^^^''Wi HARVARD UNIVERSITY LIBRARY OF THE Museum of Comparative Zoology Iw OCT 13 1927 U^i^q /IDemotrs of tbe /IDuseum of Comparative Zoolog? AT HARVARD COLLEGE. Vol. XLV. No. 1. REPORTS ON THE SCIENTIFIC RESULTS OF THE EXPEDITION TO THX-; TROPICAL PACIFIC, IN CHARGE OF ALEXANDER AGASSIZ,H BY; ( P^IIj: U. S. FISH COMMISSION STEAMER "ALBATROSS," FROM AUGUST, 3.^9?^ to JUNE, 1900, COMMANDER JEFFERSON F. MOSER, U. S. N., COM- MANDING. XIV. THE SOLENOGASTRES. By HAROLD HEATH. WITH FORTY PLATES. (Published by permission of George M. Bowers, U. S. Commissioner of Fish and Fisheries! C.\MBRIDGE, U. S. A.: printe& for tbe /iDuseum. June, 191L TROPICAL PACIFIC. The foUowhuj I'lihlicdfioiis uf the Museum confain Repurtf: on the Dredging Operations in charge of Alexander Agaxsiz, of the U. S. Fish Conunission Steamer "Albatross," during ISO!) and 1900. Commander Jefferson F. Moser, U. S. N., Commanding. T. A. A(;assiz. Preliminary Report and List of Stations. Witli Remarks on the Deep- Sea Deposits by Sir John Murray. Mem. M. C. Z., Vol. XXVI. No. 1. January, ],902. 114 pp. 21 Charts. I!. A. G. Mayer. Some Species of Partiila from Tahiti. A Study in Variation. Mem. M. C. Z., Vol. XXVI. No. 2. January, 1902. 21 pp. 1 Plate. III. A. Agassiz and A. G. Mayer. Medusse. Mem. M. C. Z., Vol. XXVI. No. 3. January, 1902. 40 pp. 13 Plates, 1 Chart. , IV. A. Agassiz. The Coral Reefs of the Tropical Pacific. :\Iem. M. C. Z., Vol. XXVIII. February, 1903. 33, 410 pp. 238 Plates. V. C. R. Eastman. Sharks' Teeth and Cetacean Bones from the Red Clay of the Tropical Pacific. Mem. M. C. Z., Vol. XXVI. No. 4. June, 1903. 14 pp. 3 Plates. VI. W. E. HoYLE. Cephalopoda. Bull. M. C. Z., Vol. XLIII. No. 1. March, 1904. 71 pp. 12 Plates. VII. H. LuDWiG. Asteroidea. Mem. M. C. Z., Vol. XXXII, July, 1905. 12, 292 pp. 3.5 Plates, I Chart. VIII. W. E. RiTTER and Edith S. Bi-xbee. The Pelagic Tunicata. Mem. M. C. Z., Vol. XXVI. No. 5. August, 1905. 20 pp. 2 Plates. IX. Mary J. Rathbin. The Brachyura. Mem. M. C. Z., \q\. XXXV. No. 2. August, 1907. 54 pp. 9 Plates. X. C. H. Gilbert. The Lantern Fishes. Mem. M. C. Z., Vol. XXVI. No. 6. July, 1908. 23 pp. 6 Plates. XI. A. Agassiz. Echini. The Genus Colobocentrotus. Mem. M. C. Z., ^'ol. XXXIX, No. 1. November, 1908. 44 pp. 49 Plates. XII. J. Murray and G. V. Lee. The depth and marine deposits of the Pacific. Mem. M. C. Z., Vol. XXXVIII, No. I. June, 1909. 170 pp. 5 Plates. 3 Maps. XIII. W. C. Kendall and E. L. Goldsborough. The Shore Fishes. Mem. M. C. Z., Vol. XXVI. No. 7. February, 1911. 106 pp. 7 Plates. Xn'. H. Heath. The Solenogastres. Mem. M. C. Z., Vol. XLV. No. 1. June 1911, 180 pp. 40 Plates. /IDemotrs of tbe /iRuseum of Comparative Zoolo^v AT HARVARD COLLEGE. Vol. XLV. No. ]. REPORTS OX THE SCIENTIFIC RESULTS OF THE EXPEDITION TO THE TROPICAL PACIFIC, IN CHARGE OF ALEXANDER AGASSIZ, BY THE U. S. FISH COMMISSION STEAMER "ALBATROSS," FROM AItGUST, 1S99, to JUNE, lOno, COMMANDER JEFFERSON F. MOSER, U. S. N., COM- MANDING. XIV. THE SOLENOGASTRES. By HAROLD HEATH. WITH FORTY PLATES. [Published by permission of George M. Bowers, U. 9. Commissionep of Fish and Fisheries | CAMBRIDGE, U. S. A.: prtnteC* for tbe fllMiseum, June, 19n. TABLE OF CONTENTS. Paoe. INTRODUCTION 9 HISTORICAL REVIEW H GENERAL FEATURES, iMETHUDS, ETC 19 External Characters 19 Internal Anatomy 19 Methods '22 Occurrence 23 Mode of Life, Food 24 Color, .Size 25 Lenoth Index 2(') COMPARATIVE ANATOMY .... 20 Foot and Glands 26 Hypodermis and Product.s .... 27 Spicule Development 2( a If II Lnphoiiii'iiia spiralis 4176 2 537-672 Hawaiian Ids., near Niihau Id. Alcxanilroiiifiiia agassizi 2992 f 2980 6 1 460 Revillafiigedo Ids. off Mexico. " valiila 4382 4389 •■ 4391 1 1 1 603-1350 Off S. California. Halomvnia nra\iila 4804 2 229 Kurile Ids., near Simushir Id. Heipoiiii.'uia plalypoda 4781 11 482 Aleutian Ids., near .Agattu Id. Dondersia californii;a 4303 1 21 California, off San Diego. Irlilhyiiiiii'iiia pnnisa ( 4400 1 4402 20 2 500-542 II II II II Up to the present time iKine of tliese molluscs has been describeci from the North Pucifie. A few species are known to occur about Australia an(.l the Philippines, and an extensive collection was made in the East Indian Archipelago Ijy the Siboga Expedition. The present collection embraces thirty-one species of which all are new with the exception of Limifosaor lalpoideus Heath; the species belong to fourteen genera of which eight are new. From this material it becomes increasingly evident that this group of molluscs is cosmopolitan; but there is no evidence of a bipolarity or any indication that the north and south poles were centres of distribution. And furthermore there is no apparent relation between size and geographical distribution. In concluding this section of the paper it becomes a most pleasant duty to express my obligation to those who have aided in its completion. To Hon. HISTORICAL REVIEW. 11 G. M. Bowers, Dr. D. S. Jordan, and Dr. C. H. Gilbert I am indebted for a portion of the material, to certain data relating to a few of the species and for many courtesies while on board the "Albatross." I am likewise under deep obligation to Drs. E. J. Nolan and H. A. Pilsbry for the use of the magnificent library of the Philadelphia Academy of Sciences, and for several suggestions of a most helpful character. To Dr. W. K. Fisher I am indebted for the identi- fication of the alcyonarian hosts and for a specimen of Chaciodcrma hmmiiensis. I also wish to express my gratitude to my assistants. Miss R. M. Higlcy, and Mr. F. W. Weymouth who have greatly lessened the burden necessarily involved in such a study as this. And finally my indebtedness to the late Mr. Alexander Agassiz is very great; in every way possible he helped the work along. HISTORICAL REVIEW. The first known reference to any species of Solenogastre occurs in the works of Loven, who in 1844 briefly described as a gephyrean worm Chaeto- derma nitidulum. During the next thirty years a number of systematists adopted this scheme of classification though there was some difference of opinion concerning the exact position of the species within the group. Diesing ('59), Keferstein ('65), Quatrefages ('65), and Baird ('68) allied it to Sipunculus or Priapulus; Theel ('75) created for it a new family (Chaetodermidae), while M. Sars ('69) placed it among the gephyreans without any comment. Dalyell ('53) in "The Powers of the Creator" gives under the name Vermiculus crassus an abbreviated description and one figure of an undoubted Chaetoderma, ac- cording to Koren and Daniellssen ('77) C. dahjelli, but the description is much too indefinite to make the determination certain. During this time M. Sars discovered another species of Solenogastre, be- longing to a new genus and ultimately to a new family, which he placed ('69) among the MoUusca without any comment whatever, merely giving it the name Solenopus nitidulus. Some years later Tullberg ('75) described what is considered to be the same sjjecies under the name Ncomenia carinata. This last named author's investigations mark a distinct advance in our knowledge of these forms, since they are concerned not only with the study of the external char- acters l)ut with the internal organization as well. In certain respects, especially in the treatment of the urogenital system, the work is seriously at fault, but never- theless it was thoroughgoing enough to lead Tullberg to conclude that, while the species is vermian in what he considered to be probably the most important characters, it is on the other hand decidedly similar to certain of the Mollusca. 12 HISTORICAL REVIEW. In 1876 von Cli'aff siil)jeete(_l ClKntodcnnn nitidulutit to an t'xaniiniition, which considering the amcjunt of material and the methods then in vogue, \\as more than ordinarily searching; and while he, like Tullberg, fell into eiioi' regarding the urogenital organs, his results relating to the other systems, espe- cially the nervous, were of the greatest importance. While not entirely com- mitted to any particular belief regarding the animal's relationships he was inclined to uphold Keferstein, Diesing, and others; and yet he drew attention to the fact that the spicules, gills, mode of egg development, and musculature are so unique that the genus may in reality belong elsewhere, possibly in close proximity to the Turbellaria owing to the close correspondence in the nervous systems. In 1S77 and the following year von Ihering proposed a new classification of the Mollusca based on extensive anatomical researches largely concerned with the nervous system. He drew attention to the very important fact, not pre- viously recognized, that in many fundamental respects the iSolenogastres are allied to the C'hitons. In his opinion the ancestral neomenian was probably not distantly related to the gephyreans or nemerteans and accordingly lacked a true shell, and Chitonellus, with its small shell and extensive girdle, is thus more closely related to the Solenogastres than other Chitons and must there- fore be looked ujion as a connecting link. Owing to the presence of lateral nerves the C'hitons, Chitonellus, and the Solenogastres are clearly differen- tiated from the gephyrean worms ami annelids, so that in this and other respects they approach the molluscs. On this account a new phylum of Vermes, the Amphineura, was created for their reception. The following year ('78) this same author reviewed the woik of Tullberg, Koren and Daniellssen, and von Graff in a suggestive paper, and was more strongly convinced than before that his conclusions were justified. These papers of von Ihering's created much criticism, some of it decidedly hostile, but it is undoubtedly true that, while some of the results have not proved to be correct, the work as a whole had a stimulating effect and has been i)roductive of much good. During this same i)eriod ('77) Koren and Daniellssen described a number of species collected along the Scandinavian peninsula. In most cases the descriptions are so brief that they are not even of generic value, and the animals remain practically imknown down to the present day. Neomenia carinata {Solcnopus nitidulus) is described at some length, but the facts adduced are not strikingly different from those presented by Tullberg. By these authors the Solenogastres were considered to be true molluscs, and were placed among the opisthobranchs in a new order, Telobranchiata. HISTORICAL FlKVIFAV. 13 During this same year Lankester (77) placed himself on record t(i the effect that members of the genus Neomenia are among the most gcncrnlized molluscs, related to the Chitons and Cliitonellus, yet belonging a|)ait in a jjliylum which he termed Scolecomorplia, llie first tlivision of Mollusca eiiccphala. Another of the important pajx'rs which appeared during (his year came from the liand of von Oraff, who investigated the anatomy of Neoitiviiid viir'ntidd and i-ccxaniincd Chactodviinii nilitluhun. iM'oin the facts disclosed this auttior was convinced that not only arc tlic two forms constructed upon tiie same plan, but that the estal)lishment of the Ami)hineura as a separate phylum by von Ihering was wholly justified. "Wir erkennen in Neomenia mid ("haeto- derma Modificationen einer sehr alten Urform, vom denen die letztere naher den Wiirmer, die erstere naher den Mollusken ankniipft." In (juite a remarkable way the facts discovered in the study of these two species supported von Iher- ing's contention that the Solenogastres and the Chitons are allied forms, though they ditl not force one to the belief that the Amphineura are necessarily to be placed in the phylum Vermes, and von Ihering himself abandoned this position a short time later. About the same time Hansen (77) published a most important paper on the anatomy of Chaetoderma niiiduhun, and in the treatment of all the systems, especially the urogenital, advanced our knowledge to a considei'al)le extent beyond the old position. For the first time the mode of development of the sex products was determined, and their route traced to the exterior; in other words the broader features of the anatomy of Solenogastres now became com- l)n'liensible. In the opinion of the author, Chaetoderma does not clearly belong lo any definite place in the existing system of classification; that wliilc certain niiilluscan characters appear, others are strongly suggestive of annelid relation- slii])s, so that its exact position is yet in doubt. The following year Gegenbaur (78) in the "Grundrisse" made a few very guarded statements regarding the relationships of Neomenia and Chaetoderma which may, provisionally at least, be designated the Solenogastres and may be regarded as a division of the group Vermes. The ventral groove of Neo- menia "represents the first stage in the formation of that pedal surface of the body which is seen in the lowest Mollusca." On the other hand the nervous system, while decidedly different from that in the worms, nevertheless presents some fundamental resemblances. With the exception of two or three short notices no other papers appeared until ISSl, which in some respects is the most important year in the history of 14 HISTORICAL RKVIEW. the group. At this time JSpengcl describod tlie innervation of the osphradiiun, and in several species of molhiscs discovered deep-seated resemblances in the elements of the nervous system and tlieir arriuigcnu'nt. (_)n the l)asis of this work, which includes the examination of Neomenia, C'haetoderma, and two unidentified species of Solenogastres, antl with the additional helj) alfoi'ded l)y the work of Tullberg, von Graff, and Hansen he emphatically claimed, in ojipo- sition to von Ihering and Gegenbaur, that the Amphineura are true molluscs. Accordingly he established the Amphineura (Chitons and Solenogasti'es) as a class of the MoUusca. A very short time afterward the masterly work of Hulirecht appearetl, and in some respects it continues to be the most important work that has ever been published on the subject. The study was based chiefly upon a gigantic species, ProneoDioiia sluidri, which occurs in Barents Sea north of Scandinavia, and embraced a careful examination of its external and internal anatomy. The results, with a few relatively imimiiortant exceptions, have been confirmed by the study of many other species, and form a most substantial foundation for studies of more i-ecent date. Concerning tlie relationships of Proneomenia, Neomenia, and Chaetoderma the author has no hesitancy in agreeing with Spengel that they constitute one order (Solenogastres) of the class Amphineura, the Chitons belonging to the other (Polyplacophora). In a number of suc- ceeding publications this position is held without modification, and the few additional facts of importance that are presented still further emphasize the correctness of the theory. From this time forth scarcely a year has elapsed withmit one or more papers appearing on the subject of the Solenogastres. Deep-sea researches or work along the shore line beyond the littoral zone have brought to light an ever increasing number of species whose anatomy is now for the most part fairly well known. Without exception all are built upon essentially the same funda- mental plan though in detail each species presents, as is to be expected, some new and interesting modifications. To the majority of zoologists the accumu- lated results point unmistakably to the true molluscan nature of these animals, but a glance through some of the succeeding paragraphs will show that there is far from being a unanimity of opinion regarding their position in the phylum and their relation to other groups. The great mass of anatomical details which have been published during the past twenty-five years, serving chiefly to dis- tinguish genera and species, new and interesting though they may be, can be but briefly considered in a review of this character. HISTORICAL REVIEW. 15 In 1SS2 Kowalcvsky and Clarion called in ciucstion the work of all pre- ceding authors, claiming that they in every case had wrongly oriented the ani- mals, that the anterior end is in reality posterior and I'icc versa. Tullberg's lateral glands (portion of the coelomoducts) are accordingly the salivary glands, the penis with its appendages is clearly the radula, the mouth cavity is the rectum, the "egg bag" (pericardium) is the intestinal coecum above the pharynx, the branchia are the buccal cirri and finally the protrusible pharynx is the com- bined uterus and oviducts. This paper called forth an immediate rejoinder on the part of Hubrecht, who reviewed the work of the authors in cjuestion, and showed that the orienta- tion of the animals in question is correct, and that Kowalevsky and Marion have created confusion worse confounded owing, for one reason at least, to the fact that they had not seen the species under discussion. During the next four or five years Kowalevsky and Marion published, either separately or conjointly, several papers preliminary to their chief work wliich appeared in 1887. In this study the authors describe to a certain extent the habits of five new species of these molluscs collected along the shores of France, and accompany it with a very detailed description of the external and internal anatomy. Some of these last named results are referred to else- where in the jiresent pajier. In the meantime Selenka ('85) inil)lished an account of the gephyrean worms collected by H. M. S. Challenger, and therein l)iiefly described Chaetoderma tnililiirr from the Malay Archipelago, adding the remark that he was unable to give any tlata that might settle its systematic position. In 1888 Hubrecht described a new genus of iSoIenogastres (Dondersia) taken in the vicinity of Naples. It is a fairly close relative of Proneomenia and Neomenia, and the anatomical characters are accordingly not strikingly differ- ent from those presented in the paper on P. sluiteri. In this same year Hansen ('88) made a stud}' of several species of Soleno- gastres long before described by Koren and Daniellssen ('77). His researches chiefly concern Neomenia cariiiata, which is shown more conclusively than be- fore to be similar to P. sluiteri. Chaetoderma nitidulum was found to pass the sex prnducts iiilci the jieiMcardium from whence they pass through ducts into the anal cavity (Hubrecht) or branchial cavity (Hansen). Pruvot ('90) denied the existence of a heart, or pericardimn or dorsal aorta in the Solenogastres. The blood moves in lacunae of which a large one passes dorsally along the mid line propelled by contractions of the body. The paired 16 HISTORICAL REVIEW. gonad, postei-ioily becomes single, and opens into what has l>een termed the pericardium, in i-eaUty an accessory part of the reproductive system; while the dorsal sinus courses in a tube (hanging partially in the so-called pericardium), which is called, l)y other authors, the heart. In DGndcrsia banyulensis sperma- tozoa develop on the external walls of what has l^een termed the heart, while the lateral walls of the pericardium are ciliated and serve to convey the sperms as in many hermaphroditic gastropods. Eggs are temporarily stored in the "poche accessoire" (pericardium), and the kidneys are in reality genital ducts without renal fimction. In a later paper ('90) the same author describes a few very interesting stages in the development of Dotidcrsia haui/itlensif^, and two years afterward adds some further observations regarding the embryology of Proncotuenia agUiopheniac. Ova in the pericardium (of other authors) lack membranes; as these are present in extruded eggs it follows that the suj^posed kidneys are in reality shell glands. The segmentation stages resemble those of scaphopods and pelecypods, and to some extent this similarity is visible in the later develop- ment. A gigantic coat of ciliated cells (a highly developed velum probably) is formed, and within this the embryo forms by a process certainly not primitive or at all events unlike that of other molluscs studied up to the present time. When the velum is thrown off the \iw\i\ resembles to some extent a young Chiton, possessing seven imbricnted calcareous plates along the dorsal surface and later- alh' situated Hatteneil sjiines in what appears to be the girdle. The internal organs at this stage are practically undeveloped, and as the later growth is wholly unknown the present results throw but little light ujion the imjiortant subject of the jihj'logenetic development of the group. Pelseneer ('90) considered (contra Hubrecht and otliers) that Chitonelhis is not a jirimitive form intermediate between the Solenogastres and more typical Chitons, Init on the other hand is highly specialized. These conclusions were based on data supjilied by a study of the branchia, nervous s_ystem, ami shell. In an introduction to the extensive work of Blunnich ("91) Hatschek seconds Von Ihering in his attempt to place the Chitons and the Solenogastres apart from the gastropods, and agrees also with Pelseneer in regard to the position of Chitonelhis. Owing to the studies noted in the preceding paragraphs the broader fea- tures of the anatomy of these animals have been settled beyond dispute, and conse([uently the papers from this time forward serve in large measure to supply details, and to a limited extent indicate the phylogenetic relationships of this HISTORICAL REVIEW. • 17 group of animals. The more special features relating intimately to the various systems are noted to some extent in the main body of the present paper, while the general considerations are discussed on p. 164-173. Among the more exten- sive of these works are those of Pru\'ot ('91), who has described several species from the shores of France; Wircn ('92) whose study of .several species from the Scandinavian coast is among the Ix'st that has ever appeared ; Sinuoth ('93) and Pilsbry ('98) whose systematic works are of the highest value; Thicle whose various papers dui'ing \hv past lifteen years have added materially to our knowl- edge of the anatomy of molluscs, including several species of Solenogastres; Nierstrasz (1902. etc.), who with an abundance of material collected chiefly by the Siboga Expedition, has added extensively to our information of these animals; and finally the present writer wlio has contributed some data relating especially to the nervous system. Since the above was written Nierstrasz has pul)lished an important report (1908) reviewing the work of various investigators since the apjiearance of Sim- roth's paper in 1893. It is a valuable contrilnition, and the scheme of classifica- tion there adopted will be of nuich service. Turning now to the broader features of the classification of these animals we find that practically every investigator in this field of research is agreed with Spengel that among the Mollusca their nearest relatives are the Chitons; Init regarding their more accurate position wdthin the phyhmi differences of opinion appear. \'(in Ihering in 1877 called attention to the fact that the Solenogastres and the Chitons are not distantly related, especially- if we consider Chitonellus to be a connecting link and therefore a primitive animal. According to this line of reasoning the Solenogastres, tlevoid of a ladula (none had been discov(>red at that time) and shell, are the more ancestral and are so closely related to the worms that both they and the Chitons constitute a special phylum (.\mphineura) of Vermes as noted in a previous paragraph. Hubrecht. without laying much stress on the ancestry of the Amphineura, though lie hints at their derivation from a platyhelminth ancestor, was likewise of the opinion that the Soleno- gastres are primitive, and that Chitonellus is a link connecting them with the more highly modified Chitons. It may be mentioned also that Haller in '94 modified some of his previous ideas, having become convinced of the correctness of von Ihering's and Hubrecht's position. The above idea was combated by Pelseneer ('90) who claimed that Chito- nellus is nothing more than a highly modified Chiton and in no direct way related to the modified grou]i of the Solenogastres. Hatschek ('91) agreed with this 18 HISTORICAL REVIEW. theory thougli ho did claim with von Ilierins (a theory abandoned hy him in '90) that tlie Am])hineura are Vermes. (Irobl^en ('94) hkewise considered this the correct vicnv tliougli he beheved the Amphineura to be true molluscs. This notion is implied in the work of Haller ('82), who made the claim that the Chitons and the Soleiiogastres are distinct groups of animals which have been derived from a common vermian ancestor. In a more vigorous fashion Thiele argues from the same standpoint. With one or two exceptions those who argue along the line just indicated regard the Solenogastres as primitive animals, and are accordingly opposed to several investigators who hold a diametrically opposite view. Simroth, Wiren, and Heath believe that the Solenogastres early branched off from some primitive polyplacophore and while retaining several jirimitive features are in other re- spects degraded organisms. Pelseneer and Ciarstang take jiractically the same view. Marion, in a sense, does the same as he compares the adult Solenogastre to the larva of the Chiton. Plate traces the Solenogastres and Chiton lines of descent to some ancestral mollusc which may have given rise also to the present classes. In regard to the derivation of the molluscs, and the Solenogastres especially, from some premolluscan ancestor there are a number of widely divergent theories. In 1877 von Ihering b(>lieved that among the worms the gephyreans ai'e most closely related to the Solenogastres. Haller ('82) on the other hand i-egarded them as more closely allied with the nemerteans. Hubrecht, Thiele, Plate, and a number of other writers consider that the molluscs, or at all events the Soleno- gastres, arose from a turbellarian-like ancestor. This idea has been most fully developed by Thiele. According to him the progenitor of the molluscs and the Solenogastres (which are considered to l)e worms) was in the fundamental characters of its organization similar to the modern cotylean polyclad. The often frilled sensory margin of the body became the mantle, which for jiui'poses of protection, developed a cuticular covering and ultimately a shell, while the ventral sucking disc expanded into the moUuscan foot which in its least modified form occurs in Haliotis and similar species. GENERAL FEATURES, METHODS, ETC. 19 GENERAL FEATURES, METHODS, ETC. External Characters. — The Solenogastres constitute a group of marine animals, which combine with features more primitive than in any other molluscs numerous others indicating a high degree of modification. All ai'e bilaterally symmetrical, worm-like in form, usually nearly round in cross section, and vary in shape from short thick set to very slender greatly elongated types. While the average length is not far from twenty-five millimeters several species such as Notonicnid clavigcra and Kruppomenia minima are from one to four or five millimeters long, and on the other hand Proneomenia sluiteri reaches the giant size of one hundred and forty-eight millimeters. The mouth, or more properly the atrial opening, usually in the form of an elongated slit, holds an antero-ventral position, and is clearly separated from a ventral median fvu-row, in the Neonieniina extending throughout the entire length of the body. This latter structure is generally considered to be a true pedal groove, the small fold included therein and abundantly provided with glands being the foot. In the Chaetodermatina no external trace of these organs exists, but a gap in the ventral musculature, and a thickening of the muscle bands on each side of the mid line, and in Limifossor a definite pedal sinus, indicate that they were present at a former time. In what must be con- sidered a primitive condition the ventral furrow is posteriorly continuous with what has usually been termed the cloacal cavity, which contains the openings of the urinogenital apparatus, anus, and the respiratory organs. As is more fully shown in a succeeding paragraph, the cloaca is in reality a mantle ca\aty, and the two branchiae it contains in the Chaetodermatidae are undoubtedly true ctenidia. On the other hand the folds of the cloacal wall, sometimes ex- cessively developed and highly vasculai-, do ncjt appear to be rudimentary nor degenerate ctenidia. A cuticular sheath, often of great thickness, envelops the body, and con- tains from one to seven or eight layers of calcareous spicules. Where more than one layer is present groups of cells constitute papillae or organs of j^rob- lematical use. In the adult condition all traces of a shell are absent, but in the development of Myzomenia (Dondersia) banyulensis, as determined by Pruvot, a stage occurs in which the embrj'o bears on its dorsal surface seven slightly imbricating, calcareous plates. Internal Anatomy. — In regard to the internal organization there are numerous features that indicate a degraded condition, due probably to parasitic 20 GENERAL FEATURES, METHODS, ETC. hal^its or an adaptation for a life in the bottom ocjzc. In Chaetoderma the ali- mentary tract is a comparatively simple tul)e passing directly through the body, (as with other Solenogastres), ]iro\'itleave. Wiren ("!)2) who kept over one hundretl specimens of C. niliduluvi in captivity says they progress through the ooze by means of the alternate contraction ami expansion of the prothoi'ax aided by movements of the entire body; and that when at rest they ordinarily tlirect the liody vertically with the cloacal chamber widely expanded, and the branchiae fully cx])osed at the opening of tln^ biu'row at the siu'face. When disturlied they disajipear instantly several inches into the nuiil. In most respects these observations answer for ('. cruiVttd which I kcjjf some time in cajitivity. This animal from time to time advanced through the ooze in the manner described liy Wiren but it never appeared at the surface. For hours together it remained ([uietly in its burrow with the gills fully expaiuled and when disturbed retreated but slowly, though the gills were at once retracted and the cloacal chamber closetl. The indivitluals acting in this manner were apparently in a normal condition since the alimentary canal of several kept in cajitiNity for nearly one month contained tjuantities of food. The members of the suboi'dci' Neomeniina, on the other hand, are not known to bun'ow, Init are usually found on some species of hydroid, coral, and exceptionally (Proncomcnia vagans, P. desidei-ata) , on plants. Now and then specimens have been dredged unattached and it may be that they, like Ncor7icnia GENERAL FEATURES, METHODS, ETC. 25 carinnla and Slulonwnid sdlirilori. crawl alioul freely over the sea Ixittoin, tliiiui:;li it is possible also that tliev haxc been loosened IVoni some host. The relation of mollusc and coelenterate has not Ix'cn thorouji;hl_v worked out, but there are many indications and some definite proof that it is a genuine case of parasitism ami not an accidental association or a case of commensalism. In Dn-panomenid nnnpjii-illd from Ihe Hawaiian Islands the jjroboscis of one indi\idual was inserted into the body wall of some s])ecies of E])izoantIius, many of whose reproductive and other cells had been withdrawn so that htn-e therc> is no doulit that this Solengastre is a i)arasite. The ])resence of nematocysts in the alimentary canal of se\'ei'al other species, includinij; most of the sjiecies of Strophomenia, indicates that they likewise are in the same category. Iliibrecht ('80) states that a bit of alcyonarian coi-al was found in the mouth (if /'. sluiliri, but he calls attention also to th(> fact that diatoms and entomostracans occurred in the faeces in the cloacal cavity, and Heusclier ('1)2) recoids the presence of Entomostraca in the gut itself. It thus develops that the diet of sucli species is varied, and it is possible that such forms as P. r<((/(ms and P. ckfiidcrutu which were found crawling about on plants belong to this same class. yo far as known the food of the C'haetodermatidae consists of mic-roscupic organisms and organic I'emains which they scoo]) up while bui-row ing through the ooze. Wireii ('92) believes that the buccal plate (Mundschild) acts as a digging organ, and this may intleed be the case but the exceedingly small amounts of inoi'ganic matei'iaJ, which make their wa\' into the digest i\'e tract, indicate that in addition to functioning in a jiurely mechanical fashion it manifests a decided selectix'c action. \Mien selected the food is cari'ied backwju'd by the great conical tooth in the genus Chaetodei'ma, atui in the foi'm of a more or less spherical bolus, mixed with the secretion of the salivary glands, is carried into the mid gut. In Prochaetodernia and especially Limifossor the I'adula and its supports are of large size and indicate active, predatoiy habits, l>ut the contents of the gut are essentially the same as in Chaetodernui. Color, Size. — In a ])reserved state the skin of the Solenogast res is usually unpigmented, the light yellow or yellowish brow n tint of [Iw animal being due to the cuticle investing the body. In many species, especially of thcChaetoder- matidae, this may be obscured to a greatei' or less degree by the nuiltitudes of refringent sincules imbedded in it or by some of the internal organs. The liver for example is often dark brown, and shining through the translucent body wall and the overlying cuticle and spicules, giv(>s a decided frosted gray tint to several species. The red color of the Ijlood plasma may imj)art a pinkish tinge espe- 26 COMPARATIVE ANATOMY. cially in the head and cloacal regions. An incrustation, rusty red or l)lack, may cover the entire animal though it is usually restricted to the posterior extremity. In a very few species some of the hypodernial cells contain pigment, red, yellowish red, lilac, or yellow in color. Echinojnenia coralliophUa, a species living on CoraUium rnhruin, is provided with movable scales which when de- pressed give the body a whitish tint resembling the tentacles of the liost, and this may possibly be the case with Slrophonicuia spinosa. I^jion I'aising tlie spines the pigmented hypodermis becomes less ol)scuied and the animal assumes a reddish color similar to the coral stalk. The smallest sexually mature Solenogastres are not over 5 mm. long, and on the other hand Proneoinenia sluiteri attains, as ])reviously stated, the great length of 148 mm. The average length is probal)ly not far from 30 mm. Length Index. — In the discrimination of species the so-called length index, or the ratio of length to breadth of body, has lieen used to a c(insiderable extent, but from several ex]iei'im(Mits in the j)reser\alion of fresh material, I am convinced that it is of little use, certainly not with closely I'elated forms. For example nearly sixty ('huctdikrmd inonk'nuinni^i, which had come uji in the same dredge haul, were treated with slow alcohol in juecisely the same fashion and yet the length indices varied fully twenty i)er cent. Some si)ecimens must invariably be subjected to a greater j^ressure than otiiers in the dredge load, and these are more flaccid and less contractile and with them the length index is relatively greater. COMPARATIVE ANATOMY. Foot and Glands. — It is now a generally accepted fact that the ventral furrow and its included fold rfjiresents a greatly reduced pedal furrow and foot. In the C'haetodermatina all extei-nal tiaces of these structures have disappeared completely, but internally a gap in the veiitial nuisculature and a thickening of the longitudinal muscles on each side of the mid ventral line and in Limifossor a well-developed pedal sinus in the head region indicate their foinicr existence. In what appears to lie the least mcidificd species, the foot consists of a single fold, but in several other species this is accompanied on each side by a fokl of almost ecjual height and length, and in the Necmeniidae the creeping surface is often comparatively broad and is developed into se^'eral folds. Whether one or more of these plaits exist each is bounded by a single la^cr of ciliated COMPAHA'ri\"K ANATOMY. 27 epithelium \\hich as a rule is continuous with the lining of the cioafa. The last named space is certainly a true mantle cavity and the plume-like i)ranchia it contains in the Chaetodermatidae are ctenidia. On the other hand it is questionable if the folds developed in the cloacal walls, as in Alexandromenia for example, are homologous organs. As in the Chiton embryo two sets of pedal glands exist, termed by Hubrecht ('80) the anterior and posterior pedal. The first named is a highly developed organ filling the greater part of the space between the body wall and gut in the head region. It is composed of pyriform cells whose ductules lead into the anterior end of the pedal furmw, which is usually develojK'd into a cavity of considerable size, where they make their exit liy separate intei-cellular openings. The postei'ior ])etlal gland is situated alxive and slightly to each sitle of the foot throughout its entire extent. Its cells aic likewise pear shaped and o])en inter- cellularly into the pedal furrow. Hypodermis and Products. — What is sometimes termed the skin consists of two main elements, the hypodermal cell layer and the overlying spiculose cuticle. In the majority of species the hypodermis consists of a single layer of cells, antl exceptionally { Paramenia palifcra. Iclilliijamenia ichthyodes, for exam- ple) forming a more or less irregular many cell layer. Concerning the nature of the elements entering into its formation there are numerous differences judg- ing from the accounts of the various authors, and the functions ascribed to them are equally diverse. The ordinary hypodermal cells, those responsible for the formation of the cuticle, are usually cubical or low columnar in form with round or oval nuclei imbedded in a finely granular cytoplasm pigmented in a few spe- cies (f. g. C. niUdulum) which blends with the overlying cuticle. In the same general situation gland cells are present in several species together with more slender elements which may jierform a sensoiy function. The papillae occur in all Neomeniina in which there is more than one layer of spicules imbedded in the cuticle, (lenerally speaking each consists of a compara,tively slender stalk which is attached to the hypodermis, and on the other hand expands into a more or less globular mass in contact with the free surface of the cuticle or may even project above it. The cells composing the swollen portion are apparently filled in life with a highly spongy, possibly vacuo- lated protoplasm which in preserved material may shrink greatly, producing radiating pseudopodia-like processes. In Halomenia gravida outpouchings of the gut occur at fairl>- regular intervals along the dorsal side of the animal on each side of the mid line. These penetrate the somatic musculature (Plate 32, 28 COMPAUATn'E ANATOMY. fig. 5) and come in contact with tlic under sui'faco of modified papillae, which appear to be capable of a cei'tain amount of protrusion t)wing to the agency of a surrounding blood sinus. The significance of this remarkable state of affairs is very obscure; and for tliat matter the various fimctions such as excretion, touch, and pressure relations which ha\e been asci'ilied to these organs as yet rest upon no direct experimental evidence. That they ari' the liomologue of the aesthetes in the Chiton shell is a reasona])le assumption, but this carries with it no trustworthy evidence regarding their office. The spines form from one to several layers in the cuticle, and present a great variety of forms. In the Chaetodermatina spearhead-types prevail, and in the Neomeniina, where there is but one layer, this shape may likewise occur. In those species with more than one layer the usual type is needle-shape, and with it may be associated radially directed spicules usually with truncated ba,ses. Spicule Development. — In a number of species of the present collection, notably Pronrumvnid lutwuilcnsh, Slrophonicnia scandens, and Hahtmcnia (jravidii, certain of tlie more important details of the formation of the spiculose investment of the body appear with unusual distinctness, and to avoid needk^ss repetition the results are discussed once for all in the following paragraphs. Speaking first of Proneomenia hawaiicnsis, in the earliest stages of the spicule formation, where the calcareous product is no larger than the neighboring hypodermal elements, several cells are seen to \n' taking part. One of somewhat larger size than the others, and with clear finely vacuolated cytoplasni and distinct granular sphei'ical nucleus, rests underneath the base of the s]iine. Its g(>ncral appearance is essentially like that of the cell beneath the spicules of (,'. in'tiilnluin as figured by Wiren or the s]Mcule fornung cells in the mantle of c(>rtain sjiecies of Chitons, and is par excellence the lime secreting element. Wiren is of the opinion that the basal cell is a modified wandering cell that has left the blood stream and migrated to the hyixidernfis. In all of the Soleno- gastres under discussion the wandering cells are of a granular character with no distinct cell membi'ane anil cl(\irly different from any of the hyiiotlci'nial cells. Furthermore in the species undi'r discus-idu I have never seen these jilasma cells outside of the somatic muscle layer, anil there are never any indications that the spicule fornung elements are tlerivetl from any other soui'ce than the hypodermis. In very early stages, perhaps from the first, the spicule is surrounded by a delicate cuticular sheath whose reaction to the ordinary stains indicates a com- position unlike the material in which the spicules are imbedded. This spicule ro^rPAIJATIVK ANATOMY. 29 sheath is probably fonned l)y seven or eight eells, slender in form, indistinet in outlin(\ with deiis(> nuclei and attenuated bases which are imbedded in the hypodermis proper. They entirely suiMound the basal cell, and distally their membranes become continuous with the s])icul(> sheath, which as Plate 36, liij;. 5, shows, is thus interrupted a short distance above the base of the s])ine. At a relatively earlj' stage in tln^ development of the spine a minute cell (Plate 36, fig. 11) may be detected Ixtween the l)asal cell proper and the spicule sheath. At first it appears to be connected with the deeper jiortions of the hypodermis by a single stalk that passes to one side of the basal cell; but in later stages such a connection disappears antl the cell in ciuestion l)econies closely applied to the base of the sjiine. It appears to be responsible for the formation of the spicule sheath immediately above it and in the following way for the cavity of the spicule itself. At first the cuticle above this small basal sheath cell is of uniform thickness and the lime salts, deposited presumably by the basal cell proper completely fill the spicule sheath, but very soon a minute knob-like elevation appears on the basal part of the sheath, ami, pei'haps owing to this increased thickness of the cuticle, it interferes with the deposition of calcareous material, for from this time on a cavity develops in the spine that in size and position corresponds to the cuticular knob. As the latter increases in size the spicule cavity enlarges, and when in later stages there is a decrease the cavitj^ becomes proportionately narrowed until both finally disappear together. In the earliest stages of its existence the long axis of the spine is at right angles to the hypodermis, but as development progresses it becomes more and more inclined until it reaches the final horizontal position. This rotation is probably due in large measure to the uneciual elongation of the attached cells, while the continual advance of the free tip of the spine through the surrounding cuticle is due to the addition of new material basally. During this whole forma- tive period and after its completinn the entire spicule is migrating also toward the free surface of the body. Hubrecht and several other subsequent observers consider that this movement is caused l)y the continual cuticular current, so to speak, brought about by the peri)etual addition of new material in contact with the hypodermis, a belief with which I concur for on any other supposition it would be difficult to account for the perfectly vertical, unbent position of the slender renuiant of the matrix cells in late stages. Until the spine has been carried outward for a distance eciual to one fourth or one thii-d the thickness of the cuticle the matrix cells retain their usual con- nections and generally are fairly distinct though showing more or less shrinkage 30 COMPARATIVE ANATOMY. (see Plate 36). This latter feature becomes strongly marked beyond this stage and as the cells shift to a distinctly subterminal position their boundaries dis- appear, and in the neighborhood of the spine they become dense and fibrous. Still later they migrate to a point fully one eighth the length of the spicule from a terminal position, their attachment to the spicule sheath becomes reduced to a minute knob-like disc and all l>ut the basal portions basal cell ceases to be functional as a spicule forming agent it may transmit iminilscs to the central nei'Xdus system as the spines and scales in the mantle of the ( 'hitons are sup- ]:)Osed to do. In Ilalomenia, Lophomenia. Dorymenia anil all of the species of Stroph- omenia described in the present jxiper this type of develo]mient ]:)revails. In some species the matrix cells become detached from the fully formed sjiine, but otlierwise no fundamental differences exist. Regarding the species described by other authors, Hubrecht and Heuscher agree that in Proneomenia sluiteri a cup of several cells clasps the base of the spicule, and Kowalevsky and Marion and to a certain extent Pruvot, and Wiren have made similar observations on other species. Through the generosity of Professor Hulirecht I have been able to examine a portion of the type of P. sluiteri and though the cells are not so clearly defined as in P. hawaiiensis there is no doubt that in both the spines follow the same course of development. Nierstrasz writes of Cyclomenia halo- COMPARATIVE ANATOMY. 31 scricca "Tlio s])icula remain in conticclidii with the liypodermis In' tliin threads, met with everywhere in the spicuhi" (cnticnla?) and of P. discoveryi he states that ''The spicules are formed in small accumulations in th(> epidermis" though one cell only is said to be active. Thiele has consistently argued in favor of one matrix cell and it is probable that he has focussed his attention on the type of spine I am about to describe. As noted in a preceding paragraph, there are in addition to the tangentially placed spicules in some species others that from the beginning to the close of their development are dii-ected i-adially. This latter type of spine, so far as I have determined, has a mode of development comijletely different fi-om the one just described. In SlroplmDienin tn'miguhiri.