs. Into this region have shifted the most important parts of the intestinal canal and of the respiratory and circulatory systems, as well as the genital organs. Fig.192. — Scolopendrella im- ■maculata (after Latzel, from Lang's Text-book). 392 IXSECTA. It should lie mentioned that the boundary between the thoracic and abdominal regions is, in many Insect larvae, less sharply marked. This is connected with the fact that, in larvae, the thorax is frequently of less significance for the locomotion of the whole body than in the imagines, either because locomotory organs develop on the abdomen also {e.g. , in caterpillars), or that such organs are altogether wanting on the thorax as well (maggot-shaped larvae). More careful examination, especially of the inner organs, will, however, reveal in these cases also important differences between the thoracic and the abdominal segments. As we find that the separation of the thorax from the abdomen is very marked in the Thysanura, we may regard it as a feature inherited long ago by the Insect phylum, and may consider the apparent obliteration of these boundaries in certain larval forms as merely a second- ary phenomenon. The loss of extremities in the abdominal region is an important feature which dis- tinguishes the Insecta from the Myriopoda. With regard to the derivation of the Insecta from the latter group, or from forms re- sembling the Myriopoda, the fact that the rudiments of abdominal extremities appear in the insect embryo and disappear later is of importance (pp. 296-300). The ventral stylets found on the abdomen in the Thysanura have repeatedly been regarded as vestiges of extremities, and this seems all the more probable as, in Maehilis, these stylets actually function as locomotory organs. Recently, however, following Haase (No. 153), and supported by the occurrence of similar stylets on the coxae of the thoracic limbs of Maehilis, and on most of the limbs in Scolopendrella, zoologists have been inclined to regard these appendages merely as coxal spurs (p. 299). On the first abdominal segment of Campodea, on the contrary, there is a true limb-rudiment. While, in the Myriopoda, the number of the body-segments varies greatly in the different genera and species, the number seems to be fixed and universally prevalent in the Insecta. The thorax is always composed of three segments, each of which carries a pair of legs (a fact which gave rise to the name of Hexapoda). In the same way it seems to be clearly established by ontogeny that the abdominal region is universally composed of ten trunk-segments and one sub- Fio. 193. — Campodea staphy- linus (after Lubbock, from Lang's Text-booV). GENERAL CONSIDERATIONS. 393 sequent anal segment (telson). Greater difficulty arises in reckoning the number of segments which have been drawn into the formation of the head. Three maxillary segments (a mandibular and a first and second maxillary segment) here combine with an anterior primary cephalic section. The segmentation of the brain leads us to suppose that the latter is composed of three segments (p. 325), while between this section and the mandibular segment a vestigial so-called pre-maxillary segment seems to be intercalated. In reckoning the segments here, however, we are on somewhat hypothetical ground. It may be mentioned that the antennae belong to the second brain- segment, and, by their originally post-oral position, as well as by their relation to the coelomic sacs belonging to that segment (in Orthoptera, p. 295), in all respects resemble true trunk-limbs. This is in entire agreement with what has been learnt of these limbs in connection with Peripatus and the Myriopoda. One of the most interesting questions in the phylogeny of the Insecta is that of the rise of the wings. The rudiments of the wings appear on the meso- and meta-thorax as dorsal integumental out- growths, the inner cavities of which receive later the tracheal ramifications. It is an interesting fact that similar lateral fold-like widenings of the dorsal plates, which recall the first rudiments of wings, also occur on the pro-thorax (Machilis and Blatta). These are most clearly visible on the larvae of Calotermes (Fig. 194, F. Muller, Xo. 158), in the youngest stages of which outgrowths of the pro-thorax and meso-thorax are first evident, these being originally devoid of tracheae. While the anterior pair of these outgrowths degenerates, the posterior pair is supplied with tracheae, and is thus transformed into the rudiment of the fore-wing, the rudiment of the hind-wing appearing simultaneously on the meta- thorax. The great similarity in position and structure between the wing-rudiments and the leaf-shaped tracheal gills, as found on the abdominal segments of the Ephemerid larvae (Fig. 177, h, p. 358), has led to many attempts to consider them as homodynamous structures. This view, which was adopted by Gegenbauer and Lubbock (No. 156), has recently also received the support of Eedtenbacher (Xo. 165). F. Muller, who also supports the above, is inclined to hold that the original function of the wings was respiratory. This view, which seems well supported by the structure of the wing-rudiments, within which are found blood- spaces and tracheal ramifications, involves the assumption that the winged Insects are derived from an aquatic form. The phyletic series 394 INSECTA. mentioned above, leading from Peripatus through the Myriopoda and Thysanura to the Orthoptera, contains throughout only forms living on land and adapted for terrestrial life. We have no reason for assuming that an aquatic ancestral form has been introduced into the series of ancestors of the winged Insecta (Pterygogenea). The manner of life of the aquatic larval forms of the Hemimetabola, as well as their respiratory organs, which are suited to life in water, must be regarded as secondarily acquired. For the same reasons Ave cannot adopt the view of Dohrn, who, going still further back in the phyletic series, is inclined to refer the tracheal gills of the Ephemerid larvae as well as the wing- rudiments to the elytra of the Anne- liclan ancestors of the Insecta (Dohrn, the Pantopoda). It must be pointed out that, in Peripatus, as well as in the Myriopoda, corresponding integumental folds are altogether wanting. We therefore consider that Grassi (Xo. 150) is justified in regarding these organs as new acquisitions by the Insectan phylum, and in tracing them back to integumental outgrowths of the lateral margins of the tergal plates that have been constricted off and have become independent, the wing-muscu- lature being derived from the system of dorso-ventral muscles, which is also represented in the other segments of the body. We may perhaps assume that the transition from the creeping method of locomotion to flight was made through the acquisition of a climbing habit, in which distances Avould occasionally be overcome by springing, a circum- stance which gave rise to the development of parachute-like widenings of the thoracic segments. The transition from such integumental folds, used as a parachute but still immovable, to independent active locomotory organs seems to us fairly plausible. It is perhaps not without significance that the capacity for rising above the surface on which they rest is common among the Thysanura, the Collembola, and the Orthoptera, and that in the latter (e.g., in Psop>hus stridulus) the wings are scarcely used other- wise than as parachutes. The limitation of wings to the meso- and Fig. 194. — Larvae of Ccdotermes rugosus (after F. Muller). /', wing-like appendages of the pro- tliorax ; /", rudiment of the fore- wing ; /'", rudiment of the hind- wing. GENERAL CONSIDERATIONS. 