ete " ee orsiors rent tes ol rhagie soe ss srdobhet tot este rate a tate Berane ee sighip atid ative ta eaten at ned i x forth shite Tees Faget aia uate: ++ iy its; Sue ay ne (iy aa nti! agit Wo oe Rid pli aohts Hai sas 1 ra - e Aare Rape , ~. Tay ee \s e uA f ; 5 ig 7 af LS i, me ds [Aber ane lel uv | K i Hr si | Pr Ray oF NP fils ey \ a ra) i i= ie iw Fo ee 4 : | owner thea Ae aes eign A) 4 kA L tO IVF : 4 ee STUDIES hovvtk Wagan ae, Migs | OF THE 147} DEVELOPMENT AND LARVAL FORMS INVZO ECHINODERMS Dr. TH. MORTENSEN en WITH 33 PLATES AND 102 FIGURES IN THE TEXT PUBLISHED AT THE EXPENSE OF THE CARLSBERG FUND | \ ae SS SS ) WE a —- SS, Se ne Fig. 27. Posterior transverse rod of the larva of Echinometra lucunter; 2%/,. p. the first rudiment of a pedicellaria. Fig. 28. Spines of the young Echinometra lucunter; the left an embryonal spine, the right a spine of the adult type. *°°/,. spheridia are remarkably spinelike. Of pedicellariz only the ophicephal- ous form has appeared even in the oldest of the specimens. For the rest I cannot enter on the study of the postlarval development on this occasion. Echinometra oblonga (Blv.) Pl. XII, Figs. 2—3. In the rock ponds at the coast of Hawaii, near Hilo, this species occurs in considerable numbers, together with the other pacific species, Echino- meltra Mathxi. It was found to have its breeding season at the time of my stay there, in April 1915; fertilization was undertaken on the 6th, and pro- ved to be successful. The eggs are very small, opaque. In the course of | hours the 16—32 cell-stage was reached, and after about 16 hours the gastrula stage. On account of its opaqueness it forms a poor object for microscopical examination in these young stages of development. On the second day the embryos had assumed the Pluteus shape; for a few days more the larve were thriving well, but then the culture degenerated, doubtless because the water was not very good. In spite of repeated attempts I did not succeed in getting another culture, so that information can be given only of the first larval stage. The shape of the larva (Pl. XII. Figs. 2—3) is somewhat unusual, the upper edge of the anal lobe projecting almost at a right angle from the body, like a lip. The postoral band does not follow the edge of the anal lobe, or rather it has the appearance that it does not do so, the part inside 713 the band rising over it — in the preserved specimens at least. The suboral cavity is very distinct. The body is short, obliquely truncated as in Ech. lucunter. The postoral arms are somewhat unusually long. On the sides of the body the ciliated band reaches far backwards. The pigmentation is not very conspicuous; there is an indication of pigment spots in the point of the arms and scattered red pigment cells over the body. No figures can be given of the skeleton on account of its having been al- most dissolved in the preserved specimens; but I have noticed that the body skeleton forms a complicate basket structure, the recurrent rod being double; it is not very thorny. The postoral rods are fenestrated. As seen in Pl. XII. Fig. 2 there are still traces of the skeleton preserved, showing the double recurrent rod, but it is insufficient for giving detail figures. The postoral rods have been restored in the figure, but it is distinctly seen in the specimen that they are fenestrated and also that they are somewhat thorny. Echinometra Mathzi (Blv.). Fertilization of this species was undertaken repeatedly in April 1915, but never very successfully, only once the young Pluteus-stage being obtained. Moreover, the skeleton has been completely dissolved in the specimens preserved; no more information can therefore be given than what is found in my notebook, namely that the body skeleton forms a basket structure, very thorny, and very oblique, the recurrent rod being consider- ably shorter than the body rod. — Unfortunately it is not stated in my book, whether the recurrent rod is double as in E. oblonga and lucunter. The postoral rods are fenestrated. The larva is rather much pigmented, almost opaque. In view of the fact that de Meijere?) and H. Lym. Clark?) are inclined to regard Ech. oblonga only as an extreme form of Ech. Mathei, while Déderlein®) makes it the type of a separate genus (on account of its peculiar triradiate spicules), it is interesting to notice that there appears to be quite a conspicuous difference between the larve of the two forms, which is decidedly in favour of their being, at least, distinct species. Even in the first cleavage processes I noticed a marked difference between them, the cleavage cells lying much more closely pressed against one another in E. oblonga than in Mathei (while the eggs did not appear to be different in size or color). On account of the incompleteness of this record of ') Siboga-Echinoidea 1904, p. 101. *) Hawaiian a. other Pacific Echini. The Pedinide .... and Echinometride. Mem. Mus. Comp. Zool. XXXIV. 1912, p. 370. 3) Echinoidea d, deutschen Tiefsee-Expedition. p. 233. 10* 76 their development the question cannot, of course, be regarded as settled; but there is reason fo expect that a complete study of their development will show them to be quite distinct. Echinometra van Brunti A. Ag. Fertilization of this species was tried repeatedly in the beginning of November 1915 at Taboga, Panama, but always with poor success, it being at the end of its breeding season. Some of the embryos reached the young Pluteus stage. In the few preserved specimens the skeleton has been dissolved. No remarks about its structure being found in my notebook, I can only state from memory that the body skeleton forms a basket structure, but whether the recurrent rod is double or not, unfortunately, must remain uncertain. The body is very obliquely truncated. Colobocentrotus atratus (Linn.) Pl. XII, Figs. 4—5. This species occurs in great numbers on the rocky coasts of Hawaii, near Hilo, always in places exposed to the most violent surf. Fertilization was undertaken on April 4th; the culture proved successful. About the early development stages I have only noticed that gastrulation had taken place at the age of 24 hours and that the young Pluteus stage was reached on the second day. The larvee developed normally for about two weeks, when all died away. New fertilization was then undertaken, but with poor result, it being about the end of its breeding season. Information van thus be given only of the first larval stage, and, moreover, only in- complete, the skeleton having been completely dissolved in all the pre- served specimens. (There was no possibility of making drawings from the living specimens). The shape of the larva in the first stage is quite characteristic on account of the unusual width of the posterior end (Pl. XII, Fig. 5). A sketch made from a living specimen shows that the skeleton passes close to the stomach, the posterior end of the body being thus much broader than the supporting skeletal structure. The body is otherwise short, obliquely truncate as in the Echinometra larva. The postoral band has a downward sinuation at each side, which indicates the presence of ciliated lobes; on the sides of the body the vibratile band goes down almost to the posterior end, which fact, combined with the unusual width of the posterior end, indicates the pre- sence of posterolateral processes in the fully formed larva. The postoral arms long, diverging, rather broad (probably somewhat broader than shown in the figure). The suboral cavity is very deep, with a little eleva- tion in the middle above the stomach. The digestive organs show the un- usual feature of a small intestinum, well separated from the rectum by a constriction. The pigmentation consists of scattered, red grains; there are no distinct pigment spots in the arm points. Regarding the skeleton I can only quote from my notebook, that it is very robust, forming a basket structure, “somewhat extra complicate”’ and strongly thorny. (This probably means that the recurrent rod is double as in other Echinometrids). The postoral rods are fenestrated, the anterolateral rods somewhat stronger than usual. In one of the oldest specimens the posterodorsal rods had begun to form, so that it could be ascertained that also these rods are fenestrated. — That the fully formed larva will prove to have a posterior transverse rod there is not the slightest reason to doubt, but none of the larve developed far enough to show it. It is quite a problem to understand how the young metamorphosing Echinoid can succeed in attaching itself to the rocks, without being washed off or otherwise crushed by the surf.. Also the development of the peculiar spines, so well adopted to its unique habits'), would be exceedingly interesting to follow. Heterocentrotus mamillatus (Klein.) Fertilization of this species, which was found here and there under stones and coral blocks in the rock ponds near Hilo, Hawaii, was under- caken on the 8th of April. The culture was not very successful, but some specimens developed normally and reached the first Pluteus stage, but not beyond that stage. For want of sufficient material no new culture could be started. The skeleton of the preserved specimens of the larva having been dissolved, no information can be given beyond the observations from the living specimens stated in my notebook. About the first developmental processes it is only noticed that the em- bryos were in the gastrula stage 24 hours after fertilization; they were rather opaque. — The young Pluteus is only slightly pigmentated, some scattered red grains are found in the body, nothing in the point of the arms. Judging from the preserved specimens the shape of the young Pluteus appears to be much like that of Colobocentrotus. The body skeleton forms a complicate basket structure. The postoral rods are fenestrated. The statement that the basket structure of the body skeleton is “com- plicate’”” makes it highly probable that also this species has a double recurrent rod like the other Echinometrids. ') The same habit of living on rocks exposed to the surf was most probably adopted by the extinct Hemicidarid, Acrocidaris nobilis Ag. as also by the peculiar extinct Asteroid, Spheraster. 78 Echinopluteus transversus nova forma. Pl. XIII, Figs. 1—4. In the material collected by Mr. Blegvad in the West Indies in the early part of 1911 there was a number of specimens of that most remark- able Echinoid larva (Fig. 29) which I described in 1912 in the “Festschrift fir Spengel’!) and which I thought must be the larva of Echinometra lucunter. That this reference was erroneous | have proved myself by rearing the larva of that Echinoid from the egg through metamorphosis. It therefore still remains an open question to which Echinoid this larva belongs. Meantime I have found in the plankton samples collected in the Indian Ocean on my Expedition to Siam in 1900 a specimen of a larva A SS = = fs) = ‘ Sooo e crore Scre eee Saonnemeetnenee oS POTS CO SSP CSSOC CEOS ORS SOG Fig. 29. Echinopluleus transversus, species e. A. 4/,, B. %/;. Figure A shows the excessive length of the postoral arms, which are not even quite complete here, the point being broken in the specimen from which the figure was drawn. exactly similar to the Westindian form, and another one was taken there in 1914 0n my way to the Pacific. Furthermore two related forms were taken in the Gulf of Panama in 1916, and two more were found in the material collected in the West Indies by Mr. Blegvad and Mr. Faye. Thus I have now no less than six different species of this interesting larval form, which I shall designate as Echinopluteus lransversus. The species described in the paper quoted above is, together with the one from the Indian Ocean, the most spezialized of all, the other forms showing more or less primitive features. Still they all have so many characteristic features in common that I have deemed it desirable to keep them together under the same name, even though it may seem very probable that they will prove to belong to different genera, but within the same family, beyond doubt. This larval type is eminently characteristic through the excessive devel- opment of the postoral arms, while the other arms remain undeveloped, ‘) Th. Mortensen. Uber die Larve von Echinometra lucunter (L.) (?). Zool. Jahrb. Suppl. XV. 2. Bd, 1912, p. 275—88, Taf, 19—20, 79 excepting only the antero-lateral arms, which are well developed, though small, in the four first species. The postoral rod is fenestrated, the rudi- mentary posterodorsal rod is simple. There is a recurrent rod, but no typical basket-structure. The posterior transverse rod has a very peculiar bow-shape. A small, unpaired, posterior process (perhaps a spine of the young sea-urchin) is present in species e—f; whether it is present also in the other species remains uncertain. The vibratile band is not strongly developed on the body, in accordance with the reduced state of the arms; there are no vibratile lobes or epaulets. In species a—d the preoral lobe has the typical shape; in species e—f it bends forwards so as to cover the mouth and the oral area. In the two latter species the esophagus appears to be folded. The stomach is very wide; it has been impossible to disting- uish with certainty the rectum and the anal opening. The postoral arms are directed outwards so as to recall the postero- lateral arms of Ophiurid larvee, to such a degree even that in young slages, where the characteristic skeletal parts of Echinoplutei (dorsal arch, post- erior transverse rod) have not yet been formed, it is really very hard to see, whether it is an Echinoid- or an Ophiurid-larva. From the ventral transverse rods, which are more or less transformed, a supplementary transverse rod proceeds, serving for the attachment of an adductor muscle, which (in the species e—f) together with a similar muscle on the dorsal side, serves as antagonist to a powerful abductor muscle connecting the ends of the short, widened body rods. By means of this muscular apparatus the long postoral arms can be moved actively and thus apparently serve as an active locomotor organ, besides the usual ciliary movement of the vibratile bands. The immense length of the postoral arms together with the active swimming movements, which the strong development of the musculature in the species e and f indicate almost certainly to take place, make this larva one of the organisms most highly adapted to pelagic life. Species a. (Pl. XIII. Fig. 2). The single, well preserved, specimen of this species ,which was found in a plankton sample from the surface taken by Mr. Blegvad in the West Indian Sea, at 18°59’ N. 65°05’ W. (a little to the North of the Virgin Islands), (1/III. 1911), is a young one, not yet having the hydrocoel distinctly differentiated. The postoral arms are Ca. 1.5 mm long, or nearly 7 times the length of the body (the point is broken, so that the total length remains unknown, but it is not likely to be much more than the preserved part). The antero-lateral arms are well developed, a little more than body length. They are somewhat widened at the base and with a slight widening at the point. The preoral ciliated band is well developed and forms a beautiful curve: the postoral band forms a simple 80 bow and passes nearly straight on to the postoral arms?). Along the dorsal side the band goes almost straight down to the base of the postoral arms, where it then makes an almost right angle before passing out along the arms. There is no indication of ciliated lobes or epaulets. The oral area is very large, the large mouth opening lying free, contrary to what obtains in the species e and f, where it is covered by the preoral lobe. The rectum I have not been able to distinguish with certainty, but it seems that the Fig. 30. Skeleton of Echinopluteus transversus, species a. A. from the ventral, B. from the dorsal side. */,. al. anterolateral rod; b. body rod; cr. connecting rod; po. postoral rod; r. recurrent rod; str. supplementary transverse rod; t. thorn; vr. ventral recurrent rod; vtr. ventral transverse rod. anal opening lies in the posterior part of the body, below the pair of supple- mentary transverse rods lying here over the middle of the stomach. — No trace of pigmentation is seen. The skeleton (Fig. 30). The postoral rods are of the usual fenestrated type, with rather large holes and with some few, small thorns along both edges. The body rod is short, with some irregular thorns along its outer edge, and apparently with a simple, not widened point, showing no special adaptation for the attachment of the abductor muscle. The ventral *) In the preserved specimen the postoral band is even‘more straight than in the figure, passing somewhat below the points of the ventral transverse rods. That this is due to con- traction on preservation is evident, the natural position of the band being above the point of the ventral transverse rods; in that position it is also seen in species c and d. I have there- fore not hesitated in showing the band in its natural position in the figure. 81 transverse rods are directed nearly straight upwards, bending inwards at the point, so that they meet and together form an elegant arch; they are slightly widened and serrated at the point. From their base issues a backward prolongation, a “ventral recurrent rod’, which soon bifurcates, each sending one branch, the “supplementary transverse rod’’, upwards, where they join over the middle of the stomach, the other branch going straight downwards, crossing the body rod. Both these branches are simple rods in this species, while in the other species they undergo a re- markable specialisation, species b being, however, peculiar in this regard. The antero-lateral rods are well developed, simple, with few small thorns irregularly disposed. The recurrent rod is well developed, smooth, slightly longer than the body rod; from about its middle a short, simple, somewhat widened transverse rod proceeds recalling the transverse rod of an Ophiopluteus, with which it is, hovewer, not to be compared, of course; it is homologous with the connecting rod of Echinoid larvee with a typical basket structure, as e. g. the larve of Astropyga or Tripneustes. At the base of the postoral rod, close outside the point of issue of the antero-lateral rod, a rather long, simple thorn is found. The postero-dorsal rod, the posterior transverse rod and the dorsal arch have not yet appeared. There is some probability that the “Auricularia paradoxa’ (from 0°4’ N, 46°6’ W) which I described in the “Echinodermenlarven der Plankton-Expedition” (p. 21 Taf. I, Fig. 7) really is a decalcified specimen of this species. It may seem strange that an Echinoid-larva should have been mistaken for an Auricularia; but a glance at the figure will show that there is really nothing which could indicate that it might be an Echinoid-larva, the shape being just as unusual for an Echinopluteus as it is for an Auricularia; as long as the Echinopluteus transversus was un- known, hardly anybody could have fancied that this might be an Echi- noid-larva. Now it is easy to see that a form like Echinopluteus lransversus, species a, when decalcified and otherwise not too well preserved — as was the case with that larva — would get very much the same appearance as the “Auricularia paradoxa’. The complicated folds of the vibratile band, of course, do not very well suit the rather simple course of the band in the single specimen of the present species of Echinopluleus trans- versus; as, however, this is a younger stage, it is quite conceivable that the band may be more developed in the older stages of this larva. The species c would hardly come into consideration, the band being simple there also in a further advanced stage, so that it could certainly not give rise to the complicated folds of “Auricularia paradoxa’’ when decalcified; the postoral arms of this species are also much longer. Although still somewhat problematical I have deemed it well worth 1. 82 while to offer this suggestion as to the true nature of the rather mysterious Auricularia paradoxa, For the definite solution of the problem direct observations of the living larve or, at least, considerably more material than at present available will be necessary. Species b. A single specimen of this species, which is, evidently, closely related to species a, was found in a plankton sample from the Gulf of Panama, at Taboga, (January 1916). It is in a poor state of pre- Fig. 31. Skeleton of Echinopluteus transversus, species b. A. From the ventral, B. from the dorsal side. *°/,. Letters as in fig. 30. servation so that only the skeleton can be figured and described; but there? is no reason to expect that it would show noteworthy differences from the species a in regard to the general shape of the larval body. Ap- parently it differs markedly in shape from species a, the postoral arms being nearly erect; this, however, evidently is due only to muscular con- traction (comp. species c, Pl. XIII, Fig. 1). The skeleton (Fig. 31), although in its main features like that of species a, shows some very characteristic differences from that species. The ventral transverse rods form an arch as in that species, but are not so high and somewhat more widened and thorny at the end. The ventral recurrent rods are especially characteristic, being hookshaped, not bifur- cating at the point, with one or two small thorns along the sides. They 83 are directed obliquely outwards, projecting considerably from the body; it seems fairly certain that a muscle is passing between the ends of the hooks. The body rods are simple, with irregular thorns along their outer edge, not widened in the point, which is apparently not specially adapted to the attachment of an abductor muscle. The postoral rods are in the main as in species a; only the left being slightly abnormal in its basal part. Both these rods are broken so that their length cannot be ascertained. The anterolateral rods are well developed, slightly thorny. The recurrent rods are somewhat irregular; the right one bifurcates, one branch going medially, the other backwards, reaching the end of the body rod; the left ends with a widening in which a rather large hole. A somewhat long thorn proceeds from the base of the postoral rods, just outside the point of issue of the anterolateral rod. Posterodorsal rods, dorsal arch and posterior transverse rod are not yet developed. Species c. (Pl. XIII. Fig. 1). One beautifully preserved specimen of this larva was found in a plankton sample from Christiansted, St. Cruz, West Indies, taken on the 16. VI. 1915 by Mr. H. Faye. In its main features it agrees with species a, but the postoral arms are very much longer, no less than 12 times the body length; moreover the point is broken off, so that they must be rather considerably longer. The preserved part is 3mm long. The specimen has been preserved in such a state of contraction of the adductor muscle, that the arms are directed almost straight upwards, which gives the specimen an aspect very different from that of the other species. That this is, however, really due only to muscular contraction is beyond doubt, as is evident only from a comparison with the two related species, b. and d., from the Gulf of Panama, of which species b has the arms in the same upright position, while the other, species d, has them horizontally directed. The contraction of the muscle connecting the two body rods (which is quite distinctly seen in the specimen) would give the postoral rods the usual horizontal position. The anterolateral arms are distinct but considerably shorter than in species a, hardly more than one fourth of the body length, while in the latter species they are somewhat more than body length. The preoral band is well developed, though not so strongly curved as in species a. The postoral band has a concavity in the middle and makes a rather deep downwards bend before passing on to the postoral arms; however, this configuration may probably be due only to the upward direction of the postoral arms; it seems certain that on the postoral arms being hori- zontally directed the postoral band must assume the same simple shape as in species a and probably also be lowered down to the level of the 11* 84 ventral transverse rods, while in the present state of preservation it pas- ses at a distance above these rods. The band along the dorsal side is straight and simple; there is no indication of vibratile lobes or epaulets. No indication of pigmentation is seen. The oral area is very wide as in species a, the mouth opening remaining uncovered. The rectum and anal opening are not to be distinguished. Nothing can be discerned of the enterocoel or hydrocoel. The skeleton (Fig. 32). The postoral rods are of the same character as ptr Fig. 32. Skeleton of Echinopluleus transversus, species c. A. from the dorsal, B. from the ventral side. *°°/;. Letters as in fig. 30; further: dtr. dorsal transverse rod; pd. postero- dorsal rod; ptr. posterior transverse rod. The left anterolateral rod is broken and pushed some way backwards (Fig. A). in species a, fenestrated, with rather large holes; they are set with some few small thorns in the lower part, entirely smooth in the outer part. The body rod is short, smooth; the point bent a little outwards, apparently somewhat widened; it appears to be finely fenestrated, or perhaps only irregularly serrate —— this cannot be definitely ascertained in the side view presented by the single specimen at hand. But, at any rate, it is clear that there is some adaptation for the attachment of the abductor muscle. The ventral transverse rod is short, robust, with a simple, straight point. The ventral recurrent rod is short, bifurcating, one branch going along with the body rod and perhaps uniting with it at the point — this cannot be definitely ascertained, but it looks so — the other branch, the sup- plementary transverse rod, proceeding over the stomach, where it meets the corresponding rod from the other side in the midline. The end of this supplementary transverse rod again bifurcates, one branch going up- wards, the other downwards; the latter branch is somewhat widened, flat, with some few thorns along the edge, whereas the upwards directed branch appears to be simple or at most carrying a single small thorn. To this branch the ventral adductor muscle is attached in the species e and f; it seems beyond doubt that there must also be such adductor muscle attached to these rods in the present species, but on account of its excessive contraction it is impossible to see it. The anterolateral rods are fairly robust, simple, or with a very few small thorns. Just outside the point of issue of the anterolateral rod a fairly long thorn proceeds from the postoral rod. The recurrent rod is short, simple, with a small medially directed branch at the end. The posterodorsal rod is fairly well developed, almost as long as the anterolateral rod; it is simple and smooth: at the base it is somewhat curved, with a well developed medially directed process, the dorsal transverse rod, and another downwards directed pro- cess. Its whole shape is somewhat unusual for a posterodorsal rod, recall- ing to some degree an irregularly shaped dorsal arch: but I have hardly any doubt that it is really the posterodorsal rod, — especially because it is a paired structure — the dorsal arch not having made its appearance as yel. However, attention must be called to the fact that in species d, which is evidently in the same stage of development, the dorsal arch is well developed, while there are no posterodorsal rods. Full certainty re- garding the true interpretation of these skeletal parts can hardly be acquir- ed from the scanty material available at present. — The posterior trans- verse rod is bow-shaped, with a few small thorns along the posterior edge. The ends of this rod appear to be simple, not widened or fenestrated as in species e and f. Species d. One specimen was found in a plankton sample taken at Taboga, Gulf of Panama, in January 1916. The state of preservation is not good enough to allow giving a full figure of it, especially because the preoral lobe is destroyed. But it is evident that this larva resembles the species c very closely in the shape of the body. The postoral arms are horizontally directed, thus giving the larva a very different aspect from the specimen of species c shown in Pl. XIII. Fig. 1; but this depends, as stated above, only on the state of contraction of the muscles moving these arms. The arms are broken, the remaining part being 7 times the body length; that they must be really a good deal longer is evident, but whether as long as in species c cannot be ascertained. The anterolateral arms are 86 well developed, longer than in species c, nearly as long as the body. The postoral band makes a simple curve above the ventral transverse rods, apparently rather unusually broad. The skeleton (Fig. 33). The postoral rods are like those of the species a—c, with few small thorns and fairly large holes. The body rod is short, \ Fig. 33. Skeleton of Echinopluteus transversus, species d. A. from the dorsal, B. from the ventral side. 2°°/,. Letters as in fig. 30; further: da. dorsal arch; ptr. posterior tranverse rod. a little curved at the point, which appears to be somewhat widened, probably fenestrated, and somewhat specially adapted to the attachment of the abductor muscle. The ventral transverse rods are developed so as to form a high, beautiful arch; the point is bent inwards, so that they overlap one another in the middle. They are somewhat thorny, a little widened at the base and towards the point. From the ventral recurrent 87 rod a large, upwards directed Supplementary transverse rod proceeds; in the specimen they are broken at their base, but it appears that their position must be as shown in the figure 33, crossing one another over the middle of the stomach. The point is widened and has an indication of the same structure as in species c and the species e—f, viz. a simple, outer process and a widened, somewhat serrate inner process. (On account of the crossing of the rods the latter branch here becomes the outer one). By this position of the rods it is hard to see how the adductor muscle could be attached here (perhaps to the outer side of the simple process, not to the inner side as in the other species(?); the muscle cannot be observed in the specimen). Another branch of the ventral recurrent rod goes downwards, where it meets (and joins?) the end of the body ’ rod. The anterolateral rods are well developed, slender, with few small thorns. Just outside their base a long simple thorn proceeds from the postoral rod. The recurrent rod is fairly long, smooth, bifurcating at the - end. The posterior transverse rod is simple, bowshaped, with a small prominence in the middle of the posterior edge, otherwise smooth; the ends are not widened. Posterodorsal rods have not been formed, but there is a well developed, quite normally shaped dorsal arch; its branches are simple and smooth; there is a small median backward prominence, as usually found in the dorsal arch. Species e. This is the species which was described in the paper quoted above. Several specimens were taken by Mr. Blegvad off the South Coast of Haiti and off the West Coast of Portorico in the end of January and the beginning of February 1911. Furthermore some specimens were found in plankton samples taken by Mr. P. Kramp off Madeira (32°10' N. 17°20’ W.; 30/X. 1911) and off the Canaries (27°10' N. 21°53’ W.; 3/X1.1911) and in the same neighbourhood (34°22' N., 14°57’ W., and 28°43’ N. 20°40’ W.) by Mr. H. Fogh in October 1912 (on the 16th and 22nd). A pair of specimens were also taken by Mr. H. Faye in the harbour of Christiansted, St. Cruz, on the 9th and 16th of June 1915. The larva is thus seen to occur nearly all the year round. Referring to the paper quoted I may give here only a pair of textfigures (Fig. 29 on p. 78) in order to show the general appearance of this extra- ordinary larva, which, with its extremely long horizontally directed, slightly downwards curved postoral arms, in fact reminds one of a rope- dancer with an immense balancer. In the specimen with the arms best preserved they are 12 mm long, no less than 24 times the body length, and still the point is broken off. The body has the shape of a truncated four-sided pyramid. The postoral arms alone are developed; a pair of 88 short processes from the anterior end of the body cannot really be regarded as antero-lateral arms, partly because the irregular rods by which they are supported cannot be homologous with the anterolateral rods, partly because they appear not to be bordered by the vibratile band. The post- ' = awe Fig. 34. Skeleton of Echinopluteus transversus, species e. A. seen from behind; B. from the dorsal side. 18°/,. abm. abductor muscle. b. body rod; dam. dorsal adductor muscle; pd. posterodorsal rod; po. postoral rod; ptr. posterior transverse rod; str. supplementary transverse rod; vam. ventral adductor muscle; t. thorn; * rudimentary anterolateral rod. oral band makes an elegant curve or rather half circle round the whole anal lobe, and then makes a nearly right angle in passing out along the postoral arms. Along the dorsal side it makes a pair of folds. The preoral band is difficult to observe; it passes close along the postoral band, the 89 preoral lobe bending downwards over the oral area so as to cover the mouth opening completely, the entrance to the mouth being the narrow slit left between the anal area and the projecting frontal area. The eso- phagus is short, somewhat folded (comp. Species f, Pl. XII. Figs. 3—4); the stomach is very large, filling almost the whole body; the rectum appears to be small and short, situated in the lower part of the body, apparently within the space limited by the posterior and the supplementary ventral transverse rods. I have, however, been unable to make out its shape quite distinctly, and likewise the anal opening could not be located. The development of the enterocoel and hydrocoel could not be made out; "y a single specimen in a younger stage is, unfortunately, in too poor a con- oh SHY é Be : aR P dition for showing anything of the st), “Up ——, i =~ Gare enterocoel or hydrocoel formation; \( ) year it does, however, show that the oral 4 és E lobe is quite narrow and somewhat 6352 0 ofp J 0°09 g ogc), elongate, so that from the beginning PERO the mouth-opening is free, the oral Fig. 35. Half of the body skeleton of Echino- Bei Pavel A Be pluteus transversus, species e, seen from the area Deing not yet covered up y side; showing the widening of the body rod(b.). the preoral part of the body. This The postoral rod (po.) has been cut close to then takes place in a somewhat its base. al. the rudimentary anterolateral ‘ rod; t. the thorn to the point of which the later stage, though still rather early, dorsal adductor muscle is attached; str. sup- being completed already before me- plementary ventral transverse rod. vr. ventral Bane AT 2 : nt I, eel tamorphosis is beginning (Fig. 29). Peis 1 ah : In the course of the proceeding metamorphosis the body becomes quite oblique, the whole anterior part being pushed over to the right side, as seen in the figures of species f. The skeleton (Figs. 34—35). The postoral rods differ markedly from those of species a—d in having only very small holes, the unfenestrated part of the component rods being considerably larger. They are entirely smooth. The body rod is quite short, at the point widened into a large fenestrated plate (Fig. 35), to which the abductor muscle is attached. Seen in ventral or dorsal view this plate appears as a thin, slightly out- wards curving rod; seen from the hind end of the body (Fig. 34 A) it is found to be directed somewhat towards the dorsal side. The ventral transverse rods are entirely lacking; the ventral recurrent rod is quite short. The supplementary ventral transverse rod is highly developed, dividing in the end into two branches, one upwards directed apparently having the shape of a simple thorn, but really, as may be seen in side view, having the shape of a small, mostly fenestrated plate, to which the ventral 12 90 adductor muscle is attached, the other medially directed, having the shape of a thin plate with serrate edges, but otherwise unfenestrated. No muscle is attached to this plate. The posteriorly directed branch from the ventral recurrent rod coalesces with the end plate of the body rod. (In the paper referred to above the supplementary transverse rod was regarded as the true ventral transverse rod; this was the only possible explanation at that time, before the more primitive forms of this larval type were known. The presence of both the supplementary and the true ventral transverse rods in the species a—d leaves no doubt that the interpretation of these skeletal parts in species e (and f) given here is the correct one). The anterolateral rods are quite rudimentary. As will be remembered there is in the species a—d just outside the base of the anterolateral rod a long thorn proceeding from the postoral rod. In the present species there is only a thorn at the corresponding place, but generally a distinct knob is seen on the proximal side of the base of this thorn. Very probably this knob is all that remains of the anterolateral rod, the thorn itself corresponding to the thorn outside the anterolateral rod in the other species. The point of this thorn serves for attachment of the dorsal adductor muscle. If it could be proved that this thorn also in the other species serves for attach- ment of this muscle, the interpretation here suggested would be definitely proved to be correct; unfortunately I have been unable to ascertain the presence of such dorsal adductor muscle in any of the four other species, so that this point must remain unsettled for the present. — The postero- dorsal rods are simple, somewhat irregularly thorny. The posterior trans- verse rod is situated somewhat dorsally, not in the middle line (Fig. 34 A). It is a most elegantly shaped piece of calcareous structure, a gracefully curved bow, the ends of which are slightly widened and fenestrated. In the middle part it bends over ventrally, forming a triangular fenestrated plate, which covers the abductor muscle. At its posterior side a small separate skeletal part is situated resembling a small young spine (also recalling the posterior process of the Spatangoid larve, with which it can, however, scarcely be homologous, the latter being in direct connection with the posterior transverse rod and an outgrowth therefrom, while in the present case it is a separate piece, independent of the posterior trans- verse rod). Whether a dorsal arch is present or not is hardly ascertainable. The upper part of the body contains some irregular branched rods support- ing the processes at the anterior edge, which look like but are not really rudimentary anterolateral or preoral arms. It is quite possible that one of these rods really represents the dorsal arch (see Pl. XIII, Fig. 4 of species f); that it is in a nearly vertical position may not be a serious objection to regarding it as homologous with the dorsal arch of other Echinoid 91 larvee, where it is always found ina horizontal position: but the fact that there is a piece exactly corresponding to it on the ventral side makes the homology more doubtful. The question must remain unsettled for the present. Species f (Pl. XIII, Figs. 3—4, Textfigure 36). One specimen was taken off Minicoy (Maldive Islands; 73° E, 7° N.; 26/1V. 1900), another in the Bay of Bengal (89° E. 6° N.; 5/I. 1914). Both specimens are in begin- ing metamorphosis; in both of them the postoral arms are broken, so that their length cannot be ascertained; the remaining piece of one of them is Fig. 36. Skeleton of Echinopluleus transversus, species f. Seen from the dorsal side. '*°/;. Letters as in Fig. 54. 6 mm long. There is evidently no reason to doubt that they will prove to be of about the same length as in species e. The two species are so very closely alike that it is hardly possible from the scanty material available of the species from the Indian Ocean to point out specific differences. That they are, however, really different species is evident from the fact that there are no species of regular Echinoids known to occur both in the West Indian Sea and the Indian Ocean. To enter on a detailed description of the present species seems entirely unnecessary; reference to the figures must be sufficient. The interesting problem to which Echinoids these remarkable larve must be referred now needs some discussion. As mentioned above I came to the result, when describing the first of these larvae, the species e, that it was probably the larva of Echinometra lucunter. The fact that it has in the metamorphosis-stages ophicephalous pedicellariz of the type 12* 92 found in the regular Echini at once excludes both the Spatangoids, the Clypeastroids and the Cidarids, the two former having quite another type of ophicephalous pedicellariz, the latter being entirely devoid of that type of pedicellaria. Reviewing the regular Echini, other than Cid- arids, occurring in the West Indies, it seemed evident that Echinometra lucunter was the only form that could come into consideration, taking for granted that only littoral forms could come into regard when seeking for the parental origin of these larvee, which were found to occur in fairly good numbers in the littoral waters. The premises were apparently quite correct. Nevertheless the conclusion was wrong. It is not the larva of Echinometra lucunter; that larva, reared from the egg through metamor- phosis, as described and figured in the present work (p. 71; Pl. I, Figs. 1—2) has quite the typical shape of Echinoid larve. — There was at that time also some uncertainty regarding the larva of Diadema antillarum, it being only from inference concluded that the Diadema-larva would be found to resemble the Echinopluteus Miilleri, viz. that larva from the Mediterranean which has been, though without sufficient evidence, referred to Centroslephanus longispinus, a near relative of Diadema. Al- though the development of Diadema is still only incompletely known, the shape of the young larva, reared by the present author (see above, p. 29 eS P1.V, Fig. 5) gives sufficient evidence that it has nothing [s xo -) with the Echinopluteus transversus to do, and Diadema is Bee Ry - end of the body, where it bends over to the ventral side and then proceeds downwards over the frontal area unto the pre- oral band. While the dorsal part of this rod is only slightly thorny, the part on the ventral side is provided with very long side branches, so that it has the appearance almost of a spinal column with its ribs. (The name cos/alus refers to this peculiarity). This remarkable skeletal rod supports the | unusually large preoral part of the larval body, which forms if like a large vault over the mouth of the larva. — The ends of the median process and the preoral rod nearly touch one an- Fig. 62. Part other, but do not join and remain separately movable. In this eee place furthermore a quite small, styliform skeletal rod, the Ophiopluteus accessory rod, is found, which may perhaps have some func- costatus, tion in connection with the movements of the oral lobe. Speciesa./- One specimen of this larva, fully developed, showing the five lobes on the hydrocoel, was taken in the Red Sea, 17° 40’ N. 40° 10’ E. 12/X1, 1899. Besides there are two younger specimens, taken at the same time, which may probably belong to the same species. The preoral rod has merely an indication of side branches, and there is no accessory rod. The backwards directed processes from the transverse rods have not yet developed. The posterolateral rods are canaliculate unto the very points in one specimen, not at all canaliculate in the other; but in this latter specimen they are, upon the whole, somewhat abnormal. — Both spec- imens being young and in a poor state of preservation, it is impossible to ascertain whether they belong to species a or to a separate species. Also from 25 miles S.E. of Minikoi (1/I, 1914) there is a specimen of QO. costatus, which probably belongs to species a; but it is in too bad a state of preservation for definitely ascertaining to which species it belongs. Species b. (Pl. XXI, Fig. 2). The size and general shape of the body are as in species a. The arms are somewhat longer, the posterolateral ones six times, the other arms about twice the body length, the antero- 137 lateral arms being somewhat shorter than the postoral and posterodorsal arms. Skeleton (Fig. 63). The body skeleton is somewhat more robust than in species a. The ventral median process is long, reaching to the anterior edge } / J Henny, IY ( | Fig. 64. Part of postero- lateral rod of Ophiopl. | } costatus, Fig. 63. Skeleton of Ophiopluteus costatus, species b. *°/). species b. Letters as in fig. 61. aaviie of the anal lobe; it has some conspicuous side branches, while the dorsal median process, which is of about the same length, is simple, pointed, only with a few backwards directed thorns in the lower part. The posterolateral rods are set with much larger thorns than in species a, and the “canal” is quite short, reaching only from the lowermost to the third spine (Fig. 64), and there is no widening at its upper end. The preoral rod reaches much farther down on the ventral side, beyond the preoral and even over the postoral band, so that a flap of the frontal area proceeds downwards over the oral area, covering the mouth. The exact shape of the accessory 18 138 rod could not be made out; but in any case it is somewhat curved and distinctly longer than in species a. One specimen from the Malacca Strait, 7°20’ N. 98°4’ E. 7/IV, 1900. Another specimen from the Malacca Strait (5° 53’ N. 95° 43’ E, 28/XI, 1899) is intermediate in its characters between the two above species. The ventral median process is as long as the dorsal, but very slightly spinous. The posterolateral rods are strongly thorny as in species b, but the canal cannot be discerned. The preoral rod ap- pears to be as in species a, but as it lies in a different position in the preparation, direct comparison with that of the two other species is impossible; on the other hand it affords an excellent view of this rod in side view (Fig. 65). The accessory rod appears to be lacking. The length of \ ! the arms is about as in species a. In this \ specimen, as well as in the type of species b, the primary lobes of the hydrocoel have | not yet been formed. Fig. 65. Side view of preoral rod On account of the insufficiency of the of Ophiopluteus costatus. ™/;. material available it is impossible at pre- dm. dorsal median process. é : : sent to decide whether this latter specimen belongs to species a or, perhaps, represents a third species. That, at least, the two forms described as species a and b are distinct species seems, however, beyond doubt. It is quite impossible to give any reasonable suggestion as to which Ophiurid may be the adult form to which this very interesting larval type belongs. Ophiopluteus undulatus nova forma. This larval form is especially characterized by its peculiar, undulated posterolateral rods. The body skeleton is of the compound type, very short and compact; the transverse rods are simple, without processes. The rods of the arms are very smooth; only on the posterolateral rods there may be an indication of thorns. On account of the character of the body skeleton the posterior end of the body is rounded, the end rods forming only a very slight prominence or none at all. The frontal area apparently as a rule low, but wide. The stomach is very wide, the esophagus short. There appears to be a slight widening in the point of all the arms. Pigmentation unknown. 