^ for example as soon as it becomes clearly recognizable it rests upon wliat appears to be a single cell, and as long as it remains in the cuticle no additional elements put in an ajjpearance. In the later stages the matrix cell usually becomes more cup-like closely clasping the base or side of the spine delate 36, fig. IS), and it may elongate to form a slender stalk but it is ;ilways luiicellular. It is to be noted, however, that in this species, and perha]:)s in others, there are additional, radially directed spines of nuich larger size whicli appear to be formed by more than one cell though this is not certain owing to the fact that the base of each spicule is crowded against the somatic musculature. In the Chaetodcrmatidae Wiren has shown that but one formative cell exists though in its early stages the spine is surrounded by three hypodermal cells wliich may exercise a moulding influence. From my studies I doubt this last statemiMit In both Chaetoderma and Limifossor the formative cell is surrounded by hypodermal elements but there is no evidence that in the develop- ment of the spine they take any active part. It thus becomes evident that there are various tyjies of spicule formation among the Solenogastres just as there are among the C'hitons, but it is a most interesting and significant fact that the most comuion tyjie of Chiton-spine development (according to Plate, '01, Theil C, p. 372) is almost the precise counterpart of what exists in P. Iitiwdiiensis and several other species. Wiren is inclined to the belief that after the spicule forming cells have performed their function they become transformed into the hypodermal papillae. There is confessedly no definite evidence to substantiate such a theory, and on the other hand there are one or two facts that tend to discredit it. In the first place there is no definite relation between the numlier of papillae and spines; in the majority of species the latter considerably outnumber the former, notal^ly 32 COMPAIIATIVK ANATOMY. ill Alexandromenia. Ami ii^aiii wlioro tlie siiiciilc retains its connection with the original formative cells such a fate is out of the question. And finally a few authors, notably Heuscher, have obser\('d the origin of these (jrgans directly from the hypodermis. I have seen many timers papillae-like elevations (Plate 36, fig. 18) such as Heuscher figures, and I am strongly of tlie opinion that they have no connection with the matrix cells. Digestive Tract. — TIk^ anterior division of the alimentary canal consists of a cavity whose walls are provided with twD folds (Mundleisten), usuafly of a horseshoe-sha]ie that tlehne the cirrose area wiiere the wall is modified into numerous finger-shaped filaments. In the greatiM- number of species this and the succeeding jiortions of the gut are intimately united, but in Ehop. aghiD- pheuiac, Dondersia, and a few other species they ai'c distinctly separated by a ridge covered with the spiculose cuticle investing the body generally. Thiele considers that this sensory atrium is an ectod(>rmic invagination of the integu- ment corresponding ]iossibly to the ('hitdu snout. Accordingly \\here the sejiaration is com]ilete the true mouth is jiosterioi' to the opening of tlie atrial cavity, and where the latter is fused with th(> gut the mouth is situated behind tli(^ internal buccal I'idge. In the ( "haetod(>i'matidae this anterior division is absent or is represented, as Thiele assumes, by the Iniccal sensory ]ilate and possibly the semicircular groove (halbmondforniigc^ (Irube). The limits of th(> jiharynx are difhcult to dehne but it is generally assumed that it contains the radula and the outlets of the salivary glands. As a rule its epithelial lining is unmoditied liumgh it ma>' be greatly folded and, in some s]ie- cies form paiiillac which afford an outlet for the doi-sal sa,li\'arv glands. If an oesojihagus exist it is usually not clearly diffei'entiated from the ]ihai'ynx, ami in the following jiages I lia\'e disregarded it. In the Neomeniina tlie (ligesti\'e gland or "liver" is not differentiated from the stomach or intestine. 'I'his l;ist named organ may jiossibly be represented liy the ciliated tract attached to the under surface of the gonad. In the C'haeto- dermatina the digestive gland, stomach, and intestine are clearly defined and are not essentially different from what exists in other molluscs. The cloacal chamber is probably an ectodermie invagination ami a true mantle cax'ity. There is no evidence that it is an expanded rectum. Muscui-AU System. — This system has been described in a V(M'y few sjjecies though most authors I'efer to the more ob\'ious features especially the nature of the body wall. This consists';Kif an outer circular layer, resting in some spe- cies u]ion a hn'er of diagonal fibrils that in turn is in contact with a system of COMPAHATIVK ANATOMY. 33 longitiidinul luiiuUcs. In (ho anterior and i)()storior regions these are subject to various modifications, forming numtli and cloacal sphincters and (Uiators, and afTorcHng attachment for the gill retractors or anteriorly for the nudtitudinous muscles operating the forward section of the digestive tract. These last named muscles have never been fully studied in any species though they are fairly well known in Chacioderma nilidulum anil Limifossor talpoideus. As the matter rests at present there is a similarity between the various species with respect to the somatic musculature and to certain of the more cons]iicuous bundles elsewhere in the body, but beyond this our data are too insufficient to permit of close comparisons. Phimauy Body Cavity and Septa. — The space between the alimentary canal, gonad, and body wall, the jiiimary body cavity, reaches a varying height of development according to the species. In the Chaetodermatina it constitutes a comparatively limited pseudovascular system and in some of the Neomeniina it is likewise much reduced, but in several species it becomes much more extensive. This haemocele is divided by a horizontal septum that extends between the longitudinal fibres on each side of the body beneath the gut and so forms a ventral blood siiuis. In tlie Neomeniina it is small, but is bounded, in every species descrilied in the present ]iaper, by connective-tissue fibi-es per- forated here and thei'e to permit of comnumication with the oxcilving blood spaces. In C'haetoderma a vertical sejituni separates heart, gills, and i)allial complex from the remainder of the primary body cavity. It is not present in the Neomeniidae nor in Limifossor. In the last named genus there is an addi- tional partition, essentially the same as in the Chitons, which separates the head cavity from the succeeding portions of the haemocele. It is perforated by the aorta, alimentary canal, and a pedal sinus which passes forward into close proximity to the mouth. Circulatory and Respiratory Systems. — The heart is develo])ed as a fold of the doi'sal ])('ricardial wall reinforced by a varying numl)er of nuiscle fibres, and in some species differentiated into an auricle and ventricle. From its anterior end the aorta arises and passing dorsal to the gonad leads into a more or less definite head cavity. In the Chaetodermatina and a vei-y few Neomeniina this vessel possesses definite walls; in the others it is part of the general lacunar network. In Limifossor a distinct connective-tissue septum bounds the head cavity posteriorly, but in the greater number of other species the blood spaces in the head region communicatoJi^'ectly with those surrounding the mid gut. These last named sinuses are in coiTTlnunication also with a ventral 34 COMPARATIVE ANATOMY. modiaii pedal sinus oven thoujih (he loot bo abspiit. In the posterior end of the body the blood is collected in a branchial sinus, if gills be present, or is con- veyed to the posterior end of the heart by means of clearly defined channels in the neighborhood of the rectum. In the Neomeniina the haemoglobin is contained in the corpuscles, and in the Chaetodermatina liy the plasma. In the Chaetodermatina two plume like respiratory organs, which spring from the anterior wall of the cloacal chamber, are in all essential respects like those of the Chitons and are doubtless true ctenidia, the space wherein they are contained being the mantle cavity. In the Neomeniina such organs are absent, though the cloacal wall may be developed into folds in some species of large size, penetrated by numerous sinuses and covered with a i-ichly ciliated ejiithe- liiim. It is believed by some investigators that such lamellae are incipient or degenerate ctenidia but there is little to support such a tiieory. Apart from these organs it is probable that respiration takes place over the surface of the body, especially along the ventral furrow if such be present. It has been suggested also that the great buccal folds may possess a combined respiratory and sensory function. Nervous System. — In the Solenogastres there is a more pronounced con- centration of the nerve cells to form definite ganglionic enlargements than in the Chitons, anil the nerves supjilying the mantle-cavity complex arise from a more restricted section, l)ut in the arrangement of the brain, ]iedal and lateral ganglia, and the labio-buccal system there is a very distinct fundamental resem- blance between tlie two orders. In every case the lirain, usually if not always bilobed, is situated on the dorsal side of the alimentary canal about the inter- section of the mouth and pharynx. From its anterior face three pairs of nerves originate in the Neomeniina and innervate the l^uccal wall and the adjacent regions of the body. About the bases of the atrial cii-ri these fibres connect with accumulations of nerve cells which may l)e the homologue of the great ganglionic masses in contact with the brain in the Chaetodermatina. In this last named group these accessory nerve masses, ten in numbei- in Limifossoi-, are connected with the lii'ain by several nerves, and on the other hantl give rise to fibres which innervate the sensory buccal plate (Mundschild). In what appears to be the most primitive condition three pairs of connectives, the pedal, lateral, and labio-buccal, take their separate origin from the brain as in the Chitons. Such a state of affairs is the rule in the Neomeniina. In Chaetoderma erudita these cords imite innnediately before plunging into the accessory gangli- onic masses attaclied to the l)i'ain l)ut retain a delicate connective-tissue sheath. COMPAKATIVK ANATO:\lY. 35 and hence a perfect individuality, until they pass into the lirain. In some other species of the genus, as for example Chaetoderma nitidulum, they fuse indistin- guishably as they enter the brain. Finally in Limifossor the pedal and lateral cords fuse a comparatively long distance from the brain, and at a less remote position are completely united with the labio-buccal connective. In every case the pedal ganglia are almost as long as the animal and hold a ventral position on each side of the pedal furrow though not always in close proximity to it. In several species commissures at fairly regular intervals have been seen uniting these cords, and equally numerous connectives have been traced to the lateral nerves whicli hold a more dorsal position along the sides of the body. In the posterior entl of the animal the relation of these elements varies considerably in the different species. In Proncomenia haumiiensis, for example, the pedal cords become reduced in calibre, and finally break up into small nerves which have not been shown to have any connection with the lateral cords though they come into the neighborhood of some of the small nerves originating from them. In Strophomenia scandens the last two or three latero-pedal connectives are of relatively large size and the union of the lateral and pedal ganglia is clearly established. Furthermore in several species, such as Lcpidomenia hystrix, and Neomcnia carinata, the posterior ends of the pedal cords terminate in ganglionic enlargements (ganglion posterius inferius, Wiren) united by a commissure of more than usual size situated ventral to the rectum. The posterior ends of the lateral ganglia are also frequently enlarged (ganglion posterius superius) and are invariably united above the rectum, thus completing in several species a circum- rectal nerve ring. In the Chaetodermatidae the lateral and pedal ganglia are united, at least in the anterior end of the l^ody, by commissures and connectives but more posteriorly those main ganglionic cords terminate in a large nerve mass, the so-called gill ganglion, continuous across the mid line above the rectum. In what probably represents a typical condition the sublingual or l)uccal system, in reality the labio-buccal, holds essentially the same relations as in the Chitons. Connectives lead from the brain along the pharyngeal wall and unite with ganglia about opposite the forward end of the radula or in the neighborhood of the openings of the ventral salivary glands. These nerve masses probably represent the labial and buccal ganglia of other molluscs, and in at least one species, Strophomenia scandens, are united by three commissures and the sub- radular system. One of these commissures, the dorsal buccal, crosses the dorsal side of the pharynx, while two pass ventral to it. In Proneomenia hawaiiensis, where the subradular system is most highly developed, a connective arises from 36 COMPARATIVE ANATOMY. the inner face of each Uxbio-buceal gangUon and unites with a small subradular ganglion which is in close contact with a subradular organ. These two ganglia are in turn united by a subradular commissure. In the genus Chaetoderma also there is a well-defined subradular system (p. 57). In several species of Solenogastres various authors have foinid what corre- s]5onds to tiic labio-buccal connectives and ganglia, and in most of these cases have found one connnissure which is either the ventral buccal or labial. In the Neomeniidae the subradular organ is usually wanting together with the customary nerve sup]ily; and in most species it is impossible to find more than one buccal commissure. However these nerves are usually very small and difficult to trace so that negative evidence in this case may not be entirely trust- worthy. Sense organs. — In the majority of Solenogasti-es a dorsal sense organ exists in the mid line in the cloacal region. In the Neomeniina it consists of a circular depression, naked or covered with a thin cuticular layer, and surrounded by spicules which in its contracted condition overarch it. When expanded by an underlying blood sinus a disc-like projection is elevated from the bottom of the depression and is raised above the surrounding spines. In the Chaetodermatina a groove, likewise overarched by spi.cules when contracted, is probably a homol- ogous structure. In both families these organs are innervated by nerves from th(^ dorsal commissure uniting the lateral nerve cords. Various functions have been assigned, but without any experimental evitlence. Thiele believes that the ventral fiu-row may be tactile but no sense cells have been shown to exist. In the gills on the other hand stiff hairs have been found among the cilia and are considered to be parts of sensory elements. Osphradia are unknown. In the Chaetodermatitlae the anterior sensory i)late is innervated by a heavy set of nerves and probably acts as a tactile organ. Owing to the heavy cuticular covering it may act also as a digging organ, operating in a general way like a hog's snout. In Proneinuoiid lumHuiinisis a low sensory ridge encircles the atrial wall innnediately within its outer opening. It is coni])osed of high colunuiar cells which rest throughout their entire extent upon a rod-like mass of nerve cells. Internal to this are the usual atrial ridges (Mundleisten), of which the more external closely parallels the sensory tract just mentioned, while the inner one passes nearly around the canal at the commencement of what is probably- the mouth cavity. Both are usually well-developed ciliated folds, capable, in some COMPARATIVE ANATOMY. 37 species, of great clisti'iitioii owing to tlie laige blocxl sinuses contained witliin. Between the ridges the hning of the atrium is developed into simple or branched, finger-like processes termed cirri. Each is composed of glandular and sensory cells frequently pigmented, and is penetrated In* a slender canal travciscd hy a nerve fibre. Regarding the function of these organs and the ridges it is vari- ously considered that they are gustatory, olfactory or tactile, or even respiratory. The cirri may be jjrotnided from the oi)ening, and this fact in connection with their glandular character has suggested that they may serve also to collect food; but judging from Drepanomcnia rninpiirrUa (page 79) there are times when they are inoperative in this respect. Regarding the papillae in the hypodermis of the Neomeniina various hypotheses have been suggested. They appear to be connected with nerve fibres, a!id may reasonably be considered the homologue of the aesthetes in the sliell of the Chitons, l)uf up to the ])resent time there is no proof that such is the case, nor that they are tactile, or secretory, or excretory, organs as some authoi's ha\'e maintained. A subradular organ, normally located and innervated, is known to exist in a few species (jiage 8!j). In its finer details it bears a striking resemblance to its homologue in the Chitons and perhaps functions in the same way. CoELOM. — The secondary body cavity comprises the gonad, pericanlium, and gonoducts wliose relations have been determined in most of the known species of Solenogastres although their physiological significance remains very incomplete. The gonad, usually paired, is situated along the dorsal side of the animal between the body wall and alimentary canal. Posteriorly i( is continu- ous, by mean> of two ducts, with the pericardium which in turn is in comnuinica- tion with the cloacal cavity by means of two canals, the coelomoducts. So far as known all the Chaetodermatina are dioecious while the Neomeniina are hermaphroditic, and generally speaking the gonad is of the same length as the liver and therefore nearly as long as the body in members of the first named suborder. In the young the reproductive gland is paired, and in the adults with the exception of the genus Chaetoderma this contlitioii of affairs persists, tliough sometimes partially obscured l)y the development of numerous germinal folds. In the Neomeniina the sperms arise along the outer walls of the gland while the ova, in some species surrounded by a follicle, are more inwardly ])lacetl. In the mature state the ova and spermatozoa make their way through the short, ciliated canals terminating the gonad posteriorly, and enter the pericardium from whence they pass to the outside through coelomoducts of various degrees 38 C'OMl'AliATlVE ANATOMY. of complexity. In thfe young of some of the Neomeniiiia these last named canals are simple tulies of about e(iual calibre throughout, but in later life they become modified into a shell gland ami one or more seminal receptacles. In adult Chaetodermatina these canals remain comparatively simj^le, and there is some evidence that they function as excretory organs as well as genital ducts. In every case the coelomoducts originate from the hintler wall of the peri- cardium as comparatively small, ciliated tubes which pass anteriorly to about the level of the forward wall of the j^ericardium where they make a shai'p licnd and join the so-called shell gland in the Neomeniina. At the intersection of these two divisions from one to twenty-five vesicular appendages are usually attached, which have usually been considered seminal receptacles though the arrangement of the sperms in a few species indicate that for at least a part of the breeding season they may function as seminal vesicles. As the distal or ventral portions of the coelomoducts usually unite before entering the cloaca the shell gland which they form is a bicornuate, comparatively swollen structure with excessively thick walls and contracted lumen. It is reasonably certain that this organ functions as a nidamental gland, forming possibly an all)iunenous envelope before the egg passes to the exterior. In intimate connection with the genital apparatus just described there are a number of problematical organs which in some cases at least appear to have some role to play in the reproductive process. These include the genital spicula or penes such as occui- in Neomenia carinata, Donjmenia acuminata, Parnrhopalia pruvoti, Stijlomcnia salvalori, and several other species. These are either rela- tively large calcareous jiairetl spines, which ordinarily are concealed in sheaths formed as anteri absence of any other well-defined organs in or near tlie l)uccal cavit>' it is nol improbable that the plate acts as an olfactory or more than usually delicate tactile organ. These same activities or possibly the sen.se of taste have been assigned to the frontal sense organ noted in the preceding paragraph. The dorso-terminal groove in the ( "haetodermatidae and its homologue in the Neomeuiidae is usually consiilered to be an organ of special .sense, Heu.scher alone alleging the contrary owing to the fact that the depression in Proncomenia sluiteri was filled witli detritus. This condition is not infrequently encountered in animals which have Ijeen excavated from the material in a dredge, but it is certainly not a normal state of affairs. Concerning its function we have abso- lutely no positive evidence. It is reported by Pruvot that it may hold the same office as the frontal sense organ though the belief ajjpears to rest upon nothing more tangible than a certain similarity of structure. CLASSIFICATION. While the modification, by Nierstrasz, of Simroth's scheme of classification doubtless fails of necessity to indicate accurately the phylogcnetic relationships of the Solenogastres it has the virtue of being more convenient than any other now in use and hence has been adopted in large measure. The family name Para- meniidae must be discontinued. Cockerell ('03) has shown that Pruvot's genus Paramenia is preoccupied and has proposed the name Pruvot ina, hence in the following table I have used a new family name Pruvotiniidae. The family 42 CLASSIFICATION. Lepidomeniidae Nstr. oontainiiiji Dondersia must retain Simroth's name Dondersiidae. As has been noted by other authors the genus Proneomenia as ci'eated by Hubrec'ht, has been much enlarged to inchide a number of species some of which probably belong to other genera. Proneomenia loc'neri appears to be very closely related to Dorymenia acuta and may in reality belong to Dorymenia. In the genus Strophomenia the long pharynx, the papillae anil the numeious seminal recei)tacles are so characteristic and similar in the species of the present collection that I have no hesitancy in placing all in the same genus. Pruvot's material was jioorly preserved, as he states, and untlcr such circumstances the peculiar condition of the ventral salivary glands is readily explained as I know from experience. The radula sac is likewise incorrectly placed, being much too far forward. Evidently in Pruvot's species all traces of this organ have dis- appeared. With this revision the species are quite similar. It is probable that Rhopalomenia indica Nierstrasz belongs to this genus. Order Aplacophora v. Ihering. Suborder. I. Chaetodermatina Siniroth. Spiculose integument continuous all aroinul the body. Chaetodermatidae, p. 42. Suborder II. Neomeniina Sinn'otli. Spiculnse integument interrui)ted licncath by a longitudinal ventral furrow. Neomeniidae, p. 44. Proneomeniidae, p. 45. Piuvotiniidae, p. 47. Dondersiidae, p. 48. CHAETODERMATIDAE Simuoth. Opening of mouth and anal chamber terminal. Body with more or less sharply marked regions. Ventral furrow antl fold lacking. Two highly devel- ojied plume-like gills. Radula tlistichous, polyserial or strongly reduced to a large unpaired cuticular tooth. The mid-gut possesses, in most cases, a well- developed digestive gland. Copulatory apparatus lacking. Coelomoducts remain separate. Cuticle thin, spicules flat, often needle-form, but solid. Inhabit ^bottom ooze. CLASSIFICATION. 43 Chaetoderma Loven, 1S45. Body vermiform, witliout ventral groove; mouth and anal chamber ter- minal. Two gWh. Sexes separate. Radula reduced to conical peg. Type of genus C. nitididitni. C. aryentea, s]). no v. Length 24 nun. Ijy 1.(1 antl 2.ti mm. through the metathorax and preal)domen respectively. Silvery white. Spines, in side view, usually bent. Alaska. (p. 02.) C. attcnudld, s]). nov. Body long and slcndci-, measuring 61 mm. in length by 1.5 through the metathorax and 2.7 through the ]ireabdomen. Buccal jilate relatively small. Alaska, (p. 55.) C. californica, sp. nov. Body measuring 2-4 mm. in length by 1.6 mm. greatest diameter. Larger spines with exi)anded bases. California, (p. 64.) C. crudita, sp. nov' . Average length, in preserved state, 27 mm. by 2.5 nnn. average thickness of preabdomen. Buccal plate deeply cleft by mouth opening. Brain of large size. Alaska, (p. 59.) C. haicaiicnsis, sj). nov. Body slender, measuring 12-19 mm. in length by 0.5-0.6 mm. greatest thickness. Buccal plate broadly elliptical, perforated in centre by mouth. Hawaii, (p. 49.) C. japonica, sp. nov. Length 17 nmi. by 1.5 mm. greatest thickness, buccal plate shield-shape, perforated by mouth opening. Japan, (p. 67.) C. montereyensis, sp. nov. Length 45 mm. by 3 mm. greatest diameter. Buccal plate unusually large. Tooth stout. Monterey Bay, California, (p. 61.) C. nanuln, .sp. nov. Body small, comparatively stout, measuring 9 mm. in length by 1.4 mm. greatest diameter. Spines keeled and of heavy appearance. California, (p. 66.) C. robusta, sp. nov. Body heavy, measuring 60 mm. long by 3.5 and 4.7 nun. through the meta- 44 CLASSIFICATION. thoriix and preabdomen n's])ectivoly. Buccal plate shiold-shape. Largest spines relatively .slender without tlefinite keel. Alaska. (]i. (iS.) C. scabrd, sp. uov. Small, measuring 12 mm. in length by 2 mm. greatest diameter. Body wall relatively thin. Brownish, becoming olive-green in region of digestive gland. Monterey Bay, California (p. (j3.j Limifossor Heath, llt(l4. Body short. Radula very large, distichf)us, with twenty-eight transverse rows in L. tdlpoidcus (about the same mnnb<'r in L. fntluhi). Doi'sal salivary glands present. Stomach antl digestive gland well develojied, and distinct from intestine. Type of genus, L. talpoidcus. L. fratula, sp. no v. Slaty gray with yellowish cast. Spines from middle of body Cf) nnn. long. Length index 1.3-4.7. California, (p. 72.) L. tdlpoulvufi. Slaty gray in color. Spines, from all i)arts of the body, 0.02-0.38 mm. in length. Length index l-li. Alaska, (p. 09.) NEOMENIIDAE Simhoth. Body compressed, more or less crescent-shaped, without distinct divisions. Index 7 at most. Opening of atrium ventral, of the anal chamber ventral ( r terminal. Ventral furrow present, usually with several folds. Cuticle some- times comparatively thick, spines mostly needle-like, flat, grooved, or hollow. A circlet of gills in the anal chamber. Radula and salivary glands usually lack- ing. Epidermal i)a]iillae, cf simjile structure, usually ])resenl. Fore gut i)r(i- trusible. Coelomoducts separate or united to form shell gland or copulatory organ. Digestive gland lacking. Penial spines usually present. Free, creeji- ing abont over bottom. Drepanomenia, ixrn. nov. Body short and thick. Hollow needle-like si)in( s with truncated bases; .slender stalked pajjillae. Wntral salivary glands long and tubular. Coelomo- ducts simple, without appendages. No co]nilatory apparatus. Type of genus D. vamprjrclla , s]i. nov. With characters of the genus. Hawaii, (p. 77.) CLASSIFHATION. 45 Pachymenia, gon. nov. Body stout, measuring 27 l)y 4.") mm. Ouo layer of awl-like spines, papillae nuilticeilulai' with broad bas(>s, ])osteriiirly ill defined. Pharynx very large and muscular with numerous glands some of which open by one pair of duets at forward border of j^harynx. Dorsal and xciitral limbs of eoelomoduets providetl with numerous glands. One ])air of small seminal receptacles. Cloacal wall covered with glands e.xcept region of branchial folds. No eopulatory spines. Type of genus P. cibijsftorum, sp. nov. With characters of the genus. California, (p. 72.) PRONEOMENIIDAE Simroth. W(jrm-like. Radula distichous or polystichous, sometimes lacking. Sali- vary glands tuljular, lobed or lacking. Cuticle thick, spicules mostly needle- like in several layers. Epidermal papillae present. Gills usually lacking. Coelomodui'ts usually united into shell gland, sometimes separated. Copulatory apparatus may be present. Free living, partly or entirely parasitic. Proneomenia Hubrecht, ISSO. Body elongated, vermiform, the length 9-50 times the diameter. Cloaca opening ventral. Foot present, the groove passing into the cloaca. Cuticle thick with crowded spicules. No gills. Radula multidentate. Two salivary glands. Copulatory organs present or absent. Type of genus, P. sluiteri. P. lunraiieiisis, sp. nov. Length 36 by 2 mm. Doi-sal and ventral salivary glands. Radula with 3S-40 teeth in each transverse row. One pair of seminal receptacles. No copulatory organs. Hawaii, (p. 82.) P. insu'aris. sp. nov. Anterior end similar to the foregoing species (posterior end missing). One ])air long, tubular salivary glands. Radula polystichous with 28 transver.se rows. Hawaii, (p. 90.) Driomenia, yen. nov. Measurement 9 by 1 mm. Cuticle thick, papillae present, spines needle- shaped, slightly curved. Atrium separate fiom mouth opening, no radula, one pair globular ventral salivary glands. Antero-lateral pericardial wall prolonged 46 CLASSIKICATION. into a pair of horn-like pouclies. One pair seminal receptacles. No co])ulatory spines. Gills absent. Type of genus D. pacijica, sp. nov. With characters of the genus. Japan, (p. 93.) Dorymenia, gen. nov. Vermiform, body terminating posterioi'ly in a finger-shaped elongation. Radula polystichous, with 48-51 longitudinal rows of 22 teeth each. One pair of seminal receptacles. A pair of long copulation spicules closely associated with a pair of globular coeca likewise opening separate!}' into the cloaca. Type of genus D. acuta, sp. nov. With the characters of the genus. California, (p. 95.) Strophomenia Pruvot, 1899. Body elongated, cylindrical, tlie thick cuticle penetrated by acicular spicules and closely crowded vesicular ]iapillae; radula and salivary ducts, caudal sense organ sometimes absent (?); two distinct genital openings into the cloaca. Type of genus, S. lacazei. S. farcimen, sp. nov. Length 18 by 2 mm. Papillae smal', closely crowded. Radula with 15 transverse rows of 24-28 teeth. 19 seminal receptacles. Japan, (p. 119.) S. ophidiana, sp. nov. Length 43 by 2.5 mm. Pajiillae fairly numerous. No radula. 23 small seminal receptacles. Japan, (j). 112.) S. reguhiris, sji. nov. Anterior end missing. Papillae small, closely crowded. 12 seminal recep- tacles. Japan, (p. 116.) .S'. scandcns, sp. nov. Length 32-39 by 1.6-2.1 nun. diamter. Papillae crowded. No radula. 15-18 seminal receptacles. Hawaii, (p. 106.) S. spinosa, sp. nov. Length 28 by 1 mm. Papillae few. Radula small, monoserial or biserial with 8 transverse rows. 12-31 seminal receptacles. Jajian. (p. 122.) S. triangularis, sp. nov. Length 12 mm. by 1.6 mm. diameter. Triangular in cross section. Radula CLASSIFICATION. 47 distichous, apparently cDiuh-likc witli Hi cusps in each row. Seminal receptacles 10-12. Japan, (p. 125.) PRUVOTINIIDAE, iiiiin. iiov. Worm-like. Cuticle, as a rule, thick; spines as in I'roncomeniidae or hook-like. Epidermal papillae present or absent. Radula distichous, simple or double comli-form, or l;ickin]iarate. C'opulatory apparatus may be j^resent. Ventral fold and fvuidw may be absent. Living free, or on corals, etc. Herpomenia, fjcn. nov. Length 11-lS mm. by ().()-(). 9 mm. Foot smoothed out, ciliated. Ventral salivary glands very large, encircling the thick walled \'ery muscnlai' jihai'ynx. Radula lacking. 1 jiair seminal rece]itacles. Shell gland almost globular. Type of genus H. platjipoda, ^Y>. ncjv. With the characters of the genus. Aleutian Ids., Alaska. (]). 151.) Dondersia Hiji)hk<'iit, 18SS. Body long, length index 10-4S. 1 ventral fold. S]iicula, needle- or spatula- shaped, flat. Cuticle very thin, no pai)illae. Dorso-terminal sense knots present. 1 ]iair of short ventral s;di\'ary glands, which unite before opening out into the pharynx. Radula small, nionoserial or biserial. Mouth ca\'ity and pharynx opening separated frcmi each other. No copulation spicula. Type of genus, D.fesliva. D. califiiniivd, sp. nov. Dorsal salivary glands very scant; ventral small. Radula with at least 12 i)airs of teeth. Inunature. Califoi'nia. (p. 155.) Ichthyomenia I'li.siun, Isiis. Body cyliiulric-couic, broader behind, narrowed in front. Cloaca opening a terminal transverse slit, a jiromincMice in front of it. IuhA groove and foot present, dis;i])])earing posteriorly. Cuticle not jtapillose, the ventral spicules Ieaf-sha]ied, the rest scale-like, imltricating. No gills. Radula well developed or i-uilimenta,ry, with appai'cntly two rows of teeth. Length 5 to 13 times the breadth. Tyjie of genus, /. icltlhyodes. I. iMronii, s]i. nov. Body pale yellowish white, Ki mm. long by 1.2 nnu. thick. Scales fish-like and club-shaped. U])wards of 50 sensoiy pits in the anterior end. No radula. California coast, (p. 159.) CHAKTODKRArA HAWAII FASTS. 49 DESCRIPTION OF SPECIES. Chaetoderma hawaiiensis, sp. nov. Two specimens of tliis species were dredged in tlie vicinity of Kauai Island. The first came from the western end (Sta. 4130) at a depth of 288-309 fath.; the second from the northern extremity, Mokuaeae Islet (Sta. 3992), at a depth of 528 fath. The first specimen was found in a mass of polyps of Epizoanthns (Plate 2, fig. 3) elevated at least a foot above the Ijottom, and densely matted together in such a manner as to preclude the possibility of accidental lodgment. The second individual was found by Dr. W. K. Fisher among the spines of a species of starfish {Odinia pacifica Fisher) also in such a ])osition that it could scarcely be due to accidental shifting. There is no especial reason for considering this species a parasitic form nor indeed a commensal, for the food in the alimentary canal consisted of small ciuantities of plant spores, sponge spicules, and organic debris such as is ordinarily found in those species that burrow in the ooze. It seems more probable that it, like Chaetoderma 7iitiduhmi, as described byWiren, may leave its burrow to crawl about on the bottom, or as in the present case even on the bodies of other animals. In its external features this species displays the usual characteristics of other members of the family. The body including the globular and apparently non-retractile prothorax, itself about 2 mm. in length, is 12 mm. long. Imme- diately behind the swollen part of the prothorax the diameter of the body is 0.65 mm., and this continues with little change through the anterior half of the ani- mal. Beyond this point the calibre gradually increases to I3 mm. in the neigh- borhood of the cloaca, beyond which a slight decrease occurs that continues to the end of the body. In the other specimen the size of the prothorax is the same, but the length of the metathorax ( 19 nun.; diameter 0.5 mm.) and the abdomen (7 mm. ; diameter 1 mm.) is considerably greater and bears witness to the futility of using the length index in the discrimination of some species. The color of both specimens is a slaty gray, though this is usually obscured by an inorganic incrustation covering the body generally. In the region of the metathorax the larger spines are completely hidden in a granular deposit that gives this part of the animal a brick-red shade. The same substance, in one ease black in color, is present in several other species of Chaetodermatidae in my possession, and may perhaps be an excretory product thrown out from the coelomoducts. 50 CHAETODERMA HAWAIIENSIS. Tho cuticular plate covering the fi'oiital sense organ is almost circular in outline and is elevated above the general surface of the prothorax. This is especially the case with the lateral portions which assume the form of pronounced folds decreasing in height as the centre of the organ is approached. The boun- dary between the cuticle and the underlying epithelial cells is not sharp and the outlines of the cells themselves are not clearly visible. The greater number appear to be sen.sory elements and are distinguished by their relatively slender appearance (the diameter being to the height as one to four) and by darkly staining elongated nuclei placed in the basal half of the cell. Among these are a few cells of the same height, but of greater diameter, which contain spherical centrally placed nuclei with a small amount of chromatin. Great numbers of ganglion cells are situated in immediate contact with this sensory plate, and some of the more deeply seated clearly connect with nerves passing out from the precerebral ganglia (as I have termed the great accumulations of ganglion cells in contact with the brain in the Chaetodermatidae) , and on the other hand originate fibres that pass down to the frontal organ (Plate 28, fig. 1). A very few small pyriform gland cells, staining almost black in haematoxylin, extend from the midst of the ganglion cells to the sense organ where they probably open to the surface. Large numbers of muscle fibres attach also to the sensory epi- thelium and at several points there are indications that they pass between the hypodermal cells and connect directly with the overlying cuticle. The hyjiodermis is i)ractically the same as in other species antl not jjarticu- larly favorable f(or the solution of any of the several problems connected with it. Its cells differ considerably in form and appearance. In th(> swollen jsart of the prothorax they are slightly higher than broad; in the metathorax the reverse is true; while in the abdominal region the height is about twice the diameter. Everywhere their boundaries are indistinct, and thus unlike the sharply defined central nuclei. Here and there are more slender elements with elongated darkly staining nuclei, and somewhat more numerous are the basal cells in contact with the base of the overlying spicules. These latter cells vary widely in general appearance from very small compact elements to others large, globular, and much vacuolated, owing to different stages of development and probably to some extent to mechanical compression. No pigment cells exist, nor wandering cells nor other elements that are sufficiently different from the usual constituents to be jjlaced in a separate catagory. The spicules are of the usual spearhead shape, and form a continuous series of increasing size, from those of the prothorax with a length of .0275 mm. CHAETODERMA IIAWAIIEXSIS. 51 to others on the postalxloiuen .22") iiiiii. lonjz;. Plato 37, fig. 12, gives an accurate idea of the usual type of spine, tliesc heing from the niitldle of the metathorax. The mouth opening is situated about the centre of the frontal sense organ and is rcMnarkable for its minuteness. From here to the region of the radula the canal is relatively small, not exceeding one .seventh of the greatest diameter of the prothorax. It is invested by a thin layer of muscles and is attached also til a considerable number of scattered fibres that pass outward and are inserted in the body wall. Ganglion cells arranged in groups as in the Neomeniidae are also fa.stened to the buccal mass. The epithelial lining consists of high columnar cells with basal nuclei imbedded in moderately dense cytoplasm, that more distally becomes filled with a finely granular colorless secretion. In one speci- men ])articles of a golden yellow color occurred in, or between, some of these elements, but whether they were developed in .sZ/m, or had l)een j)ro(luced by cells more externalh^ placed, it is impossible to determine. While in one specimen the digestive tract continues of about the same calibre throughout, the other expands widely in the region of the radula, and at the anterior end of this enlargement a circvdar fold is present that is probably homologous with the proboscis of Drepanomenia, for example. In one indi- vidual this swollen section is almost completely filled with diatoms, plant spores, sponge spicules, and organic remains some of which appear to have l>een mixed with some viscous secretion and mouUled into globular masses. From a point corresponding to the hinder border of the globular part of the prothorax to the anterior end of the preabdomen, the wall of the digestive tract is relatively thin due to the scarcity of muscle fibres and the lowness of the epi- thelial cells. The latter are columnar elements of medium height with spherical basal nuclei and an abundance of a finely granular, light yellow substance filling the entire distal half. At various points throughout the prothorax and meta- thorax this substance is in the act of escaping in the form of droplets constricted from the cell, antl all stages exist between this and the development of a finely granular secretion filling the canal. Cells freed of this secretion are cubical in form and are usually relatively dense. Cleared specimens show that while in this part of the gut true pouches do not exist the canal is by no means entirely straight, wrinkles and folds occurring throughout, thoHgh they lack the definite- ness and regularity of the dilations in the Neomeniidae. At the level of the front end of the gonad the development of digestive fluids becomes relegated to a set of cells that form a large diverticulum extending to the hind end of the body. In this digestive gland, judging from the material 52 CHAETODERMA HAWAIIENSIS. in hand, two distinct kinds of matorial are secreted, but in widely differing quantities according to the locality. The cells attached to the gonad are usually more or less pyriform with comparatively small, dense almost homogeneous nuclei placed basally, while the remaining protojjlasm is closely packed with innumerable granules. By far the greater number of these ai-e sjiherical and oi yellow or slightly greenish yellow tint. They are liberated, as is the secretion in the more anterior pai'ts of the gut, ])y the constriction of (lro])lets fi'om the distal end, and may be seen undergoing disintegration and solution in many different places. Among the graiuiles of this character are others in the foi'm of small particles of a distinct jjink or violet color after treatment with haematoxylin. They have every appearance of being an end product and nf)t a stage in the development of the more abundant secretion. Elsewhere in the gland the cells are of looser texture, tlie basal nuclei are relatively larger and granular and many if not all contain the two species of secretion just described. (Jenerally the yellowish product is scant in amount, while in many cells the violet tinted substance accumulates in spherical or elliptical masses, surrounded l)y a vacuole in preserved specimens, that almost tills the cell. These secretory products are passed out entire, and in a single section as many as twenty-five may hold positions in the lumen of the gland. Making their way forward many if not all pass into the intestine, and here may be seen in various stages of solution, forming at first a vacuolated product that before dissolving completely transforms into a finely granular material of maroon color after treatment with haematoxylin. The cells of the intestine are cul)ical in form and in front (if the pericardium show signs of slight glandular activity. Behind this p(/int this phase of activity disappears, and cilia l:)ccome developed and contimie to the opening into the cloacal cavity. The large size of the ganglia and the abundance of the nerve cells that envelop them and also the sharpness of even the smaller nerves renders it pos- sible without much effort to gain a very clear idea of the nervous system of this species (see Plate 7, fig. 2). As is there shown the brain, located some dis- tance above the digestive tract, is tlistinctly envelo|)ed in a delicate connective- tissue sheath and is clearly bilobed in form though its outlines are somewhat obscured l)y great masses of ganglion cells (forming the precerebral ganglia) attached chiefly to its lateral and ant(n-ior surface. A considerable number of delicate fibres, passing out from the brain, attach to these ganglionic bodies which in turn are connected with large numbers of nerve fibres that pass chiefly to the walls of the mouth and the frontal sense organ. CHAETODERMA H WVAIIHXSIS. 53 This part of the iicrvous system is tiius t'sseiitially as \vc find it in otlicr Solenogastres. In Pronemurnld hawaiiensis, for example, the supraoesophageal gangha are comiecteil willi several nerves some of which unite with groups of ganglion cells attached to the bases of the cirri, and from these again other nerves pass to the digestive tract and ])ii)l)al)ly to th(> Ixidy wall. In C'haetoderma the chief difference is that the nerves uniting the brain and precerel)ral ganglia are very short. In the present species the pedal and lateral connectives unite immediately before plunging through the precerebral ganglia, and as Plate 7, fig. 2 shows the labio-buccal cord unites with them before the brain is reached. This same condition of affairs exists also in two species of this genus taken in Alaska, though much more obscured than in the present species. At the posterior end of the prothorax the pedal and lateral cords that have gradually approached each other actually come in contact and in one specimen even fuse for a short ilistance and lose the usual sheath of ganglion cells. Anterior to this point two pedal com- missures exist, but until the hindmost pai't of the body is reached no farther trace of such nerves has been found. On the other hand latero-pedal coimectives are present throughout the entire length of the animal. The relations of the labio-buccal ganglia are represented in Plate 7, fig. 2. The non-ganglionic connectives imbedded in the pharyngeal wall unite with the superficially attached ganglia, that are also imited by a connnissure passing behind the median tooth. In front of the ladula there are connectives giving rise to nerves passing dorsally to what is probabh' the subradular organ and in addition attaching to a ganglionic mass in the mid line. As this part of the nervous system appears with greater distinctness in ('. atknmiUi it is more fully described in that connection. Throughout the entire metathorax the lateral and pedal cords pursue their course almost in contact, here and there giving rise to nerves that soon disappear, and finally join in the extreme posterior end of the body. Shortly after their union they are connected by a heavy ganglionic commissure jiassing dorsal to the intestine. In the mid line it develops a single nerve that makes its way into the tissue of the rectum, wliile on the dorsal side four fibres originate, two of which pass at once ;nt(j the gills while the others attach to the inner side of the cloacal epithelium, and branching repeatedly supply this membrane and the dorsal body wall and a well-marked fold of the hypodermis to be described presently. At the junction of the latero-pedal cord and the commissure a nerve arises that passes backward and appears to supply the ventral body wall of the cloacal region. 54 CHAETODERIMA HAWAIIENSIS. Owing to the deliris encrusting; the posterior end of the body it is impossible to determine the position of the dorsal sensory groove in entire specimens. In sections it is seen to occu])y the usual position, that is from the extreme hinder end of the animal forward to a point almost immediately above the level of the anal opening. As is represented Plate 6, fig. 8, it consists of a relatively deep fold of the hypodermis, that anteriorly rapidly disappears but is continuous with a ridge-like elevation in the mid dorsal line extending for a short distance more anteriorly. Some of the spines of the immediate neighborhood are of compara- tively small size and overarch the depression, which is also covered by a thin continuation of the cuticle investing the body. The cells of this organ consist of those common to the hypodermis, and others which are much more slender and compact with spindle-shaped nuclei usually subcentrally placed. The latter elements are probably sensory and connect with small groups of ganglion cells holding positions immediately beneath the circular somatic muscles in the neighborhood of the organ, and on the other hand are undoubtedly related with nerves from the branchial ganglia. That this is a definite sense organ and th(^ homologue of the dorsal organ of the Neomeniidae, as maintained by various authors, there is little doul)t, l)ut there is nothing that more definitely establishes its function. The gonad extends from the front end of the metathorax to the pericardium with which it is united Ijy two short and comparatively wide ducts. Both speci- mens were sexually mature males and considerable cjuantities of spermatozoa occupied positions in the pericardial cavity, and at various points in the coelomo- ducts. These last mentioned organs arise from the postero-lateral borders of the pericardial chamber in the form of clearly defined tubes, whose cells are nearly cubical in form and support an al)undance of large cilia. Bending sharjily inward each becomes continuous with a canal, of much larger size and ditTerent structure, that after extending forward for a short distance pursues an irregular coiu'se opening symmetrically on each side of the rectum. The lai'ge non-ciliated por- tion of the ducts is composed of rather low cells with well-defined, basally placed nuclei, in which the chromatin exists in the form of a moderate number of sharply- defined granules. In the more distal part of each cell is a sharply defined vacu- ole, in which are one or two light greenish yellow bodies, having the appearance of concrements such as occur in the kidneys of several molluscs. At various places these are in the act of escaping through the ru])tLued or ilissolved end of the cell or having become free are nndergoing a process of solution. Such an appearance in the kidney of other molluscs would not in any way appear unusual. CHAETODERMA ATTENUATA. 