395 meta-thorax may be connected with the position of the centre of equilibrium of the body. We agree with Brauer (No. 146) in considering the wingless condition as a primary characteristic only in the Apterygogenea, whereas in those wingless orders of Insects (the Mallophaga, Siphonaptera, etc.) which are placed with the Pterygogenea it must be regarded as secondarily acquired. The segmental arrangement of the tracheal stigmata should be noted. It appears that originally a pair of stigmata occurred on each of the three thoracic segments, as well as the eight following abdominal segments, at least, the respiratory system of the Thysanura, as investigated by Grassi and Haase, is favourable to such an assumption. In most Insecta, however, the number of thoracic stigmata is reduced. There does not appear to be a true pair of stigmata in the head. We have already given the reasons (pp. 323 and 335) why neither the encloskeletal invaginations of the head nor the salivary glands can be regarded as homodynamous with the tracheal invaginations. It should, however, be mentioned, on the other hand, that the presence of a pair of stigmata belonging to the head has been maintained in Scolopendrella (Haase) and in Sminthurus (Lubbock). We have still to mention the compound eyes (facet-eyes) as one of the features which raise the Insecta to a higher level than the Myriopoda. The most primitive form of eye in the Insecta is evidently represented by the ocellus (Fig. 165, p. 332), the structure of which, according to Grenacher (No. 151), may still, in a few cases, be traced back direct to the simple cup-shaped eye, while, in other cases, through the development of a vitreous body-layer (lentigen layer), it becomes a bilaminar complicated eye (Fig. 164 B, p. 331). We shall hardly err in deriving the Insectan ocellus direct from the cup-shaped eyes of the Annelida (Kennel, No. 154). The compound eye, on the contrary, appears to correspond to an accumu- lation of ocelli, in which the number of ocelli has been increased while the single ommatidia have sunk to a lower level of functional capacity. We have already seen (p. 242) that, in the Myriopoda, an almost complete series of transitions is to be found between the aggregations of ocelli and the true facet-eyes. We shall therefore be justified in assuming this derivation as highly probable for the facet -eyes of the Insecta. For the relation of the facet -eye to the ocelli of the same animal, cf. p. 333. Bearing in mind the fact that Machilis already possesses facet -eyes, we must regard the latter as a somewhat ancient acquisition among the ancestors 396 INSECTA. of the Insecta, and shall feel inclined to regard the absence of the facet- eye (whether in larvae or in imagines) as the result of degeneration. We must, in conclusion, point out a few more important factors in the embryonic development of the Insecta. The first of these is the development of the embryonic envelopes, the acquisition of which (like the development of flying) proves the Insecta to be the most highly developed of all Arthropods. It therefore seems remarkable that the Insecta in other respects, especially with regard to the way in which the germ-layers form, have retained very primitive char- acters. The long blastopore which extends over the whole of the ventral side, the presence of a distinct invagination-gastrula which leads to the development of an archenteric tube, and the manner in which the mesoderm separates from the entoderm must here be mentioned in this connection. "With regard to the last point, it should be mentioned that the separation of the mesoderm from the entoderm is accomplished by a process which can be traced back to that of infolding, so that Kowalevsky (Xo. 49) quite correctly compared the formation of the germ-layers in the Insecta with their formation in Sar/itta, a proceeding in which he was afterwards supported by Rabl. The coelomic sacs of the Insecta may thus, according to their development, be regarded as archenteric diverticula. Another point of interest is the transformation undergone in later stages by the primitive segments, which were treated in detail in the chapter on the development of the heart and the genital organs. LITERATURE. Embryonic Development. Older treatises by Herold, Kollicker, Leuckart, Rathke, Robin, Suckow, and Zaddach. 1. Ayers, H. On the development of Oecanthus niveus and its parasite Teleas. Mem. Boston Soc. Nat Hist. Vol. iii. 1884. 2. Balbiani, E. G. Sur la signification des cellules polaires des Insectes. Compt. Mew J. Acad. Sci. Paris. Tom. xcv. 1882. 3. Balbiani, E. G. Contribution a l'etude de la formation des organes sexuels chez les Insectes. Recueil Zool. Suisse. Tom. ii. 1885. 4. Barrois, J. Developpement des Podurelles. Assoc. Franc. A cane. Sci. 7e Sess. 1879. LITERATURE. 397 5. Blochmann, F. Ueber die Richtungskbrper bei Insecteneiern. Morph. Jahrb. Bd. xii. 1887. 6. Bobretzky, X. Ueber die Bildung des Blastoderms und der Keimblatter bei Insecten. Zeitschr. f. Wiss. Zool. Bd. xxxi. 1878. 7. Brandt, A. Beitrage zur Entwicklungsgeschichte der Libellu- liden und Hemipteren. Mem. Acad. St. Petersbourg (7). Tom. xiii. 1869. 8. Brandt, A. Ueber das Ei und seine Bildungsstatte. Leipzig, 1878. 9. Brandt, A. Commentare zur Keimblaschentheorie des Eies. I. Die Blastoderm-elemente und Dotterballen der Insecten. ArcMv. /. Mikro. Anat. Bd. xvii. 1880. 10. Bruce, A. T. Observations on the Embryology of Insects and Arachnids. A memorial volume. Baltimore, 1887. 11. Butschli, 0. Zur Entwicklungsgeschichte der Biene. Zeitschr. f. Wiss. Zool. Bd. xx. 1870. 12. Butschli, 0. Bemerkungen iiber die Entwicklungsgeschichte von Musca. Morph. Jahrb. Bd. xiv. 1888. 13. Carriere, J. Die Entwicklung der Mauerbiene (Chalicodoma muraria Fabr.) im Ei. Archiv. f. Mikro. Anat. Bd. xxxv. 1890. 14. Carriere, J. Kurze Mittheilungen aus fortgesetzten Unter- suchungen iiber die Sehorgane. 7. Die Entwicklung und die verschiedenen Arten der Ocellen. Zool. Anz. Jahrg. ix. 1886. 15. Carriere, J. Die Driisen am ersten Hinterleibsringe der Insectenembryonen. Biol. Centralhl. Bd. xi. 1891. 16. Cholodkovsky, X. A. Ueber die Bildung des Entoderms bei Blatta germanica. Zool. Anz. Jahrg. xi. 1888. 17. Cholodkovsky, X. A. Zur Embryologie von Blatta germanica. Zool. An?:. Jahrg. xiii. 1890. 18. Cholodkovsky, X. A. Zur Embryologie der Hausschabe (Blatta germanica). Biol. Centralhl. Bd. x. 1890. 19. Cholodkovsky, X. A. Studien zur Entwicklungsgeschichte der Insecten. Zeitschr. f. }Yiss. Zool. Bd. xlviii. 1889. 20. Cholodkovsky, X. A. Ueber die Entwicklung des centralen Xervensy stems bei Blatta germanica. Zool. Anz. Jahrg. xiv. 1891. 20a. Cholodkovsky, X. A. Die Embryonalentwicklung von Phyllo- dromia (Blatta) germanica. M* in. Acad. St. Petersbourg (7). Tom. xxxviii. 1891. (Seen only after this volume had been printed.) 398 INSECTA. 21. Dohrn, A. Notizen zur Kenntniss der Insectenentwicklung. Zeitschr. f. Wiss. Zool. Bd. xxvi. 1876. 22. Ganin, M. Beitriige zur Erkenntniss der Entwicklungsgeschiclite bei den Insecten. Zeitschr. f. Wiss. Zool. Bd. xix. 1869. 23. Ganin, M. Ueber die Embryonalhiille der Hymenopteren- und Lepidopterenembryonen. Mem. Acad. St. Petersbourg (7). Tom. xiv. 1870. 24. Graber, V. Vorlaufige Ergebnisse einer grb'sseren Arbeit iiber vergl. Embryologie der Insecten. Archiv. f. Mikro. Anat. Bd. xv. 1878. 25. Graber, V. Ueber die Polypodie bei Insectenembryonen. Morph. Jalivb. Bd. xiii. 1888. 26. Graber, V. Ueber die primare Segmentirung des Keimstreifs der Insecten. Morph. Jahrb. Bd. xiv. 1888. 27. Graber, V. Vergleicbende Studien iiber die Keimhiillen und die Riickenbildung der Insecten. Denkschr. Acad. Wiss. Wien, Bd. lv. 1888. 28. Graber, V. Vergleichende Studien iiber die Embryologie der Insecten und insbes. der Musciden. Denkschr. Acad. Wiss. Wien. Bd. lvi. 1889. 29. Graber, V. Ueber den Bau und die pbylogenetiscbe Bedeu- tung der embryonalen Bauchanhange der Insekten. Biol. Centralis. Bd. ix. 1889. 30. Graber, V. Vergleicbende Studien am Keimstreif der Insecten. Denkschr. Acad. Wiss. Wien. Bd. lvii. 1890. 31. Graber, V. Ueber die embryonale Anlage des Blut- und Fett^ewebes der Insecten. Biol. Centralbl. Bd. xi. 1891. 32. Grassi, B. Intorno alio sviluppo delle api nelF uovo. Atti Acad. Gioenia Sci. Nat. Catania (3). Tom. xviii. 1884. 33. Grassi, B. Breve nota intorno alio sviluppo degli Japyx. Catania, 1884. See also : I. Progenitori degli Insetti e dei Miriopodi. 1. L' Japyx e la Campodea. Atti Acad. Gioenia Sci. Nat. Catania (3). Tom. xix. 1885. 34. Hallez, P. Orientation de l'embryon et formation du cocon chez la Periplaneta orientalis. Compt. Rend. Acad. Sci. Paris. Tom. ci. 1885. 35. Hallez, P. Sur la loi de l'orientation de l'embryon chez les Insectes. Compt. Rend. Acad. Sci. Paris. Tom. ciii. 1886. 36. Hatschek, B. Beitriige zur Entwicklungsgeschichte der Lepi- dopteren. Jen. Zeitschr. f. Natunr. Bd. xi. 1877. LITERATURE. 399 37. Heider, K. Ueber die Anlage der Keimblatter von Hydro- philus piceus L. Abh. k. Acad. Wiss. Berlin, 1885. 38. Heider, K. Die Embryonalentwicklung von Hydrophilus piceus L. Theil. i. Jena, 1889. 39. Henking, H. Die ersten Entwicklungsvorgange im Fliegenei und freie Kernbildung. Zeitschr. f. Wiss. Zool. Bd. xlvi. 1888. 40. Henking, H. Ueber die Bildung von Richtungskorpern in den Eiern der Insecten und deren Scliicksal. Nachr. Ges. Wiss. Gbttingen, 1888. 41. Hertwig, 0. and R. Die Coelomtbeorie. Versuch einer Erklarung des mittleren Keimblattes. Jena, 1881. 42. Heymons, R. Ueber die hermaphroditische Anlage der Sexual- driisen beim Mannchen von Phyllodromia (Blatta) germanica. Zool. Anz. Jahrg. xiii. 1890. 43. Heymons, R. Die Entstehung der Geschlecbtsdriisen von Phyllodromia (Blatta) germanica L. In.-Diss. Berlin, 1891. 44. Jordan, K. Anatomie und Biologie der Physapoda. Zeitschr. f. Wiss. Zool. Bd. xlvii. 1888. 45. Kadyi, H. Beitrag zur Kenntnis der Vorgange beim Eierlegen der Blatta orientalis. Zool. Anz. Jahrg. ii. 1879. 46. Korotneff, A. Entwicklung des Herzens bei Gryllotalpa. Zool. Anz. Jahrg. vi. 1883. 47. Korotneff, A. Die Embryologie der Gryllotalpa. Zeitschr. f. Wiss. Zool. Bd. xli. 1885. 48. Kowalevsky, A. Embryologische Studien an Wiirmern und Arthropoden. Mem. Acad. St. Petersbourg (7). Bd. xvi. 1871. 49. Kowalevsky, A. Zur embryonalen Entwicklung der Musciden. Biol. Centralhl. Bd. vi. 1886. 