139 Species a. (Pl. XXIV, Fig. 3; Pl. XXV, Fig. 1). The posterolateral arms are fairly upright; they are about four times the body length, con- siderably longer than the three inner pairs of arms, which are scarcely twice the body length. The preoral ciliated band is straight, bending abruptly downwards at the sides. The frontal area is distinct. Fig. 66. Skeleton of Ophiopluteus undulatus. *°/,;. A. species a; Fig. 67. Part of B. species a, var. (off Pulo Pisang); C. species b. Letters as in fig. 61. posterolateral rod of Ophiopluteus un- The skeleton (Fig. 66, A; Fig. 67, A, B). The body re neta eee rods are rather strongly curved, the end rods very short, form of species a: rounded, with three or four short prominences. The — ©. from an abnorm- transverse rods are short and thick, forming together =? Specimen a small arch; the ends of the transverse rods abut on one another quite simply, without any sort of widening or indentations. The undulations of the posterolateral rods begin at about the level of the postoral ciliated band and continue unto the point; in the outer part the undulations are somewhat longer. A few short, straight thorns are found on the top o! each undulation on the adoral side, sometimes also on the outer side. i8* 140 Specimens of this species were found in plankton samples from the fol- lowing localities: Of Koh Kong, Gulf of Siam, 25/I. 1900 (1 specimen); off Koh Chuen, Gulf of Siam, 3/III. 1900 (2 spec.); 7° 2’ N. 75°32’ E. 22/X1. 1899 (1 spec.); 7°37’ N. 73°34’ E. 26/IV. 1900 (1 spec.); 8°9’ N. 71°51’ E: 27/1V. 1900 (1 specimen, peculiar through the outer fourth part of the posterolateral rods being straight); 6° 40’ N. 77°30’ E. 2/1. 1914 (1 spec.); off the East end of Sokotra, 28/XII. 1913 (1 spec.); 12° 25’ N. 46° E. 26/XII. 1913 (1 spec.). Also the larva in metamorphosis, figured in Pl. XXIV, Fig. 3, from the Malacca Strait, off Pulo Pisang (1/XII. 1899) must be referred to this species. Another specimen having very nearly completed its metamorphosis, only with the long beautiful posterolateral arms still intact, was found in a plankton sample taken by my brother, Mr. H. Mor- tensen, at Aden, in February 1898. In a plankton sample from off Pulo Pisang, Malacca Strait, 1/XI11. 1899, there is a young specimen differing from the typical form through the posterolateral rods being slightly concave at their base (Fig. 67, B); also the body skeleton is slightly different from that of the type. Perhaps this may represent another species. One specimen from the Malacca Strait (5° 58’ N. 95° 43’ E. 28/X1. 1899) shows a characteristic feature in the postero-lateral rods, the lower part being distinctly thicker than the outer part, an abrupt narrowing being found a little beyond the beginning of the undulating part (Fig. 67, C). Whether this is only an individual abnormality or it represents a separate species, cannot be stated from the present material; the same, of course, holds good for the specimen mentioned above with the unusually formed lower part of the posterolateral rods. As seen from Pl. XXIV, Fig. 3 this larva belongs to the type of Ophi- uroid-larve, where the right anterolateral arm is preserved for some time after the other inner arms have disappeared during the process of meta- morphosis. It also presents the very interesting feature that the vibratile band of the posterolateral arms continues directly across the body, below the developing Ophiuroid, so as to pass without interruption straight from one arm to the other. The same feature is seen in the metamorphosing Ophiopluleus formosus, species a, (Pl. XXX, Fig. 2) and Ophiopluteus pu- stillus, species b (Pl. XXLX, Fig. 3), indicating the existence of a sort of pupa-stage also in the Ophiuroid-larve. (Comp. p. 125). A noteworthy fact is observed in the young Ophiurid in the said figure, viz. the oral shields; they have appeared in four of the interradii but not in the fifth, and they are not quite of equal size. It should be emphasized that they make their first appearance on the ventral side, which proves that it is not a general rule that they are originally formed on the dorsal 141 side and then, during growth, transferred to the ventral side, as was stated by Ludwig.') Species b. (Pl. XXV, Fig. 2). This species is characteristic mainly by its remarkably short posterolateral arms, which are shorter than the anterolateral arms, and also broader, the vibratile band along them being distinctly broader than on the other arms. The posterior edge of the body is more rounded than in species a, quite semicircular, the body skeleton not at all projecting beyond the line formed by the posterolateral rods. The body is very short and broad, though perhaps hardly so low as shown in the figure; especially the frontal area may be somewhat higher in the living specimens. — The skeleton (Fig. 66, C) differs only very little from that of species a. The transverse rods are slightly widened at the end, and the end rods appear to differ somewhat in the shape of their thorny outgrowths. Only two specimens, both in a poor state of preservation, are at hand, from the Gulf of Aden (12° 48’ N. 50° 23’ E. 27/XII. 13) and from off the East end of Sokotra (28/XII. 13). Both specimens agree in the characters pointed out, so that it seems beyond doubt that this is a separate species, not an individual abnormality. It should especially be emphasized that there is no doubt regarding the peculiar character of the posterolateral arms; the skeletal rod is complete, not broken, and although none of the two specimens are so far in their development as to have begun meta- iorphosis, there is no reason to expect that the posterolateral arms would ultimately reach the same length as in the other species. The fact that the body skeleton (Fig. 66, C) has the recurrent rod devel- oped only on one side is, of course, an individual abnormality; in the other specimen there is also a slight abnormality in the recurrent rod, but in this case it is on the right side. Species c. (Pl. XXIV, Fig. 2). This third species has a considerable resemblance to species a. The posterolateral arms are long as in that species, but more divergent; the arms are, upon the whole, relatively shorter than in species a. Also the body is shorter and broader, and the frontal area is mainly confined to the edges. (The figure is drawn from a fairly well preserved specimen, so that it is not likely to be much different in the living specimens). In the skeleton (Fig. 68) only small differences from species a are to be noticed; the body rods and the transverse rods are more straight and the end rods have apparently only one small process on each side. 1) H. Ludwig. Jugendformen von Ophiuren. Sitz.ber. kgl. Preuss. Akad. d. Wiss Berlin. 1899. p. 212. 142 One specimen of this species was found off Taboga, in the Gulf of Pa- nama, in December 1915. Another specimen from the same locality very probably also belongs to this species; it is, however, abnormal, the right posterolateral arm being quite short, while the left is normally developed. Also the postoral arms ey /] are quite short and \\ | / Wy rudimentary. The body EN \ } y V is considerably smaller than in the normal LY, specimen. i / Besides this species \ ff : | there are two more UW _ larve from the Gulf — \\\ / if ie ase r = Wi es SS) a F aa F eee SS sell aes sae which appear to repre- ~ ye — b sent two more species TS of the undulatus-type, related to species b. As they are not in a good state of preservation, I do not think it worth while describing them, the main thing being the fact already sufficiently established that there are several distinct species within the undulatus-type. The fact that this larval type occurs both at Panama and in the Indian Ocean affords an important indication for the determination of the Ophi- urid to which it belongs. It may be concluded that it must be a type represented by more than one species both in the Indian Ocean and at the tropical west-coast of America. However, it would hardly be advisable to speculate more on this question at the present state of our knowledge. Fig. 68. Skeleton of Ophiopluteus undulalus, species c. *°°/;. Ophiopluteus fulcitus nova forma. The outstanding feature of this larval type is the branch proceeding obliquely upwards from the postoral rod towards the postoral vibratile band, giving thus a skeletal support to the anal area. There may be one or two more processes from this rod, situated farther down. The antero- lateral rod is furnished with distinct, curved thorns along its outer side, from about the level of the frontal area. The posterolateral rods are pro- vided with large, curved thorns along the inner edge. The postoral and the posterodorsal rods are smooth or finely thorny. Body skeleton com- pound; the transverse rods nearly straight, hardly widened towards the 143 point, but with a few short, irregular processes at the point. The end rods fairly long, with the point curved outwards, trifid, the median part being the largest. The anterolateral arms distinctly longer than the postoral and posterodorsal arms. Posterolateral arms fairly long, forming an obtuse angle. Arms slightly widened at the point. Species a. (Pl. XXIII, Fig. 1; Pl. XXIV, Fig. 1). Anterolateral arms distinctly longer than the length of the body, the right one generally dist- inctly longer than the left. The front- al area is fairly large and distinct, the preoral band forming a beautiful curve. The post- oral band makes a strong upward curve, on account of the supporting rods. Probably it will be found to reach even farther up in the living specimens; the fact that the supporting rods go beyond it is probably due to contraction on pre- Fig. 69. Skeleton of Ophiopluleus fulcitus, species a. *%°/,. Letters as in fig. 61. s. supporting rod. servation. The skeleton (Fig. 69). The posterolateral rods are entirely smooth along the outer side. The supporting rods of the anal area are long, thorny in the outer part, their points generally crossing one another; another, shorter process issuing near the base of the postoral rod, directed ventrally. The anterolateral rods with one process near the base, somewhat below the point of issue of the posterodorsal rod; another somewhat higher up, both directed obliquely dorsally and inwards. Postoral and posterodorsal rods smooth. Body rod and recurrent rod with a prominence near the lower end. — The prominences on the anterolateral rod wanting in one young specimen, while in another specimen, also rather young, the upper one is wanting. One abnormal specimen (right postoral arm rudimentary) has only the upper one of these prominences developed. 144 This larva was taken in the Gulf of Panama, near the island of Taboga, in December 1915. 8 specimens. Species b. (PI. ' ‘ XXIII, Fig. 2). This VA Se species differs from ee i the preceding one in the general shape of the body being shorter and broad- er, and the postero- lateral arms some- what more diverg- ing. The supporting rods of the anal area are shorterand upon the whole more distant. The posterolateral rods are thorny on the Fig. 70. Skeleton of Ophiopluteus fulcitus, species b. *°°/,. outer side in the S- supporengared: middle part. The anterolateral rods only with one process, corresponding to the upper one of the two found in species a. Postoral and posterodorsal rod slightly thorny, but to a variable degree. (Fig. 70). This species was found, together with species a, in the Gulf of Panama, near the island of Taboga, December 1915. 6 specimens. Speciesc. (Pl. XXIII, Fig. 3). The main character distinguishing this species from the two preceding ones is the presence of three processes from the postoral rod, the median and the lower ones being directed ventrally (Fig. 71). The anterolateral rod with only one process, corresponding to the lower one of those in species a. The posterolateral rods thorny on the outer side, as are those of species b; the postoral and posterodorsal rods slightly thorny. The shape of the body of this species is more like that of species a. Gulf of Panama, near the island of Taboga; December 1915. 2 specimens. Although there is some variation in the length of the supporting rods of the anal area, the three species are easily distinguished by the characters pointed out above, which are very constant, judging from the fairly large and well preserved material in hand. There can then hardly be any doubt that we have here three distinct species, not merely individual variations of one species, of the same eminently characteristic — lar- val type, occurring together in the same locality and at the same time of the year. It is useless to meditate over the problem to which genus of Ophiurids these larve belong. The fact that it is known only from the Gulf of Panama is not a sufficient indication for solv- ; Fig. 71. Skeleton of Ophiopluteus fulcilus, species c. *9°/,. ing the problem. s. supporting rod. Ophiopluteus opulentus nova forma. A very marked feature distinguishing this larval form from all the Ophiurid larve hitherto known is the presence of a pair of short extra arms on both the ventral and dorsal sides, proceeding from the lower part of the postoral and the anterolateral rods, below the point where the posterodorsal rod issues from the latter. These extra arms are very much shorter than the other arms and, upon the whole, hardly of the same morphological value as these; but their structure is the same, the ciliated band passing out along them in the same way as in the usual arms. The body skeleton is of the compound type. There is a pair of small processes from the middle of each transverse rod. The posterolateral rods are canaliculate to a greater or lesser extent. The end rods are of medium length, straight, trifid at the point. The posterolateral arms are about 2—3 times as long as the inner arms. Species a (Pl. XXI, Fig. 3). The posterolateral arms about twice the length of the inner pairs of arms, which are thin and round, ending in a small widening. It is uncertain whether there is any widening at the point of the posterolateral arms. The body is fairly large, the inner arms measur- ing about 11/, the length of the body. The frontal area is distinct, but not large. 19 146 The skeleton (Fig. 72). The posterolateral rods are canaliculate from about the middle unto the point. They are provided with small thorns, somewhat irregularly placed, along both the outer and inner sides (Fig. 73, A). The other rods are smooth or only here and there with a very small thorn. The transverse rods have a small thorn directed forwards and a somewhat larger one projecting backwards; the ends of the transverse rods are only slightly widened. The dorsal extra rods are a little thorny. Only a single, not very well preserved specimen of this larva is at hand, taken off Koh Kam, in the Gulf of Siam, 4/II. 1900. x | eA Se a tr aS —— je | Fig.73. Outer part a a a Fig. 72. Skeleton of Ophiopluteus opulentus, species a. *°°/;. of posterolateral Letters as in Fig. 61. rod of Ophioplu- teus opulentus; Species b. There are two specimens from off the East A. species a; B. species b. 79/,. end of Sokotra (28/XII. 1913), both in so poor a condition that no whole figure can be given of this larval species. It can only be stated that the general form appears to correspond to that of species a. The skeleton (Fig. 74) affords some very good characters distinguishing it from the former species. The posterolateral rods are entirely smooth along the outer side (Fig. 73, B) and the canal goes almost from the base to the point of the rod. The processes from the transverse rods are directed more outwards in one of the specimens. Species c. (Pl. XXII, Fig. 3). The relative length of the body and the interior pairs of arms about as in species a, but the posterolateral arms are conspicuously longer, about three times the length of the interior arms. The frontal area appears to be rather larger than in species a. The preoral vibratile band is somewhat upwards curving, while in species a it is nearly 147 straight. However, much stress cannot be laid on this character, which may to some extent depend on preservation. It appears that the small extra arms are less developed than in species a. In the specimen figured the ventral ones are distinct, but quite small; the dorsal ones are not distinct. In another specimen one of the dorsal arms is very distinct. That they will be developed, both pairs, in the fully formed larvae (— the hydro- coel has not begun to form lobes in any of the specimens—) I have no doubt. The skeleton (Fig. 75). The posterolateral rods are canaliculate almost from the base to near the tip. They are naked along the outer side, ex- cepting one or a very few small prominences so placed off the thorns on \ Fig. 74. Skeleton of Ophiopluleus opulentus, species b. *°°/,. the inner side, that a slight thickening is formed (Fig. 76); there is gener- ally such a thickening in the lower part, sometimes also a few farther out, and in that case the rod gets a peculiar, somewhat nodulose appearance. The dorsal extra rods are smaller than the ventral ones. The processes from the transverse rods are well developed; the ends of the transverse rods are distinctly widened. Of this species there are three specimens from Christiansted, St. Cruz, West Indies, taken by Mr. H. Faye, 16/VI. 1915. There are also some younger specimens, which may possibly belong to the same species, but the distinguishing characters not yet being fully developed (— the extra rods have not appeared —) I cannot state definitely whether they do really do so. To this species, however, some curious “specimens” (Pl. XX, Figs. 3—5) undoubtedly belong, which consist only of the body skeleton and the po- sterolateral arms, these arms still having their vibratile band intact. All the rest of the larva has disappeared. This is a case analogous to what obtains in the Ophiothriz-larva, viz. that the posterolateral rods are not 19* 148 absorbed during metamorphosis, but are kept intact as a floating appa- . ratus and then dropped, when the young Ophiurid has completed meta- im orphosis and is ready to give up pelagic life. The long posterolateral arms remain in connection, after the Ophiurid has gone, their ciliated band being intact and in the present case, moreover, joining in the middle, below the Ophiurid (this takes place before the Ophiurid leaves the “larva’’). While in the case of Ophiothrix the abandoned posterolateral arms must evidently soon perish, this appears not to be the case with the present Fig. 76. Basal part of posterolateral rod of the same species. *%/,. larva. As stated above (p.124) it seems fairly certain that a new larval body regenerates in the place of the former larval body. How far the process of regeneration goes cannot be ascertained; but in any case Pl. XX, Fig. 5 shows that it may go on so far as till the formation of a new mouth and esophagus; it is also evident from the numerous nuclei seen in the anterior part of the new body that a vigorous growth is going on here, so that it would seem most probable that the process may continue the short while, until the new digestive organs are able to assume normal function and then there seems to be no reason to doubt that a new complete and ultimately metamorphosing larva may be the result. Thus we would here have a true case of metagenesis, otherwise totally un- known in Echinoderms. — As explained above (p. 124) the suggestion that 149 we might have to do here only with abnormal larvee is quite inacceptable ; it would be impossible for a larva with such incomplete digestive organs, and accordingly unable to feed, to form the long postoral arms and the whole body skeleton in a perfectly normal way. (That the transverse rods are somewhat different from those of the fully formed larve is evidently due to the absorption during metamorphosis). This, however, leads to an- other important problem. It is evident that the considerable growth of the regenerating larval body could not take place without some food sup- ply. The “larva” itself being unable to feed until the new digestive organs are ready to function, there must be some nourishment stored somewhere. This also appears to be the case. In the vibratile band of the posterolateral arms there is generally seen a great number of very fine refractile grains. These might well represent the food supply; in fact, it is hard to see, where it could otherwise be found. Of course, I do not mean to maintain that definite proof of this astonish- ing regeneration has been given. But the available material certainly in- dicates that it does take place. The problem most urgently invites closer investigation. The fact that this larval type is found both in the West Indies and in the Indian Ocean gives a valuable hint as to the question to which Ophi- urid type it belongs. Still it is not sufficient for solving the problem, and an attempt to find out the parental forms on the base of our present know- ledge would be essentially guess work of very doubtful value. Ophiopluteus similis Mrisn. Pl. XXVII, Fig. 3. Th. Mortensen. Die Echinodermen-Larven d. Plankton Exped. p. 60. Taf. V. 8; VI. 4. The specimen figured, which was taken at Misaki, Japan, 29/VI. 1914, I am unable to distinguish by any noticeable character from the 0. similis from off the Cape of Good Hope described in the work quoted. The more upright position of the posterolateral arms is hardly a reliable character but probably due to preservation, as, in fact, the arms of many Ophiurid larve are apt to bend and curve when the larve are mounted in Canada- balsam, especially such forms as have long arms with slender supporting rods. (In the present specimen all the arms are much curved, but have been restored to their natural shape in the figure). There is a distinct sub- oral cavity. The skeleton (Figs. 77, 78) presents some minor differences in the shape of the transverse rods, which may perhaps indicate that the specimen from 150 Japan represents another species than that from Cape, but the differences are too insignificant for making it necessary to designate it as a separate species at present. In any case the two forms must be very closely related. Another, younger specimen, from the same locality and the same day, probably also belongs to this species, although it differs in the antero- lateral arms being longer than the postoral ones and in the suboral cavity being indistinct. In this specimen the posterolateral arms have the same outward direction as in the type. Thesuggestion made in the work quoted that this larva may belong to some Am- phiura-species is with- out real support. It was made on account of the similarity be- tween this larva and Ophiopluteus bimacu- Fig. 77. Skeleton of Ophiopluteus similis. *°°/;. latus (Joh. Mill.), Letters as in fig. 61. which latter was re- ferred to the genus = se a Bs cee Amphiura on cae of the fact that the young Ophiurid, as shown in the figures of Joh. Miller (V. Abhandl. Taf. V, Fig. 6) has two papille on each mouth angle, a feature mainly characteristic of Am- phiura. However, as I have shown recently!) that Ophiopluleus mancus is the larva of Amphiura filiformis, the suggestion that O. bimaculatus might belong to an Amphiura becomes improbable. The two papille of the young Ophiurid are hardly the typical infradental papille of the Amphiurid type, but more probably tooth papille situated in the inner part of the mouth; in Amphiura filiformis I have found the infradental papille to appear only at a much later stage of development. The fact of the occurrence of this larval type both at Cape and Japan is not of sufficient zoogeo- graphical importance for indicating to which species of Ophiurids it belongs. Whether it is more nearly related to Ophiopluteus bimaculatus must also remain uncertain for the present. Fig. 78. Part of posterolateral rod of Ophiopluteus similis. ?%°/,. 1) On the development and the larval forms of some Scandinavian Echinoderms. p. 138. 151 Ophiopluteus formosus nova forma. Body skeleton compound; no processes from the transverse rods. The posterodorsal rod issues near the base of the anterolateral rod; postero- lateral rods simple, with numerous, fairly large, curved thorns along the inner side; the other rods smooth. This larval type is not distinguished by any specially marked characters, the low point of issue of the posterodorsal rods being the most conspicuous feature. The larva is a very typical, regular form, of a beautiful, graceful shape. Species a. (Pl. XXVII, Fig. 1; Pl. XXX, Fig. 2). The body is fairly large; the inner arms hardly as long as the body, the posterolateral ones Fig. 79. Skeleton of Ophiopluteus formosus, species a. *°/,;. Letters as in fig. 61. about twice that length, gracefully bent. The preoral band is straight in the middle, turning rather sharply downwards at the sides; the postoral band with a slight sinuation at each side. At the base of the posterolateral arms the band makes a deep downward curve, forming an incipient vibratile lobe. The frontal area is not very large. Arms fairly broad and flat. In the skeleton (Fig. 79) the somewhat widened, lobate ends of the trans- verse rods are noticeable; otherwise there are no very prominent features. 5 specimens from the Gulf of Panama, near the island of Taboga, in December 1915. One specimen is in the metamorphosis stage (Pl. XXX, Fig. 2) the inner arms, except the right postoral one, having been absorbed; also the left posterolateral arm has begun to undergo absorption. The outstanding fea- ture of this stage is the arrangement of the vibratile bands. The two bands 152 of the posterolateral arms are continuing across the body, below the devel- oping Ophiurid. At the anterior end there is another transverse band, ap- parently derived directly from the preoral band. Thus a remarkable rear- rangement of the vibratile bands takes place in this form, something corre- sponding to what takes place in the transformation of the Auricularia into the pupa-stage (Cf. p.125). A similar case was observed in Ophiopluteus pusillus, species b (Pl. XXX, Fig. 3) and O. undulatus, species a (Pl. XXIV, Fig. 3). Species b. (Pl. XXVII, Fig. 2). This species differs so conspicuously from species a as regards its shape that it may seem doubtful, whether it Fig. 80. Skeleton of Ophiopluteus formosus, species b. *%°/;. really belongs to the same type. The inner arms are one and a half to two times, the posterolateral ones more than four times the body length. The arms are narrow, rounded, not broad and flattened as in species a. On the other hand the skeleton (Fig. 80) so closely resembles that of species a that only very small differences can be pointed out, the main thing being that the transverse rods are somewhat more robust, the indentations at the ends being less deep. Perhaps the shape of the meshes of the body skeleton will prove a little different. In species a the meshes are somewhat narrower at the lower end, which does not appear to be the case in the present species; but as there is only one specimen, and that one even showing a slight difference in the shape of the two sides of the body skele- ton, this feature is not to be relied upon as a specific difference. The thorns on the posterolateral rods are blunt and more distant than in species a. (Fig. 82, B). One single specimen, fairly well preserved, from the Red Sea, 24°43’ N, 35° 45’ E. 10/XI. 1899. 153 Speciesc. Ina plankton sample taken by Mr. Gruelund in the neigh- bourhood of the Cape Verde Islands at 16°11’ N. 21°57’ W. 14/VI. 1920, a single specimen of a larva of the O. formosus-type was found. It is in a very poor state of preservation, only the body skeleton and the postero- lateral arms being intact. These parts, however, clearly show that we have here another, distinct species of the formosus-type, which may be easily recognized by the characters olf its skeleton, even though the shape of its body is un- known. The skeleton (Fig. 81) differs from that of species a and b mainly in the pec- uliar shape of the end rods, the points of which are bent : ; - ~ ere nearly at a_ right PTET E HTN Ne x ey: angle; they are longer than in the j \ N \ s two other species in and also provided with longer side- ¢. >t branches. One _ of the transverse rods Fig. 82. Part of posterolateral rod of Ophiopluteus formosus, *°),. A. of species a; B. of species b; C. of species c. Fig. 81. Skeleton of Ophiopluteus formosus, species c. *°°/;. carries a short pro- cess, which is, however, broken, so that its total length and the shape of its point cannot be ascertained; but it is evidently very nearly as shown in the figure. The point of issue of the posterodorsal rod is not so close to the base of the anterolateral rod as in the other species. (Apparently the distance is not the same on both sides; this, however, is due to the fact that the right one is broken and lies in a somewhat different position). The thorns on the posterolateral rods are finer and more distant than in species a and b. (Fig. 82). The posterolateral arms are very long and much diverging. The occurrence of this larval type in the Red Sea, at Panama and in the Atlantic Ocean is, of course, of importance, but it is not sufficient for venturing on a suggestion as to the Ophiurid to which it belongs. That Ophiopluteus formosus and similis are nearly related is fairly evident, the 20 154 characters distinguishing them from one another being very slight and un- important. Upon the whole the larval types O. similis and formosus are among the less interesting on account of their more generalized features. But I have thought it desirable to include them in this report in order not to exag- gerate the impression that the Ophiurid larve are easily arranged into natural groups. That is the case with some forms, like e. g. undulatus, costa- lus; but the present forms are equally important as representing types which do not lend any special support to a natural classification. Ophiopluteus fusus nova forma. Although there is only one specimen in hand, and that even in a very far advanced stage of metamorphosis, as seen from the figure 83, A, I have deemed it advisable to give a description of it, partly because the char- acters to be seen on the remnants of the larva still attached to the young Ophiurid are sufficiently marked for recognizing the species, partly because it seems to show affinities to another Ophiurid-larva which stood hitherto remarkably alone, viz. Ophiopluteus paradoxus. The main character of this larva is the peculiar thickening of the postero- lateral rods a little above their basis. While the thickening increases very gently from the base upwards, it ends rather abruptly outwards, the outer part of the rod being very thin, with fairly long, thin, straight thorns on either side (Fig. 83, C); the basal and thickened part is smooth. Also the postoral rod has the same shape, although the thickening is less pronounced. The other rods have already been absorbed, so that nothing can be said about their structure. The body skeleton is simple, the end rods moderately long, slightly bent in their outer parts. The transverse rods are irregularly widened at the point; there is no median process. The shape of the posterolateral arms appears to be rather broad and flat, as in O. paradoxus. There is a distinct vibratile tuft at the posterior end, as in the latter species. The single specimen in hand was taken at the Azores (38°14’ N., 24° 35’ W. 19/1 19ia- Ey Blesvad). As stated above this species would appear to be nearly related to Ophio- pluteus paradoxus, the larva of Ophiura albida Forb. They agree in the characters of the body skeleton, the presence of a posterior vibratile tuft, as in the general structure of the young Ophiurid, and probably also the larval arms have the same shape in both; the only essential difference is the swelling of the basal part of the posterolateral and postoral (a. 0.?) 155 rods in O. fusus. If it should ultimately be proved that these two larve are really nearly related, it will be a fact of considerable importance, showing that also within the genus Ophiura, even as recently restricted ——— Fig. 83. Ophiopluteus fusus. A. specimen in metamorphosis; '°/;. B. body skeleton of the same; C. outer part of posterolateral rod. *°°/,. Letters as in fig. 61. by Matsumoto and H. L. Clark, different groups of larvee may occur. The fact that the larve of the two species O. albida and texturala are so different would then prove that these species are not so nearly related as hitherto supposed; the other conclusion that might be drawn, viz. that the larve cannot show anything regarding the natural affinities, is much less probable, considering the facts brought to light in the present work. 20* 156 Ophiopluteus serratus Mrtsn. Pl. XXVIII, Figs. 1—2. In the memoir “Die Echinodermenlarven der deutschen Siidpolar Ex- pedition 1901—3” p. 96 (Taf. XIII, Fig. 1) an Ophiurid larva was described under this name, characterized especially by its very large median process with a series of more or less branched prominences along its outer side. In the material collected during my expeditions to Siam 1899—1900 and to the Pacific in 1914—16 there is a considerable number of Ophiurid larve which belong, evidently, to the same type. They may possibly re- present two different species or even more; but as they show considerable variation, especially in regard to the size of the median processes, it is hardly possible to come to a definite result as to their specific limits. I have therefore preferred to designate them all simply as Ophiopluleus ser- ralus. Only the study of living material and tracing their origin to the parental forms will give the clue to the question of the specific limits within this, evidently, very variable larval type. This larva is characteristic through the great length of the inner arms and the nearly upright position of the posterolateral arms. These latter arms are ca. 5 times, the inner arms about 3 times the body length. There appears to be a very slight widening at the tip of the arms. The preoral area is very low; it may be distinct in its whole width or only at the corners. This is possibly a specific difference, which cannot, however, be decided from the preserved material. The preoral band is more or less covered by the postoral band, the whole oral area-being almost totally covered up by the postoral or anal lobe!). A conspicuous feature in some of the spec- imens is the very large size of the stomach (Pl. XXVIII, Fig. 1), which fills the whole space inside the posterolateral rods. In other specimens it is much smaller, occupying only the space inside the rods of the inner arms (Pl. XXVIII, Fig. 2). It is very well possible that this is really a specific character, but in the preserved material it is too often impossible to decide, what is due to preservation and what not, so that, although there would seem to be no doubt in some cases that we really have to do with two different species, it is better to leave the species question undecided at present. In the skeleton (Fig. 84) there appears to be no distinct characters eventually corresponding to the mentioned differences in the size of the stomach and the preoral area. A noteworthy fact is the very great variation in the size of the median processes from the transverse rods. In some specimens they are quite small and inconspicuous, in others as large as ') The suggestion given in the original description that the peculiar feature of the oral region of the type specimen was due to compression is certainly correct. 157 in the type. The spinulation, of course, varies with the size of the pro- cesses. Some specimens observed alive at Misaki, Japan, in June 1914, were found to have the stomach green, while some tiny red pigment grains were scattered along the body skeleton. No pigment spots in the tip of the arms. Numerous specimens were found from Japan to the Red Sea: Misaki, Japan, 29/1V, 2-17/VI. 1914; Zamboanga, 1/III. 1914; Malacca Strait, Fig. 84. Skeleton of Ophiopluleus serratus. */,. Letters as in fig. 61. 30/XI. 1899; S. of Koh Chuen, Gulf of Siam, 3/III. 1900; 89°15’ E. 5°55’ N. (Bay of Bengal), 5/I. 1914 (1 specimen); 73°34’ E. 7°37’ N. 26/IV. 1900 (1 specimen); Gulf of Aden, 14/XI. 1899; Strait of Bab-el-Mandeb, 5/V. 1900 (1 specimen); off Jebel Zukur, Red Sea, 5/V. 1900; Suez, 31/1. 1898 (Mr. H. Mortensen). A very interesting feature is connected with the metamorphosis of this larval type. The suboral cavity continues some way down in the body, asa pouch on each side of the stomach. In the wall along the bottom of this pouch a thickening appears at the time, when the hydrocoel begins to form lobes (Pl. XXVIII, Fig. 2); this thickening grows very considerably and forms lobes, which ultimately combine with the lobes of the hydrocoel. 158 The process having been described in details by Joh. Miiller in his Memoir “Die Ophiuridenlarven des Adriatischen Meeres” in the case of the larva, which he names Pluteus bimaculalus, there is no reason for me to give a detailed description of the process here again, the less so since the partly poor state of preservation of the material available would prevent the completeness of the description. But it is an exceedingly interesting fact that we have in this larvaa typeof development rather different from that found in the Ophiothrix-larva, the only larval type the development of which has been studied in detail as yet. A careful study by means of sections of the developmental processes of the present larval type would be very desirable and would be sure to yield most interesting results. —- This is another instance serving to show that it is not justi- fiable to conclude from the study of the development of one single form that the whole of the group follows exactly the same type. (Cf. Antedon). As stated above the mediterranean Ophiopluteus bimaculatus agrees with the present larva in the type of its development, and there is certainly no doubt that these larve are very nearly related. They are alike in struc- ture and shape, and only in minor features specific differences can be pointed out. It might be suggested that perhaps also Ophiopluleus affinis (Echinodermenlarven d. Plankton-Expedition, p. 61, Taf. VI, 1—3) will prove to belong to this group. Its general appearance is very much like that of O. serratus, but its oral structure is incompletely known. The skeleton affords the same general structure, the only essential difference being that the posterolateral rod has a longitudinal ridge, dividing it into two parallel rods. But this would hardly be sufficient to disprove its relationship to the serratus-bimaculatus group. — No suggestion can be given as to which Ophiurid is the parental form of these larve. As stated above, under O. similis (p. 150) the suggestion that O. bimaculalus might belong to an Amphiura is untenable, since it was found that the larva of Amphiura filiformis is of quite another type (Ophiopluteus mancus). The presence of the larval type both in the Indo- Pacific Ocean and in the Mediterranean and the Atlantic (— it was taken by Mr. Gruelund, VI. 1920, off Cape Blanco and at the Cape Verde Islands —) will, of course, be a fact of importance for judging of its parental origin, but it is not sufficient for making a definite suggestion as to that problem. Ophiopluteus arcifer nova forma. This larval type is distinguished by the very long median processes from the transverse rods, which in their highest development, in species a, when seen in side view, recall the shape of an oldfashioned cross bow (Fig. 85). 159 They cause a very prominent elevation of the body-wall, like a hump, on both the ventral and the dorsal side of the body. The body is comparatively low, with a low frontal area. The preoral band is generally almost straight. A noteworthy feature is the more or less conical shape of the esophagus (especially so in species a). The arms Fig. 85. Body skeleton of Ophiopluteus arcifer, species a. A. in side view. *°/,; B. oblique dorsal view. al. anterolateral rod; b. body rod; dm. dorsal median process; e. end rod; pd. posterodorsal; pl. posterolateral; po. postoral rod; r. recurrent rod; tr. transverse rod; vm. ventral median process. are slightly widened at the point. The posterolateral arms are 3—95 times the body length, rather obliquely outwards directed. The other arms are not much more than body length, their point being at about a level with the point of the posterolateral arms. The body skeleton is of the com- pound type. Species a. (Pl. XXVI, Fig. 1). The posterolateral rods are set with bilaterally arranged small thorns, which give the rod a peculiar, almost 160 banded appearance, when seen from the ventral or the dorsal side. Near their base they have a few small thorns along the outer side, and also the body rod generally has a few small thorns. The end rods are provided with several small branches at the point. The transverse rods are hardly widened at the point, but end in a number of very short thorns. The median pro- cesses, which proceed from the transverse rods of the same side, some- times the left, sometimes the right, are very long, slightly curved and irre- gularly thorny along the outer side (Figs. 85, 86). The postoral and postero- Fig. 86. Skeleton of Ophiopluteus arcifer, species a. *°/,. Letters as in fig. 85. dorsal rods have small, bilaterally arranged thorns, the anterolateral rods have thorns only along the outer side in their outer part. — The postero- lateral arms are about 4 times the body length. This species was found S. of Koh Chuen, in the Gulf of Siam, 3/IIT. 1900 (1 specimen), in the Malacca Strait (101°12’ E. 2°45’ N. 30/X1. 1899 (3 specimens) and off Jolo, 20/11]. 1914 (5 specimens). Species b. (Pl. XXVI, Fig. 2). The main difference between this spe- cies and the preceding one consists in the posterolateral rods being here provided with thorns along their outer and their inner side, not along the ventral and dorsal sides. The thorns on the outer side are irregular, slightly branched, those along the inner side simple, with the point bent a little forward. The end rods have only one side branch at the point, and the transverse rods are provided with only few, short prominences at the point. 