55 and leads to tlic irresistil)l(> ('nnflusioii that here the coelomoducts are not only morphologically related to the n>nal organs in the Chitons or other molluscs, but i)hysiologieally also as Wiren first clearly stated. Chaetoderma attenuata, sp. nov. Eight specimens of this species were dredged near the southern limit of Alaska, buried in glacial mud brought down chiefly by the Stikine River. One from Kasaan Bay (8ta. 4244) occurred in green mud at a depth of 50-54 fathoms; five opposite the mouth of the Stikine River (Sta. 4250) were in the same habitat at a depth of 61-66 faths. ; while two in the waters of Stephens Passage (Sta. 4252) were buried in gray mud at a depth of 198-201 faths. Their appearance in life, Plate 4, fig. 3, answers closely for preserved material. The type specimen measured 45 nun. in length by 1.7 mm. through the metathcrax, and 2.6 mm. through the abdomen. The color of alcoholic material, which is the same as the living save for the pinkish tinge due to haemaglobin in the head and gill region, is almost white with a tinge of yellow, becoming grayish where the liver is located. The body wall, including muscular, hypodermal, and cuticidar layers, is of median thickness and is typically located, but in specimens killed in vom Rath's fluid certain elements appear that have not been fully described, though they probably occur in all species of the genus. These are the so-called giant cells (Riesenzellen Wiren) which in ordinary material present the form of empty vesicles with the nucleus mibeddetl in the wall. In life this cavity is tilled with a secretion, that after treatment with fluids containing osmic acid, is granular as Wiren has remarked. In favorable situations it may readily be seen that fibres, muscular at least in part, e.xtend from the somatic musculature and penetrating the hypodermal la.yer attach to the sides of these cells (Plate 25, fig. 7). Ap- pearances suggest that the secretion, upon the contraction of the fibres, is forced into the neighboring lacunae, but in no case has this been actually observed though proximally the cell may be produced into a comparatively slender, short stalk. Distally the cells are usually in close contact with the free surface of the cuticle and present a sharply defined rounded appearance. Posteriorly these elements become somewhat less numerous and of smaller size. In alcoholic killed material the fibres may be distinguished, but their attachment to the cell body is very indistinct. Wiren ('92) states that these giant cells are not sharply differentiated from the basal matrix cells of the spicules, but this refers, so far as I am able to judge, 56 CHAETODERMA ATTENUATA. merely to their form as the spicule mother cells do not contain any clearly de- fined granular secretion. On the other hand the matrix cells shade much more perfectly into the cubical elements that probably form the cuticle. The mouth, placed in a cleft on the dorsal side of the buccal plate, opens into a tube whose form and general appearance are represented (Plate 25, fig. 1). The lining epithelium consists of the usual high cohmmar cells produced into several irregular longitudinal folds, through which the outlets of the buccal glands make their way. These last named organs are comparatively abiuidaiit, especially on the ventral side of the pharynx, and extend from the region of the brain to the radula. A subradular organ certainly exists in this and several other species, if position and innervation be any criterion. In material killed in vom Rath's fluid it appears with the greatest distinctness as a sharply defined median area composed of high columnar cells situated immediately in front of the peg-like tooth (Plate 25, fig. 10). In alcoholic material the ajjpearance is not so striking, and yet there is very little difficulty in distinguishing the organ. However, with such material it is sometimes a task to determine its innervation. Nerves in the immediate vicinity are usually visible, but to trace these into the ganglia is frequently a perplexing o]ieration. In vom Rath's material on the other hand the entire system is clearly differentiated (see section on nervous system). The radula and its supports (Plate 25, fig. 2) are of comparatively large size but arc typically arranged and require nt) especial description. Beyond the radula the gut becomes circular (Plate 25, fig. 3), the epithelium relatively high and a iim^ly granular secretion fills the distal (wo thirds of the comjionent cells. Among these are a very few more globular elements with a darkly staining more granular secretion. Beyond the pharynx the gut widens, the cells become lower and slightly glandular with the exception of a very few cells containing a yellow secretion. Beyond this point the relations of stomach, intestine, and liver are typical and require no detailed description. The brain and anterior portion of the nervous system closely resemble what is found in C. erudita, and so require but little additional description. It appears that the labio-buccal connectives have an origin independent of the lateral and pedal, which as in C. erudita unite before entering the brain. The commissures of the pedal cords are relatively more slender than those of C. erudita, save the anterior one which is of exceptional thickness. In some cases nerves arise from the commissures and are distributed to the body wall. The labio-buccal system is of unusual interest since it possesses what may CHAirroDKR^rA attentata. 57 be considered to l)e a siihiadular system with ganglia and connectives with fibres passing into the above described organ in iVdnt of the raduhi. The labio-buccal connectives pass backward as usual and unite with tlie well-known ganglia imbedded in the ])haryngeal niusculalure: and these bodies are in turn united by a cord in which two small ganglia are intercalated. A nerve which appar- ently has escaped observation arises from the posterior surface of each of the larger ganglia (Plate 13, fig. )?), and ind)edded in the pharyngeal wall may be traced to the forward border of the stomach. A short distance in front of the labio-buccal ganglia a clearly defined fibre arises from each of the coimectives (Plate 13, fig. 3), and, iml)eddi'd in the muscle of the pharynx, courses downward and inward and joins a ganglionic mass that gives slight evidence of being paired (Plate 25, fig. 10). To the outside of the ganglion (or ganglia) a nerve arises from each of these connectives uniting with the labio-buccal connectives, and coursing dorsally attaches to the base of the subradidar organ. There is absolutely no doubt of the existence of this system, the grayish nerves showing with great distinctness against the yellowish muscle fibres in maiei'ial killed in vom Path's fluid. In material fi.xed in alcohol on the other hand it is sometimes difficvdt to trace. The ganglionic mass may closely re- semble a salivary gland and the nerves from it counterfeit muscle fibres; never- theless with an oil innnersion lens I have been able to demonstrate its presence in all the species of the genus described in the present paper, and in a sjiecimen of ('. nitiduiiDH kindly sent me by Professor Hubrecht. As is more fully noted on page 172 I believe that the ganglion and its connectives constitute a subradular system the homologue of the one I have described in some of the Neomeniidae. Posteriorly the pedal and lateral ganglia unite in the customary fashion, and at the ]ioint of union give rise to two small nerves which become closely applied to the body wall, and aftei- branching are lost to sight among the longi- tudinal somatic muscle fibres. From the suprarectal ganglionic mass (Plate 12, fig. 4), several branches arise some of which apjiear to have escaped obser\ation, or at all events have not been traced to any considerable extent. Of tlu^se the larger i)air originate from the ventral side of the ganglion and make their way ventrally to the sides of the rectum, where according to Wircn's figures and description they diminish very rapidly in diameter and form a delicate subrectal commissure. In the present species this is certainly not the case, nor is it true of ('. erudila, for arri\ing at the rectum each follows it posteriorly to the anal opening, antl then passes outward almost at a right angle and becomes imbedded 58 CHAETODEliMA ATTEXr'ATA ill the ventral gill retractors and in this position may he traced almost to the apex of the gill. Between its point of origin and its attachment to the rectum at least four small nerves arise and extend fan-like into the ventral gill retractors which they probably innervate. I have been unable to find any sulirectal commissure. From the dorsal side of the suprarectal commissure four nerves arise, of which the outermost ])air extends dorsally through the superior gill retractors, and imbedded in the dorsal cloacal wall, whicli it jirobably inner\ates, may be followed i'' comjiressed tulies sejiai'ated l\y the aorta. The pericardium is of umisual size, extending behind the heart nearly to the posterior end of the l)ody. As may he seen Plate 36, tig. 2, it is interrupted l)y the dor.sal gill retractors, but behind these muscles the cavity again becomes continuous across the mid line, extending down the sides of the cloacal cavity (Plate 25, fig. 5) and posteriorly forming a horn-like extension in the mid line. The heart is the usual tubular organ but posteriorly it unites with an atrium, which may be considered an auricle or an invagination of the ventral pericardial wall continuous jiosteriorly with the efferent bi'anchial sinus. The openings of the coelomotlucts hold the usual position, at the sides of the supi'arectal connnissui'c, but the tubes with which th(\v comnumicate are in the first ]iart of their course very slender, ciliated, and somewhat convoluted. In this condition they extend ventrally and join the glandular portion (Plate 36, fig. 2). The cells of this .secretory portion are of the usual type, almost cubical vacuolated elements containing a small concrcment. The jiosition of the external opening is shown (Plate 25, fig. o). Wiren ('92) has accurately described a patch of glandular e])ithelium, a modification of the cloacal wall, which on each side of the body suridunds the openings of the gonoducts and extends to a certain extent ovci' the base of the gills. The cells composing it are high and consist of very slender supporting cells and glandular elements filled with an almost homogeneous substance, con- CHAETODERIMA ERUDITA. 59 tainiiis in favorable pr«])arati()iis gi-()vi];)s of small jirismatic crystals. This description answers for the present species with the exception of the crystals which have not been found, ^\'iren (■?>inpar(>s this j2;lan(lular area with the shell gland of the Ncomeniidae, thmiuli clainiinfi; it acts as an excretory or^an. Be- yond certain histoloj;ical resenihlances there are no cogent reasons for acce]iting such a theory. Chaetoderma erudita, sp. nov. Ten specimens of this species were taken in Lynn Canal, Alaska (Sta. 4258) at a depth of 300-313 fathoms; and forty-one were dredged in Chatham Strait, Alaska (Sta. 42()4) at a depth of 282-293 fathoms. In both cases the bottom consisted of very tenacious green mud. A numbei- of individuals were kept in an aquarium aboard ship and lived apparently in a normal state, burrowing extensively and in some instances feeding on organic debris. Two males gave ofT considerable quantities of sperms during a period of over an hour. It escaped from the sides of the cloacal cavity, lateral to the gills and soon diffused into the surrounding water. Much care was taken in the preservation of these animals, and yet the shrinkage was considerable, in the case of some of tlic more active ones, amounting in six individuals to a decrease in the body length of one fourth. It thus becomes apparent that the length index or ratio of length to diameter is not to be depended on in the discrimination of species. The entire animal is represented (Plate 4, fig. 9); the buccal plate (Plate 4, fig. 11), and the spicules (Plate 37, fig. 15). The hypodermis comprises three fairly distinct types of cells. Of these the most conspicuous, in alcohol killed material, is the basal cell of each spine whose nucleus is placed considerably above the level of the other types. In material killed in vom Rath's fluid the Reizenzellen of Winni, well-detined globular cells, are very distinct and con- tain a highly vacuolated material which almost totally disappears in alcohol kille mouth, a relatively wid(^ opening in the deeply cleft buccal plate, opens into a laterally compressetl ]:)uccal tube that beneath the brain develops longi- tudinal fokls and a more circular outline (Plate 29, fig. 4). As far as the forward end of the radular supports buccal glands in great abundance are attached to its walls. The subradular organ is not as sharply defined as in C. attoDiala, yet is clearly distinguishable as a median ventral elevation composed of slender 60 CHAETUDEllMA ERUUITA. columnar cells of greater height than those of the adjoining epithelium. In one specimen -killed in vom Rath's fluid the protoplasm of the comjionent cells is much vacuolated or in a very small number contains a granular secretion. Poste- rior to the radula, whose general appearance is sufficiently shown (folate 29, fig. S), the pharynx becomes dorso-ventrally compressed, then circular and opens into the stomach. This is a relatively spacious organ with tliin unfolded walls that posteriorly become thicker and folded. The relation of intestine and liver are typical and require no description. In the proximal part of the liver, and throughout the major portion of the intestine, there are considerable quantities of organic remains, diatoms, sponge spicules, a few fragments of entomostra- cans, and several chambered Foraminifera whose protoplasm was only partially digested. The nervous system of this species is exceptionally clearly tlehnetl in one specimen killed in vom Rath's fluid and for this reason has been more thoroughly studietl than any other species of the genus described in the present jiaper with the exception of C. (tttvnudld. The brain is very distinctly bilobed, a deep indentation occurring on its anterior surface. From its lateral and forward borders nerves pass into the ])recerebral ganglia which in tiu'ii send tremendous bundles of fibres to the Iniccal sensory plate. In some sjiecies the connectives to the pedal, lateral, and labio-buccal systems have distinct origins in the brain, l)ut in the present case they are united for a considerable tlistance (Plate 13, fig. 3). Each of these compound coimectives after leaving the brain and pass- ing forward a short distance gives rise to the labio-buccal connective and con- siderably farther on the pedal anil lateral connectives lieconie differentiated. The pedal and lateral ganglia are in \\w usual positions and are unit(>d l)v freciuent connectives and connnissures. At the points of origin of thes(> nerves there are no very cleai'ly dehned enlargements though anteriorly the pedal and lateral cords are of large size and gradually taper posteriorly, attaining their average size about the hinder border of the prothoi'ax. As these ganglia diminish in size the connectives and commissures become reduced in calibre and are difficult to follow yet they may be traced here and there throughout the entire length of the animal. As in C. aUenuala a nerve arises from each lahio-ljuccal connective about the level of the forward border of the radula and ])a,ssing inward and downward joins a small subradular ganglionic mass. In this sjiecies the gan- glion shows no indication of being paired. Each of these subradular connectives gives rise to a ner\e distributed to the subratluiar organ and more laterally swells niAETODKinrv moxtereyexsis. 61 somewhat though ganglion cells are lacking. From each of these enlargements a nerve is developed, and after branching in tlie jiharyngeal musculature becomes lost to view. This suhradular system does not appear with the diagrammatic clearness of the one in ('. itlU intnUi, l)ut there is no esjiecial difiicuHy in determin- ing its relations which arc essentially the same in the two species. In the posterior regions of the hrxly the nervous system very closely resembles that of ('. attcnunta. The gonad, with the usual characteristics and relations, opens into the f(jr- ward end of the pericardium by means of comparatively large tul)es in sexually matui'e animals. In som(> animals, possibly owing to killing fluids, the peri- cardial cavity is greatly distended with spermatozoa wliich have made their way into the proximal half of the coelomoducts. These last named tubes com- municate by wide openings with the pericardium and on the other hand extend forward as ciliated tubes for a short distance. Bej'ond this point their walls become glandular and are thrown into numerous convolutions which render it impossible, without much effort, to determine their exact relations. Posteriorly each duct becomes more simple, though of fairly large calibre, so that it con- tracts the cavity of the cloacal chamber; and on the ventral border of the fold thus developed the outlet canal is formed (Plate 29, fig. 5). Chaetoderma montereyensis, sp. nov. This species is evidently abundant in the deeper waters of Monterey Bay, California, as 155 wei'e taken distributed through the following stations: nine from 4485, seven from 4o()S, fifty nine from 4522, fifteen from 4523, thirty-one from 4524 and thirty-foiu' from 4525. In every case the bottom was nuid and the depth varied from 39 to 35G fathoms. C'hloretone (aceto-chloroform) was used with good results as a narcotizing agent and 70 % alcohol served as a fixing agent. The length of a medium sized specimen ' is 45 mm. with an average diameter of 2 mm. through the prothorax and 3 nmi. through the preabdomen. The color in life and ]ireserved material is yellowish white. The hypodermis \ery closely resembles that of C. atleiuiata. The spines are represented in Plate 37, figs. 2, 3. The mouth opens through a slit in the subelliptical buccal plate (Plate 4, fig. 17) and leads into a laterally comjiressed tube which becomes circular in ' Generally speaking the larger siieciniens ooiiic from the shallower depths. Thi.s is very marked in romparing those from Sta. 4.52,5 with others from Sta. 451)8. These size tlifferciices, however, do not appear to be eorrelated with any constant st met ural peeuliarilies. (32 CHAETODERMA ARGENTEA. outline slightly in front of tho radula (Plate 27). Throughout its entire course to the hinder borders of the radula its walls, more than commonly muscular, afford lodgment for numerous salivary glands whose secretion stains dai-kly with haematoxylin. The radula, its supports and nuisculature are typically situated but are exceptionally heavy and powei-ful. The remaining divisions of the digestive tract are related as usual and are represented on Plate 27. Countless thousands of diatoms, together with nondescript organic and inorganic remains, fill the intestine and in some specimens, the stomach. The pericardium is a comparatively spacious chamber, extending backward some distance over the cloacal cavity (Plate 27, fig. 9), and is perforated by the superior gill retractors; but otherwise neither it nor the tubidar heart and the connecting sinuses are peculiar in any important i)ai'ticular. The nervous system has been studied in considerable detail, and in all essential respects has been foimd to resemble that of ('. attcnuuld for example. The gonad, with the usual characters, opens into the ])ericardium l)y com- ]iaratively wide, dorso-ventrally compressed tubes. The imier openings of the coelomoducts are likewise of large size (Plate 27, fig. 8) and the adjacent ciliated section also though the latter is unusually shoit. This ciliated section unites with a division of the glandular part (shown on the left, Plate 27, figs. 2, S). The (uitlet (Plate 27, fig. 9) occurs in the customary position and is sur- roundetl by the glandular modification of the cloacal epithelium as in C. attcn- uata and a few other species. Chaetoderma argentea, sji. ikiv. One specimen (Plate 4, fig. 7) of this species was taken in southern Alaslca in the green mud of Behm Canal (Sta. 4231) at a depth of 82-113 fathoms. It was in a moribund condition and with the exception of slight movements of the body and gills gave no signs of life. The measurements are, total length 24 mm. diameter of the prothorax 1.6 mm. while the greatest diameter of the preabdomen was 2.6 nun. The color in life and in a preserved state was a silvery white. The cuticle is scant in amount and the hypodermis is comparatively low and is composed of small cells cubical or low colunmar in form. Among these are the giant cells (Reisenzellen) fi'om which the secretion has tlisappeared but they are attached to faintly staining fibres whose exact relations ha\-e not been determined. The spines are represented (Plate 37, fig. 6). Although the animal when captured was alive it never relaxed sufficiently to allow the buccal sensory jilate to become exposed. In sections this last named CHAETODEiniA SCAHRA. 63 organ appears to hv typical thousli the fi;laiuls that open aloii^ its margin are more than usually developed. Tlio proradular scH'tion of the gut is of average diameter, fairly muscular and is provided with numerous glands uncommonly compact except on the dorsal side behind the brain. Violent contractions of the prothorax have apparently been responsible for the dislodgment of the epitlielial lining of this part of the digestive tract; but there are indications that a subradular organ exists and the nerve supply is distinctly evident. The radula is constructed on th(> usual plan as may be seen (Plate 26, fig. 2). Bej'ond the radula the gut Ix'comes narrow, circular in section and very soon unites with the cajjacious stomach whos(^ i-elations to the intestine and liver are of the usual tyjie. The stomach and especially the intestine contain a considerable amount of inorganic and organic material, diatoms being especially abundant. The single specimen is a male with the gonad distended with sex products in all stages of development. \'iolent nuiscular contractions have forced a mass of sex cells, many of them immature, into the pericardium; and at various points along the coelomoducts fully developed sperms are present. The reno-peri- cardial openings, at the lc\'el of the posterior border of the sujirarectal commis- sure, are relatively wide and lead into coi-resjiondingly spacious, highly ciliated tubes wliich ])ass almost directly ventrally to a point about opposite the mid lateral line where they unite with the glandular portion (Plate 36, (ig. 1). This last named section extends as a slightly conx'oluted tul)c to a jioint about oppo- site the posterior border of the gonad where it bends sharply upon itself and venti-al to the dorsal ciliated section opens into the cloacal chamber at the usual point. The nervous system shows with distinctness and has been traced in con- si(leral)le detail, but as the results show it to be essentially the same as in C. ttttciiNdtd and ('. crudiUi it demands no especial description. The subradular ganglion with the usual connections is clearly a single mass. Gland cells in the gills are very definitely distributed, in cro.ss sections being disposed along the transverse axis of the body (Plate 26, fig. 5). Chaetoderma scabra, s]). nov. One individual was dredged in Monterey Bay, California at a depth of 795-871 fatli. It measures 12 mm. in length, 1 mm. through the metathorax and 2 nun. through the greatest diameter of the preabdomen. The expanded poilion of the ])rotliorax is light brownish yellow; more posteriorly the brown shade is mmc i)ronounced, becoming oli\-e-green in the region of the liver which 04 CHAETODERMA CALIFORNICA. shows through the transhicent Iwdy wall. An orange-brown substance incrusts the spines about the cloacal opening. Tlie s])ifules are rejiresented (Plate 37, fig. 19). The form of the buccal plate and the position of the nujuth oj^ening are shown (Plate 4, fig. 10). The adjacent section of the digestive tube rapidity as- sumes a circular form in section, and a few comi)act groups of gland cells become applied to the dorsal, and to a less extent the lateral walls. Innnediately liehind the brain these lobules become larger (Plate 30, fig. 3), but soon disappear moi-e posteriorly. The radula was cut obliquely and it is therefore somewhat difficult to determine its exact relations. The tooth appears to be relatively slender, but its supports and musculature are typical. Behind the radula the pharynx again becomes circular and in this form joins the stomach (Plate 29, fig. 9). This last named organ is at first tliin walled, but the epithelium soon grows higher, becomes folded and soon smooths out at the level of the jiosterior end of tlie prothorax (Plate 29, fig. 11). Again becoming thick walled and of small calibre it unites with the liver and intestine. From this point onwaril these last mentioned organs are of the usual type. The ])ericardial cavity is of moi-e than average size (Plate 29, fig. 10) and the heart is highly muscular; otherwise neither these organs nor the connecting sinuses are unusual. The specimen is sexually matiu'e ami nuiltitudes of sperms have made their way from the gonad through wide tubes into the pericardium and the ]iroximal portion of the coelomoducts. The ojienings of these latter organs into the peri- cardial cavity are comjiarativel.y large and the ducts themselves are relatively simjile. As in some other small sjiecies the glandular portion is a sim]ile canal extending as far forward as the posterior end of the gonad whcic it bentls abruptly and making its way posteriorly opens by an inconsjiicuous pore into the cloacal chamber. The nervous system is distinct and sharply defined and has been carefully examined, but it does not differ in any important respect from that of C erudita for example. Chaetoderma californica, sp. nov. One specimen was collected in the neighboihood of San Diego, C'alifoi-nia, (Sta. 4381) at a depth of 618-667 fathoms. It measures 24 mm. in l(>ngth by l.() mm., the average thickness of the metathorax, and 2 mm. the average diame- ter of the preabdomen. The general ajijiearance of the animal and the relative length of the vai'ious divisions of the body are shown (Plate 4, fig. (i). The color CHAETODERMA CALIFORNK^V. G5 of the protrusible portion of the prothorax is yellowish brown, while the remainder of the bod}^ is yellowish green. A rusty rod substance, possibly excreta, incrusts the spines in the cloacal region. The cuticle is of moderate thickness and rests upon a hypodermal layer whose nuclei, placed at various levels in the region of the prothorax, have at first sight the appearance of being more than one cell thick. The most common type of cell is relatively slender and contains an oval granular nucleus. Among these are other elements, probably spicule matrix cells, each of which contains a spherical nucleus larger in size than those of the foregoing class of cells and placed in the neighborhood of the base of a spine above the general level of the hypodermis. As in ('. attcnuaia spaces exist at frecjuent intervals in the hypo- dermis and from the vicinity of each fibres pass into the underlying muscular layer. As noted on page 55 there are reasons for the belief that these are gland cells, of unknown function, whose secretion is dissolved through the agency of alcohol when used as a fixing agent. The shape of the spines is shown (Plate 37, fig. 14). The alimentary canal opens through the dorsal half of the buccal plate; its first section (Plate 27) is a narrow canal that rapidly widens in the neighbor- hood of the brain. To its lining epithelium tlie usual muscles attach and afford lodgment for numerous buccal glands. These last' named organs extend from the mouth to the level of the brain and are similar to those of C. nilidulum save that the cells are less compact and of larger size. In the neighborhood of the radula the walls of the pharynx become more folded than in C. nilidulum and are unique in possessing a pouch (Plate 31, fig. 8) of considerable extent, into which the glands of the dorsal side ojien. Behind the radula the canal gradually narrows, its folds become smoothed out wheroujion it imitcs with the stomach. The salivary glands consist of several globular cells surrounding a small lumen that in .some instances is in direct communication with the digestive tract. However each cell communicates with a small ductule which gives evi- dence, not of passing the secretion into the lumen of the gland, but directly into the digestive tract through intercellular channels of the lining epithelium. The radula presents no especially noteworthy features. Its conical tooth is slightly more slender than is usual (Plate 31, fig. 1), but the cuticular wing- like .supports and musculature are entirely typical. As usual the pharynx opens by a comparatively narrow pore (Plate 31, fig. 2) into the stomach whose relations to the liver and intestine are normal. Large quantities of organic remains occur in the gut, Radiolaria, diatoms, and sponge 66 CHAETODERMA NANULA. spicules being distinguishal)le. Associated with these are numerous rounded cells (Plate 35, fig. 11) that occur also in the digestive gland. In most cases these are free but occasionally one may be seen that is encysted in the cells of the organs mentioned. Rarely they are associated in pairs as though in the process of conjugation. The gonatl holds the usual position and like the sperms, in all stages of development, presents no noteworthy characters. Posteriorly the halves of the organ diverge, become rather indistinct though their route may be traced with certainty, owing to the presence of spermatozoa, passing lateral to the heart or the expanded base of the aorta and opening into the pericardium. This last named space lacks the almost diagrammatic outline as in C. nitidu- lum and is much more limited in extent, but its relations to the gonoducts are very similar. Chaetoderma nanula, sp. nov. One specimen of this species was dredged off the coast of southern California (Sta. 4369) at a depth of 260-284 faths. It is 9 nun. in length by 0.9 and 1.2 the average diameter of the metathorax and preabdomen respectively (Plate 4, fig. 1). The color of the globular, protrusible portion of the prothorax is light brownish yellow (though this may have been produced by tannin from the cork), while the metathorax and preabdomen are considerably darker, the latter region becoming olive-green. A dark brownish substance incrusts the spines about the cloacal opening. The hypodermal layer is comparatively thin, the cells small and somewhat indistinct, yet are typical so far as may be determined. The spines are represented (Plate 37, fig. 18). The mouth opening represented (Plate 4, fig. 12), leads into a relatively spacious tube lined with slender columnar cells except along the dorsal side where they are almost cubical. A median fold, located immediately in front of the radula, probably represents a subradular organ since it is typically innervated. Salivary glands are almost wholly lacking, a small group attached to the pharyn- geal wall adjacent to the radula being all that is visible in the present specimen. The radula is small but ty])ical. Beyond it the tul)e narrows considerably, the lining becomes folded and in this form it unites with the stomach. At the outset this last named organ is plain walled but near its union with the liver becomes considerably sacculated. The intestine is distended with fragments of Radiolaria, sponge spicules, and organic debris. Parasitic Protozoa, resem- bling those from Chaetoderma calif ornica, are abundant and are imbedded in the epithelial lining of the stomach, intestine, and liver throughout their entire extent. CHAETODERMA .TAPOXICA. 67 The animal is a female, not perfectly matur(>, and the ducts leading into the pericardium are accordingly small. The openings into the coelomoducts are likewise minute, and the ciliated tube with which each connects is relatively long and slender. The glandular division with which it unites is a comparatively simple tube, at first directed forward until it reaches the level of the front end of the heart whereupon it bends abruptly and makes its way to the opening into the cloaca] chamber (Plate 28, fig. 10). The nervous system is in an excellent state of preservation and is clearly defined, but a careful study has failed to disclose any noteworthy feature. It may be mentioned that a subradular system exists similar in all respects to that of C. aUenuala. Chaetoderma japonica, sp. nov. One specimen (Plate 3, fig. 7) was dredged off Honshu Island, Japan (Oi Gawa, Sta. 3721) at a depth of 207-2.50 fathoms. The body is comparatively slender, measuring 17 mm. in length by l.o mm. through the nietathorax and 1.5 mm. through the preabdomen. The co^n- is almost white with a slight tinge of yellow. A slight incrustation, brick-retl in color covers the spines in the cloacal region. The spines are of the usual type. The mouth opens through a distinct pore in tlie l^uccal plate (Plate 3, fig. 8) which, like the neighboring section of the digestive tube, is abundantly supplied with glands, small celled and more than connnonly compact. These continue, for a considerable distance behind the radula, ajiparently unchanged in character though in many cases closely applied against the bases of the buccal and pharyn- geal epithelium. This first named organ with its supports and musculature is typical, as may be seen in Plates 30, 31. As the major portion of the body was not sectioned the union of liver and stomach has not been .seen; otherwise these organs conform to the usual i)lan. The jinTcctal iiortion of the intestine is lined with an exceptionally high epithelium so that the lumen is very small where it is not distended with pellets of faecal matter consisting principally of tliatoms and sponge spicules. The nervous system is not especially favorable for study and accordingly only its more general features have been examined. In this respect it is tyiiical. The specimen is a female and the fully formed ova present the customary appearance and are developed in a gland holding the usual position. The ducts leading into the pericardium, are large and as in the case of the last named space, and the gonoducts, are filled with eggs mostly disintegrated, due perhaps to violent movements of the somatic nuisculature. The gonoducts open by rcla- 68 CHAETODEKMA ROBUSTA. tively large pores into the pericardium and as moderately spacious tubes without any marked convolutions, extend to their openings into the cloaca (Plate 30, fig. 7). Surrounding these pores the cloacal epithelium is modified to form the glandular area similar to that of C. attemiata. Chaetoderma robusta, sp. nov. Four s])ecimens of this species were taken south of the Alaskan peninsula (Sta. 3210) in green mud at a depth of 483 fathoms. The largest specimen (Plate 4, fig. 5) is GO mm. long with an average diameter through the metathorax and preabdomen of 3.5 aiul 4.7 mm. respectively. The smallest is 35 mm. long with an average metathorax diameter of 2 mm. and 3 through the preabdomen. Where the spines have not been dislodged the general color of the body is slaty gray shading to buff at the anterior end of the body. A yellowish brown sub- stance incrusts the spines about the cloaca. The hyjiodermis consists of numerous small cells rather closely crowded so that the cells lack distinctness. However giant cells are visible and faint fibres, connective tissue or muscle, springing from the underlying body wall appear to attach to them. Spicule-matrix cells in all stages of development are visible, and in each case the spines are attached to only one cell so far as it is possible to judge. The remaining elements are comparatively slender, compact and lack any noteworthy features. The spines are represented in Plate 37, fig. 4. The buccal plate is shield-shaped in outline (Plate 4, fig. 19), and is pierced by the mouth opening. The buccal and pharyngeal cavities are slender, and the walls of more than average thickness (Plate 30). The lining cells are accordingly very slender, ciliated and are thrown into a few prominent folds. The ductules of a very large number of salivary glands make their way between the cells and in some cases are in the act of pouring their secretion into the canal. A sub- radular system is present, and as usual two nerves are distributed to a median fold of pharyngeal epithelium that probably functions as a subradular organ. However with the exception of affording scarcely any outlet for the salivary glands its cells are not clearly distinguishable from the general epithelium. The radula consists of the usual conical tooth, rather heavier than usual, but with sujiports and musculature of the customary type. Beyond the radula the tract becomes circular in cross section before uniting with the stomach whose relations to the liver and intestine are typical. The circulatory system presents no noteworthy features beyond the fact that the heart is suspended by a fold of the pericardial wall reinforced by a few LIMIFOSSOR TALPOIDEUS. O!) connective-tissue fibres, and is surrouiulecl hy a pericardial cavity that postoii- orly extends as a slit-like space between the cloacal and btnly walls far along toward the posterior end of the hotly. In its general features the nervous system closely resembles that of ('. erudita and C. attenuata. The labio-buccal system has been worked out in detail, but it is no exception to the statement just made. The gonad is of large size and is distended with spermatozoa that have made their way through wide canals into the pericardium. As usual the pericardial openings are situated close to the suprarectal commissure, and lead into clearly defined ciliated ducts which very soon unite with the glandular portion. In this species the glandular portion is at first relatively slender, and but little convoluted yet it soon enlarges greatly, becomes much folded and extends as may be seen in Plate 30, figs. 5, 7, from the posterior limit of the gonad to its opening into the cloacal cavity. Limifossor talpoideus Heath. Zool. Anz., 1904, 5, p. 2S. Zool. .Julirb. Ahth. An:it. Ontog., 1905, 5, p. 21. Several specimens of this species were taken in Alaska in the Lynn Canal (Sta. 4258) and in Chatham Strait (Sta. 42()4) at depths ranging from 282-313 fathoms. The general appearance of these animals is shown (Plate 10, fig. 1). The length ranges from 6-12 nun. and the diameter from 1-2 mm., the ratio 1 : 6 being constant. The mouth, almost terminal in position, is bounded by the sensory plates (Mundschild) and more dorsally by the type of sjiine covering the prothorax generally. The plates in life undergo rapid changes in form, but histologically and in their innervation they resemble their homologue in Chaetoderma. The deep semicircular groove (halbmondformige Grube) situated beneath the mouth and sensory plates, is lined throughout with the spiculose integument of the body. The spines are triangular or leaf-like and range in length from 0.02 nnn. in the region of the mouth to those about the cloacal chamber 0.38 mm. long. The hypodermis is relatively very thin, th(> boundaries of the cells indis- tinct and similar in general to that of other species of Chaetoderma. The somatic musculature likewise is very similar in the two genera. The mouth leads into a comparatively narrow canal with longitudinal folds covered with a well-defined cuticle. In the region of the radula the canal enlarges, develops a subradular pocket (Plate 10, fig. 4) and dorsally continues as a cir- cular tube to its junction with the stomach. Attached to a dorsal diverticulum numerous cells pour their secretion into the pharynx. A clearly defined subradu- 70 IJMIFOSSOR TAI.rOIDEUS. lar organ docs not exist and yet the fact that in the mid line tlie cells are more than usually high and slender and arc in close proximity to nerves from the labio- buccal ganglia indicates that the area exercises a sensory function. The radula and its supports and attendant musculature are enormously developed and indi- cate active predatory habits, but in every case the alimentary canal contains little besides a few diatoms, sponge spicules, and a small quantity of inorganic detritus. The radula with twenty-eight transverse rows is of the distichous type (Plate 34, figs. 3, 6), the long claw-like teeth being united while in the radula sheath by a clearly defined basement membrane. When freely exposed this membrane splits along the mid line and the teeth become located on each side of a deep cleft in the forward end of the radular supports (Plate 10, fig. 10). Odonto- blasts, in typical fashion, form the teeth which are subsequently enveloped by numerous enamel cells. The radular supports comprise two great masses of muscle and connective tissue which together form an ovoid mass grooved dorsally to hold the radula tube. To these numerous muscles attach that are in part responsible for the movements of the teeth. A detailed description of these and other muscle bands has been given in another place (Heath '05) and an attempt has been made to determine their functions. The stomach is sharply differentiated from the remainder of the digestive tract (Plate 10, fig. 4) and occupies practically all of the space between the end of the radular supports and the forward border of the gonad and digestive gland. Its epithelial lining is produced into a number of heavy folds that gradually blend with those of the oesophagus. In most cases the intestine leaves the pos- terior end of the stomach in the mid line, and immediately ventral to this union the liver opens by a single pore. This last named organ is relatively voluminous, filling much of the space beneath the gonad between the stomach and forward cloacal wall where it ends blindl>'. The intestine, of practically the same calibre throughout, makes its way by a fairly direct route to the front end of the peri- cardium. Here it bends abruptly downwartl and jxissing under the cloacal wall opens to the exterior in the mid line. A clearly defined connective-tissue septum bounds the head cavity pos- teriorly as in the Chitons. It passes immediately behind the radular supports and is penetrated by the alimentary canal, dorsal aorta, and pedal sinus. The pericardial cavity is of trihedral form and encloses a tubular and more than usually muscular heart without any distinct subdivisions. The aorta passes out from its forward border, and as a distinct tube with definite walls LIMIFOSSOR TALPOIDEI^S. 71 makes its way between the halves of the gonad to an opening in the septum bounding the head cavity. This latter space communicates with a well-defined pedal sinus, which perforates the septum and pursues its course posteriorly, communicating here and there with the general visceral cavity, to the neighbor- hood of the cloacal cavity. Here both sinuses unite on their way to the gills from which the blood passes above the dorsal gill retractor to enter the heart. The brain, clearly bilobed, develops fibres which unite with five pairs of precerebral ganglia that in turn give rise to nerves passing to the sensory plate. The lateral, pedal, and labio-buccal connectives unite before entering the brain. The last named are first to be cKfferentiated and holding the usual position at the sides of the pharynx, thej' unite with the ganglia lateral to the dorsal salivary glands. These nerve masses are united Ijv the usual commissure and by another passing dorsal to the pharynx in the neighborhood of the salivary glands. What appears to be a complete one passes ventrally into the neighborhood of the subradular organ. A nerve from each ganglion passes backward and probably innervates a portion of the digestive tract. The lateral and pedal ganglia, with the usual relations, extend to the region of the cloaca where they unite to form on each side a well-defined enlargement connected by a suprarectal commissure. From each swelling several nerves arise that are distributed to the cloacal and body walls; while from the commissure branches are developed, dorsally and ventrally, that innervate the ctenidia. The gonad extends from the stomach to the pericardial cavity into which it opens by relatively long and slender ducts. The coelomoducts have the foi-m of simple tubes extending from the pericardium to separate exits in the cloacal chamber. Their inner openings are situated in the infero-lateral angles of the pericardial cavity and are guarded by high pyriform cells devoid of cilia. On the other hand the succeeding portion of the canal, of very small calibre, is composed of cubical elements covered with a heavy ciliated coat. This division makes its way forward to the outside of the dorsal gill retractor and luiites abruptly with the glandular portion, which although a single tube is so con- voluted that it becomes a relatively voluminous structure. Its walls are com- posed of more or less cubical cells of which the cytoplasm is scant in amount owing to the presence of one or two great vacixoles. The general structure bears a fairly close resemblance to certain kidney tissue yet there is no positive proof that it possesses an excretory function. The outer openings are on each side of the anus a short distance anterior to it, and though very minute in pre- served material they are nevertheless clearly defined. 72 I'ACHYMKNIA A15VSSORUM. Limifossor fratula, sp. nov. This species is represented by two iiuUviduals taken off the coast of southern CaUfornia (Sta. 4369) at a depth of 260-284 fathoms. In general it so closely resembles the foregoing species that a very brief description will suffice. The body, slaty gray in color with a slight yellowish cast, is shorter and thicker than in L. tdlpoideus, and owing to a heavier body wall is much firmer. The spines of the two species are very similar in form, but in the present species they are of considerably larger size. Spicules from the middle of the body are 0.5 mm. in length while in L. ialpoideus the largest of the body do not exceed 0.38 mm. The hypodermis is also proportionately thick and what are probably matrix cells are frequent and sharply differentiated from the other elements of the hypodermis, and hence different from L. talpoideus in this respect. The digestive tract in the two species is, neglecting minor differences, built upon the same jilan. Heavy as is the radula and its supports in L. tal- poideus it is even heavier in the present case, and the teeth are of larger size, making it so difficult to section thein that at present there are no clear indications of their exact shape though it appears certain that the smaller cusp of each tooth is larger than in the preceding species. The muscles that operate the radula are typical but are unusually heavy. The nervous, circulatory, and reiaroductive systems are very similar in the two species. This species is readily distinguished from the foregoing by the size of the spines, the structure of the hypodermis, and the heavier musculature and con- sequent firmness of the body. Pachymenia abyssorum, sp. nov. One specimen of this species was dredged off the southern coast of Cali- fornia (Sta. 4397) in 2196-2228 fathoms, the greatest depth recorded for any Solenogastre. In bringing the animal to the surface the consequent decrease in pressure upon the body resulted in the active release of gases from the blood, causing the displacement of the cuticle to a considerable extent, the shrinkage of the hypodermal cells and the partial destruction of the foot at various points; otherwise the tissues are in a good state of preservation. The body is thick set, externally resembling Alexandromenia valida, and measures 27 mm. in length by 4.5 mm. average thickness. The color is a light yellowish white. As is indicated (Plate 39, fig. 4), the foot is exposed for a considerable distance, and is unusually broad and doubtless in life is capable of forming a relatively large PA( HV.MKMA ABVSSORUM. 