50. Kupffer, C. Ueber das Faltenblatt an den Embryonen der Gattung Chironomus. Archiv. f. Mikro. Anat. Bd. ii. 1866. 51. Lemoine, V. Recherches sur le developpement des Podurelles. Ass. Fra?i£. Avanc. Sci. Congres de la Rochelle. 1882. 52. Leuckart, R. Ueber die Micropyle und den feineren Bau der Schalenhaut bei den Insecteneiern. Archiv. f. Anat. u. Phys. 1855. 53. Melnikoff, N. Beitrage zur Embryonalentwicklung der In- secten. Archiv. f. Naturg. Bd. xxxv. 1869. 54. Metschnikoff, E. Untersuchungen iiber die Embryologie der Hemipteren. Zeitschr. f. Wiss. Zool. Bd. xvi. 1866. 400 INSECTA. 55. Metschnikoff, E. Embryologische Studien an Insecten. Zeitschr. f. Wiss. Zool. Bd. xvi. 1866. 56. Miall, L. C, and Denny, A. The structure and life-history of the Cockroach (Periplaneta orientalis). London, 1886. (The section on embryonic development by Nusbaum.) 57. Nusbaum, J. Vorl. Mittheilung liber die Chorda der Arthro- poden. Zool. A?iz. Jahrg. vi. 1883. 58. Nusbaum, J. Die Entwicklung der Keimblatter bei Meloe proscarabaeus. Biol. Centralbl. Bd. viii. 1888. 59. Nusbaum, J. Zur Frage der Segmentirung des Keimstreifs und der Bauchanhange der Insectenembryonen. Biol. Centralbl. Bd. ix. 1889. 60. ISusbaum, J. Zur Frage der Ruckenbildung bei den Insecten- embryonen. Biol. Centralbl. Bd. x. 1890. 61. Nusbaum, J. On the development of the efferent ducts of the genital glands in the Insecta (Polish, with resume in German, pp. 39-42). Kosmos, Lemberg. Jahrg. ix. 1884. 62. Nusbaum, J. Zur Entwicklungsgeschichte der Ausfiihrungsgange der Sexualdriisen bei den Insecten. Zool. Anz. Jahrg. v. 1882. 63. Nusbaum, J. The Embryology of Meloe' proscarabaeus, Marscham (Polish, with Latin explanation of plates). Kosmos, Lemberg. 1891. 64. Packard, A. S. Embryological Studies on Diplax, Perithemis, and the thysanurous genus Isotoma. Mem. Peabody Acad. Set. Vol. i. 1871. 65. Patten, W. The development of Phryganids, with a pre- liminary note on the development of Blatta germanica. Quart. Journ. Micro. Sci. Vol. xxiv. 1884. 66. Patten, W. Studies on the eyes of Arthropods. Develop- ment of the eyes of Vespa, with observations on the Ocelli of some Insects. Journ, Morphol, Boston, Vol. i. 67. Patten, W. Studies on the eyes of Arthropods. II. Eyes of Acilius. Journ. Morphol. Boston. Vol. ii. 1888. 68. Patten, W. On the origin of Vertebrates from Arachnids. Quart. Journ. Micro. Sci. (2). Vol. xxxi. 1890. 69. Pedaschenko, D. On the formation of Notonecta glauca (Russian). Revue Sci. Nat, St. Pvtersbourg. Annee i. 1891. 70. Platner, G. Die erste Entwicklung befruchteter und partheno- genetischer Eier von Liparis dispar. Biol. Centralbl. Bd. viii. 1888. LITERATURE. 401 71. Ritter, R. Die Entwicklung cler Geschlechtsorgane und des Darmes bei Chironomus. Zeitsclir. f. Wiss. Zool Bd. 1. 1890. 72. Ryder, J. A. The development of Anurida maritima Guerin. Amer. Natur. Vol. xx. 1886. 73. Schmidt, F. Die Bildung des Blastoderms und des Keimstreifs der Musciden. Sitz. Naturf. Ges. Dorpat. Bd. viii. 1889. 74. Schneider, A. Ueber die Entwicklung der Geschlechtsorgane der Insecten. Zool. Beitrdge, v. A. Schneider. Bd. i. 1883. 75. Selvatico, D. S. Sullo sviluppo embrionale dei Bombicini. Boll. Bachicoltura, Ann. viii. 1881. 76. Sommer, A. Ueber Macrotoma plumbea. Beitrdge zur Anatomie der Poduriden. Zeitschr. f. Wiss. Zool. Bd. xli. 1885. 77. Stuhlmann, F. Die Reifung des Arthropodeneies nach Beobach- tungen an Insecten, Spinnen, Myriopoden und Peripatus. Ber. Freib. Naturf.-Gesellsch. Bd. i. 1886. 78. Tichomiroff, A. Ueber die Entwicklungsgeschichte des Seiden- wurms. Zool, Anz. Jahrg. ii. 1879. 79. Tichomiroff, A. On the Ontogeny of the Bombyx mori within the egg (Russian). Izvyest. imp. Obshch, Lyubit. Estest. Antrop. i. Efhnog. Moscow. Tom. xxxii., 4. 1882. 80. Tichomiroff, A. Ueber die Entwicklung der Calandra granaria. Biol. Centralbl, Bd. x. 1890. 81. Tichomirowa, 0. S. Zur Embryologie von Chrysopa. Biol. Centralbl. Bd. x. 1890. 82. Uljaxix, V. Notes on the Ontogeny of the Poduridae (Russian). Isvyest. imp. Obshch. Lyubit. Estestv. Antrop. i. Efhnog. Moscoio. Tom. xvi. 1875. 83. Uljaxix, V. Sur le developpement des Podurelles. Archiv. Zool, Exper. Tom. iv. 1875. Tom. v. 1876. 84. Viallanes, H. Sur quelques points de l'histoire du developpe- ment embryonnaire de la Mante religieuse. Prelim, Conr. Rev. Biol, die Nord de la France. Tom. ii. 1889-1890. 85. Voeltzkow, A. Entwicklung im Ei von Musca vomitoria. Arb. Zool, Zoof. Inst. Wilrzburg. Bd. ix. 1889. 86. Voeltzkow, A. Melolontha vulgaris, ein Beitrag zur Entwick- lung im Ei bei Insecten. Arb. Zool. Zoot. Inst. Wiirzbwg. Bd. ix. 1889. 87. Weismann, A. Die Entwicklung der Dipteren im Ei, nach Beobachtungen an Chironomus sp., Musca vomitoria und Pulex canis. Zeitsclir. f. Wiss. Zool, Bd. xiii. 1865. 2d 402 INSECT A. 88. Weismann, A. Zur Embryologie der Insecten. Archiv. f. Anat. u. Physiol. 1864. 89. Weismann, A. Beitrage zur Kenntniss der ersten Entwicklungs- vorgange ini Insectenei. Beitriige z. Anat. u. Embryol. etc. Bonn, 1882. 90. Weismann, A., and Ischikawa, Ch. Ueber die Bildung der Richtungskorper bei tbieriscben Eiern. Bet: Naturf. Ges. Freiburg. Bd. iii. 1887. 91. Wheeler, W. M. On the Appendages of the first Abdominal Segment of the Embryo Cockroach (Blatta germanica). Proc. Wiscons. Acad. Set. Vol. viii. 1890. 92. Wheeler, W. M. Ueber driisenartige Gebilde im ersten Abdominalsegment der Hemipterenembryonen. Zool. Ait::. Jahrg. xii. 1889. 93. Wheeler, W. M. Ueber ein eigenthiimliches Organ im Locus- tidenembryo (Xiphidium ensiferum). Zool. Ariz. Jahrg. xiii. 1890. 94. Wheeler, W. M. Neuroblasts in the Arthropod Embryo. Journ. Morphol. Boston. Vol. iv. 1891. 95. Wheeler, W. M. The Embryology of Blatta germanica and Doryphora decemlineata. Journ. Morphol. Boston. Vol. iii. 1889. 96. Will, L. Zur Bildung des Eies und des Blastoderms bei den viviparen Aphiden. Arb. Zool. Zoot. Inst. Wilrzburg. Bd. vi. 1883. 97. Will, L. Entwicklungsgeschichte der viviparen Aphiden. Zool. Jakrb. Abtli. f. Anat. Bd. iii. 1888. 98. Witlaczil, Em. Entwicklungsgeschichte der Aphiden. Zeitsclir. f. Wiss. Zool. Bd. xl. 1884. Postembryonic Development. On pupal development cf. the older treatises of Swammerdam, Herold, and L. Agassiz. 99. Bemmelen, J. F. van. Ueber die Entwicklung von Farben und Adern auf den Schmetterlingsfliigeln. TijdscJirift der Nederlandsclie Dierliundige Vereeniging (2). Deel ii., Afl. iv. 1889. 100. Brauer, Fr. Die Zweifliigler des Kais. Museums zu Wien. III. Systemat. Studien auf Grundlage der Dipteren-Larven Denksclir. Acad. Wiss. Wien. Bd. xlvii. LITERATURE. 403 101. Brauer, Fr., and Redtenbacher, Jos. Ein Beitrag zur Ent- wickrung des Fliigelgeaders der Insecten. Zool. Ana. Jahrg. xi. 1888. 102. Dewitz, H. Beitrage zur Kenntniss der postembryonalen Gliedmaassenbildung bei den Insecten. Zeitschr. f. Wiss. Zool, Bd. xxx. Suppl. 1878. 103. Dewitz, H. Ueber Bau und Entwicklung des Stachels und der Legescheide einiger Hymenopteren und der griinen Heuschrecke. Zeitschr. f. Wiss. Zool, Bd. xxv. 1875. Bd. xxviii. 1877. 104. Dewitz, H. Ueber die Fliigelbildung bei Phryganiden und Lepidopteren. Berl. Ent. Zeitschr. Bd. xxv. 1881. 105. Fabre, L. L'liypermetamorphose et les moeurs des Meloides. Ann. Sci. Nat. (4). Vol. vii. 1857. 106. Frenzel, J. Einiges iiber den Mitteldarm der Insecten, sowie iiber Epitbelregeneration. Archie, f. Mikro. Anat. Bd. xxvi. 1885. 107. Gaxin, M. On the post-embryonic development of the Insecta (Russian). Warsaw, 1876. Syezd. Russk. Estestv. Vrachei v. Varshavye. Abstract by Hoyer. Jahresh. der Anat, und Phys. von Hoffmann und Schiralhe, Bd. v. 1876. See also Zeitschr. f. Wiss. Zool. Bd. xxviii. 1877. 108. Haase, E. Zur Entwicklung der Fliigelrippen der Schmetter- linge. Zool, Anz. Jahrg. xiv. 1891. 109. Hurst, H. The post-embryonic development of a gnat (Culex). Trans. Liverpool Biol. Soc. Vol. iv. 1890. 110. Jacobi, A. On the development of markings on the wings of the Lepidoptera (Russian). Protofo Zasyed, Obshch. Estestv. pri imp. Kazanck. Univers. 1888-89. 111. Kowalevsky, A. Beitrage zur nachembryonalen Entwicklung der Musciden. Zool. An::. Bd. viii. 1885. 112. Kowalevsky, A. Beitrage zur Kenntniss der nachembryonalen Entwicklung der Musciden. Theil. i. Zeitschr. f. Wiss. Zool, Bd. xlv. 1887. 113. Kunckel d'Herculais, J. Recherches sur l'organisation et le developpement des Volucelles. Paris, 1875. 114. Landois, H. Beitrage zur Entwicklungsgeschichte der Schmet- terlingsnugel in der Raupe und Puppe. Zeitschr. f. Wiss. Zool. Bd. xxi. 1871. 115. Lowne, B. Th. Anatomy, Physiology, Morphology, and Development of the Blow-Fly. London. Yol. i. 1890-92. Vol. ii. 1893-95. 404 INSECTA. 116. Metschnikoff, E. Untersuchungen iiber intracellular Ver- dauung bei wirbellosen Thieren. Arb. Zool. Inst. Wien. Bd. v. 1883. 117. Metschnikoff, E. Untersuchungen iiber die mesodermalen Phagocyten einiger "NYirbelthiere. Biol. Centralbl. Bd. iii. 1883. 118. Muller, E. Die Flugeladern der Schmetterlingspuppen. Kosmos. Jahrg. i. 1877. 119. Pancritius, P. Xotiz iiber Fliigelentwicklung bei den Insecten. Zool. An::. Jahrg. vii. 1884. 