161 The median processes are not so large as in species a, and set with only few thorns along their outer sides. The anterolateral rods are much more thorny than in species a (Fig. 87); the postoral and posterodorsal rods as in the preceding species. The length of the posterolateral arms is only about 3 times the body length, the other arms as in the preceding species. Only two specimens of this species were found, in the Gulf of Panama, off the island of Taboga, in December 1915. J I} 4 ] x | Ye /y \ Whi /, SX \\ YZ a \\ \ l ae ia \\\ | / Mh ww B \ \ \ A / / A CON Ae eae Fig. 87. Skeleton of Ophiopluteus arcifer, species b. *°°/,. A. Part of anterolateral rod, B. part of posterolateral rod; the adoral side of both is to the left. C. Body skeleton. The processes from the transverse rods are restored, being broken in the specimen, and are represented in a more upright position than normal. Speciesc. (Pl. XXVI, Fig. 3). This species is nearest to species b, but differs from it in the character of the posterolateral rods, which have fewer, small thorns along their outer side, and not in their whole length, while those along the inner side are larger and continue until the point; they are partly (mainly in the lower part of the arm) somewhat spinulose. The rods of the other arms are almost as in species b. The median processes from the transverse rods are shorter than in that species (Fig. 88). The posterolateral arms are about five times the body length. One specimen was taken at 17°47’ N. 70°51’ W., 1/II. 1911. (H. Bleg- vad). This species recalls to some degree O. bimaculatus (Joh. Miller), but differs conspicuously from it in the presence of thorns along the outer side of the posterolateral rods and in the much greater length of the median processes from the transverse rods. Unfortunately, the only specimen in hand does not give any evidence, whether the development of the Ophi- 21 > | Ir 162 urid proceeds in the same way as in O. bimaculatus and O. serratus (see under the latter species). Neither are any of the specimens of the species a and b in the metamorphosis stage. In any case it should be kept in mind for future investigation, that these larve possibly have some nearer affinity to one another. At our present stage of knowledge it is impossible to form an opinion of any value as to which forms of Ophiurids the larval type of Ophio- pluteus arcifer belongs. The existence of species of this larval type in the East Indian Seas, in the Gulf of Panama and in the West Indies is, of Fig. 88. Skeleton of Ophiopluteus arcifer, species c. *%/,. A. Part of posterolateral rod; B. the same rod, probably showing regeneration. C. body skeleton, seen slightly from above. course, a fact of importance, but not in itself sufficient foundation for a reasonable suggestion as to the solution of this problem. Ophiopluteus monacanthus nova forma. Pl. XXX, Figs. 3—4. From off Jolo (21/III. 1914) there are two specimens of a very peculiar Ophiurid larva, which seems to me well worth describing, although the fact that the skeleton has been dissolved prevents giving a complete description of it. The outstanding feature of this species is the presence of a very strong median process on the ventral side alone; it projects at a right angle to the body, forming a large hump on the ventral side, while the dorsal side of the body remains flat, there being evidently no median process from the dorsal transverse rods. There is a widening on the rectum, filling out the hump. In its general appearance the larva otherwise recalls Ophiopluteus arcifer. The postoral band makes a small, but very distinct backwards curve at each side. The posterolateral arms are somewhat more 163 than three times, the other arms about twice the body length. They have a slight widening at the point. Although the skeleton is dissolved, it can be definitely ascertained that the posterolateral rods are thorny along their outer sides. There seems to be reason to think that this species is related to O. arcifer, but, of course, this cannot be definitely ascertained until its skeletal struc- ture is made known. By the unique character of the ventral hump this species is easily recognizable. Ophiopluteus retrospinus nova forma. The main distinguishing feature of this larval type is afforded by the peculiar backward projecting processes from the lower part of the postero- Fig. 89. Skeleton of Ophiopluteus retrospinus, species a. *°°/,. Letters as in fig. 85. m. median process. lateral rods. — The arms are fairly long; the posterolateral arms are about three times, the anterolateral arms only a little more than body length, their point being at a level with the point of the posterolateral arms. The two other pairs of arms are somewhat shorter, their points not reaching the same level as that of the other arms. There is a very slight widening at the point of the arms. They are all gracefully curved, especially the posterolateral ones, which gives a touch of elegance to the shape of the larva. The frontal area is well developed. The preoral band is somewhat upwards curved, more or less straight in the middle. The postoral band makes a little downward curve at each side (more distinct in species b), and from here the nervous system is seen to proceed. (In species b only 21* 164 the ventral part of it could be discerned). There is a distinct suboral cavity, as in QO. similis. It is not represented in species b, owing to the less satis- factory preservation; but it is distinctly discernible also in that species. The body skeleton is of the compound type; there is a more or less developed median process from the transverse rods. The posterolateral rods are provided with strong thorns along their inner side, the other rods have fine thorns. \ —S Fig. 90. Skeleton of Ophiopluteus retrospinus, species b. **°/;. Species a. (Pl. XXII, Fig.1). The anterolateral, the postoral and the posterodorsal arms are about 11/, the body length. The backward processes from the posterolateral rods are very prominent and conspicuously ser- rated, especially along their inner side (Fig. 89). Besides these there is a smaller process (more rarely two, or none at all) from the lower end of the body rod, also a little serrate. The end rods are of moderate length, curving outwards in the lower part; the small side branch is at some distance from the point. The transverse rods are hardly swollen at the point, with merely an indication of interlacing branches. The median rod is long, slightly thorny, the ventral rod somewhat longer than the dorsal one. The rods of the inner arms are only faintly thorny, also the antero- lateral rod carries only small thorns along its outer side. This larva was found in the Gulf of Siam, S. of the island Koh Chuen, 3/III. 1900 (2 specimens), W. of Koh Kong, 25/I. 1900 (1 specimen), 165 Malacca Strait, off Pulo Pisang, 1/XII. 1899 (1 specimen) and 101°12’ E. 2°45’ N., Malacca Strait, 30/XI. 1899 (6 specimens). Species b. (Pl. XXII, Fig. 2). The inner arms are distinctly shorter than in species a, only of about body length, and especially the postoral and posterdédorsal arms are conspicuously shorter than in species a. The backward process from the posterolateral rods is much shorter than in the former species and has only very few serrations; sometimes it is quite rudi- mentary; there is no process from the body rod, and the median process from the transverse rods is quite short. The anterolateral rod is provided with rather large thorns along the outer side (Fig. 90). The specimen figured presents a series of well defined spots in the vibratile bands along the arms, especially the posterolateral ones. Most probably these are pigment spots; but this can only be definitely ascertained by the study of living specimens. This species was found off Jolo, 20/III. 1914 (1 specimen) and off Co- lombo, 13/II. 1898 (6 specimens) (Mr. H. Mortensen). It is not quite easy to distinguish this species from O. pusillus species c¢, and the possibility cannot be denied that both are really the same, very variable species; the fact that they were taken together (Colombo, 13/II. 1898) would also point in this direction; but no definite conclusion can be reached from the preserved material, and I have thought it preferable to keep them as two separate species in view of the marked difference between typical specimens. A suggestion as to the Ophiurid genus to which these larve belong, cannot be made for the present. Ophiopluteus pusillus nova forma. The characters which distinguish this larval type are the following. The body skeleton is of the compound type, often with a process from the recurrent rod; the transverse rods are slightly widened at the point, with some short obtuse prominences, fitting into one another; no median pro- cess. The anterolateral rod is set with rather coarse thorns along the outer side in the outer part. The posterolateral arms are rather short, only about twice the body length, not much outwardly directed; the anterolateral arms are somewhat less than body length; their points are at about the level of the points of the posterolateral arms. The postoral and the postero- dorsal arms are much shorter than the other arms and hardly reach beyond the frontal edge. Characteristic of this larval type is also its small size, which, in the fully formed larva, is only ca. 0.5—0.7 mm in total length, 166 from the posterior end of the body to the point of the anterolateral arms. Several species of this larval type are at hand. Speciesa. (Pl. XXIX, Fig. 1). The length of the body from the post- erior end to the frontal edge is ca. 0.34 mm. The frontal area is rather large. The postoral band may curve much downwards, as seen in the figure; this is, however, no constant feature and, evidently, to some degree at least, depends on preservation. The preoral band curves upwards, pl Fig. 91. Skeleton of Ophiopluteus pusillus, species a. Letters as in fig. 85. but remains straight in the middle part. The posterolateral arms are about 11/, body length. There is no distinct widening at the point of the arms. The body skeleton (Fig. 91) is slender, the meshes being rather large, rectangular; the end rods are fairly long and slender, with one or a few small side branches somewhat above the tip, which is very gently curved; the outer edge near the point may be finely serrate. There is no process from the recurrent rod. The posterolateral rods are provided with short thorns of about equal size on both the outer and the inner side, the postoral and posterodorsal rods with small thorns, placed mainly along the ventral side of the former and along the dorsal side of the latter, so as to appear mainly as transverse lines on these rods in ventral or dorsal view of the larva. 167 A dozen specimens of this larva were found in a plankton sample from Misaki, Japan, 17/VI. 1914. A specimen from off Koh Kam, Gulf of Siam (4/11. 1900) differs from the Japanese specimens in lacking the small side- branches of the end rods, but otherwise so closely resembles the latter that I have no doubt in referring it to the same species. Species b. (Pl. XXIX, Figs. 2—3). This species agrees with species a in the character of the spinulation of the posterolateral and the other rods, Fig. 92. Skeleton of Ophiopluteus pusillus, species b. *°°/,. but differs markedly from the latter species in the shape of its body skeleton (Fig. 92). The meshes are not quite rectangular, but generally distinctly narrowing downwards, and there is a conspicuous process from the recurrent rod. The transverse rods are somewhat longer than in species a, in accordance with the insertion of the recurrent rod nearer the body rod. The end rods are somewhat shorter and quite straight, and the side branches are placed close to the tip. Otherwise it agrees very closely with species a in its general shape, only the postoral and posterodorsal arms being still somewhat shorter; they reach only to about the level of the preoral band. A very interesting stage of metamorphosis is represented in Pl. NXIX, Fig. 3. As in Ophiopl. formosus, sp. a (Pl. XXX, Fig. 2) and O. undulatus, 168 sp. a (Pl. XXIV, Fig. 3) the vibratile band is seen to pass directly across the body, below the developing Ophiurid, so as to continue directly from one posterolateral arm to the other. A distinct anterior band is represented by the former preoral band of the larva, and moreover in this specimen distinct traces of other vibratile bands are seen, the accordance with the pupa-stage of the Holothurians being thus still more emphasized. There can then be no doubt, that we have really in these Ophiurid larve a true pupa-stage. Also in the metamorphosis-stage of O. formosus there are Fig. 93. Ophiopluteus pusillus, species c. A. body skeleton; B. outer part of postoral rod. 79/,. very slight indications of small parts of vibratile bands in the median part of the body. Most probably they will prove to exist in all forms having a distinct pupa-stage. This species was found in the Gulf of Panama, at the island of Taboga, in December 1915; 12 specimens. Species c. (Pl. XXIX, Fig. 4). The present species differs markedly from the two preceding species in the character of the posterolateral rods, which are provided with long thorns along the inner side, but entirely smooth on the outer side. The body skeleton (Fig. 93, A) forms a pair of rather small, oval meshes, and there is a long often somewhat thorny pro- cess from the recurrent rod; this process is directed ventrally or dorsally, and it is therefore generally broken in the preparations, as also seen in the figure. This thorn may, however, sometimes be lacking. The end rods are rather long, gracefully curved, with a small side branch at some 169 distance from the point. The backward projecting process from the trans- verse rods seen in the figure is no constant feature. The thorns on the postoral rods and apparently also on the posterodorsal rods are very long, and lend this rod a very peculiar pectinate appearance; but as they are directed ventrally they can be seen distinctly only when they (or the rod itself) are (is) broken. As seen in Fig. 93, B they may be bifid or even trifid at their base, and there may further be a small side branch near the point. Whether this complicate structure is a constant feature is hard to ascertain, as it can only be seen when the rod is broken and lies in a favourable position. Neither can the possibility be denied that these minor differences really are specific characters, so that more than one species may be included under the larval form designated here as O. pusillus, species c. This cannot be decided from the study of preserved material alone. None of the other species of the O. pusillus type appear to have such long thorns on the postoral rods. — The body of the present species is somewhat shorter and comparatively broader than that of the two pre- ceding species. Also the frontal area is comparatively lower and the preoral vibratile band less arched. A considerable number of specimens were found in a plankton sample taken by my brother, Mr. H. Mortensen, off Colombo, Ceylon, 13/I1. 1898. Further I have specimens from the Gulf of Aden, 3/V. 1900 (1 specimen), and the Malacca Strait, 30/X1I—1/XII. 1899 (8 specimens). This species has several points of resemblance with Ophiopl. retrospinus, species b, and probably they are nearly related. The possibility that they are really only one, very variable species cannot even be denied. Only the study of living material can decide the question. Species d. (Pl. XXIX, Fig. 5). The shape of the body is like that of species c, short and rather broad, with a very small frontal area; it differs from the other species in the posterior extremity of its body being more rounded. As seen in the figure the nervous system may be distinctly observ- able, showing the typical arrangement?). This species differs from the preceding ones in the shortness and rather robust character of the body skeleton (Fig. 94); there is no process from the recurrent rod. The end rods are short, straight, and with the small side branch close to the point. The posterolateral rods are provided with long thorns along the inner side, the outer side being smooth. The thorns of the postoral and posterodorsal rods appear to be more bilaterally ar- ranged than in the other species and are especially much less developed than in species c. ‘) Th. Mortensen. Notes on the development and the larval forms of some Scandinavian Echinoderms. p. 158—160. »*) 170 This species was found in the Gulf of Panama, at the Island of Taboga, in December 1915, a dozen specimens. Evidently it stands more apart from the other species and possibly does not really belong to the same type: still I have thought it preferable to include it under the type of O. pusillus, to which it bears in several regards a close resemblance. It is impossible to have any opinion of the relationship of this larval type. That it occurs both at Panama, Japan and in the Indian Ocean is by no means a sufficient foundation for any suggestion as to this point. Too many Ophiurid genera are common to the Pacific coast of America and the Indo-Pacific region to allow any such conclusions at our present stage of knowledge. Fig. 94. Skeleton of Ophiopluteus pusillus, species d. **°/1. Ophiopluteus diegensis nova forma. Pl. XXIX, Fig. 6. In its general appearance this small larva (0.27 mm body length) very much resembles Ophiopluleus pusillus. The posterolateral arms are scarcely longer than the body, and the point of the anterolateral arms is at about the level of that of the former ones, while the postoral and posterodorsal arms remain very short. The frontal area is quite low. The posterolateral rods are set with short thorns along both the outer and the inner side, and the anterolateral rods have rather long thorns along their outer side, as in pusillus. The postoral rods are strongly curved, with few, small thorns. The body skeleton is simple; there is generally a small thorn on the outer side near the base of the body rod. The end rods are straight, with a small branch close to the point. The transverse rods have the same structure 171 as in O. pusillus, with small obtuse, interlacing branches at the slightly swollen end; there is no median process (Fig. 95). — In the posterior part of the body some comparatively large grains, evidently pigment cells, are seen. In the specimen figured the hydrocoel has not yet begun to form the primary lobes; this is, however, the case in other specimens, and as the arms are not longer in these than in the one figured, it seems beyond doubt that the arms will not be much longer in the fully formed larva. Fig. 95. Skeleton of Ophiopluteus diegensis. **°/,, Letters as in fig. 85. A number of specimens were found in a plankton sample from San Diego, California, 7/IX. 1915. This species bears so close a resemblance to Ophiopluteus pusillus, espe- cially species b, that the suggestion lies at hand that it is really related to it, in spite of the apparently essential difference in the structure of the body skeleton. On the other hand, it also bears some resemblance to Ophiopluteus fusus. At our present state of knowledge it is impossible to ascertain which is the correct reference, especially so long as we do not definitely know the value of that conspicuous character of the body skeleton, the simple or the compound structure. Ophionereis squamulosa Koehler. Pl. XXXI, Figs. 1—6. During my stay at Tobago, B. W. I., with the Carnegie Expedition I had the good fortune of seeing (on the 13th of April 1916) a specimen of this species discharge its eggs, which were at once fertilized, a male speci- 99* ae 172 men among those kept in the same dish discharging its sperm at the same time. The eggs were few in number, amounting only to ca. 90; the specimen, however, still contained some more, so that, evidently, not all the eggs are discharged at the same time, as is the case in Ophiolhrix angulatus, in which they are all discharged at one time, almost as by an explosion. The specimen was not observed to assume a special attitude while dis- charging its eggs, such as is the case in Ophioderma brevispina, according to the observations of Grave. Owing to the small number of eggs the material was insufficient for a complete study of the details of the interior changes during the meta- morphosis, a study which would have been of the greatest interest for a comparison with the results obtained by Caswell Grave!) in Ophioderma brevispina. Although I tried many times to get another culture, I did not succeed, so I had to make a very careful use of the little material avail- able and to preserve no more specimens of the different stages than strictly necessary. The size of the eggs is 0.2 mm; they are red coloured. They were lying free at the bottom, not floating at the surface, as did those of O. brevispina. The cleavage is regular and total. In the course of 15 hours the embryos were found to be in the gastrula stage, pearshaped, of a beautiful pink colour, especially towards the pointed anterior extremity; they were swim- ming close to the bottom. In the course of ca. 24 hours it became evident that the larva would turn out to be of the vermiform type, like that of QO. brevispina, as described by Grave. At that age the embryo has an elongate, triangular shape, with a narrowing below the anterior end, which thus becomes like a rounded knob (Pl. XXXI, Fig. 6). It is ciliated all over, without any ciliated bands as yet: what looks like a transverse band in the figure quoted is not really one; it is due to the fact that a transverse fold is formed in the posterior end of the body, on the ventral side, the apparent band being only the edge of the lobe. In this fold the gastrula mouth lies which is thus no longer in the posterior edge of the embryo: it appears to be closed already at this stage. The embryo keeps swimming close to the bottom as in its youngest stages. Pl. XXXI, Fig. 5 represents a frontal section of a larva 26 hours old. It shows a double sac at the upper end of the archenteron, apparently still in connection with the latter — but the histological preservation is not sufficiently good to allow the definite ascertaining of this question. The larger of these sacs doubtlessly represents the hydrocoel, while the smaller apparently is the anterior right enterocoel. In the posterior end two ") Caswell Grave. Ophiura brevispina. Mem. Nat. Acad. of Sc. VIII. Baltimore 1900. (Mem. Biol. Lab. Johns Hopkins Univ. 4). 173 accumulations of nuclei are seen which, no doubt, represent the posterior coelomic vesicles. There is an indication that one of them is in connection with the ectoderm, but it cannot be stated definitely to be the case. It is very unfortunate that the material available does not allow a detailed study of the process of formation of the hydrocoel and the coelomic vesicles, so that a comparison with Grave’s researches on O. brevispina might be made. Especially his statement!) that the epigastric coelom originates as a separate invagination from the ectoderm is so remarkable that it would have been of the greatest interest to see, whether it originates in the same peculiar way in this species also. There is, as stated above, an indication that it really does so, but it can by no means be ascertained as a fact. The specimen represented in Pl. XXXI, Fig. 6, although of the same age, is somewhat more advanced in its development, the hydrocoel having formed the five primary lobes. The small sac lying at the upper end of the stomach apparently is the right anterior coelomic vesicle; the posterior coelomic vesicles could not be discerned. — At the age of 40—45 hours the larva has reached its full development and has a very characteristic shape (Pl. XX XI, Figs. 1—2). The anterior end forms an elongate, rounded lobe with a ciliated band around it, situated a little below its middle; below and above the band the lobe is narrowing, at the point there is a large ciliated tuft, which is of a bright pink colour. In the middle of the posterior part of the body there is another spot of pink colour, larger than the apical spot, but less intense. It coincides with the central part of the young Ophiurid, which has already developed its first pair of tubefeet. Two ciliated bands are found in the posterior, main part of the body, one at the level of the mouth of the developing Ophiurid, the other at the posterior end. Both are incomplete; the band in the middle of the body consists of a median part on the dorsal side and two lateral parts, while the posterior band consists of two parts, a larger one on the right side and a smaller one to the left. Possibly the posterior band has been complete, surrounding a posterior lobe in a former stage, before the developing Ophiurid has reached so far in its development. That the band in the middle of the body is not the broken up rests of a formerly complete band would seem evident from the fact that in the next stage figured (Pl. XX XI, Fig. 3), of a larva 21/, days old, the band is much more developed and nearly complete across the dorsal side. At the age of 40—45 hours the skeletal plates begin to appear. In the 1) Caswell Grave. Ophiura brevispina. II. An Embryological Contribution and a Study of the effect of the Yolk substance upon development and developmental processes. Journ. of Morphology. Vol. 27. 1916. 174 specimen figured in Pl. XX XI, Fig. 2 the six primary plates of the dorsal side have appeared and three of the terminal plates have just begun to develop, beginning with the posterior radius, as is evident from the relative size of these three rudiments. On the ventral side no skeletal plates have been formed as yet. There is not the slightest trace of a larval skeleton. In specimens 21/, days old (Pl. XXXI, Fig. 3) the terminal plates have been fully formed, the unpaired, terminal tentacle lying within it as in a tube. (Although the skeleton has been dissolved in all the specimens of this age preserved, this fact can be distinctly ascertained). Inside the first formed pair of tentacles indications of another pair of much smaller out- growths from the radial canals are seen. This is in accordance with the interesting fact observed by Grave in Ophioderma brevispina, that the buccal tentacles are formed after the second pair. The same observation was also made by Krohn?) in a similar Ophiurid larva which was found at Madeira. It seems highly probable that this is really the usual order of appearance of the oral tentacles in Ophiurids. As stated above, the median band is nearly complete across the dorsal side, while on the ventral side it is, of course, interrupted by the mouth opening of the Ophiurid. The left part of the posterior band has developed towards the mouth so that there are now three distinct bands reaching the mouth edge, one to the right, two to the left side. The rest of the posterior band remains at the right side of the posterior end. At this stage the larva evidently has reached its fullest development, and it is worth mentioning that it was observed now to swim fairly actively, while on the other hand, it could also now attach itself to the bottom by means of its tubefeet, so firmly that it was hardly possible to wash it off by means of a pipette. — Also in the stage of 40—45 hours the larve were observed now and then to swim free in the water. At the age of 6 days the anterior lobe of the larva is still distinct, with its ciliated bands, but it is very much shortened and evidently in the course of being absorbed (Pl. XXXI, Fig. 4). The median ciliated bands have nearly disappeared, but a small trace of each of the three bands reaching the mouth in the foregoing stage is still observable at the edge of the body; the posterior band is still quite distinct. —- The primary skeletal parts have all been formed, and the teeth and the first spines have appeared. Very prominent is the torus angularis, a fact the more remarkable since this plate is hardly observable in the adult specimens. In the two anterior radii two ventral plates are seen, one inside the other; the inner one of them doubtlessly represents the first, rudimentary, inner 1) A. Krohn. Uber einen neuen Entwicklungsmodus der Ophiuren. (Miillers Archiv. 1857. p. 370. Taf. XIV. B. Fig. 1). 175 ventral plate; it is not formed in all the radii; I find it present sometimes in three radii, sometimes only in two or one, but not in more than three radii in any of the specimens available. — In fig. 96 part of a young specimen, 12 days old is represented. The second pair of lateral plates has begun to form; the inner (in reality the second) ventral plate has assumed an elegant hourglass-shape. The buccal plates have not yet been formed. Inside the ventral plate a pair of small, irregularly halfmoon-shaped plates has appeared, representing the adoral plates. (This interpretation is not in conformity with that usually adopted; I cannot, however, enter here on a discussion of the mor- phology of the oral skeleton of Ophiurids). When the young Ophiurids had reached this age, the so- journ at Tobago ended and put a stop to the experi- ments, and the few speci- mens left were preserved. But I cannot have the slightest Fig. 96. Part of a young Ophionereis squamulosa, 12 doubt that it would be very days old. ad. adoral plate; p. mouth papilla; L. kube- : foot; to. torus angularis; v. ventral plate. *°°/:. easy to rear the young Ophi- urids of this species to a much later stage, since it is very easy to give them suitable conditions. The species is found in quite shallow water on the coral reef, in the sand under old coral blocks, together with a number of other species of Ophiurids. The experiments carried out thus far indicate that here is a lot of interesting work to be done. Quite recently a notice on the development of another species of Ophio- nereis, O. Schayeri (M. Tr.), was published by H. B. Kirk?) indicating the highly remarkable fact that in this species the development is absolutely direct, no larval form apparently occurring at all, the young star forming already within the tough envelope by means of which the eggs are attached to the under surface of stones. Unfortunately, the identification of the species is not beyond doubt, the deposition of the eggs not having been observed and the young Ophiurids not having been reared until they could be identified with certainty. It would be exceedingly interesting to study the development also of other species of this genus; although thus (probably) two species of the genus have been shown to develop without the Pluteus larval stage, it is 1) H. B. Kirk. On the much-abbreviated development of a sand-star (Ophionereis Schay- eri). Preliminary Note. Transact. & Proc. New Zealand Inst. Vol. XLVIITI. 1916. p. 383—384. 176 quite possible that other species may have a Pluteus-larva, corresponding to the fact shown in this work that within the genus As/erina some species have a typical Bipinnaria-larva, while others have a shortened, direct development as in the classical case of Asterina gibbosa. — The larva of Ophionereis squamulosa belongs to the same type as that of Ophioderma brevispina, described by Caswell Grave. There are some minor differences in the shape of these two larve, that of Ophionereis squa- mulosa having the anterior lobe distinctly longer than that of Ophioderma brevispina. Also the arrangement of the vibratile bands is somewhat dif- ferent, especially the posterior band, as a comparison of Pl. XXXI, Fig. 3 with Pl. III, Fig. 22 of Grave’s Memoir will make evident. The statement of Grave (Op. cit. p. 83) that the usual Pluteus skeleton is formed in the larva of Ophioderma brevispina, while there is no trace of a larval skeleton in Ophionereis squamulosa, would appear to indicate a very important difference between these two larve. Later on, however, Grave has announced!) that he mistook the beginning skeletal plates of the Ophiurid for the larval skeleton. There is thus no larval skeleton in the: larva of Ophioderma brevispina either. A larva closely resembling that of O. brevispina was taken pelagically at Taboga, Gulf of Panama, 18/XI. 1915. Probably it belongs to one of the Panamic species of Ophioderma. It would seem superfluous to give a closer description or figures of it. — To the same larval type belongs that describ- ed by Joh. Miller?) from Triest under the designation “wurmférmige Asterienlarve’’, and also those described from Madeira by Krohn®). It may perhaps be suggested that the Mediterranean form belongs to Ophio- derma longicauda. In the “Echinodermenlarven d. Plankton-Exped.” I designated also these worm-shaped larve as Ophiopluteus. Hamann‘) objects to this designation, because there is neither a ciliated band nor a larval skeleton in these larvee, not to mention that the usual larval arms are totally lacking. I would, however, still think it justifiable to use this designation. In the larva of Ophiura affinis®), Ophiopluteus Metschnikoffi and O. Claparéedet we have different stages of the reduction of the typical Pluteus-shape. From 1) Caswell Grave. On the occurrence among Echinoderms of larve with cilia arranged in transverse rings; with a suggestion as to their significance. Biol. Bull. V. 1903. p. 173. 2) Joh. Miiller. Uber die Larven u. die Metamorphose d. Holothurien u. Asterien. Il]. Abhandlung. Abh. d. Akad. Berlin. 1850. p. 26. Taf. VI. 8—12. VII. 1—4. 8) A. Krohn. Uber einen neuen Entwicklungsmodus d. Ophiuren. Miiller’s Archiv. 1857. p. 369, 373. Taf. XIV. B. Fig. 1—4. Th. Mortensen. Echinodermenlarven d. Plankton-Exp. p. 65—66. ‘') O. Hamann. Die Schlangensterne. Bronn. Klassen u. Ordn. III. 1901. p. 860. 5) Th. Mortensen. Notes on the development and the larval forms of some Scandinavian Echinoderms. p. 135. 177 the latter species, in which the skeleton is reduced to a pair of simple, small rods, there is only a very small step to the one of Krohn’s larve, (O. elongatus) which was observed by that author also to have a rudiment- ary skeleton (— unfortunately, he does not give any figures of it —-). We have thus an unbroken series from the typical Ophiopluteus through the worm-shaped Ophiopluteus elongatus with a larval skeleton to the larve of Ophioderma brevispina and Ophionereis squamulosa with no larval skeleton, and further on to Ophionereis Shayeri with the larval stage completely eliminated. Amphiura vivipara H. L. Clark. Pl. XXXI, Figs. 7—9. The discovery of this new viviparous Amphiura was one of the results of the Carnegie Expedition to Tobago, B.W. I., in April 19161). As empha- sized by Ludwig in his paper on “Brutpflege bei Echinodermen’’*) the viviparous Ophiurids hitherto known all belong to the colder regions (arc- tic-subarctic or antarctic-subantarctic), with the sole exception, besides the nearly cosmopolitan Amphiura (or, more correctly, Amphipholis) squa- mata, of Hemipholis cordifera Lyman (elongata Say); moreover, the vivi- parity of the latter is, no doubt, only apparent, as I have shown recently. *) It was then of considerable interest to find a truly viviparous Ophiurid living here, even under the most tropical conditions, viz. on a coral reef. (It was mainly found in thick cushions of Corallina, which covered large patches on the reef lying dry at ebb tide). It was at once evident from the considerable size of the eggs, 0.5 mm, that the interest attached to this case was more than that of having found a tropical viviparous Ophiurid. In the only viviparous Ophiurid hitherto studied, Amphipholis squamata, the eggs are very small, only ca. 0.15 mm, and the cleavage total and regular; the embryo has a rudimentary larval skeleton and is, evidently enough, only a reduced Pluteus. In the present species all this is different. First of all the cleavage is meroblastic. The nuclei, in the first cleavage stages, lie irregularly spread in the yolk substance and there is no trace of cell division; then they gradually arrange themselves in a more regular order along the surface of the egg, forming thus the ectoderm (Pl. XX XI, Figs. 7—8), while in the interior they remain without distinct order till a much later stage. The details of the embryonal and postembryonal development cannot be given here, as | cannot spare 1) H. L. Clark. Brittle-Stars, new and old. Bull. Mus. Comp. Zool. LXII. 6. 1918. p. 268. *) Zoologische Jahrbiicher. Suppl. VII. 1904. 3) Th. Mortensen. On Hermaphroditism in viviparous Ophiurids. Acta Zoologica. lI. 1920. p. 4—5. 23 178 the time, at present, for this apparently very diffieult study. It must suf- fice at the present occasion to state only that there is formed at the place destined to form the oral side of the Ophiurid an accumulation of nuclei, like a germinal disk, from which the formation of all the main structures of the Ophiurid proceeds, while the large yolk mass remains undifferent- iated, simply as a reservoir of food, covered only by a thin ectodermal coat. The arms of the young Ophiurid gradually as they grow are folded up over the yolk mass (Pl. XXXI, Fig. 9). The embryos proved unfavourable objects for a detailed study of the development of the oral skeleton. The usual primary plates are formed in the skin of the dorsal side. There is no trace of a larval skeleton. — The embryos are not liberated until they have reached a comparatively large size, with ca. 10 armjoints; the paired infradental mouth papille have then already appeared; they do not appear till the teeth have been fully formed, the latter being among the first parts of the oral skeleton to appear. That this type of development will prove to be of more general occur- rence in viviparous Ophiurids can hardly be doubted, viz. in such forms as have large eggs, rich in yolk; but not in all of them. In Slegophiura nodosa (Ltk.), the eggs of which are 0.5—0.6 mm in diameter and rich in yolk the cleavage is total and regular. A detailed study of the develop- ment of these two Ophiurids, both having large, yolk-laden eggs, but one of them having superficial, the other total and regular cleavage, would, evidently, be of exceptional interest. This is, however, entirely out of the scope of the present work. But it is worth emphasizing that Amphipholis squamata is so far from representing the usual type of development among viviparous Ophiurids that it appears in reality to be exceptional in this regard. The interesting observation that Amphiura vivipara is hermaphroditic like Amphipholis squamata may be recalled here. It led to an examination of other viviparous Ophiurids which resulted in the establishment of the surprising fact that out of the twenty viviparous forms of Ophi- urids known sixteen are hermaphroditic, three alone having separate sexes (one remaining unknown as regards its sexual characters). This fact, the more conspicuous since not a single case of hermaphroditism is known otherwise in Ophiurids, shows that there is some relation between viviparity and hermaphroditism in Ophiurids, though it remains enigma- tical, wherein this relation consists. (Comp. the author’s paper “On herma- phroditism in viviparous Ophiurids’’). 179 Ophionotus hexactis (I. A. Smith). Pl SOIT In the paper quoted above “On hermaphroditism in viviparous Ophi- urids” (p. 17) the observation was mentioned that this species, in spite of being viviparous, hasa true, well formed larva, the develop- ment being intraovarial. While in that paper the hermaphroditic cha- racter of the species was established, I may here take the opportunity of giving a description of the remarkable intraovarial development of the no less remarkable larva. The previous observers, Studer!) and Lyman?) assumed that the young were enclosed in the much widened and very thin-walled bursz, or “rather, perhaps, in pockets leading out of the burse’’; this latter expression shows that Lyman has seen something of the real fact; the understanding of the whole astonishing fact could, however, not be gained through simply examining how the large young are lying within the parent specimen. Only an examination of the ovaries could disclose the fact, almost unique among the Echinoderms’), that the young are really lying within the distended ovaries. The structure of the young ovaries differs somewhat from that generally observed in the female gonads of Ophiurans. While generally they are completely filled with eggs so as to be solid organs, they are in O. hexaclis hollow vesicles, the eggs attached to the walls not nearly filling out their lumen (Pl. XXXII, Fig. 9). The eggs are of various sizes, one or a few of them being larger than the rest; the size of the ripe egg is about 0.2 mm. When ripe the egg falls into the cavity of the ovary, but it does not pass out into the bursa, as should be expected; it remains within the ovary and, after being fertilized there, goes on developing, the embryo passing its whole development within the ovary. Gradually as the embryo grows, the walls of the ovary are distended. The other eggs remain attached to the wall as flat buttons, gradually wider and wider apart from one another. In the same time they gradually di- minish in size and ultimately they disappear completely. It seems beyond doubt that they must serve as nourishment for the young contained within the ovary. Traces of eggs on the ovarian walls may still be seen when the young has about 4—5 armjoints (Pl. XXXII, Fig. 8). As the young grow ') Th. Studer. Uber Geschlechtsdimorphismus bei Echinodermen. Zool. Anzeiger. 1880. *) Th. Lyman. Ophiuroidea of the “Challenger”. 1882. p. 41. Pl. XLV, Fig. 1; PI. XLVII, Fig..2. %) Intraovarial development is known to occur only in one other Echinoderm, Chiro- dola contorta Ludw. Ludwig. Holothurien d. Hamburger Magelh. Sammelreise. 1598. p. 77—81. to * 180 to a relatively very large size, — about 8 mm diameter of disk and ca. 20 mm length of the arms, which may have up to about 40 joints — before they are liberated, the ovaries must be distended to quite an enormous extent, and on account of the growing long arms of the young, which need con- siderable room even if bent up, must occupy any space left between the stomach and the body wall. These are the sacs observed by Studer and Lyman and taken by these authors to be the burse themselves (Studer) or pockets from the burs (Lyman); as a matter of fact they have nothing at all to do with the burs; they are nothing but the distended ovaries. The burs themselves are of the typical shape, not at all especially widened, and they never contain young. The question, how the young are liberated cannot be directly answered. Considering the great number of young found in adult specimens, it should be expected that specimens having emptied out their young would show distinct traces of the way the young have been passing. Only three ways are possible; either they must pass directly through the body wall, rup- turing the sac within which they have developed, or they must pass through the bursal wall, likewise after rupturing the ovarial sac, or finally there may be a natural opening from this sac through the bursal wall, through which they may pass. As there is never seen any traces of scars on the disk, and not on the bursal wall either, in specimens which have emptied the young, and the empty ovarial sacs are likewise unruptured, there can hardly be any doubt that there must be an opening from the ovarial sac through the bursal wall. As no such opening can be seen, it must be capable of widening to a very large extent and then completely closing again. — Through this opening also the spermatozoa must enter for fertilizing the egg. — It is a remarkable fact, however, that nobody appears to have observed specimens with the young in the act of being liberated, such as is seen so very often in other viviparous Ophiurids. After the liberation of the young the ovarial sacs shrink very consider- ably, the wall contracting and becoming much thicker; it remains attached to the body wall and the stomach and bursal walls through numerous trabecules of connective tissue, as observed by Lyman. But they do not disappear or reassume the original shape of an ovary. The propagation begins at a very early stage; already in a specimen of 13mm diameter of disk a large young was found, so that the sexual activity must have started here at a size of only ca. 12 mm diameter of disk. In another specimen of 16 mm only larve, not yet metamorphosed, were found. But this specimen was unusually late in beginning its sexual activity, which may be stated to begin, as a rule, at a size of ca. 13—14 mm diameter of disk, that is to say very soon after they are born. In such 181 young specimens generally only the ovary at the outer corner of the bursal slit is found to contain a young, and mostly of the same size in all the radii. A very large specimen, 39 mm diameter of disk, was found to have all the young, except two of them, liberated. The ovarial sacs were con- tracted as described above. In the same time new ovaries, containing young developmental stages, had been formed; so far as could be ascer- tained these new gonads were formed within the old ovarial sacs. In this specimen in two of the radii testes had been formed in the former female gonad on the corner, even two testes in each. Otherwise the testes were in the normal position and in the usual number of three. Another specimen of 37 mm diameter of disk was quite empty; but here no embryos were found in the ovaries; the gonads were evidently abortive and the specimen quite senile. It was probably such specimens that were taken by Studer to be male specimens. Generally only one young is found in each ovarial sac. In the said large specimen, however, there were in several cases two embryos in the same ovary, mostly in exactly the same stage of development. One ovary was found to contain three embryos in different stages and another even six embryos, in two different stages of development. The embryonal development proves to be of exceptional interest. Clea- vage stages were not observed, but it cannot be doubted that it is total as in Amph. sguamata; this may be safely concluded from the fact that a regular blastula is formed. The embryo develops into a compara- tively well formed larva, with a distinct vibratile band anda rudimentary larval skeleton (Pl. XXXII, Figs. 1—6). The vibratile band remains in a primitive condition, corresponding, in fact, exactly to the diagrammatic figure of the ideal type of an Ophiurid larva represented in Joh. Miiller’s classical memoir “Uber den allgemeinen Plan in der Entwickelung der Echinodermen” (Taf. II, Fig. II. 2")). No larval arms are developed, and the larva thus retains a shape like a young Auricularia; the rudimentary larval skeleton, however, shows it to be a rudimentary Ophiopluteus. Nothing can be said about the typical character of this larval skeleton, it being altogether too rudimentary to be relied upon in this regard, and no conclusions can be drawn as to its relation to other, free Ophiurid larve. It is almost always seen to consist of two main parts, as in typical Ophiurid larve, but I never saw an embryo with both parts equally developed. The posterior end of the larval body is produced into a tip, bearing a rather large tuft of cilia, as it is found, e. g. in the larva of Ophiura albida and in several other Ophiurid larvee. Mostly the larve are quite regularly 1) Abh. d. Berliner Akad. 1853. 