73 surface possibly (Mialiliiig (he animal to crawl about on the bottom ooze. As in the case of .AJcxandromeniu this imlividual is unattached and may be accord- ingly a roving form. The cuticle is approximately three times the thickness of the hypodermis, but is scant in amount owing to the vast numbers of needle-like spicules, of varying sizes, imbedded in it. As noted above the hypodermal layer is not in a good state of preservation, but it may reatlily l)e discovered that the cells are unusually slender, and laterally and vcntrally form papillae in the head region. There are low elevations at other points over the body but it is not certain that they are definite papillae. The external opening of the anterior peilal gland is a cavity of large size whose walls are provided with folds of uiuisual height. On the posterior wall these are approximately seven in number, the outermost on each side being very large. Behind the cavity the five included folds disappear while the large lateral ones unite in the formation of a foot with a creeping surface of greater width than in any other known species of Solenogastre. In the posterior end of the body the foot decreases in size and becomes continuous with small folds of the cloacal wall. The anterior pedal gland is a voluminous organ lying at the sides of the body opposite the external outlet. The cells composing it are exceptionally small but otherwise present no noteworthy characters. Behintl it shades into the posterior pedal gland without any appreciable change in the character of the cells. Throughout the entire extent of the foot the gland is unusually large and the ductules appear to open over the entire creeping surface. The external atrial opening is subterminal and large, and leads into the customary cavity provided with i-idges and cirri typically situatetl. As may Ije seen in Plate 39, tig. 1, the external ridge is contiiuious across the mid line in front of the external opening of the atrium and though relatively small at this point it rapidly increases in height, iinally becoming of such a size that it may be seen in external view. The inner ridge is likewise small anteriorly but behind becomes as extensive as the external fold. Behind these two folds are con- tinuous with each other and are connected with several long plaits in the hypo- dermis which extend to the external opening of the pedal gland. The cirri are simple unbranched processes, slightly pigmented and contain a muscle or nerve fibre extending, in some cases at least, throughout their entire extent. The atrium communicates dorsally with the succeeding section of the digestive tract whose general relations may be determined from an examination 74 PACHYMENIA ABYSSORUM. of Plate 39, figs. 1, (i. The walls of this division are provided with numerous muscle bundles of irregular distribution between which are nniltitudes of glands staining actively when treated with Delafield's haematoxylin. The cells com- posing these glands are without distinct cell boundaries, are made up of vacuo- lated protoplasm containing droplets of various sizes and are grouped into lobules of various bulk. In many places they extend into the folds of the epi- thelial lining of the jiharynx and give evidence of opening through intercellular channels. Some distance toward the dorsal side of the animal a fold of large size ap- pears in the wall of the digestive tract which narrows the pharyngeal cavity to a relatively small tube. At this point the epithelial lining becomes thicker, a char- acter which it retains to the stomach-intestine, and the walls become surrounded by a heavy sheath of circular muscles to which vast munber of gland cells attach. These gland cells are grouped into slender lobules, and owing to the fact that they are much vacuolated their tint is fainter than in the case of those of the pre- ceding division of the tract. The nuclei also are of larger size and more distinct, but the secretion presents the same general ajipearance. A slender duct on each side of the pharynx (Plate 40, fig. 6) extends from the region of the ventral labio-buccal commissure (Plate 39, fig. 8), to its outlet (Plate 39, fig. ()). Poste- riorly each ends blindly and anteriorly is provided, as in the case of Alexandro- menia, with a papilla which is doubtless capable of being protruded into the pharyngeal cavity. Throughout its entire extent the ductules from these glands attach to the canal, but behind it they connect with intercellular channels and so pour their secretion directly into the pharyngeal cavity. The glands with this last named outlet present the same appearance as those communicating with the ducts except in the neighborhood of the stomach-intestine where they become more compact. No trace of a radula or radula sac exists. The pharynx or oesophagus projects for a great distance into the stomach- intestine which is provided with several longitudinal lidges instead of the cus- tomary sacculations. The middle portion of the body was not sectioned but as these ridges are present in the posterior end of the animal it is probable that they extend throughout the entire length of the gut. Many of these folds contain blood sinuses which often produce a marked distention. The lining epithelium is composed of more than usually slender cells many of which con- tain more or less spherical, granular masses. Posteriorly the intestine narrows, passes between the coelomoducts and opens into the cloacal chamber. No PACHYMENIA ABYSSORUM. 75 traces of food wore found in the tract and accordingly we arc without any knowl- edge of the animal's feeding habits. The walls of the cloacal chamber are provided with a number of slender outpouchings and to these are attached multitudes of gland cells grouped into lobules of different sizes. Each cell is pyriform and contains a somewhat granu- lar slightly vacuolated secretion that makes its way by a delicate ductule through an intercellular opening into a diverticulmn of the cloacal wall. The general arrangement of these structures is shown in Plate 39, fig. 2. The pericardial cavity is comparatively sjiacious and the contained heart, consisting of two divisions, is moderately muscular. The aorta in the present specimen is of small size but in its relations to the gonad and the anterior end of the body it is typical. Owing probably to the size of the foot the ventral sinus is large and connects in the usual fashion with the head sinuses and here and there throughout the body with the visceral sinus. In the posterior part of the body it divides, passes dorsally on each side of the intestine and after passing posteriorly for a short distance breaks up into a small number of lacunae which connect with the gills. From these organs the blood passes through rather ill-defined channels in the somatic musculature to the posterior end of the heart. Five or six pairs of relatively large folds appear in the cloacal wall riuming more or less parallel to the outer opening near which they are situated (Plate 39, fig. 2). Here and there these develop numerous minor wrinkles (Plate 40, fig. 7) which pass from one main fold to another or extend some distance over the cloacal wall. As usual they all contain blood sinuses but otherwise are not especially modified. The brain, imbedded in the numerous glands attached to the forward wall of the pharynx, is an unusually elongated structure and without distinct signs of being bilobed. From its anterior face the usual nerves, heavy in appearance, are distributed to the body wall and the ganglionic masses about the bases of the cirri. The lateral, pedal, and labio-buccal connectives arise from the ex- treme lateral boundaries of the brain and follow the usual course. A very slight enlargement marks the point of union of the lateral ganglion with the corre- sponding connective, while one of twice the diameter occurs in the case of the pedal cords. The last named structures are united at fairly definite intervals by clearly defined commissures and about the same number of connectives attach to the lateral ganglia. A nerve from the anterior pedal enlargement passes to the wall of the outlet of the anterior pedal gland, and two branches originate from a corresponding point on the lateral ganglia and applied to the 76 PACHYMENIA ABYSSORUM. souKitic musculature extend far iorwanl into the anterior end of the body. The lahio-huccal connectives, imbedded in the pharyngeal glands, pass backward about half the length of the pharynx where they join the ventrally placed ganglia (Plate 39, fig. S). These masses are in turn united by a strong commissure providcnl in its mid section with several ganglion cells from which a nerve arises and extends backward for a considerable distance attached to the pharyngeal glands. What appears to be a dorsal commissure springs from the upper side of the ganglia, passes dorsally and may be traced here and there amid the glands over the ilorsal surface of the pharynx. It has not been followed throughout its entire extent yet I have but little doubt that it is a definite commissure. From the posterior borders of each labio-buccal ganglion a nerve arises and imbedded also in the glands of the phai-ynx extends for a considerable distance posteriorly before it becomes lost to view. Finally it may be said that the arrangement of the brain and anterior jiortion of the nervous system is more regular than in the case of Alexandromenia for exanijile, but otherwise there is no fundamental difference. In the posterior end of the body the pedal ganglia continue to be united by commissures of large size and practically the same number of connectives unite them with the lateral ganglia placed high vip on the sides of the body. In front of the anterior cloacal wall the pedal cords liend dorsally and jirovided through- out with ganglion cells attach to the lateral nerve masses at the sides of the peri- cardial cavity (Plate 40, fig. 4). From the posterior end of the lateral and pedal ganglia nerves arise and extend backward along the body wall and in some places pass into the cloacal folds. A well-defined dorsal sense organ is present whose location is represented (Plate 39, fig. 2). Owing to the fact that it is of small size, that the cells like those of the hypodermis arc not well preserved and because of the obliciue direction of the sections its structure has not been accurately tletermined yet so far as the examination has gone it apjiears to conform to the usual type. The ovo-testis occupies the usual ]iosition on the dorsal side of the animal and extends forward to the posterior end of the jiharynx. The ova are unusually large and are surroimded by a chorion, but with these exceptions neither they nor the spermatozoa present any especially noteworthy featvu'es. The ducts leading from the gonad into the pericardium are of comparatively large calibre. In several respects the coelomoducts are remarkable structures and unlike those of any other known Solenogastre. In the vicinity of their inner openings DREPANOIMEXIA VAMPYRELI,A. 77 the pericardial wall becomes thickened and numerous small folds appear which converge and in some cases at least become continuous with the adjacent sec- tion of the duct. A short distance beyond the pericardial cavity and as far distally as the seminal receptacle each duct affords attachment for a vast num- ber of glands of unknown function. These are slender diverticula (Plate 40, fig. 2), composed of very small cells witii indistinct boundaries filled with a finely granular faintly staining secretion. In the present specimen large quantities of spermatozoa are present in the coelomoducts and frequently these have maile their way into some of the diverticula where they form masses without definite arrangement. Whether this is a normal occurrence or a past mortem effect cannot be definitely decided with the material in hand. Beyond these glands the dorsal section of the coelomoduct becomes thin walled, without folds and soon joins the ventral division which as Plate 39, fig. 2, indicates is of large size, thin walled, with few foUls and is crowded with sperms. Immediately ventral to the union of the dorsal and ventral linih of each duct there is a small globular outpouching to whose internal wall large numliers of spermatozoa are attached so that in position and function it is to be considered as a seminal receptacle. Sperms with the same mode of attachment are found in considerable numbers adjacent to the seminal receptacles and rarely at much greater distances, even as far as the undivided section or shell gland. The cavity of this last named organ is not much larger than that of the dorsal .section and its epithelial lining is relatively thin but a multitude of glands, attached tliroughout its entire extent, give it a heavy appearance. These glands are composed of compact, pear-shaped cells arranged in loljules that open by intercellular channels in the epithelial lining of the shell gland. As may be seen in Plate 39, fig. 2, the shell gland pushes inward the anterioi- wall of the cloacal chamlier so that its outlet is far within this last named cavity. Drepanomenia vampyrella, sj). nov. This species is represented by a single specimen dredged off the southern coast of Oahu Island (Sta. 3907) at a depth of 304-315 fath., where the tempera- ture was 43.7 F. It was coiled tightly about a solitary polyp of Epizoanthus, and further examination showed that the jiroboscis of the mollusc was protruded through the body wall of the coelenterate, whose reproductive and other tissues had been drawn into th(» alimentary canal f)f its captor. There is therefore no doubt that this species is carnivorous and that its association with the actinian is not an accidental one or a case of connnensalism. 78 DREPANOMENIA VAMPYRELLA. The body (Plate 2, fig. 2) measured 9 mm. in length, was slightly compressed laterally, })articularly its anterior half, and in form was somewhat spindle shaped, being largest about the middle section of the body and gradually tapering off toward each end, especially posteriorly where the body becomes quite slender before terminating in a truncated extremity. A well-defined keel extends along the entire animal in the mid dorsal line. The color was faint j^ellowish white. The cuticle surrounding the body is of medium thickness, measuring 0.35 mm. in the keel and 0.28 mm. elsewhere in the dorsal region, but gradually decreasing to half this amount on the ventral surface. It includes a single layer of radially directed spicules (Plate 32, fig. 6), ranging in size from those in the first stages of formation to others of the keel 0.129 mm. long. All are of the same general form represented in Plate 37, fig. 7. It is to be noted that many of the spicules are not in contact with the hypodermis, even the matrix cells having disappeared, but are situated far out toward the surface of the body. Beneath each developing spine are several cells apparently instrumental in its formation. The cells of the hypodermis are very small and indistinct and accordingly have been examined only superficially. The prevailing cells are slender with sul)central nuclei, and are developed into numerous small elevations, some of which connect by strands with the overlying papillae (Plate 32, fig. G). These last named organs contain upwaiils of eight cells in the enlarged portion; none appear to exist in the exceedingly slender stalk. As in other species of the family the ventral furrow commences close to the hinder border of the lip, and extending the entire length of the animal becomes continuous with the cloacal opening. Anteriorly it contains a relatively deep excavation into which the anterior pedal gland opens by the usual numerous intercellular ducts. This last named gland occupies jjractically all of the vis- ceral cavity between the region of the brain and tlie anterior end of the gonad. The cells comjiosing it are generally pyriform, with an average diameter of .021 mm. and are filled, save for the small compact nuclei, with a finely granular substance that stains intensely with logwood dyes. In some cases this secretion api^ears to be undergoing solution, and presents a vacuolated appearance, a character that is very pronounced among the cells of the posterior pedal gland. Tliese latter elements are related also to the foregoing in general form, size, and appearance, and extending to the cloaca and opening on each side of the foot, are thus seen to hokl the usual position. At the hinder Ijorder of the crypt into which the anterior pedal gland opens, DREPAXOMEXIA VAMPYRELLA. 79 the foot arises as a single, small prominence and rapidly assumes its fully devel- oped condition. Posteriorly it gradually decreases in size, disappearing, so far as may be judged from longitudinal sections, at a point immediately in front of the gonoduct openings. The general relations of the anterior section of the digestive tract are fully represented in Plate 7, fig. 4. As will be seen the atrial opening is subterminal in position and of medium size. The relations of the succeeding parts, while much the same as in Proneomenia, for example, are somewhat obscured by the protrusion of the pharynx. An outer ridge, composed of the usual type of col- umnar cell though apparently lacking cilia (Mundleist), is present. The inner ridge, probably related as in other species, has in the present specimen been carried out on the tip of the pharynx, an interesting fact as it indicates that these cells may be sensory, and of service in determining the character of the animal's food or sunoundings. Between these two prominences the usual cirrose area is present, the cirri being of relatively large size and unbranched. It has been suggested that the cirri, secreting a viscous material, may serve in the capture and retention of food. Here, however, is an animal killed in the act of feeding with its proboscis penetrating its host. The material drawn into the pharynx does not come into contact with the cirri, which in this case must certainly have some other function, though it is difficult to say what this may be. In its present extended condition the pharynx is relatively slender, almost wholly devoid of longitudinal folds and is relatively muscular. Heavy retractor muscles attach to the buccal wall and serve to withdraw the pharynx, that ap- pears to lack sjiecial retractors of its own. In some species of Neomenia the ventral salivary glands are described as being more or less coiled; in the present species this would probably be the case, but with the protrusion of the pharynx their openings into the canal have been carried forward until they are very close to the end of the proboscis. Each gland is unbranched, composed of excessively spongy cells and is possessed of a relatively large lumen, and a length of fully one fourth that of the hotly. In view of tlie fact that Drepanomenia has no radula it appears probable that the salivary secretion exercises a solvent action on the tissues of its victim, and the lifjuified material is then sucked in. The digestive tract in the present case is well filled with a finely granular substance in which one may recognize here and there the remains of cells, chiefly repro- ductive, belonging to its host. As may be .seen (Plate 7, fig. 4), the stomach-intestine extends forward a considerable distance in front of its union with the pharynx, thus forming an 80 DUEPANOMENIA VAMPYRELLA. extensive anterior coecinn. On this sack a number of short outgrowths are developed chiefly on the dorsal side. A short distance behind the front end of the gonad gut jiouches appear arranged with great regularity, and from this point on digestive cells attain their fully developed condition. In the region of the pericardium the sacculations vanish, the canal narrows rapidly and the liver cells are replaced by low columnar cells thrown up into longitudinal ridges extending to the opening into the cloacal chamber. In this specimen the pericardial cavity is relatively large (Plate 6, fig. 3), and the lieart it contains is considerably distended with blood, rendering it possible to some extent to determine the coiu'se of the circulation. The blood returning from the gills, and another smaller jiortion that appears to come directly from the hinder portions of the body, pours into a well-defined auricle situated at the hinder end of the pericardial cavity beneath the ventricle. Its walls are only slightly less muscular than those of the ventricle, and owing possibly to muscular contractions, are developed into several pouches that do not have the appearance of blood glands. From the auricle the blood passes into the ventricle through a comparatively large opening guarded by a well- developed muscular flaji jirobalily functioning as a valve. From the front end of the ventricle a clearly defined vessel arises, and passing forward unites with the dorsal aorta. This latter vessel holds its usual position between the l)ody wall and gonad, but it extends backward over the dorsal side of the pericardium as far as the posterior end of the ventricle. Anteri- orly the relations of the vessels in the gonad and of the aorta to the head cavity are essentially as they are in P. haivaiiensis. This appears to be the case also with the sinuses in other parts of the body, though using longitudinal .sections through the somewhat twisted body, it is not possible without much labor, to determine their connections accurately. With the jirotrusion of the jiharynx the brain has been carried some dis- tance ventrally, Ijut luuler ordinai'y circumstances its position and the relations of the nerves which it develops are probably not umisual. As Plate 7, fig. 4, shows three pairs of nerves pass to the atrial wall as in other of the Neomeniina, and are probably destined, here as there, to supply the cirri, anterior muscula- ture, and hypodermal sense organs. So far as may be judged from sections, the labio-buccal connectives originate some distance from the pedal and pallial, and may be clearly seen to jiass down to ganglia situatetl on the sides of the pharynx where it luiites with the buccal wall. From the hinder border of each ganglion a fibre originates that may be the inferior or vcntial commissure, but DRKPANOMEMA \AMPYRELLA. 81 owing to inniimerablo muscles it was not possible to determine this conclusively. Neither was it possil)le on this accoimt to determine if any subradular system exists. The relations of the pedal and lateral ganglia call for few special remarks. In some places it was possible to demonstrate pedal commissures, and to trace connectives between the jiedal and lateral cords, especially in the hinder regions of the animal where the cords are closer together. The most posterior connective is especially heavy (Plate 11, fig. 1), and develops two or three fibres whose branches may be traced to the musculature of the body wall. On the inside a very few exceedingly delicate nerves pass to the terminal section of the coelomo- ducts. From the posterior swollen section of the lateral cords {ganglion superior posticus) several nerves arise chiefly distributed to the body wall. The dorsal commissure is relatively' heavy and closely applied to the dorsal wall of the rectum. In the median line it gives rise to a nerve that may be traced to a point near the dorso-terminal jjortion of the body. In position it corresponds to the nerve supph'ing the dorsal sense organ in other Solenogastres described in this paper, but no such well-defined sensor}* area appears to be present in this species. In this species the paired gonad, without any special peculiarities, terminates rather abruptly at a point about as far forward as the hinder border of the atrial opening, and on the other hand passes by two relatively large ducts into the spacious pericardial chamber (Plate 11, fig. 1). From the lateral portions of a small recess at the posterior end of this cavity each coelomoduct arises, and after pa.s.sing downward for a short distance then passes forward, gradually increasing in size until it reaches a position about level with the hinder tip of the gonad. Here it bends abruptly and coursing backwaid unites with the one of the opposite side, and as a short common duct makes its waj' to the cloaca. Each canal commences its course with an epithelial lining essentially like that of the pericardial wall, being composed of low flat cells entirely devoid of cilia and lateral cell boundaries. These deficiencies are soon overcome, however, and there are evidences in some of the cells a short distance from the pericardium of a slight glandular activity. In proportion to the increasing diameter of the duct the cells show a greater width and height and the cilia become a strongly marked feature. This holds true for only a portion of the canal however for at a point slightly below the level of the lateral nerve the character of the lining changes abruptly. At this point the cells become high and columnar along the dorsal side of the tube and form a ridge, extending forward to the most anterior 82 PRONEOMENIA IIAWAIIENSIS. turn in the duct. The cilia with wliich this part of the tube is provided are probably operative in driving the sex products toward the exterior. There are no evidences that they ever foi-m a groove such as has been described in a few other species, and it must rather be supposed that Ixith sex ]ii-odies wliich are in contact with a stalked cell of the hypodermis. In the ca.se of the larger spicules of this character the sharp distal point may protrude freely above the cuticle. The anterior pedal gland is relatively voluminous, extending forward as far as the cirrose area, posteriorly to the front end of the oesophagus and filling nearly all of the visceral space between these two boundaries. As has been noted in other species of the genus this gland opens separately into a comparatively large space situated behind the mcnith opening and continuous with the front entl of the pedal furrow. Posteriorly this organ passes without a sharp line of demarca- tion into the pedal gland proper that holds the same relation to the animal as in the preceding species, but in bulk and in size of its cells it is probably twice as large. Behind the opening into which the anterior pedal gland pours its secre- tion the foot commences as a low median ridge that gradually assumes its full size though this is considerably short of that of the foregoing species. The opening of the atrium holds the same position and is of the same rela- tive size as in P. hawaiiensis (Plate 8, fig. 4). The ridges and the cirrose area are likewise very similar in the two forms. The ciliated ridges are not so high in this species owing possibly to the amount of contained blood and the cirri, somewhat more .slender than in the foregoing species, are united t)y their bases into groups of two or three. The junction of the atrial cavity and the pharynx is characterized l>y a ridge similar to that of the preceding species, but is not farther markod by the presence of a dorsal salivary gland. A very few relatively large cells are situated among the nerve fibres passing out from the brain, but while they are in the correct position for the unpaired gland n(j ducts have been discovered. The paired salivary glands present the same general appearance as in P. hawaiiensis. In the present specimen each organ extends from its opening at the sides of the for^vard end of the radula backward twice the distance from the front end of the animal to the opening of the salivary duct. Bej-ond this point the remaining portions of the body are missing. The cells are of the usual high columnar type and are more vacuolated than in any specimens of the preceding species. 92 PRONEOMENIA INSULARIS. After extending backward for about half its length the pharynx bends abrujitly upwards and unites with the stomach-intestine. In the angle thus formed is placed the radula that is considerably shorter than in the foi-egoing species. The teeth also are nnich smaller though of somewhat similar shape and judging entirely from longitudinal sections there are probably not less than twenty-eight transverse rows with at least twenty-four teeth in each row and possibly more. On the other hand the radular supports, in the foi'm of several transverse rods of compact connective tissue, are more highly developed. The relation of the pharynx and the stomach-intestine are sufficiently shown (Plate 8, fig. 4). The anterior coecum, the gut pouches, and the digestive cells are also essentially like those of P. hawaiiensis and require no comment. In connection with the circulatory system there are no unicjue characters. Every blood space was crowded with corpuscles, spherical in foi'm, with dense nuclei and cytoplasm, colorless after treatment with Delafield's haematoxylin, and containing several refringent granules. In this species the brain is of medium size, its greatest diameter equaling one eighth the average diameter of the body, and is situated some distance behind the union of the pharynx and mouth. As usual three pairs of nerves pass from its forward and latei'al regions to the front part of tlie l)ody, becoming lost to view in the region of the cirri or more laterally among the body nuiscles. The connectives passing backward are comjiletely ensheathetl in the anterior pedal gland whose granular substance rendei's it ^'ery difficult to follow them to their destination. It has been possible to trace the relatively large pedal fibres to the pedal ganglia, and the lateral connectives to their position along the sides of the body, but the labio-buccal connectives are exceedingly difficult to differen- tiate. However with the aid of high magnification their coui'se has been traced beyond question to the ganglia situated on each side of the pharynx about the level of the radular supports. Each is characterized by a much elongated spindle shape, the connective uniting with its anterior end and the ventral connnissure attaching posteriorly. This latter nerve arches over the dorsal side of the radula and otherwise presents the usual ajipearance. A dorsal buccal commissure, subradular system, and labial commissure were not found. Immediately in front of the radula is a ridge of columnar cells that may correspond to the subradular organ. In several sections it is also ]iossible to distinguish fibres that have the peculiar I'efraction of other undoubted nerves; j'et in spite of ]iersistent effort it has not been possible to determine their relations. Three pedal connnissures have l)een proven to exist and eight palliopedal connectives, all with the usual relations. DRIOMKNIA PACIFICA. 93 The paired Ronad cxtcnuls forward to a point slip;htly hohind tho level of the hinder end of the radula sui)poi'ts. Antorioi'ly its cc'lls are entirely male, niinnte o\a apjx'arinfi; only in the most ])ostei'ioi' i)art of the fragment. Driomenia pacifica, s]). nnv. Three individuals of this species were taken off the southern end of Japan, two fi-om Ose Zaki (Sta. 371()) at a depth of 05-125 fms. and one from Kago- sliinui (!ulf (Sta. 4935) at a depth of 103 fathoms. All were imbedded in a mass of hydroids belonging to the genus Sertularella. The body is of uniform diam- eter, measuring 1 mm. by 9 nun. in length. The color in alcohol was a yellowish white. No dorso-terminal sense organ has been discovered. The cuticle is thick and contains an innumerable number of hollow sjiicules of varying sizes but witli the general form represented in Plate 39, fig. 5. The usual hypodermal cells are not especially favorable for study, but on the other hand those of the papillae (Plate 38, fig. 10) are exceptionally clear. In the distal portion of each papilla the cells are spindle shaped, usually compact and finely granular, and appear in many cases at least to be continuous with a slender fibre which traverses the stalk and may be followed reacUly into the tissue beneath the h^ypodermal layer. Beyond this jxiint their course cannot be determined with certainty and accordingly there is no clear evidence as to whether they are muscle or nerve. The anterior pedal gland, occupying the customary position, is composed of the usual pyriform cells of average size. Posteriorly it passes without any change, except a decrease in the number of cells, into the posterior ])edal glanil which accompanies the foot throughout its entire extent. The outlet of the anterior pedal gland (Plate 38, fig. 1) is a plain walled, globular cavity heavily ciliated. Posteriorly two lateral and one median fold arise on its walls and soon unite to foiin the foot which continues to the cloacal opening though the two lateral folds become of small size. The atrial chamber, distinctly separated from the remainder of the digestive tract by a spiculose ridge, is a cavity witli walls fashioned into the usual sensory organs. The inner and outer ridges are moderately low, horseshoe-shaped ciliated folds neaily encircling the chamber and bounding the cirrose area. Each cirrus is very slender, without a distinct cavitj' and is usually united at its base with one or two others. The true mouth opening is l^orne on the summit of a low yet broad ])ro- boscis which is separated bj' a very narrow spiculose ridge from the outlet of 94 DItlOMKXIA I'AriFK'A. tlic anlcrior jicHlal gland. From external view this proboscis is not visible, the ventral furrow appearing to extend to the atrial chamber. The mouth leads into a slender tube, with irregular longitudinal folds, and becoming gradually larger as it courses dorsally it opens into the stomach-intestine. About midway it connects with two short ducts from the salivary glands wedged between the stomach-intestine and the anterior pedal gland in the general position rejiresented (Plate 38, fig. 1). Each salivary gland cell is pyriform, slightly vacuolated with distinct compact nucleus and unites with the main outlet by means of a very slender ductule. The anterior coecum and the adjoining section of the gut, is a plain walled tube lined with the usual vacuolated, gramilar digestive cells. More posteriorly lateral pouches of irregular form appear (Plate 38, fig. 1) and may be found at fairly regular intervals as far as the anterior end of the pericardium. Here the canal rapidly narrows, becomes circular in outline (Plate 38, figs. 7, 9) the epithe- lial lining is reduced in height and by a slender pore it ojiens into tlie cloacal chambei' whose form and relations are represented (Plate 38, fig. 2). It may be added in this coiuiection that the walls of the cloaca are devoid of folds, glandu- lar appendages, or modifications of any definite character. The pericardial cavity is of moderate size but in one respect differs from that of any other Solenogastre. Immediately behind the opening into the gonad the pericardial wall forms two latero-ventral outpouchings of considerable length (Plate 38, figs. 2, 7). The component cells are low^ columnar in form, without definite signs of glandular activity and contain relatively large dense nuclei. It is impossible to determine their function though they may be seminal vesicles since cells of the same general appearance compose the lateral jieri- cardial wall and connect these diverticula with the innei' ojienings of the coelo- moducts. The heart is of average size and consists of two distinct divisions. The anterior one, without any sharp boundary line is continuous with the aorta which, throughout its entire length, is a tube of more tlian usually great size. Its connections with the gonad and the anterior end of the body are normal as are those of the visceral sinus. Large blood spaces occur about the cloacal cavity and as the walls of the latter ai'e thin the exchange of gases may readily take place at this point. The blood corpuscles vary considerably in shape, in some cases being similar to the elongated type found in Strophomenia and at other times appearing almost globular. This may be a post morktn effect but the cells are very well preserved. DORYMENIA ACUTA. 95 The nervous system is difficult to trace and accordingly has been cxaniined in its broader features only which sliow it to l)e of the usual tyi)e. The sing;le s])ecinu>n examined is sexually mature and the reproduclix'e gland extends as far forward as the level of the salivary glands. While the eggs are attached as usual to the median wall of the gonad the s])erms develop in lateral pouches. In the posterior end of the body these crypts are of large size, extending in some cases far down the sides of th(> intestine, and they are connected with the gonad by small pores (Plate 38, fig. 2). The eoelomoducts arise as relatively small tubes from tlie posterior Ijortlei' of the pericardium and extend forward to the region of the seminal receptacle where as usual they unit(> with tlie shell gland. The lining epithelium is low, the cells cubical and ciliated and without indications of being glandular. Each seminal receptacle is a comparatively large club-shaped sac pro\id('d with several small outpouchings especially on its distal extremity. In these small ])ouches multitudes of spermatozoa are attached to the lining epithelium which is com- posed of slender columnar cells. The shell gland is in the form of a thick set Y and as may be seen, Plate 38, it contains a cavity of moderate size. The greater number of gland cells are of one type, high columnar elements containing, large numbers of sphei'ical granules. In the neighborhood of the opening into the seminal receptacle these are associated with a small number of cells containing, after treatment with Delafield's haematoxylin, a homogeneous violet colored substance. In close proximity to the opening into the cloacal chamber the dorsal wall of the gland contains a considerable number of cells, which secrete a coal black substance when treated with the above mentioned stain. All of these glandular elements are in contact with slender supporting cells containing mesially placed spindle- shaped nuclei. Dorymenia acuta, sp. nov. Eleven specimens of this species were dredged in the vicinity of the Santa Barbara Islands, off southern California, at depths ranging from 302-638 fathoms. The three largest specimens measure 35 nun. in length by 1.25 average diameter, with one exception having a thickness of 2.25 mm. The two smallest indi- viduals are 14 mm. in length by 1 mm. average diameter, and with one excep- tion, a slightly spindle-shaped individual, all of the specimens are slender and of about uniform diameter throughout (Plate 3, fig. 11). Their color varies from a grayish white, where the brick-red color of the liver shines through the cuticle, to light lemon-yellow. The head is rather sharply pointed, and but slightly 90 DOUVMENIA ACITA. differentiated from the body proper, wliieh ]iosteriorly terminates in a very ])(>iiited extremity as in I'roucdnienia webcri Nierstr. The atrial opening is rehxtively small, subterminal and surrounded by tumid lips whieh separate it from the ventral furrow. As noted in a succeeding paragrajih, the anterior pedal gland is highly developed, but its outlet is not especially modified externall.y. Posteriorly the pedal groove is continuous with the cloacal cavity. The cloacal opening is relatively large, venti'al and is overarched liy the posterior ])ointed end of th(> body, whose lateral margins are involuted, but may perhaps be flared occasionally to expose the genital s])icula, the appearance of the hinder end of the animal resembling at such a time Iclilhyonienia ichthyodes Pruvot. A well-developed dorsal sense organ (Plate 15, fig. 11), visible in sections only, is ])resent a short distance from the posterior end of the body, and is sup- plied with special nerves and blood sinuses as in P. Iimimiensix. The cuticle investing the body is relatively thick (Plate 33, fig. 4), and is developed })y a hypodermal layer in which the component cells are of small size. Those not instrumental in the formation of the spicules or pa])illae are more or less cubical in form antl consist of vacuf)lated cytoplasm in which the luicleus, usually spherical, holds a central position. At various points the nuclei are tlense and elongated and may possibly belong to sensory cells. The spicules are hollow needle-like structures (Plate 37, fig. 10), those of the alternate layers crossing the others almost at right angles. In their forma- tion no points of especial interest appear. As usual several cells take part in the process as in P. hawaiiensis. The papillae are of average number antl present the aj^pearance represented (Plate 33, fig. 4). Three or four s])indle-shai)ed nuclei occur in the slender fibrous stalk while from five to seven are present in the swolhni distal jioi-tion. In this last nametl situation the nuclei are frequently of two sizes, small dense bodies, and one or two of twice their size with a more vesicular ajipearance. Judging from many sections both the number and character of these elements are due to different stages in the development of the papillae. In advanced stages these latter organs may open to the exterior and become so filled with debris that the cellular elements save those of the stalk, become obliterated. This, however, is undoubtedly an abnormal condition and marks the close of an active functional existence on the ]iart of the papilla. The anterior pedal gland, as in various other species of Neomeniina, is a voluminous organ extending anteriorly as far as the brain, posteriorly as far as the forward border of the radula and filling jiractically all of the space between DORYMENIA ACUTA. 97 the gut and Hr' Ixjdy wall. The colls, where not compressed, are pear shaped with a diameter ranging from .OlSo-.OSol nun. In the early stages of their existence the cytoplasm is vacuolated and not affected with haematoxylin hut with the assumption of glandular activity th(> secretion, in the form of fine darkly staining granules, appears in the peripheral portions of the cell gradually filling the more central portions with the exception of the small, and at this stage, much shrunken nucleus. Delicate ducts, as usual, lead from the cell body to their intercellular opening into the anterior end of the pedal furrow. From external view the opening of the anterior pedal gland is not marked by any noteworthy peculiarity, but from sections it may be seen that the pedal groove soon expands inwardly into two extensive lateral diverticula (Plate 15, fig. 1), whose anterior walls, in some specimens, are thrown into low folds and more posteriorly are supplied with very heavy cilia, ranging from one to three times the length of the supporting cell. Along the median dorsal line a large fold exists which more posteriorly is continuous with the foot. Everywhere throughout this fold and over the anterior folds of each crypt the secretion makes its exit in the form of a very finely granular almost homogeneous substance and after treatment with haematoxylin of a slightly pinkish tint. The posterior pedal gland is also well developed and consists of a rod of cells on each side of the mid ventral line continuous in front with those of the anterior pedal gland with which they are identical save for their slightly smaller size. Posteriorly thej^ gradually diminish in bulk and number, and in the region of the cloaca finallj' disappear. The foot consists of very little more than a V-shaped epithelial fold, the included muscle and connective tissue being very scanty, and entirely devoid of blood sinuses or at all events those of sufficiinit size to include blood corpuscles in preserved material. Throughout its entire extent it is accompanied by two small epithelial ridges which are to ])o considered special modifications of the hypodermis. The inter-cellular openings of the pedal gland occur in the angle formed by these ridges and the foot. The atrial opening, holding the usual subtcrminal ])osition, leads into a cavity possessing essentially the same relations as in various species of Neo- menidae. Like Proncomcnia hawaiiensis, for example, there are two conspicu- ous ridges surrounding the cirrose area, and external to the outer buccal ridge a low elevation encircles the cavity save in the mid line jiosterioi-ly. From this elevation numerous delicate fibrils may be traced to a rod-like accunnilation of ganglion cells coextensive with the ridge itself. On the other hand these 98 DOIJYMENIA ACUTA. nerve cells (•(imiect by I'elatively large nerve bundles with the ganglia located near the bases of the cirri. Of the ciliated I'idges (Mundleisten) the more external are in the form of two elevations wliich approach each other very closely in front and behind, at which ])oints they become low and inconspicuous though in their mid section they are comparatively high. The inner ridge has the form of a horseshoe, the free extremities connecting posteriorly with the ends of the outer ridge. This inner prominence is relatively short yet high, and like the outer con- tains a loose mesh work of muscle and connective-tissue fibres among which are numerous lilood coriniscles. The cells composing them present much the same appearance as those of the outer low elevation described in the preceding jiaragraph. Practically all are slender and contain small amounts of pigment and elongated nuclei. Nerve fibres may be followed into the ridges which thus seem to be sensory. In P. hawaiiensis these cells are richly ciliated, but in this species all traces of cilia are absent and, it may be added, the material is excellently jireserved. Williin the cirrose area and lying behind the innermost ridge the atrial wall in the mid line is develojied into a fold, of large size, which is closely packed with multitudes of l)lood corpuscles. If the buccal ridges serve as respiratory organs, as some authors would have us believe, this structure is certainly more efficient since it is not only voluminous but its epithelial covering is not more than one third as thick as that of the general atrial cavity. The cirri are prominent structures in this species, being not only numerous but of considerable length and calibre. Each is composed of cubical or low cohunnar cells filled to a considerable extent with the usual yellowish pigment which more or less conceals the small centrally placed spherical nucleus. At various points these organs may arise singly from the buccal wall, but usually the bases of from four to six are fused, and into this stalk muscle and occasionally nerve fibres may be traced. The cavity of the single cirrus is usually so small that the relation of these fibres remains unknown and, it may 1)6 noted, efi'ec- tually blocks the entrance of blood cells, so that these organs are rather to be considered retractile sensory organs with little respiratory function. A shoit distance behind the cirrose area the pharynx originates as a tube with somewhat smaller diameter than the atrial cavity. However, immediately behind the region of the brain the canal from external view expands considerably, but sections of this region show that a great fold develops in the pharyngeal wall which it entirely encircles reducing the cavity to'a crescentic slit (Plate 15, DOUYMEXIA ACUTA. 