120. Pancritius, P. Beitrage zur Kenntniss der Fliigelentwicklung bei den Insecten. Inaug.-Diss. Konigsberg, 1884. 121. Rees, J. van. Beitrage zur Kenntniss der inneren Metamor- phose von Musca vomitoria. Zool. Jahrb. Abth. /. Anat. Bd. iii. 1888. 122. Rees, J. van. Over de post-embryonale ontwikkeling von Musca vomitoria. Maandbladvoor Natuurwetensckappen. Juli, 1885. 123. Rees, J. van. Over intra-cellulaire spijsverteering en over de beteekenis der witte bloedlichampjes. Maandblad voor Natuurweienschcuppen. Jaarg. xi. 1884. 124. Rehberg, A. Ueber die Entwickelung des Insectenniigels (in Blatta germanica). Marimwerder. 1886. 124a. Schaffer, C. Beitrage zur Histologic der Insecten. Zool. Jahrb. Abth. f. Anat. Bd. iii. 1889. 125. Schiemenz, P. Ueber das Herkommen des Futtersaftes und die Speicheldriisen der Biene etc. Zeitschr. f Wiss. ZooL Bd. xxxviii. 1883. 126. Semper, C. Ueber die Bildung der Fliigel, Schuppen und Haare bei den Lepidopteren. Zeitschr. f. Wiss. ZooL Bd. viii. 1857. 127. Verson, E. Der Schmetterlingsniigel und die sog. Imaginal- scheiben desselben. Zool. Am. Jahrg. xiii. 1890. 128. Viallanes, H. Recberches sur l'histologie des Insectes et sur les phenomenes histologiques qui accompagnent le developpe- ment post-embryonnaire de ces animaux. Ann. Sci. Nat. (6). Tom. xiv. 1882. 129. AVeismann, A. Die nachembryonale Entwicklung der Musciden nach Beobacbtungen an Musca vomitoria und Sarcophaga carnaria. Zeitschr. f. Wiss. Zool. Bd. xiv. 1864. 130. Weismann, A. Die Metamorphose von Corethra plumicornis. Zeitschr. f. Wiss. Zooln Bd. xvi. 1866. LITERATURE. 405 131. Wiblowiejsky, H. v. Ueber den Fettkbrper von Corethra phnnicornis unci seine Entwicklung. Zool. Anz. Jahrg. vi. 1881. Parthenogenesis and Heterogeny. 132. Adler, H. Ueber den Generationswechsel der Eichen-Gall- wespen. Zeitschr. f. Wiss. Zool Bd. xxxv. 1881. 133. Balbiani, E. G. Observations sur la reproduction du Phyllo- xera du Chene. Ann. Sci. Nat. (5). Vol. xix. 1874. 134. Blochmann, F. Ueber die Geschlechtsgeneration von Chermes abietis L. Biol Centralbl Bd. vii. 1887-88. 135. Blochmann, F. Ueber die regelmassigen Wandertmgen der Blattlause, specieli liber den Generatiqnscyclus von Chermes abietis L. Biol Centralbl Bd. ix. 1889. 136. Carriere, J. Die Reblaus. Biol Centralbl Bd. vii. 1888. 137. Dreyfuss, L. Ueber Phylloxerinen. Wiesbaden, 1889. 138. Grimm, 0. v. Ungeschlechtliche Fortpflanzung einer Chiro- nonius-Art und deren Entwicklung aus dem unbefruchteten Eie. Mem. Acad. Petersbourg. 1870. 139. Leuckart, R Die ungeschlechtliche Fortpflanzung der Ceci- domyialarven. Archiv. f. Naturg. 1865. 140. Leuckart, R. Zur Kenntniss des Generationswechsels und der Parthenogenese bei den Insecten. Frankfurt, 1858. 141. Lichtenstein, J. Zur Biologic der Gattung Phylloxera. Stettin. Entorn. Zeitung. Jahrg. xxxvi. 1875. 142. Siebold, C. Th. v. Beitrage zur Parthenogenesis der Arthro- poden. Leipzig, 1871. 143. Signoret, V. Phylloxera de la Yigne. Ann. Soc. Entomot. de France (4). Tom. ix. 1869. 144. Wagner, N. Beitrag zur Lehre von der Fortpflanzung der Insectenlarven. Zeitschr. f. Wiss. Zool Bd. xiii. 1863. General. 145. Brauer, Fr. Verwandlung der Insecten im Sinne der Descendenz-Theorie. Verb. Zool Bot. Ges. Wien. i. 1869. ii. 1870. 146. Brauer, Fr. Systematisch-zoologische Studien. Sitzungsber. Akad. Wiss. Wien. Bd. xci. 1885. 147. Carriere, J. Die Sehorgane der Thiere, Munchen und Leipzig, 1885. 148. Fernald, H. T. The Relationships of Arthropods. Johns Hopkins Univers. Stud. Biol Lab. Vol. iv. 1890. 406 INSECTA. 149. Graber, Y. Die Insecten. Miinchen, 1877. (Die Naturkrcifte. Theil. ii. Yergl. Lebens- und Entwicklungsgescbiclite cler Insecten. 1879.) 150. Grassi, B. I progenitori dei Miriapodi e degli Insetti. Memoria viii. Anatomia comparata dei Tisanuri e considerazioni generali sull' organisazione degli Insetti. Atti Acad. Lincei. (4). Tom. iv. 1888. See also Archiv. Ital. Biol. Tom. xi. 1889. 151. Grenacher, H. Untersucbungen iiber das Sehorgan der Artbropoden, insbesondere der Spinnen, Insecten und Crustaceen. Gottingen, 1879. 152. Haase, E. Die Yorfahren der Insecten. AbJi. Ges. Isis. Dresden. Bd. xi. 1887. 153. Haase, E. Die Abdominalanbange der Insecten mit Bertick- sichtigung der Myriopoden. Morph. Jcdirb. Bd. xv. 1890. 154. Kennel, J. v. Die Ableitung zunachst der sog. einfacben Augen der Artbropoden, namlich der " Stemmata " der Insectenlarven, Spinnen, Scorpioniden etc. von Angen der Anneliden. Sitzungsber. Nat, Ges. Dor pat. Bd. viii. 1889. 155. Iyorschelt, -E. Ueber die Entstebnng und Bedeutung der versch. Zellenelemente des Insectenovariums. Zeitschr. f. Wiss. Zool, Bd. xliii. 1886. 156. Lubbock, J. Origin and Metamorphoses of Insects. Nature Series. London, 1883. 157. Mayer, P. Ontogenie und Pbylogenie der Insecten. Jen, Zeitschr. f. Natunciss. Bd. x. 1876. 158. Muller, F. Beitrage zur Kenntniss der Termiten. Jen. Zeitschr. f. Natunciss. Bd. vii. 1873. 159. Nassonow, X. Welche Insectenorgane -diirften bomolog den Segmentorganen der Wiirmer zu balten sein1? Biol. Centralbl. Bd. vi. 1886. 160. Packard, A. S. Guide to the Study of Insects, etc. New York, 1889. 161. Palmen, J. A. Zur Morpbologie des Tracbeensystems. Helsing- fors, 1877. 162. Palm£n, J. A. Zur vergleichenden Anatomie der Sexual- organe bei den Insecten. Morph. Jahrb. Bd. ix. 1883. 163. Palmen, J. A. Ueber paarige Ausfiihrungsgange der Gesch- lechtsorgane bei Insecten. Eine Monographische Unter- suchung. Helsincjfors, 1884. 164. Ratzeburg, J. Th. C. Die Forstinsecten. Berlin, 1837-1844. LITERATURE. 407 165. Redtenbacher, Jos. Vergleichende Studien iiber das Fliigel- geader der Insecten. ^4???*. Hofmus. Wien. Bd. i. 1886. 166. "Westwood, J. 0. An introduction to the modern classification of Insects. London, 1839-1840. APPENDIX TO LITERATURE ON INSECT A. The Student will do well to consult Packard's Text-book of Entomology (1898), which deals most thoroughly with the mor- phology and ontogeny of the Insecta. I. Bruel, L. Anatomie und Entwicklungsgeschichte der Gesch- lechtsausfiihrwege sammt annexen von Calliphora erythro- cephala. ZooL Jahrb. (Anat.). Bd. x. 1897. II. Carriere, J. Die Entwicklungsgeschichte der Mauerbiene (Chalicodoma muraria) im Ei ; herausg. u. vollendet von Otto Burger. Nova Acta Abh. d. k. Caes. Leop.-Akad. Bd. xlix. 1897. III. Cholodkovsky, 1ST. A. Beitnige zu Monographic der Coni- feren-Lause. Horae Soc. ent. Ross. xxx. IV. Cholodkovsky, 1ST. A. Contributions a la theorie du meso- derme et de la metamerie. Congr. Zool. ii., 1. 1892. V. Cholodkovsky, X. A. Embryonic development of Phyllo- dromia germanica (Russian). Trud. St. Petersburg est. xxii. VI. Cholodkovsky, X. A. On the morphology and phylogeny of Insects. Ann. Nat. Hist. (6). Vol. x. VII. Cholodkovsky, X. A. Zur Biologie der Larchen-Chermes- Arten. Zool. Am. Jahrg. xix., pp. 37-40. Zur Kenntniss der auf Fichte (Picea excelsa) lebenden Lachnus-Arten. T. c. pp. 145-150. VIII. Claypole, A. M. Some points on cleavage among Arthro- pods. Trans. Amer. Micro. Soc. Vol. xix. 1897. IX. Graber, V. Beitrage zur vergleichenden Embryologie der Insecten. Denkschr. Akad. Wien. Bd. lviii. IXa. Graber, V. Ueber die Morphologische Bedentung der ven- tralen abdominalanhiinge der Insekten-Embryonen. Morph. Jahrb. Bd. xvii. 1892. X. Heider, K. 1st der Keimblatterlehre erschiittert 1 Zool. Centralbl. Bd. iv. 1897. XI. Hexking, H. Untersuchungen iiber die ersten Entwicklungs- vorgiinge in den Eiern der Insekten. Theil. i., ii., iii. Zeitsch. f. Wiss. Zool. Bd. xlix., Ii., liv. 408 INSECTA. XII. Henneguy, L. F. Contributions a l'ernbryogenie des Chal- cidiens. Corrupt. Rend. Acad. Set. Paris. Tom. cxiv. 1892. (Translated in Ann. Nat. Hist. (6). Vol. x.) XIII. Heymons, R. Die Embryonalentwicklung von Dermapteren und Orthopteren unter besonderer Beriicksichtigung der Keimblatterbildung. Monographisch bearbeitet. Jenaf 1895. XIV. Heymons, R. Die Entwicklung der weiblichen Geschlecht- organe von Pbyllodromia germanica. Zeitschr. f. Wiss. Zool. Ed. liii. 1897. XV. Heymons, R. Die Fortpflanzung der Ohrwiirmer. Sitz- ungsber. Akad. Wiss. Berlin. 1894. XVI. Heymons, R. Entwicklungsgeschichte untersuchungen an Lepisma saccharina. Zeitschr. f. Wiss. Zool. Bd. lxii. 1897. XVII. Heymons, R. Fliigelbildung bei der Larve von Tenebrio molitor. Sitzungsber. Ges. naturf. Berlin. 1896. XVIII. Heymons, R. Grundziige der Entwicklung und der Korper- baues von Odonaten und Ephemeriden. Abh. Akad. Berlin. 1896. XIX. Heymons, R. Ueber die Bildung der Keimblatter bei den Insekten. Mit. Akad-. Berlin. 1894. XX. Heymons, R. Ueber die Bildung und den Bau des Darm- kanal bei niederen Insekten. Sitzungsber. Ges. naturf. Berlin. 1897. XXI. Heymons, R. Ueber die Fortflanzung und Entwicklungs- gescbicbte der Ephemera vulgata. Sitzungsber. Ges. naturf. Berlin. 