182 shaped, sometimes, however, they are more or less irregular, the vibratile band forming lobes, as in the specimen figured in Pl. XXXII, Fig. 7. The interior organisation of the larva is quite typical, though, as might be expected, the digestive organs are to some degree rudimentary, in ac- cordance with the fact that the larva has no opportunity of feeding on other organisms, as do the free pelagic larve. The esophagus is somewhat less developed than in the typical larvee, and especially the mouth opening is smaller, sometimes irregular in outline. The stomach is large and of the typical shape; the rectum is sometimes quite unusually large, but apparently there is no anal opening (Pl. XXXII, Fig. 6). A more or less distinct line of nuclei in the oral area (Pl. XXXII, Fig. 3) would appear lo represent the larval nervous system. The enterocoel and hydrocoel vesicles are large and very distinct, the hydrocoel having at first only 5 lobes; the sixth appears to form between the 2nd and 3rd primary lobes. On metamorphosis the hydrocoel grows upwards over the esophagus. In some cases there is a distinct right anterior vesicle (Pl. XXXII, Fig. 1); in two cases a normally lobed right hydrocoel was found, no trace of a left hydrocoel being observable (Pl. XXXII, Fig. 4). There is not the slightest reason to doubt that these “right’’-specimens would have developed normally. This case is quite unique in Echinoderm embryology. Larve with both a right and a left hydrocoel have been observed repeatedly, rarely in the free, but not so rarely in cultures, and, as shown by MacBride?) can be produced artificially; but, so far as I know, larve with a right hydrocoel alone have never before been observed, in nature or in artificial cultures. The theo- retical interest connected with this case is very considerable. This is, how- ever, not the place to enter on a discussion of this matter. Regarding the structure of the young Ophiurid I-would only call atten- lion to the fact that there is not one prominent central plate as is usually the fact in young Ophiurids, but two or three of somewhat irregular shape (Pl. XXXII, Fig. 8). Mostly the young Ophiurids have the oral tentacles very much extended (— the specimen figured is exceptional in this regard, having the oral tentacles for the greater part retracted —); they are further conspicuous by the granulated appearance of their point. The sug- gestion lies near that these tentacles may perhaps have some kind of absorbing function. In any case it is clear that the food contained in the substance of the small egg can by no means account for the large size 05 attained by the young within the brood sac. Nourishment from the parent ‘) E. W. MacBride. The artificial production of Echinoderm Larve with two water- vascular systems, and also of Larve devoid of a water-vascular system. Proc. R. Soc. B. Vol. 90. 1918. p. 323—48. Pls. 4—10. 183 specimen must be absorbed in some way or other. At first it is evidently the other, not developing, eggs within the same ovary that serve as food for the growing young, not directly, but by being dissolved gradually and thus absorbed, perhaps by the digestive organs of the larva and the young Ophiurid. But since these eggs have completely disappeared a long timé before the young Ophiurid has reached the size at which it is to be liberated, it must be fed on some sort of nutritive fluid secernated by the parent specimen. It would not seem improbable that the oral tentacles of the young have something with the absorption of this nourishing fluid to do. The remarkable fact that this species, although representing one of the most highly specialized cases of viviparity, passes through a distinct Plu- teus-stage, more developed than many a truly pelagic larva, evidently has an important bearing on the question of the phylogenetic meaning of this larval form. But that is for another opportunity. Ill. Asteroidea. The Asteroids, upon the whole, lend themselves rather easily to artificial fertilization and are among the favourite objects of experimental embry- ological studies, although much less so than the Echinoids. A curious fact which renders them somewhat less favourable objects for such a study is this that the spermatozoa are very often found to be quite immovable when taken from the testes and diluted with water, though the specimens be evidently perfectly ripe and ready to empty the content of the gonads. It is, of course, impossible to obtain fertilization by means of such im- movable sperm. On adding, however, some drops of sodium hydroxyd (NaOH) to the water and thus raising the alkalinity it is easy enough to stir the spermatozoa to activity and then fertilization is easily ob- tained!). But it is surprising how much NaOH must sometimes be added, before they begin to move. The spermatozoa of Asteroids are evidently exceptionally sensible to changes in the alkalinity of the water, that is to say, they cannot stand any lowering of the normal alkalinity of the water, while they do not object to an amount of alkalinity considerably beyond that of the water of their normal surroundings. The eggs are very rarely found to be fully mature when taken out of the ovaries, even though the specimens look ever so ripe; they have nearly always a large germinal vesicle. But if they are left for some hours in the water, the germinal vesicle disappears, and they are then ready for fer- tilization. 1) Comp. the author’s paper “On the development of some British Echinoderms’, p. 7 (Luidia ciliaris). 184 Many starfishes have large eggs, rich in yolk. It appears to be next to impossible to obtain fertilization of such eggs, unless they are shed in the natural way by the females. In this regard the Asteroids — together with the Holothurians — are in remarkable contrast to the Echinoids, in which large and yolky eggs are as easily fertilized as are the small, transparent eggs, with little yolk substance. The development of the Asteroid-larve generally takes a rather long time, which fact partly accounts for the comparatively poor results achieved in the rearing of these larvee. A stay of two—three months in some place is, in general, not long enough for obtaining satisfactory results in the rearing of Asteroid-larvee. The fact that the main characters of the larvee do not appear until they have reached their full size and are near metamorphosis, makes this only more evident. The rearing of the young larval stages alone does not serve to give much more than the proof that the species in question has pelagic larve, while in the Echinoids even the first larval stage may give very important information of the larval characters. To rear the young starfish from larve in beginning metamorphosis, taken pelagically, is generally easy enough, and much may certainly be achieved in that way, under suitable laboratory conditions. But a fairly long time will be necessary for rearing them far enough to identify them with cer- tainty; this cannot be achieved during a stay of two—three months. — The experiments of Mead?) showing how exceptionally rapid growth may be in these animals, if a rich food supply is given them, while on the other hand a poor food supply delays growth very considerably, is of special interest in this connection. (That the food supply acts in the same way on the larve, accelerating their development when available in good quality and sufficient quantity and delaying it when poor and insufficient, is a well known fact, which I have also had a rich opportunity of observing). The lacking of a skeleton interferes in some way with the preservation of Asteroid larvee. Not that they are difficult to preserve well, if carefully treated; on the contrary! When taken up a little cautiously in a pipette and then dropped directly into alcohol or formalin they will generally not contract at all; even such forms as have long, contractile arms may be preserved perfectly in this way?). Thus far the Asteroid larve are much more easily preserved in a satisfactory way than Echinoid- and Ophiurid larvee, which will nearly always contract as much as the skeleton allows. 1) A. D. Mead. On the correlation between growth and food supply in starfish. American Naturalist. XXXIV. 1900. (No. 397). 2) Still better results may be achieved with such larve if they are narcotized, by means of magnesium sulfate, drops of alcohol or the like, before they are killed. — 185 But on a less careful preservation the result will mostly be very poor; which accounts for the fact that Asteroid larve from plankton samples are generally of little use, being often unidentifiable, in contradistinction to the Echinoid- and especially the Ophiurid-larve, where the skeleton affords sufficient characters for identification, so that such larve may be very well utilized, even if the soft parts of their bodies are in a poor state of preservation. — In the present work I have made little use of such material of Asteroid larve from plankton samples, although I have quite a good deal of them, some of them representing types of Bipinnarie hitherto un- known. They are, however, not of so considerable importance that I deem it desirable to publish descriptions of them, which cannot be in any way satisfactory. Astropecten scoparius Miiller & Troschel. Pl. XXXII, Figs. 3—5. This species, which occurs abundantly in shallow water near the Bio- logical Station at Misaki, was found to have ripe sexual products in June; fertilization was undertaken on June 21st 1914, the sperma being treated with NaOH. The embryo is in the blastula-stage very much folded, look- ing almost like a walnut, the folds disappearing when the embryo leaves the egg-membrane. I have no notices on the progressive development of the embryos, so that the details of the transformation of the embryo into the Bipinnaria-larva cannot be given. At the age af 7 days they had devel- oped into beautiful Bipinnarie, the enterocoel pouches having, however, not yet united in the anterior end of the larval body. At the age af 19—21 days the larve were in beginning metamorphosis. Beyond this stage it was not possible to rear them, my stay at Misaki ending by that time. The larva (Pl. XX XIII, Figs. 3—5) is of a very typical Bipinnaria-shape, with short, rounded processes. The preoral band forms a high arch, while the postoral band curves only slightly downwards in the middle. The postoral processes are very distinct, as are also the posterolateral, postero- dorsal and anterodorsal processes. The median processes are of about equal size, short, rounded, the ventral one being somewhat ventrally directed (this accounts for its apparently being much shorter than the dorsal pro- cess in Fig. 5). Preoral processes are not distinct, the frontal area only being narrowed at the base of the median process, especially so in the metamorphosis stage; also at the base of the dorsal median process there is a distinct narrowing, the vibratile band from the two sides sometimes nearly joining in the midline. The frontal area is distinctly convex, the sides lying much deeper than the middle part, as is especially distinct in the metamorphosis stage. — There is a distinct suboral cavity, narrow 24 186 but fairly deep, reaching down to the upper edge of the stomach or even some distance beyond. Fine, redbrown pigment grains are scattered all over the body. The length of the fully formed larva is ca. 0,5—6 mm. Although metamorphosis was not completed, the development pro- ceeded far enough to show definitely the important fact that this larva has no Brachiolaria-stage; this is what was to be expected from the evidence of the only other Astropeclen-species, the metamorphosis of which has been studied, viz. Astropecten aranciaca (the “Bipinnaria von Triest”’ of Joh. Miiller, IV. Abhandl. Taf. V.). Astropecten polyacanthus Miller & Troschel. A few specimens of this species, containing fairly ripe sexual products, were obtained at Misaki on the 18th of June 1914, and fertilization was undertaken, resulting, however, in only a small cul- ture, the eggs not being as ripe as desirable, and therefore only comparatively few being fertilized. The sperma, as usual, had to be treated with NaOH in order to obtain active movement of the sperma- tozoa. Only a few of the embryos survived the ga- strula stage, and only a single specimen reached far enough (at the age of 10 days) to show the characters of the larvee. Moreover, this specimen was abnormal, its right side remaining in a rudimentary condition. But the left side of this larva was typically devel- oped, and, in spite of all, this specimen (Fig. 97) Fig. 97. Abnormal larva makes it sufficiently evident that the larva of A. of Astropeclen polyacan- ; 3 ; : thus. 10 days old, '%,, Polyacanthus is so closely alike that of A. scoparius that it will be hard to distinguish them. The fact that the postoral band is straight in the A. polyacanthus-larva, while it is concave in the larva of A. scoparius, is not likely to be a reliable disting- uishing character. The polyacanthus-larva differs, however, from the A. scoparius-larva in being devoid of pigment. It was also found to be con- siderably more active in its movements than the latter species. Asterina (Patiria) pectinifera (Miiller & Troschel). Pl. XXXII, Figs. 1—2. Th. Mortensen. On the development of some Japanese Echinoderms. p. 550. This strikingly beautiful starfish occurs fairly abundantly on the rocky shores near the Biological Station at Misaki. Fertilization was undertaken in the end of May and the beginning of June, but the cultures obtained 187 were not very good, the embryos developing only to the beginning forma- tion of the Bipinnaria; the somewhat unexpected fact that this species has a typical pelagic larva was, however, already proved by this first attempt. On the 18th of June I had the pleasure of seeing a female spec- imen, kept in a dish together with a number of other specimens, discharge its eggs, from which an excellent culture was obtained. I regret not having noticed the exact size of the eggs, only stating that they are small; I think I remember that they were of yellowish colour. The cleavage began very soon after the fertilization; the fact that I have made no notices about the cleavage process most likely indicates that it is of the usual, regular type. After 16 hours the gastrula stage was reached, and at the age of two days the embryos had the typical Bipinnaria-shape. The larvee were generally found swimming close to the bottom of the dish. At the age of 10 days the enterocoel pouches had united in the anterior end of the body. At the age of 18 days some few of the larve were in the metamorphosis- stage and, accordingly, had reached the full larval shape. The young larva (Pl. XX XIII, Fig. 1) looks very much like the As/ro- pecten-larva, with small, non-contractile processes; there are no postoral or preoral processes. The suboral cavity is small, not very deep. It is espec- ially noticeable that the vibratile band of the ventral median process is as distinctly developed as the rest of the bands, contrary to what is gener- ally the case in larve having a Brachiolaria-stage. It was, therefore, a great surprise to me to see that this larva in its final stage really is a Brachiolaria (Pl. XX XIII, Fig. 2), though of another type than the Bra- chiolaria of Aslerias. At each side of the fairly large sucking disk there is a small brachiolarian process with a few papille at the end. I was un- able to ascertain the exact number of these papille, as also to ascertain whether the vibratile band continues along the paired brachiolarian pro- cesses. The median process is only slightly transformed, retaining its flat shape, and its vibratile band bordering it as in the younger stage; only a row of small papillae, 3—5 in number, along each side, inside the band, indicates its brachiolarian character. The other processes remain unaltered, short and small. The length of the fully formed larva is ca. 0,6—7 mm. It is unpigmented. Asterina (Patiriella) regularis Verrill. Among a large number of specimens of this species which | collected on the rocky shore at the Island Bay, outside Wellington, New Zealand, on the 17th of February 1915 I found a few specimens to contain partly ripe sexual products. The artificial fertilization which I undertook was not 24* 188 very successful, still I succeeded in obtaining a small culture of larve, which I carried along with me to Sydney. They developed fairly normally unto the Bipinnaria stage, but I did not succeed in rearing them unto metamorphosis. — Later in the year, in June, Professor H. B. Kirk again undertook a fertilization of this starfish and sent me the larve he had reared. None of them, however, had reached a more advanced stage of development than those of my own culture. There is, accordingly, no de- finite proof that also this larva has a Brachiolaria-stage, but there is, on the other hand, no reason to doubt that it will prove to have a Brachio- laria-stage of the same type as that of A. pectinifera. The young larva (Fig. 98) is of the same type as that of A. pectinifera. The preserved material, including the larve sent from Prof. Kirk, is not in a very good state of preservation, the best specimen being also somewhat abnormal in the anterior end; I shall therefore not try to point out the specific differ- ences between the two larve, these differences being, evidently, very trifling and unimportant, at least as far as the young Bipinnaria-stage is concerned. I may here call attention to the fact, which has ap- parently been overlooked, that besides Asterina gibbosa one more species of Asterina, viz. A. (Patiriella) exigua (Lamk.) has been recorded to have a shortened develop- Fig. 98. Larva of ment. This observation was made by Th. Whitelegge, Asterina regularis; who in his “List of the Marine and Fresh-water Inverte- 9 days old. '%/1. The larva is slight- ly abnormal in the anterior end; the fold across the frontal area is prob- brate Fauna of Port Jackson and Neighbourhood” (1889, p. 40) states that “the ova of this species are de- posited on stones in shallow rocky pools and are to be found from June to December. They are particularly ably due to con- traction on preserv- ation. well adapted for study, inasmuch as after the young leave the egg-case they do not swim away, but remain round about the empty egg-cases and never leave until they assume the true starfish-form. The larvee are remarkably hardy and may be kept in confinement without change of water until they have passed through the larval stages.’’!) During my stay in Sydney I had the opportunity of confirming Whitelegge’s observations, in the beginning of March 1915, which shows that the breeding season of this species is considerably longer than stated by Whitelegge: more probably it will be found to breed all the year round. (In Asterina gibbosa the breeding ') I have thought it proper to quote Whitelegge’s short statement in full, as his paper may not be so easily accessible outside Australia. 189 season is confined to the months March—July, in the Mediterranean mainly to April—May). I shall not enter on a description of the developmental processes in A, exigua; no doubt, it would be of considerable interest to see how far this species agrees in this regard with A. gibbosa, but such a study would be out of the scope of the present work, the more so as the discovery that this species has a direct, shortened development, like A. gibbosa, is not mine. I would only call attention to the fact that the brachiolarian pro- cesses are in general more distinctly developed inthe = /" 7 ‘Selene ey . Divs embryos of A. evigua than eae abs) ey aS in those of A. gibbosa (Fig. ) Nee AEN ( eee ON | 99). The sucking disk is - | iF ‘laa not very much developed, a y \ y} still it is distinct enough. —— es ee interpretation, first Fig. 99. Embryos of Aslerina exigua, in different views. "5/1. set forth by Ludwig!) that these processes of the Aslerina-embryo are homologous with the Brachiolarian processes of the Astferias-larva, is definitely proved to be correct through the fact that the larva of Asterina pectinifera passes through a Brachiolaria-stage. In his paper “Uber die Genitalorgane der As/erina gibbosa’*) Ludwig mentions the fact (previously observed by Gasco) that the genital open- ings of Asterina gibbosa are found on the ventral side, while in all other Asteroids thus far observed they are found on the dorsal side. The fact of the two species A. pectinifera and regularis having pelagic larve raises the question whether the genital openings in these species are found in the same position as in A. gibbosa or on the dorsal side, as is the rule in Asteroids. It is very easily seen that there are no genital pores on the ventral side in the two said species; on the dorsal side it is not easy to ascertain, which of the pores seen there are genital pores, as they do not differ in size from the papular pores; on following the genital ducts they are however easily found, and it is thus ascertained that the genital pores are really situated on the dorsal side. In A. exigua they are found on the ventral side as in gibbosa. This contradicts the statement of Ludwig, in the paper quoted, who maintains that in A. pentagona v. Mar- tens, which is generally regarded as a synonym of A. exigua, the pores are found on the dorsal side. The discrepancy is probably due to As/erina 999 !) H. Ludwig. Entwicklungsgeschichte der Asterina gibbosa Forbes. Zeitschr. f. wiss. Zool. Vol. 37. 1882. (Morphologische Studien an Echinodermen. II. p. 154). *) Zeitschr. f. wiss. Zool. Vol. XXXI. (Morphol. Studien an Echinodermen. I. p. 290). 190 pentagona not being the same species as A. exigua as it is generally thought to be. It is also on zoogeographical grounds very probable that the spec- imens from Cape are not the same species as the Australian form. Un- fortunately, Ludwig does not state from which locality his specimens came. But, in any case, the fact thus ascertained that the genital pores are on the ventral side in the Australian specimens, on the dorsal side in specimens from some other locality (Cape?) shows that what has hitherto been included under the name of Asterina exigua is not all one and the same species. It can hardly be doubted that the ventral situation of the genital pores?) in A. gibbosa and exigua is not a character of greater systematic value (— the two said species are referred to different genera in Verrill’s revi- sion of the Asterinine?) —), but only an adaptation to their breeding habits, and it seems then legitimate to draw the conclusion that the same habits will be found in other species showing the same position of the genital pores. An examination of the material of Aséerina in the collections of the Copenhagen Museum, which contains most of the known species of the genus (s. lat.), shows that, with one single exception, these species have dorsal genital pores, and it may therefore be concluded that pelagic larve are the rule in the Asferina-group, direct development exceptional. Only in one case, besides A. gibbosa and exigua, I find ventral genital pores, viz. in a small Asferina from the littoral region of Misaki, Japan, which appears to represent a new species; at least it is not identical with any of the species mentioned from Japan by Goto?). It may accordingly be expected that also this species will prove to have a shortened develop- ment. — This species, moreover, is interesting in being selfdividing in its younger age, the only other case of selfdivision hitherto known among the Asterinide being Asterina Wega Perr. While the species is normally 5-rayed, the specimens arising through divisions are often 6- or 7-rayed. Only very small 6—7-rayed specimens were found. (Possibly this self- dividing form is not really the same species as that with the ventral genital pores). 1) It is interesting to notice that also the male specimens of A. exigua have ventral genital openings; Ludwig does not state, whether this is likewise the case in A. gibbosa, but it may seem probable from analogy with A. exigua that it is so. 2) A. E. Verrill. Revision of the genera of starfishes of the subfamily Asterinine. Amer. Journ. Science. 4. Ser. XXXV. 1913. p. 477. 8) Seitaro Goto. A descriptive Monograph of Japanese Asteroidea. I. Journ. Coll. Science Tokyo. Vol. XXIX. 1914. 191 Gymnasteria carinifera (Lamk.). Pl. XXXII, Fig. 6. This species was found, though not very commonly, in the rock-ponds near Hilo on Hawaii, and proved to have ripe genital products during the time of my stay there. Fertilization was undertaken on the 7th of April and proved successful. The next day the embryos were in the blastula- stage; the ectoderm is folded as in Astropecten, Luidia etc. At the age of two days they were in the gastrula-stage, being somewhat elongate in shape; on the third day the formation of the mouth and the enterocoel- pouches was beginning, and on the fifth day the embryos had the shape of young Bipinnariz, showing, when disturbed, the usual contraction of the dorsal side, so characteristic of young Bipinnarie. At the age of three weeks the larve were well developed Bipinnarie of the rather characteri- stic shape shown in Pl. XXXIII, Fig. 6. The most conspicuous feature is the great width of the dorsal side in the anterior part, which very con- siderably surpasses that of the frontal area. The preoral band is much raised in the middle, forming almost an acute angle; the frontal area is much narrowed at the base of the median process, but there are no preoral processes. Both the ventral and dorsal median processes are broadly rounded and rather unusually short, though not quite as short as they appear in the figure; the specimen figured is somewhat contracted on the dorsal side, these processes thus being somewhat dorsally directed so as to appear shorter than they are in reality. The posterolateral processes are very short, merely indicated, the posterior end of the body being rather narrow. There are no postoral processes; the postoral transverse band is straight. — The larva is unpigmented; I have noticed in the living larve the existence of fine, isolated muscle fibres. The body length of the larva is ca. 0,4 mm. Beyond the stage described the larve did not develop. The culture was kept for a little while yet (the specimen figured is 26 days old), but no further development was observed. The important question must then be left undecided, whether this larva has a Brachiolaria-stage. This is to be expected, since another form of the same family, Porania pulvillus, has a Brachiolaria-stage; but the proof is wanting. Ophidiaster Guildingii Gray. During my sojourn at Tobago, B. W. I., with the Carnegie Expedition in April 1916 I found some few ripe specimens of this species on April 19th, which were used for fertilization. The eggs are very small, only 0.11 * 0.12 mm. The cleavage is quite regular; the gastrula is formed, before the 192 embryo leaves the egg membrane. At the age of four days the embryos had assumed the shape of young Bipinnarie. By that time I had to leave Tobago; the culture was carried along with the other larval-cultures to Trinidad, but after a few days it was found to be all dead. Accordingly I can give no further information about the shape of the larva; but the fact here established that this species has a typical pelagic larva is in itself of no small interest. — I have not thought it worth while giving a figure of the young larva, partly because the few specimens preserved are in a rather poor condition, partly because the larve are not yet far enough developed for giving any information about the special characters of this larva. Asterias calamaria Gray. Th. Mortensen. Development of some Japanese Echinoderms; p. 551. The breeding season of this widely distributed pacific species was found to begin (in Japan) towards the end of June, and fertilization was under- Fig. 100. Young Bipinnariz of Astferias calamaria (Fig. A), Pisaster ochraceus (Fig. B), Euasterias Troscheli (Fig. C.) and Orthasterias leptolena (Fig. D.). A—C. '%/1; D. 1/1. taken on June 25th. The development of this species being rather slow, as usually in the Asterids, there was no hope of rearing the larve to full shape, as I had to leave Misaki already on the 12th of July. By that time the larve had reached the stage represented in Fig. 100, A. It is, of course, not sufficiently advanced for showing the specific characters, and it is, especially, not to be seen, whether it will develop into a Brachiolaria. The result achieved with this species is then in the main confined to the proof of its having pelagic larve. I have no notices about the first developmental processes. 193 The observation mentioned in the preliminary notice may be recalled here, that this species is very often infested with a parasitic organism (Cirripedian?), which castrates its host, the parasite occupying the place of the genital organs of the starfish. Pisaster ochraceus (Brandt). Among the numerous Asteroids occurring near the Biological Station at Nanaimo, Vancouver Isl., this species is especially conspicuous on account of its remarkable gregarious habits, generally assembling in shoals on rocks, where it may be above the water for several hours during ebb tide; the specimens are very closely aggregated, with interwoven arms. It was found to have ripe sexual products by the end of May, and fertilization was undertaken on May 24th. It was unusual in the spermatozoa being very active in natural sea-water, without addition of NaOH. The eggs are small, not very transparent. The blastula-stage was reached after about 24 hours (the ectoderm is not folded) and the gastrula-stage in the course of the second day; at the age of four days the embryos were beginning to assume the Bipinnaria-shape, and at the age of six days they were beautiful small Bipinnariz of typical shape. Although the culture was kept alive and apparently in good health for about four weeks the larve did not develop beyond this stage; in some of the specimens of this age the enterocoel pouches had begun to grow forwards into the anterior end of the body, but they had not yet fused. The shape of the larva is shown in Fig. 100, B (the specimen figured is 30 days old); it does not offer any marked special features. The fact that the ciliated band of the ventral median process is rather weak indicates that the larva is going to develop into a Brachiolaria. It is pigmented all over with fine, yellow spots. Euasterias Troscheli (Stimpson). This species, which is also very common at Nanaimo, although not so much so as the preceding one, (— it may also be found above the water for several hours during ebb tide, but it is not gregarious —), was likewise found to have ripe sexual products by the end of May. Artificial fertiliza- tion was undertaken on May 25th. The eggs are exceedingly numerous, very small, not very transparent; the spermatozoa were very active in natural sea-water, without addition of NaOH. The cleavage stages were very beautiful, perfectly regular. The embryos were not liberated from the egg membrane till after 24 hours, the gastrula stage being reached on the 95 194 second or third day; on the fourth day the embryos were beginning to assume the Bipinnaria-shape. The culture was kept for nearly four weeks apparently in good health (— the larve kept swimming near the bottom of the dish, not rising to the surface as usual —); but none of the larve reached beyond the young Bipinnaria-stage, the enterocoel-pouches not yet at that age having begun prolongating forwards. The shape of the Bipinnaria (Fig. 100, C; the specimen figured is 23 days old) does not offer any very marked features distinguishing it from the young larva of Pisaster ochraceus. It is provided with small, yellow pig- ment spots, especially along the vibratile band. Orthasterias leptolena Verrill. Fertilization of this species was undertaken repeatedly in June 1915 at the Biological Station, Nanaimo, but the cultures were not very successful, only in one case the young Bipinnaria-stage being reached. The eggs are of a faint reddish tint; the cleavage stages, blastula and gastrula are very beautiful, perfectly typical. The young Bipinnaria (Fig. 100, D; the spec- imen figured is 7 days old) differs rather considerably from that of the other forms in being broadly rounded in the anterior end; the vibratile band is not developed along the anterior edge of the body, which indicates almost certainly that the larva will develop into a Brachiolaria. Pycnopodia helianthoides (Brandt). This large and strikingly beautiful Asterid, which occurs quite com- monly at Nanaimo, was found to have its breeding season in May—June. Fertilization was undertaken repeatedly, but a good culture never resulted. The embryos never reached more than just beyond the gastrula stage. All the information acquired of the development of this species thus amounts only to this that it has pelagic larvee; but what they look like or whether they have a Brachiolaria-stage remains an open question. As stated above (p.185) the material of pelagic Echinoderm larve found in the plankton samples includes a number of Bipinnarie, among which some interesting new forms which would very well deserve to be figured and described. Unfortunately, they are nearly all of them in a poor state of preservation, so that it is hardly possible to give a reconstruction of their true shape. I have therefore thought it better to desist from describing them here. Only one form I shall briefly mention, viz. the larva in the Brachiolaria-stage represented in Pl. XX XIII, Fig. 7. It affords the inter- 195 esting feature that the ventral median process is transformed, carrying papille, only in its lower part, while its outer part retains its normal shape unaltered. This larva was.taken in the Red Sea, 6/V. 1900. Quite a considerable number of Asteroids apparently have direct devel- opment. This may fairly safely be concluded from the fact that they have large, yolky eggs. It is to be expected that these forms will have only a rudimentary larva as it is known in e. g. Solaster and Echinaster sepositus. I have observed the eggs to be of this character e. g. in Clenodiscus cris- patus, Hippasteria phrygiana, Ceramaster japonica, Mediasler levis, Psil- aster andromeda, Pontaster tenuispinus. In spite of many attempts I have never succeeded in obtaining artificial fertilization of any of these forms. IV. Holothurioidea. The Holothurians are, next to the Crinoids, the poorest of all Echino- derms for artificial fertilization. The numerous Dendrochirotes have, pro- bably, all of them large, yolk-laden eggs, which completely resist artificial fertilization; but also in the Aspidochirotes, where the eggs are generally small and transparent, artificial fertilization is rarely successful. I have myself succeeded in fertilizing and rearing the larva of Holothuria nigra (in Plymouth, 19131)), and in the present work I have three more fertilizations and rearings to record, viz. Slichopus californicus, St. Kefersteini and Holo- thuria sp. But in all these cases it was only a very small percentage of the eggs which were fertilized, and with many other species I tried in vain to ob- tain fertilization. It is also rather troublesome to get the necessary material for the study of the different stages of development, when only a small cul- ture is available. A much more easy way to obtain fertilization and normal development of the larve is that employed by Selenka?) and Edwards’), who put a number of specimens together in a large live-box, the cracks and upper side of which were covered by gauze, the box then being sunk to the bottom in an easily accessible place in the sea, in a depth of ca. 1'/, meters. Fertilized eggs were then soon after found in considerable numbers lying on the bottom of the box, and the embryos thus secured were found to develop normally. For obtaining the normal larve this method will doubtless prove to be excellent in many cases, not only in Holothurians but in many other kinds of marine animals which do not ') Th. Mortensen. On the development of some British Echinoderms. Journ. Mar. Biol. Ass. X. 1913. p. 17. *) E. Selenka. Zur Entwickelung der Holothurien (Holothuria tubulosa und Cucu- maria doliolum). Zeitschr. f. wiss. Zool. NXNVII. 1876. p. 157. 5) Ch. L. Edwards. The development of Holothuria floridana Pourtalés with especial reference to the ambulacral appendages. Journ. of Morphology. XX. 1909. p. 212. 95* a 196 lend themselves very easily to artificial fertilization nor to breeding in aquaria, I was not prepared to use that method during my voyage, partly because my former experiments with Holothuria nigra did not make clear to me the difficulties generally met with in the study of the development of Holothurians by means of artificial fertilization, partly because it would have been rather troublesome to carry along such a live-box on the long voyage. As is the case with the Asteroid-larvee, the larve of the Holothurians are as a rule not found very well preserved in plankton samples, and the material which I have gathered in that way is not very important, present- ing no new larval types. The contribution to the knowledge of the Holo- thurioid-larve which I can give here, is accordingly rather small, especially as compared with that of the Echinoid- and Ophiuroid-larve. Stichopus californicus (Stimpson). Pl. XXXII, Figs. 8—9. This species, which occurs fairly commonly along the rocky shores near the Biological Station at Nanaimo, was found to have ripe sexual products in June. Fertilization was undertaken repeatedly and with good success, though the percentage of the fertilized eggs was always rather small. I have no notices about the first developmental processes, except that the embryos were found in the blastula stage one day, in the gastrula stage two days after fertilization; the formation of the Auricularia may begin on the second day. At the age of about 6 days the larve were typical Auriculariz, provided with a starshaped calcareous body in the left postero- lateral process. In this stage the larve remained till the age of ca. 3 weeks, no further development taking place. The shape of the larva (Pl. XXXIII, Figs. 8—9) Denne? is somewhat elongated. The preoral band is highly Bee e ° ( arched; there are no preoral processes, but the frontal D ‘ area is distinctly constricted at the level of the upper iv end of the oral cavity, the anterior part forming a are enue’ ee rounded lobe; the corresponding part on the dorsal jornicus. ®);. The smal) Side is much narrower. The dorsal and posterolateral spicule to the rightfrom lobes are fairly distinct. The postoral band is strongly the right posterolateral . : a ieee convex in the middle so as almost to suit the shape of the preoral band. The anal area is markedly con- stricted at the lower end of the body, the posterior edge of which is distinctly concave. The calcareous body (Fig. 101) has the shape of an irregular star; very rarely there is a small, irregular calcareous body also 197 in the right posterior corner. The larva is colourless, excepting for a slight yellowish tint along the vibratile band. The larval nervous system was not distinctly observed. Stichopus Kefersteinii Selenka. Of the rather numerous Holothurians occurring along the rocky shores of Taboga, in the Bay of Panama, several species were found to contain ripe sexual products in October—November 1915; many attempts were made to obtain their larve through artificial fertilization, but the results were very poor. With the present species a fertilization was undertaken on October 23rd, which resulted in a small number of embryos. At the age of two days these latter were beginning to assume the Auricularia-shape; on the 3rd day all the embryos had died. A number of specimens were examined in order to have started a new culture, but all were empty by this time. The information acquired about the development of this species accordingly does not go beyond the fact that it has a typical pelagic larva. Holothuria n. sp. (?)') Pl. VI, Figs. 5—6. Like the preceding species this one was found to contain ripe sexual products in October; fertilization was undertaken on the 21st of this month 1915. The eggs were found to ripen after lying about 3 hours in the water; they then readily accepted the spermatozoa and quite a large percentage was fertilized. The cleavage is perfectly regular; the gastrula-stage was reached at the age of one day, and at the age of two days the embryos showed the beginning formation of the Auricularia. Three days old they were typical young Auricularias of the characteristic shape shown in PI. VI, Fig. 5, with the anterior part very broad, almost bell-shaped. At the age of 8 days the larve had reached the stage shown in PI. VI, Fig. 6. On the dorsal side the vibratile band now almost joins in the midline in the anterior end; otherwise the larval shape is still quite simple, without any indication of processes. The frontal area may perhaps be somewhat 1) This appears to be an undescribed species, allied to H. kapiolaniw Bell. It may be shortly thus characterized: Length 5—6 cm. Dorsal side brown, with two rows of small, black spots; ventral side pale; tentacles yellow. Deposits in the skin: slightly curved, very spinous rods; well developed end-plates in the pedicels. Calcareous ring high, without a deep notch posteriorly. Genital tuft long; only one Polian vesicle. The madreporite is single, fastened on the mesentery, with lancetshaped head; the stone canal is s-shaped. Cuvierian organs present. Locality: Taboga, Gulf of Panama. Littoral. I am indebted to Miss E. Deichmann for the above remarks on this species. In the list of the Echinoderms the development of which has been studied in the present work, given in the Introduction (p. 17—18), this species was forgotten. 