99 figs. 1, 7). Numerous muscles inserted in the tissue of the fold and on the other hand to the body wall doubtless serve to dilate the canal when the animal is in the act of feeding. Inunediately behind this fold, and therefore in the region of the radula, the canal becomes much wider but more posteriorly again narrows and by a comparatively small opening connnunicates with the stomach-intestine. Throughout the entire extent of the ])iKuynx its epithelial lining is tludwn into numerous longitudinal folds, especially in the neighborhood of the raduhi whoie they become wavy and in sections present a most complicated appearance. In general the cells of the pharyngeal epithelium are low columnar elements dcNuid of cilia and overlaid with a clearly defined euticular layer. Two sets of salivary glands are present, a ventral pair and a group of cells imbedded in the large fold just mentioned. The cells of the last named gland, which is probably the homologue of the dorsal salivary gland of several other species of Solenogastres, are not grouped compactly as in P. hawaiiensis, for example, but are scattered throughout the tissue of the fold and ojien by separate pores over its entire extent. All the cells are i)yriform and in early stages are composed of a finely vacuolated cytoplasm in which the secretion ultimately makes its appearance in the form of distinct granules of comparatively large size. These rarely accumulate to a sufficient extent to hide the nucleus but make their way bj^ well-defined ducts to open by intercellular channels into the pharyn- geal cavity. The ventral salivary glands open into the pharynx at the sides of the extreme tip of the radula, and in the form of tubular outgrowths extend backward for a distance of 3.5 mm. As may be seen (Plate 15, fig. 2), the ducts are of large size and are bounded by slender cells densely filled with a secretion having nnich tlu> same appearance as that developed in the dorsal set. The radula is of the polystichous type, and judging wholly from sections consists of 48-51 rows with twenty-two teeth in each row. All the teeth are of essentially the same shape (Plate 34, figs. 7, 11) and size with the exception of those on each side of the mid line which appear to be slightly smaller. A very thin yet clearly defined basement nicnibiane is i^resent. The radula sac rests upon a support consisting of a series of globular cells of which two are of relatively large size and are located synunetrically on each side of the mid line. All of these contain nuclei and a finely granular cytoplasm, which in the larger cells is usually greatly slu-unken. To these supports numer- ous muscles attach, but from sections it is most difficult to describe their relations and define their function. 100 DORYMENIA ACUTA. The matrix cells of the radula are comparatively small and the teeth nuiucioiis so that an accurate determination of the development of the teeth and hascinciit mcmljranc is difficult. Odontoblasts, holding the ordinary position and with the usual appearance are present and numerous enamel cells arise from (he bottom of the radular sac. These latter elements may be traced forward a short distance amonf-; the newly developed teeth where they disappear. The colls responsible for the formation of the basement membrane are not unlike the odontoblasts all of which blend with the epithelial cells of the ventral side of the radular sac. In all essential respects therefore the development of the radula in this species is not unlike what exists in Limifossor talpoideus. Immediately behind the opening of the pharynx into the intestine the walls of the latter develop a circular fold (Plate 15, fig. 7) which in life may be less contracted and serve as a valve. Anterior to it the intestinal coecum extends as far forward as the brain. It is almost wholly devoid of diverticula, though its low epithelial lining is thrown into numerous small folds on its ventral surface. Slightly behind the valve-like fold just mentioned the intestine proper arises, characterized by diverticula of almost mathematical regularity lined by the usual high club-shaped digestive cells, except underneath the gonad where the intestinal epithelium loses its glandular character and its low columnar cells supp(jrt a coat of cilia. This state of affairs continues to the front end of the coelomoducts where the canal rapidly decreases in diameter and becomes ciliated throughout. Making its way to the dorsal side of the animal it passes between the pericardium and gonoduct to open into the cloacal chamber. The pericardial cavity is a comparatively large space whose general shape and lelations may be determined from Plate 6, fig. 4. The cells composing its epithelial lining are indistinct in outline, yet, judging from the nuclei, are more numerous than in any other species of Solenogastre described in the present paper. The heart also is of large size (Plate 15, figs. 4, 6) in both specimens which were sectioned, and very clearly consists of an auricle and ventricle. The first named division is much distended and the walls are thin and delicate, consisting externally of an epithelial sheet resembling that lining the pericardium, internally supjiorted by a few muscle fibres. These form a loose meshwork from which occasional fibres pass across the auricular cavity to be inserted elsewhere in the wall. The long, sharply defined median dorsal sinus, extending from the pos- terior end of the body, enlarges as it passes forward and enters the auricle on its posterior bolder. As usual the ventricular walls are of greater thickness and the spaces formed by the interlacing fibres relatively small and filled with groups DORYMENIA A( TTA. 10 1 of corpuscles. Anteriorly it passes into the aorta whicli, passing between the two widely separated ducts from the gonad, becomes a vessel of large diameter. The vessels to the gonad, and the exit of these into the visceral cavity and the communication of the dorsal sinus with the numerous channels in the head region are of the usual type. The sinuses of the head are relatively small yet may readily be followed through the anterior pedal gland and about the buccal wall to a small median ventral sinus situated above the outlet of the anterior pedal gland. Above the forward end of the foot this median sinus widens greatly, and during its journey to the posterior end of the body communicates here and there with the visceral sinus and at various points is divided horizon- tally by a muscular septum, thus forming two fairly complete sinuses one above the other. In the region of the coelomoducts these channels become sharply defined, though as they approach the cloaca, the ventral one, in frequent commu- nication with the dorsal, gradually diminishes and at the termination of the foot vanishes completely. The remaining ventral sinus has likewise greatly decreased in size in this same region, and communicating frequently with the outlj'ing visceral cavity disappears immediately in front of the cloaca. At the sides of the cloacal cavity and posterior to it the blood probably passes backward in the ventral half of the body, and by means of numerous channels passes into the dorsal half where it is transferred to the heart by several lacunae, the median dorsal one being most clearly defined. The corpuscles are about two thirds the size of those of Proneomenia which otherwise they closely resemble (Plate 35, fig. 13). It will be seen (Plate 6, fig. 4 and Plate 9, fig. 2), that the coelomoducts arise from the postero-lateral borders of the pericardium by relatively large openings, and extending forward as far as the anterior extremity of the heart connnimicate by a narrow canal with the conical seminal receptacle, and by a larger opening with the last section, of larger calibre, which unites with a corresponding tube of the opposite side and by a single median opening communicates with the cloacal cavity. Immediately beyond its inner opening the epithelial lining of each duct is thrown into prominent ridges composed of slender ciliated cells in which there are faint traces of glandular activity, which may possibly become more pronounced during the breeding period. Half way to the seminal receptacle the ridges disappear and the cells become lower, more cubical, and apparently are possessed of cilia. Furthermore throughout this same section of the canal some of the cells of its outer half become much elongated and form strands which 102 DOliVMENIA AtlTA. bridfiic tlic cavity. In the neighborhood of the seminal receptacle these are numerous hut what their office may l)e it is difficult to conjecture. 'rhc seminal receptacle is a thin-walled subconical sac provided with very slijihf internal folds and composed of low cubical cells which bear no trace of cilia nor secretory ])roducts. During the time it is filled with sperms, however, the cells become more or less liquified, the nuclei relatively large and pale after treatment with haematoxylin and great numbers of spermatozoa become im- bedded in their substance. It unites by means of a short narrow canal with the coelomoduct at the point wliere the first and second sections meet. Tlie second section of the coelomoduct, or shell gland, is a tube of compara- tively large size (Plate 15, figs. 4, 6). In the anterior third its walls are thin, almost devoid of folds and the cells composing it vary from cubical elements to others of low columnar form, if the animal be immature or out of the breeding season. As this last mentioned time approaches the cells become greatly thick- ened and the meagre secretion becomes abundant, filling the cell as a dark, almost black, substance like that of the muciparous gland in many molluscs. This condition of affairs continues along the mid ventral line of the duct for a consideral)le distance posteriorly. The same modifications occur in the suc- ceeding portions of the coelomoduct, but as the time of sexual activity approaches the cells become greatly elongated, are thrown into large transverse folds and are filled with a faintly yellowisli secretion which at other times is scarcely visible. In this species, as in P. wehcri, two genital spicula are present and of large size (Plate 6, fig. 4). Each is inserted in a deep sheath, a diverticulum of the wall of the cloacal cavity, which extends forward and slightly upward to a point about level with the base of the seminal receptacle. The cells of the distal extremity, which are probably the spicide-matrix cells, are very slender elements (Plate 15, fig. 10), with dense spindle-shaped nuclei imbedded in an almost homogeneous cytoplasm having somewhat the appearance of the odontoblasts of various molluscs. Throughout the remainder of the sheath, especially on its inner half, the cells are considerably smaller and their distal portion appears to be more or less cuticularized (Plate 15, fig. 5). Two powerful sets of muscles, the retractors and protractors, attach to the sheath (Plate 9, fig. 2). The first named consists of a large number of minor bands inserted in the distal end, and on the other hand to the body wall, after having spread out fan-like, a short distance anterior to the seminal receptacle. The protractors are more numerous and attach at various points within a narrow^ DORYMENIA ACITA. 103 zone iminediatcly behiiul tlic ii'tiactDis. On tlic otlicr liaiid, alter passinjr backward, the various groups of fibres become attached to tlie Ixxly wall or to the cloacal wall in the neifihborhood of the opening of the s))ieule sheath. Numerous other strands occur in tiiis region whose function it is to widen tlie cloaca, enabling the spicule to be exposed while others hiing aliout a counter movement. Those active in the first ojieration consist of many bands passing radially from the wall of the cloaca to become inserted in the body wall, and others which pass from the cloacal wall anteriorly to blent! with the somatic muscles. The remainder, responsible for the reduction of the cloacal cavity, comprise many fibres which attach to the walls of the cloacal cavity, and passing backward unite with the body wall on each side of the forward border of the cloacal opening. A pair of curious vesicles, irregular in foi-ni Init of comparatively large size, occur one on each side of the bod}' wedged in between the cloacal wall, spicule sheath, and coelomoduct and, as sections show (Plate 15, fig. 4), they are sepa-; rated from each other by a thin vertical wall. On the posterior face of each a short slender tube communicates with the cloaca (Plate 9, fig. 2). In immature individuals or those not sexually active the walls are comparativel}' thin, and are composed of cells cubical or low columnar in form without any distinct signs of glandular activity; but as the breeding season apin'oaclies the walls become much thickened and each cell develops some substance whicli gives it a longi- tudinally striated' appearance. This material remains unstained in haematoxy- lin and as it forms crowds the nucleus to the distal end. There is some evidence, though scanty, that this secretion is poured into the diverticulum and there becomes transformed into a darkly staining mucus-like substance which every- where lines the walls. Here and there are blood sinuses filled with corpuscles especially in the region of the opening into the cloaca where the cells are lower, without secretory products and covered with an abundance of cilia. The fact that these modifications occur simultaneously with those of the gonoducts strongly suggests that these organs in some way play a part in the reproductive process. They may function as uteri Init olniously such con- jectures are of very little value at the present time. There is some reason to believe that the type of coelomoduct found in Chaetoderma is more like that of Chiton, and accordingly of a more primitive type than in the Neomeniina where they are provided with seminal receptacles, glands often of enormous development and spicula, in some cases, provided also with glandular appendages. It is interesting to note that in an immature 104 DOIIYMKXIA A( TTA. condition sonic of these more modified types present a simpler condition than at a lat(M- stajje. In small specimens of the present species, about 14 mm. in length the various organs, connected with the reproductive system, hold the relations described above but they are far from being functional. The gonad, for example, is clearly paired throughout its entire length and the epithelium shows the merest traces of reproductive activity. The folds of germinal epithelium, that form a most characteristic feature of the adult organ, are commencing to appear on the latero-ventral surface of the gland, and there are slight evidences of a proliferation of cells on the inner wall of each gland. The dorsal aorta or sinus is of umisual size and in some places separates the two halves of the gonad completely, especially in front of the heart, where they are distant from each other by an interval ecjual to one third the iliameter of the body. Thus widely separated they open into the pericardium, which, as in the adult, is of large size (Plate 15, fig. 9). The heart likewise is typical. The ducts leading to the exterior are of essentially the same calibre throughout; the seminal receptacle terminates anteriorly in a relatively long flagellum-like process; the spicule sheaths have developed though there are no traces of the organic basis of the spicules themselves as in decalcified specimens of larger size; and the vesicles opening into the cloaca are both present though their outlets are relatively large. Above all there are no signs of glandular activity. As noted on page 169 if these ducts are in part excretory this phase of activity should appear long before sexual maturity and its absence indicates that these tubes are merely for carrying off sex products. In this species the sheath surrounding the nerve bundles is of unusual density or at all events stains with uncommon intensity in haematoxylin, so that branches not over 0.002 mm. have been followed. Owing to this fact more than usual care has been taken to determine the distribution of the more impor- tant trunks. The brain, holding the customary position, dorsal to the pharynx, is of medium size and very clearly bilobed. From its anterior half the usual three pairs of nerves arise and at their origin each is connected with two ganglia, one very minute in size. These fibres extend laterally and anteriorly and after branching several times connect with ganglionic masses about the bases of the cirri. A pair of very small nerves spring from the middle section of the brain close to the junction of its lateral and ventral surfaces. Each of these proceeds lat- erally and ventrally, and coming in contact with the sides of the pharynx branches and becomes lost among the numerous muscle fibres. DORYMENIA ACUTA. 105 From the posterior half of the brain the lateral, pedal and lahio-huccal connectives take their rise from independent, distinct roots (I'latc 13, fig. 1). In side view the first two appear to be relatively short but in reality they extend laterally for a considerable distance at the same time bending ventrally to join the corresponding ganglia. All are practically devoid of ganglion cells. At the junction of each lateral connective with the ganglion there is a well-defined en- largement which anteriorly gives rise to a strong fibre passing forward closely applied to the somatic musculature. In the neighborhood of the atrium it branches repeatedly and the resulting fibres connect in some cases, at least, with the ganglia in the neighborhood of the cirri. At the union of this cord and the lateral ganglion another nerve appears which likewise rests against the body wall, and after passing forward and downward becomes lost to sight after branching a few times. This same anterior enlargement develops one or two very small laterally directed nerves which soon become indistinguishable among the somatic muscle fibres. The lateral and pedal cords traverse the body holding the usual positions. Throughout their entire extent pedal commissures exist and approximately the same number of connectives unite the pedal and lateral ganglia. In the front end of the body, where the connectives are unusually distinct, they are often found to be united b}^ commissures and accordingly lack the regular arrangement sometimes seen in figures of other species. As seen (Plate 13, fig. 1) the anterior connective gives ofT a branch that passes forward and seemingly unites with the ganglionic rod of cells attached to the base of the outer atrial ridge. This appears also to be the destination of another nerve originating from the front end of each pedal ganglion. The labio-buccal ganglia are ellipsoidal bodies resting on the dorsal surface of the ventral salivary glands a short distance behind their outlet into the pharynx. From the anterior surface of each a strong nerve arises and in the usually con- tracted state of preserved material is considerably twisted before it expands and breaks up into three strong branches. Of these the more dorsal one is the buccal connective attached to the brain. The one immediately ventral to it extends anteriorly, slightly imbedded in the pharyngeal musculature until it arrives at the great dorsal fold. Here it bends sharply inward and deeply im- bedded in the muscle bands crosses the pharynx to unite with its fellow giving off one or two delicate fibres on the way. A second dorsal commissure is formed by two relatively small nerves each of which springs from the anterior face of the labio-buccal ganglion. These, in closer proximity to the mid line than the 106 STROPHOMENIA SCANDENR. others for most, of their course, become united al)ove the pharynx shghtly behind the dorsal pluiryngeal fold. At the juncition of the labio-buccal connective and anterior dorsal commis- sure a third nerve arises probably to be considered the subradular connective. A short distance distal to its origin it originates two, sometimes three, small nerves which may be tracetl deep into the pharyngeal musculature in the neigh- borhood of the radula. Still farther outward a small ganglion is attached to it by a small stalk and between these two bodies the main fibre continues to com- plete the commissure. In the posterior part of the body the pedal ganglia decrease rapidly in size, and become lost to view without being directly connected with the lateral cords. In this region the last four or five commissures are more than usually crowded together. The last three connectives on each side become united before enter- ing the posterior end of the lateral ganglia, which here break up into four strong branches that pass backward, and after dividing repeatedly become lost in the walls of the cloaca and body including the posterior elongation. At the point where these nerves arise the lateral cords are joined by means of the customary suprarectal commissure. Strophomenia scandens, sp imv. Three specimens of this species were taken attached to a colony of Acantho- gorgia armata dredged in the vicinity of Bird Island (Sta. 4156) at a depth of 286-568 fath. where the bottom was white mud and foraminiferous rock and the temperature was 45.8 F. The bodies of these animals were wrapped about the stems of the corals as shown (Plate 2, fig. 1), but none of the polyps in their immediate vicinity exhibited a shrimken appearance as though these molluscs had been indulging their appetites as in the case of Drepanomenia. The con- tents of the alimentary canal consisted only of a small amount of a finely granular substance. The largest specimen measured 39 mm. in length and 2.1 nun. in diameter; the smallest was 32 mm. long with a thickness of 1.6 mm. The two ends of the body are similar in appearance, the posterior being slightly more slender and pointed (Plate 2, fig. 1). In cross section the body is in general nearly circular, but in both of the specimens at hand the ventral surface is slightly flattened. The atrial opening is an elongated slit surrounded by rounded lips, behind which the ventral furrow commences and posteriorly is continuous with the doacal opening also subterminal in position. STROPHOMENIA SCANDENS. 107 The anterior portion of the venti'al fuiTOw forms a relatively deep depn^ssion (Plate 12, fig. 1) with corrugated walls, the opening of the anterior pedal gland. The gland itself occupies most of the visceral cavity in front of the forward end of the foot (Plate 16, figs. 1, 4). Its cells are of large size relatively, and are charged with a darkly staining secretion that makes its way by slender ducts to the ventral furrow, where each terminates in an intercellular opening. The secretion appears to be viscous, and in one of the specimens carefully dislodged it extended backward for a distance of 22 mm. as a narrow band. The cuticle enclosing the body is about 0.2 nnn. in thickness and as Plate 16, shows this measurement is very uniform save in the immediate neighborhood of the ventral furrow. Imbedded in its substance are, roughly speaking, six to eight layers of spicules, those from the Ijack and sides of the animal being represented (Plate 37, fig. 17). Among these is a much smaller number of radially directed spines that become more minute and more abundant in the neighbor- hood of the foot. The hypodermal layer is remarkably sharply dehnetl, and not including the spicule forming elements and those connected with the papillae, consists of cells about twice as high as broad with greatly vacuolated protoplasm and well- defined nuclei occupying a more or less central position. Occasionally more slender cells are encountered but these may in reality belong to the i)apillae. The development of the spicule is essentially the same as in P. Ii(nvaiicnsis (page 28) both as regards the number and arrangement of the operating cells which also retain their connection with the spine as long as it remains in the cuticle. As is shown (Plate 32, fig. 3) the pajMllae are numerous and closely crowded together at the surface. The expanded portion contains not far from twenty nuclei and the stalk from 2-5; otherwise there are no especial features of impor- tance. The foot arises in the extreme anterior end of the pedal furrow as a well- defined median ridge whose bounding cells are apparently covered with small cilia; but these are usually obscured by the huge cilia of the cells situated lat- erally. In the anterior portions of the vential furrow the secretion from the anterior gland passes through intercellular openings at all points; more poste- riorly it passes through the foot and the epithelium in immediate contact with it. The secretion of the anterior pedal gland when treated with Delafield's haematoxylin contains one substance which stains almost black and another of light blue tint. These appear to be two distinct secretions, for the dark sub- 108 STROPHOMENIA SCAXDENS. stance escapes high up on the sides of the ventral furrow while the light blue product passes out more ventrally. Behind the opening of the anterior pedal gland each side of the foot is accompanied by a longitudinal fold, which persists to near the hinder end of the animal when the lateral ridges, decreasing in size, pass into the general hypo- dermal covering of the body and are covered in large measure by the cuticle in which small spicules may be formed. The atrial opening leads into a well-defined cavity (Plate 12, fig. 1) whose walls are provided with organs not unlike those in P. haivaiiensis for example. The outer sensory prominence and the outer and inner ciliated ridges are not so elevated, but the last two are composed of the same type of ciliated cells. At certain points the outer sensory ridge surmounts groups of ganglion cells from which fibres may occasionally be traced into close proximity to nerves that inner- vate the cirri. Along the median dorsal line another well-defined elevation is present between the two limbs of the inner ridge, but its cells are like those of the ordinary buccal epithelium and are probably not highly sensory. In cross section the pharynx is roughly semilunar in shape, appearing rela- tively narrow when viewed from the side, but with a diameter of one fourth that of the body when seen from the dorsal or ventral surface. In front of the open- ing of the anterior pedal gland its cells are relatively high and slender, finely granular and contain more or less spindle-shaped nuclei placed at different levels. Behind this point the lining retains the same general character, but is developed into numerous and relatively low transverse folds supported by a small amount of connective tissue. Slightly in front of the labio-buccal ganglia the ridges become higher, the underlying connective tissue more abundant and among the cells of the usual type are a few others of more slender appearance with darkly staining elongated nuclei. It is possible that these elements are sensory in function, but owing to the large quantity of muscle and connective tissue in the neighborhood no special nerve supply has been distinguished. Beyond these supposed sensory ridges the digestive tract bends abruptly upon itself and coursing upward and forward unites with the stomach-intestine (Plate 12, fig. 1). There is reason to suppose that this section between the labio-buccal ganglia and the gut represents the oesophagus, but with the excep- tion that its epithelial lining consists of more spongy cells not produced into folds there is little to distinguish it as such. The relations of the paired ventral salivary glands (Plate 6, fig. 6) are some- what peculiar and except in very well-preserved material are difficult to determine, STROPHOMENIA SCANDENS. 109 a fact that may be responsible for some of the remarkable relations of these glands as described in some other members of the genus and in Rhopalomenia. In the present species these organs are placed side by side beneath the intestine and extend backward from their outlet for a distance equal to at least four times the diameter of the body. Anteriorly they diverge and enter the pharyngeal wall almost directly above the labio-buccal ganglia. Imbedded in muscle and connective-tissue fibres and some of the outlying pharyngeal glands, each canal now bends sharply upon itself (in two individuals) and becoming much more slender courses downward and somewhat backward to enter the pharynx on its ventral side. From the bend to a point close to their outlet these ducts are not only relatively slender but their lumen is of very small size and save in excellently preserved material, as in one of the specimens, is scarcely to be distinguished from the surrounding tissue. They are also composed of remarkably delicate material for in one specimen in which the organs are in a fairly good state of preservation this section of the gland has macerated and disappeared completely, leaving only the short and firmer outlets into the pharynx and the free main part of the gland. On the ventral side of the pharynx the glandular portion of each of these organs ceases abruptly and becomes continuous with a canal of somewhat larger calibre composed of small cubical cells, that proceeds almost directly inward and opens ventrally into the pharynx close to the mid line. This terminal sec- tion of the salivary glands is a conspicuous object in sections, but its connection with the adjoining glandular part may be readily overlooked in which case the terminal section appears to be a short diverticulum of the ventral wall of the pharynx while the gland proper seems to unite with the pharynx near the dorsal side. In the present species there is in addition to the various divisions of the salivary glands a median diverticulum whose relation to the digestive tract and buccal commissure support the belief that it is a rudimentary radula. The ventral wall of the pharynx forms a shallow pocket and the salivary ducts open into the bottom of this at each side. Exactly in the mid line and immediately above this pouch there is a small posteriorly directed diverticulum of the ventral wall of the pharynx and in two examples the buccal commissure passes dorsal to it (Plate 6, fig. G), the relations being the same as in the radula of P. hawaiiensis for example. Its epithelial lining consists of low cells composed of highly vacu- olated cytoplasm in which the nuclei hold a basal position, but there is no sign whatever of teeth. 110 STROPHOMKNIA SCANDENS. CorrelaU'd with (he jjosterior attachment of the pliarynx is the excessive devclopiHcnt of the anterior coecum (Plate 12, fig. 1), that extends forward to file hiniler limits of the atrial cavity. In one specimen its walls are lined tlii-o\ighout with pyriform digestive cells, whose clear basal portion holds a small compact nucleus while the vacuolated distal part contains numerous granules characteristic of this type of cell. In another individual treated in precisely the same way the granular jiortion of the cells was absent as was the case in other parts of the digestive tract. As was noted in the case of P. hawaii- ensis this appears to be the normal method of ridding the cells of their secretion. In the hinder jiortions of the coecum gut pouches appear and becoming more fully developed a short distance more posteriorly they continue with much regularity to the region of the heart where they abruptly disappear. Beneath the anterior end of the pericardium the intestine continues as a tube of relatively large calibre, but at the posterior end it rapidly narrows down to open into the cloaca (see Plate 12). As is represented (Plate 13, fig. 2, and Plate 16), the pericardium of this species is of considerable size and contains in addition to the heart a number of mature ova. Histologically the differentiation of auricle and ventricle is not clearly marked and save for a constriction there is nothing to distinguish these two divisions. The blood occupying the interstices between all the organs in the posterior part of the body passes into the auricle by a wide opening and thence pours into the ventricle through an aperture not guarded by clearly defined valves. From the forward end of the heart the aorta takes its rise and with the usual position makes its way to the head after having supplied the gonad. The lilood spaces at the anterior end of the body are very limited, more so in fact than is indicated in the figures which have omitted the intrinsic muscles of the digestive tract as well as retractors and protractors that attach to the body wall. However, the course that the blood takes in passing through the body projier differs in no essential particular from that of P. hawaiiensis. The corpuscles are spindle shaped (Plate 36, fig. 12) and the small densely staining nucleus is superficially placed, in some cases being elevated above the surface of the slightly yellowish homogeneous cytoplasm. In a preceding paper (Heath '04) attention has been called to the fact that in its more important details, especially in the relations of the labio-buccal sys- tem, the nervous system of this species shows a striking similarity to the Chitons and certain prosobranchs. The brain is relatively small and holds the usual position on the dorsal side of the jiharynx behind the mouth cavity. As usual STROPHOMENIA SCANDENS. Ill three pairs of nerves arise from its anterior borders and are distributed to the various structures of the atrial and body walls. From its lateral margins the lateral, pedal, and labio-buceal connectives arise side by side. The lateral present the usual appearance as is also true of the pedal whose union with the pedal ganglia is marked by a well-developed enlargement. In nearly all cases if not invariably the labio-buccal ganglia of the Solenogastres are located near the openings of the paired ventral salivary glands when these exist, and since in the present speeies these are situated far back the connectives imbedded in the muscular pharyngeal wall are characterized by a relatively great length. The ganglia imbedded in the wall of the pharynx give rise to three commissures and to the subradular connectives. The dorsal and ventral buccal commissures are sufficiently indicated (Plate 6, fig. G) to require no farther comment. In one specimen, possibly both, a well-developed ganglionic enlargement occurs on the ventral commissure and may correspond to the buccal ganglion, the other larger ganglion with which it is in close proximity and from which the subradular system is connected representing the labial. From this last named nerve mass another commissure passes ventral to the alimentary canal and in its course gives rise to two nerves which pass backward for a short distance and then become lost in the tissues of the pharynx. As has been noted a typical subradular system is present in P. hawaiiensis. In Limifossor talpoideus it holds the customary position, but ganglia are appar- ently entirely absent. In one specimen of Slrophomenia scayidens very small masses of ganglion cells seemingly represent the subradular ganglia of which no sign exists in the other specimen. In this species no well-defined sul)radular organ occurs hence the ganglia are perhaps in a state of degeneration. Owing to the comparatively small size of the latero-pedal and pedal com- missures they are not readily followed yet in certain places they have been traced in dissections and sections, so that so far as may be judged they exhibit no unusual features. In the postei'ior part of the body the relations shown (Plate 13, fig. 2) exist. As there indicated the lateral and pedal ganglia are united by two strong con- nectives, and postei-iorly give rise to several branches that pass backward and become lost in the somatic musculature. From the middle of the suprarectal commissure a nerve arises that has been traced to the base of the dorso-terminal sense organ. This last named structure (Plate 32, fig. 9) is protruded above the general level of the cuticle. As indicated the cells are slender, naked, and rest upon a small accumulation of what are probably ganglion cells. Muscle 112 STROPHOMENIA OPHIDIANA. fibres attaching to (lie base, are probably retractors, the pressure of the blood in (he uiuleriying sinus being responsible for the protrusion of the organ. The hermaphrodite gland holds the usual position and anteriorly extends to a sliort distance behind the level of the union of the oesophagus and gut. In all the specimens the sex elements are fully formed, some of the large ova having been dehisced and carried into the pericardial cavity. Posteriorly fhe gland gradually decreases in size finally passing into two long slender canals that unite with the front end of the pericardium (Plate 13, fig. 2). This last named cavity communicates posteriorly with the coelomoducts that extend backward a short distance before proceeding forward. At first the epithelial lining of these canals consists of low cells similar to those bounding the pericar- dium, but these are soon replaced by others almost cubical in form supporting an abundance of long cilia that continue to the seminal receptacles. These latter organs consist of from fifteen to eighteen sacs attached by short ciliated stalks to the gonoduct. In every case they were completely filled with sperma- tozoa. The section of the coelomoduct, extending backward from the seminal receptacles, is lined throughout with relatively high columnar cells filled with a darkly staining granular secretion. Immediately back of the seminal receptacles these are developed into several longitudinal folds that quickly disappear more posteriorly. Still farther backward the two ducts unite a short distance in front of their outlet and in this single canal the internal folds reappear and persist to the cloacal cavity. Immediately ventral to the outer opening of the reproductive system is a short diverticulum of the anterior wall of the cloaca (Plate 13, fig. 2, dt). Its cells are cubical in form and essentially like those of the cloaca at this point, but there is no sign of spicules or any secretion that indicates its possible use. Strophomenia ophidiana, .sp. nov. One specimen of tins species (Plate 1, fig. 2), attached to an alcyonarian coral, Acanlhoyuryia angiisliflora, was taken off the southern end of Honshu Island, Japan (Sta. 3755) in water 52-77 fath. in depth. It measured 43 mm. in length and 2.5 mm. through the thickest part of the body. The color is ci-eamy white shading to very light l^rown in the neighborhood of the head. A well-defined dorsal sense organ is present of the usual type. Tlie cuticle is relatively tliick, measuring on an average 0.19 mm., and rests upon a thin hypodernial layer whose constituents present no especially note- STROPHOMEMA OPHIDLVNA. 113 worthy features. The papillae are fairly numerous, especially ventrally where they project somewhat above the external surface of the body. Their shape and general character are represented in Plate 33, fig. 9. Small yellowish gran- ules are scattered throughout the cuticle, but of their origin or function it is impossible to make anj- definite statements. The spines are shown (Plate 36, fig. 17). The sensory atrium is exceptionally small (Plate 8, fig. 5) and the two outer ridges appear to be lacking altogether. The inner ridge, on the other hand, is clearly developed and typically placed but is of limited extent. The cirri in this genus are united as usual in groups of 2-5 and are composed of the customary cubical pigmented cells. Posteriorly the atrium opens into the buccal-pharyn- geal tube, a long slender structure of about the same calibre throughout. Its lining is composed of slender epithelial celfe produced into longitudinal folds especially in the anterior half. External to the epithelium is a well-developed layer of circular muscles and in contact with this a sheet of longitudinal fibres which become more abundant in the neighborhood of the stomach-intestine. In its anterior fourth this section of the digestive tract is attached to a consider- able number of fibres which extend more or less radially to the body wall. Pos- terior to this region gland cells, arranged in small pear-shaped groups, are imbedded in the longitudinal muscle fibres and by intercellular canals open into the gut. The beautifully regular arrangement of these organs is shown (Plate 17, fig. 13). Ordinarily the stomach-intestine connects with the posterior end of the pharynx or oesophagus, but in the present species such is not the case, for this junction is considerably in front of the ventral salivary glands which are append- ages of the pharynx. However the ventral wall of the gut from the dorsal intestinal coecum to the openings of the salivary ducts and for an equal distance posteriorly is clearly pharyngeal in character, lacking the hepatic cells but pos- sessing the characteristic muscle layers. The ventral salivary glands are relatively long tubular structures penetrated eccentrically by a thin epithelial tube through which the outlying glandular cells open. Close to the outlet into the pharynx this glandular portion disap- pears, and the delicate duct makes its way through the ventral side of the pharynx to open into a small cul de sac. No radula is present. The extent and relations of the anterior intestinal coecum are well repre- sented (Plate 8, fig. 5) and the intestine conforms so closely to the usual type that 114 STKOPHOMKNIA OIMIIDIANA. it demands no doscription. Posteriorly the gut narrows rapidly, becomes some- what rectangular in cross section as it passes between the limbs of the shell gland, and shortly before its outlet in the cloaca develops moderately high folds. The pericardial cavity is of very large size (Plate 18, figs. 2, 3), and the con- tained heart is of the usual greatly elongated type. There are no distinct signs of a division into uurich- and ventricle though a valve-like flap near its anterior end may indicate such or possibly the commencement of the aorta, which for a considerable distance is of as great diameter as the heart itself and even in the head region ])itli(>lial lining is usually fashioned into low longitudinal folds and rests upon a circular muscle layer of moderate thickness external to which are a few longitudinal bands 124 STROPHOMENIA SPINOSA. and radial fibres attaching to the body wall. As may be seen in Plate 8, fig. 2, numerous pyriform masses of cells attach to the pharynx throughout its entire ext(>nt and pour the secretion into the digestive canal through iuun(>rous inter- cellular openings. The ventral salivary glands are long tubular bodies with a very slender duct through wliich the attached gland cells pour their secretion. Their openings into the pharynx are exceedingly narrow but occur in normal position at the sides of the radula (Plate 17, fig. 4). The long dorsal coecum, the fairly regularly pouched mid gut, and the relations of the rectum are all typical. The radula is evidently in a highly degenerate condition, and differs con- siderably in different specimens. In the individual represented (Plate 34, fig. ter, and open into depressions on the pharyngeal wall on each side of the radula (Plate 18, fig. 9). A thin epithelial lining borders the lumen while the outer surface is in contact with gland cells, pyriform and filled with a secretion difi'ering considerably in different specimens and parts of the same gland. .\t some points the cells are closely packed with a bluisli or innkish secretion or at others this material is collected into roundi'd i)articles, dark Itrownish yellow in color, surrounded by a vacuole of considei-able size. The radula, typically located, seemingly belongs to the distichous type, and yet is r(>adily related to the polystichous form occurring in other genera if we assume that the bases of the once distinct teeth have secondarilv fused. STROPHOMENIA TRIANGULARIS. 127 From transverse sections it appears that there are fifteen rows of teeth whose general fmin is shown (Phite 34, fig. 4). They are thin and dehcate, or at all events are not easily displaced in sectioning, and stain readily in haematoxylin. As usual with this genus the anterior intestinal coecuni is of great lengtli and considerably in front of its union with the pharynx exhibits most of the essential characters of the stomach-intestine, possessing fairly regular out- pouchings and an epithelial lining of the customary glandular type, but lacking the cubical ciliated cells beneath the gonad. The coecum and most of the .suc- ceeding portions of the gut contain nettle and gernr cells extracted from the host. Opposite the seminal receptacles the intestine narrows, becoming gradually smaller until it opens into the cloaca (Plate 18, fig. 10). .\t the same time the dorsal ciliated ejiithelium gradually extends round the sides of the rectum, finally meeting in the mid ventral line shortly- in front of the anal opening. The pericardial cavity in this species is of very large size and the heart of unusual length. The posterior division, however, is unusually small and peculiar in being paired, save at its junction with the ventricle with which it communi- cates by a single small pore apparently furnished with a valve. The aorta, at its origin, is of the same calibre as the ventricle and occupies the entire though narrow space between the halves of the gonad. The branches, passing from it around the ventral and lateral surfaces of the gonad to unite with the visceral cavity, are likewise of large size and very distinct. More anteriorly these branches become much smaller and in the region of the head all but disappear in the present specimen. The course of the sinuses in the head, their union with the ventral sinus and the relation of the latter vessel with the heart are typical. The corpuscles are more than usually compact, but in their ellipsoidal form resemble those of other species of the genus. The gonad extends as far forward as the rackila and presents the usual features. Posteriorly it opens, in both specimens examined, by unusually large ducts into the pericardium, owing possibly to the fact that the time the animals were killed ova were present in considerable numbers in the pericardial cavity and along thi^ cloacal passages. These canals, arising typically from the poste- rior end (if tiic pericardium (Plate 9, fig. 3), average approximately 0.095 mm. in diameter and are lined with cubical and low columnar richly ciliated cells. Toward the median line of the bod,y these elements are more glandular and the secretion may direct the course of the sperms. The seminal receptacles, numbering 10-12, vary in size as may be seen (Plate 9, fig. 3), and possess unusually long stalks. Both vesicle and duct are 128 LOPIIOMENIA SIMRALIS. coiiiiiosocl of cubical colls, those of the latter being twice the height of the others. Ill the receptacle spermatozoa are numerous and are attached by their heads to the wall. The anterior half of each ventral section of the cloacal passage (the shell gland) is composed of cells of the appearance represented (Plate 18, fig. 8), filled with a distinctly granular, darkly staining secretion. In the remaining half the api)earances are much the same save that the glandular material is of a pink or reddish color. The halves of the shell gland open into the cloaca by separate pores and a fold, distended with blood, separates in large measure these openings from that of the intestine. The nervous system is typical in its general features. In the labio-buccal system two commissures unite the ganglia ventral to the pharynx; one of these bears a pair of small ganglia as in Strophomenia scandens (Plate 6, fig. 6). One conunissure likewise passes dorsal to the pharynx. The relations of this system are essentially the same as in Plate 6, fig. 6, with the exception of the most l)osterior ventral commissure that appears to be lacking in the present species. Lophomenia spiralis, sp. nov. This species, represented by two specimens (Plate 2, fig. 4), was taken in the vicinity of Niihau Island (Sta. 4176) at a depth of 537-672 fath.; bottom, gray sand and foraminiferous mud; temp. 38.3 F. Both individuals were closely wrapped about the stalk of a hydroid colony (Cryptolaria operculata Nutting) and sections disclose the presence of nettle and other cells in the alimentary canal from which we may infer that these forms, like Drcpanomenia vampyrella, prej' upon the polyps. Both specimens were of almost identically the same shape and size, measuring approximately 24 nun. in length and 1.5 mm. in average diameter. Each end of tile body terminates in a fairly sharp point but as the mouth is subterminal and surrounded by well-developed lips it may readily be distinguished from the cloacal opening that extends a short distance on to the dorsal surface. The ventral furrow holds the usual position, being continuous with the cloaca poste- rioily and in front terminating immediately behind the atrial opening. In one individual where the pharynx was slightly retracted this groove appeared to be directly continuous with the atrium but sections prove conclusively that such is not the case. The opening of the pedal gland as usual occupies the anterior end of the ventral furrow. Its position, shape, and general appearance are accurately rep- LOPHOMENIA SPIRALIS. 