1896. XXII. Heymons, R. Zur Morphologie der Abdorninalanhange bei den Insekten. Morpli. Jahrb. Bd. xxiv. 1896. And Biol. Centralbl. Bd. xvi. 1897. XXIII. Hyatt, A., and Arms, J. M. The meaning of Metamor- phosis. Nat. Sci. Vol. viii. 1896. XXIV. Karawaiew, W. Die Xachembryonale Entwicklung von Lasius flavus. Zeitschr. f. Wiss. Zool. Bd. lxiv. 1898. XXV. Knowa, H. The development of a Termite. Johns Hopkins Univ. Circ. Vol. xv. And Ann. Nat. Hist. (6). Vol. xviii. XXVI. Korotneff, A. Zur Entwicklung der Mitteldarmes bei den Arthropoden. Biol. Centralbl. Bd. xiv. LITERATURE. 409 XXX. XXXI. XXVII. Krassilstschik, J. Zur Entwicklungsgeschichte der Phytophthires. Ueber Viviparitat mit geschlechtlicher Fortpflanzung bei den Cocciden. Zool. An::. Jahrg. xvi. XXVIII. Kulagin, X. Beitrage zur Kenntniss der Entwicklungs- geschichte von Platygaster. Zeitschr. f. Wiss. Zool. Bd. lxiii. 1897. XXIX. Lecaillon, A. Recherches sur l'oeuf et sur le developpe- ment embryonnaire de quelques Chrysomelides. These* presentees a la Faculte ties Sciences de Paris. Ser. A. 1898. Also Zool, Centralbl. Bd. v. 1898. Lemoine, V. Etude comparee du developpement de l'ceuf cbez Puceron vivipare et ovipare. Bull Soc. ent. France. 1893. Lemoine, V. Etude comparee du developpement de l'oeuf dans la forme agame aptere dans la forme agame ailee et dans la forme sexuee du Phylloxera. Zool. Anz. Jahrg. xvi., pp. 140-142 and 145-149. XXXII. Lemoine, V. Xote complementaire sur l'ceuf du Phyl- loxera agame aptere. T. c. pp. 247 and 248. XXXIII. Marchal, P. La dissociation de l'ceuf en un grand nombre d'individus distincts et le cycle evolutif chez l'Encyrtus. Compt Rend. Acad. Sci. Pans. Tom. cxxvi. 1898. XXXIV. Mobusz, A. Ueber den Darmkanal der Anthrenus- Larve nebst Bemerkungen zur Epithelregeneration. Archiv. f. Naturges. Jahrg. lxiii. 1897. XXXV. Xussbaum, M. Zur Parthenogenese bei den Schmetter- lingen. Archiv. f. Mikro. Anal Bd. liii. 1898. XXXVI. Petrunkewitsch, A. Ueber die Entwicklung des Herzen bei Agelastica alni. Zool. Anz. Jahrg. xxi. 1898. XXXVII. Rengel, C. Ueber die Veranderungen des Darmepithels bei Tenebris molitor wahrend der Aletamorphose. Zeitschr. f. Wiss. Zool, lxii. 1897. XXXVIII. Saint-Hilaire, K. K. On the formation of the egg in Dytiscus. Compt. Rend. Soc, St. Petersh. 1895. (Russian, with resume in German : Ueber die Enstehung des Eies bei Dytiscus.) XXXIX. Tichomirow, A. Aus der Entwicklungsgeschichte der Insecten. Festschr. Leuckart. 1894. pp. 337-346. 410 INSECT A. XL. Uzel, H. Beitriige zur Entwicklungsgeschichte von Cam- podea staphylinus. ZooL Anz. Jahrg. xx. 1897. XLI. Verson, E., and Bisson, E. Die postembryonal Ent- wicklung der AusfuhrungSgange und der Nebendriisen beim mannlichen Geschleclitsapparat von Bombyx niori. Zeitschr. f. Wiss. Zool. Bd. lxi., pp. 318-337. XLII. Verson, E., and Bisson, E. Die postembryonal Ent- wicklung der Ausfiihrungsgange und der Nebendriisen beim Aveiblicben Gescblechtsapparat von Bombyx mori. T. c. pp. 660-694. XLIII. Viallanes, H. Sur quelques points de l'Histoire du Developpement Embryonnaire de la ]\Iantis Religiosa. Ann. Sci. Nat. (7). Tom. xi. 1891. XLIV. "Wheeler, W. j\1. A contribution to Insect Embryology. Journ. Morph. Vol. viii. 1893. XLV. Willey, A. Trophoblast and Serosa. A contribution to the morphology of the embryonic membranes of Insects. Quart. Journ. Micro. Sci. Vol. xli. 1899. CHAPTER XXVIII. GENERAL CONSIDERATIONS ON THE ARTHROPODA. Ik reviewing once more the ontogeny of the various divisions of the Arthropoda, we are specially struck by the uniformity of character found among them. In the constitution of the eggs in the cleavage, the method of formation of the germ-layers and the shape of the embryo, there are so many points of resemblance that we are justified by ontogeny in regarding the Arthropoda as phyletically distinct, i.e., as forming a natural group, even though, as will be shown below, the common stock divides near its root and gives rise to three great branches known to us under the names of the Crustacea, the Arach- nida and the Myriopoda-Insecta.* The eggs of the Arthropoda are as a rule distinguished by the large quantity of yolk contained in them, and the equal distribution of the latter (centrolecithal eggs of the Arthropoda). The typical method of cleavage in the Arthropoda is a superficial one, which has developed from total and equal cleavage, as may be seen from the ontogeny of various Crustacea, f We also see that the Arthropodan eggs, in those cases in which the food-yolk has secondarily degenerated, undergo total cleavage (Clacodera, Peripatus Edwardsii,\ parasitic Insects). In other cases the total cleavage perhaps still represents a primitive condition, e.g., in Branchipus. In a few Arthropoda, the egg appears to be telolecithal, and the cleavage is at first restricted to only a small part of the egg (e.g., in Mysis, Cuma, some Isopoda, and the Scorpiones). This apparently different method of cleavage is, however, to be traced back to superficial cleavage. * [See Natural Science, Vol. x. "Are the Arthropoda a Natural Group?" — Ei>.] f The statements in this chapter are based upon the facts already given in connection with the different divisions of the Arthropoda. The reader must refer for these to the preceding chapters. J [See footnote, p. 165. — Ei>.] 412 GENERAL CONSIDERATIONS ON Superficial cleavage, as a rule, occurs only in the Arthropoda. Where other forms, e.g., Renilla, Glavularia (Vol. i., p. 76) show a similar method in the first stages, this does not lead to the same results as typical superficial cleavage, viz., to a unilaminar blastoderm covering the whole surface of the egg with a uniform layer and an accumulation of food-yolk filling the cleavage-cavity.* The formation of the germ-layers is introduced by gastrulation, which, in many cases, is of the invagination-type (Moina, Lucifer, Astacus, Peripatus, Hydropliilus), in others, on the contrary, gastru- lation is replaced by a solid ingrowth of cells (Ligia, Limulus, Scorpiones, Araneae, Myriopoda). The position of the blastopore varies in the different groups. As a rule, the blastopore corresponds to the ventral side of the body. In Peripatus and the Insecta, the blastopore is an exceedingly long slit, the anterior end of which corresponds in position to the mouth, and the posterior end to the anus (Figs. 99, 134, and 145). In the Crustacea, on the contrary, the blastopore is said to belong to the posterior end of the germ-band, and to correspond more or less in position with the later anal aperture. The accounts given of the Arachnida seem to indicate that, in position, the blastopore may be related to the anus. The act of gastrulation leads to the breaking up of the common rudiment of the entoderm and the mesoderm. The rudiment of the mesoderm in the Arthropoda is always multicellular, except perhaps in a few quite isolated cases, such as Cetocliilus. In the Insecta, the formation of the mesoderm may be traced back to a folding of the lateral diverticula of the archenteron (Figs. 154 and 155, p. 314). The processes that take place in Peripatus may perhaps be interpreted in the same way, although in this form we are inclined to assume, in agreement with the Annelida, the development of two mesoderm- bands advancing from behind forward through the multiplication of cells. The facts as yet known of Peripatus seem rather to support this last view. The question whether the condition found in the Insecta (i.e., the rise of the mesoderm from the archenteron through folding) represents a primitive or a derived condition, is connected with the as yet unsolved problem of the first (phylogenetic) rise of the mesoderm. In the Crustacea, the mesoderm arises in the form of a growth at the lips of the blastopore. The same is most probably the case in * [For a comparison of the cleavage and formation of the germ-layers in the Arthropoda, see Wagner (No. X.). — Ed.] THE ARTHROPODA. 