198 loo narrow in the figure, as the specimen has apparently contracted some- what on preservation (it is slightly distorted, which has been corrected in the figure). The posterior end is rather elongated, with a somewhat in- distinct, round body at one side, and some accumulated cells which prob- ably indicate the presence in the fully formed larva of some kind of cal- careous bodies. It has a faint greenish colour. — Beyond this stage I did not succeed in rearing the larve, and no more specimens containing ripe sexual products were obtained, the breeding season of the species being evidently passed by the beginning of November. Auricularia nudibranchiata Chun. Since Chun observed this magnificent larva at Orotava in March 1887 it has never been recorded again from the Atlantic. I was therefore very pleased in finding a specimen during my stay at Tobago, B. W. L., in April 1916, proving thus its occurrence also in the West Indian seas. It has further been found in plankton samples from the following localities in the Sar- gasso Sea and in the vicinity of the Azores: 45°32’ N. 25°50’ W., 24/VI, 1911 (1 specimen); 40°47’ N. 21°10’ W. 21/III. 1911 (1 specimen); 39° 22’ N. 22°49’ W. 20/III. 1911 (2 specimens); 36°13’ N. 33°50’ W. 15/III. 1911 (1 specimen); 34°39’ N. 40°54’ W. 13/III. 1911 (several specimens); 33°55’ N. 43°40’ W. 12/III. 1911 (several specimens); 30°30’ N. 49°57’ W. 22/II. 1911 (2 specimens). This shows that it is fairly widely distributed over the ocean, the time of its occurrence extending at least from February to June. None of the specimens exceed the largest specimens found by Chun in length (6 mm), and none of them are in a further stage of development than the more advanced of Chun’s specimens. While the species had not been rediscovered in the Atlantic till now, it was recorded in 1911 by H. Ohshima?) as occurring in the seas of Japan (the Sagami Sea), and quite recently MacBride?) has recorded it from off the North end of New Zealand. It is, of course, beyond doubt that these larvee from the Japanese and New Zealand seas really belong to the type of A. nudibranchiata; but it is equally certain that they are not the same species as the Atlantic form. In accordance with the nomenclature here adopted I shall designate the three forms respectively as species a, b and c, the Atlantic form being, of course, designated as species a. ‘) Hiroshi Ohshima. Note on a gigantic form of Auricularia allied to A. nudibranchiata Chun. Annot. Zool. Japon. VII. 1911. p. 347. 2) E. W. MacBride. Echinoderma (Part II) and Enteropneusta. Larve of Echinoderma and Enteropneusta. British Antarctic (“Terra Nova’’) Expedition. 1910. Zoology. Vol. IV. p. 88. 199 Species b, the Japanese form, differs from the Atlantic first in its much larger size, reaching 15 mm, the latter not being known to surpass a length of 6 mm, the stage of development being the same in both. Further I find on a comparison of some specimens, kindly sent me by Dr. Ohshima, with those of species a, that the body is distinctly narrower in species b than in species a, and also the projecting lobes of the vibratile band are larger and more prominent in the Japanese species. Through this strong development of the vibratile band, in fact, the sides of the larva are wider than its dorsal and ventral sides, and the preserved specimens therefore generally lie on the side, while specimens of species a generally lie on the dorsal or ventral side. Whether the small difference in the structure of the calcareous deposits pointed out by Ohshima — viz. that the edge is finely serrated in the Japanese specimens, while Chun states it to be smooth in the Atlantic species — is a reliable difference needs to be verified by a reexamination of the latter; I am sorry to be unable to settle this question, all my specimens having been preserved in formaline, and the deposits thus having been dissolved. Species c, the New Zealand form, is likewise considerably larger than species a, the single specimen known measuring “at least a centimetre in length’. Otherwise it is impossible to state by which other characters it is distinguished from species a and b, no description or figures having been given of it. Only in one feature it is stated to differ from the two other species, viz. in the median ventral pouch from the intestine being paired, while in the two other species it is unpaired. It would appear from the expression used by MacBride (Op. cit. p. 89), running thus: “This single diverticulum is supposed to be a rudimentary representative of the gill- (lung-)trees of other Holothurioidea; but the discovery that the pouch is paired deprives this argument of its ground”’, that he thinks Chun’s state- ment of the pouch being unpaired in the Atlantic form to be wrong. I would then take the opportunity of confirming Chun’s statement. There is not the slightest doubt that the pouch is really unpaired, as figured by Chun, and the same is the case in the Japanese species. On account of this intestinal pouch Chun was inclined to refer this larva to some Elasipod, and I have joined him in this opinion. MacBride, however, thinks this a mistake. Having found in a species of Cucumaria the gill-trees to appear only in a rather advanced stage of development, he finds it “exceedingly unlikely that a rudiment of these gills should appear in the Auricularia larve; they probably only begin to develop when the cloacal function of sucking in and ejecting water has been established. I conclude, therefore, that the intestinal pouches of A. antarctica and A. 200 nudibranchiala have nothing to do with gills; and if this be admitted, there is no reason why A. nudibranchiata, like A. antarclica, should not be the larva of a Synaptid, to which group its wheels naturally ally it.” Without entering on a discussion of the time of appearance of the gill-rudiments in Cucumaria I would merely point out the fact that there is no structure in the Synaptids into which this pouch could possibly develop. This would then mean that this large organ of the Auricularia should completely dis- appear again on metamorphosis. This is certainly not very likely. The question about the parental origin of this larva can hardly be definitely settled except through the rearing of the larva; it seems only safe to say this much that the intestinal pouch proves that it must belong to some form with gill-trees, and accordingly cannot be a Synaptid, in spite of the Auricularian wheels. GENERAL PART I. Classification. FTER the description given above of the new larve reared or other- A wise gathered it now remains to see, whether this material, together with that known from previous researches, gives any answer to the original question, which was the guiding principle in the author’s studies, namely this: is there any correspondance between the different larval forms and the natural groups of the adult forms — or to put it more definitely: is there a larval classification corresponding to that of the adult forms, each group of the adults (genus, family or order) having a distinct type of larvee differing from that of the other groups? It is known already from Joh. Miiller’s magnificent researches that at least each class of Echinoderms has its own distinct larval type — the folothurians the Auricularia, the Asteroids the Bipinnaria, the Ophiu- roids the Ophiopluteus and the Echinoids the Echinopluteus!), the Crinoids alone having no separately named larval type, which is also unnecessary as long as we do not know a single case of a typical pelagic Crinoid-larva. But are “orders’’ and “families” to be distinguished within these four larval types? The question may at once be answered, at least partly, in the affirmative. A more detailed discussion is, however, necessary. Starting with the Echinoid-larvz it was clear already from the pre- vious researches that the Spatangoid-larve form a very distinct type, char- acterized by the unpaired posterior process, which does not normally occur in any of the other Echinoid-larve, so far as hitherto known?), They also differ from all other Echinoid-larve in having well developed anterodorsal arms. We now know the larve of quite a number of Spatangoids, viz. of ') Joh. Miller designated both the Ophiurid- and the Echinoid-laryie simply as Plu- teus; the separation of these two larval types under the names of Echinopluteus and Ophio- pluteus. was introduced in the present author’s work “Die Echinodermenlaven d. Plankton- Expedition.” *) Prouho in his “Recherches sur le Dorocidaris papillata’’ (Arch. de Zool. expér. gener. V. 1888) describes and figures (p. 141, Pl. XXV, fig. 9) an abnormal larva of this species with an unpaired posterior process. 26 202 the genera Spatangus, Echinocardium, Moira, Brissus, Brissopsis, Meoma, and besides a number of other Spatangoid-larve (chinopluteus fusus, soli- dus and others) are known to exist. They all of them agree in having the unpaired posterior process, and there is not the slightest reason to expect that there will be any exception to this rule. We have then here a larval type corresponding to the order of the Spatangoidea, forming a distinct larval “order.” Is it possible to distinguish larval “families” within this order, corres- ponding to the families of the adult Spatangoids? We are not yet in pos- session of sufficient facts for answering this question definitely; but the facts which are known would seem to point in the direction that also “families” may be distinguished within this larval order. The Spatangoid-larve may be divided into two main groups, viz. those with posterolateral arms and those without. Of the latter group only one species has been referred to its parental form, viz. that of Brissopsis lyrifera; but there is some reason to suggest that also the larvee of the genus Brissus belong to this type (comp. above, sub Brissus Agassizi, p. 119). A larva of the same type, which was found in the Gulf of Panama I would be inclined to refer to Meoma grandis. That would tend to indicate that this larval type is characteristic of the family of the Brisside. But it can be nothing more than a suggestion at the present state of our knowledge; moreover the fact that Metschnikoff?) has found a larva of this type occurring in great numbers at Triest, where Brissus unicolor, otherwise common in the Mediterranean, apparently does not occur, would seem to prove that this larval type may also be found in other Spatangoids. Metschnikoff refers this larva to Schizaster canaliferus “da sie dort unter den Echinoidenlarven ebenso vorherrschend ist, wie Schizaster unter den erwachsenen Spatangoiden.”’ The larvee of the genera Echinocardium, Spatangus and, probably, Moira?) agree in having posterolateral arms, supported by a simple rod, not widened at the base. This would point in the direction of this larval type being characteristic of one group of Spatangoids. The existence of several species of the type of Echinopluleus fusus might indicate that this 1) E. Metschnikoff. “Studien tiber die Entwickelung d. Echinodermen u. Nemertinen.” Mém. Acad. St. Pétersbourg. VII. Sér. XIV. 1869. p. 46. Taf. VITI—IX. *) Caswell Grave has reared the larva of Moira atropos through metamorphosis, but unfortunately does not give any information of the shape of the larva. On my applying to him for the information wanted, he kindly sent me some slides containing some of the larve. It turned out, however, that there were two different species of lary in the slides, one with posterolateral arms, the other without. Having called his attention to this fact I was informed by him that he thought it almost certain that the form with the postero- lateral arms was the Moira-larva, but full certainty could not be obtained any more from the material still preserved. 203 larva is characteristic of another group. But all this can, of course, be nothing more than mere suggestions—working hypotheses in the light of which future investigations should be carried out. It must be remembered that the classification of the Spatangoids is very far from being settled. In fact, H. Lyman Clark in his work on the Spatangide (“Hawaiian and other Pacific Echini’’)") states that “a satisfactory classification of the Echini here included in the Spatangina is in the present state of our know- ledge simply impossible.’ I agree with Clark herein as far as the Amphi- slernala are concerned, while I think the classification of the Meridostler- nata in fairly good order. The amphisternous ‘families’? Hemiasteride and Spatangide, as comprised in Clark’s work, are certainly unnatural, which is, however, not to be wondered at, his classification being “purely one of convenience, worked out in the endeavour to make an artificial key to all the genera and species of living Spatangoids.”’ But no better classi- fication exists, and it seems at least that the family Paleopneustide, as comprised by Clark, is fairly natural, and thus far his classification re- presents a progress. — I would expect that the study of the larve will prove of great importance for establishing the natural relationship between the numerous forms of Spatangoids. That the As/ernata, or at least the Cassidulids, do not really belong to the Spatangoids, but are more nearly related to the Clypeastroids seems rather probable alone from the study of the Echinobrissus-larva. The study of the development of forms like Echinolampas and Echinoneus will be of the greatest importance for settling the question of the natural relationship of these interesting forms. Another very distinct larval type is that of the Clypeastroids. The order of the Clypeastroidea being a very well limited one, and there being no doubt that the forms referred to that group are really naturally related, the study of the Clypeastroid-larvee will afford a critical test to the theory of the value of the larve as lending proof of the natural relations of the adult forms. It must be claimed, of course, that if the larve really are of such classificatory value, the larve of the Clypeastroids should form a uniform group, as do the adult forms. As we now know, through the previous researches and those recorded in the present work, larve belong- ing to 10 different genera, viz. Echinocyamus, Clypeasler, Echinarachnius, Dendraster, Encope, Mellita, Astriclypeus, Arachnoides, Laganum and Pero- nella, we can form a fair judgment of the character of the Clypeastroid- larve in general. The result fully bears out the expectations. These larve form a very uniform group, so uniform that it is even in some cases hard to distinguish the larve of different genera. The Clypeastroid-larve ') Mem. Mus. Comp. Zool. XLVI. 1917, p. 98. 26* 204 form a distinct larval “order’’, corresponding to the natural order of the adult Clypeastroids, and accordingly testify the correctness of the suggestion of the classificatory value of the larve. The Clypeastroid-larve are characterized by the body skeleton forming a basket structure, which very often develops into a large complicate, fenestrated plate in the posterior end of the body. Although the body skeleton is partly absorbed in the fully formed larva, there is no change in the body shape, so that we cannot speak of a first and a second larval stage. No posterior transverse rod is formed, and there are no vibratile epaulets, whereas there are small ventral and dorsal vibratile lobes. The postoral and posterodorsal rods are generally fenestrated. The question whether different groups of these larve, corresponding to the families of the Clypeastroids, are to be distinguished, cannot be an- swered at present. It is noticeable that the single Clypeaster-larva known as yet differs rather conspicuously from the other larve, excepting the Echinocyamus- and the Laganum-larva (— the modified larva of Peronella Lesueuri does not count in this connection —), in its skeleton being much simpler, apparently not forming a fenestrated plate in the posterior end of the body; this may possibly indicate a family character. But very much more information is needed before we can form a real judgment of the value of this difference. Another very noticeable fact is this that the larva of Arachnoides placenta, so unique among the Clypeastroids through the aboral position of its periproct, does not differ in any way markedly from the larvee of the other Scutellids. It is worth pointing out that the peculiar property of turning green on preservation in alcohol or on being damaged or dying, so characteristic of Clypeastroid tissue, is found also in the larve and the young, meta- morphosing sea-urchin. Within the Regular Echinoids we may first dismiss the order of the Cidaroidea, altogether too little being known of their larve for judging about their essential characters. If the unexpected should turn out that Echinopluteus transversus belongs to Cidarids, it will be very hard to recon- ciliate its highly specialized characters with the primitive character of Cidaroid structure, as also it would be most surprising to find so different larvee as Echinopluleus transversus and the larva of Cidaris cidaris (Doro- cidaris papillata) within the same family (the Cidarids apparently forming only a single family). — But there is no reason to trouble with this pro- blem so long as it has not been definitely settled that Echinopluteus lrans- versus really is a Cidaroid-larva. The few facts known about the development of Diadematids tend to show that their larve are characterized by their skeleton forming in their 205 first stage a typical basket structure, while in their full shape they have the characteristic shape of the larva represented in Joh. Miller’s VII. Abhandl. Taf. V, Figs. 1—3, and in the present work, Pl. V, Fig. 7, with very long postoral and posterodorsal arms, a posterior transverse rod, very large vibratile lobes but no epaulets. There is then reason to expect that the Diadematid-larve will prove to represent a distinct, well char- acterized larval “family.” The Echinothurids, unfortunately, remain entirely unknown as regards their development. While at least some of the deep-sea forms may be ex- pected, on account of the large size of their eggs, to have a direct develop- ment, the small eggs of Asthenosoma varium indicate that this species will have pelagic larve. It will be exceedingly interesting to see whether its larva will prove to resemble the type of the Diadematid-larve, as it might perhaps be expected on account of the relation which undoubtedly exists between the Echinothurids and the Diadematids, in spite of the remark- able difference in their test structure. Entirely unknown as regards their development are also the Salenide. Judging from their small genital pores they may be expected to have small eggs and, therefore, probably pelagic larve, in spite of their being mainly deep-sea forms. It will be most interesting to learn whether their larve resemble those of their nearest relatives among recent Echinoids, the Arbaciide. Within the family Arbaciide we now know the larve of three species of the genus Arbacia, viz. A. lixula, punctulata and stellata, these three being so closely alike that they are to be distinguished only by very un- important characters. One more larva of the same type is known (PI. VI, Fig. 1), which may perhaps belong to another genus; in any case there can be no doubt that it is also an Arbaciid-larva. The facts thus far made known decidedly indicate that there is a very distinct type of larva belonging to the family of the Arbaciide, characterized in its first stage by the skeleton forming a basket-structure, in its second stage by the existence of a posterior transverse rod which ends in a pair of long posterolateral arms of the same shape as the other arms. The postoral and posterodorsal rods are of the fenestrated type, but the holes are generally small or entirely lacking in a greater or lesser part of the rod. There are large ventral and dorsal vibratile lobes, but no epaulets. It will be of very great interest to learn whether the larve of the com- paratively few other forms belonging to this very well limited and un- questionably quite natural family are in conformity with the Arbacta- larve, as they should be, according to the views of the present author. The polyporous Telrapygus niger is easily enough accessible for study, and 206 also Coelopleurus may be got at without too much trouble, while the rare, small deep-sea forms Podocidaris, Habrocidaris etc. will, at any rate, not be within easy reach. But the study of the two first named forms will suffice for meeting reasonable claims to the knowledge of this larval “family.” It is, however, especially the larve of the Camarodonta which we must consider in discussing the classificatory value of the larvee. This great group comprises the bulk of the recent regular Echini, and here the views of the various authors regarding what would be a natural arrangement are most diverging. May not perhaps the larve give us the clue to their real interrelations? We now know, more or less completely, the larve of so many of these forms that we are at least able to form a fair judgment, whether the larve will give us any support for the solution of the problem. Two main views regarding the classification of the Camarodonta are prevailing among recent authors. One is that of H. Lyman Clark and R. T. Jackson, the other that of Déderlein and myself. While according to the two former authors the test structure alone affords valid characters by which to distinguish genera and families, the minor microscopical struct- ures, the pedicellarize and spicules affording characters of no higher value than for distinguishing species, I have expressed the view that also the microscopical structures, especially the globiferous pedicellarie, afford a valid basis for the higher systematic divisions, the genera and even the families, and that these characters should be taken into consideration as well as the structural characters of the test. The latter characters are, of course, regarded as the more important, but it is maintained that where the structural characters of the test do not give the clue to the natural interrelationship of the forms, as is mostly the case in the great and uni- form group of the Camarodonta, there the microscopical characters set in. The classification worked out on the base of these principles, which was set forth in the author’s work on the Echinoidea of the Danish “Ingolf’’- Expedition, Part I—II. (1903—7) was fully accepted by Doderlein, who is even inclined to ascribe still more importance to the microscopical characters than I can agree to. — In the said work I have especially pointed out that the character of the ambulacra, the oligoporous or poly- porous condition, is of no value as a character of larger groups, as it has, evidently, developed separately in different groups. Also the elongation of the test in Echinometra and some allied genera is regarded as unvalid as a family character. H. L. Clark on the contrary regards these two char- acters as of primary value for the classification, although he cannot deny that the polyporous condition has developed in at least two separate groups (his families Strongylocentrotide and Echinometride). 207 The two classifications are evidently irreconcilable. It is clear that we know perfectly well all the characters which are at all available as a base for the classification. The internal anatomy is so uniform within the whole group that there is no hope of finding therein any characters of value for the classification. Only the structural characters of the test and the organs attached to it, viz. the spines, tubefeet, pedicellariz, and the spicules found in the skin of these organs are available. (The spheeridize do not afford characters of classificatory value). The divergences in the two main views on the true relationship of these forms then depend on subjective appre- hension, and further discussion will hardly alter the views on any side’). Only new arguments may be expected to settle the problem and give us a clue as to which of the views regarding the interrelation between the forms within this uniform group comes nearest the truth. Such an argu- ment I have expected to find in the study of their larval forms. The present researches have added so considerably to our knowledge of these larval forms that it is possible to form a reasonable judgment, whether that ex- pectation is justified or not. One of the families within the Camarodonta is unquestionably a quite natural group; viz. the Temnopleuride, or rather the subfamily Temno- pleurine, the other subfamily, the Temnechinine or Trigonocidarine being less distinctly circumscribed. If the larve are of real value to classifica- tion it must be a just claim that the larve of the Temnopleurine should form a corresponding natural group. While hitherto not a single Temno- pleurid-larva was known, we now know, more or less completely, the larve of Temnopleurus toreumaticus, Temnotrema sculpta and Mespilia globulus, and it is indisputable that they do really agree in several important char- acters, so that it has the appearance that there is really a distinct larval type peculiar to the Temnopleurids (or at least the Temno- pleurine). Some larval forms taken pelagically and most probably, belong- 1) A classification differing again from both the above mentioned has been set forth by J. Lambert & P. Thiéry in their “Essai de Nomenclature raisonnée des Echinides” (I—IV. 1909—14). These authors have made a point of working out the natural classification of the Echinoids, taking both the recent and the fossil forms equally into consideration. This is, of course, in itself an excellent principle; in fact, nobody could deny this to be the only natural way — if only the fossil forms were sufficiently well preserved for this use. But this, unfortunately, they are not. When the other authors have confined themselves mainly to the recent forms, it is because they recognize the fact that the fossil forms are imperfectly preserved, so that their characters can only be incompletely ascertained, and accordingly the relation between the recent and fossil forms must remain more or less doubtful. Lam- bert & Thiéry, however, have adopted the view that the classification of the whole group, recent and fossil forms, has got to be based alone on the characters which can be found in the fossil forms, viz. on the test structure alone. Regardless of all facts known they con- struct their classification strictly on this principle. As, moreover, the nomenclature is worked out on the principle of the priority rule in its strictest, most litteral sense, it is no wonder that the result is so absurd that it can hardly be taken seriously. 208 ing to the same group serve to emphasize the distinctness of this larval type. This larval “family’’ appears to be characterized in the first stage through the skeleton forming no basket-structure; the fairly elongate body rod divides at the end into two rather long horizontal branches. In the second stage a posterior transverse rod is formed, from which a pair of short, branched posterolateral rods may issue; the latter may, however, be lacking. There are four vibratile epaulets, but no vibratile lobes. The rest of the Camarodonta was arranged by the present author in the families Echinide, Toxopneustide and Echinometride, while Clark will not acknowledge the family Toxopneustide, the forms referred to that family being redistributed in the Echinide and in a new family Strongylo- centrotide. The family of the Echinometride is regarded by Clark as containing only the polyporous, oblong forms, the other, more primitive forms referred to this family by the present author being also redistributed by Clark in the Echinide and Strongylocentrotide. According lo my view Clark’s families Echinide and Strongylocentrotide are quite hetero- geneous, while his family Echinometride is quite natural, only, in my opinion, not wide enough. We shall see now what the larve will teach us about this matter. It may first be stated that, as regards the various genera and the species referred to them, Clark and I in general agree, only with regard to the genus Strongylocentrotus we disagree in some important points, as is stated below. To the family Echinidee Clark refers the genera Psammechinus, Lyle- chinus, Echinus, Parechinus, Nudechinus, Evechinus, Toxopneustes, Tri- pneusles and Gymnechinus; according to my view only Psammechinus, Echinus and Parechinus belong to this family, Evechinus being an Echino- metrid, while the rest of them belong to the Toxopneustide. No larve belonging to the genera Parechinus (in the sense of Clark, which I adopt), Nudechinus or Gymnechinus are known, while of the other genera we know quite a fair number of larvee, so that we can see whether they favour Clark’s views or those of the present author. The two species known of the genus Psammechinus, miliaris and micro- fuberculalus, have both been studied as regards their development, miliaris being very completely known, while there is no description of the fully formed larva of microluberculalus. The larva is characterized by having, in the first stage, long body rods, widening in the end; no basket structure. In the second stage the larva has four epaulets, situated at the base of the four main arms. There is no posterior transverse rod or posterolateral processes. 209 Of the species referred to the genus Echinus s. str. two have been stu- died as regards their development, viz. esculentus and aculus. The larvee of these two species are so closely alike that they are hardly distinguishable. In the first stage they have elongated, clubshaped body rods, but no basket structure; in the second stage they have four epaulets, at the base of the four main arms, and besides a pair of laterally placed epaulets which may form an almost closed ring round the posterior end of the body. There is no posterior transverse rod or posterolateral processes. The larva of the nearly related Sterechinus Neumayeri (which Clark regards only as a synonym of Echinus margarilaceus or Slerechinus Agassizt, as I think its correct name should be) is known in its second stage. It agrees with the two Echinus-larve in all essential features. The larve of these forms accordingly agree very well in their main characters as should be expected, since they are indisputably closely rela- ted. But now those of the other genera! Here the matter lies quite differ- ently. Within the genus Lylechinus we now know the larve of the species variegatus, anamesus and pictus in both the first and second stages, and of the species panamensis and verruculatus in the first stage. The young larve have a short, rounded body, the body rod being short, branching at the end; the recurrent rod is well developed but does not unite with the branch from the body rod so as to form the basket structure, except in the verruculatus-larva. In the second stage there is a transverse rod, and posterolateral lobes, while the presence of epaulets is not definitely settled. — The larva of Tripneustes esculentus agrees with the Lylechinus-larve, dif- fering only in the body skeleton of the first stage forming a typical basket structure. The Tripneustes gratilla-larva in its first stage has the same structure, while the second stage is unknown. Of Toxopneusles we know with certainty only the young stage of the larvee of T. pileolus and roseus: they agree completely with the Tripneusiles-larva in the skeletal structure. Finally the larva of Evechinus chloroticus has in its first stage a more com- plicate body skeleton than that of the Tripneustes-larva, though essentially of the basket-type, recalling that of the Echinometra-larva, and in the second stage (almost certainly) a posterior transverse rod and postero- lateral processes. It must be conceded, at least, that the facts from the larve do not lend support to Clark’s view regarding the relationship of these genera. Clark’s family Strongylocentrotide encloses the genera Echinostrephus, Pseudoboletia, Paracentrolus, Loxechinus, Cxnocentrolus, Pachycentrolus, Heliocidaris and Strongylocentrotus. Of these the genera Echinostlrephus, Cenocentrotus, Pachycentroltus and Heliocidaris according to me belong to the family Echinometridz, Pseudoboletia to the Toxopneustide, Paracen- 97 al 210 frolus and Loxechinus to the Echinide; the relation of Strongylocentrotus appeared uncertain, though I have thought it probably nearest related to the Toxopneustide. Unfortunately, the larve of most of these genera are sill unknown; but we know those of Paracentrotus lividus, Heliocidaris luberculala (— the reduced larva of H. erythrogramma does not count in this connection —) and some “Strongylocentrolus’’-species. The larva of Paracentrolus lividus agrees with the Echinus-type; the body rod is elongated, clubshaped, and no basket structure is found in the first stage; there is no posterior transverse rod or posterolateral processes in the second stage. Anterior but no posterior epaulets (as in Psammechinus miliaris). In the larva of Heliocidaris luberculala the body skeleton in the first stage forms a very complicate basket structure, as in the Echinometra- larva; in the second stage there is a posterior transverse rod and postero- lateral lobes; vibratile lobes are present, but no epaulets. Within the genus Strongylocentrolus we know the larve of the species drobachiensis and franciscanus fairly completely, that of pulcherrimus nearly so and that of purpuratus in its first stage. They all agree in the body skeleton forming no basket-structure, the body rod being clubshaped, elong- ate (less so in franciscanus); in the second stage there is no posterior trans- verse rod or posterolateral processes, but both anterior and posterior epau- lets as in the Echinus-larva (presence of epaulets in the pulcherrimus-larva uncertain). To the genus Slrongylocentrolus Clark also refers the mediterranean species granularis, otherwise unanimously regarded as the type of a separate genus, Sphaerechinus. According to my view this genus belongs to the family Toxopneustide, representing a polyporous development within that family. The larva of this species differs most conspicuously from those of the above named Slrongylocentrotus-species. The body skeleton forms a basket-structure in the first stage; in the second stage a posterior trans- verse rod is formed, and there are posterolateral processes, vibratile lobes and anterior but no posterior epaulets. That the postoral rods are fene- strated is another conspicuous difference from the S/rongylocentrotus-larve, though of minor morphological importance. Thus it is evident that the characters of the larva of Sph. granularis bear decided testimony against classifying that form with the “Strongylocentrotide’’, or even referring it to the genus S/rongylocentrotus itself, while on the other hand there is per- fect agreement between the larval characters of S. granularis and those of the other Toxopneustid-larve thus far known. To the genus Strongylocentrotus Clark also refers the Japanese species depressus, which, according to my view, represents the type of a separate genus, Pseudocentrolus, of the family Toxopneustide, The little bit of in- 211 formation that has been given here of its larva is sufficient to show that it differs very markedly from the. true Strongylocentrotus-larve in its body skeleton in the first stage forming a basket-structure; there is then very good reason to expect that also in its second larval stage it will prove to agree with the Toxopneustid larval type. At least the facts known are decidedly not in favour of Clark’s views. Regarding the genera referred by Clark to the family Echinometride there can be no doubt that they really form a natural group, Parasalenia alone being doubtful. It must then be claimed that the larvee of these forms should be in conformity with one another in their essential features. It is a pity that we do not know much about these larve, only one species, Echinometra lucunter, having been reared to its full larval shape and through metamorphosis. But it is known that the larvee of Echinometra oblonga and Colobocentrotus atratus agree with the Ech. lucunter-larva in the very char- acteristic feature of the recurrent rod being double, and it is highly prob- able that the same character applies to the larvee of Echinometra Mathexi and Helerocentrotus mamillatus. Thus the facts hitherto known regarding the larve of the Echinometride agree very well with the results derived from the study of the adult forms as to their natural affinities. Summarizing now the preceding discussion it must be stated that the study of the larve most decidedly lends support to the author’s views as to the classification of the Camarodonta, and thus also gives proof of the correctness of ascribing comparatively greal importance to the minor microscopical characters of pedicel- larie and spicules in the classification of this group. The classi- ficatory results reached on using these characters, combined with the characters of the test, have been splendidly confirmed through the study of the larve. Disregarding these characters, Clark was led e. g. to such an absurdity as to include Spherechinus granularis in the genus Slrongylo- centrolus, otherwise so naturally circumscribed through the peculiar char- acter of its globiferous pedicellarie. There is no reason to enter here on a discussion of the objections raised especially by Clark against using these microscopical characters in classification, as it may now be regarded as an established fact that they are really of eminent classificatory value. Thereby I do not mean to maintain that my classification was correct in all details. Thus e. g. the position assigned by me to the genus S/rongylo- centrotus, mainly on account of the structure of its globiferous pedicel- larie, as being probably related to the Toxopneustide, was evidently in- correct. The study of the larve shows that it is most likely to be an offshoot from the Echinide s. str. Upon the whole I would emphasize that the true position especially of the forms belonging to the Camarodonta, (excepting 7* 212 the Temnopleurine), cannot be taken as definitely established, until all the characters available, including the structure of the larve, have been studied and duly taken into consideration. The struc- ture of the pedicellariz, of course, does not in itself give definite proof of their affinities; also these characters may, doubtless, have developed separ- ately along different lines, as is seen from the case of Strongylocentrotus, not to mention Glyplocidaris crenularis'). Therefore | would not consider e. g. the position of the genus Lowechinus in the family Echinide as definitely established, until it has been proved that its larva also belongs to the Echinoid type. Still less can the position of forms like Nolechinus magel- lanicus and Pseudechinus albocinclus, where the characters of the pedicel- lari are very indistinct, be regarded as settled, as long as we do not know anything about their larve. Upon the whole, the natural classification can- not be found through arranging the forms diagrammatically after this or that character; the interrelations and affinities being the result of organic evolution and descent it is not likely that we will find the true expression thereof by means of a diagram. As a definite result it may be stated that there is a definite larval type characteristic of each of the three families, the Echinide, the Toxopneustide and the Echinometride. They may be shortly characterized thus: , The larve of the Echinide have in their first stage an elongated body, supported by a long, more or less club-shaped body rod, the skeleton forming no basket-structure. In the second stage there is no posterior transverse rod, no posterolateral or vibratile lobes; epaulets are developed at the base of the four main arms, often also one pair at the posterior end of the body. Fenestrated rods are not known to occur within this type. The larve of the Toxopneustide have in their first stage a short body, the body rod being short and, together with the recurrent rod, forming a basket-structure (excepting the genus Lylechinus). In the second stage there is a posterior transverse rod; posterolateral and vibratile lobes are found, and there are often (always?) epaulets at the base of the four main arms. The rods of the main arms simple or fenestrated. The larve of the Echinometride have in the first stage a short, obliquely truncated body, supported by a complicate basket-structure, the recurrent rod being double. In the second stage there is a posterior trans- verse rod; posterolateral and vibratile lobes are found, but no epaulets. The rods of the main arms are fenestrated. ') Comp. the author’s “Echinological Notes.’ Il. A new principle of Classification. Vid. Medd. Naturh. Foren. Kobenhayn. 