129 resented (Plate 6, fig. 5). The hypodermal cells bounding the cavity are higher and more slender than the usual type and are richly ciliated. All contain spindle- shaped nuclei and lightly staining cytoplasm. Immediately behind the outlet of the anterior pedal gland the foot arises and extends to the opening of the cloaca. At the outset it is well developed but gradually decreases in size posteri- orly until near the cloaca when it enlarges to twice its average size. A short distance behind its front end the hypodermal cells lateral to the base cease to develop spines and assume the form of a ridge (Plate 19, fig. 3) that increasing slightly in size continues to the cloaca where each like the foot enlarges scjmcwhat before disappearing. At all points the cavity of the foot is very small, never spacious enough to permit the entrance of blood corpuscles. The relations of the anterior section of the digestive tract are shown (Plate 6, fig. 5). Here the pharynx is somewhat protruded, but it is evident that the structures borne on the buccal wall are not unlike those of the other Neomeniina hitherto described. Immediately within the lips the usual ciliated atrial ridge is present. However it is relatively short, its contained blood sinuses small and the component cells low and very slightly pigmented. This is also true of the inner elevation with the exception that in both specimens it is of somewhat greater height. The cirri of the included area are relatively slender and are usually attached separately to the buccal wall. Their cavities are of extremely small calibre and it is onlj- in exceptional cases that the contained nerve fibre may be determined. Also their cell boundaries are indistinct, but (otherwise these organs are not unlike those in Proneomenia hawaiiensis. The pharynx consists of two distinct portions, the first a relatively slender tube leading from the mouth into the second enlarged section that opens in turn into the stomach-intestine. Throughout the epithelial lining consists of columnar cells, forming at first low longitudinal ridges that graduallj^ increase in height attaining their maximum size at the opening into the stomach. At the junction of the first and second sections the jiharyiix is produced into a much folded diverticulum that affords an outlet for the dorsal salivaiy gland (Plate 6, fig. 5). This last named organ is relatively very voluminous, larger in fact than in most species of Solenogastres hitherto described. With the exception of two narrow lateral areas held by the anterior pedal gland it occupies the dorsal surface of the digestive tract, and to a nuich less extent, the lateral borders between the brain and stomach-intestine. In the most favorable sections the component cells are clearly pyriform and are connected by a narrow duct filled with granules that leads to the pharynx. Some of the smaller cells 130 LOI'IIOMENIA SPIRALIS. arc fairly well filled with small particles which in the larger cells are applied to the cell inenibranc, the remaining contents consisting apparently in life of a fluid. In the majority of cases the nucleus is spherical, finely granular and contains a distinct nucleolus. The position of tiic radula is shown in Plate 6, fig. 5, while a cross section of the radula tube is represented (Plate 34, fig. 1). From these it will be seen that this organ l)eioiigs to the distichous type and that there are about twenty rows of teeth. While the shape of each tooth is sufficiently shown in the figures it is worthy of note that in the radula tube the base of each tooth is connected by a narrow cuticular bridge. When the radula opens out into the pharynx each plate ai)pears to split in half, at all events the exposed teeth in one specimen at: least are fully three times farther apart than in the radula tube and they do not app(>ar to be connected by a basal plate. Immediately in front of the radula there is a pair of short diverticula of the ])liaryngcal wall which serve as outlets of the salivary secretion. The mass of the outlying gland comprises several divisions bounded by connective tissue and located chiefly at the sides of the pharynx. The cells themselves are pyri- form, highly vacuolated and their ductules attach chiefly to the dorsal side of the main duct. The stomach-intestine assumes its average diameter at once and gut pouches appear close to the anterior end. No forward coecum is present. The epithelial cells lining the tract at first form a flat surface livit a short distance backward they become arranged in the form of longitudinal folds (Plate 19, fig. 3) that continue to the neighborhood of the accessory reproductive organs. In the latter region the diameter of the intestine decreases, its lining walls are smooth, and wedged in between the cloacal passage and the pericardium, it makes its way to the cloaca. The cells of the rectum are pear shaped with basally situated dense nuclei and the distal portions are sw'ollen with a secretion that in the form of a darkly staining finely granular mass fills the lumen of the gut. In this species the pei'icardium is long and the contained heart is relatively slender (Plate 8, fig. (i). The blood from the posterior part of the body enters the hinder extremity of the heart which here has the form of a very slender tube (Plate 19, fig. 9), attached to the dorsal wall of the pericardium. About midway in the jiericardial cavity this canal enters another section, of four or five times greater diameter, that for a short distance hangs freely in the pericardium but more antei-iorly unites with the pericardial wall and with gradually diminishing calibre holds this relation until it passes into the aorta. This latter vessel LOPHOMEMA SPIRALIS. 131 supplies the gonad in the usual way and in the head region breaks up into a system of sinuses that become continuous witii the blood spaces lying at the sides and beneath the intestine. In the region of the accessory reproductive organs the lacunae become very circumscribed, but ,so far as they have been traced they exhibit essentially the same relations as in P. hmvaiiensis. The corpuscles of this species (Plate 36, fig. 14) exhibit no features worthy of note. The paired gonad extends forward to a point some distance behind the junction of- the pharynx and stomach-intestine. Posteriorly it ends abruptly slightly in front of the j)ericardium with which it is connected by two short ducts. The arrangement of ova and sperms are the same as in the foregoing species. From the posterior end of the pericardial cavity the coelomoducts arise and proceeding forward (Plate 8, fig. G, and Plate 9, fig. 5) almost hori- zontally unite with the shell gland. In one specimen the first section of these tubes is of small diameter and ill defined, while in the other it is well developed and is filled with spermatozoa that also crowd the pericardium. In this same individual there is a circumscribed area including the extreme posterior tip of the pericardial cavity and the neighboring section of the coelomoduct where muii- bers of spermatozoa are attached to the epithelial lining. Beyond this jioint for some distance the duct continues of small calibre, with an epithelial lining composed chiefly of goblet-shaped cells charged with a clear secretion, and then suddenly terminates in a bulb-like enlargement that in turn luiites with tlie seminal receptacle and the slime gland proper. At first the cells in this swollen division are of greater height and more slender than those atljacent, and attach vast numbers of spermatozoa, l:)ut more anteriorly sex cells are lacking and the epithelium consists of goblet-shaped elements like those just descril^ed save that they are of larger size. The same type of cell also occurs in the elongated seminal receptacle except at its distal end where the cells are lower, more compact and also attach large numbers of spermatozoa. The ventral limbs of each coelomoduct extend backward from the point of attachment of the seminal receptacle to a point a short distance in front of the cloaca where they unite and communicate with the exterior by a single opening. The epithelial lining in this part of the duct is relatively very high and gives evidence of forming two distinct secretions. The cells in the anterior third of the canal are much vacuolated and contain a small amount of some faintly staining secretion. More posteriorly they gradually shade into more slender and elevated elements that, in the posterior half of the cloacal passage, contain numerous granules densely crowded in their distal portions. Such ];V2 L()I>H()MK.\1A SPIRALIS. cells arc conliiuHl almost exclusively to the ventral and lateral walls. The dorsal side in l)oth specimens consists entirely of cells whose entire substance, with the exc(>i)tion of a small basally placed nucleus imbedded in a scant amount of protoplasm, consists of a vacuolated homogeneous secretion staining like the muciparous cells of certain gastropods. It is possible that such a secretion is due to the transformation of products similar to those of the ventral granular cells, but there is no trustworthy evidence that such is the case. On (>ach side of the ventral furrow immediately in front of the cloacal open- ing arc two indentations of the hypodermis each of which contains not less than fifty needle-shaped spicules that are clearly modifications of the spines produced in the adjacent territory. As is shown (Plate 19, fig. 5) they are directed inward toward the mid line and sections show that multitudes of muscle fibres attach to the diverticulum chiefly on its blind extremity. Among the more prominent of these is a transverse band, another extending outwardly and attaching to the body wall and, most prominent of all, numerous strands that pass in a postero-lateral direction and attach to the body wall slightly behind the level of the cloacal opening. In its broader features the nervous system of this species corresponds closely to that of tlie other species described in this paper. The brain is of the ordinary biloljed type and is placed above the buccal cavity (Plate 6, fig. 5). From it six nerves arise that passing forward and downward jjrobably innervate " the buccal and body walls with the attendant sense organs. The relations of the pedal, lateral, and buccal connectives also conform to the usual type. The last named cord is exceedingly difficult to follow owing to its almost exact re- semblance to the dense masses of muscle and connective tissue that accompany it, l)ut with liigh magnification it may be traced to the elongated ganglia placed at the sides of the radula and a little below it. The commissure attaching to the posterior ends of the ganglia passes dorsal to the radula. No clearly defined subradular organ exists and no nerves or ganglia belonging to this system are present so far as it is possible to judge from the material in hand. The lateral cords, holding the customary position at the sides of tlie animal, pass backward through the body and finally terminate in ganglionic enlarge- ments on each sitle of the pericardium near the tip of the seminal receptacle. The commissure uniting them passes out from the hinder end of each ganglion and crosses dorsal to the intestine. One or two nerves also pass out from each enlargement, hut in the confused mass of muscle they have been traced but a short distance. ALEXANDROMENIA AGASRIZI. 133 The pedal cords after traversing the body gradually approach each other in the posterior end of the animal and termijiate in two ganglionic masses on each side of the mid line a short distance in front of the two groups of spines that project into the ventral furrow. There are strong indications that these poste- rior ganglia are united by a conmiissure. Owing to the difficulties of observation no other pedal commissures have been discovered and for the same reason latero pedal connectives have not been found with certainty though at various points there are indications that such exist. The posterior sense organ is located at the extreme hinder end of the animal in the mid line. At this point the hypodermal cells, unchanged in appearance, approach near to the outer surface of the body and there become continuous with a sensory epithelium composed of slender fairly dense cells in which the elongated nuclei hold an almost median position (Plate 32, fig. 11). Over the exact centre of this area the cuticle is exceedingly thin but gradually increases in thickness as the outer limits of the organ are approached, and contains con- siderable numbers of small spines that in both specimens overarch the sensory cup. Numerous muscles and connective-tissue fibres attach to its under surface and in the meshwork thus formed blood corpuscles and nerve cells occur in moderate quantity, the latter probably connecting with branches from a strong nerve that may be followed into close proximity to the posterior pallial com- missure. Alexandromenia agassizi, sp nov. Six specimens, one badly mutilat(xl, of this species were dredged in 4r)0 fathoms (Sta. 2992) near the Revillagigcdo Islands off the coast of Mexico. All save one, clinging to a fragment of some land plant (Plate 2, fig. 5), were unattached and nothing is known of their mode of life. The smallest specimen is yellowish white; the remainder, of larger size, are yellowish brown. The largest individual measures 25 mm. in length and 5 mm. in average diameter in the middle of the body; and of the five remaining sjiecimens three are of about this same size. The smallest is in an uncontracted state and is 22 mm. in length and 3 mm. in average diameter. As may be seen in Plate 2, fig. 5, the head is not distinct, usually bluntly pointed, and readily distinguish- able from the posterior end where the borders of the cloacal opening are widely expanded, in one specimen especially (Plate 5, fig. 5), exposing the gill plates, about 40 in number. A dorsal sense organ is visible in sections but not in surface view. i;}.| AI.KXANI)U()MP:NIA agassizi. Tlie Itoily is siirrmiiidcil l>y a cuticle O.IOS iiiiii. in thickness, and as in the prccpiliufi; species, this is iarjiely occupied by pai)iliae (Plate 33, tig. 5) and spicules of two varieties (Plate 37, (iji. '•>). Of the latter those of one type project from the hypoderniis, with which tliey i-emain connected, almost at right angles and l)r()tru first named spines. The papillae are fully as numerous as in the succeeding species and the constituent cells are approximatt'ly half as abundant; but the differentiation into stalk ;uul ex])antled part is mA so sharply defined (Plate 33, fig. 5). In many cases the base of the stalk is of great width and expands but slightly as the sur- face of the botly is approached, the jKipilla in such circumstance having a club- shaped appearance. Even in the more typical forms the departure from such a state of affairs is not marked. The cells appear to be all of one type in the distal portion, at all events the nuclei are of essentially the same size, though they vary considerably in shape, and ai'e surrounded by masses of yellowish green pigment. On the ventral surface, especially in the region of the mouth, gland cells a])pear in the hy])odermis. In their early stages each is pear-shaped, the stalk being inserted among the hypodermal cells, while the ilistal portion contains a lightly staining almost homogeneous secretion. Later this product becomes more abundant, swelling the c(>ll to twice its original size, and a granular mass appears to make its way by a very delicate pore to the exterior, though this has not been demonstrated to mj' entire satisfaction. The main portion of the anterior pedal gland is located between its outlet into the ventral furrow and the radula and its supports. At this point the cells are contiiuious across the mid line and laterally extend as thin plates compressed between the l)ody wall and the proiligious salivary glands (Plate 20, fig. 4). Posteriorly they separate into two groups which pass without any recognizable line of tlemarcation into the posterior pedal ghuul. The cells composing it are of the usual i)yriform type, densely filled with intensely staining secretory products, and are arranged into irregular groups or lobules. In the customary fashion the ductules from each cell open by an intercellular canal into the anterior end of the pedal furrow. .\t the point where the anterior pedal gland opens to the exterior the ventral groove becomes a deep excavation (Plate 7, fig. 3) the area of whose walls is incrcas(>d by the presence of extensive dorso-lateral outpouchings and numerous folds corn-sing from the roof half way down the sides. On the posterior face ALEXANDROMENIA AGASSIZI. 135 several folds arise and posteriorly extend along the groove to the cloacal cavity. These are not constant for at their first appearance they are eleven in number (Plate 20, fig. 9), soon decreasing to nine and gradually to five in the jxistcrior half of the animal. Each is penetrated by a loose meshwork of muscle and con- nective-tissue fibres, through which the multitudinous ductules of the pedal gland take their course, together with many corpuscles from tlieovcrlying sinuses. The component cells are high and columnar, especially the outermost which con- tain small ciuantities of yellow pigment. The opening into the atrium, holding its customary subterminal position (Plate 7, fig. 3), leads into the atrial cavity whose walls are difTerentiated into the usual ridges (Mundleisten) and cirrose area. As in P. hawaiicnsis the outer- most ridge is accompanied throughout its anterior half by a prominence, ill defined, and yet evidently sensory since it is composed of slender cells with basal nuclei resting uj^on a rod-like group of ganglion cells. In the posterior half of the lips this structure becomes more indistinct and finallj' blends indis- tinguishably with the outer atrial ridge. Of the two large atrial ridges the outer is of large size and skirts the cavity save on its posterior face. It is supportetl by an abundance of connective- tissue fibres associated with a scant amount of muscle bands among which small blood sinuses make their way. The halves of the dorsal ridge arise independently of each other in the mid line on the roof of the atrial cavity. At first very low they rapidly increase in height (Plate 20, fig. 1), but behind gradually disappear in the neighborhood of the opening into the pharynx. Blood sinuses penetrate into their interior and probably in life increase these organs to a very considerable degree. On each side of the mid line in front and hanging from the roof of the cavity are two pairs of large papillae springing from the outer and inner ridges respectively. The epithelial covering is composed of colunmar cells wlu)se distal half is filled with golden yellow pigment. Each cirrus arises Independently as a finger-shaped process of the atrial wall with an average length of 0.3 mm. It is composed of cells about twice as high as wide, closel}^ packed with yellow pigment, arranged about a central cavity within which it is occasionally possible to follow a nerve fibre. The opening of the mouth into the pharynx is guarded by a circular fold beyond which the canal passes dorsally until in the neighborhood of the radula where it bends at right angles and passes directly backward to join the stomach- intestine. Throughout its entire extent its internal lining is developed into many folds, large and small, often exceedingly wavy and of most complicated appear- 13(3 ALKXANDHOMKNIA AdASSlZI. iiiicc ill sections. Ill front oi the raduhi specially, large projections appear and, as will 1)0 more fully (Icscribcd presently, afford an outlet for the two sets of sali- vary glands. The character of the pharyngeal epithelium undergoes minor modifications at various points, but in general it may be said that the constituent cells range in siz(> from cubical bodies to others three times as high as wide, are devoid of cilia and are bounded by a well-defined cuticular sheet. The salivary glands, which probably are homologous with the dorsal set in other Neomenina, are in this species remarkable for their size and extent. With the exception of a few scattered groups of cells the main portion is distributed in the form of a wide band encircling the pharynx from immediately behind the brain to the radula (Plate 7, fig. 3, ds). So far as may be determined from a single specimen all the cells are pear-shaped and are grouped into club-shaped lol)ules of various sizes. In the expanded part the cells are one layer thick and surround a central canal down which the ductules make their way to open by intercellular canals into the pharynx. From transverse sections there are indications that a portion of the gland located on the dorsal side of the pharynx immediately behind the brain differs in character from the remaining ]X)rtiou. The follicles are more slender than those of adjoining regions and the glandular products in the early stages of their development are of lavender color and markedly different from the yellow- ish jiink substance elsewhere. All the follicles of the dorsal salivary glands of whatever character are supported by numerous muscles forming the pharyngeal wall and by connective-tissue bands between which numerous blood sinuses are present. The glands which probably correspond to the ventral salivary of other Solenogastres although in this species their outlet into the pharynx is somewhat more in front of the radula than usual, are enormously developed. As may be seen (Plate 7, fig. 3, sg.) they extend from the brain to a point considerably behind the radula where they entirely surround the alimentary canal and else- where overlap it to a very considerable degree. As in the dorsal group the gland is composed of thick-walled follicles, of large size, and from each pyriform cell a ductule leads to its intercellular outlet. Owing to the great bulk of the lobules and the remoteness of the greater number from the main duct (leading into the pharynx) this latter canal is provided with numerous branches (Plate 26, fig. 9) which come in contact with the majority of the follicles where each ends blindly. These minor ducts are lined by an epithelium in wdiich the cells are of two distinct types. In the inner half, through which the ductules make ALEXANDROMENIA AGASSIZI. 137 their way, the cells arc very high and slender and the cytoplasm vacuohited. In the outer half the cells are not over one third as high and the protoplasm is relatively compact. As just noted the branching ducts of each sitle unite into a single tube, which opens into the bottom of a dcej) depression on the sitle of the pharynx. From the base of this pit a large conical papilla (Plate 36, fig. 10), whose surface is thrown into five or six circular folds, projects inward, in life its ti)) ])rolial)!>' extending as far as the opening into the pharynx. It is practically solid, con- sisting of connective tissue and several muscle fibres which probably act as retractors. The ductules of a very few adjacent follicles make their way into this protuberance and are accompanied by similar tubules from some of the dorsal salivarj^-gland cells. These slender canals make their exit at several points, from the tip and by means of pores ventrally placed on the circular folds. The last named openings connect with small canals (perha])s one sixth the length of the circumference of the fold), hjcated between the cuticle covering the papilla and the underlying epithelium. There may be other exits but if so they are of small size and are invisible in the material in hand. Attached to the pharynx, between these large glands (sg.) and the stomach- intestine another extensive set occurs (Plate 7, fig. 3, and Plate 20, fig. 2) that are difficult to homologize. They consist of numerous pyriform cells, highly vacuo- lated or containing masses of some secretion of a pinkish shade, arranged in the form of lobules bounded by muscle and connective-tissue fibres. The ductules pass to the pharyngeal epithelial lining through which they open by intercellular channels. The radula is relatively small and is located on a tongue more than usually pointed. It is of the monoserial type (Plate 34, fig. 5) and as far as may be determined from transverse sections, comprises between forty-five and fifty- three transverse rows. Each tooth is a narrow rectangular plate lient to form a very obtuse angle. The odontoblasts are of the usual columnar type. The radula sac is supported by ten or fifteen pairs of large cells, probably turgid in life, filled with a highlj^ watery secretion, surrounded by connective tissue and muscle fibres. These last named elements are part of bands, too complicated to allow of reconstruction, that attach to the pharyngeal wall, the radula, or the sheath of the salivary glands. The gonad, with the usual relations, extends from a short distance posterior to the radula to within a short distance of the front end of the pericardial cavity. In common with other hermaphroditic Solenogastres the organ in this species 138 ALKXAXDKOMKXIA ACASSIZI. is i)air('d and llic cfiK^ are developed alon-z; the septum while the s])ennat()Z(ia arise more externally wlune tiie wails are often greatly folded. Posteriorly the iialvTS of the gland narrow greatly and assinne the form of comparatively slender canals (Plato 7, fig. 5), which communicate with the pericardium by wide oi^en- ings. Their epithelial lining is apjiarently ciliated and is fashioned into several low longitudinal ridges. The coelomoducts arise from the postero-lateral borders of the iierieardium as conijiaratively wide canals, which first extend downward and then forward to join the so-called shell gland at the point where the seminal recei)tacles are located. As usual the shell gland of one side joins the cori'esponding organ of the other and after narrowing to a sleiuler tube opens into the cloaca in the mid line. The walls of the coelomoduct (see Plate 20) in the region of the pericardium are comparatively thin, but one third the distance to the seminal receptacle they become thicker, the cells more slender and the ten to lifteen longitudinal folds more pronoimced, a state of affairs which continues to the shell gland. Cilia are certainly present at various jioints antl it is probable that they exist tlu-oughout the duct between the jiericardial cavity antl the seminal receptacle. In the same section small ([uantities of a glandulai' secretion are tleveloped having the form of minute granules which show at first a distinctly acid reaction but after their discharge become more or less confluent and alkaline. Minute (luantities of spermatozoa are also distriliuted throughout this same division of the duct. In the single specimen examined the seminal recei)tacle is a small disc- shaped sac attached to the coelomoduct where the iimer and outer portions meet. It is wedged between the bod\- musculature and the shell gland and the slit-like lumen contains a few spermatozoa only. The cells composing the walls are comparatively low and are glandular, the clear seci-etion, small in amount, giving the cytoplasm a vacuolated appearance. In this species the shell gland is of enormous size, filling practically all the space between the digestive tract and pericardium dorsally and the body muscu- lature ventrally and laterally. As figured, Plate 20, each half is penetrated by a duct, of about the diameter of the foregoing section, thi'ough which the secre- tion from the surrounding glandular portion makes its way. In the neighbor- hood of the cloaca these ducts emerge from the gland, unite with each other, and forming an S-shaped loop in side view open into the cloaca by a very narrow pore. ALEXANDROMENIA AGASSIZI. 139 The component gland cells are arranged into numerous lol)ules, which gener- ally extend from the surface of the gland tu tlie main central duct. These are separated from each other by delicate connective-tissue sheaths, which fre- quently contain small blood sinuses, and are traversed by a slender canal which apparently does not function as a duct. The evidence is not altogether conclusive but from a careful study of sections it appears that the secretion does not pass through the cavity of the lobule, but is contained in delicate ductules which arise in the gland cells. Arri\ing at the central cavity of the gonoduct the darkly staining secretion makes its escape through the lining epith(>liuni by means of intercellular openings. The epithelium of the main duct is composed of high colunniar cells which contain a finely granular lightly staining glandular substance. About the termi- nal section of the coelomoduct, where the eiMthelium become lower and the secretion scant in amount, a heavy sheath of circular muscles appears and con- tinues to the cloacal cavity. As might be expected in an animal of tlie size of this species the circulatory system is well developed and of more than ordinary complexity owing to the large number of sinuses which hold po^sitions in and aroimd the various systems. The heart, having the usual position, consists of two distinct divisions (Plate 7, fig. 5), an auricle and ventricle. The walls of both of these are of uncommon thickness, but otherwise present no especially noteworthy features. They are connected by a tubular stalk which projects slightly into the cavity of the ven- tricle and may serve as a valve. From the front end of the ventricle blood passes through numerous openings in the somatic musculature to unite into one tube, the aorta, which from the out- set appears to lack any trace of an endothelial lining. As it coui'ses forward dorsal to the reproductive gland it originates numerous vessels which ramify through the body wall or passing ventrally between the halves of the gonad forms a subgenital sinus. In the head region it expands considerably and communicates with extensive lacunae in and between the body wall and the alimentary canal. The blood in these spaces makes its way to the ventral surface where it unites into one main canal, immediately ventral to the gut, which connects also with several small channels coursing along the folds in the ventral furrow. At the posterior end of the Ixjdy this ventral sinus courses dorsally, keeping close to the under surface of the gut, and when opposite the junction of the auricle and ventricle it divides into two short branches which pass backward on each side of the rectum. These last named vessels are very 140 ai,kxaxi)|{()Mi:m.\ acassizi. short and s.m.ii divUlc into a larj^e nuinl)ci- of minor sinuses conninmicatiiifi; with the fiili folds. Each fiill |)latc is nicri-ly a cihatcd fold of the cloacal wall with which it connects anteriorly and laterally. The blood enters the anterior border usually, and coursing through the narrow enclosed spaces finally makes its way laterally to the body wall. Here it unites with other vessels of similar origin and finally by means of a large canal passes dorsally and enters the heart immediately after uniting with the corresponding sinus from the other half of the body. The blood corpuscles are spherical bodies, .0074 mm. in diameter, hyaline in ajipearance and containing a small, dense nucleus. The leucocytes are remark- ably infrequent, unusually compact, but otherwise devoid of any noteworthy features. About the bases of the gills there are great accumulations of gland cells (Plate 36, fig. 19) occupying spaces in the meshwork of muscle strands between the folds and the body wall. They project somewhat into the sinus of each gill plate, and are occasionally penetrated by l)lood sinuses, Init there is no trace of any outlet nor is there any indication of their possible function. Each cell is approximately 0.01 mm. in diameter, and contains a small, spherical nucleus imbedded in slightly vacuolatetl, granular cytoplasm. The nervous system of this species is in an excellent state of preservation, and as the nerves in many parts of the body are more than usually well defined considerable attention has been devoted to this portion of the anatomy. The location of the brain and principal ganglia"are the same as in Dorymenia acuta for example, but in the distribution of certain of the nerves difTerences appear which are here described. The brain (Plate 12, fig. 5) is relatively small, but anteriorly gives rise to the usual number of nerves distributed to the atrial cavity and the neighboring body wall. The miited coimcctives of the lateral, pedal, and labio- buccal systems enter from the side. The first two comiectives become differen- tiated a short distance laterally and pursue their usual course through the body. The labio-buccal connective springs from the lateral connective posterior to the union with the pedal and holds its customary position at the sides of the pharynx. In Dorymenia there is a large nerve fibre arising from the anterior end of the lateral ganglion and is distributed to the pi'ecerebral ganglia about the bases of the cirri; in the present species it is totally lacking. In l)oth species nerves arising from the pedal ganglia are distributed to the walls of the pedal-gland outlet. Numerous connectives imite the lateral and pedal ganglia, and are far from being regular, in several cases uniting with neighboring connectives by ALEXANDROMENIA AGASSIZI. 141 delicate branches. Usually the most anterior connective is of the largest calibre, but in the present instance the first two or three are extremely tenuous, not over one third the diameter of the succeeding connectives. Dorsal nerves from the lateral ganglia are numerous but in no instance have they been traced to the mid dorsal line. The labio-buccal system is probably more extensive than is shown in Plate 12, fig. 5, for owing to tlie great width of the muscular joharynx and the abundance of salivary glands closely crowded together it is very difficult to trace nerves for any considerable distance. The connectives may be readily followed to the ganglia at the sides of the pharynx, and the commissure uniting these is as readily demonstrated, but a different state of affairs is met with elsewhere. Before uniting with the la]:)io-buccal ganglia the connectives far out on the external face of the salivary glands enter a ganglionic enlargement on each side from which two nerves originate. One of small size disappears almost immediately among the glands; the other of much larger size passes in toward the mid line, Ijut Ije- comes lost in the darkly staining secretion of these same glands. From the dorsal side of each labio-buccal ganglion a small nerve arises that gradually extends to the lateral border of the pharynx which it crosses to form a com- missure. At various points throughout this entire system exceedingly small nerves arise and probably innervate the neighboring regions but their destina- tion is not certain. In the ]:)osterior end of the body the lateral nerves become so crowded against the body wall, owing to the huge shell gland, that it has been impossible to trace connectives in this region. Opposite the posterior end of the heart they expand greatly (Plate 12, fig. 6), and originate several nerves distributed more posteriorly as well as the suprarectal commissure. Two connectives, the pos- terior one of large size, unite with the posterior end of the pedal cords. In the mid line the suprarectal commissure develops a nerve that passes to the base of tiie dorso-terminal sense organ, to which it sends a small nerve, whereupon it proceeds backward distributing fibres to the dorsal gill plates. The large branches, springing more laterally from the posterio-lateral enlargements, branch repeatedly and in many places delicate offshoots have been found enter- ing the branchial folds. Plate 12, fig. (> represents the more important of these whose number ami origin is correctly shown though the branching is somewhat diagrammatic. With this exception the nerves and ganglia shown are recon- structed carefully from micrometric measurements. The pedal ganglia have not been examined in the middle of the body, but 142 ALKXAXDllOMENIA VALIDA. elsewhere (hey are united by stronu; eonimissures. In the region of the shell gland this organ is entered, and probably iiniervated, by several fibres which arise from these ganglia. Poslerioi- to the last of the latero-pedal connectives three or four small bi-anches pass into the muscles about the external reproductive opening. No nerves have ever been followed from this source into the region of the branchial huncllac, and if any are derived from this source they are of V(>ry small size. Alexandromenia valida, sp. nov. Four specimens of this species (Plate 3, fig. 3) were collected off the coast of southern California from the following Stations: 2980, 4382, 4389, 4391. T\\v d('i)th langes from 603-1350 fms. and in every case the bottom was green nuid. All the specimens were unattached so that nothing was learned of their habits. The measui'ements of the largest specimen (Sta. 4389) are as follows: length 32 mm., dorso-ventral diameter of head 3 mm., of cloaeal region 3 mm., average diameter of l)ody 3..") nnn. The head is not distinct from the liody but is characterized by a more pointed ajjpearance than the posterior end, and in all the specimens the borders of the cloaeal opening are slightly expanded, exposing, to a slight extent, the gill folds. The atrial opening is relatively small and its forward border is almost level with the front end of the animal. A pedal furrow extends along the ventral surface of the l)o(ly and posteriorly becomes continuous with the cloaeal cavity. Anteriorly, for about 1 nnn. extent, it expands and allows the escape of the anterior pedal gland seci'etion fPlate 21, figs. 2, 4) which, in the type, fills the ojH'ning and extends posteriorly some distance along the ventral furrow. The gen(>ral color of the Ixxly is light yellow. A well-dcveloiied dorsal sense organ is jiresent, situated in the type about 2 mm. from the posterior end of the animal. It is in the form of a shallow pit and is especially conspicuous in surface view on account of the numerous hypo- dermal jiapillae which surround it. The body is surrounded liy a cuticle, O.ii; mm. in thickness, which is crowded with inmuncniblc si)icules and papillae of large size (Plate 33, fig. S). The hypodermal layer is concei-ned almost wholly with the development of these structuR's, the |)ortion i)rol)al)ly responsible for the formation of the cuticle being limited to a fcnv cells packed between the bases of the papillae and develop- ing spines. Everywhere the hypodermal elements are of small size, and though excellently preserved are not favorable for study. As just noted the i)ai)illae are of enormous size and in a fully developed ALEXANDROMENIA VALTDA. 143 condition contain not less tiian 100 colls. Of those fully twonty-five aro located in the stalk while the i-eniaindor hold positions in the expanded portion. This last named section is relatively comjiact and lacks the spaces and psoudopodia- like processes characteristic of the niajoiity of Solenogastres. The cells com- posing it are of two distinct types, one containing small spherical dense nuclei and another in which the luiclei are of twice the diameter of the first, and stain but faintly in haematoxylin. The cell boundaries are invisible and it is conse- quently impossible to determine if these differences are correlated with others. The cells with the small nuclei contain a few relatively large yellowish pigment granules, and there is some evidence, though not wholly conclusive, that this secretion is absent from the cells of the remaining type. Of the two kinds of spicules the larger (0.189 mm. long) projects from the hypodermis almost at right angles and its ]:)ointed extremity projects slightly above the surface of the body. As Plate 37, fig. 11 shows these spicules are hollow and remain constantly in contact with their matrix cells, which are several in number and in their general appearance and relations are not unlike those of P. haivaiiensiii. The second type of spine is many times more numerous than the one just described. Over the body generally all are of essentially the same size (0.135 mm. average length) and in their needle-like appearance resemble the long spines of Proneomenia, Strophomenia, etc., with which they are probably homologous. Without any very definite arrangement they form several layers parallel with the hypodermis. The anterior pedal gland occupies practically all the space between the body and pharyngeal walls between the brain and the radula. The cells are of various sizes but average 0.189 mm. in diameter and are so densely filled with secretory products that all other elements of the cell are invisible. From each a duct leads to its intercellular opening int(j the anterior end of the ventral furrow. The outlet of the anterior pedal gland, the widened end of the pedal furrow, is of large size, and its walls are fashioned into numerous folds (Plate 21, fig. 2) to afford sufficient surface for the exit of the many ductules. On the roof of the cavity two short longitudinal ridges, almost papilla-like, are pi-esent and from these numerous much smaller folds extend down the sides of the chamber to the external opening. After treatment with haematoxylin the secretion of the pedal gland becomes almost black, indicating an alkaline reaction, while the cells lining the outlet of the pedal gland (between which the ductules make their exit) are bright pink in color and therefore distinctly acid in their reaction. [IJ ALRXANDROMENIA VALID A. Tlio colls of the posterior pedal gland are of comparatively small size, but otherwise are essentially tli(> same as those of the anterior pedal gland. As usual they open by separate intercellular exits into.the pedal furrow. In several sp(>cies of Solenogastres the foot is accompanied by two longi- tudinal ri(lg(-s, modifications of the hypodermis. In the present species the median projection is bordered by two pi'ominences on each side. All are of essentially the same size and appearance, being thin folds of epithelium into wliicii a few muscle antl connective-tissue fibres project. No blood spaces occur within tiiem and at their bases the secretion from the pedal gland finds its outlet. The atrial cavity in the only specimen examined is of limited extent (Plate 11, fig. '.i), but it possesses ridges of large size that together with the cirrose area presents a very characteristic appearance. The indistinct prominence, which in some species accompanies the outer ridge, is not sharply defined, being recog- nizul)le solely by the rod-like group of ganglion cells in the customary position. The outer atrial ridge projier is of large size, and as a much folded, horseshoe- shaped elevation surrounds the atrial cavity except posteriorly. The inner fold is of inferior size and consists of two ridges which arise independently on the roof and posteriorly diminish in size and gradually disappear. Their epi- thelial covering consists of high slender cells with elongated subcentrally placed nuclei distal to which the cytoplasm contains quantities of light yellow pigment. Internally the folds are sujiported by strands of connective tissue with a sriiall adniixtiu'e of muscle f Hires among which well-defined l)Iood sinuses pursue their course. The cirri arise singly from the atrial wall and are of more than ordinary thickness. As usual each consists of an outgrowth of the buccal wall, composed of more or less cubical cells in which the yellowish brown pigment is so abundant that it usually conceals the nucleus. The cavity of each cirrus is very narrow, allowing the passage of a nerve strand l)ut not of the blood. The liuccal cavity or jiharynx, seixirated from the atiial cavity by a circular fold (I'late 11, lig. 3), is an irregular cavity whose general appearance and rela- tions are represented in Thite 21. The walls throughout are produced into numerous wavy, more or Ic-ss longitudinal folds, lined with a thin cuticular sheet. In the middle third, whicli contains the radula, the folds become more distinctly longitudinal, but moi-e posteriorly they once more become very irregular. As ni .1. (ifidsxizl there are three sets of salivary glands, and as may be seen ni I'late 21, figs. 2, 4, in arrangement and size, they are essentially the same as ALEXANDROMENIA \'ALIDA. 145 ill the foregoing species. The smallest dorsal glands (dsg) are more or less im- bedded in the pharyngeal wall from the eirrose cavity to the posterior end of the radula sac. Each consists of an aggregation of well-marked pyriform cells usually charged with a finely granular darkly staining secretion. In some of the larger groups the secretory products are not so clearly granular and have a more reddish cast, in this respect and in general appearance resembling the second type of dorsal salivary gland. The second species of gland (sg) is in reality a paired structure each half consisting of about two dozen lobulate glandular bodies united by as many branches of a main duct whicli opens into the pharyngeal cavity. It appears probable that each organ arose in the embryo as a diverticulum of the gut, and subsequently developed outgrowths in which some of the cells became glandular. These retaining their connection with t\w lumen of the duct elongated greatly, became pyriform and formed the lobule of the completed gland. As in A. agassizi the canal in each lobule develops small lateral branches and in any case the duct holds a superficial position. The course of the main duct, which lies to the outside of the glands is shown in Plate 21, fig. 2. It opens at the base of a corrugated papilla enclosed in a diverticulum of the pharjaix that in turn opens at the forward border of a broad papilla on the pharyngeal wall (Plate 11, fig. 3). As in the foregoing species the papilla contains a few small canals which open on its surface, but their inner con- nections are difficult to trace. They appear to be the outlets of a number of small glands belonging to the first type which, as noted above, approach the second in the form of the cells and the character of their secretion. The tubules of the third set (gl) are in form and position like those of .4. agassizi. The secretion is more abundant and more granular and darkly staining, giving them a denser, more compact appearance, but this set is not voluminous as in the preceding species. The radula is of the distichous type (Plate 11, fig. 3, Plate 34, fig. 14) and contains approximately tliirty-four rows of teeth. These are developed by large numbers of exceedingly slender odontoblasts, and immediately after their forma- tion are enveloped in sheaths composed of numerous so-called enamel cells. Both of these groups blend with cells that become smaller as the opening of the radula sac is approached. At the forward border of the radula the cells of the pharyngeal wall become more columnar, less dense, and their nuclei assume a more slender shape. Ap- pearances suggest a subradular sense organ, but it lacks the definiteness of this 146 HAI,()MKXIA (;K.\VII)A. structure in Pnmcoinenia hmmiiends for exaniplo, and is apparently not inner- vated by a well-defined subradular nervous system. The oesophagus opens at the summit of a papilla (Plate 11, fig. 3) into the stomach-intestine, which manifests no especially noteworthy features save that its lining is of such thickness that in ])reservericardial cavity is smaller, and the inner ends of the coelomoducts are moic slender, but they rapidly increase in size and their walls l)ecome more than usually folded. The shell gland, especially its posterior half, is more distinctly lobulate and somewhat more acid in reaction. The seminal receptacles are considerably larger; but neglecting these differences the two species agree closely so far as this system is concerned. The nervous system is not especially favorable for study and for this reason only the more obvious portions have been examined. In all essential particulars these closely resemble homologous structures in the foregoing sjjecies. nalomenia gravida, up. nov. This spe(ues is represented liy two individuals taken off Simushir Island of the Kurile group at a depth of 229 fathoms (8ta. 480-4). Both were discovei-ed in dead barnacle shells and are evidently free roving forms. The larger speci- men is 11 mm. long by 1.6 mm. average diameter and the length index 7:1 is characteristic also of the smaller one. The anterior cirrose section of the gut, or the atrium, is separated from the succeeding portion by a ridge (Plate 5, fig. 3) covered with the spiculose cuticle investing the body. Posteriorly the pedal groove is continuous with the cloaca which like the atrial opening, extends far up toward the dorsal surface of the body. The color is light yellow shading to nearly white in the head and cloacal regions (owing to com]-)act muscles) and along the mid dorsal line where the gonad is situated. .V dorsal sense organ is present (Plate 22, fig. 12), and is remarkable for its large size and from the fact that it is more anteriorly locateil than is usually the case with other species, being placed opposite the forward cloacal wall. HALOMENIA GRAVIDA. 