413 the Arachnida, In the latter, the mesoderm runs forward from the point of origin in the form of two bands (mesoderm-hands) on either side of the middle line. These two bands are also found in PeripatuS, the Myriopoda, and the Insecta, as well as apparently in the Pantopoda, while, in the Crustacea, the arrangement of the mesoderm is less regular. Some Crustacea, however (Branch) pics, Cymothoe), show a similar regular form of mesoderm-rudiment. The paired rudiment of the mesoderm breaks up into segmental divisions in a somewhat similar way in all Arthropoda. These divisions are the primitive segments (mesodermal somites), which either become hollow, and are then known as coelomic sacs, or are not thus modified, but soon break up into mesenchymatous tissue. This latter is the case in most Crustacea, in which coelomic sacs are rarely to be found, but the former condition occurs in the Xiphosura, Arachnida, Pantopoda, Onychophora, Myriopoda, and Insecta. Although the primitive segments have as a rule a very similar fate, and undergo similar modifications in all Arthropoda, certain differ- ences are to be found in the various classes in the size attained by them, and in the time at which their further differentiation begins. The most primitive condition is exhibited in Peripatus, in which the primitive segments in their large size resemble those of the Annelida (Fig. 100, p. 200). The Myriopoda and the Orthoptera follow next in the conspicuous development of the primitive segments within the germ-band (Figs.' 168 and 169 A, p. 343), while in the other Insecta the coelomic sacs are from the first small, a considerable part of the mesoderm being, as a rule, excluded from participation in the formation of these sacs (Fig. 158, p. 321). In the Crustacea, the development of the coelomic sacs is almost entirely suppressed. The Arachnida, on the contrary, which in many other respects appear as a modified group, are distinguished by the fact that in them the coelomic sacs are unusually large, and even in the later stages of embryonic development (at the time when the heart is forming) extend almost to the dorsal middle line (Fig. 45 and 46, p. 88). The appearance and further development of the organs in the different groups of the Arthropoda show remarkable and important agreement. In the case of the nervous system it has been proved that an invaginate middle strand and two lateral strands almost universally take part in the formation of the ventral chain of ganglia. The fibrous substance appears on the inner surface of the ganglionic rudi- ments, and is only later taken into the latter, a process which must 414 GENERAL CONSIDERATIONS ON be regarded as specially primitive, and is to be found taking place in a somewhat similar manner in all the various groups. "While the formation of the chain of ganglia takes place, as a rule, by a process of delamination from temporary ectodermal grooves which afterwards vanish, permanent invaginations occur which take part in the formation of the brain, leading no doubt chiefly to the formation of the optic ganglia. The appearance of these more or less extensive depressions, known as cephalic pits, is specially character- istic of the various divisions of the Arthropoda (Perijxittis, the Myriopoda, the Insecta, Limulus, the Arachnida). In Peripatus, indeed, another significance has been attributed to these depressions, and it is doubtful whether they participate in the formation of the brain. The depressions in the cephalic region in Peripatus corres- pond to similar pit-like invaginations arranged in pairs which recur in each of , I- . the trunk - seg - ments. The na- ture of these remarkable struc- tures, which are very characteris- tic of Peripatus, has not yet been established, but similar depres- sions have been described in the Myriopoda and the Pantopoda. The development of the eyes may be closely connected with the cephalic pits just mentioned which, as it appears, chiefly give rise to the formation of the optic ganglia (Scorpiones, Araneae). How- ever much the permanent Arthropod eyes vary with regard to structure, they may, in the first instance, be traced back to pit-like depressions of the ectoderm, and in explaining them we must start from such simple eyes as those occurring in the larvae of Insects and in many Myriopoda. This simplest form of Arthropodan eye, the ocellus (Fig. 195), consists of a depression of the hypodermis, the cells of which have become differentiated into the so-called vitreous body (gl), and retinal cells (rt), secreting rods. The unilaminar character of the hypodermis has, however, been retained in this simple eye, so Fig. 195. — Section through the ocellus of a Dyiiscus larva (after Grenacher). cli, chitinous covering of the body ; ;(/, vitreous body; hyp, hypodermis ; I, lens; n, optic nerve; rt, retina; st, roils. THE ARTHROPODA. 415 that it appears as a mere continuation of the hypodermal layer ■(Fig. 195, hyp, gl, rt). Over the eye lies the lens which has arisen by the thickening of the outer chitinous covering of the body, and is secreted by the hypodermis (lentigen or vitreous body-layer). From such a simply constructed eye we have to derive the complicated eyes found among the Arthropoda, but in so doing we must dis- tinguish sharply between the various phyletic ontogenetic series of the Arthropoda, and we must remember that it is not possible to regard as directly related one to another the various forms of com- pound eyes found in the separate divisions such as the Crustacea and the Insecta, although the eyes in these groups are very similar in structure. 7KO; LI . w* Fig. 196.— Three ommatidia of the lateral eye of Limulus (after Watase). In A the retinula is supposed to be cut through medianly, in B and C it is retained whole, c, central ganglion- cells; eh, chitinous covering; hyp, hypodermis; I, lenticular sphere; mes, mesodermal tissue; n, nerve; rh, rhabdom ; rt, retinula. It may appear at first sight unreasonable not to regard the compound eyes of the Crustacea and the Insecta, which are so remarkably similar in organisation, as directly related one to the other, but when the phylogenetic course of develop- ment of the two divisions is taken intoaccount weshall have to takeup this position. It can only be assumed that the development of compound eyes is a character of the Arthropodan organisation, and that it takes place in the different divisions (Crustacea, Arachnida, Myriopoda, and Insecta) independently, and yet may lead, as in the Crustacea and the Insecta, to almost the same result. The eyes of Peripatus differ altogether in structure from those of other Arthropoda. The eyes of this form also, indeed, originate as 416 GENERAL CONSIDERATIONS ON 3. 0 depressions, which, however, close to form vesicles and become separated from the hypodermis. The lens is secreted within the optic vesicle. The eyes of Peripatus thus, in their onto- geny, pass through the stage of the sim- plest Arthropod eye, but then rise to a higher level — K, than that attained by the latter, and can far better be compared with the higher forms of eye found in the Annelida. In any case, Ave do not recognise in them the Arthropodan type of eye. The facet -eyes of the Insecta must be regarded from a together of simple ocelli in the way already indicated in the Myriopoda. The latter group, in the simplest cases, have only a few ocelli on each side (Scolopendra four), but their number may in- crease (Lithobms, Julus, thirty to forty on each side), and in some forms (Scutigera) there may even be as many as 200 ocelli on each side, which, by their close approxima- as arising massing Fig. 107.— A-D, diagrams illustrating the development of an ommatidium from a depression of the hypodermis. D represents an ommatidium from the compound eye of an Ampliipod (Talor- ehestia, after Watase). c, central cell; ch, chitinous covering of the head ; h, hypodermis ; k, crystalline cone ; Jtz, crystalline cone-cells; I, lens; lg, lentigen cells; n, nerve; rh, rhabdom ; ft, retinula cells. THE ARTHROPODA. 417 tion, recall the appearance of facet-eyes, although a group of eyes does not possess the true structure of the latter. Each ocellus in this way becomes a single ommatidium of the facet-eye. The diminution in number of its elements which it then undergoes, and the simultaneous formation of the rhabdoms are consequences of the subordination and loss of individuality of the originally distinct single eyes on becoming merged in the complex eye, of which organ they now form a part. Attempts have been made to trace back the facet-eye to the more primitive form from which it originated, by regarding the ommatidia which, according to the view mentioned above, were derived from single ocelli, as sim- C. J8. pie hypodermal de- (fl pressions which, in consequence of the length of the omma- tidia, became very deep (Fig. 197 D). In making such an attempt to explain the structure of the ommatidia it is best to start from a de- pression of the hypodermis which corresponds to a simplified ocellus (Fig. 197 A). As the depression deepens and, instead of rods, rhabdoms begin to form in the retinal cells, this eye reaches a grade of development (Fig. 197 B) essentially equivalent to that of an ommatidium in the lateral eyes of Limulus (Fig. 196). The lateral eye of Limulus is composed of a number of single eyes formed of only a few cells (Fig. 196). These unilaminar eyes are quite continuous with the hypodermis, but already show rhabdom- formation (Fig. 196 A, rh). It is indeed not certain whether the eyes of Limulus should really be regarded as primitive eyes, or as degenerate forms of the compound eye ; in any case, however, we can imagine that the higher facet-eyes passed through a similar stage (Fig. 197 B). When the depression deepens, another series of hypodermal cells 2 E Fig. 19S.— Diagrams illustrating compound eyes in longi- tudinal section. A, Limulus; B, a larva of Agrion; C, Branchipus (after Watase). The thick black line represents the hypodermis, and each of the depressions formed in it represents an ommatidium. 