1910. p. 51 note. 213 Turning to the Ophiurid-larvz we find matters much less clear than is the case with the Echinoid-larve. In fact, our previous knowledge would hardly seem to warrant any other conclusion than this, that these larve form a rather chaotic assemblage; thus e. g. the larve of forms so closely related as Ophiura affinis, albida and texturata are so different, that they would rather seem to belong to different families than to closely related species. The present researches, however, tend to prove that matters do not stand quite as badly. It may be regarded as an established fact that within the genus Ophio- thrix the larve are of a very uniform character, so uniform, indeed, that it is in many cases impossible to distinguish the various larval species. Whether this type is peculiar to the whole family of the Ophiothrichide, and not to the genus Ophiothrix alone, remains to be proved. But in any case we have here a distinct larval type common to the very numerous species within this genus, and there is not the slightest reason to believe that the larve of any of these species should prove to differ markedly from the common type. This is then in perfect accordance with the legitimate claim, that nearly related forms should agree in regard to their larval characters. Another conspicuous case is aflorded by the Ophiocoma-larve. The four (or five) larve hitherto known are so closely alike that it may be safely stated that there is a distinct larval type of this genus, characterized through the transformation of the central part of the body skeleton into a sort of link, and through the existence of vibratile lobes. Whether this larval type is peculiar also to the family Ophiocomide remains uncertain. The study of the great variety of Ophiuroid larve of unknown origin has disclosed the important fact that also among these larve several distinctly characterized types may be distinguished, each comprising seve- ral species. After the analogy of the genera Ophiothrix and Ophiocoma it may very reasonably be concluded that these types represent various genera, the larve of the different species within these genera being upon the whole closely alike, as should be claimed upon the theory of the classi- ficatory value of the larve. Such larval types are especially Ophiopluteus undulatus, O. arcifer, O. pusillus, O. serratus-bimaculatus. It is then hardly to be denied that definite proof has been given that, also within the Ophiuroids, the larve of closely related species are in general similar, agreeing in their main characters. The conclusion to be derived from this fact must needs be this that the con- spicuous difference between the larve of Ophiura albida, affinis and lexturala proves these forms not to be so very closely related, in spite of the facl 214 that, even after the recent subdivision of the great genus Ophiura (Matsu- moto, H. L. Clark), these two species are left in the same subdivision, the genus Ophiura s. str., which would mean that a further subdivision of this genus is necessary. In this connection it is important to notice that most probably Ophiura affinis and the mediterranean O. Grubet, which has unjustly been made a synonym of the former, have larve of almost iden- tical structure!), Further the Ophiopluleus fusus, described in the present work, recalls the Ophiura albida-larva to a considerable degree, so that it might well seem that here again we have the larve of two really related forms. (— O. fusus being found at the Azores, one might perhaps think of Ophiura Thouleli Koehler as its parental form —). Regarding Ophiura leviurala, the larva of which is quite unique among the Ophiurid-larvee hitherto known through its fenestrated posterolateral rods, it is worth pointing out that this species is very peculiar by its series of pores along the ventral midline of the arms, therein differing markedly from the other species of the genus. Upon the whole, I would think it very probable that the genus Ophiura (or Ophioglypha) in its wider sense is a parallel to the Camarodonta within the Echinoids, that is to say, representing the most specialized type, in which the “coarser’’ characters, if I may use such an expression, have reached such a point of specialization and uniformity that they do not afford a sufficient base for further classification. The generic characters would then have to be looked for among the minor, morpho- logically less important structures, just as in the Camarodonta the char- acters of the pedicellariz and spicules set in, where the morphologically much more important characters of the test structure are failing as a suf- ficient base for classification. This is, of course, only meant as an idea to be tested by further studies. But in any case, it would seem perfectly absurd, in view of the facts here made known regarding the Ophiurid larve, to draw the opposite conclusion that, since the larvee of forms so closely alike as Ophiura albida, affinis and texturata differ so markedly from one another, the Ophiurid larve have no classificatory value at all. While it thus seems an established fact that within the Ophiuroid-larve several very distinct generic types may be discerned, it is, at the present state of our knowledge, impossible to say whether “family’’-types also may be distinguished. There is some probability that the family Ophiothrichide has a distinct larval type of its own, perhaps also the Ophiocomide have a distinct larval type — but very much more knowledge is required, before we can form a safe judgment of this problem. ') Comp. the author's paper “Notes on the development and the larval forms of some » Scandinavian Echinoderms”, p. 135. Within the Asteroid-larvz the facts previously made known, together with those given in the present work, decidedly tend to show that there are distinct “family” types of the larve, corresponding to the families of the adult. Unfortunately, our knowledge is still very limited, so that there is only a poor foundation for the conclusions; but the little we know speaks rather clearly. Within the Astropectinide we know the larve of four different species, all agreeing in their main features; the processes are very short, rigid, and the larve do not develop into a Brachiolaria. By the metamorphosis the whole larval body is absorbed. — Within the Luidiidee we know the larvee of the two species Luidia Sarsi and ciliaris, which both agree so closely in their characters that it is difficult to distinguish them. They are char- acterized by the great elongation of the anterior part of the body which apparently forms an active swimming organ; all the paired arms, which are long and movable, are situated close together in the posterior part of the body. There is no Brachiolaria-stage, and the larval body is not ab- sorbed by the metamorphosis but apparently simply cast off, when meta- morphosis is completed. — The fact that “Bipinnaria aslerigera” is re- corded from the Celebes-Sea (in the “Summary of Results” of the “Chal- lenger’’, p. 836) indicates that at least one more species of Luidiz has a similar larva, which lends support to the conclusion that we have here a distinct larval type, characteristic of the genus Luidia. Whether we may conclude this much that these two larval types represent the larve of the families of the Astropectinidee and Luidiide, is, of course, doubtful. But, at least, there is nothing to disprove it. Within the other families the few facts known of the development of the Asterinide and the Gymnasteride hardly form a sufficient base for any conclusion as to the characters of their larvee. I would only say that the character of the rudimentary larve of Asterina gibbosa and exigua, com- bined with the fact that the larva of Aslerina pectinifera develops into a Brachiolaria, tend to show that within this genus (s. lat.) the larvee typic- ally have a Brachiolaria-stage. Within the genus Aslerias the development of the species A. rubens and vulgaris has been studied very completely, and of a third, A. glacialis, enough is known to state that its larva agrees very closely with those of the other two species. It appears that we have here a very distinct larval type, characterized through its long, movable processes, and developing into a Brachiolaria with round arms, crowned by a circle of papille. On metamorphosis the larval body is completely absorbed. The characters of the young larve of some other forms, Orthaslerias leptolena, Euaslerias Troscheli, seem to indicate that these forms also have a Brachiolaria-slage; 216 further the species with shortened development, Astlerias Miilleri, groen- landica, hexaclis, through the fact that their rudimentary larve have dist- inet Brachiolarian processes, lend support to the suggestion that this larval type is characteristic of the whole family Asteriide. So little is known as yet of the development of Holothurians that hardly anything can be concluded as to how far the larvee may be arranged in groups, corresponding to the orders or families of the adults. A single larva is all that is known of the Synaptids; three larve of the genus Holo- (huria are known, but none of them in their full shape, and it is uncertain whether they adopt a more complicate structure in their more advanced stages. That they are very much alike in the younger stages, all being of a very simple shape, is in good accordance with what should be expected, but much stress cannot be laid on this fact. Of more weight is the fact that three different species are known of the type Auricularia nudibranchiata, resembling one another so closely that it is difficult enough to distinguish them. This proves that we have here a case of the larvee of closely related species agreeing very closely in their characters. (Although we do not know to which genus the A. nudibran- chiata belongs, there can hardly be any doubt that all these larval species belong within the same genus, or, at least, — in case MacBride’s state- ment that the intestinal pouch of species c is a paired organ be correct very closely related genera). It would not seem unreasonable to suggest that Auricularia antarctica, plicala and minor belong to the same group as A. nudibranchiala, and then we would here really have a “family” of Holothurian-larve, fairly well characterized through several features (e.g. the peculiar shape of the oral region, the peculiar folding of the vibratile band). But as long as we do not know anything with certainty about the parentage of these larve, we may only hint at this as a possibility. It appears to be a rule that the Dendrochirotes have a shortened devel- opment, without a typical pelagic larval stage. Without entering here on a discussion of the question, whether the simple larva of the Dendrochi-. rotes is a secondary adaptation or whether it represents the most primi- tive condition of Echinoderm-larve, it is clear that it has no bearing on the problem of the classificatory value of the typical pelagic larve. The Crinoids are the poorest of all Echinoderms in regard to the pre- sent subject, as we do not know one single truly pelagical Crinoid-larva, corresponding to the other four main types of Echinoderm larve. Till now only Comatulids have been studied as regards their development, and it seems very likely that within this whole group the development is of the 217 type without a typical pelagic larva, the Comatulids thus offering a case analogous to that of the Dendrochirote Holothurians. It may perhaps be expected that typical pelagic larve will be found to exist in some of the stalked Crinoids; but till now nothing is known about that, and accord- ingly the Crinoids do not contribute to the solution of the problem here discussed and must be left out of consideration for the present. After this summary review we may state as the result of these researches that the facts hitherto brought to light are decidedly in favour of the view that the larve of the Echinoderms have animportant bearing on the classification of the adult forms. It seems a suf- ficiently established fact that the larve of closely related species are upon the whole very much like one another, often hardly distinguishable, and, further, that the larve of allied genera agree in important characters, so that we get groups of larve, larval “families’’, “orders’’, corresponding to the groups (families, orders) of the adult forms. Accordingly the very im- portant conclusion seems justified, that when forms, which were sup- posed to be nearly related, prove to have essentially different larve, these forms are not in reality nearly related’). The study of the larvee will thus form a most important criterion for our classifica- tion, giving, so to say, the final judgment of its true value, and we are justified in stating that no classification which isin contradiction to the evidence given by the larve can give an adequate expres- sion of the natural relationship of the adults. — These results probably apply also to other animal groups than the Echinoderms; but this side of the problem will not be discussed on the present occasion. Although the facts already known seem to warrant the above conclu- sions, | would maintain most emphatically that the present researches are only to be regarded as a reconnaissance in this field. Proof is given that the way entered upon is the right way, and we may be sure that further researches will yield important results. But there is a long way to go. What has been obtained as yet is nothing more than random samples, taken here and there. It must be claimed that all the various forms should be studied, and not only the genera; it will be equally important tohaveall thespecies within the various genera made the object of study, even though they are, apparently or really, ever so closely related, and it is equally desirable that the larvee should be studied and figured alive. Thus we will gradually acquire a sum of knowledge of the greatest importance for reaching the end of all systematic work, the com- prehension of the multitude of living forms in their natural relationship. ‘') Provided, of course, their embryological conditions be the same, That closely related species may differ in one having typical larvie, the other having abbreviated development or being viviparous, is quite a different thing, and does not interfere with the above conclusion, 92 218 Il. Morphology; Phylogeny; Biology. The general homology of the four main types of the Echinoderm larve was so clearly pointed out by Joh. Miller in his classical memoir “Uber den allgemeinen Plan in der Entwicklung der Echinodermen’’!) that very little has had to be corrected?) or added, the main thing being that the separate frontal band of the Asteroid-larvee does not represent an essential difference from the other larve, as Joh. Miiller thought it to do. In the light of the more extensive knowledge of the larval forms now acquired it may be worth while to pay some attention to the special development within each of the four types, to see how they develop from the primitive generalized shape into more or less highly specialized forms or, on the other hand, are reduced into forms so simple that they are hardly recognizable as forms derived from the true pelagic larval type. The Auricularia is the type in which the least amount of specialization appears to have taken place, due allowance being made for the fact that we know comparatively less about the Holothurioid larve than of the three other main groups. The simplest form is that found within the genus Holothuria (— unless it should turn out to be much more specialized in the more advanced stages of development, which does, however, not seem very likely —). The vibratile band hardly shows an indication of the typic- al larval processes, the larva in fact looking almost like a diagram of the primitive Echinoderm-larva, as regards its outer shape (— as to its inner structure it is less simple —). Only simple calcareous bodies are known to occur, irregular stars or spherical bodies*), no wheels. — The larve of Synapla and Stichopus are only a little more specialized, the vibratile band forming small processes corresponding to those of the other Echinoderm- larve. In the Synapta-larva calcareous bodies are formed in the shape of elegant wheels. The Auricularia paradoxa being in all probability only a decalcified Echinopluteus transversus, only one group of Holothurian larve is known, which represents a more specialized form of the Auricularian type, viz. the larvee Auricularia minor, plicata and nudibranchiata. While A. minor is still comparatively easily referred to the primitive type, A. plicata is considerably more specialized, but the climax is reached by A. nudibran- chiala in which the vibratile band develops to such an extraordinary degree of complication that it is difficult enough to recognize the original type. Auricularia antarctica, though somewhat specialized in other directions, ') Abhandl. d. Akad. Berlin. 1853. Taf. II. *) Comp. the author's Memoir “Die Echinodermenlarven d. Plankton-Expedition” p. 8. 8) The Auricularia sphaerigera (Joh. Miillers’s “Auricularia mit Kugeln’’) may well be suggested to belong to the genus Holothuria. 219 seems fairly evidently to belong to this same larval group. As might be expected in such a specialized type, the calcareous bodies are wheels of a very elaborate structure. While in A. nudibranchiala the Auricularian type has reached its highest perfection (— it is very interesting that also in regard to size this larva represents the climax —), the Dendrochirote larva represents the opposite extreme, being simply worm-shaped, with none of the characters of the Auricularia, its vibratile rings recalling those of the Auricularian pupa, not the band of the Auricularia itself. For deciding the question, whether this simple larva is a true primitive form (as maintained by Caswell Grave')) or merely a reduced form, it would be of importance to know whether transitional forms exist, as is the case in the Ophiurans. Of this, however, we are ignorant as yet; but considering the very imperfect stage of our knowledge of the development of Holothurians, it would not seem unreasonable to expect that such an interesting case might ultimately turn up. The Bipinnaria in its simplest type differs so slightly from the simpler forms of the Auricularia that it may sometimes be difficult enough to decide to which of them such a form belongs. This primitive type is, as far as hitherto known, peculiar to a genus of starfishes which is unanini- ously regarded as one of the more primitive of recent Asteroids, viz. Astro- pecten, It is especially an important fact that no Brachiolaria-stage occurs in this type of Bipinnaria. The same fact applies to that more specialized type, the Luidia-larva, in which the anterior part of the body has under- gone a development into an apparently active swimming organ. Also the remarkable Bipinnaria antarclica recently described by MacBride?) evidently agrees with the Astropecten- and Luidia-larve in having no Brachiolaria-stage, as MacBride justly infers from the fact that even in a fairly advanced stage of metamorphosis the ventral median process retains its typical shape, no trace of Brachiolarian arms having yet appeared. If the reference of this larva to Cheirasler gerlachei is correct, we have here another case in support of the assumption that in the more primitive Asterids upon the whole the larvee do not develop into Brachio- larie. This said larva represents the most specialized form of all known Asteroid-larva in regard to the vibratile band, which forms dense folds or crenulations along the strongly developed arms, a case analogous to Auricularia nudibranchiata. ') Caswell Grave. On the occurrence among Echinoderms of Larve with cilia arranged in transverse rings, with a suggestion as lo their significance. Biol. Bull. V. 1903. p. 169. *) E. W. MacBride. Echinoderma (Part II) and Enteropneusta. British Antarctic (“Terra Nova’) Expedition, 1910. Zoology. IV. 1920. p. 90. 28* 220 While it would thus appear to be a rule that the larve of the Phanero- zonia have no Brachiolaria-stage, the facts known of the development of the Spinulosa and the Forcipulata (Cryptozonia) seem to indicate that their larvee are characteristic through having a Brachiolaria-stage!). The dif- ferences between the various types of Brachiolarias are very unessential, consisting mainly in the median Brachiolarian process being now round, with a crown of papilla, now flat with a series of papillae along the edges; this need not be considered here. The point to be emphasized is that the Brachiolaria-stage is known to occur only in the more spec- ialized Starfishes, and the Brachiolaria-arms and the suck- ing disk connected therewith are therefore later acquired, specialized structures. Accordingly the homology generally supposed to exist between the sucking disk of the Brachio- laria and the Pelmatozoan stalk is only apparent, and the greal part it has played in phylogenetic speculations is un- justified, not being supported by facts acquired from an extended study of the development of Asteroids. Of course, I agree that very much more knowledge is needed for finally establishing this statement as a fact beyond dispute. The metamorphosis of the Astropecten- and Luidia-larve need being studied in a much more detailed way than has hitherto been done. It would also be of the greatest interest to study the development of such Astropectinids and other Phanerozonia as have large, yolky eggs and to see whether the larvee of such forms possibly develop Brachiolarian pro- cesses and a sucking disk, as do e. g. the Solasler-larve. If my views on the development of Asteroids are correct, they should not develop such processes, or, at least, no sucking disk. (Processes might not necessarily be homologous with the Brachiolarian arms). Anyhow, the facts hitherto known of Asteroid development seem to me to enforce the above conclu- sions. It is of importance to notice the fact that during the metamorphosis of ') If the larva figured on p. 149 of my paper “Notes on the development and the larval forms of some Scandinavian Echinoderms” really belongs to Stichasler roseus as supposed by me there, that will be an exception to the rule indicated by the direct observations on the development of the said groups of starfishes. I would not be inclined to think this prob- able and therefore now doubt the correctness of that suggestion. The larva figured would then belong to Astropecten irregularis. Gemmill in his paper on “The larva of the Starfish of Porania pulvillus (O. F. M.) (Qu. J. Micr. Sc. Vol. 61. 1915) concludes from the fact that he has found this larva to have a Brachio- laria-stage that “it is evident that the division of Asterids into Phanerozonia and Cryptozonia is not necessarily associated with fundamental differences of development.” As, however, the position of the family Gymnasteride (or Asteropid), to which Porania belongs, within the Phanerozonia is doubtful, this conclusion may not be justified. The fact of the Porania- larva being a Brachiolaria at most may serve to prove that this larval stage makes its appear- ance in the most specialized group of the Phanerozonia. 221 Asteroids no breaking up and subsequent rearrangement of the ciliated band into transverse rings has been observed, such as it occurs especially in Holothurians, but also, more or less distinctly, in Ophiuroids and Echi- noids; accordingly we cannot speak of a pupa-stage in Asteroids corresponding to that of Holothurians and (more or less modified) ‘of Ophiuroids and Echinoids. Neither do the forms with direct development (Henricia, Solaster etc.) develop ciliated rings, as do the Crinoids and Holothurians with abbreviated or direct development; there is only a general ciliation in these reduced starfish-larve. It is, of course, possible that larve with ciliated rings may occur among those of the Phanerozonia which have direct development; but as yet we do not know a single case of such larve within the Asteroids — and this fact is not in favour of the suggestion that the simple, wormshaped larva provided with ciliated rings, like the Comatulid- and the Dendrochirote-larve, represents the primitive Echinoderm-larva. Two more forms of starfish-larve have been described, which would appear to represent special types differing essentially from the main types treated above. They must be briefly mentioned here. In my Memoir on “Die Echinodermenlarven d. deutschen Sudpolar- Expedition’’!) p. 89 notice is given of a larva with more than 20 Brachio- larian processes, the general shape of the larva being otherwise like that of a typical Asferias-larva; the larva itself I had not examined, the descrip- tion being founded on observations of the living specimen made by Van- héffen during the stay of that Expedition in the Antarctic Sea. As later on the specimen of this larva was examined by me, those numerous Bra- chiolarian arms were found to be a mistake; they were nothing but the spines of the young metamorphosing starfish. —- A correction of the de- scription was given in the “Vorwort”’ to the said volume of the “Deutsche Stidpolar-Expedition”’, p. VI; but as this correction is very likely to be overlooked I have found it my duty to mention it here. In 1906 Koehler & Vaney?) published the description of a remark- able new Asteroid-larva, taken in the vicinity of the Azores, which was named Stellosphera mirabilis. Only two stages were observed, the younger being provided with clusters of large spines, which were absent in the later stage, so that a remarkable, regressive development of the calcareous bodies in the larval skin takes place during the progressing development ') Deutsche Siidpolar-Expedition 1901—1903. Bd. XIV. Zoologie VI. 1913. *) R. Koehler et Cl. Vaney. Description d’une nouvelle larve d’Astérie appartenant trés vraisemblablement a une forme abyssale. Bull. Mus. Océanogr. Monaco. No. 64. 1906. The description was reprinted in Koehler’s great work “Echinodermes provenant des campagnes du yacht Princesse Alice (Astéries, Ophiures, Echinides et Crinoides), Reés. d. Camp. Scient. Monaco. Fasc. XXXIV. 1909. p. 131—136. Pl. XXIV. 222 of the larva. The crossed pedicellarie, present in both larval stages, de- finitely prove that it belongs to the suborder of the Forcipulata, and rea- sons are given for the suggestion that it is the larva of a deep-sea form. It is a noticeable fact that no traces of a vibratile band are found, Through the kindness of Professor Jules Richard I have received from the Musée Océanographique of Monaco some specimens of this very interesting larva. The examination of these specimens has led me to an interpretation differing rather considerably from that given by Koehler & Vaney. Both the stages described are far advanced metamorphosis stages, so that the true shape of the larva is not to be made out, the larval body most probably (— as may be inferred from a comparison with the later metamorphosis-stages of Aslerias —-) already being for the greater part absorbed. But this much is still clear, that the larva belongs to the Brachiolaria-type. In fact, there can be no doubt that what Koehler & Vaney take to be the mouth of the larva is really the sucking disk of the Brachiolaria, round which are found the three typical Brachiolarian arms. They are of quite typical shape, ca. 2—5 mm long, cylindrical, the point being covered with small papille; furthermore there is a close series of similar small papille from the base of the median, anterior Brachiolarian arm backwards to each side of the sucking disk. On the larval body there are besides found the last vestiges of the other larval arms, mostly reduced to mere papilliform processes, which can no longer be identified with certainly, excepting only the unpaired, median dorsal process, situated behind the median Brachiolarian arm. The larva is in a stage correspond- ing very closely with that represented in Pl. 21, Fig. 13 of Gemmill’s Memoir on the development of Aslerias rubens'). If only the body and arms were a little more contracted — as they would certainly be on a not very careful preservation — we would here have a figure very well representing the Slellosphera. If we now look upon the figures on PI. XXIV of Koehler’s work, it is evident that the figure 10 represents a larva seen from the apical pole, showing the sucking disk in the middle, with the three Brachiolarian arms (very poorly represented — they have evidently been unusually contracted in that specimen; the papille on the point of these “pédicelles péribuccaux’’ were also observed by Koehler & Vaney). The smaller papilla scattered over the body are the remnants of the larval arms, not at all the beginning tubefeet of the young starfish, as would seem to be the opinion of the authors. The six large calcareous plates with the clusters of spines are the terminal plates of the starfish. Another curious result of this examination is that the I. stage of Koehler & Vaney is really the older, more advanced stage, their II. stage being 1) Phil. Trans. Ser. B. Vol. 205. 1914. the younger. Accordingly the remarkable absorption of the plates and spines, supposed by the authors to take place during the progress of development, does not take place at all, the facts being quite the inverse. The peculiar stage represented in fig. 1 of Koehler’s work, showing what appears like two filaments at the “mouth” is only a later stage of development, where the Brachiolarian arms have been almost absorbed, the other arms completely so. In 1915 Gemmill*) described a new Asteroid-larva, Brachiolaria hiber- nica, characterized by the point of the Brachiolarian arms being wholly covered with small papille, and through the presence of a row of small papillae on each side of the sucking disk, which latter is transversely elong- ated. The single specimen observed (taken in the Atlantic, 50 miles to the W. of Ireland, in a vertical haul from 2165 meters) was in an advanced stage of metamorphosis, the larval arms being reduced to short cylindrical prominences, but Gemmill thinks it very probable that in its younger stages it has the shape of a typical Brachiolaria larva. The characters of this larva exactly agree with those of Slellosphera mirabilis pointed out here, and the figures of the larva given by Gemmill also recall the S/el- losphera to such a degree that I cannot doubt that it is really the same larva. It is true that Gemmill does not say anything about its being six-rayed, as is S/ellophera. But his Fig. 1 shows, besides the five young arms, a sixth prominence, which is indicated in the explanation of the figure only as a “prominence of soft tissues in the aboral notch.” I can hardly doubt that this is really the beginning 6th arm. (The skeleton, which would have settled the question, was, unfortunately, dissolved). Thus, I would think, the ~Séellophera mirabilis’ has been definitely deprived of its remarkableness and been reduced to a, probably, quite ordinary Brachiolaria in an advanced stage of metamorphosis. The suggestion of Koehler & Vaney that it belongs to a deep-sea Asteroid was apparently for a great part due to its supposed fundamental difference from the typical pelagic larvee of littoral Asteroids. The larvee were taken in vertical hauls with open nets, from 3000 m. to the surface: there is thus —as Koehler & Vaney of course admit — no certainty about the depth in which these larve were taken. As also other larvee, belonging undoubted- ly to littoral Echinoderms, have been observed far away from the coasts (see below, the chapter on the geographical distribution of the larve), there seems thus far to be no special reason for regarding the ~Stellophera” as the larva of a deep-sea Asteroid. Still it is possible that it was really taken in greater depths, and if we look up the forms of Starfishes to which it ‘') James F. Gemmill. On a new brachiate Asteroid larva and on the advanced Bipin naria of Luidia ciliaris (Philippi) Gray. Proc. R. Phys. Soc. Edinburgh. NIX. p. 191. 224 might possibly belong, it seems indeed probable that it does belong to one of the forms occurring in greater depths. The facts that the young starfish is 6-radiate and that it belongs to the Forcipulata (as proved by its pedicellarie) considerably restrict the number of forms which may come into consideration. Among the starfishes known to occur in the littoral regions of the Azores there is not one having normally 6 arms; Asterias lenuispina Lamk. however sometimes has this number of arms, though there are generally seven or eight. In case the “Stellosphera”’ should belong to this species, it must evidently start with six arms, since the metamorphosing starfish apparently always has this number of arms. The larger number of arms in A. lenuispina then must be due to additional arms formed by regeneration after selfdivision. This is by no means improbable, and it is therefore thus far very well possible that we have in the “Stellosphera” the larva of Aslerias lenuispina. If this is not the case, the only other forms known to which it could belong are Aslerias Richardi Perr. (not recorded from the Azores, but not unlikely to occur there), Pedicellasler sexradiatlus Perrier and Freyella sexradiala Perrier, all of them deep-water forms. As the larva was found in considerable numbers it is very unlikely that it belongs to the rare Freyella sexradiatla or to Asterias Richardi, not known with certainty to occur in the Azorean waters. Thus Pedicellaster sexradialus alone remains. The fact that this species has small eggs (— as I have had the opportunity to ascertain —) makes it highly probable that it has a typical pelagic larva (in contradistinction to Pedi- cellasler typicus, which has large, yolky eggs). One might expect that the structure of the pedicellariz would decide the question to which of these species the larva belongs. This is, however, not so, the crossed pedicellarie of the two forms being hardly different al all. On the other hand the pedicellariz of the young starfish, as figured by Koehler & Vaney, differ from those of the two species mentioned in having only one series of teeth along the edge, while there are two series in both Asé. lenuispina and Pedicellaster sexradiatus. This structural differ- ence might rather bear testimony against the referring of the larva to any of the two said species. It is, however, very well possible that the single series of teeth in the young is an embryonal character. I therefore would not deem this structural difference of sufficient importance to counter- balance the fact that no other starfish, which might come into considera- tion as its parental form, is known from the locality where the larva occurs. (The specimens of the larva examined by me all had their calcareous structures dissolved, so that I have been unable to ascertain myself the correctness of the figures of the pedicellarie given by Koehler & Vaney). ~ The occurrence of the larva off Ireland, which may be regarded almost 225 as an established fact through the “Brachiolaria hibernica’ of Gemmill, affords another important aid for settling the question of the parentage of the larva. Aslerias tenuispina does not occur in the Atlantic to the North of 40° Lat. N. Accordingly its larva could hardly be found off Ireland. On the other hand Pedicellaster sevradialus is not known to occur farther North than the Biscayan Gulf; but it is by no means improbable that it does, however, occur in the deep-sea off Ireland. The result is then that it seems highly probable that the “S/ellosphera mirabilis” is the larva of Pedicellaster sexradiatus. The Ophiopluteus, in spite of the great number of “species” known, appears to show only unimportant deviations from the original simple type, which may perhaps be most nearly represented by the Ophiolhriv- larva. In the shape of the body there are only few features which may be regarded as a higher specialization; such feature is the presence of vibratile lobes at the base of the posterolateral arms in the Ophiocoma-larva. The presence of a ciliated tuft at the posterior end of the body, occurring in various forms, may also be a specialized feature, probably acquired separ- ately in various groups. But too little is known for forming a definite judg- ment of the value of this character. The ciliated ring in the posterior end of the body of Ophiopluteus coronatus (“Echinodermen-Larven d. Plankton- Exped.” Taf. VI, Fig. 6), as yet quite a unique feature in Ophiurid-larve, may perhaps represent a further specialization of the apical tuft. But so long as we do not know anything about the affinities of that larva (— its skeletal structure is still unknown —) this question must remain unsettled. The length of the arms varies very considerably. The posterolateral arms are generally much the longer and form the main floating apparatus of the larva; this is especially evident in the Ophiothrixv-larva and Ophiopl. opulentus, where these arms are entirely unaltered during metamorphosis, the young Ophiurid remaining attached to them and simply leaving them by the time it is giving up pelagic life. The arms of Ophiurid-larve never become actively movable swimming organs, no muscles connecting the two parts of the body skeleton being present, contrary to what is often the case in Echinoid larve. Generally the arms are very thin and narrow, more rarely broad and flat (as in Ophiopl. paradoxus and the Ophiocoma- larva). The widening of the arm points often found in preserved speci- mens is probably only due to preservation. It is a curious fact that the right anterolateral arm is often longer than the left one. A very conspicuous feature is the absence of posterodorsal arms in the larva of Amphiura filiformis (Ophiopluteus mancus). Also in Ophiopluleus dubius (Echinodermen-Larven d. Plankton-Exped. Taf. IV. Fig. 9 —10) the 99 226 posterodorsal arms would appear to be absent. It is very tempting to sug- gest that this larva may belong to Amphiura Chiajet. In that case the absence of this pair of arms would seem to be characteristic of the Am- phiura-larve in general. However, the fact that among the numerous Ophiurid-larve examined by me from so many parts of the world, there is no one lacking the posterodorsal arms, makes this suggestion very doubt- ful, as it is, indeed, hard to assume that in all that multitude of forms there should not be at least some larva of one or another of the numerous Amphiura-species occurring in nearly every locality. But, anyhow, the lacking of the posterodorsal arms, being quite a normal feature at least in the larva of Amphiura filiformis, is a very noteworthy, probably regres- sive, specialization. In contradistinction to this stands the formation of small extra arms in Ophioplutleus opulentus. It seems beyond doubt that these arms are due to the branches from the postoral and anterolateral rods pushing out the vibratile band so as to form a beginning arm; but as this is apparently the case also with the other arms, the growing skeletal rod being the primary factor in the formation of the arms, it seems per- fectly justified to regard these small extra arms as having the value of true arms, and one might very well fancy that other larval forms may exist in which these arms reach a size similar to that of the other arms. These additional arms of Ophiopluleus opulentus form an interesting analogy with the formation of anterodorsal arms in the Spatangoid-larve. It is, however, the skeletal structure that shows the more interesting diversities within this larval type. Two main types are to be distinguished, one having simple body rods, the other a compound body skeleton, a ventral and a dorsal recurrent rod forming together with the body rod two meshes in each side of the body. It would seem fairly certain that the simple type is the more primitive. The only larva with a compound body skeleton, the origin of which has been definitely ascertained, is that of Ophiactis balli*); but it is clear that all the many different forms haying a compound body skeleton cannot possibly belong to the family Ophiac- tidee. It is evident, therefore, that this type of skeletal structure must have developed independently along various lines, and accordingly the character of the body skeleton, whether simple or compound, cannot be of primary importance from a classificatory point of view. If I was right in referring the Ophiopluteus gracilis, described in my Memoir on the Echinoderm larve of the German South Polar-Expedition (p. 89), to Ophiura gelida, it becomes evident that the compound type of body-skeleton is at most a genus-character, not a family-character. The fact that the larva of ‘ Ophiactis balli has a compound body skeleton, while the larva of Ophio- ') Th. Mortensen, On the development of some British Echinoderms; p. 11. 227 pholis aculeata, apparently a near relative of it, has a simple body skeleton tends to prove the same thing. The end rods and the transverse rods reach an extraordinary degree of specialization in the Ophiocoma-larve, in which they are so much reduced in size as to form together merely a sort of link between the two halves of the skeleton. Otherwise the end rods show only a small amount of diversity, less so than the transverse rods which often afford important specific characters. More interest, however, is attached to the median pro- cesses from the transverse rods, present in a number of forms: these rods may show a high degree of development, especially in the larvee of the serralus-arcifer-group; also Ophiopluteus monacanthus must evidently re- present a remarkable specialization of this skeletal part. These median processes are generally directed horizontally outwards from the transverse rods and thus considerably enlarge the body volume, which may serve to increase the floating power of the larva. In no case the vibratile band is produced along the body prominences formed by these rods. The remarkable preoral rod of Ophiopluleus coslalus, as yet quite a uni- que feature in Ophiurid-larve, may possibly have originated as a further development from the median process; but nothing definitely can be stated about this at our present state of knowledge. There is a certain analogy between this preoral rod and the dorsal arch of Echinoid larvee — but, of course, it is only a quite superficial analogy, by no means a homology, as, upon the whole, there is no direct homology between the skeleton of Ophiurid- and Echinoid-larve, the skeleton having certainly been acquired separately in each of these two larval types. The rods of the inner pairs of arms (postoral, anterolateral and postero- dorsal) show but little diversity. The various degree of development of their thorns is a feature of interest here; it is not improbable that the strong development of thorns along the aboral side in the outer part of the antero- lateral arms in forms like Ophiopluleus pusillus may prove of some import- ance. A more conspicuous feature is, however, the development of special rods supporting the anal area in Ophiopluleus fulcilus; but this is also a feature as yel unique among Ophiurid-larve, about the morphological meaning of which il is not safe as yet to state anything definite. It recalls the ventral transverse rods of Echinoid-larve, with which there is, how- ever, no homology, of course. The posterolateral rods show the greatest diversity of all skeletal parts of the Ophiurid-larvee, and afford especially important diagnostic char- acters. The more usual type is a simple rod provided with strong, curved thorns along the inner (adoral) side; more rarely thorns are found also along the outer (aboral) side. A most remarkable specialization is the un- 99 228 dulating rod of Ophiopl. undulatus. In some forms the posterolateral rod is canaliculate to a greater or lesser extent (O. coslalus, opulenius), in others il is composed of two closely apposed, parallel rods (O. a/finis), and finally it reaches its highest specialization in the larva of Ophiura texlurata, in which it is fenestrated — another feature unique among Ophiurid-larve. This character also recalls the Echinoid-larve in which fenestrated rods are of general occurrence; but, as was pointed out already by Metschni- koff't) (comp. also “Echinodermen-Larven d. Plankton-Exp.” p. 53) the structure of the fenestrated rods of the O. levturata-larva is markedly dif- ferent from that found in the Echinoid-larve, in accordance with the fact that this character has been acquired separately in the two types and is a highly specialized feature in the Ophiurid-larve, while in the Echinoid- larvee it is, evidently, a primitive character. Some Ophiurid-larve are more or less rudimentary, showing various degrees of reduction from the typical Pluteus-shape to a mere worm-shape which has no longer any likeness whatever to the Pluteus-form. The first stage of reduction may perhaps be represented by the larva of Amphiura filiformis, in which the posterodorsal arms have disappeared, the larva otherwise being typical in shape. A somewhat more advanced stage of reduction is probably represented by the larva of Ophiura affinis or that figured by Joh. Miller on Taf. VII, Fig.5 of his VI. Memoir on the development of Echinoderms. Unfortunately, only the metamorphosis- stage is known, however, so that we do not know how far the reduction has gone; in any case the posterolateral arms are completely developed. A much more reduced stage is represented by the larva figured by Metsch- nikoff (Op. cit. Taf. NIL. B. Fig. 4—5) (Ophiopluteus Metschnikoffi) and a still further reduced stage by the larva figured by Claparéde?) (Ophiopl. Claparédei). In these two larvee, which I have reproduced in figure 102, the Pluteus-shape is, especially in the latter, hardly recognizable any more; the rudimentary skeleton of O. Metschnikoffi still distinctly recalls the typical Ophiopluteus-skeleton, while in O. Claparédei it is reduced to a pair of simple rods. In the larva represented in fig. 102, C, which was dis- covered by Krohn (Ophiopluteus elongatus) the Pluteus-shape is no longer recognizable, but traces of the larval skeleton are stated by Krohn®*) to be found. Through these forms we reach to the larvee of Ophioderma brevispina, 1) E. Metschnikoff. Studien tiber die Entwickelung d. Echinodermen u. Nemertinen p. 29. *) E. Clapareéde. Beobachtungen tber Anatomie u. Entwickelungsgeschichte wirbel- loser Tiere. 1863. Taf. 1 Figs. 11—12. %) “Im Embryo liess sich soweilen ausser kleineren Kalkablagerungen .... deutlich ein gerader, starker, bis in die beiden Enden reichender Kalkstab unterscheiden.” A. Krohn. Uber einen neuen Entwicklungsmodus d. Ophiuren. Mill. Archiv 1857. p. 373. 229 Ophionereis squamulosa (and doubtless many other Ophiurids), in which also the larval skeleton has disappeared completely, no trace of the Plu- teus-form remaining. In these latter larve ciliated rings are formed, as in the barrel-shaped larve of Comatulids and Dendrochirotes, and in the pupa-stage of other Holothurians. The existence of a true pupa-stage in Ophiurids was first shown by Caswell Grave!). (Very likely the larva, in which the rearrangement of the vibratile band into an indication of rings was observed by Grave, was a species of the type described here as Ophiopluteus opulentus). My own observations in the main agree with those of Grave, however, differing Fig. 102. A. Ophiopluleus Melschnikoffi (after Metschnikott); B. Ophiopl. Claparédei (after Claparéde; modified so as to represent the larva in ventral view); C. Ophiopl. elongatus (after Krohn). from them in some minor points. I find the rings less distinctly indicated than shown by Grave, which may, however, be due to my not having ob- served the metamorphosing larve in the same lateral position as that in which they are represented in Grave's fig.8, but from the ventral or dorsal side. The anterior “ring’”’ I have not found to be complete, but it may very well be so in a later stage of metamorphosis such as that figured by Grave, the stages in which I have noticed this rearrangement of the vibratile band being less advanced (Pl. XXIX, Fig. 3; Pl. XXX, Fig. 2). I have been able to ascertain that the anterior band of the “pupa’’-stage is derived directly from the preoral band of the larva. Another interesting fact is that in some cases, at least, the vibratile band of the posterolateral arms in an advanced stage of metamorphosis coalesces in the midline of the larval body, below the developing Ophiurid; how far this is a general rule remains to be ascertained. ') Caswell Grave. On the occurrence among Echinoderms of larve with cilia arranged in transverse rings, with a suggestion as to their significance. Biol. Bulletin. V. 1903. p. 179. 