147 The cuticle, surrounding the body, is of average thickness (Plate 32, fig. 4), but is actually rather scant in amount owing to the large numbers of spicules (Plate 22, fig. 13) imbedded in it anil to the papillae many of which are of unusual size. On the ventral side of the body the unmodified hypodermal cells, those which are probably largely responsible for the development of the cuticle, are comparativelj' few in luunber and are crowded between the bases of the papillae, but dorsally they become more numerous and may be seen to possess a cubical form and no especially noteworthy features. Owing possibly to differences in age the papillae vary greatly in size but all are constructed on essentially the same plan. As may be seen (Plate 32, fig. 4) the stalk is composed of an out- pushing of hypodermal cells and is usually shorter in the larger papillae. It is surmounted by the usual balloon-shaped group of cells, which certainly number not less than one hundred in the larger organs. Each cell is greatly elongated, vacuolated in its outer portion in jireserved material, and contains a spindle- shaped, basally placed nucleus. In several cases the cavity of the stalk is traversed by a delicate fibre, sometimes enclosing a nucleus, which appears to be a nerve. In this species the stomach-intestine is related in a remarkable way to the cuticle and papillae but for what purpose I cannot say. On the dorsal surface of the smaller specimen and in the region traversed by the mid gut there are not less than twenty pairs of small rounded knobs of light yellow color, forming one longitudinal line on each side of the mid line. These are not distinctly visible in surface views of the larger individual, but in sections they are seen (Plate 22, fig. 1) to be evaginations of the dorsal wall of the intestine which protrude through definite openings in the somatic musculature and extend half way to the outer surface of the body. Each is in contact with the under surface of a circular disc-like patch of hypodermal cells having the appearance of a modified papilla. The stalk is absent and the cells are relatively low, but some are distally vacuo- lated and are not the compact, cubical elements of the unmodified hypodermis. Surrounding the point of attachment of gut and papilla is a small ring-shaped blood sinus, frequently containing corpuscles. The relations of these various elements are represented in a typical condition in Plate 32, fig. 5. Concerning their mode of operation it appears probable that the pressure of the blood in the sinus cau.ses a protrusion of the papillae and the attached liver lol)e, but for what possible reason I cannot say. The spicules, whose general shape is shown (Plate 22, fig. 1.3), are of various sizes in a mature condition even in the same locality. These are intermingled 148 HALOMKNIA GRAVIDA. and lack any very definite arrangoiiu-iit further than that they encircle the enormous papilhie, giving the animal as seen under low magnification, a mottled appearance. The anterior pedal gland occupies the major portion of the space inckuled between the gut (Plate 22, fig. 1), and body wall from the level of the brain to the stomach-intestine. Its cells are comparatively small but the secretion is abundant, enabling one to follow in many cases the slender ductule to its opening into the anterior end (jf the pedal furrow. The walls surrounding this external outlet are unfolded (Plate 22, fig. 1), highly ciliated and form a cavity of more than usual size. From its posterior border two I'idges develop (Plate 22, fig. 2) and extend along the pedal furrow to within a short distance of the cloaca, where they disappear though the furrow, reduced in size, becomes continuous with the cloacal cavity. In surface view it is jjossible to detect slight folds in the exposed cloacal wall which in sections may be seen to become of much greater height within. These, twenty-six to thirty in number, at first hang freely in the cloacal chamber (Plate 22, fig. 11), but more anteriorly they become attached to the wall of the rectum, antl yet farther forward some of the dorsal ones extend into the cavity above the rectum dividing it into a corresponding number of small crypts (Plate 22, fig. 8). In these last named spaces, and between the folds for some distance more posteriorly, upwards of twenty embryos have taken refuge and undergone the first stages of their development protected by the parent. The epithelium of the basal half of each io\d in contact with the embryo is low and seemingly non-ciliated, while that of the distal half assumes the high colunmar, ciliated appearance characteristic of such organs in several other species. As noted previously the cirrose section of the gut, or the atrium, opens subterminally and has no direct connection with the remainder of the digestive tract. Its cavity is largely obscured by the thick-walled, large ciliated ridges which hold the usual position and define the cirrose area. The cirri are com- paratively short and thick set and are usually united by their bases in groups of two or four. Their cells are of the customary pigmented type and surround a lumen of exceedingly small calibre. Immediately behind the opening into the atrium the cuticle surroumling the body becomes continuous across the mid line for a short distance, and still more posteriorly breaks through to form a second opening, probably the true mouth. This aperture leads into an irregularly shaped cavity (the general arrangement shown (Plate 5, fig. 3), whose walls, seemingly ciliated throughout. HALOMENIA GRAVIDA. 149 are developed into a complicated series of ridges. In the customary position a distichous radula is present and though of small size is typical in all essential respects. It rests upon a delicate though perfectly distinct basement membrane and is produced by odontoblasts at the bottom of a shallow sac. There are, so far as may be determined from cross sections, about twenty-four rows of teeth which present the appearance represented (Plate 34, fig. 12). On each side of the forward limits of the radula sac the narrow ducts of the ventral salivary glands open into the digestive tract and on the other hand lead right and left into a reservoir extending far toward the dorsal side of the pharynx. Each of these cavities is surrounded by a gland (Plate 22, figs. 2, 5), composed of multitudes of small, pj'riform cells grouped, by means of delicate connective- tissue septa, into lobules of various sizes. The ductule from each cell makes its way to the wall of a reservoir into which it pours its granular, moderately stain- ing secretion by way of an intercellular opening. Beyond the radula the alimentary canal courses dorsally and opens into the stomach-intestine. In front of this junction a coecunv extends far forward and as noted previously develops on each side of the mid line diverticula which pierce the body wall and come in contact with the under side of what appear to be modified papillae. These are developed also by the stomach-intestine throughout its entire extent. A further peculiarity of the mid gut exists in the form of numerous small secondary outpouchings of the wall of the ordinary gut pouches which give in cross section a complicated appearance to this region (Plate 22, fig. 3). The dorsal wall of the intestine, in contact with the gonad, is relatively low and heavily ciliatetl. .Vt the level of the forward pericardial wall the gut narrows rapidly, loses its glandular character, becomes ciliated and by a relatively small pore opens into the cloaca. The position of the heart, its relation posteriorly to the branchial folds and anteriorly to the aorta are typical. As is seen (Plate 3, fig. 5) it consists of two divisions united bj^ a small canal apparently provided with a valve. Both portions are moderately muscular, and lodge in the meshes of the muscle bands numerous large, irregular cells which have the appearance of blood forming elements. The aorta, extending anteriorly and dorsal to the gonad, supplies the last named organ with many well-defined ventral branches which, as in P. hawaiiensis , pass ventrally along the mid line and reaching the neighborhood of the gut course outwardly and then dorsally to unite with the larger spaces beside the aorta. In the head region this main vessel communicates as usual with sinuses which carry the blood in turn to spaces between the gut and body j-Q HALOMKNIA GRAVIDA. wall. OiMH.sito tin- anteri..r ennts in the anterior third are wholly male, an.l in the neigh- borhood of the iKM-icardium also are great accumulations of spermatozoa while in an intermediate position the great ova, ().2(i nun. in diameter, are most con- spicuous objects. Correlated with their great size the ducts leading into the pericardial cavity are of unusually large calibre, being 0.175 mm. in their greatest diameter. The lining cells are low and heavily ciliatetl. From the postero-latcral borders of the pericardium the coelomotlucts arise, extend outward and forward and enlarging somewhat unite with the shell gland (Plate 3, fig. 5). In the early part of their course the cells, like those lining the pericardium, are low, but more outwardly they l^ecome more colunnuu- and form longitudinal ridges of considerable height. At the anterior linuts of the shell gland is the opening of the seminal receptacle, which is a simple unbranched tube, empty in the present instance, and is provided with a high ridge extending, like a typhlosole, throughout its length. The cells composing this latter organ are slender, triangular elements which when combined form a fan-shaped structure in cross section. A heavy layer of circular muscles en- sheaths the seminal receptacle, and a few radiating bands extend from it chiefly to the body wall. As usual the shell gland is U-shaped ami in the present exam]ile is fully functional. In the neighborhood of the seminal receptacle its cells are rather low and their secretion snuill in amount, but half way down toward the mid line they become greatly elongated, and distally contain a finely granular secretion which escapes in large ciuantities into the lumen of the duct. Upon the fusion of these tubes in the mid line the tlorsal wall of the undivided section is composed of cells, also high in form, which during the early stages of glandular activity are filled with a darkly staining vacuolated secretion (Plate 15, fig. 8 and Plate 22, fig. 6). This condition of affairs, mucous cells dorsal and an)umen forming elemcTits below, continues to the single median opening in the cloacal cavity. As noted previously this species broods its young. The eggs, about two dozen in luunber, have been retained in spaces between the great branchial folds in the cloacal wall and evidentlv they have come down at different periods HERPOIMENIA PLATYPODA. 151 as all stages of development are represented from a 4-cell condition to advanced larvae where the shell and foot are indicated, the fore and mid gut clearly differ- entiated, the anterior pedal gland developed and to some extent functional, the central nervous system partly mitlined (Plate 22, fig. 7). A fuller account of the embryology is planned for a later paper. The brain, of usual size, holds the customary position above the pharynx. The anterior nerves passing to the cirri, etc., and the pedal, lateral, and buccal connectives have the usual relations. Anteriorly the pedal cords enlarge and are united by a more than usually heavy commissure, and each gives rise to a nerve that passes to the wall of the outlet of the anterior pedal gland. Here and there pedal commissures may be detected, as well as connectives with the lateral cords, but these are usually small and often very difficult (o follow. The labio-buccal connectives are imbedded in the walls of the pharynx and connect with ganglia, of rather small size, located at the sides of the radula. These ganglia are united by a conmiissure ventral to the pharynx posterior to the radula. The ganglia, and more anteriorly the comiectives, give rise to small fibres that may form commissures, as in other species, Init their lack of sharpness renders it impossible to trace them more than a short distance into the pharyn- geal wall. In the posterior end of the body the pedal nerves diminish in calibre and disappear beneath the shell gland. The lateral nerves, upon reaching the forward border of the shell gland, pass diagonally inward toward the mid line until they reach the level of the outlets from the pericardium. Here they en- large (Plate 3, fig. 5) and are united l)y a conmiissure passing dorsal to the rectum. The last two connectives between the jiedal and lateral cords are of about twice the usual diameter and pass to the inner side of the shell gland. From the posterior superior ganglion several nerves arise that course over the cloacal passage to which they give off delicate fibres, then extend into the somatic musculature and probably are distributed in part to the hypodermis. Near the mid dorsal line two other nerves originate and attached to the forward wall of the cloaca j)ass dorsally and are distributed through the somatic musculature in close proximity to the hypodermis. Herpomenia platypoda, sp. nov. Eleven specimens of this species were taken in the neighborhood of Agattu Island of the Aleutian chain (Sta. 4781) in water 482 fath. in depth. All were attached (Plate 1, fig. 4) to a colony of some unidentified campanularian hydroid 152 HKUrOMKNIA PLATVFODA. and varied in size fmni 1 1 nun. in Icnglli and (I.O nnn. diameter to 18 mm. long and O.U mm. in thickne.ss. A well-defined dorsal keel extends throughout the entire length of the animal with the exception of the extreme pos- terior end. In a eontraeted state both ends are pointed and quite similar, but usually tlie atrial cavity is opened so that the outer ridges are exposed, thus giving the front end a blunt appearance. The color is white or yellowish white, depending to considerable degree upon the food contained in the alimentary canal. A single layer of spicules envelops the body, the majority being leaf-like in form and biconvex in cross section. A second type with a short stem occurs in the neighborhood of the ventral furrow, and to a less extent over the body generally. The cuticle is more highly developed than is usual with species in this family, and is underlaid by a hypodermal layer of more than usual height beneath the dorsal keel. In their general appearance the component cells resemble those of some of the Chaetodermatidae, being of cubical or rectangular shape, or in the case of what appear to be sjiicule matrix cells more or less globular except in the region of the keel where they are much elongated. The anterior pedal gland (Plate 19) occupies practically all of the head region not held by the ventral salivary glands from slightly in front of the middle of the pharynx to the forward end of the body. The crypt into which it opens is typi- cally placed (Plate 8, fig. 1), possesses unfolded walls and is profusely ciliated. Posteriorly the anterior gland joins, without any sharp line of demarcation, the posterior one which continues to the cloaca. In the anterior end of the body the foot is the merest fold or it may be en- tirely smoothed out. This latter condition obtains in the posterior half of the body (Plate 19, tig. 10). However, all of the cells retain their ciliated condition though they are more colunmai' than the ordinary pedal cells of other species. The atrial cavity is unusually small (Plate 8, tig. 1) and the outer ridges are lacking or are without clearly defined boundaries. The inner elevation on the other hand is a prominent fold, penetrated by blood sinuses, and is composed of cubical cells save along the free border where they are higher, more sjwngy and heavily ciliated. The cirri, presenting the customary appearance, are ar- ranged in groups of 3-7. The opening from the atrium into the buccal-pharyn- geal section is comparatively narrow, being reduced by a large fold springing chiefly from the dorsal side of the digestive tract. Beyond this the canal widens. Its hnmg epithelium becomes considerably folded, and is composed of columnar cells endowed with considerable glandular activity. Still farther inward the wall again develops a circular fold which forms the outer boundary of another circu- HERPOMENIA PLATYPODA. 153 lar groove between it and the muscular pharynx. Through the posterior wall of this groove the ventral salivary glands find their outlet by means of two ducts very closely situated to the mid line (Plate 8, fig. 1, Plate 19, fig. S). The ventral salivary glands are uniusually large organs completely encircling the gut for part of their course, and occupying most of the space between it and the body wall from their outlet to the junction of the pharynx and stomach- intestine. The cells composing them are pyriform, 0.024 nun. in greatest diame- ter and are filled with a homogeneous, moderately staining secretion in which the relatively large nucleus holds a more or less central position. The cells are arranged in lobules, and their delicate ductules open by intercellular channels (Plate 19, fig. 13) on each side of the body into a reservoir whose superficial extent is much increased by several folds. The outlet from each reservoir passes to a small diverticulum of the pharyngeal wall into which it opens opposite its fellow and very close to the median plane. The pharynx is a prodigiously heavy tulie whose walls are composed of a compact mass of muscle fibres chiefly circular (Plate 19, fig. 7) lined with high columnar cells developed into upwards of a dozen longitudinal folds. These last named elements contain a weakly staining secretion and are in contact distally with a distinct cuticular membrane. Posteriorly this tube projects some distance into the stomach-intestine, Plate 8, fig. 1. A radula is completely lacking in this species, and it is probable that, as in Drcpanomcnia rampijrcUn, the secretion of the salivary glands exercises a solvent action on the tissues of the host that are then sucked up by the jiowerful jiharynx though its exact mode of operation is difficult to understand. Nettle cells, from the host, are present in the intestine, in some cases seemingly imbedded in the epithelial cells. The lining of the stomach-intestine, where the distal portions of the cells bearing the secretion have not been detached, is excessively high, in one specimen almost occluding the lumen. Beneath the gonad is a small median fold appar- ently ciliated, and showing evidence of slight glandular activity. Ventral to the pericardial cavity the gut rapidly narrows to an almost circular tube of small calibre and opens with the shell gland into the cloaca. The pericardial cavity, as may be seen (Plate 8, fig. 3), is of moderate size only, and is in large measure filled by the simple tubular heart. The aorta is exceptionally small as are the sinuses generally with which it connects, yet so far as they have been traced their relations appear to be perfectly normal. The gonad holds the customary position, but is remarkable in several respects. In the first place it is sharply differentiated into two zones in the two 154 IIKItl'OMEMA Pl.ATVrODA. specimons {'xamined, :i forward section in which the sex products originate, and a posterior division that scrv(>s merely as a duct. This latter portion is exceedingly narrow (Plate 8, lig. 8), ui)wai'(ls of twice the bodily diameter in length, is with- out any signs of developing germ cells and contains throughout fully developed sperms. The ova and sjierms develop in the customary position, but the former are scant in amount and unusually large, almost completely filling the gland. Especially about the periphery each ovum contains imbedded in the yolk large numlx'rs of clear vesicular botlies (Plate 35, fig. 9), approximately 0.0068 mm. in diameter with an eccentrically placed darkly staining mass usually super- ficially placed. At first these bodies appeared to be remnants of nutriti\e cells, possibly modified follicl(> cells, but subsequent study leads strongly to the con- viction that they are portions of fragmented nuclei. In the early stages of ova development these same bodies occur, but are of extremely small size (Plate 35, fig. G). Still earlier (Plate 35, fig. 7) it has been possible in several cases to find imbedded in very small ova from one to three cells resembling primordial germ cells, and probably corresponding to follicle cells that are known to occur in a few Solenogastres. These nuclei in a slightly older stage become somewhat larger antl stain blue instead of light red or pink. The granules assume the ves- icular appearance characteristic of later stages and slightly later upon the rupture of the nuclear membrane become scattered throughout the egg. Against the belief that these structures are cells may be urged the fact that in their early development they are much smaller than any cell of the body, measuring not more than 0.0008 mm. in diameter, and .secondly there is at no time any sign of a cyto- plasmic mass. These granules correspond closely in size and number to the chromatin bodies, possil)ly chromosomes, that occur normally in the spermato- cytes. In the pericardial cavity the spermatozoa are attaclied in considerable numbers to the wall especially along the dorsal surface, or at the time of the animal's capture were being swept along in a current passing beneath the heart and outward through the coelomoducts arising from the posterior wall. The dorsal section of each duct (Plate 8, fig. 3) is a simple tube of even calibie through- out, passing tlownward and forwartl from the pericardium to unite with the shell gland. Shortly before this union it unites with the duct of the seminal receptacle, which resembles a flask with a long curved neck. This last named organ like the dorsal limb of the coelomoduct is lined with cubical cells possibly ciliated, to which are attached a small number of spermatozoa. The shell gland is a globular body and almost totally lacks the cornua DONDERSIA CALIFORNICA. 155 prominent in the majority of species. On the .sides of the organ the epithelium is comparatively thin but it rajiiclly becomes thicker above and below owing to the excessive development of numerous gland cells. In this highly developed condition each cell is a goblet-shaped body with very slender stem and a slightly expanded base in which the nucleus is placed. The secretion consists of a granular mass much like yolk in appearance and staining reaction. Among these larger elements slender supporting cells, usually with subcent rally ])lac('d spindle-shaped nuclei, occur in considerable numbers. Both of these elements occur on the sides of the organ but as mentioned previously they are very lf)w and cubical or rectangular. Since the distal portions of these cells are dislodged in the apparently normal process of liberating the secretion it is probable that at times these lining cells are of greater height. At the postero-ventral surface of the gland a short narrow duct, whose position and general appearance are represented (Plate 8, fig. 3), makes its way into the cloaca. Owing to the similarity of the nerves and connective tissue, and muscle fibres, and the consequent difficulty of tracing these to any extent, the nervous system has not been examined. Dondersia calif ornica, sp. no v. One immature specimen taken at a depth of twenty-one fathoms off south- ern California (Sta. 4303) is the sole representative of this sjxH-ies. Owing to the fact that it bore a superficial resemblance to several small nemerteans it was killed with them in corrosive acetic destroying totally all traces of cal- careous structures. The general form of the body is shown (Plate 3, fig. 9). The length is 7 mm. and the greatest diameter 1.2 nmi. The pedal groove, and single, included fold, is continuous with the cloaca though at the point of union the former has be- come very indistinct. The outlet of the anterior pedal gland is a well-marked invagination with highly folded, ciliated walls. In this genus more than one dorso-terminal sense organ is present, two being found in D. festiva and three in D. annulata according to Nierstrasz ('02). In the present species eleven exist, all constructed on the same plan (Plate 35, fig. 12). Of these five occur along the mid line, and the others are not far removed from it. As Plate 6, fig. 2, shows they are not symmetrically disposed for of the six not in the mid line five are on the left side of the body. Each organ consists of a globular mass of slender cells, with elongated mesially placed nuclei, covered distally with a thin continuation of the cuticle investing the body. From the bases of the cells 156 DONDKHSIA ( ALIFOKMCA. fibres, probably muscle aiul nerve, judginK from other species, pass into the undcrlyinf; tissue. It is evident from the thinness of the cuticle (Plate 32, fig. 8) that but a single layer of spicules exists in this species and from spaces in the decalcified cut icie it is evident that they are of small size. The hypodermis is comparatively thick and comi)rises several classes of cells. The most conspicuous, and at the same time the most rare, are gland cells which are generally more deeply placed than the other elements between whicli their delicate ductules pass to the exte- rior. These are most abundant on the ventral surface. .\t all ]7oints there are almond-shaped spaces in the hypodermis which appear to have been filled with a calcareous product, and judging from the cell remnant usually in connection with them, it is probable that they are sjiicules in process of formation. The cells apparently responsible for the formation of the cuticle are columnar, non-staining elements containing a centrally placed nucleus. Between them are very slender cells with spindle-sha])(tl mesially placed nuclei which may possess a sensory function tliciugli tliis is not definitely established. No jiapillae are present. In this sjiecies the atrial cavity, pi'ovided with cirri and ridges, is entirely distinct from the radula-bearing region which communicates with the exterior by an opening immediately in front of the outlet of the anterior pedal gland (Plate 5, fig. 4). In this forward division the dorsal or innermost of the buccal ridges is lacking; the external one on the other hand is prominent and abundantly ciliatcnl. The cirii are united by their bases into groups of three or four or rarely six. In this enlarged basal part it is sometimes ]K)ssible to distinguish a few l)i])olar cells which connect with fil)res passing distally thi'ough the cavity of eacli cirrus, and in a reverse direction become lost to view in th(^ vicinity of the ganglionic musses surrounding the cirrose division of the digestive tract. .\s just noted this anterior end of the alimentary canal is separated from the succeeding portions by a narrow tract bounded by hypodermal cells and covered with a spiculose cuticle. Whether this division line disappears later in life, the cirrose section then communicating with the remainder of the gut, as is usually the case, it is impossible to say, though judging from the size of the specimen and the profound changes required to l)ring al)out such a state of affairs it seems probable that the jirescnt arrangement obtains in the sexually mature indivitlual. From the foregoing it develops that what Thiele assumes to be the true mouth is an independent opening communicating in the present instance with a com- paratively narrow jilicated tube leading into the larger pharyngeal, radula- DONDERSIA CALIFORNICA. 157 bearing portion whicli in turn connects with the stomach-intestine l^y a short oesophagus. Dorsal saUvary ghinds are represented l)y a small number of pyriform cells commimicating with the pharynx immediately behind the level of the brain. Slightly more posteriorly there are other similar cells, but they cannot with certainty be differentiated from the anterior pedal gland. The ducts of the ventral salivary glands open close to the mid line on each side of the forward end of the radula (Plate 23, fig. 5). Distally they make their way, as slender tubes, in a lateral direction and then expanding to twice their initial diameter proceed for a short distance posteriorly. To this expanded portion are connected multitudes of pyriform gland cells arranged somewhat intlefinitely into lobules attachetl to the lateral and ventral walls of the jiharynx. In 1). aiundatu Nier- strasz ('02) finds numerous cells situated about the ventral ducts; it is probable that they are the salivary cells whose ductules have been destroyed owing to faulty fixation. The radula is comparatively small and the teeth very transparent so that it is somewhat difficult to discover their exact form. Judging from cross sections each tooth consists first of a basal plate (Plate 8, fig. 8), narrow rectangular in form, and without any connection with the plates of neighboring teeth. This basal bar supports what appears to be a triangular median tooth, but high powers resolve this into a pair of elements closely appressed. It thus apjiears that the radula is monoserial, each l)ar l)earing a pair of conical cusps. On the other hand the radula may be considered biserial, the basal bar representing a basement membrane, but against this is the fact that the bars are not united with each other. There are not less than twelve teeth if the radula be con- sidered monoserial. Beyond the radula the digestive canal narrows, becomes folded longitudi- nally and opens abruptly into the stomach-intestine that after forming a short dorsal and ventral coecum develops the deep, characteristic lateral pouches with glandular walls. In the mid dorsal line the epithelium is differentiated into a fokl composed of high, richly ciliated cells which laterall.v become reduced in height ;ind gradually shade into the non-ciliated digestive cells. The relations of the gut to the cloacal cavity are indicated (Plate 6, fig. 2). The pericardium is spacious and the heart is of more than average size. The blood from the posterior regions of the body pours into its posterior division corresponding to an auricle (Plate 6, fig. 2), thence into a ventricle-like portion and from there is driven into the aorta. This vessel throughout its entire extent 15S I)()XI)KI{SIA ( ALIKOIJXICA. (lovol()])s branches, in roality ojicninss, coninuinicating with iiumcrous lacunae in the dorsal crcsl-liivc portion of the liody. Those in turn connect with others of less extent in the lateral regions and through these with the pedal sinus. In the head region the aorta breaks up into several sinuses which make theii' way through the anterior pedal gland to connect with the pedal sinus antl the more lateral spaces just described. In the posterior end of the animal the blood accuMiulates in large sjiaces surrounding the intestine and coelomoducts and pours into th(> lieart l)y means of a sinus passing dorsally on each side in the neighborhood of the reno-pericartlial openings. The pedal sinus contimies backward to the cloaca then ixxsses dorsally into a space beneath the shell gland and from tliere into vessels leading to the heart. Tlie gotuxd is distinctly paired, the two divisions being in contact only in I lie middle of the liody. Elsewhere they are widely separated by means of the dorsal aoita. In th.e mid section spermatogonia are fairly numerous and at all points ova are conunencing to develop. In the heart region the glands narrow and (•(imnnuiieate with the pericardium, which ]iosteriorly conununicates also with the co(>lonio(lucts opening into the cloaca. The dorsal division of these tubes is comparatively slender and is composeil of cubical ciliated cells without signs of glandular activity. No trace of a seminal receptacle is visible unless what appears to lie the anterior end of the ventral section may be so considered. This lower division, or shell gland, is composed of rather low colunmar cells, tending to form longitudinal ridges, but they likewise are inactive. Beneath the single opening of the coelomoducts the cloacal wall is developed into an outjiouching which in the adult animal may develop copulatory spicula or some gland connected with the egg-laying jirocess, though in the present specimen such functions are purely hypothetical. In shape this outgrowth resembles a thick set Y, having a median undivided section which opens into the cloaca and on the other hand connects with a blind ])oiu'li on each side of the mid line. The walls are not unlike those of the shell gland, consisting of columnar cells which are richly ciliated. Th(- cloaca or mantle cavity in this sjieeimen is of unusually small size though it may increase in diameter as the mature condition is approached. A glance at Plate 6, fig. 2, will show that in this species the dorsal commissure uniting the lateral nerve cords is placed unconunonly near the cloacal opening. If in the adult the commissure is customarily placed it might readily be shifted by the active growth and enlargement of the cloaca. The nervous system is typical. The brain is situated posterior to the atrial ICHTHYOMENIA POROSA. If)!) cavity, or cirrose portion of tho digestive tract, but witli reference to the i)luir}iix it is normally situated. As usual branches pass out from the anterior surface of the brain to unite with ganglionic masses about the bases of the cirri which they appear to iimervate and three pairs of connectives unite with the labio- buccal, pedal, and lateral nervous systems. Connectives and commissures, agreeing closely in number with the gut pouches, join the pedal and lateral cords throughout the body. In the region of the cloaca the pedal cords diminish in size and finally disappear, and the lateral cords likewise diminish considerably in calibre, and are united by a commissure which unlike the usual type, is devoid of ganglion cells. The labio-buccal comiectives, imbedded in the pharyngeal wall, attach to large ganglia lateral to tlie salivaiy ducts. Owing to the munbers of salivary ductules it is very difficult to trace nerves in this region and determine if there be more than the one commissiu'c imiting these ganglia beneath the pharynx. Ichthyomenia porosa, s]i. iiov. Upwards of twenty individuals of this species were taken in one dredge haul (Sta. 4400) off the coast of southern California, and two additional specimens were captured in the same locality at Station 4402. In both cases the bottom consisted of green mud at a depth of 500-507 and 542 fathoms respectively. All were unattached and there is no evidence whatever regarding their mode of life though it is possible that they may be parasitic upon some of the sea pens (Pennatul'idae), of which three species abound in this locality. There are slight inequalities in size due to differences in age and sexual maturity but the average length is approximately 10 mm. with a diameter of 1.2 mm. The head region is indistinct (Plate 3, fig. 4) and externally is chai'ac- terized merely by a very slightly greater diameter than that of the body. Poste- riorly, in an uncontracted state, the body terminates in a pointed extremity, but in other cases it may become blunt and where the cloaca is opened widely, trumpet-shaped. A pedal groove is present and as usual extends from the hinder border of the mouth to the cloaca! opening with which it is continuous. The opening of the anterior jiedal gland is usually very distinct, having the appear- ance shown in Plate 5, fig. 6. Of the spines covering the body by far the most abundant are exceedingly delicate, of a pointed ovate shape (Plate 37, fig. 1), 0.024 mm. long, and are imbricated, forming a single layer. In the neighborhood of the ventral furrow these are associated with a somewhat similar type, 0.0594 mm. long, with 1(;0 ICHTHYOMENIA POROSA. thickened edges especially on tlie rounded extremity. Scattered fairly regu- larly among the first variety are those of the second type, paddle shaped, with short handle and a length of 0.054 mm. Along the ventral furrow they are of a greater length, 0.01 mm. The hypodermis is a])parently one cell thick but the species is peculiar in iiaviiig the layer d('Vclo])ed into many transverse folds (Plate 32, fig. 7) especially on th(> dorsal surface, and in section these ridge-like elevations render the cell relations ol)scure. These wrinkles are more pronounced in some specimens than in others, and are usually more prominent in the anterior half of the body. In some cases they are doubtless due to reagents but usually they are certainly normal. The ordinary type of hypodermal cell is very slender, especially in the ridges and is provided with a relatively dense subcentrally placed nucleus. Accompanying these are numerous larger, more globular cells, apparently in large measure empty in preserved material. This may be due to the precipita- tion of some highly watery secretion, oi- more probably to the decalcification of some calcareous jiroduct. This species possesses upwards of fifty remarkable organs, apparently sensory, located chiefly about the anterior end of the body in front of the outlet of the anterior pedal gland. All are situated in the ventral half of the animal. Each consists of an invagination of the hypodermis (Plate 24, fig. 12) with an average depth of 0.1 mm. The lining cells are low, very indistinct and are pro\ills filled with a darkly staining, finely granular secretion clearly distinguished from (hat of the foregoing group. Anteriorly it rests against, and opens through, the posterior wall of the outlet of the anterior pedal gland, and more posteriorly forms a thin sheet resting against the ventral body wall and opening between five folds in the ventral furrow. Posteriorly these folds very soon disappear save one, the foot, and the accompanying glands diminish greatly. ICHTHYOMENIA POROSA. 161 The atrial opening is subterminal and opens into a comparatively large sized cavity (Plate 5, fig. G) in which the ridges (Mundleisten) appear to have no existence. On the other hand cirri are present in great abundance and in some specimens project from the mouth for a short distance. These organs may spring directly from the buccal wall, Ijut especially on the sides of the mouth they are borne in groups of from three to seven on stalks containing muscle and connective-tissue fibres between which there are extensive blood sinuses, enabling the animal to project the cirri through the mouth opening. Each cirrus is an unbranched process consisting of small cells with very indistinct nuclei and cell boundaries owing to the large amount of yellowish brown pigment. In all of the specimens sectioned numerous pigment granules have escaped from the cirri and at various places form small accumulations, but whether this is a normal process it is impossible to say. This secretion renders it also impossible to determine their innervation, a difficulty that is increased by the small calibre of the contained canal which in preserved specimens is too narrow to permit the entrj' of blood corpuscles. The cirrus-bearing section of the digestive tract passes abruptly into the succeeding region, the junction in every case being guarded by a distinct fold which thus appears to be a permanent structure. At first the buccal-pharyngeal walls are almost smooth and the epithelial lining, composed of cubical cells, is thin but opposite the mid section of the brain the lumen narrow^s, becoming trefoil shaped in section, and numerous small transverse folds have developed which now are of greater thickness. This condition of afTairs continues with slight modifications to the region of the ventral salivary glands where the canal becomes increasingly narrower and the corrugations more pronounced. As Plate 5, fig. 6, shows a clearly defined fold is present at the junction of the pharynx and stomach. Throughout its entire extent the walls of the pharynx are thick and are composed internally of heavy circular muscles to which are attached numerous radiating bands in.serted on the other hand to the body wall. All signs of a radula are absent and appearances suggest that this species like Drepanonicnia vampyrella subsists on some delicate organism, such as the sea pens, whose juices are extracted by powerful sucking movements of the pharynx. Two ventral salivary glands are present in the form of small tuliular out- growths opening on the underside of the pharynx about ojiposite the level of the hinder border of the brain. The cells are small but are filled with an abun- dant secretion, indicating that though these organs are diminutive they are functionally active. I(i2 K irniVOMKNIA POROSA. As indicated (Plate 5, fis- G), a dorsal coecum is but slightly developed, and I ho intestine from its junction with the pharynx to the cloaca is of uniform size and characlcr, antl is pouched in regular and characteristic fashion. Its walls are composed of liigh club-shaped cells in which the small nuclei are basally situated while the remaining portions are filled with large droplets of some secret ion unaffected by haematoxylin. Immediately beneath the gonad the cells ar(> nuich reduced in size, and are possessed of little if any glandular activity l)ut l)ear a heavy coat of cilia. Near the anterior end of the coelomoducts the digestive canal narrows abruptly (Plate 6, fig. 1) to form a small canal which arising near the ventral side of the animal makes its way dorsally on the under side of the pericardium (Plate 24, fig. 5) to open into the cloaca. The cells composing it are essentially the same as those lining the intestine beneath the gonad. The herma]ihiddite gland extends forward as far as the posterior limits of the pharynx or slightly beyond (Plate 24, fig. 4) and in a sexually mature animal contains ripe sex products throughout its entire extent. These originate in the usual fashion and are in no wise peculiar save that the fully developed eggs are unusually large, measuring 0.176 mm. Clearly defined tubes lead from the gonad into the pericardium, which in the .specimen represented in Plate 6, fig. 1, contains both ova and sperms. The antero-ventral pericardial wall is ciliated and elsewhere cilia appear to be present though the true condition of affairs is masked by the abundance of precipitated secretion. From the posterior end of the pericardium the coelomoducts arise as rela- tivc'ly narrow canals lined with almost culiical cells bearing a coat of ciha similar to those of the pericardium. Coursing downward and forward they gradually increase in size and the walls, retaining their ciliated coat, develop several folds before they unite with the limbs of the huge gland a short distance behind their anterior boundaries. At about one fourth of the distance from the pericardial opening to its outlet into the shell gland each tube originates what probably functions as a seminal vesicle (Plate 6, fig. 1). In calibre and histological fea- tvu-es each is similar to the neighboring jiarent canal with the exception of the distal extrcMuity which forms an enlarged, almost globular dilation. From Ijegimnng to end the rcsicula .s(mi))ale.s contain spermatozoa in most cases attached by theii' heads to the epithelial lining. Distal to the openings nito the vesicles the coelomoducts contain small quantities of spermatozoa, unattached. M the junction of the dorsal and ventral limbs of the coelomoduct on each ICHTHYOMENIA POROSA. 163 side a very large seminal receptacle is attached. Each originates as a slender duct, which pui'sues a tortuous course anteriorly, and opens on the lateral or latero-ventral surface of a vesicular enlargement with folded walls. The cells composing the duct are histologically essentially the same as those of the seminal vesicles save that they are of almost twice the height. Those of the dilation are likewise columnar and contain a glandular secretion which escapes distally in the form of moderately staining droplets. In addition to this secretion the receptacle contains numbers of sperms some of which are deeply imbedded in the walls. Whether these last named structures are intact or not it is im- possible to state; they show no clear signs of disintegration. The Y-shaped ventral section, or shell gland, is of large size and its walls, developed into many irregular folds, are unusually thin. Distal to the median, undivided section the cells of the epithelial lining are chiefly glandular, hemmed in by slender supporting cells, and are filletl with a violet colored vacuolated secretion in haematoxylin preparations. In the adjacent undivided region this type of cell is replaced by another of much greater length (Plate 24, fig. 2) in which the secretion is more vacuolated and stains less deeply. Associated with these are comparatively few elongated cells filled with a dark, coarsely granular secretion and very many containing in each a granule of a dark brownish color. These occur in the anterior half of the undivided part of the shell gland; from it the transition to the posterior half is very abrupt, especially dorsally where the cells become higher and are filled almost completely with a substance of varying character, depending probably on the nearness of the egg-laying season. In one specimen with sex products in the pericardial cavity these cells near their free surface contain one or two large spherical dark blue or violet globules, while the remaining cytoplasm, is packed with an almost homogeneous mass. Ventrally the secretion is more granular and -the more distal products are yellow- ish brown in color. In another specimen treated in identically the same manner these products have much the same appearance, but stain slightly. The lumen of the shell gland is spacious and opens in the dorsal part of the cloacal cavity near the anus. Opening by a wide pore jiosterior to the rejiroductive outlet is a large diver- ticulum of unknown function (Plate 6, fig. 1). Its walls are somewhat folded and are reinforced by a thick muscular coat (Plate 24, fig. 2). The epithelial lining consists of columnar cells of average height covered externally with a thin, sharply defined cuticular layer. Among the cells of this character are others, fairly numerous, very slender, with dense elongate nuclei, that in especially ICl GKMKRAl. CONSIDERATIONS. favorable matoiial may ho so(>ii to tpimiiiatp proximally in fibres passing into till" muscle layer. Distaliy they attach to the bottom of minute depressions ill the cuticle and therefore probably are sensory elements. In two sp(>cim(Mis the cloacal cavity has been widened greatly, completely exposing the anus and the openings of the shell gland and the more ventral diverticulum. This last named organ has been almost completely everted. These individuals appear in all respects to be normal. With th{> stains employed the nervous system is not sharply differentiated from the surrounding tissue, and accordingly but little has been determined save that relating to the larger ganglia and nerves. As may be seen (Plate 5, (ig. (')), the brain holds the usual position and gives rise to the customary nerves distributed to the atrium and the body wall and more posteriorly to the pedal, lateral, and labio-buccal connectives. The pedal cords are considerably enlarged at their anterior ends, while the lateral show scarcely any modification. Pedal commissures and latero-pedal connectives occur at fairly regular intervals through- out the entire length of the animal. The labio-buccal ganglia are located at the sides of the pharynx a short distance behind the level of the salivary glands, and are united l)y at least one ventral commissure. In the posterior end of the body the relations of ganglia and nerves are not especially clear, but the pedal cords appear to end in small enlargements, united with the termination of the lateral ganglia by one or two slightly enlarged connectives. As usual nerves pass into the hiiuler jiart of the animal from the ends of the lateral ganglia which are united by the usual suprarectal commissure. GENERAL CONSIDERATIONS. If unanimity of ojiinion be any criterion whereby we may judge the correct- ness of a theory it must be admitted that we are yet a long way from the solution of the origin of the Mollusca, for scarcely any two investigators hold identically the same views. In their development or in their adult organization many of the meml)eis of the phylum exhibit features which are the close counterpart of others in the fiatwonns and annelids and it may well be that, generally speak- ing, those students are correct who hold to the idea that all have descended from a common ancestor, though the details of the process are most obscure. Narrowing down the problem to the Solenogastres there are few who dissent from \ho opinion that they are true molluscs, though it cannot be said their I)osition witiiiii the group is definitely established. However it is becoming GENERAL CONSIDERATIONS. 