418 GENERAL CONSIDERATIONS ON may be drawn into the formation of the eye (Fig. 197 C), these representing the crystalline cone-cells (kz) of the ommatidium. A series of lentigen cells may also be utilised in the formation of the eye (Fig. 197 C, l.g). The further deepening of the optic pit, and the great lengthening of the cells lead finally to the form of the ommatidium (Fig. 197 D). The hypodermal cells, the lentigen cells, the crystalline cone-cells, and the retinal cells thus appear as a uni- larninar layer of long cells penetrating far down, and having the same arrangement as in the simple ocellus (Fig. 195). The lumen here, however, is not open as in the ocellus, but filled by the mass of the crystalline cone and rhabdoms, but this does not constitute an essential difference between the two eyes. The grouping together in larger or smaller numbers of these single eyes which arise as simple depressions of the hypodermis is elucidated by Fig. 198, which at the same time represents the arrangement of the ommatidia on a convex base usual in most facet-eyes, and determined by the functional re- quirements of the eye. The method of composition of the facet-ej'e here described is essentially in keeping with the view long ago maintained by Gkenachek. This author starts from a simple eye consisting of few elements, such as is represented by an ommatidium of an acone facet-eye of the Tipulidae, and derives the facet- eye through the increase in number of these eyes, and the ocellus through the multiplication of the elements with the retention of the single lens. In the simple eye, which here forms the starting-point, we have an ocellus of specially simple structure. It has already been stated that the compound eye of the Crustacea must be regarded as belonging to another ontogenetic series. It will therefore not be a matter of surprise to find that it deviates in many ways from the above in its development. The character of the compound eye is, in the Crustacea, always preserved. In some cases, e.g., in the Isopoda, it might appear as if we had before us transition stages between the simple and the compound eye, but it is more than probable that, in this branch of the Crustacea, we have to do merely with a simplified form of the facet-eye. This view of the Isopodan eye was adopted long ago by Gkenachek, who attempted to solve the question as to how the very simple eye of the Isopoda was related to other Arthropodan eyes, by maintaining that the former was to be regarded as a compound eye in consequence of its possessing a double crystalline cone and a retinula forming rhabdoms and divided into seven parts. It cannot therefore be doubted that, in the Isopodan eye, which is not unlike a group of single eyes, we have a secondary form, and this is in itself very probable, inasmuch as the Isopoda are, in many respects, a highly modified group of the Crustacea. A degeneration of the facet-eye, which was originally stalked in the Malacostraca, has taken place in any case in this order. THE ARTHROPOPA. 419 We have no indication of the manner in which the facet-eye has arisen in the series of the Crustacea. Xone the less must we consider that this eye, which closely resembles the facet-eyes of the Insecta, arose in the same way as the latter. Any deviations that may occur, such as the presence of another cell-layer in the ommatidium (Fig. 197 D, l.g), are to be explained simply by the inclusion of another row of cells in the hypodermal depression, as already shown. The structure, development, and relations of the unpaired median eyes in the Crustacea are still little understood. It has recently been asserted that they arise by inversion (Claus, No. 3), and since this method of formation is characteristic of some of the eyes found in Limulus and the Arachnida, relations between the median Crustacean eye and the median eye of Limulus and the Scorpiones, as well as the so-called principal eyes of the Araneae, are suggested. The eyes of the Arachnida belong to a third ontogenetic series. They have only one lens, and are thus devoid of the characteristic feature of facet-eyes, but in the eyes of Scorpio we find a grouping of the cells into retinulae and the formation of rhabdoms within these latter, and in this respect they may claim to be compound eyes. We considered ourselves justified in explaining the common lens as having arisen by the flowing together of distinct corneal lenses (p. 71, etc.), and find in the lateral eyes of Limulus, which also show rhabdom- formation, an indication of such a fusing of the lenses. We tried further to show it to be probable that the eyes of the Araneae, which in their present form appear to be simple eyes, are to be derived from compound eyes, this origin being still indicated in their development and their structure. It is highly probable also that the compound eyes of the Arachnida, like those of the Insecta, arose through the accumulation of simple hypodermal depressions resembling ocelli. When we turn to the ontogenetic formation of the Arthropod eyes, we find that the simple forms arise as pit-like depressions of the ectoderm. In the higher forms, i.e., in the compound eyes, this primitive method of formation is obliterated. The single eyes here arise merely through the differentiation of a cell-layer without special depressions. Where such a depression is found in the development of a compound eye, it leads to the formation of the eye as one whole. In this last process, as in the differentiation of the single eyes out of a multilaminar cell-layer, we have secondary phenomena representing a simplified method of formation of the 420 GENERAL CONSIDERATIONS ON compound eye. It should also be noted that the ontogeny of the Arthropod eye is as yet not satisfactorily explained. The respiratory organs of the Arthropoda must be dealt with separately, according to the different phylogenetic series into "which they are to be divided. Since we derive the Arthropoda from forms which live in water, it appears to us that the most primitive form of respiratory organ must have been a tubular or leaf-like outgrowth of the body-surface. Such a simple form of respiratory organ is found in the "ills met with as branchial tubes in the Annelida and O Crustacea. These branchial structures appear, as a rule, as append- ages of the extremities. The gills of Limulus are also leaf-like appendages of the abdominal limbs. From these we have to derive the lung-sacs of the Arachnida (Scorpiones, Araneae), a fact indicated by the method of development of these latter. In the transformation of gills into lungs Ave recognise an adaptation to life on land. When this adaptation goes further, it leads to the development of unbranched tracheal tufts (Araneae) which finally ramify in a dendriform manner and develop a spiral filament (Pseudoscorpiones, Solifugae). In this way is attained the same type of tracheal system as is produced in different manner in other groups of Arthropods otherwise very far removed from the above, viz., Peripatus, the Myriopoda, and the Insecta. In the forms which were the starting-point of this last series, the tracheae appeared as depressions of the body-surface, which at first were irregularly distributed over the body {Peripatus), but later attained to definite segmental arrangement. The tracheae in the Myriopoda and the Insecta arose as such segmentally-arranged depressions. The branches of the tracheal system are formed by the splitting and branching of the original invaginations. In the Insecta these tracheal rudiments appear very early, in the Myriopoda, on the contrary, much later, and, as Peripatus in this way resembles the Myriopoda, this late appearance of the tracheal rudiments has been regarded as an indication of their having been recently accpuired. The similarity in structure between the tracheae of the Arachnida and those of the Myriopoda and Insecta is remarkable. The presence of the spiral filament in these two forms of tracheae, which must be regarded as having arisen independently in the two groups, is specially striking, but this feature loses its value as an indication of a common origin when it is seen that such a spiral thread also occurs in other tubes lined by a chitinous intima, e.g., the efferent ducts of glands (salivary and spinning glands of the Insecta) and the vas deferens of the Cytheridae, p. 335). THE ARTIIROPODA. 121 The so-called closed tracheal system of many aquatic larvae, e.< 03 CO H o g4-a ca 03 03 o ea 03 rS cc ^4—1 CO cm g .a -a ■ cS ci c*h _j o ^ s r CO -a O g '_ o -m * ce — , — 1 '3 "-1 —.3d < — +a 5 s ca g ^2 "Si ■j. 0 PS co o 3 -M c CO .- .a o 1 — '0 la 0 o E o 03 £ C3 03 PH.a 03 ^3 CO += H5 a 03 eS 03 03 -M co- CO co a £j M ,u o 03 CO o 03 03 o ra £ a so 03 o so « so 5 E CO 5 Pi eS •-■ 3 'o £ ■>« 03 5 03 < CO « pa r* ° CCy ^-J •3"B 2 § •so - 1 a 03 . o t-o a " ;_< o S i go ea . fc 03 "S co =0 — M Cm O * r-1 0 ^ S 03 SO 03 ci CO ~ 'd a ~ +J 0 1 LITERATURE. Further information as to literature will be found at the ends of the chapters dealing with the different divisions of the Arthropoda. 1. Anonymous. Bilden die Arthropoden eine natiirliche Gruppe? Kosmos. Bd. xiii. 1883. 