230 The Echinopluteus affords the greatest diversity of forms of all Echino- derm larvee, and several well marked types are to be distinguished. It does nol immediately appear which of these represents the more primitive Lype. Considering, however, the fact that the larve in which the body skeleton in the first stage forms a basket-structure, and which have in their second stage a posterior transverse rod and more or less developed posterolateral processes, are characteristic — so far as we know — of the Cidarids, Dia- dematids and Arbaciids, that is to say of the more primitive forms of Echinoids, it can hardly be disputed that we have got to regard this larval type as the more primitive form. Consequently the larval type char- acteristic of the family Echinide s. str., with the elongated, club- shaped body rods, with the recurrent rod rudimentary or absent, and with- oul a posterior transverse rod or posterolateral processes, is a highly specialized and exceptional larval type. It is therefore not at all justifiable to make this larval type represent the Echinoid larve in general, as is done in most text-books. Characteristic of the larval body of the primary type are the vibratile lobes; in the more specialized types, the larve of the Echinide s. str., and of the Spatangoids these lobes have disappeared, while the Clypeastroid- larve have retained them to some degree. A further specialization from the lobes are the epaulets occurring in the higher types of the Regularia. Both the vibratile lobes and the epaulets evidently serve to increase the floating power of the larva. This object is attained to a still higher degree in several larval forms of Regular Echini and Clypeastroids in which muscles connect the lower ends of the rods of the four main arms, so that these arms become actively movable. These larve, when floating, keep the four main arms in a more or less horizontal position, raising them when disturbed. This is not yet an active swimming movement, the muscular apparatus being too simple for performing regularly repeated movements. Only one Echinoid larva appears to be able to swim actively, viz. the remarkable Echinopluteus transversus, in which a complicate muscular system is developed, as described above (comp. fig. 54, p. 88), the body- skeleton being most extraordinarily adapted for serving as a support to the muscles. The four main arms, the posloral and posterodorsal, are the most divers- ified of the larval arms. They are always rather long, but sometimes (Dia- dematide) attain a very great length. This is carried to an extreme in Echinopluteus transversus, as regards the postoral arms, while the postero- dorsal arms have disappeared. In some forms these arms are broad and flat, especially so in the Mespilia-larva. The anterolateral and preoral arms are very uniform in character throughout the whole class. The postero- lateral arms, always the most prominent of the Ophiopluteus-arms, are developed as true arms only in Arbaciids and Spatangoids, in the rest of the larvee they have the shape of broad, earshaped lobes or are entirely lacking. As regards the skeleton it is a noteworthy fact that the fenestrated rods represent a primary structure, as must be concluded from the fact that this type of rods (always confined to the four main arms) is found in the larve of the more primitive forms, simple postoral and posterodorsal rods, upon the whole, occurring only quite exceptionally (Lylechinus, Echinobrissus) outside the family Echinide, the larve of which are also otherwise among the most specialized of all Echinoid-larvee. In some cases (e. g. Temnopleurid larva, species c, fig. 22, p. 61) these rods begin as fenestrated, but terminate as simple rods. In Evechinus chlorolicus they appear to have only two component rods. The preoral and anterolateral rods are always simple, more or less spinous, presenting no features of special interest. The dorsal arch also affords little diversity; the long lateral processes from it, supporting lobes of the dorsal side of the body (e. g. Temnopleurid-larva, species ¢; Echino- metra lucunter), are very probably homologous with the anterodorsal rods of Spatangoid-larve. — The primitive type of the body skeleton is, as stated above, that in which the body rod and recurrent rod unite so as to form a basket structure. This has been given up only in the Temno- pleuride and Echinide, where the body rod has been more or less specially developed, and, exceptionally, in the genus Lylechinus of the Toxopneu- slide. A special development of the basket structure is found in the Echino- metride, where the recurrent rod has become double, and in the Clypea- stroids, where the posterior part of the “basket” often develops into a large, fenestrated plate. The element of the Echinopluteus-skeleton displaying the greatest divers- ity is the posterior transverse rod, or rather the branches from its ends, the posterolateral rods. These may be simple rods, now very short, now extraordinarily long (Pl. VI, Fig. 1), or branching structures of exquisite form (e. g. Temnopleurid-larva, species a, fig. 20, p. 58; Heliocidaris luber- culata, fig. 24, p. 65). While it has completely disappeared in the Clypea- stroid-larve, it has been further specialized in the Spatangoid-larve, car- rying the unpaired posterior process, so characteristic of that larval type. The ventral transverse rods are generally of a very uniform character in the Regularia, whereas in the Clypeastroids and Spatangoids they are somewhat specialized, being reduplicated in the former, often assuming a broad, flat shape in the latter. This apparently unimportant skeletal ele- ment has been made the starting point of an extraordinary development in Eehinopluleus transversus, where it is transformed into a complicate 932 supporting apparatus for the muscular system. (Comp. e. g. fig. 34, p. 88). It is of considerable interest to trace this unusual development of a quite indifferent element into a structure of the greatest perfection. Reduced larval forms, so common in the other classes of the Echino- derms, are only rarely met with in Echinoids, in fact were hitherto quite unknown. I have had the good fortune of discovering two most interesting cases of reduction not counting the Echinopluleus transversus —, in which all the arms, excepting the postoral ones, have disappeared. One of them, the larva of Peronella Lesueuri, is still recognizable as an Echino- pluteus, the postoral (and sometimes also the posterodorsal) arms. still remaining, and a larval skeleton, rudimentary but still distinctly referable to the normal skeletal type, being developed. The vibratile band has dis- appeared, the larva being covered only by a uniform ciliation; only in some exceptional cases a rudimentary band is developed (p. 116; fig. 48). In the other case, Heliocidaris erythrogramma, the reduction has gone so far that there is not the slightest trace of the Pluteus-shape left. This larva recalls the barrel-shaped larvee of Comatulids and Dendrochirotes, but differs markedly from them in having only one, not very distinctly differentiated ciliated ring. The existence in Echinoid-larvee of a pupa-stage corresponding to that occurring in Holothurians and Ophiurans has been emphasized by Cas- well Grave, who has observed transverse ciliated rings on the newly metamorphosed sea-urchin of Mellita lestudinala. (Op. cit. Fig. 10, p. 178). | have not made any corresponding observations and shall therefore refrain from commenting on the possible existence in Echinoids in general of such a pupa-stage. Whether the Heliocidaris erythrogramma-larva is to be regarded as an indication in that direction I shall leave undecided. The few Crinoid-larve known being all of the vermiform type there is no reason to mention them in the present connection. With the greatest expectations we may look forwards to the future discovery of some typical pelagic Crinoid-larva; the study of its relations to the other four main types of Echinoderm-larve will be of extreme interest. The existence of such a pelagic Pelmatozoan-larva would seem almost beyond doubt, though possibly not in any of the recent Crinoids. But these latter repre- sent — in spite of the enormous development of the Comatulid-type — only a diminutive fraction of the whole group of the Pelmatozoa. It would hardly seem too bold to fancy that at least some of the fossil Pelmato- zoans, especially of the numerous Cystideans, had truly pelagic larve. It is very sad that we have no hope of learning anything about them, it being in the highest degree unlikely that such larve should have been preserved in a fossil state. Fritsch, it is true, has described an organism from the Lower Silurian of Bohemia, which he thinks must be a fossil Crinoid-larva (“ein Pluteus eines Crinoiden’’), the “Furca bohemica’’ of Barrande!). But provided first that it be an Echinoderm, anybody who has the slightest knowledge of Echinoderm-larve will see at a glance that it is quite absurd to speak of a “larva” ina case like this, where a single or double series of distinct marginal plates is represented along the whole border of the animal. But, moreover, it is certainly no Echinoderm at all. Professor O. Jaekel, with whom I have discussed this matter, informs me that he knows the “Furca bohemica’ quite well, and states that the re- presentation which Fritsch has given of it is quite erroneous. In fact, Professor Jaekel does not hesitate in declaring it to be identical with the organism from the Middle Cambrian (Stephen formation) of British Columbia which Walcott has described?) under the name of Marella splendens, and which is a Crustacean, probably allied to the Trilobites. As regards the anatomical structure of the Echinoderm-larvee (— leav- ing the reduced, barrel-shaped forms out of consideration here —) attention may be called to the fact that the ventral depression of the body, in which the mouth lies, the oral area, generally continues some way down below the upper edge of the anal area, thus forming a cavity, the lower limit of which is mostly very distinctly seen as a line passing across the anal area at about the level of the lower end of the esophagus. It is sometimes very large, as in Ophiopluleus serratus (bimaculatus), sometimes also of a more complicate shape, as in the larva of Colobocentrolus atratus. This “suboral cavity’, as I have termed it, was well figured and described by Joh. Miiller in his Memoir “Uber die Ophiurenlarven d. Adriatischen Meeres”’, p. 4, Taf. I, Fig. 1, and also Metschnikoff (Op. cit.) has care- fully studied this cavity, the “subumbrella’’ as he names it, in the same larva, where it plays an important part during the process of metamor- phosis. Also in a Spatangoid-larva he has noticed it (Taf. VIII, Fig. 12). The same structure was observed, but apparently not understood, by Mac- Bride in the larva of Echinocardium cordatum; he only mentions it as “two large cavities’, one on each side in “the web of skin connecting the post-oral arms.”’ — This cavity has some connection with the question about the origin of the “amnion” of Echinoid-larve, and accordingly some morphological interest is attached to it. In the epidermis of the sunk oral area is found a pair of nerve-streaks, ‘) A. Fritsch. Uber eine Echinodermenlarve aus dem Untersilur Bbhmens. Zool. Anz. 33. 1909. p. 797. *) Ch. D. Walcott. Cambrian Geology and Paleontology. II. No. 6. Middle Cambrian Branchiopoda, Malacostraca, Trilobita, and Merostomata. Smiths. Miscell. Coll. Vol. 57. 1912. p. 192. Pl. 25—26. 234 which appears to be a structure characteristic of Auricularize and Ophio- plutei. It was discovered by Metschnikoff and made the object of a closer study especially by Semon in his paper on “Die Entwickelung der Synapta digitata’!); furthermore Chun described the nervous system of Auricularia nudibranchiata, while it was observed by myself in Auricularia antarclica, all these forms agreeing very closely in regard to the shape and position of this structure. Quite recently MacBride, (Echinoderm larve of the British Antarctic (“Terra Nova’’) Expedition) gives a repre- sentation of the nervous system of Auricularia antarctica (Pl. I, Figs. 1—2) quite different from that given by me (Echinodermen-Larven d. deutschen Siidpolar Expedition p.81, Taf. X, Fig.1;Taf. XI, Fig.2). There is, however, hardly any doubt that what MacBride here represents as the larval ner- vous system is the edge of the atrium, which continues some way on to the oral area as a streak of thickened skin (comp. my memoir quoted above, p. 79, Taf. IX, Fig. 1), such as it was also found by Chun in Auri- cularia nudibranchiata. 1 do not want to deny the possibility that this structure may be of a nervous character, but that has not yet been proved, and, in any case, it is not the homologue of the nervous system of Auri- cularias in general; this latter structure is found more laterally in the oral area and is not in direct connection with the mouth?). In the Ophiurid-larve the nervous system is essentially like that of the Auricularia, forming a band across the oral area to each side of the mouth; for a more detailed representation of this structure in Ophiurid-larve I may refer to my paper, “Notes on the development of some Scandinavian Echinoderms” p. 158—160. — In Echinoid-larve the nervous system is ') Jen. Zeitschr. f. Naturwiss. XXII. *) I may take the opportunity here to object to a statement by MacBride concerning my representation of the coelomic structures of Auricularia antarctica. On describing the stage represented in his Fig. 2 MacBride says (Op. cil. p. 87) that “this stage corresponds with that represented in text-fig. 1 in Mortensen’s latest paper, but M. has quite misunderstood it. He overlooked the rudiment of the posterior coelom altogether, and has figured a sac entirely detached from the anterior coelom as the posterior enterocoele. This sac, which lies above the junction of the stomach and intestine, has nothing to do with the coelom, but is one of the intestinal pouches characteristic of Auricularia antarctica.”’ It is true that I have not observed the developmental stages of the posterior coelom (— on reexamining the material still at my disposal I do not find any such stages represented and I feel confident, therefore, in saying that I have not overlooked these stages, but they were not present in my material —). It is possible that the vesicle which I represented as the posterior coelom really is the intestinal pouch (— which I have, otherwise, not at all overlooked, stating, on p.80, that the rectum “schwillt mehr oder weniger plétzlich an und kann bisweilen den unteren Teil des Magens ganz Uberdecken’’ —). But, anyhow, I have at least not misunderstood the anterior coelom and the hydrocoel. It must certainly be acknowledged that the representation of the internal structures of Auric. anlarctica, given by me, means a very considerable progress beyond that given by MacBride in his first description of this larva (National Antarctic Expedition. Natural History. Vol. VI. Echinoderma. 1912), and I might perhaps suggest that the author of that description was not quite right in giving as his only comment on my representation of that larva that I had “quite misunderstood” its coelomic structures. 235 not of quite the same kind as in Auricularize and Echinoplutei. In the larva of Echinocyamus pusillus | have found a simicircular streak in the epidermis of the oral region, which is evidently a nervous structure (Op. cit. p. 157, fig. 5); the same structure appears to exist also in some other Clypeastroid-larve, and also in the larva of Echinometlra lucunter I have observed something evidently corresponding to it (Pl. XII, Fig. 1). But it does not appear as a rule to exist in all Echinoid-larve, but rather as an exception. On the other hand MacBride has shown an apical nervous system of more ganglionic structure to exist in the larvee of Echinus escu- lentus and Echinocardium cordatum (for references, see my paper quoted above). It will be an object of considerable interest to study the relation between these two nervous systems in Echinoid-larve, how far they are distributed within this larval type and whether they may occur together in one and the same larva or exclude one another. It is a very remarkable fact that the Asteroid-larve do not appear to possess anything corresponding to the nervous system of the three other larval types. (Also the Crinoid-larve have a fairly well developed nervous system, but that is, of course, of quite another type than that of the pelagic Echinoderm-larve). Only a sub-epidermal network of nerve fibres and neuro-muscular cells have been observed by Gemmill in the larve of Asterias rubens and Porania pulvillus. This fact tends to indicate that the Bipinnaria represents a more primitive type than the other larve. The intestinal organs are of a very uniform structure throughout the whole of the Echinoderm-larve; only a few rather unessential points may be mentioned here, viz. the existence of a small distinct intestine in the larva of Colobocentrotus atratus and of a kind of gizzard-like structure in the larva of Arachnoides placenta and, less distinctly, in some other Cly- peastroid-larve. The main interest as to the internal structures of Echi- noderm-larve is attached to the enterocoel and hydrocoel, and their transformations during the growth and metamorphosis of the larva. Of this, however, I have only made very few observations, it being out of the plan of the present work to pay special attention to these structures. That would have meant such an increase in the extent of the work that it was out of question merely for this reason, not to mention the time that would have been required for such a study. But, of course, it would be of the greatest interest to study these developmental processes in so many dil- ferent forms of the typical larve as possible, this being the only way to ascertain which features are of general validity and which of only more special value. — In the present work it is only the more exceptional forms, with abbreviated development, Peronella Lesueuri, Heliocidaris erylthro- gramma and Ophionereis squamulosa, which have been made the object of 30* 236 a somewhat more detailed study as regards the development and gradual transformation of the inner structures. The results acquired from the study of these objects, though they cannot claim to be of more general value, are in themselves of very great interest and throw important light on several problems connected with the developmental history of Echino- derms., The most interesting of the facts discovered by the study of the said forms with abbreviated development would appear to be that of the am- nion developing from the pharynx in Peronella Lesueuri, while other- wise it develops as an invagination from the ectoderm on the left side of the larva, above the hydrocoel. (In Heliocidaris erythrogramma it could not be decided whether it likewise develops on the left side, there being appar- ently no means of identifying the right and the left sides in this perfectly cylindrical larva). This may perhaps give some indication as to the way in which the amnion of Echinoid-larve originated. Mac Bride has come to the conclusion that the amniotic cavity of the Echinopluteus may be interpreted as a portion of the stomodeeum which is formed separately from the rest of it'), because in the metamorphosing larva of Ophiothrix fragilis he has found the primary tubefeet protruding into the stomodzum. The fact that the amnion of Peronella Lesueuri does originate as an oul- growth from the pharynx would seem to lend an emphatic support to MacBride’s theory. Still I would not take it as having been definitely proved as yet by these facts. Peronella Lesueuri is a very highly special- ized type: may we then really rely upon this remarkable development of the amnion in the much reduced larva as meaning a reversion to the pri- mitive mode of development? It is perhaps not quite safe to draw this conclusion as yet. If other similar cases were found, the conclusion would be very considerably strengthened; but this isolated, evidently very spe- cialized case does not seem to me to afford sufficient proof. Neither does the Ophiothrix-larva appear to me to be a sufficient proof of the theory. It is the only Ophiuran with a typical pelagic larva the metamorphosis: of which has been adequately studied as yet. But the little we know about the metamorphosis of other Ophioplutei does not point towards the stomo- deeum as having generally the function of an amnion in Ophiurid-larve. In Ophiopluleus bimaculatus, so carefully studied by Joh. Miiller?) and Metschnikoff*) — though not by means of sections, of course — it ap- pears that the suboral cavity has got the function of an amniotic ') MacBride. Textbook of Embryology. I. Invertebrata. 1914. p. 522. 2) Joh. Miiller. Uber die Ophiurenlarven d. Adriatischen Meeres (V. Abhandl. 1852), Taf. II—III. ‘) E. Metschnikoff. Studien tiber die Entw. d. Echinod. u. Nemertinen. Taf. VI—VII. 237 cavity. Further it is a fact that the growth of the hydrocoel in Ophiurid- larve takes place in different ways. In the larva of Amphiura filiformis (Ophiopluleus mancus) e. g. it grows upwards over the mouth, the closing of the hydrocoel ring taking place at the lower end of the pharynx; in the larva of Ophiura albida (Ophiopluleus paradoxus) the growth proceeds in the opposite direction, the closure of the hydrocoel ring taking place above the mouth. (Comp. Pl. III, Fig. 4; Pl. IV, Fig. 28 and Pl. VI, Fig. 40 of Chadwick’s Memoir on the Echinoderm-larve). The larva of Ophio- thrix fragilis appears to be about intermediate between these two forms, as regards the formation of the hydrocoel ring. These facts at any rate show this much, that we must be very cautious in drawing such important conclusions from facts derived from the study of only one single form. With this I do not mean to deny the possibility that MacBride’s theory of the origin of the amnion of Echinoplutei may be right. On the contrary, it appeals to me as a very reasonable suggestion. Also a comparison with the vestibulum of Crinoid-larvee would seem to lend support to this theory. The fact that the hydropore is not formed in the larva of Peronella Lesueuri nor apparently in Heliocidaris erythrogramma until at a rather advanced stage of metamorphosis, may be recalled here. I would, how- ever, not regard this fact as being of greater morphological importance, it being probably a modification caused by the exceptional conditions ob- taining in these larvee. In the same way I would not regard the unusual way in which the enterocoel develops in these forms as of essential im- portance from the point of view of comparative morphology. On discussing the question of the original type of the Echino- derm-larve Caswell Grave’) comes to the conclusion that the larve with transverse ciliated rings (Anledon, Cucumaria) represent the primitive condition from which the other larvee have been specialized “and carried far out of the path of phylogeny, as a result of their independent life. To this type of development the specialized larve tend to return at the time when their free-swimming life is given up.’’ He makes an attempt to show, how the ciliated rings were useful to the free-swimming animal not only as organs of locomotion, but also as organs of feeding (comp. his textfigure 11. b.). The attempt does not appeal to me as very successful, apart from the fact that it could hardly always have been, as he states, the two anterior rings that were lost when fixation on the bottom took place. It is true that larvee of this type occur in various groups of the Echinoderms, being even apparently the rule in Comatulids and Dendrochirotes. The fact 1) Caswell Grave. On the occurrence among Echinoderms of larvie with cilia arranged in transverse rings, with a suggestion as to their significance. Biol. Bull. V. 1903. p. 183. 238 that Heliocidaris erylhrogramma has such a larva, although with only one ciliated band, might be adduced here as proving the occurrence of this larval type also in Echinoids, where it was otherwise unknown. But it is a very strong objection that in the whole class of the Asteroidea not one case of larvee with ciliated rings is known, neither is anything correspond- ing to the pupa-stage known in this class, which is otherwise one of the more primitive of Echinoderms, at any rate more primitive than Ophiu- roids and Echinoids. Another important fact is this, that in all the cases where larve with ciliated rings occur, the eggs are large and yolk-laden. This evidently means that the larve with ciliated rings, developing from such eggs, are modified in accordance with the fact that they have food enough in store in the yolk and therefore need not trouble with catching food. The ciliated rings of these larve decidedly would serve very badly as food-gatherers. If that type of larve were really the primitive form we should have the remarkable fact here of an organism having arisen evidently unable to subsist by its own means. The fact that the stage with the ciliated rings is never indicated in the beginning of the larval development, but always at its end in those forms which have larve of the typical pelagic shape, is also very hard to understand on the assump- lion that the form with the ciliated rings is the more primitive. Also the fact that in the viviparous Ophionolus hexactis the embryos develop into larvee corresponding exactly to the generalized, primitive type of Echi- noderm-larvee, is of considerable importance in this connection; this larva decidedly needs no special adaptation to pelagic life, and it is hard to see why it should adopt this form instead of that with the ciliated rings, were it not of phylogenetic importance. In my opinion there is then no doubt that the larva with a simple circumoral band, as it is found in the younger stages of all the four main larval types of Echinoderms, is the primitive form, and that it represents a true phylogenetic stage in the ancestry of Echi- noderms — that is to say, the larva in its simplest type, the Dipleu- rula. Of course, the larve of recent Echinoderms, with their more or less highly specialized characters, do not represent ancestral types of the various classes of the Echinoderms, Bipinnaria of the Asteroids, Echino- pluteus of the Echinoids ete. They represent special adaptations of the original pelagic ancestral form, having been modified along with the adults, so as to form groups corresponding with the natural groups of the adults, the result being that there is a larval classification exactly parallel with that of the adults. But the primitive type, the Dipleurula must, I have no doubt, represent the organism from which the whole of the Echinoderm stem developed. 239 Neither time nor space permits me to enter on a detailed discussion of the phylogeny of Echinoderms. I must content myself with stating my perfect agreement with Bather’s views on this matter !). (— I would only call attention to the difficulty pointed out already by Bury 2) that no stage of fixation occurs in the embryogeny of Echinoderms, excepting the Crinoids; the fixation of Brachiolarie by means of the sucking disk is, as hinted at by Bury, evidently a secondary adaptation, which hardly counts in this connection —). Consequently I am decidedly opposed to the theories of the ancestry of Echinoderms more recently set forth by A. H. Clark %) and J. E. V. Boas *) — not to mention that of Simroth °). In my memoir on the development of Crinoids I have objected to some few points in the theory of Clark — who maintains the barnacles to be the ancestors of Echinoderms; but otherwise I do not think this theory more worthy of a refutation than that of Simroth, deriving the Echino- derms from Myzostoma! My few counter-remarks to some of the inter- pretations of Crinoid morphology, set forth — rather emphatically — by A. H. Clark on the base of his theory, are meant as a tribute called forth by my admiration for the eminent specialist in Crinoids, not as a wish to refute his theory, which seems to me a superfluous task. Both the said theories, besides bearing evidence of most unusual conceptions of mor- phology, are at variance with the fundamental principle of phylogeny, that evolution goes from the lower towards the higher organization, not the inverse way (due allowance being made, of course, for the rarer cases of regressive development, as e. g. the Acoela, which is, however, only an apparent exception to the rule). This objection does, of course, not apply to the theory of Boas, that the Echinoderms have developed from some fixed form of Coelenterates: but otherwise this theory is, in my opinion, no more acceptable than are those of A. H. Clark and Simroth. As stated above, I cannot enter on a detailed discussion of the reasons given for this theory, but must con- fine myself to making a few objections, which, however, would appear to suffice for proving the untenability of the theory. One of the main facts adduced by Boas as support for his theory is this that in the Crinoids — and, mind well, not the more primitive Pelmato- zoa, the Cystids, the simpler forms of which, at least, do not show any external signs of a radiate structure®) — the water-vascular system re- ') F.A.Bather. Echinoderma, in Ray Lankester’s Treatise on Zoology. Part III. 1900. *) H. Bury. The metamorphosis of Echinoderms. Quart. Journ. Mier. Se. 38. 1895. p. 93. %) A. H. Clark. A Monograph of Existing Crinoids. I. U.S. Nat. Mus. Bull. 82. 1915. ‘) J. E. V. Boas. Zur Auffassung der Verwandtschafts-Verhaltnisse der ‘Tiere. I. 1917. 5) H. Simroth. Uber den Ursprung der Echinodermen. Verh. d. Deutsch. Zool. Ges. 1904. *) The radiate structure of Echinoderms is regarded as requiring their derivation from “einer exquisit radiir gebauten Abteilung festsitzender Tiere’? (Op. cit. p. 21) 240 mains in open connection with the body cavity throughout life, which is taken as a proof that the Crinoids are the most primitive of all Echi- noderms'). It would appear that the author has entirely overlooked the fact, established beyond any doubt, especially through Seeliger’s researches 2), (— his results being fully confirmed by my own researches —) that the connection between the water-vascular system and the body cavity in Antedon is only a secondary character, arising at a rather late stage of de- velopment (in the Pentacrinoid-stage). Accordingly this is no primitive feature and cannot afford any proof of relationship with the Coelenterates. The Crinoid-larva is regarded as the most primitive of Echinoderm- larvee, “nahert sich viel mehr dem Coelenteraten-Larven-Typus, ist jeden- falls durchaus anderer Gestalt als die der anderen Echinodermen, und da diese Larve der .urspriinglichsten Echinodermen-Abteilung angehort, hat sie bei der Beurteilung der Verwandtschaftsbeziehungen ein grésseres In- teresse als die andere, aus deren Bau sich nicht auf die Abstammung der Echinodermen Schliisse ziehen lassen’. The typical pelagic, bilateral larval form “kann ich (Boas) nattrlich nur als eine ganz sekundare beurteilen. Dieselbe ist bekanntlich . . . noch nicht bei den Crinoiden ausgebildet, deren Larve mit fiinf Wimperringen umzingelt ist, von denen vier allen radidren Anforderungen gentigen.” (Op. cit. p. 23). — I may refer to the remarks given above in criticism of the view, also expressed by Grave, that the larva with the ciliated rings is the primitive larval form of Echi- noderms. Especially the structure of the Antedon-larva is anything but primitive, but only intelligible as a special adaptation of the simpler type of the pelagic larve to the conditions of the yolky eggs, as it is found likewise in gther barrel-shaped larve developing from eggs with a rich content of yolk. On the contrary, as stated above (p. 232), there is every reason to suppose that there was also a typical bilateral larval form at least in some of the numerous fossil Pelmatozoa, should it even be proved that none of the few surviving types have such a larva. The rather bold assertion that no conclusions as regards the ancestry of Echinoderms can be drawn from the structure of the typical larve does not do away with the fact that all these larve pass through the “Dipleurula’’-stage, that is to say, begin as a simple, bilateral, wormshaped organism with a segmented body cavity consisting of at least two, probably three segments, and pro- ') From the Crinoids the Asteroids are derived and from the latter again the Ophi- urids and Echinoids, the Holothurians having arisen from the Echinoids. Without entering on a discussion of this view of the genetic interrelations of the classes of recent Echi- noderms I would merely express my astonishment at a fancy bold enough to conceive the derivation of Asteroids from the Crinoids. *) O. Seeliger. Studien zur Entwicklungsgeschichte der Crinoiden. Zool. Jahrb. Abt. f. Anat. u. Ontog. VI. 1895. 241 vided with a simple circumoral ciliated band. This fact is hardly intelli- gible except on the assumption that it represents the original larval type and an ancestral stage in the evolution of the Echinoderm stem. If there is any connection between Coelenterates and Echinoderms, the vestiges of it must be sought for in the structure of the Dipleurula, not in the Crinoids. But I fail to see how this could be found. Possibly the Cteno- phora might have given rise to the Dipleurula —but then these latter are. in my opinion, no Coelenterates. It is an established fact that some animal forms have a different mode of development under different biological conditions. As typical instances may be named the shrimp Palemonetes varians, which has much larger eggs in Southern Europe than in Northern Europe, a corresponding rather con- siderable difference obtaining in the larval development, and Musca corvina which is recorded by Portchinski to be oviparous in Northern Russia, while in Southern Russia it is viviparous in summertime, oviparous in spring!). Such a remarkable diversity of development, for which Giard has created the name Poicilogony, was maintained by this author?) to occur also among Echinoderms, Ophiothrix fragilis being especially named as an instance of poicilogony. This species is stated to develop “suivant les condi- tions éthologiques ... tantot par des Pludeus normaux (comme dans la Mé- diterranée), tantot par des Plu/eus imparfaits tels que ceux étudiés par Apo- stolidés (a Roscoff), tantot méme par des embryons trés condensés, inca- pables de nager et qui donnent une Ophiure presque sans métamorphoses (a Vimereux etc.).”” (Op. cit. p. 240). This would appear to apply also to a number of other Ophiurids of the North Atlantic, since he states in a previous paper®), likewise reproduced in the “Oeuvres diverses” I. p. 509, that “presque toutes les Ophiures que j’ai observées dans la Manche sont .... Vivipares. Je citerai entre autres: ’Ophiothrix fragilis et l Ophiocoma neglecta que j'ai plus particuliérement étudiées au point de vue de la re- production”’.... “A uncertain moment de l’année, on trouve des embryons dans toutes les Ophiures que l'on ouvre indistinctement (excepté celles qui sont infestées. par les Orthonectida).” As I have shown in my paper “On Hermaphroditism in viviparous Ophiurids” (Acta Zool. I. 1920. p. 7—8) this statement of these Ophiu- rids being at times of the year viviparous, while at other times they have ') J. E. V. Boas. Kleinere carcinologische Mittheilungen. Zool. Jahrb. Abt. f. Syst. 1890. *) A. Giard. La Poecilogonie. Congrés Internat. de Zool. Bern. 1904. Bull. Scientif. Dép. du Nord de France. 39. 1905. — Reprinted in Oeuvres diverses. I. Biologie Générale. 1911. p. 420. 8) A. Giard. Particularités de Reproduction de certains Echinodermes en Rapport avec l'éthologie de ces animaux. Bull. Scientif. Dép. du Nord de France. IX. 1878. p. 296. 31 242 pelagic larvee, is quite without foundation, resting, no doubt, on misinter- pretations and being due to most imperfect methods of research. (The statement of the same author that these Ophiurids also are hermaphroditic is equally phantastic, as I have shown in the paper quoted above). — As for his statement that in Ophiolhrix the larvee now develop into the typical pluteus shape, now develop through a very imperfect larva, it rests on the researches of Apostolidés!), who expressly says (Op. cil. p. 76) that the nine tenths of the larve reared by him did not develop into the true larval shape but showed “une forme plus ou moins arrondie, qui continue a vivre, et dans laquelle nous avons. suivi le développement complet de lanimal’’. Although he says to have “des raisons de prétendre que, jusqu’a la fin du développement de Vanimal, il continuera a en étre ainsi” it is evident that he has simply failed to rear the larvee beyond the stage with the posterolateral arms, the larve dying when they had reached this stage — in fact, he states himself that “cette forme singuliére ne continuera plus a s’augmenter, mais peu a peu a se dégrader, jusqu’au moment de sa disparition.”’ While thus we need not trouble with this imperfect larva with only two arms, the direct statement of Apostolidés that he has followed the com- plete development of the still simpler embryos cannot simply be done away with. It does, however, seem very hard to believe it to be correct — and he does not give a single figure to accompany these remarkable observations. That author moreover being anything but a first rate author- ity it is impossible simply to endorse his statement. It is a fact that Giard’s statement of these Ophiurids as being at times of the year vivi- parous, rests on misunderstood observations; it is therefore hardly too bold to suggest that Apostolidés’ statement likewise may rest on misappre- hension. In any case, we cannot accept the statement, until it has been confirmed by renewed researches by a competent investigator. It would appear that Apostolidés (Op. cit. p. 77) has also observed those incomplete larval forms in the free. I would, however, suggest that it may have been the rudimentary larval forms, Ophiopluteus Metschniko/fi or O. Claparédei, which he has observed. But that these larve should also belong to Ophiothrix, like the typical Ophiopluleus, is exceedingly hard to believe. The whole question, however, very much needs reinvestigation. As a further instance of poicilogony in Echinoderms Giard mentions Asterina and Aslerias. “Asterina cephea habitant les mers chaudes, aban- donne au hasard de la vie pélagique des oeufs qui doivent donner naissance A des larves nageuses.... A. gibbosa n’est done auwune forme poecilogé- 1) N. Apostolidés. Anatomie et développement des Ophiures. Arch. Zool. expér. & génér. X. 1881. 243, nique de PA. cephea. De méme Aslerias Miilleri West sans doute qu'une variété poecilogonique septentrionale du vulgaire Aslerias glacialis des cotes de France.’ rectness of the observations. But the said forms are far from being only “poicilogenetic varieties” of one and the same species of Aslerina or Asle- rias; as anybody having some systematic knowledge of these great “genera” of starfishes will agree, they are very well separated species which, on modern classification, should at least be referred to different subgenera. In spite of the bad luck of the instances of poicilogony among Echino- derms produced by Giard I do not pretend to maintain that poicilogony does not exist within this animal group. In fact, we have an almost typical instance in Aslerias Miilleri—groenlandica. These two forms are so closely related that they are hardly distinguishable and appear to be merely var- ieties of the same species. But while A. Miilleri, as already observed by M. Sars, hatches its eggs in a brooding cavity made by the raised arms, A. groenlandica, according to the beautiful observations of I. Lieber- kind?) hatches its eggs within the stomach. It is unknown as yet whether there is any difference in the embryonal development of these two forms: but at any rate the breeding habits of these two forms are remarkably different. The cases of nearly related species showing quite a different mode of development, although not to be termed directly poicilogony, are of very considerable interest as related phenomena. Such cases were known hith- erto only in the genus As/erias, but have been proved now to exist also within the genus As/erina (s. lat.) and among the sea-urchins in the genus Heliocidaris, where H. luberculata has a typical pelagic larva, H. erythro- gramma a barrel-shaped larva showing no likeness whatever to a Pluteus. (It must be emphasized that there can be no doubt that the two said sea-urchins are really nearly related species, decidedly belonging to the same genus). In a good many cases of nearly related species one has pel- agic larve, while the other is viviparous (Echinocyamus pusillus and nu- lfrix, several Ophiurids). Upon the whole more extended researches will be sure to disclose many cases of dissimilar development in nearly related forms. In this connection mention must also be made of the interesting observation by Nachtsheim?) that in Echinaster seposilus the eggs may differ very conspicuously in size; the development, however, is the same ’ In these two cases there is, of course, no doubt as to the cor- in both larger and smaller eggs, and the larve developing from them differ only in size. 1) I. Lieberkind. On a starfish (Asterias groenlandica) which hatches its young in its stomach. Vid. Medd. D. Nat. Foren. 72. 1921. *) H. Nachtsheim. Uber die Entwicklung von Echinaster sepositus (Gray). Zool. Anz. XLIV. 1914. p. 601. 31* 244 The observations set forth above (p. 148) tending to prove the existence of a case of true metagenesis in an Ophiuroid-larva, Ophiopluteus opu- lenlus, may be recalled here. Although this is certainly quite an isolated case in Echinoderms, considerable interest is attached to it both froma morphological and a biological point of view. The observations on the development of the various forms studied in this work convey a good impression of the great diversity obtaining, as regards the length of the time required for the development. Although I have not taken care in all cases to notice the exact time at which each stage of development was reached, still a good deal of inform- ation may be gathered from what has been noticed. It may be practical to give the facts in a tabular form. Young larva Metamorphosis 0 Blastula Gastrula I. stage Il. stage beginning completed Diadema antillarum ......... Se aye 2 days —_— Astropyga pulvinala ......... 1 — we Be Tripneustes esculentus ....... See 1(2) 12 days 22 days Lylechinus variegalus ........ 6 hours ae 1 5 - 13 — (CEOUTIESUS Sr. aseletonecie 20 - (?less) 24 hours 2—3 7 - panamensis....... 1 Toxopneustes pileolus........ me 24 2 Strongylocentrotus pulcherrimus os 24 2 a — franciscanus er 24 2 = Temnopleurus loreumaticus . . . ste ‘ 22 hours so as Evechinus chloroticus ........ 16 hours 24 3 days 14 — ca35 — Heliocidaris tuberculala ...... 20— 24 2 11 — 35 — xe erythrogramma ... ae 18 3 Bc <5 4—5 days Echinometra lucunter ........ fe ar 2 4 — 19 — _- OGlLONGO ara chisier eye 16 2 Colobocentrotus atratus....... se 24 2 Clypeaster japonicus......... ss 14 3 11 — 1 Arachnoides placenta ........ oh 16 — ca LP/, 2} /s— 3} /.— Echinarachnius excentricus ... oe oe 2 . 8 ETICOPE TTICKOPOLG) 101. 1-712) 6 ciel ae ae 12 hours 4 — 14 - Mellita 6-perforala .......... Me 6 — 1 day “on 5 — Bie Astriclypeus Manni ......... “ 14 — a3 3 = 6 — 13 — Peronelia ‘Lesueurt <.,. 0... -\- 5— 12 - 18—20 hours ma 2 3—4 — Echinobrissus recens......... aun 24 - 2 days 3t ca.15 == Brissus) :AigassiZiios «352 is =) oe 14 — Nie 7 — WVLCOIMGs QRONGIS se reurneus «iss om 14 — 20 hours a ve Ophiothrix angulata ......... a 18 - 1 day .. 40—45 hours Ophionereis squamulosa...... oh 15 — we sis Ee ca.6 — Astropecten scoparius ........ ae : 7 days (less?) .. 19—21 days A SLENUNU DECLULIPERG wae tenet oir ve 16 2 18 = Gymnasteria carinifera....... ca.24 hours 2 days 5 Ophidiaster Guildingii ....... 1 PisasterOchraceus’, . 1. +011 24— 2 \ Euasterias Troscheli......... 24— 2—3 1 . Stichopus californicus.,...... 24- 2 6 - It is evident that the amount of yolk substance contained in the egg has an important bearing on the length of time required for the develop- ment. When food enough is contained in the egg for sustaining the embryo until metamorphosis is completed, the self-feeding larval stage is done away with as unnecessary and the development thus considerably short- ened, e.g. Heliocidaris erythrogramma, Peronella Lesueuri, Ophionereis squa- mulosa and many other forms with abbreviated development, not mentioned in the present memoir. That also temperature has an important bearing on the time required for the development is an established fact; interest- ing results are sure to be obtained on studying the development of widely distributed forms in various places, where conditions are different (tropical, extratropical). The few facts known, e. g. of the development of Peronella Lesueuri in tropical seas, as compared with the observations given above, tend to show that the development proceeds at a conspicuously quicker rate at the higher temperature of the tropics than in the cooler climate of extratropical regions. The said factors, yolk and temperature, cannot, however, account for all the differences. Thus e. g. of Tripneustes esculentus and Lytechinus varie- gatus, both living in quite shallow water in the tropics and both having small eggs, poor in yolk substance, the former takes three weeks to reach the stage of beginning metamorphosis, the latter only 13 days; or En- cope micropora assuming the shape of a young Pluteus already at the age of 12 hours, while Mellila 6-perforata, living under similar conditions re- quires the double time for reaching that stage. But altogether too little is known as yet for giving a reasonable base for an attempt to find out the causes of these differences. The observations recorded here may only serve to prove that here is a problem worth studying. In a very interesting paper on “Sea-temperature, breeding and distribu- tion in marine animals!) I. H. Orton comes to the result that “in those parts of the sea where temperature conditions are constant or nearly con- stant, and where biological conditions do not vary much, marine animals will breed continuously.’ He concludes that this will be the case in the tropics, founding on the statement of Semper (“Animal life’, p. 110) that in the Philippines he could not detect a single species (of Invertebrates) of which he could not “at all seasons find fully grown specimens, young ones and freshly deposited eggs.” This phenomenon, Orton states, “appears to be generally recognized for the tropics, but it would appear that definite systematic work on the breeding and rate of growth throughout the year ‘) Journ. Mar. Biol. Assoc. United Kingdom. XII. 1920. 216 of a large number of animals in the tropics would still be very useful.” (Op. cit. p.355). With the latter sentence I most heartily agree, but otherwise I must decidedly object to the statement that in the tropical seas the animals are continuously breeding at least as regards the Echinoderms, and I can- not but wonder how Semper came to the above result. It is true that I have made only few observations, as regards this point, during my stay at the Philippines; but at any rate I found that the large Synapla Beselit had no ripe sexual products in February—March at Zamboanga. And it is certainly not to be assumed that the animals behave otherwise at the Philippines than elsewhere in the tropics, as regards their breeding. My observations decidedly prove that at least several of the littoral Echino- derms do not breed continuously. Thus e. g. I never found the opportunity for studying the development of Diadema, until I came to Tobago, B. W. I. and there found D. antillarum to have ripe sexual products in the end of March; and when, about a week later on, I wanted to start a new larval culture it was impossible to find one specimen containing ripe sexual pro- ducts, all were quite empty. A similar experience I had with Echinometra van Brunli, Brissus obesus, Slichopus Kefersteinii, and in several cases | had, to my great annoyance, to give up any hope of obtaining a larval culture, because the breeding season did not coincide with my stay (e. g. Echinoneus). —- On the other hand, I have some observations tending to show that in the tropical seas Echinoderms (some forms, at least) have more than one breeding season in the year. But for proving this defini- tely observations must be continued through a longer period, a stay of a few months being, of course, insufficient for such a task. I would recall here that I have been able to prove (Die Echinoiden d. Deutschen Siidpolar-Expedition, p. 71)"') that the Antarctic Slerechinus Neumayeri breeds at least from June to April, which fact is doubtless due to the very uniform temperature reigning in the Antarctic Sea throughout the year. This case then confirms Orton’s statement that “where bio- logical conditions do not vary much, marine animals will breed continu- ously.’ I would only object to making this a general rule; this it is cer- tainly not, especially not in the tropics. ') Deutsche Siidpolar-Expedition, Zoologie. II], 1909, 247 Ill. Geographical Distribution. In my work “Die Echinodermen-Larven d. Plankton-Exped.” p. 108— 114 the distribution of the Echinoderm larve and some of the problems connected therewith were discussed at some length. Since then only very little additional information on this subject has been obtained. MacBride in 1903 published a “Report on a small collection of Echinoderm larvee made by Mr. George Murray during the Cruise of the “Oceana” in Novem- ber 1898"’4); only one species, the larva of Luidia Sarsi, is recorded there, from the Atlantic off S. W. Ireland (52°4’ N., 12°27’ W.), Gemmill?) has recorded some larve of Luidia ciliaris likewise from off S. W. Ireland (30—60 miles off Tearaght Island), and one single specimen of “Brachio- laria hibernica” from 50 Miles N. by W. of Eagle Island, W. of Ireland. In my report on “Die Echinodermen-Larven d. deutschen Stidpolar-Expedi- tion” (p. 109) a Spatangoid-larva is recorded from 10°17’ I. 28°45’ S., that is about in the mid-Atlantic between Cape and Patagonia, at a depth of nearly 5000 Meters. Apart from this highly interesting find only larvee taken near the littoral regions are recorded in that work, as also in Mac- Bride’s reports on Echinoderm larve from the Antarctic Expeditions (National Antarctic Exped.; British Antarctic (“Terra Nova’) Exped.). The existence of quite a considerable number of pelagic Echinoderm larvee in the Antarctic Seas has been established herewith; but this fact, though of very considerable interest, does not concern us here. Referring still to the *Stellosphera mirabilis” recorded by Koehler & Vaney from the vicinity of the Azores (ca. 36—40° N. 19--30° W.) mention has been made of, so far as I know, all records giving observations on the occurrence of Echino- derm larve in the open Ocean.*) As stated in the Introduction (p. 10) I have at various occasions made efforts to bring together some more facts relating to the problem of the distribution of Echinoderm larve over the open Ocean. Observations made during the passage over the Indian Ocean from Aden to Singapore gave the result that, while Echinoderm larve were found in good numbers in the Gulf of Aden, none were observed until near the Maldives, but from there larvee occurred the whole way across the Bay of Bengal. This would appear to indicate that larve do not occur in the great Arabian Sea. I would, however, not regard the negative result as conclusive, because the samples were taken there at day time. Later on samples were taken also by night, and these were invariably richer than those taken by day, in 1) Ann. & Mag. Nat. Hist. VII. Ser. XI. p. 477—78. *) J. F. Gemmill. On a new brachiate Asteroid larva and on the advanced Bipinnaria of Luidia ciliaris (Philippi) Gray. Proc. R. Phys. Soc. Edinb. XIX. 1915. 5) See, however, the paper by Stanley Gardiner mentioned below (p. 251). 248 conformity with the well known fact that most pelagic animals avoid the strong daylight by going down in the sea, raising again to the surface at night time. I would expect that samples taken by night will prove larvee to occur also all over the Arabian Sea. But that remains to be proved. Besides the material of plankton samples collected for me on various cruises across the Atlantic by Messrs. Blegvad, Kramp and Fogh (comp. p. 10) I have also had the opportunity of examining some other samples like- wise taken on the way to and from the West Indies, kindly placed at my dis- posal by Dr. Johs. Schmidt. The samples were all taken at the surface by night time. The following larve were obtained from these samples. Spherechinus granularis... 36°13’ N. 33°50’ W. 15/111 1911 Blegvad Off Azores -— — np o0l42 30°30 = 5 LoD - — - —— Echinopluteus transversus .. 32°10’ - 17°20’ - 30/X 1911 Kramp - Madeira 36 ENO Sa DalGeh 6S" By eats — - Canaries Ba, op py melee = alps Tals Fogh - Madeira .. 28°43’ - 20°40’ - 22/X - - Canaries Echinopluteus sp.......... 25°40’ - 24°10’ - 24/X - -- sit EMPL H Gh eR this 18°45’ - 62°20’ - 19/XI 1911 Kramp Virgin Isl. IGGL TOUS. elora oa Bede 36°13’ - 33°50’ - 15/JI11 —-(Dr.Schmidt) - Azores - SGN gomtrnos Cee i Byles = Shei Se SIA — - —- Sarsi(?)2)is wn. cee BOTOO = 49957). = 227 : Sargasso Sea Bipimnnaria spesiat-eeael> - AQ SATS = 2 TSO fer. 2 TI Blegvad Off Azores =e Oh) aor 32°22 - 22°49" — 26/1TI —- . : Sey eta a autere aie: crs 20°46’ - 41°16’ - 11/XI 1911 Kramp Sargasso Sea ats pact h tom hea shay hsgs 19°36’ - 47°13’ - 13/XI' — - dae Aes eo Lue, Oreo ee LOY Led Oe PAE ROLTe OMG Oe 25°16’ - 55°37’ - 20/II : LOS OOS ee GY Onc 24°33) 60012) — 1S ois Fogh Auricularia nudibranchiata. 30°30’ - 49°57’ - 22/11 1911 (Dr.Schmidt) . 33°55’ - 43°40’ - 12/111 — , 994639 & 40°54! 55. Au , 36°13’ ~ 33950" - 15/111 — , 39°22’ .- 22°49’ - 20/1TI Off Azores : 40°47 21°10) = 21/0 — - : 45°32" © 195°50' =) 24yvi : The result is seen to be in fair accordance with that previously obtained from the study of the material collected by the Plankton-Expedition, viz. that, while in general the Echinoderm larve occur more numerously in the coastal waters, several specimens may be found also in mid-Ocean. This raises the interesting problem: how did they come there? Were they carried out from the coastal waters by the currents, or did they rise from the bottom? So long as we do not know the parental origin of the larve found in the open sea the question cannot be answered definitely. Most probably they 1) Identification of the single specimen a little uncertain on account of the very poor preservation. 249 come from both sources. The question whether any deep-sea Echinoderms have truly pelagic larve must, after what we now know, be answered in the affirmative. Laganum diplopora has been directly proved to have pel- agic larve, and it is almost certain that also Pedicellasler 6-radiatus has pelagic larve (S/ellophera mirabilis); judging from the character of the eggs many other deep-sea forms must have pelagic larve (though a great per- centage of them have large eggs and therefore certainly must have an abbreviated development!). On the other hand it is evident from the facts recorded that larve of littoral forms may be carried very far from the coasts. In this connection it is important to notice that Bipinnariz are predominant among the larve found in the Sargasso-Sea (beside the large Auricularia nudibranchiata); this is in good accordance with the fact that the Asterid-larve in general require long time for their development, much longer than do the Echinoid-larve. Thus there will be ample time for their being carried great distances by the currents. The transport of the larve by means of the currents is, of course, of considerable zoogeographical importance. Their eventual transport across the Atlantic has some bearing on the problem of a former land connection between Africa and South America (the Archhelenis-theory)?). It is a well known fact that quite a good number of littoral Echinoderms are common to. the West Indies (Brazil) and West Africa. This peculiar distribution can be accounted for only in two ways, viz. by the existence of a former land connection (or a series of islands) between the two continents along the shores of which these Echinoderms were formerly distributed, or by the transport of their larvee across the Ocean. That the latter alternative is really possible seems undeniable from the facts made known of the occurrence of larve in the open Ocean. Accordingly the Archhelenis-theory is no necessity for understanding the recent distribution of the littoral Echinoderms of the tropical Atlantic. On the other hand, it must be agreed, there are many other facts, especially in the geographical distribution of land- and freshwater animals and plants, which seem to require that theory for an explanation. The occurrence of larvee of deep-sea forms at the surface of the ocean is another problem not very easily explained (— due allowance being made for the fact that we have as yet not definitely ascertained a single case of a larva of a deep-sea form found at the surface; we do not know ') Pelagothuria natatrix has large eggs, until 1,2 mm in diameter according to Ludwig (“Albatross’’-Holothurioidea. 1894. p. 119). This large size of the eggs probably means that also this form has a direct development, and we have thus the remarkable fact that this typical pelagic Holothurian has not a typical pelagic larva. *) H. v. Ihering. Archhelenis und Archinotis. Gesammelte Beitrage zu einer Ge schichte der Neotropischen Region. 1907. 250 exactly the depth at which Siellosphera mirabilis was taken, and we have not yet definitely proved Auricularia nudibranchiala to belong to a deep- sea Holothurian). It seems impossible that the larvee could rise all the way by means of their ciliary movements. One cannot help suggesting that vertical currents may have something to do with the transport of the larve from the deeper layers towards the surface and the inverse. The existence of vertical currents is a suggestion called forth by various biolog- ical facts. Thus e. g. Prof. Ad. S. Jensen informs me that the distribution of the eggs and the young of Reinhardtius hippoglossoides (Walb.) in the Davis Strait seems to be explicable only on the assumption of vertical currents existing there. While the eggs and newly hatched young larve, with the yolk sac, are found in depths of 600—1000 Meters, the slightly older young, where the yolk sac has just been absorbed, are found at 30 Meters below the surface!). It seems simply impossible that these young fish larvee could have passed that great vertical distance alone by means of their own feeble swimming power. Hjort and Murray in “The depths of the Ocean” (p. 378—-380) think that vertical circulation has a great influence upon the distribution of pelagic plants, and upon the whole speak of vertical circulation in the Ocean as an established fact, at least for the upper 200—3090 Meters. This, however, will not suffice for explaining the ascent of the larve from the bottom at much greater depths. This inter- esting problem cannot be solved at our present state of knowledge. Upon the whole, we are hardly in possession of facts enough even for a definite formulation of the problems connected with the deep-sea larvee. We can only see this much that continued researches will be sure to bring im- portant results. ') Comp. Rapport til Indenrigsministeriet over Briggen “Tjalfe’’s praktisk-videnskabe- lige Fiskeriekspedition til Grenland. 1909. Appendix. In a paper by D.H. Tennent on “The early influence of the spermatozoa upon the characters of Echinoid larve’”’ (Papers from the Tortugas Labora- tory of the Carnegie Inst. Washington, Vol. V, 1914) some observations on the early developmental stages of Eucidaris tribuloides (Lamk) are recorded. It is evident that the larva of this species agrees closely with that of Eucidaris Thouarsi (comp. p. 22, Pl. V, Figs. 1—2). In the Pluteus, 6 days old, represented in his figure 6 (p. 133) the skeleton is shown; the postoral rods are fenestrated, horizontally directed. One cannot help being struck by the resemblance with the Echinopluteus transversus shown by this young larva, and — in spite of the ophicephalous pedicellarize —. the conclusion seems almost unavoidable that Echinopluteus transversus really is the Cidaris-larva. The paperisnot recorded either in the “Zoological Record” or the *Biblio- graphia zoologica.”’ It was therefore only per chance that I discovered it in time for mentioning this important observation by Tennent in this place. When speaking of the phylogenetic importance of the larva of the viviparous Ophionolus hexactis (p. 238) I forgot to mention the larva of Chirodota rotifera, which, although developing within the body cavity of the mother, has also fairly distinct ciliated bands like those of the typical, free-living larve.') Attention may still be called to an interesting paper by I. Stanley Gardiner: ‘Notes and observations on the distribution of the larve of marine animals’. (Ann. Mag. Nat. Hist. VII, Ser. XIV, 1904, p. 403—410), in which the problem of the importance to be ascribed to marine larvee “in distributing species and genera from shore to shore, from one littoral zone to another” is discussed. In general the views of Stanley Gardiner are not in contradiction to those set forth in the present work, and | do not see any reason for entering on a detailed discussion thereof, the more so as I most sincerely agree with the author in the main object of ‘) H. L. Clark. Development of an Apodous Holothurian (Chirodota rotifera). Journ Experim. Zool. IX, 1910, 252 his paper “to direct attention to the subject in the hope that zoologists who are dealing with plankton will not confine their attention merely to the adult groups of the same, but will, in addition, arrange for ex- amination of the larvee therein in view of the distribution of the different groups of littoral animals.’’ I would only express my astonishment at the conclusion he arrives at, viz. that “it would appear to me (Gardiner) that no results in distribution can be expected, so far as the Indo-Pacific is concerned, from Echinoderms — and probably also from Enteropneusts — other than the Crinoids,”’ the pelagic life of which latter is consider- ably shorter than that of typical pelagic Echinoderm larve. Probably this statement is not meant so absolutely as it sounds. I also venture to think that the fear expressed by Stanley Gardiner “that in the present state of our knowledge any consideration of larval distribution is premature and must be inconclusive’ has been partly removed by the present researches. EXPLANATION OF THE PLATES Some of the figures of Ophiurid-larvee were drawn by Mr. I. Lieberkind, who has also drawn some of the skeletal figures of these larvie given in the text. I beg to express here my indebtedness to Mr. Lieberkind for his valuable assistance. Fig. Fig. Fig. Fig. Plate I. Larva of Echinometra lucunter (Linn.), two weeks old, fully formed; seen from the ventral side. Drawn from life. 1°°/;. Larva of same species, same age; from the dorsal side. At the base of the left postoral arm is seen the beginning amniotic invagination. Drawn from life. '°/;. Plate II. Larva of Tripneustes esculentus (Leske), 17 days old; from the ventral side. Drawn PrOME Lites oo cs Larva of same species, 22 days old; from the dorsal side. At the base of the left postoral arm is seen a small invagination, the amnion. A pedicellaria has begun to form in a groove in the posterior end. The forward direction of the ventral ciliated lobes (v. 1.) is due to pressure of the larva under the cover. Drawn from life. 1°°/;. Plate III. Larva of Lylechinus variegalus (Lamk.), 13 days old, seen from the dorsal side. Drawn from life. The amnion has been formed and in the posterior end a pedicel- laria (p) is beginning to develop. 1°°/,. Larva of same species, 15 days old, seen from the ventral side. Drawn of life. '°°/,. Plate IV. Larva of Astriclypeus Manni Verrill, 6 days old; seen from the dorsal side. am. amnion. Drawn from life. 1°°/,. Larva of Mellila seviesperforala (Leske), 7 days old; seen from the dorsal side, turned slightly upwards; in beginning metamorphosis. Drawn from life. On account of the opacity of the larva the details of the body skeleton could not be discerned. a0, o Plate V. Embryo of Eucidaris Thouarsi (Val.), 3 days old. *°°/,. Young larva of same species, 6 days old, slightly restored. ‘The skeleton dissolved. The small body to the left of the mouth probably represents part of the enterocoel. 285) Pa Section of young embryo of Phyllacanthus parvispinus Woods, in the blastula stage. LAO 1 Section of embryo in the gastrula stage of same species. '°°/,. Larva of Diadema antillarum Phil., 5 days old; from the ventral side. Drawn from Fig. Fig. Fig. 6. o- = w life. **°/,. The line seen across the ventral transverse rods is the lower limit of the suboral cavity. Young larva of Astropyga pulvinala (Lamk.), 5 days old; ventral view. Drawn from life. 15%/,. Larva taken pelagically, probably belonging to Astropyga pulvinala. Seen in some- what oblique side view. Some dislocation has occurred; the anal area is somewhat twisted to the left so that the side area of the larva has become too broad. The figure is slightly corrected, the postoral (p. 0.) and posterodorsal (p. d.) arms being in a more upright (but unnatural) position in the specimen from which the figure has been drawn. The basal part of the left postoral rod not to be seen quite di- stinctly, that of the posterodorsal rod quite indistinct in the specimen. 1°°/;. p. tr. posterior transverse rod. Plate VI. Larva of Arbaciid; taken pelagically off the Azores. he postoral and posterodorsal arms broken; their length must considerably surpass that shown in the figure. 7*/,. The same larva, more magnified (°°/,). The larva is in beginning metamorphosis, the anterolateral and preoral arms are not broken. The large fenestrated plate at the base of the postoral (and, less developed, the posterodorsal) rods is not cha- racteristic of the larval skeleton but is connected with the metamorphosis. p. pedicellaria; p. tr. posterior transverse rod. Larva of Heliocidaris tuberculata (Lamk.), seen from the posterior end; showing the position of the ventral (v. tr.) and dorsal transverse rods (d. tr.); p. tr. posterior transverse rod; da. dorsal arch; p. 0. postoral, p. d. posterodorsal arm; p. |. postero- lateral lobe. 225/;. Larva of Temnotrema sculpla A. Ag., 20 days old. Drawn from life; ventral view. The skeleton was merely sketched in the original drawing; as it has been dissolved in the preparation, so that the sketch could not be completed, it had to be omitted in the figure. 1°9/,. -6. Young larve of Hololthuria n. sp. (?), from the ventral side. Fig. 5 represents a specimen 3 days old, Fig. 6 another, 8 days old. The rectum is indiscernible. The frontal area in fig. 6 probably too narrow, due to contraction on preservy- ation. 45%/7; Plate VII. Larva of Mespilia globulus (Linn.), 15 days old; seen from above, but in such a position that the body is leaning somewhat over to the ventral side. The fact that the posterodorsal arms are narrower than the postoral arms in this figure is due only to the oblique position; in reality these four arms are of the same width. Drawn from life. 1°/,. Larva of same species, ventral view; drawn from a preserved specimen. (This ac- counts for the narrowness of the postoral and posterodorsal arms as compared with fig. 1.) Beginning metamorphosis; two pedicellarie have appeared in the posterior end. Same age as the larva represented in fig. 1. 1°/;. Larva of Arbacia stellata (Blv.); dorsal view. °°/;. Young larva of Lylechinus panamensis Mrtsn., ventral view; 5 days old. Drawn from life? 45°/,. Young larva of Lylechinus anamesus H. L. Clark; side view; 7 days old. Drawn from life. *°°/;. Plate VIII. Young larva of Lylechinus variegatus (Lamk.), 2 days old. Ventral view. Drawn from) life. +°/; Young larva of Tripneustes esculentus (Leske), 3 days old. Ventral view. Drawn from life. 2°°/,. : Larva of Lytechinus anamesus H. L. Clark, 7 days old. Not yet fully formed. Ven- tral view. Drawn from life. %°/,. Larva of same species, same age as fig. 3; dorsal view, Drawn from life, °°/). Fig. oO Young larva of Tripneustes gralilla (Linn.), 3 days old. Dorsal view. The skeleton has been dissolved. 2°°/,. Same as Fig. 5; side view. *79/,. s.c. suboral cavity. ~ Larva probably belonging to Toxopneusles roseus (Ag.). Not yet fully formed. Slightly distorted. Ventral view. '°/,. Young larva of Toxopneustes pileolus (Lamk.), 9 days old. Ventral view. 1'8°/,. Plate IX. Young larva of Strongylocentrotus franciscanus (A. Ag.), 4 days old. Dorsal view. 150/, Larva of same species, 8 days old; ventral view. 15°/,. —- - — — 10 days old; side view. 15°/,. Fully developed larva of Strongylocentrolus franciscanus, taken pelagically. Drawn from a preserved specimen. The course of the ciliated band along the sides not quite distinct in the specimen. 1°/,. Larva of Strongylocentrotus pulcherrimus (A. Ag.), 4 weeks old. Ventral view. *°°/,. Figures 1—3 and 5 drawn from life. Plate X. Young larva of Evechinus chloroticus (Val.), 14 days old. Ventral view. 1°/,. More advanced larva of same species, 3 weeks old. Dorsal view. !°/;. Fully developed larva of same species; nearly 5 weeks old. Ventral view. 1°/, 4—5. Two abnormal young larv of Evechinus chloroticus. °°/,. 6. ~l — i) 3. 4. or —_ — Nearly fully developed ‘larva of Temnopleurus toreumaticus (Klein); ventral view. 9 days old. Drawn from life. 1°/,. Fully developed larva of Arachnoides placenta (L.), 7 days old. Ventral view. '°°/, Plate XI. Young larva of Heliocidaris tuberculata (Lamk.), 6 days old. Dorsal view. 22°5/ Fully formed larva of same species, 5 weeks old. Ventral view. In the posterior end a pedicellaria has appeared; a muscle is seen connecting the lower ends of the postoral rods. Drawn from life. 15°/,. Larva of Temnopleurid, species c. Slightly restored. The true shape of the epaulets not to be ascertained. 15°/,. Larva of Echinobrissus (Oligopodia) recens (M. Edw.), 4 days old; ventral view. !°°/, Larva of same species, 10 days old; ventral view. 1°°/,. Plate XII. Larva of Echinometra lucunter (Linn.), 7 days old; dorsal view. The body skeleton has begun to be absorbed. Drawn from life. Showing the nervous system (the oblique line on the anal area, at the base of the postoral arms). ?°°/,. Larva of Echinometra oblonga (Blv.), 12 days old; side view. Body skeleton partly dissolved. 7°°/,. Larva of same species, same age; ventral view. The skeleton has been omitted in the figure, being partly dissolved so that a correct drawing of it could not be made. (Comp. fig. 2). 7°/,. Larva of Colobocentrotus atratus (Linn.), 12 days old; side view. *°°/,. Larva of same species, same age; ventral view. Skeleton dissolved. Drawing from a preserved specimen, combined with a sketch from life. *°°/,. Young larva of Temnopleurus toreumaticus, 22 hours old; ventral view. ‘The ciliated band rather too distinct in the figure. 72°/,. Plate XIII. Echinopluteus transversus, species ¢. Ventral view. §°°/;. : a. - Slightly corrected, the postoral band having been shown in what is evidently its normal place, not in the place where it is seen in the specimen, below the end of the ventral transverse rods. '*°), 256 Fig. 3. Echinopluteus transversus, species f. Ventral view. The vibratile band not distinct over the Echinoid-rudiment. *°/,. - 4. Same larvg as fig. 3, in dorsal view. 85/;. 5. Young larva of Mellita sexiesperforata (Leske); 24 hours old. Ventral view. 15°/;. 6. Larva of same species, same age; side view. 15"/;. Plate XIV. Fig. 1. Young larva of Laganum diplopora H. L. Clark; 3 days old. Ventral view. *°°/,. - 2. Embryo of Clypeaster japonicus Déderl., 24/, days old. Showing beginning forma- tion of the skeleton. 15°/,. 3. Larva of same species, 11 days old. Ventral view. Drawn from life. 1*%/,. 4. Gastrula of Echinarachnius mirabilis (A. Ag.), 18 hours old. 18°/,. - 5. Fully developed larva of Echinarachnius (Dendraster) excentricus (Esch.), 14 days old. Ventral view. From a preserved specimen, combined with a sketch from a living specimen. Skeleton dissolved. 1°°/,. 6. Larva of same species, same age; side view. Slightly restored. 1°°/,. - 7. Larva of Encope micropora Ag., 14 days old; ventral view. From a preserved spec- imen, combined with a sketch from a living specimen. 15°/,. Plate XV. All figures of Peronella Lesueuri (Val.).. Figs. 1—4 drawn from life, *5/,; the rest 15°/,. ' Fig. 1. Embryo, five hours old, in the blastula-stage; showing irregular folding of the ecto- derm. - 2. Young larva, 17 hours old. The mouth has been formed, and the postoral arms are beginning to appear as a pair of small lateral thickenings. - 3. Young larva, 20 hours old; dorsal view. The postoral arms are distinct. - 4, Fully formed larva, 39 hours old; ventral view. The preoral lobe has been completely reduced, the mouth opening now being at the anterior edge of the body. - 5. Longitudinal section of an embryo, 9 hours old. A number of mesenchyme cells have been formed, but gastrulation has not yet begun. - 6. Longitudinal section of an embryo, 12 hours old. Gastrulation has begun. - 7—10. From a series of longitudinal sections of an embryo, 16 hours old; showing the shape of the entoderm. Between figs. 7 and § there is one section, fig. 9 following immediately after 8; there are 3 sections between figs. 9 and 10. - 11-13. From another series of longitudinal sections of an embryo, 16 hours old; showing ‘the shape of the entoderm. Fig. 11 is the more dorsal of the three. In fig. 13 is seen the apical thickening. - 14. Longitudinal, sagittal section of an embryo, 16 hours old. The mouth (m) has begun to form. - 15—18. From a series of frontal, longitudinal sections of an embryo, 18—20 hours old. Fig. 15, which is the more dorsal of them, shows the widening (am.) from the lower end of the pharynx (ph.), surrounding the entoderm (ent.); this widening gradually disappears in the following figures. In fig. 18 is seen the rectum (r) in a lateral posi- tion. There are 4 sections between figs. 15 and 16, 1 between figs. 16 and 17 and 1 sections between figs. 17 and 18. - 19—22. From a series of sagittal longitudinal sections of an embryo, 18—20 hours old. Showing the posterior prolongation (am.) from the pharynx (ph.). Also the shape of the archenteron (ent.) appears from a comparison of these figures. Fig. 20 shows the blastoporus, which has now shifted to the ventral side, to form the anal open- ing (a). Plate XVI. All figures of Peronella Lesueuri (Val.); all 48°/,. Fig. 1. Longitudinal, frontal section of a larva 23—24 hours old. Showing the amniotic prolongation (am.) of the pharynx (ph.) continuing to the posterior end of the body. The part in the middle is the raised ventral wall of the amnion (comp. figs. 14-16): 207 in its posterior part the wall already has assumed an endothelial character. The spaces below the arms are the remnants of the blastocoel cavity (bl.). Fig. 2—4. From a series of longitudinal, frontal sections of an embryo, 32—35 hours old; fig. 2 is the more dorsal; fig. 4 the more ventral of them. The pharynx has been very much shortened and now has simply the character of the opening of the amnion. The Echinoid-rudiment has begun to form lobes, representing the young spines and primary tubefeet. In fig. 4 the curved intestine (i) is seen. The distinctly limited space to the right of it is part of the coelom (c), as is also the large, irregular space in fig. 3. In this latter figure is seen to the left below the anterior part of the amnion some nuclei, serially arranged so as to appear like a canal. This is, however, a casual arrangement of no importance and there is no real canal. 5. Median longitudinal sagittal section of a larva, 21-22 hours old. The amniotic prolongation of the pharynx has reached the posterior end, covering the whole of the dorsal side (comp. Pl. XV, Figs. 20—21). Its dorsal wall is beginning to assume an endothelial character, while the ventral wall has already thickened considerably. bl. blastocoel; h. hydrocoel; i. intestine. 6—7. From a series of longitudinal, sagittal sections of a larva, 26—29 hours old. The amniotic opening, the former mouth, has now been transplaced to the dorsal side. There are two sections between the two figures. Below the thickened ventral wall of the amnion is seen the hydrocoel (h.), and on the ventral side the intestine (i.), the space separating them being the coelom (c.). 8—9. From a series of longitudinal, sagittal sections of a larva, 32—35 hours old; fig. 8 median, fig. 9 more lateral, separated from fig. 8 by 10 sections. The two parts of the coelom seen in fig. 8 are found to be in direct connection in more lateral sections. The space in the middle probably belongs to the hydrocoel (h.?). The distinctly limited part of the coelom in fig. 9 is also seen in the adjoining section to be in direct communication with the rest of the coelom. 10. Median longitudinal, sagittal section of a larva, 41—43 hours old. The amniotic opening is now entirely dorsal. The primary tubefeet are protruding into the amnion. 13. From a series of transverse sections of a larva 18—-20 hours old; fig. 11 is from the anterior end, the two lateral wings being the base of the postoral arms; figs. 12 and 13 from the middle part of the body, separated by three sections. am. amnion; bl. blastocoel; ent. entoderm; ph. pharynx; r. rectum. 14—16. From a series of transverse sections of a larva, 23—-24 hours old. Fig. 14 the more anterior, 16 the more posterior. In fig. 15 is seen the last trace of the blasto- porus, or anal opening (a.). Figs. 14 and 15 are separated by 1 section, figs. 15 and 16 by 4 sections. The beginning folding of the ventral wall of the amnion is to be noticed. am. amnion; bl. blastocoel. The cavity in figs. 14 and 15, design- ated h.c¢., apparently represents both hydrocoel and coelom, not yet separated. 17—18. From a series of transverse sections of a larva, 32—35 hours old. The young spines and primary tubefeet (t.) protruding into the amnion. The two figures are separated by 8 sections, fig. 17 being the anterior. am. amnion; c. coelom; h. hydro- coel; i. intestine. 19—20. From a series of transverse sections of a larva, 41—43 hours old. Showing the pore canal (p.c.). The two figures are separated by 4 sections, fig. 19 being the more anterior. 11 Plate XVII. All figures of Heliocidaris erythrogramma (Val.). All 7°/;. Fig. 1. Embryo, 18 hours old, showing beginning formation of the gastrula. 2. Embryo, 30 hours old, in the gastrula stage. 3. Embryo, 30 hours old, showing beginning differentiation of the archenteron, in the lumen of which is seen a mass of a slimy(?) substance. 4. Further advanced stage in the differentiation of the archenteron, a large pouch having been formed: the hydrocoel (and enterocoel?); 30 hours old. 5. Embryo, 42 hours old, showing the first indication of the amniotic invagination. The archenteron (the dark body in the middle) is connected with the blastoporus by a thin strand. The space seen above the archenteron is the hydro-enterocoel. 33 1 a pn = c 13. 14. wo ~I i. ). 258 Embryo, 42 hours old, showing a slightly more advanced stage in the formation of the amniotic invagination and a beginning differentiation of the hydrocoel. Embryo, 30 hours old, showing the archenteron completely separated off from the blasloporus. Embryo, 2*/, days old, showing a further advanced stage of the amniotic invagina- tion, which has formed a furrow nearly round the body. A distinet, broad ciliated band is seen round the middle of the body. The archenteron has been separated off from the blastoporus. The internal structures very indistinct in this specimen. Embryo, 3*/, days old. The tubefeet of the young urchin have begun to protrude through the amniotic invagination. ‘The internal structures not to be made out in this specimen. ; Embryo, 3°/, days old. The primary tubefeet and the first spines (numbering 10, both together, in this specimen) are protruding through the amniotic cavity, now split open. The pigment shows a rather distinct arrangement in bands. Embryo, 2°/, days old, showing the five primary tubefeet lying within the amnion. Embryo, 4'/, days old. The primary tubefeet and spines have further protruded, and the oral and aboral end of the larval body are pushed up on the dorsal side of the urchin. Nearly metamorphosed sea-urchin; the aboral part of the larval body has been nearly completely absorbed, forming only a small prominence on the upper side. The primary tubefeet are fully extended. Embryonal spines have begun to appear also on the aboral side. Embryo, 4'/, days old. The amniotic invagination has been widened so much that the primary tubefeet and the spines are assuming a circular arrangement. Only two of the tubefeet extended. The pigment forms a very prominent band above the widening of the body. ; Figures 1—9 and 11 are drawn from specimens mounted in balsam and thus made transparent; figures 10 and 12—14 are drawn from specimens not thus clear- ed up. Plate XVIII. All figures of Heliocidaris erylhrogramma (Val.). All, except fig. 4, 18°/. Section of an embryo, 6 hours old, in a young cleavage stage. The formation of the mesenchyme has begun. The nuclei very indistinct. Section of an embryo, 18 hours old, in a young blastula stage. The whole of the blastocoel cavity filled by mesenchyme cells, still in an embryonal condition. Section of an embryo, 18 hours old, in the fully formed blastula stage. The mesen- chyme cells have assumed the character of a reticulum, interwoven between the large vacuoles, containing a probably fatty substance which serves as nourishment to the developing embryo. Part of a longitudinal section of an embryo, 2°/, days old, showing the formation of the amniotic invagination. *°/;. Longitudinal section of an embryo, 2°/, days old; showing the primary tubefeet (t) protruding into the amniotic cavity (am.), which opens out through a small pore. c. enterocoel. Part of a section from the same series as fig. 5. The amnion is closed here. h. hydrocoel. Longitudinal section of an embryo, 42 hours old. Showing the hydrocoel (h) lying close towards the amniotic invagination (am), which is here still an open groove. Trace of the blastoporus is seen in the oral end; ent. entoderm. ~9. Two longitudinal sections of an embryo in metamorphosis, 4'/, days old. In fig. 8 is seen the stone canal and its outer opening, in fig. 9 the inner opening of the stone canal into the hydrocoel. a. ampulla; c. coelom; d. dorsal; ent. entoderm. h. hydrocoel; r. reservoir of food; st. c. stone canal; vy. ventral. Fig. Fig. Fig. ~I A) of wre wo = i) 259 Plate XIX. Ophiopluleus of Ophiothrix species a. Misaki, 17/VI. 14. °°/;. == = — - cx Jolo; 20/1DI. 14.. 807, a x = — b. Taboga, Panama. XII. 1915. see — = —— : bs 43/,, — - _- : - The young Ophiurid has been aavaladed and dropped off, the posterolateral arms remaining in connection and still continuing the pelagic life. Tobago, Panama. XII. 1915. *%/;. Ophiopluteus of Ophiothrix, species d. Somewhat restored, especially the intestinal tractus. Off Jolo. 20/III. 14. 5%°/,. Plate XX. Larva of Ophiothrix angulata Say, var. poecila H. L. Clark. Drawn from life. 40 hours old. 2/,. Same, 41/, days old. Drawn from life. 7°°/,. Ophiopluteus opulentus, species ¢. The young Ophiurid has been developed and dropped off, the posterolateral arms remaining in connection and continuing pelagic life. The arms are not represented in full length, ought to be ca. 4 cm longer by this magnification. The small swelling in the middle to be noticed. *%/;. Same as fig. 3, showing beginning regeneration of the larval body. *°%/;. Same specimen as fig. 4, more magnified. Showing beginning formation of mouth, vibratile band, anterolateral (?postoral) rods. 1%°/,. Plate XXI. Ophiopluleus coslalus, species a. °°/1. —_— -— — b. 90/1, — opulentus, species a. Slightly restored. °°/,. Plate XXII. Ophiopluteus relrospinus, species a. The neryous system is distinctly seen. °°/;. — — — b. Nervous system only partly distinct. °°/;. — opulentus, — c« 43/;. Plate XXIII. Ophiopluteus fulcitus, species a. %/;. — = = b. ee He _ — — « Plate XXIV. Ophiopluteus fulcitus, species a. ™°/,. — undulatus, species c. °°/;. — — - a; in metamorphosis. 1°/;. The details of the skeletal plates of the young Ophiurid could not be given quite accurately by this magnification. The first ventral plate has appeared; in four of the interradii is seen an unpaired plate, which evidently represents the mouth- shield. The preoral band was not very distinct and is perhaps not quite correctly represented. Plate XXV. ae undulatus, species a. Slightly restored. %°/). - b. Slightly restored. °%/,. The three inner arm- pairs probably still longer than represented. The widening on the ends not distinct, drawn only from analogy with the other species of this type. The shape of the preoral band perhaps not quite correct; the sinuation distinct only on one side. 33” Fig. Fig. Fig. i] _= © 260 Plate XXVI. Ophiopluleus arcifer, species a. *°/,. = b. 80/,. — c, 80,, Plate XXVII. Ophiopluleus formosus, species a. °°/,. - b. Slightly restored. °/,. stmilis. */;. Plate XXVIII. Ophiopluteus serratus. °°/,. — , beginning metamorphosis. °°/,. Plate XXIX. Ophiopluteus pusillus, species a. 1°/,. = == b. 15/,. - - — -; in metamorphosis. 1'%/,. The remnants of the vibratile band to be noticed. Ophiopluteus pusillus, species c. 1°°/;. smear diegensis. ‘°/,. The small spots in the posterior part of the body are pigment cells. j Plate XXX. Ophiopluteus of Ophiocoma, species c. slightly restored. 1°°/;. formosus, species a. in metamorphosis. *°/;. os monacanthus. °°/,. -= —- , in side view, somewhat distorted. ®°/,. Plate XXXI. Larva of Ophionereis squamulosa Koehler; 45 hours old. Drawn from life. Ventral view. 12°/,. Same larva, same age. Dorsal view. ‘The 6 primary plates and three of the terminal plates have appeared; the oral skeleton has not yet appeared in this specimen. c. b. ciliated band; hp. hydropore; pr. t. primary tentacle; tf., second pair of tubefeet. *45/,. Same larva, 2'/, days old; from the ventral side. The mouth opening has been formed. Letters as in fig. 2; further stc. stone canal; b.t. buccal tentacles. °15/;. Same larva, 6 days old; the metamorphosis nearly completed, but the anterior end of the larval body and traces of the ciliated bands (cb.) are still distinctly visible. The first ventral plate (v.) and the first side plate with its spine have been formed. The mouth shields have not yet appeared. t. terminal plate. *!°/,. Longitudinal, frontal, section of larva of Ophionereis squamulosa, 26 hours old; e.c. enterocoel; h. hydrocoel; st. stomach. 15°/,. Larva of same species, 26 hours old; drawn from a cleared up specimen. The hydro- coel is in a more advanced stage than in Fig. 5, showing the five radial canals. The enterocoel is indistinct. st. stomach. 18°/,. Cleavage stage of Amphiura vivipara. ™°/,. Further advanced stage of the same; the nuclei have arranged themselves in a layer along the surface so as to form the ectoderm. ™°/,. Young embryo of Amphiura vivipara, showing the arms curved over the dorsal side, in the skin of which the primary plates are observed, irregularly arranged. °°/,. Plate XXXII. All figures of Ophionotus hewactis (E. A. Smith). 1—2. Young larva with a right hydrocoel. 1 from the ventral side, 2 in side view. 1%°/,. 261 Fully formed larva; the hydrocoel has formed the primary lobes. The rudimentary skeleton (sk.) is seen. 1°°/;. Larva in a slightly less advanced stage, with a right hydrocoel. 1°°/;. Fully formed, normal larva. The round hole seen in the stomach is the entrance to the rectum, which is seen here directly from above. '°°/;. Fully formed larva, in side view. The oral lobe is bent somewhat forwards so as to cover the oral region. Abnormal larva. 1°°/;. The young Ophiuran lying within the ovarial sac, on the wall of which are seen two small, nearly absorbed eggs (e.). °°/). An ovary, attached to the genital rhachis; a trabecule of connective tissue is pro- ceeding from the ovary. °°/;. a. c. anterior coelom; e. egg; h. hydrocoel; m. mouth; n. nerv(?); p. ¢. posterior coelom; r. rectum; rh. rhachis; sk. skeleton; st. stomach; tr. trabecule. Plate XXXIII. Larva of Asterina pectinifera (M. Tr.); 18 days old. Drawn from ics) hy Fe Fully formed larva of same species; same age. Drawn from lifes 20°), Larva of Astropecten scoparius M. Tr., 7 days old. s. c. suboral cavity. *°/;. Larva of same species, 19 days old. Side view. *°/;. Larva of same species, in metamorphosis. 19—21 days old. %/;. Larva of Gymnasteria carinifera (Lamk); 26 days old. 1°°/;. Bipinnaria sp. from the Red Sea. °°/;. Larva of Slichopus californicus (Stimps.), 6 days old. Drawn from life. 75/,. Larva of same species, same age; side view. Drawn from life. Coin it, AP gen yilant _" Newey ali i te 4 ’ wit @ yi e@titrs ee i ia a iy of moa Aa ee ar ee A Ae ew { i Wr ai & : ee ny gre arr ae ied gvubige > ha ; r » ‘ mh 1 ‘. aa , = : : A 7 : ‘ two by ees RS Dansk Repr. Anst 1. Echinometra lucunter (Linn.). a? ; noderm-Larve. Fle We Th. Mortensen del. Dansk Repr. Anst. - Tripneustes esculentus (Leske). 4 Pl. II. : ws “Gs y Rts Xo . \ . * . o/s ~ ‘ C aa : pe ; Th. Mortensen dol. Dansk Repr. Anst. v Lytechinus variegatus (Lamk.) od ee en hinoderm-Larve. Pl. IV. : SS ~ A ‘ 7 & 1. Astriclypeus manni Verr., 2. Mellita peiacia (Leske). Th. Mortensen del. Dansk Repr. Anst. PR ifortansen a 1—2. Eucidaris Thouarsi (Val.); 3—4. Phyllacanthus parvispinus Woods; Danske Anse: 5. Diadema antillarum Phil.; 6—7 (?) Astropyga pulvinata (Lamk.). ‘(é) ‘ds u ermyjofoxy “9—6 “BY “YW e3d[nos ewajjoUWayT “> fo eet S(yue]) eyEpMozaqny sueppoyap] “¢ ‘ds epeqry "7—] "ysuy ‘sdey 4sUeG “WAIL T “Ula pouryoy hinoderm-Larva. Py Vil. 1—2. Mespilia globulus (Linn.); 3. Arbacia stellata (Blv.); 4. Lytechinus panamensis Mrtsn.; 5. Lytech. anamesus H. L. Clark. Th. Mortensen del. Dansk Repr. Anst. 1. Lytechinus variegatus (Lamk.); 2. Tripneustes esculentus (Leske); 3—4. Lytechinus anamesus H. L. Clark; 5—6. Tripneustes gratilla (Linn.); 7. Toxopneustes roseus A. Ag. (?); 8. Toxopneustes pileolus (Lamk.). Th. Mortensen del. Dansk Repr, Anst. ichinoderm-Larve. Pl. IX. 1—4. Strongylocentrotus franciscanus (A. Ag.); 5. Str. pulcherrimus (A. Ag.). one Th. Mortensen del, 7 = chinoderm-Larve. u fod 4 = = =| = d ; z é & 2 6 2 ed} F Echinoderm-Larve. Pls SX KI 1--2. Asterina pectinifera (M. Tr.); 3—5. Astropecten scoparius (M. Tr.) 6. Gymnasteria carinifera (Lamk.); 7. Bipinnaria sp.: 8—9. Stichopus californicus (Stimps.). Th. Mortensen del. ¥ 25 Fre Ft! a feie} Sees sc tgriesta Sgrgheresses Seat re lt atataraet thtetagerer Ss et tot be Pl Cheba ee oe) WOROTOTPSOETER! : »-+ sezncaceuestetspageRaenepeeepapeps ps isstecssschshpe gra theedseasabans pi tategsosté pesessesicgsesssersesesterssttsssisacatacseceetatssts dice esse cig eghiar peangrgeieoatesestoesesseress geaese= : 1) || my \\| Hi} \| GE SMITHSONIAN INSTITUTION LIBRARIES © Oo mo ~m w tT Oo Oo oO fo.0] ie.0) oO Oo mo