105 increasingly evident that they possess more characters in common with the Chitons than with the other classes, and these characters, interpreted in the light of community of descent of the two groups, are more readily understootl than from any other viewpoint. It is little more than waste effort at the ]iresent time to attempt to recon- struct the external characters of the ancestral Solenogastre, for it is generally agreed that the present thiy species, worm-like in form and without shell oi' well-developed foot, are highly modified in these resjiects. The aggravatingly few facts of their embryology are also without nuich value for the .solution of the problem. As the matter now stands there is no positive evidence that they ever had a shell, but in view of the fact that these animals show a close resem- blance to the Chitons in several other respects it is not unreasonable to believe that one was formerly present. It must be admitted that Pruvot's figure and description regarding the shell in the larvae of Dondersia banyulensis are very indefinite, and have led some authors to claim that the seven valves of the dorsal side are in realitj' greatly enlarged scales. There are some evidences that such is indeed the case in the young of Hohnncnia gravida, at all events the plates do not develop exactly as does the Chiton shell. It is possible that we have here the confirmation of Blumrich's theory that the original shell arose by the ex- cessive development of flattened spines along the dorsal surface of the animal. It is a significant fact that the mantle of the Solenogastres has no counter- part save in the Chitons. In the least modified condition it consists primarily of a single layer of epithelial cells overlaid by a cuticular covering often of enor- mous thickness. Those probably responsible for the formation of the cuticle, and of pigment when such is present, are comparatively simple, unmodified, more or less columnar cells. At frequent intervals throughout the layer spicule- matrix cells arise and develop a single stratum of spines, or several layers im- bedded in the cuticle. In their mode of origin the spines of the Solenogastres are essentially similar to those of the Chitons (see p. 29). Thiele declares that the Solenogastre spicule is produced from one matrix cell, but this method is certainly not frequent, and Plate states that it is rare among the Chitons. Wiren reports that in Chaetoderma there is one basal cell and three smaller ones en- compassing the young spine. Hubrecht discovered that in Proneomeiiia sbiiteri the base of each spicule is grasped by a considerable number of matrix cells, and as the spine is carried outward by the continued growth of the cuticle they continue to retain their attachment for a long period. Pruvot ('90) finds in a few species that during the early development of the spicule four or five cells |(i(i GKXKRAL CO.XSIDKIJATIONS. arc ill cunt act witli it, and the aitachnicnt may jiersist for a considerable period. 1m I'ronconienia hawnricmis and a numbiT of other species I have found that there is one i)asal cell, apparently responsible for the formation of calcareous material, surrounded l)y seven or eight smaller cells attachetl also to the base of the spine and perhaps responsiltle for the formation of the cuticular sheath. This last nanuMl mode of formation is almost the exact duplicate of the most common method of spine development of the Chitons (Plate, 1901, p. 372). Among the Chitons the matrix cells retain their connection with the spicule as long as it exists. In the Chaetodermatidae, and in those Neomeniina with a single layer of spines, there is a tendency to follow this primitive method. The same is true, though in some cases to a more limited extent, in a few species with tliick cuticle and several layers of s])iculcs, notably P. haivaiiensis and Sln>j)li<>invnia scandens. The balloon-shaped papillae dcvclojied from the hypodermis and in a fully developed condition extending to the free surface of the cuticle are of proble- matical significance. It has been suggested (Kowalevsky & Marion, Wiren) that they may be spicule-matrix cells that have assumed some new function after the formation of the spine; but the fact that in P. hawaiiensis the matrix cells retain their attachment, at least in part, as long as the spine is imbedded in the cuticle, jirecludes such a possibility. In .Vlexandromenia there are many times more spines than papillae. Heuscher on the other hand describes their origin as simple outgrowths of the hypodermis. Regarding their homologue in the Chitons nothing may be claimed definitely. They may correspond to the packets or papillae (Plate) or, with a greater degree of probability, to the aesthetes as several authors have claimed. That a foot of much larger size existed in the ancestral Solenogastre and that the ventral groove of the Neomeniidae does not "represent the first stage in the formation of that pedal surface of the body which is seen in the lowest mollusca" (Ciegenbaur) is indicated by a number of facts. In the first place although no external trace of a foot exists in the Chaetodermatidae there is a space along the mid ventral line between the longitudinal somatic muscles which are thicker here than elsewhere, and in Limifossor this same space is occupied by a sinus exactly similar to the pedal sinus of the Neomeniidae. In Limifossor the pedal sinus anteriorly penetrates a clearly defined septum and communicates with the head cavity as in the Chitons. It is much more reasonable to consider that the pedal sinus is the renuiant of the foot of the ancestral Solenogastre than that it is the first sign of the appearance of a definite creeping surface. GENERAL CONSIDERATIONS. 1G7 In the Neomeniina there is the same cleft in the ventral nnisculature, thickened as in the Chaetodermatina, and the foot is present as a small fold extending along the mid ventral line. Anteriorly it affords an outlet for the enormously developed anterior pedal gland which in position and development as far as this has been traced (see Heath '05), is homologous with the pedal gland of young Chitons. The remainder of the foot is su]ip!ied with the poste- rior pedal glands which are present in a diffuse condition in the Chiton foot. In the Solenogastres there are no eyes, tentacles and even the proljoscis or snout of the Chitons is believed generally to have no homologue in the group. Concerning this last named organ however Thiele claims (and in this he is fol- lowed by Nierstrasz and Pelseneer) that it is ]iresent, though in a highly nKxlificd condition and I am strongly convinced of the force of his argument. In the first place it is a well-known fact that the first section of the alimentary canal in the Neomeniina contains the atrial ridges (Mundleisten) and the enclosed cirrose area, all innervated by fibres originated directly by the cerebral ganglia. In Rhopalomenia aglaopheniae, Dino/ucnid hubrcchti, etc., and I have found the same state of affairs in Drioynenia pacijica, this portion of the canal exists in the form of a depression in front of the niouth, and is separated from it by a narrow bridge of spicule bearing cells continuous with the general covering of the body. Owing chiefly to its innervation Thiele considers that this "sensibles Atrium" is the homologue of the Chiton snout which is now withdrawn into a depression. Where this atrium is a direct part of the digestive tube the true inouth has been drawn into the body and is located immediately behind the most posterior atrial ridge. It now becomes an interesting fact that in the Chiton development the mouth at first is posterior to the snout and but slowly takes up its final central position (cf. Heath '99). It lends support to the belief that in these molluscs with isolated atrial cavity the position of these organs does not represent a highly modified condition but a relatively primitive state of affairs. In the Chaetodermatidae the nerves which originate from the anterior surface of the brain pass at once into the huge, compact ganglionic bodies homol- ogous, I believe, with the more diffuse nerve masses in contact with the bases of the cirri and buccal sensory ridges in the Neomeniidae. From these a rela- tively enormous num]:)er of fibres pass into the Mundschild or buccal plate, just as the cirri and ridges are innervated by nerves from the neighboring ganglia. There is thus little doubt that the buccal plate and sensory atrium are homolo- gous and if the above line of reasoning be correct, they are the homologue also of the Chiton snout, which is likewise innervated by nerves from the cerebral ganglia. IGS GHNEKAl- COXSl DERATIONS. licfiardins thi' inaiitlc cavity I Ik'Ucvc tliat it as truly exists in the Soleno- pistres as in tlic Cliitons or prosoliranchs, for example. The sole reason for considcriufi that the so-called branchial or cloacal cavity is a secondary modifi- cation apiK-ars to have chiefly originated with Thiele who claims that the bran- chial lamellae of various neomenians are highly developed rectal folds and accordingly the branchial cavity is nothing more than a greatly expanded rectum. The same is true of the Chactodermatidae, for the plume-like respiratory organs are said to have been developed from similar rectal folds, and accordingly their remarkable resemblance to true ctenidia is of no especial significance. Further- more the fact that the nephridia open into this space is likewise of no importance for it is of coenogenetic origin. Regarding the branchial plates of the Neomeniina, they certainly have every appearance of being merely folds of the walls of the branchial cavity, but that they are closely related phylogenetically to the respiratory organs of the Chaeto- dermatidae is an entirely different matter and one most difficult to substantiate. On the other hand it seems to me that the gills of this last named family are not clearly homologous with the neomenian resjiiratory organs. In a former pa])er I have called attention to the fact that in gross and microscopic appearance, blood supi)ly, and innervation they are practically identical to the Chiton gill. Such an idea brings us without violence to the belief that in the original ancestor of the Solenogastres and Chitons there was a true mantle cavity containing at least two ctenidia, the separate outlets of the urogenital system and possibly an osphradium though such an organ may well have been in a difTuse condition as in some of the modern lamellibranchs. Accordingly the connection between the pedal furrow and the mantle cavity is not secondary but primitive and similar in its broader features to what is found in the Chitons. The polybran- chiate character of this last named group yet remains a puzzle for as I have pointed out ('05) there is nothing in the development of these organs to indicate if it be primitive or not. While several fundamental differences between the circulatory systems of the Solenogastres and the Chitons have been found to exist these may be ex- plained to some extent on the supposition that originally the foot was of larger size, and in any event they do not outweigh several remarkable resemblances. In both the pericardium is dorsal, posterior, and communicates with the exterior with paired ducts. In present day species it contains the heart, a simple tubular organ or differentiated into a ventricle and single auricle, which may originally have been paired as Wiren has ingeniously suggested. From the anterior end GENERAL CONSIDERATIONS. IG!) of the heart the aorta arises and passing along the dorsal side of th(> gonad, that it supplies in Chiton-like fashit)n, it reaches the head cavity. In the great majority of species this last named organ is not clearly defined, but in Limifossor it is separated from the visceral cavity by a connective-tissue septum as clearly defined as in the Chitons and having essentially the same relations. Within the head sinus the blood makes its way Ijy irregular channels into the visceral cavity and passes backward. In Limifossor the septum is perforated ventral to the intestine and through this the blood makes its way into the ventral sinus. A special visceral artery or sinus is lacking within the group, its function being taken by the general visceral cavity and ventral sinus. This last named space communicates freely at many points with the visceral cavity and posteiiorly both unite and the combined vessel makes its way to the ctenidia and from thence into the heart. Hansen years ago noted the presence of crystals in the coelomoducts of Chaetoderma, and considered it possible that they may f unci inn as kidneys. From much more extensive studies Wiren has taken the same position, showing the close similarity of the tissue to that of the Chiton kidney. Another fact of the greatest importance is that in Chaetoderma erudita for example, where the sexes are separate the nephridia of the male are eiactly the same as those of the female. If the coelomoducts here act in the capacity of shell, mucous or other glands intimately connected with the egg-laying process it is reasonable to suppose that it would be more highly developed in the female than in the male. Since it is not it becomes much more prol^able that the resemblance of the cells of the duct to those of the Chiton kidney is not accidental, Init that they are true excretory elements and the ducts therefore have retained the excretory function derived from the ancestral form. Stating the matter in another way it appears that the coelomoducts are, from the standpoint of both structure and function, of a more primitive character in the Chaetodermatina than in the Neomeniina. In the latter family the ducts have assumed an important role in the storage of sperms, or in the development of envelopes for ova and perhaps other processes connected with egg-laying, so that the original function of excretion is effectually masked, if it exist at all. In the Chitons the kidneys become active excretory organs long before any trace of the gonad or its ducts appear, and if these tubes in the Neomeniina act as kidneys they likewise would probably assume their duties at an early stage. In Dorijmenia acuta, however, there are no signs of such activity in individuals 14 mm. in length. The reproductive glands are present, though in a very im- 170 GKNKRAL ( ONSIDKItATIONS. mature coiidition, and tlic cocloinodiicts, haviiifi Ihc i'uriii of ranals of al)out equal ealibre throuf!;li. 757-759. '88. Sur les especes de Proneomenia des cotes de Provence. Comptes rendus Acad, sci., 106, p. 529-532. bibliography: 177 Marion, A. F'. Sci' Knw.ilfvsky, A. MoBius, K. '75. Vermes. Julirefil). Kimiiii. wiss. uiitcrs. Dents. Mcrrc. Kiel, JmIir' 1S72-73, p. 153-170, pi. 3. NlERSTK.\SZ, H. F. 1902. Till' S()leiicis;astri's of the Siliosja-cxpcilitioii. Sil)oi;;i-rNi)., Monoj^r. 47, Id p[»., i; pis. 1903. Das her/, (k'r S()k'iioj;:asti-cii. \'crli. .Vkad. AuistiTdaiii, scr. 2, 10, .")1' iJp., 3 i>ls. 1903. .W'Uf Solfiiogastreii. Zunl. jalirli. AKth. aiiat. oiil., 18, p. 3.59-3S6, pi. 3.5-3(J. 1905. Knijipoiin iiid iiiiiiiiini ami llic railula ilcr Soleiiogastrcn. Zool. jalirli. Alitli. aiiat. out., 21, ]). t)r).")-7()L', pi. 30-41. 1908. Dio aiiiijhiiK'urcii. 1. Die .Solciiogastreii. Ergvlm. zool., 1, ji. 24()-3t)li. NORM.\N, A. M. '79. On the occurrence ol' Necinicnia (Solenopusj in the British Seas. Ann. mag. nat. hist., ser. 5, 4, p. l(J4-l()(i. Pelseneer, p. '90. Stir le pied dii ( liitDnelhis et des .\placophora. Bull. Sci. France et Belgi(|ue, 22, p. 4S9-495. '96. Les reins, les glandes geuitalcs et leurs conduits dans les inollusques. Zool. anz. 19, p. 140-145. 1900. Recherches niorphologi(|Ues et phylogencti(iu<-s sur les nlolh^-^(ples archa.iqucs. Mem. eour. .Vcad. my. Bdg.. 57, 1 12 pp., 24 pis. 1901. Les neonieniens ile I'e.xpedition antaretiiiuc beige et la distrihution geographique r p.'dal ganglion. peg |ircc<'rcl)r;d .nanghou prill pcrii-anliuin. pli pharyn.x. pi l.ileral ganglion. ps pi'dal sinus. r radula. r;> seminal recejitacle. rs radular su|>iiorl. s anterior vertii'al scpruin sc subradular coniniissure. sg ventral salivary gland sgl shell gland. sn svibradular ganglion. so sense organ. sp spii'ule, sr iliHs.il gill rcl ractcir. sro siilir.iclular organ. sto stoiiKicli. sv seminal vesicle. vr ventral gill retraiior, vs ventral diaiiliragm. PLATE 1. PLATK I. Fig. 1. Slrophdmcnia farciincn. X :!. Fiji. -• Slrii]''"'"!''"'''^ °P'''''''"'''- ^ ''^' Fig. 3. Slro|iliiiincnia siiinosa. X 5. Fig. 4- IIi'rpciiiKMiia plalyiioda. X 10. ]■"[«. o. Slniplidiiicnia triann'i':"'"*- ^ ^- "Albatross" Pacific Soi£Nor,AsrHF:s Pi .ATE 1. Wi,A' , !-■ , ^<- ■^^e!^ 5>^V,l/,l I \\>iV\v >/ ■Ji'^l V"'^ '^!'^i^^^^ -^-^..f.^ ^^/', 4 •sa^ '^'^t ^^•^: - -*>, ^•tt I ^ HHealh Jel 3I'!»:ise. I'tn ficisturi TLATE 2. I'Ig. PLATE 2. Fi;;. 1. Strnphnraenia srandens. X 4. Kin. -■ DiT|)anomenia vampyrella. X 7. Imp;. ;!. tlhaotoderma hawaiiensis. X 8. Fin. '• Li>|)homema spiralis. X 7. Alexaiidromenia agassizi. X 4. Fin- •)• l.iiiiifossor fratiila. X 5. "Albatross" Pacific Soi^nogastres. PiATKg. % -..■-:.,.. <^^^_m.^^ # -^^ < m.:- \ ^n, \ 'i/ . '^. ,n^ >:^ ^.«i I J'" PLATE 3. V\'A 1. IMS 2 Fii; 3 Vi'A 4 Vvi r, Mm; (1 I'isi 1 Fitf s Fifi 1.1 I'-iK U) V\)i 11 PLATE 3. H:duiiienia gravida. X 0. Strophomenia farcimi'ir, vmusuall.v llmk t^lii'i''"H'n. X 7. .Vlpxanili'omciiia valida. X ■i. Ichthyoinoma jiorosa. X 9. Ucconstruflion of poslcrior ciid cif llalomi'iiia fjiavida. ll;dimirnia M;i"ivida. AiitcTiiir I'lid. • HK Chai'lcidcrma .iaponii'a. X l. lli-ad ofChai'tudiTiua .);iiKiuica. X Hi. DondiTsia i-alil'oniica. X 9. Proiu'omi'iiia liawaiieiisis. X 2. Ddrvmi'iiia ai-ula. X 3. "Albatross" Pacific Solenogastres PlATF,3. t'\i'i '^'1/ ,iM^''^y /■ '" " " -^ / w %-'>->'^ '■^i^^p?*^' ^ '■MM r s'"^, V' t.^:«V r c. I. A ■'.' H.Heatli del PLATE 4. I'lS- 1. Fig. O Fig. 3. Fig. 4. Fig. 5. Fig. 6. Fig. 7. Fig. S. Fig. 0. Fig. 10. Fig. u. Fig. 12. Fig. 13. Fig. 14. Fig. 1'). Fig. 1(1. Fig. 17. Fis. l.S. Fis. 11). PLATE 4. Cliaclodeima luuuilu. X 7. Chaiioderma .sfabra. X 7. Cliaetodcrmaattcnuata. Livinjisppcimen. X 'i. Chaetoderraamontereyensis. X 1.3. Chaotoderraa robust a. X 1.3. Chaotoderma californica. X 3. Chai'tddermaargi'iitca. Living siwciuien. X 2.5. CliactDdennainontprcy^si..^. X 2. ClKuModrrmaorudita. Living sppi'iniPii. X 2. Cliaclodprniaatlcnviata, front and side vipws. Living spcrinirn. X 10. CluK'lodcnna pnidila. X 15. Cliaplodornia nanula. X 15. Liniifossor fral ula. X 13. Chaptiidprnia nionlcrcypnsis. X lo. Dondprsia palifornioa. Anterior end. X 22. Chaptodprniaspal)ra. X 15. Chaetoilenna nionterpyensis. X 10. , Dondersia ealifornica. Posterior end. X 22. Cliaetodprmarobusta. X S. "Albatross" Pacific Solenogastres Plate 4-. «@3Si 8 iM fl II II ^ ^^S in I ?r 11 .,i I .-"^^ II |{ 14- lb 73 , . '■i'''>'.y,it'4 , yv 7/ &?'>'"■ /.O HHealliW BMelse'.lilh Boston PLATE 5 I'LAT}'] r,. Kill;. Fii;. I'is. I'osli'iior iMiil 1)1' Ch:ietoilcnii;i altfiuiula, living speciiucii. Kcc(jiistriK-liuii of imtcriur end of I'ruuronicniii luiwiiiionsis S;unc of lI;iloini'ni:i, graviiUi. Same of Donilrrsia caUfornica. I'oslri-ior end of Alrxamlnmicnia ai^assizi. Anlcrioi- rnci of li-litliyonii'nia jjorusa. "Alhatross" Pacific Solenogastrks Plate 5. ■^^ g^^3Wwv,-.v-?'"-.»i^;^^^^v,' ?on A B C .^.=. ii'^^%^"^ )^ A ' \ A?*^"?. ^■■^ ■pf ^~,s^'fr^«^ 50 n II Heath del. B Meisi!l lith Bfjstoii PLATI': tj. PLATE (i. FiS- 1- Hpixiiislrui'tidii nf posterior onil of Ii-htliyoiuriii;i porosa. Fig. 2. Same of Dondcrsia (•alifornina. Fig. 3. Sami' «R«¥£^-. . 4 ^^^: ''~jr5>-i. 5on int ■■'■' br ■"^"'■^"' ai M:- .■-•:«s?i!F dSO '>'" 4 P^*- '"#. —^^' ^'^{ Hiieathdel. B Meir.el. U'l Boston PLATE 7. PLATE 7. ViH. 1. Ui'cDnslriii'lion iif antrrior riiil iif Sfr<>|iliniiicni:i fari'irncn. Via. 2. Saini^ of aiiti'rior rml of ClKU'lnilcnna liawaiicnsis. Vv^. '■^. Sanir of anterior ciiil of Ali'xaiulroiiK'nia asassizi. Fifi. 1. S:nnc ol anliaior rml of I )n'|iano!iii'nia vaini)yn'lla. Fis;. 5. Sainr nf |in-;| rrior I'li'l of Alcxanili'oiiicnia a;;assizi. "Albatross" Pacific Solenogastres Plate 7. ■3 t K Keath del BMeise! 'ilji Boslun PLATE 8. PLATE 8. Fis- 1. R'-con-ilnidion ol' iuitcrior oml of Hprponienia plalyiioila. Via,. 2. Same of aiitrrior end of Strophoiiionia spinosa, small sperimcn. Fig;. 3. Same of posli'rior end of HiTpomcuia platypoda. Fiji;- ^• Same of anlorioi' cuil of rroiicomt'.nia insularis. Fii;. ■). Same of aiiliTior end of Slropluimenia opliidiana. V\!i. (). Same of po-. PLATE 9. PLATK n. Reconstruction of the iKisterior vnA of tlie l.ody from Ininsverse sections Fig. 1. Strophomcnia oplu(li:ina. Fig. 2. Dorynienia acuta. F'ig. 3. Stiophonienia triangularis. Fig. 4. Stropliomenia spinosa. Fig. 5. Loiihoinenia spiralis. "Albatross" PACinc Sole-nogastres Plate 9. H Heath del. B Meisel lnh. Boslon PLATE 10. PLATE 10. Figs. 1-7, !I-U). Sections of Limifo.ssor talpoideus; Fig. S. Chaotodcrma crudita. Fij;. 1. Fig. '2 Fig. o. FiK- 4. I'ig. 5. Fig. (i. Fig. 7. iip|M irl s. V\k. S, Fig. 0. Fig. in. Side viow, living s|K'cinirn. X 5. Anterior end, living specimen, .showing extremes of motion of frontal sense organ. Lateral view of organs of jjosterior end of body. Same of organs in anterior part of body. Diagram of radiila and muscles that open and close the teeth. Section through dorsal sense organ. X 13o. Lateral view of buccal mass, showing more important nuiscles that operate the radnia anil IJrain and labio-buccal nervous system. Spicules from miildle of body, front and side views. X 11(1. Dorsal view ol radula and its supiiorts: luirtion of dorsal pharyngeal wall removed. "Albatross" Pacific Solenogastres Plate 10. H KeaUi del BKeisdlttBoslon PLATE 11. PLATE 11. Fin. 1. Roconstruction of the posterior oml of Drepiinomonia vampyrcUa. Fi};. 2. Same of anterior end of Stroi)liomenia spiiiosa, larjje speeinieu. Fig. 3. Same of anterior end of .\lexandroinenia valida. Fig. 4. Same of posterior end of Stropliomenia fareinien. Fig. 5. Same of posterior end of Proueonieuia hawaiiensis. "Albatross" Pacific Solenogastres Plate U. H.Heatkdel, BHeise] i)th Boston PLATE 12. FiS- 1. Roconst ruction of interior end of Strophonienia scandens. Fiff. "2. Section through dor.so-terminal sense organ of Alexandromenia valida. I''ifl. i!. Ixceonst ruction of (losterior cud of Stroi)honu'nia .si)ino.sa, small specimen. Fig. 1. Reconstruction of iiosterior end of C'haetoderma attenuata, showing principal nerves, two (brn) ]);issing into the gill. Fig. 5. Reconstruction of anterior i>arl of nervo\is system of Alexandromenia agassizi. Fig. I). Same of ])oslerior end. "Albatross" PAcmc Solenogastres Plate 12. / \V|-'i/Vi \ rJ H.Heati it. EM»;ser-s Btfsion I'l.ATE 13. PLATE 13. Fig. 1. Reconstruct ion of anterior rn.l of Dorynionia acuta, sliowinn; nervous system. Labio- huccal system heavily sluicli'il. Fig. 2. Same of posterior end of Slrophomenia seandeus. Fig. o. Same of aiilcrior end of Cliaetoderma ervidita. Fig. 4. Same of posterior end of I'roneonienia liawaiicnsis. "Albatross" Pacific Solenogastf'-: Plats 13. c\ . ^ T""'' ih^t^'^i PLATE 14. I'LATE 14. Cross sections of Proiicomenia liawaiieiisis. X 33. Figs. 1, 2, 7 oorrfsjiond to liiir-s A, B, C in lif!;. 2, y\. r, (anterior cml) ami figs. 5, {), S to D, E, F in fig. 4, pi. 13 (posterior end). In lig. 1. ini, oin represent inner anil ouli'r atrial riilge; os, external sensory ridge. Fig. 3. Seel inn tlirough middle of liody. Fig. 4. Section throngli animal aliout one liflU body length from anterior end. Fig. 9. Section throngh liody a slmrt dislaiiee behind pharynx. "Albatross" Pacific Solenogastres Platf.14. .^■"-'_ 5 fe'O.Xr, '/ '4:-'-'"'^ 53' =><<, V-, s%^ ■•"" ,/ / i.neath del. SMeweU'ff. Boston PLATE li; Vi'^s. 1-7, 0-12. Si'clions of Dciryiiicnia ;icul;\. X ')!); Fifj. S. Halomonia gravida. Figs. 1, 2 corros])oiul to lines A, R imlicatcd in titj. 7, ])]. 1") (anterior end), and figs. 4, 6 to D, C, in fig. 4, ])1. l) (posterior end). Seetlon throiigli dor.so-terminal sense organ. Cross section of ]ienial spine. X 20-"). Longitudinal section of anterior end of liod.\'. Cross section corresponding to line D, fig. 5, pi. 3. Cross section of yonng animal, ])osterior end. Longitudinal section through lia.se of penial s])ine. X 20.5. Section through dorso-terininal sense organ. X 20.t. Section through atrial cavity. Fig. 3. Fig. o. Fig. 7. Fig. S. Fig. 9. Fig. UI Fig. 11. Fig. 12. "Albatross" Pacific Solenogastre;; PlJ^TElS. HHeati del ? Meisel 'ith Boslor. PLATE 10. 1' iK 1. Fig 2. I'iS 3. I'ii; 1. V'i'i f). FiK i; iMK 7. FiK S. Imk 0 PLATE 1(1 Cross sfptions of SlrophomPiiia srandons. X 33. In rps'ion of lirain. Bcliinil pharynx. In niidillc of l)ody. Tliroufj;li |iliarvnx opposite anterior pedal gland outlet. Thrimnh seminal reeeptaeles. 'Ilirough raid .seelion of eoelomoduets. ()|ipusili' junction of pharynx and mid gut. Through jiiiiciioii of eoelomoduets. Througli oullet of eoelomoduets into elo.aea. "Albatross" Pacific Solenogastres Plate 16. H Heath del. ?Meisellilti Boston PLATE 17 PLATE V, Figs. 1-7. Strophoi.icnia s|)ini).sa, X 33. Figs. S, 0, 13. S. oi)hi(li;iiia, X 24. Fig.s. 10-12, 14-17, S. farciiiien, X 33. Figs. 1, 2, 3 fross srclioiis of S. spinosa (large spnrimni) along lines A, B, C indicated in fig. 2, pi. S (the lines .\, H sliould be perpendieular to the entiele). Fig. 4. Cro.ss section through radida and salivary gland outlet, S. spinosa. Figs. .'), 0, 7 are through fig. 4, ])!. !) (posterior end). Figs. S, (I, i:i are cross sections of S. ophidiana along lines A, H, C of fig. 5, pi. S. Figs. 10, 11, It) are cross sections of S. farcimen along lines. A, B, C indicated in fig. 1, pi. 7 (ante- rior end). Figs. 14, I.'), 12, of same species, corresjKind to lines D, E, F, of fig. 1, pi. 1 1 (posterior end). Fig. 17. Spines of Strophouienia farcimen. X 210. "Albatross" Pacific Solenogastres Plate 17. HHeatl. dtl. B Meisfl kli BkIoii PLATE 13. PLATE IS. FiRs. 1-4. Slr(i|ilic)inonia o|ihiiliana X 24. Fiss. .5-12, S. Iriangiilaris. X 5.5. Fig. 13. S. farciinen. X 00. Fig. 1. Cros.'; section of S. ophidiana along line D of fig. 5, pi. S. Figs. 2, 3, 4 of same species are along lines, E, F, G, fig. 1, pi. 9. Figs. 8, 12, 10, 11 of S. triangularis correspond to lines D, E, F, G, fng. 3, pi. 9. Figs. 5, 6, 9 correspond to lines A, L5, C, iig. 6, pi. 36. Fig. 7. Section through mid gut behind phar3'nx. Fig. 13. Section through mi in lig. 1, pi. ,S. >, liO. Figs. 10, 11. Seel ions llirougli ])oslerior end of H. iilalypocla along lines D, E, (wliich should incline to left) Hg. 3, pi. .S. Fig. 12. Section I hrougli dorso-terminal sen.sc organ. X 205. Fig. 13. Seclion through .salivary duels, showing entrance of ductules from gland cells. Figs. 11. 1.5. Longitudinal sections through posterior end of IL i)laty]joda. "A L 8 AT RO S S" PaC 1 Fl G S 0 LE N 0 G A ST R K S Pi.atf; 19. ' '' ao lRsko,^,(;^5" , /» / "^ V-- \ v sgl -% ■f® ''^ / .y // \ ii f--Sg TT V' ■ i!^ //I rf'" /r) Vhi .'^ ^■'"'v^^. ''mm sg ,vA^SS, |il. 11 (anterior end). Figs. 3, 5, (). Through posterior en. lalpoideus. X 255. Fig. 13. Section through heart and ])ericardiuni. X 50. "Albatross" Pacific Solenogastrks i;-:EEl. 1 ^'l^U, s ■< I / / #i?>'^^'-' It / ii? '^# ' ^'1^>-' 4 V oi* •'T' 0*' ^-'^i V. % -7 . ^; ,x /o 5.«'>P -"^Jo :' (anterior end). Fig. 4. Section through body behind pharynx. Fig. 5. Salivary gland lobules opening into main duet. Figs. 6, S, 11 (see fig. S, pi. 15) correspond to lines E, F, G in lig. .'i, pi. 3 'posterior cntl). Fig. 7. Longitudinal section through advanced larva, showing cerebral ganglion chain of nuclei, stomodaeum, st, and early stage in the development of the anterior pedal gland and outlet, f. Fig. 9. Section through anterior division of heart. Fig. 10. Section through junction of two divisions of heart. Fig. 12. Section through dor.so-terniinal .sense organ. Fig. 13. Spines from middle of body. X 300. "Albatross' Pacihc Solenogastres Plate E2. H Heatl del. PLATE 23. PLATE 23. Cross seel, ions of Donilersia californica. X 60. Fins. 1, '2, 3, 5 ('i)iT('spoii(l to lines A, H, C, D, fifj. 4, pi. .5 (antci-ior end). Fip. 4. 'Plirounh anterior end of mid s?ut. Fifrs. (■), (I are alonn lines G, F of tig. 2, pi. 0 (posterior end). Fig. 7. 'rhongh nii^>: : I /-Vf. 7 \ ) ay ■f^^-f H HcaA del PLATE 24. PLATE 24. Figs. 1-5, 11. Section.s of Iflitliyoincnia ijorosii. X 55. Fig. 0-10, 12-14. Stroplioinenia regu- laris. X 55. Fig.s. 1, 4, 11 of I. porcsa, correspond to lines B, C, A, fig. 6, pi. 5. Figs. 2, 5 are along the lines E, D, fig. 1, pi. 6 (posterior end). Fig. 3. .Section through middle of body. Figs. 6, 8, 9, 10, of S. regularis, are along the lines, C, D, H, A in fig. 8, |:)1. 26 (in fig. B read el in place of ct). Fig. 7. Section of cuticle and hypodermis of S. regularis. X 205. Fig. 12. Section fhrougli one sense organ of anterior end of I. porosa. X 555. Fig. in. Section through one seminal receptacle and several stalks of S. regularis. X 255. Fig. 14. Blood corpuscles of S. regularis. X 555. 3" Pacific Solenogastp.es Pi.ATK24. D*^^ I ^#^1 i' s§l / I ... I; f- m /« pen' ., ■ / ^ ml J3 1^^'% ii II H Heath del. PLATE 25. PLATE 25. Sections of Chaetodermaattcnuat a. X 33. 1. Section tlirougli brain. 2. Section tliroufili radiila. 3. Section througli junction of stomach and intestine. 4. Section tlirough suprarectal conunis.sure. 5. Section tliroiigli outlet of coelonioduct (pores omitted in figure), (i. Sectiim llirougli posterior end of jjrothorax. X 00. 7. Hypodermis, showinj;; gland cells and attached fibres. X 555. 8. Section through heart. Fig Fig Fig, Fig Fig Fig Fig Fig Figs. 9, 10. Two successive sections through the subradular ganglion and connectives. X 135. Alb/jross" Paciffc So:£nogast^j:s Platk25. da Ui PLATE 26 PLATE 26. Figs. 1-7. Chactodcrma argentea. ilrniniMiia agas.sizi. X 33; Fig. S. Strophcmrnia rcgulari.s. Fig. 9. .\lexan- Fig Fig Fig Fig Fig, Fig, Fig, Fig, Fig 4. 5. G. 7. S. 9. riirciugh brain region. riiriiugh radula. riiniugli heart and eoeloinoduots. riirougli suiirarcrtal commissure. rhrougli region of outlet of coelomoducts. Plirougli ])osterior end of jirolhorax. riuougli junction of ]iliar)'nx and stomach. Reconstruction of pcstcrior end of Strophomenia rcgularis. A portion of the ventral salivary gland.? and duct in AlexandroiiK nia aga.ssizi. ■.:ric So;£mogastre:s ^«'^^-'% ff^\ ,i- bt f^ ■v3i 4*1:1 i **K^N.-^^ - * r^<'. 10 ':lS ©:, PLATE 29. PLATE 29. Figs. 1-5, 8, 12. Cross sections of Chaetoderma erudita. X 33. Figs. 0, 7, 9-11. Chaetodernia scabra. X 20. Tlirough heart and roeloinoduets. Close to junction of pharyn.x and sloniach. Posterior end of prolliorax. Through brain. Througli outh'l, on right, of ciiiloMioduct. Suprarectal commissure. Tlirough gills and cloacal chamber. Through radula. Through junction of pharj'nx and stomach. Througli hcarl and coclomoducts. Through posterior end of i)rolhorax. Fig. 12. Through suprarei-tal commissure. Fig. 1 Fig. 2. Fig. 3 Fig. 4, Fig. 5, Fig. 0. Fig. 7. Fig. S. Fig. 9. Fig. 10. Fig. 11. "Albatross" Pacific Solenogastres pq •^^:^. }da ^■'.•\^ Ipcm ^ ■ .<■- bg ^ /i'; ' / N PLA'I'K :!(!. PLATE 30. Figs. 1,3. Chaetodornm srabra. X 33. Figs. 2, 4-("), 10, 13. Cm^^s sections of Chaetoder rubusta. X 33. Figs. 7-0, 11, 12. Chadoderma japonica. X 33. llirough brain region. Tlirough radula. Same. Through radula. Tlirough brain region. Tlirough supra rectal comnii.ssure. Through posteiior end of prothorax. Though outlet, on left, of coelonioduct. Through heart and coelonioduct s. Through suiirarectal commissure. Through junction of pharynx and stomach. Through posterior end of prothorax. Through junction of pharynx and stomach. Through outlet, on left, of coelonioduct. Fig. 1. Fig. 2. Fig. 3. Fig. 4. Fig. 5. Fig. (i. Fig. / . Fig. 8. Fig. 9. Fig. 10. Fig. 11. Fig. 12. Fig. 13. h'. 5A.? _i 33'pACinC S0I£N0GASTRE3 :30. w m0 , \ sto %.y/ ^ 'i .-•«% -.Xs • :v =%,P™ 114 STROPHOMENTA OPIIIDIANA. it (loinands no doscriiitioii. I'ostcriorly the gut narrows rapidly, becomes some- what rectangular in cross section as it jiasses between the limbs of the shell gland, and shortly before its outlet in the cloaca develops moderately high folds. The pericardial cavity is of very large size (Plate 18, figs. 2, 3), and the con- tained h(>art is of the usual greatly elongated type. There are no distinct signs of a division into auricle and ventricle though a valve-like flap near its anterior enti may indicate such or possibly the commencement of the aorta, which for a considerable distance is of as great diameter as the heart itself and even in the head region contiiuiesof large calil)re (Plate 18, fig. 1 )• Its relations to the gonad and visceral cavity are similar to what occurs in .'^. iridugulari!^. In the pos- terior part of the body the channels are more than usually ill defined, but the course of the blood is essentially the same as in the other species of the genus. The corpuscles possess the characteristic elliptical or pointed ovate form, and are accompanied by a relatively large number of leucocytes. The gonad is fully developed, of relatively large size and the sex products are arranged in the customary fashion. Throughout its entire extent, but especially in the posterior half f)f the animal, the normal reproductive elements are associated with large masses of eggs in all stages of degeneration. This may be due to post mortem changes, but the sharply defined character of the various stages of the spermatozoa, ova, blood corpuscles, and other cellular elements in various parts of the body argues against such a view. In some species of Chitons {e. g. Ischnochiton magdalcnensis) a considerable number of ova do not pass to the exterior during the egg-laying process, but undergo disintegration and are absorbed. Appearances indicate that this is the state of affairs in the present species, and the almost empty condition of the seminal receptacles further indicates that the breeding season has passed. The ducts leading from the jjericardial cavity are relatively slender though they enlarge somewhat before entering the shell gland, and as the cells change from a cubical to a columnar form they become increasingly glandular. An unusually large number of seminal vesicles are present, twenty-three occurring on the side of the body represented (Plate 9, fig. 1). In these the distal, usually vesicular portion is exceptionally small (Plate 18, fig. 4) but the diameter may be somewhat increased when the organs are filled with sperms. These bodies are attached not only to the forward end of the shell gland but several of them open into the dorsal section of the gonoduct. The component cells are columnar and show at various points faint signs of glandular activity. The shell gland on the other hand is highly glandular, more than usually irregular in outline and as STROPHOMENIA OPIIIDIANA. 115 usual in the genus unites with its fellow of the opposite side so close to the cloaca that two openings appear to be present. The cells are all of columnar form and are of one type judging from the darkly staining granular secretion. As in other species of the genus a diverticulum of the cloacal wall is present ventral to the outlet of the shell gland, but there are no indications that it is of any especial significance. In the other species of the genus Strophomenia described in the present paper there are from few to many diverticula extending outwardly from the cloacal wall, but they never reach the excessive development existing in this species. These are shown, somewhat diagrammatically (Plate 9, fig. 1). The cells are usually columnar and are filled with a finely granular sul)stance which in various places is in the act of escaping into the cloacal cavity. The brain, clearly bilobed, is located against the under side of the intestinal coecum at the level of the posterior border of the atrium (Plate 8, fig. 5). From it the usual three pairs of nerves originate, that after branching unite with ganglionic masses attached to the bases of the cirri or without such union pass to the body wall. The connectives to the lateral, pedal, and labio-buccal sys- tems arise in the customary situations and the relations of the ganglia themselves, so far as they have been determined, are typical. Pedal commissures and latero-pedal connectives occur at frequent, though not perfectly regular, inter- vals and a corresponding number of unusually heavy nerves arise from the dorsal surface of the lateral ganglia. These have in several cases been followed close to the mid dorsal line but that they form commissures is not assured. They probably innervate the neighboring somatic musculature and hypodermis. Posteriorly the pedal cords, united by commissures to the anterior cloacal wall, branch repeatedly in this last named locality and innervate the body and cloacal walls and some of the fibres become imbedded in the shell gland. The lateral cords at this same level likewise branch repeatedly and supply the same structures as the pedal, though more dorsally, and in addition give off a few small nerves that attach themselves to the pericardial wall. A very few branches from these last named nerves have been traced a short distance into the heart. The pedal and lateral cords are posteriorly united by one delicate branch; others may exist, but the nerves are not sufficiently differentiated from the surrounding tissue to permit of their being followed for any considerable distance. It is a peculiar fact that no trace of a dorsal posterior commissure uniting the lateral cords has been found to exist. The labio-buccal connectives arise to the inside of the connectives leading PLATE 31. PLATE 31. Figs. 1-4, 7 (lorma jniionica. S, 10. Cross .sections of Chartoilerma californica. X 33. Fig. i). Pmnroniriiia hawaiien.sis. (X 33.) Figs. 5, 0. Chaeto Fig. Fig. Fig. - Fig. Fig. Fig. Fig. Fig. Fig. glion ri'll; Fig. Through radula. Through junction of pharynx and stomacli Through heart and coelomoduets. Througli brain region. Through brain of Chart odcrnia japonica. Through I'adula, same species. Tlu'ough outlet of coelomochict, on left. Pharynx and glands in front of radula. External sen.-iory atrial ridge of Pron<-omenia liawaiiensis; os, ridge resting upon gan- >; oin, outer atrial ridge; e, cirrus. 10. Through suprarectal commissure "Albatross" AciFic SoLENOGA'irsrs ■31. 4 ^-«« %\ ^hi\ ^%.a u t) I \ fr V' !;%&. :Q^ m ,M /0- &.. . .■^^' , -•-•as.- <:3ii j,P^ \%.^>^ #••#» '^^ ■■^; PLATE 32 PLATE 32. Cuticle and hypodorinis from aide of head unless otherwise stated. Hypodermis and its products in Stroi)liomenia spinosa, large specimen. X 205. Same in Lophomenia spiralis. X 25.5. Same in Strophomenia seandens. X 155. Same in Halomenia gravida. X 225. Perforation of the somatic musculature in H. gravida by a diverticulum of the mid gut which comes in contact with a modified hypodermal papilla; s, blood sinus. X 255. Hypodermis of Drepanomenia vampyrella. X 255. Same of Ichthyomenia i)orosa. X 330. Same of Dondcrsia californica. X 555. Section through the dorsal sense organ of Strophomenia seandens. X 300. Same of Proneomenia hawaiiensis; ne, nerve. X ISO. Same of Lophomenia spiralis; ne, nerve. X 150. Fig. 1. Fig. 2 Fig. 3. Fig. 4. Fig. 5. 'h CO Fig. I116S G. Fig. 7. Fig. S. Fig. ■ '.). Fig. 10 Fig. u. -loss" Pacific Solenogast? rk;; ■iateSE. 1 "•I'VES^ ^^: ..*">- ^. -■■>,v. ^1 1Sr-*. Arfk^^y {•■^V •ii.r^gu'^i) ^N\\-;A>i *»'-• V -«e--. PLATE 33 Fig. 1. Fig. '"> Fig. 3. Fig. 4. Fig. 5. Fig. 6. Fig. 7. Fig. 8. Fig. 9. PLATE 33. Cuticle and liypodermis from sido of head unles.s otherwise stated. Strophomenia trianguUiris. X 200. Strophoraenia farcimen. X 300. Proneomeniahawaiiensis. X 300. Dorynienia acuta. X 330. Alexandromenia agassizi. X 200. Proneomenia insuhiris. X 4(X). Strophomenia spinosa, dor.sal siile. X 200. Alexandromenia valida. X 2t)0. Strophomenia ophidiana. X 200. "Albatross" Pacific Solenogastrks PU'ji:33. 7 i^m ^''1','Sj ■•^ H.Heaui d£l. PLATE 34. Fig. 1. Fig. 2, SO. Fig. 3. Fig. 4. Fig 5. Fig. G. Fig. 7. Fig S. Fig. 9. Fig. 10. Fig. 11. Fig. 12. Fig. 13. 2,55. Fig. 14. Fig. 15. PLATE 34. Cros.s spctiou through radula sac of Lophomenia .spirahs. X •'57. LongitiuUnal section through radula and subradular organ of Pronconicnia liawaiiensis. Dorsal \ic\v of radula of Liniifossor talpoideus. X 135. Two rows of teeth of Stropliomenia triangularis. X 330. Portion of radula of Alexandromenia agassizi. X 255. Side view of radula of Limifossor taljioideus. X 135. Teeth of Dorymcnia acuta (mid line to left). X 330. Teeth of Strophomenia spinosa, large individual. X 555. Same species, small individual. X 330. Same species, small individual. X 330. Side view of teeth of Dorymcnia acuta. X 330. Cross section of radula of Halomenia gravida. X 330. Proncomenia hawaiiensis, showing 7 of the 40 teeth of each transverse row; m, mid line. Tooth of Alexandromenia valida. X 200. Portion of radula of Strophomenia farcimen. X 255. "Albatross" Pacific Solenogastres Pij\te34. V %■' Z 1/ ■^/Jv ^^ i !i . ^•'l 'L /(/ 'iE>.*^ /■^ // ^i PLATE 35. PLATE 35. Fig. 1. Limifossor fratulii. Section through rrgion of salivary glands. X 28. Fig. 2. Same through must-lc 17 (fig. 4, jil. 10). Fig. 3. Same through posterior end of ratlula supjiort . X 28. Fig. 4. Same in region of heart. Fig. 5. Same through opening of dige.stive gland. Figs. 6, 7, 9. Development of ova in Herjjomenia platypoda. In fig. (i the nuclei of probable follicle cells (fc) are intact; in fig. 7 the membrane has dissolved and the scattered chromosomes are becoming vesicular; in fig. 9 the oviun is almo.st mature and the vesicles (cr) of relatively large size. X 555. Fig. S. Longitudinal section through ri'gion of radula of Chaetoderma erudita, showing .sub- radular ganglion, sn. X 135. Fig. 10. Section through brain of Limifo.ssor fratula. X 28. Fig. 11. Protozoa encysted in wall of digestive tract of Chaetoderma californica. Fig. 12. Section through dorso-terminal sense organ of Dondersia californica. X 333. Fig. 13. Blood corpu.scles of Proneomenia hawaiiensis. X 450. Fig. 14. Section through Proneomenia insularis. -S" Pacific Solenogastres '':,ATi;3S. ■>f0f y — '- • i ..''^^. \ ^'^ qjSir ■W ^$v^&~ M/ m 13 ' , % ^ q\ vi^M- v'^.fefc O A; // -?«>55^v — w. \ #. PLATE 36. PLATE 3(3. Reconstruction of posterior end of Chaetodernia argentea. 2. Same of Chaetodernia attenuata. Advanced stage in spicule development, the matrix cells retaining their attachment to Proneonienia hawaiiensis. X 555. Somewhat earlier stage than fig. 3. X 555. Completion of sjiicule development and commencement of shifting of matrix cells. X 555. Reconstruction of anterior end of Strophomenia triangularis. Early stage in development of spine in Proneonienia hawaiiensis. X 555. Hypodermis in Limifossor fratula. X 255. Posterior end of Alexandromenia agassizi. X 3. Pajiilla and outlet of .salivary gland in A. agassizi. Very early stage in development of spine in Proneomenia hawaiiensis. X 555. Blood corpuscles of Strophomenia scandens. X 450. Same of Chaetodernia hawaiiensis. X 4.50. Same of Lophomenia spiralis. X 450. Spines of Strophomenia triangularis. X 150. Same of Strophomenia spinosa. X 205. Same of Strophomenia ophidiana. X 205. Completed development of radially directed spine of Proneomenia hawaiiensis; and developing papilla. X 555. Fig. 19. Base of gill plates and attached gland cells (gc), Alexandromenia agassizi; ne, nerve. Fig. 1. Fig. 2 Fig. 3. the spine Fig. 4. Fig. 5. Fig (j. Fig. 7. Fig. S. Fig. 9. Fig. 10. Fig. 11. Fig. 12. Fig. 13. Fig. 14. Fig. 15. Fig. 16. Fig. 17. Fig. 18. "Albatross" Pacific Solenogastres ■^i,ArE3G. m ■aS'/J^ fr i-^/ _-^C^r %W#^5i :' -■■ 5 -1-,*^ ,W/'«v, \^-: /» ^ H.HEstti del. n • -«*k ; ! / l:< PLATE 37. Fig. 1. Fig. '2 Fig. 3. Fig. 4. Fig. 5. Fig. 6. Fig. 7. Fig. S. Fig. 9. Fig. 10. Fig. 11. Fig. 12. Fig. 13. Fig. 14. Fig. 15. Fig. 16. Fig. 17. Fig. 18. Fig. 19. Fig. 20. PLATE 37. Spicules of Ichthyomcnia porosa. X 4S0. (Jf Chaetoderma montereyensis, small specimen. X 80. Of Chaetoderma montereyensis, large specimen. X SO. Of Chaetoderma robust a. X 80. Of Proneomenia hawaiiensis. a X 330; b X 130. Of Chaetoderma argentea. X 135. Of Drepanomenia vampyrella. X 135. Of Chaetoderma attenuata. X 135. Of .\lexandromenia agassizi. X 335. Of Doryraenia acuta. X 135. Ale.xandromcnia valida. X ISO. Of Chaetoderma hawaiiensis. X 255. Of Lophomenia spiralis. X 335. Of Chaetoderma californica. X 135. Of Chaetoderma erudita. X 135. Of Proneomenia insularis. X 135. Strophomenia scandens. X 300. Of Chaetoderma nanula. X 135. Of Chaetoderma scabra. X 100. Of Chaetoderma sp.? Unidentified fragment, Alaska. X 135. \ ^ \'>^Si;„'; v' ' ■■^v- I r^i \ (I /' / / n ■.'!■ " ^\ / // " /f r II x^ ^ n \ w If- I- 1 /.v V /'/ :'// y^ / PLATE 38. PLATE 3S. Fig. 1. Reconstruction of anterior end of Driomenia pacifica. Fig. 2. Same, posterior end. Fig. 3. Entire animal enveloped in small portion of hydroid colony, Hertularella sp. X 5. Figs. 4, 5, 6, 8. Sections along lines A, C, U, D, in fig. 1. X 50. Figs. 7, 9. Along lines E, F, of fig. 2. X 50. Fig. 10. Section of hypodermis and cuticle. X 260. Fig. 11. Same of Pachynienia abyssorum. X 60. ALBATROSS Pla:k38 r?^ '^'^ ' ^J t- u r. "iy :-? ^^^ % % ri l/ M 1' \ V: -^. \ V--' ( /-f i r ulr-- ^;;^,^r^,-<^K»= /o - 1^j»*"-:^ VI PLATE 39. PLATE 39. Fig. 1. Reconst motion of anterior end of Pachymenia abyssonim. Fig. 2. Same, jiastcrior end. Fig. 3. Foot, same species, in middle of body. Fig. 4. Pacliymenia abyssoruni. X 3. Fig. .5. Spine.s of Driomenia pacifica. Fig. 6, S. Sections of Pacliymenia abyssoruni along lines B, D, fig. 1. X 25. Fig. 7. Section of Driomenia pacifica along line G, fig. 2, pi. 38. X 50. "Albatross" Pacific Solenogastres ■\°i ..^^^m-:. ~^&^ -ij^^r '-^-/iJKWrii. ' -0^^^?^ ' .■WiH't'i'i*'**^*'f*-^^' ;:r:^'^..-.^^ « ! = -V 6 Or*' -■'':^''v -i/V ••'-■'/ PLATE 40. PLATE 40. Figs. 1-4, 6-10. Cros.s .sections of Pachyiiienia abyssoniin. Fig. 5, Driomenia pacifica. Figs. 1, 2, 4, 7, along lines E, F, G, H, posterior end. X 25. Figs. 3, 6. Sections along lines A, C, anterior end. Fig. 5. Section of Driomenia along line H, fig. 2, pi. 38. Fig 8. Spines of Pachymonia abyssorum. X 140. Fig. 9. Section about the posterior end of gonad. Fig. 10. Section through body behind pharynx. "Albatross" Pacihc Soi.enogastr-\<-; ^;i:40 ^.*> -•"5^i?<% ■yif \ n r.iv.:t. or. -fjeacfOAl^UlaMk ««^.«M PUBLICATIONS MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE. There have been pubhshed of the Bulletin Vols. I. to LIL; of tlie Memoirs, Vols. I. to XXIV., and also Vols. XXVI., XXVIII., XXIX., XXXI. to XXXIII., XXXVII., and XLI. Vols. LIII. to LV. of the Bulletin and Vols. XXV., XXVIL, XXX., XXXIV. to XXXVL, XXXVIII. to XL., XLII. to XLVII. of the Memoirs, are now in course of publication. A price list of the imhlications of the Museum will be sent on application to the Curator of the Museum of Comparative Zoology, Cambridge, . Mass. Date Due Harvard MCZ Library II I ; II I II' I / 3 2044 066 302 06i ii;ii:!;HJ:|ilil|:i!!l'|llii(l^^^^^