2. Beddard, F. E. On the possible origin of the Malpighian Tubules in the Arthropoda. Ann. MaE. Insecta, 364. Pentastoraidae, 137. Frauenfeld, G. von. Acarina, 104. Frenzel, J. Insecta, 385. FtJRSTENBERG, M. H. F. Acarina, 126. G. Gaffron, E. Peripatus, 204, 209- 211. Ganin, M. Insecta, 309,311, 364, 365, 369, 375, 380, 385. Scorpiones, 5, 9. Gaubert, P. Arachnida, 119. Gegenbauer, C. Insecta, 393. Gkrlach, A. C. Pentastoinidac, 137. Goodrich, E. S. Arthropoda, 421, 429. Goossens, Th. Insecta, 29S. Graber, V. Insecta, 264, 266, 270- 274, 278, 282, 284- 308, 311-320, 324- 328, 336, 337, 312, 360, 378. Myriopoda, 248. Grassi, B. Insecta, 263, 268, 287, 301, 316, 323, 327, 334, 337, 340-342, 373, 394. Pedipalpi, 27. Grenacher, H. Araneae, 64, 69, 73. Arthropoda, 414, 418. Insecta, 332, 395. Myriopoda, 242. Grimm, O. von. Insecta, 405. Gudden, R. Acarina, 105. II. Haask, E. Arachnida, 118. Insecta. 299, 300, 350, 351, 373, 392,394. Myriopoda, 255. Haecker, V. Insecta, 351. Hagbn, H. A. Insecta, 359. Halle n, G. Acarina, 94, 109. Hallez, P. Insecta, 398. Hansen, H. J. Arthropoda, 431. Opiliones, 36. Hansen, H. J. and St'lRENSEN, W. Pedipalpi, 27. Hatschek, B. Insecta, 273, 302, 323, 324, 338. Heathcote, F. G. Myriopoda, 220-223, 235-252, 351. Heider, K. Insecta, 263-266, 270, 271, 287, 2S8, 290- 295, 301, 302, 306, 310, 312, 316-318, 340, 321-328, 342, 351. Henking, H. Acarina, 97. Insecta, 262, 263. Opiliones, 32. Henneguy, L. F. Insecta, 363, 366. Herbst, C. Myriopoda, 252. Herold, M. Araneae, 45. Hertwig, O. and R. Insecta, 317, 318, 351. Heymons, R. Insecta, 300, 304, 317, 320, 336, 340, 341- 352. Myriopoda, 223, 239, 241, 247, 253. Hodge, G. Pantopoda, 161. HOEK, P. P. C. Pantopoda, 139, 145, 148, 151, 153, 159. Hoyle, W. E. Pentastoinidac, 135. Hurst, H. Insecta, 403. Hl'TTON, F. W. Peripatus, 164-166. Button, T. Solifugae, 34. Hyatt, A. and Arms, J. M. Insecta, 403. AUTHORS INDEX. 439 IHI.E, J. E. W. Pantopoda, 160. J. Jacobi, A. Insecta, 403. Jacquart, H. Pentastomidae, 135. Jawokowski, A. Araneae, 51, 79-81, 111, 114. Arthropoda, 431. JOKDAN, K. Insecta, 279. Joukpain, S. Acarina, 128. K. Kapyi, H. Insecta, 399. Karawaiew, W. Insecta, 382, 385. Kaufmann, A. Cytlieridae, 335. Kaufmann, J. Tardigrada, 162. Kennel, J. von. Araneae, 69. Insecta, 395. Peripatus, 164-173, 178-201, 206-211. Tardigrada, 163. Kenyon, F. C. Myriopoda, 257. Kingsley, J. S. Araneae, 42. Arthropoda, 426. Myriopoda, 257. Kishinouye, K. Araneae, 37, 41-45, 51, 62-64, 67, 74, 81, 84-87, 93. Arthropoda, 429. Klafalek, F. Insecta, 388. Klein en berg, N. Lumhricus, 91. Peripatus, 195. Knatz, L. Insecta, 298. Knowa, H. Insecta, 408. Koenike, F. Acarina, 126. KoLLIKEK, A. Pantopoda, 139. KOKOTNEFF, A. Insecta, 266, 282, 306, 311, 324, 339-342. KOUSCHELT, E. Araneae, 51. Insecta, 347. KOWALEVSKY, A. Insecta, 266, 284-2S7, 303-316, 337, 341, 369, 372, 375, 376, 381, 396. Myriopoda, 247. Kowalevsky, A , and SCHULGIN, M. Scorpiones, 1-26. Kraepelin, C. Insecta, 350. Kramer, V. Acarina, 103, 108. Krassilsi schik, J. Insecta, 409. Kriechbaumer, J. Insecta, 296. KULAGIN, N. Insecta, 366. KtJNCKEL D'HeRCU- LAIS. J. Insecta, 369, 374, 381. KUPFFEK, G. Insecta, 283. L. Laboulbene, A., and Megmn, P. Acarina, 10S. Landois, H. Insecta, 372. Lankester. E. Ray. Araneae, 92. Limulus, 75, 78, 111. Mvriopoda, 257. Pedipalpi, 27. Scorpiones, 76. Solifugae, 36. Lankester, E. Ray, and Bourne, A. G. Scorpiones, 16. Latzel, R. Insecta, 391. Myriopoda, 219, 237- 239. Laurie, M. Pedipalpi, 81. Scorpiones, 1-26, 78, Solifugae, 35. Lebfdin.sky, J. Opiliones, 34. Lecaillon, A. Insecta, 317, 336. Lemoine, V. Insecta, 263, 268, 303, 304, 364. Lendenfeld, R. von. Pantopoda, 156. Lendl, A. Araneae, 42, 47, 111. Leuckhart, R. Araneae, 79. Insecta, 352. Pentastomidae, ISO- IS?. Leydig, F. Insecta, 324, 362. Opiliones, 111. LlCHTENSTEIN, J. Insecta, 3S9. LindstkOm, G. See Thorell. Locy. W. A. Araneae, 37-39, 42, 47, 53 55, 58-61, 64-66, 76, 79, S3, 85, 87. LOHMANN, H. Acarina, 100, 103,108. LOHIiMANN, E. Pentastomidae, 138. LOMAN, J. C. C. Araneae, 83. Opiliones, 34. Lowne, B. Th. Insecta, 403, Lubbock, J. Insecta, 354, 358, 363, 365, 392-395. Myriopoda, 239. LumviG, H. Araneae, 37-39. M. MpAlister, A. Pentastomidae, 130. MacLeod, J. Acarina, 101. Araneae, 76-79. Opiliones, 33. Scorpiones, 19. Solifugae, 36. Marchal, P. Insecta, 366. Scorpiones, 120. 440 AUTHORS INDEX. Mark, E. L. Araneae, 64-67,74-76. Scorpiones, 15, Mayer, P. Insecta, 406. Megxix, P. Acarina, 105, 109. See Robin. Meissenheimer, J. Arthropoda, 421. Melnikoff, N. Insecta, 279, 307. Metschnikoff, E. Insecta, 278-281, 284, 305, 352-354, 364, 366, 380, 382. Myriopoda, 220-236, 243 245. Opiliones, 33. Pedipalpi, 27. Pscudoscorpiones, 28, 30. Scorpiones, 1-11, 19, 47. Miall, L. C, and Denny, A. Insecta, 400. Michael, A. D. Acarina, 105, 109. Mobusz, A. Insecta, 385. Morgan, T. H. Pantopoda, 139-153, 160. MOKIN, J. Araneae, 39-46, 50, 58-62, 76, 78-91. MOSELEY, II. N. Peripatus, 175, 184. Ml'LLER, F. Insecta, 373, 393, 394. MlJLLER, JOH. Scorpiones, 1. N. NALF.r.\, A. Acarina, 107, 108. Nassonow, N. Insecta, 406. Neumann, C. J. Acarina, 107. Newport, G. Myriopoda, 219, 233, 237, 238. NlTZSCH, C. J. Acarina, 107. Nusbaum, J. Insecta, 293-298, 322, 342, 349, 388. Nu.ssbaum, M. Insecta, 409. OlJDEMANS, A. C. Acarina, 103, 109. Arthropoda, 426. Packard, A. S. Arthropoda, 429. Insecta, 357. Myriopoda, 258. Palmen, J. A. Insecta, 301, 335, 342, 349, 350, 358. Pankritits, P. Insecta, 372. Patten, W. Insecta, 267, 284, 294, 311, 323-333, 338, 340. Scorpiones. 12, 14, 18, 63. Parker, G. H. Araneae, 74. Scorpiones, 14-16. Pedasciienko, D. Insecta, 400. Perayaslawzewa, S. Pedipalpi, 121. Petrunkewitscii, A. Insecta, 341. Pictet, F. J. Insecta, 361. Plate, L. Tardigrada, 163. Platner, G. Insecta, 400 Pocock, R. I. Araneae, 37, 80. Myriopoda, 218, 257. Purcell, F. Araneae, 58, 68, 72, 74. R. Raul, C. Insecta, 314, 396. Rath, O. vom. Diplopoda, 219. Mynopodi, 232-237. Rathke, H. Insecta, 297, 338. Scorpiones, 13. Ratzeburg, J. Th. C. Insecta, 406. Redtenbacher, Jos. Insecta, 393. See Brauer, Fr., and Redtenbacher, Jos. Rees, J. VAN. Insecta, 369, 371, 374- 388. Rehberg, A. Insecta, 404. Rexgel, C. Insecta, 385. RlTTER, R. Insecta, 315. Robin, C. Insecta, 352. Robin, C., and Megnin, P. Acarina, 95. Rosenstadt, B. Myriopoda, 259. Ryder, J. A. Myriopoda, 239. S. Sabatier, A. Araneae, 37. Salensky, W. Araneae, 39, 42-48, 58, 62, 76, 81. Saint Remy, G. Araneae, 63. Peripatus, 188, 217. Sars, G. 0. Pantopoda, 160. Schaffer, C. Insecta, 335, 341, 342, 372, 386. SCHAUB, H. YON. Acarina, 128. Schiemexz, D. Insecta, 337, 385. SCHIMKEWITSCH, W. Araneae, 39, 42-55, 59, 61. 62, 76, 79, 81, 83 93, 111. Opiliones, 32. Pantopoda, 158, 160. Peripatus, 176. Scorpiones, 115. Schmidt, F. Insecta, 267, 271. AUTHORS INDEX. I 11 Schneider, A. Insecta, 353. Schubart, T. D. Pentastomidae, 130. SCHULGIN, M. See KOWALEVSKY and SCHULGIN. SCHWAMMERDAM, J. Insecta, 369. Sclater, W. L. Peripatus, 166, 172- 174. SCUDDER, S. H. Arachnida, 119. Selvatico, D. S. Insecta, 401. Sedgwick, A. Insecta, 351. Peripatus, 165, 169, 175, 177-188, 192- 214. Semper, C. Insecta, 322, 372, 373. Pantopoda, 155, 156. Sheldon, Lilian. Peripatus, 166-168, 174, 177, 178, 181, 196, 206, 208, 254. SlEBOLD, C. TH. VON. Tardigrada, 162. SlGNORET, V. Insecta, 405. Silyestki, F. Myriopoda, 257. Simmons, O. L. Araneae, 78, 114. Sograff, N. Geophilus ferrugineus, 219, 220-224. Mvriopoda, 219-226, 231, 239-251. Summer, A. Insecta, 268. Sure x sen, W. See Hansen and S6RENSEN. Spencer, B. Arthropoda, 421. Stecker, A. Pseudoscorpiones, 28. Myriopoda, 223. Stiles, Ch. W. Pentastomidae, 131- 137. Strubell, A. Pedipalpi, 121. Stuhlmann, F. Insecta, 401. Sturang, R. Araneae, 93. Scorpiones, 25. Supino, F. Acarina, 128. T. Tarnani, J. Pedipalpi, 121. Thorell, T. Araneae, 81. Pedipalpi, 27. Thorell, 1., and Lindstrum, G. Arachnida, 119. Tichomiroff, A. Insecta, 263, 285, 294, 298, 317, 322, 323, 340-342. TlCHOMIROWA, O. S. Insecta, 401. TOMUSVARY, E. Peripatus, 188. Myriopoda, 259. Troussart, E. L. Acarina, 93. U. Uljanin, V. Insecta, 263, 268, 300. Uzel, H. Insecta, 300. V. Vejdovsky, J. F. Pseudoscorpiones, 122. Verson, E. Insecta, 404. Verson, E., and Bisson, E. Insecta, 410. Viallanes, H. Arthropoda, 432. Insecta, 293, 294, 325- 328, 369. VOELTZKOW, A. Insecta, 293, 314, 315, 336, 337. W. Wagner, J. Acarina, 107, 109, 117. Arthropoda, 412. Wagner, N. Insecta, 405. Watase, S. Araneae, 70. Arthropoda, 416, 417. Limulus, 415. Weismann, A. Insecta, 264, 266, 283, 295, 302, 334, 335, 342, 351, 369-379, 385, 3S6. Weismann, A., and Ischikawa, Ch. Insecta, 402. Weissenborn, B. Scorpiones, 113. West wood, J. O. Insecta, 364. Wheelel, W. M. Insecta, 266, 282, 288, 290-301, 305-307, 312-313, 317, 324- 327, 342, 351. Wielowiejski, H. yon. Insecta, 347. Will, L. Insecta, 268, 271, 279, 281, 289, 311, 317, 324, 341, 342, 354. Willem, V. Myriopoda, 259. Willey, A. Insecta, 304. Peripatus, 165, 173, 191, 207, 211. Winkler, W. Acarina, 101, 102,107- 109, 111. Wittaczil, Em. Insecta, 280, 2S1, 306, 342, 354. Z. Zaddach, E. G. Insecta, 299. Zittel, K. A. Araneae, 71. Zograff, N. Arthropoda, 132. PLYMOUTH I W. BRENDOX AND SON, PRINTERS.