BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY VOE. It 1953—1955 The WIULS Es er. <> ~~! seer Bi 7 \ » 7 ees Nu *] 4 te Pe! } SF e/ \? rs) ~ OU ~~ Ra His PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) LONDON: 1955 DATES OF PUBLICATION OF THE PARTS No. I 17 February 1953 No. 5 March 1953 No. 7 November 1953 No. 22 January 1954 No. 22 January 1954 24 September 1954 22 October 1954 22 October 1954 v4 ° ear arry yp 31 December 1954 No. 10. 30 November 1954 No. It. 25 January 1955 No. 12. 24 June 1955 No. 13. 19 August 1955 PRINTED IN GREAT BRITAIN AT THE BARTHOLOMEW PRESS DORKING BY ADLARD AND SON, LTD. N e ds ey CONTENTS ZOOLOGY VOLUME 2 The Ostracod genus Tvachyleberis. By J. P. Harpinc and P. C. SYLVESTER-BRADLEY (Pls. 1-2) . On the Puerulus stage of some spiny lobsters (Palinuridae). By ISABELLA GORDON . On Sirpus, a genus of pigmy cancroid crabs. By ISABELLA GORDON The Macrotritopus problem. By w. J. REES (PI. 3) . Hearing in Cetaceans. By F. C. FRASER and P. E. PURVES (Pls. 4-5) . The “ Rosaura’’ Expedition 1937-1938. (Pls. 6-9) Gear, Narrative and Station List. By J. Ss. COLMAN Under-Water illumination and ecology in tropical estuaries. By J. S. COLMAN and L. H. N. COOPER The Echinodermata. By D. D. JOHN and A. M. CLARK Fishes. Part 1: Families.Carcharhiniidae, Torpedinidae, Rosauridae, (nov.), Salmonidae, Alepocephalidae, Searsidea, Clupeidae. By DENYS W. TUCKER Sponges. By M. BURTON 7. On the Polyzoan genus Crepidacantha Levinsen. By D. A. BROWN 8. Subdivisions of the genus Ptilinopus (Aves, Columbae). By A. J. CAIN . IO. 6 ages . 12. , 13: . A collection of Mesostigmatid mites from Alaska. By G. OWEN EVANS The Otariid seals of the Pacific Coast of America. By JUDITH E. KING (Pls. 10-11) The Southern Right Whale Dolphin, Lissodelphis peroni (Lacépéde). By F. C. FRASER (PI. 12) New species of Tilapia (Pisces, Cichlidae) from Lake Jipe and the Pangani River, East Africa. By ROSEMARY H. LOWE (Pls. 13-17) The morphology of the head of the hawfinch (Coccothraustes cocco- thraustes) with special reference to the myology of the jaw. By REGINALD WILLIAM SIMS Index to Volume 2. 309 339 347 369 395 ERRATA P. 65, first entry should read: Boun, G. 1902 .. Bull. Sci. Fr. Belg., 36: 178-551, etc. P. 265 (title-page) : For ‘‘ 2 Plates ’’ read ‘‘ 2 Figures.’’ _TRACHYLEBERIS a Bey fe : THE OSTRACOD GENUS TRACHYLEBERIS BY J. P. HARDING British Museum (Natural History) AND P. C. SYLVESTER-BRADLEY Department of Geology, University of Sheffield Hee Stet ] 2 VPN 1953 Pp. 1-16; Pls. 1-2; 25 Text-figures. BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 2 No. 1. LONDON: 1953 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical Series. Parts appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. Thus paper is Vol. 2, No. 1 of the Zoological series. PRINTED BY ORDER OF THE TRUSTEES ‘OF THE BRITISH MUSEUM Issued February, 1953. Price Four Shillings. fae OsTRACOD GENUS TRACHYLEBERIS By J. P. HARDING anp P. C. SYLVESTER-BRADLEY SYNOPSIS When Brady erected the genus Tvachyleberis for a species which he had described earlier as Cythere scabyocuneata, he had before him specimens from New Zealand which did not in fact belong to that species. Both the type specimens of Cytheve scabrocuneata and the specimens from New Zealand are in the British Museum (Natural History). A new name is proposed for the New Zealand species and full descriptions of the type specimens of both species are given; fortunately they are congeneric. Brapy (1880) described the shell of a new species or ostracod, Cythere scabrocuneata, from specimens collected by the “Challenger” Expedition. Later he (Brady, 1898) described the limbs and other soft parts of further specimens from Lyttelton Harbour, New Zealand, which he thought belonged to the same species, and as the limbs showed some unusual features, he erected a new genus Trachyleberis. Miiller (1912) and Skogsberg (1928) considered Tvachyleberis to be synonymous with Cythereis Jones, but, as one of us has shown elsewhere (Sylvester-Bradley, 1948b), the shell of Tvachyleberis differs significantly from that of Cythereis, and the genus was therefore revived, though without reference to the soft parts. Some of the original “ Challenger’ material is preserved in the British Museum (Nat. Hist.) in the form of dry shells gummed to wooden microscope slides, and the Lyttelton Harbour material, consisting of several specimens of both sexes, is preserved in spirit in the Hancock Museum, Newcastle, and we are much indebted to the Curator, Mr. C. E. Fisher, for letting us examine this material. He has also graciously permitted some of the material to be kept at the British Museum. We also re-examined the “ Challenger ” specimens and found that some of what Brady refers to as “empty shells” had a few appendages still inside them ; and by treating the whole ostracod with a 1% solution of tribasic sodium phosphate (Na,PO,) it has been possible to recover these appendages and make permanent mounts. It is now quite evident that these “ Challenger ’’ specimens and the ones from New Zealand are different species, and that the Lyttelton Harbour specimens must be given a new name ; we propose to call them Tvachyleberis lyiteltonensis. It is unfortunate that the only specimens in the British Museum (Nat. Hist.) of the “ Challenger ” collections are from the Inland Sea of Japan, as Brady describes C. scabrocuneata from three different localities, the Bass Straits, the Inland Sea of Japan and Wellington Harbour, New Zealand, and it is quite possible that the Wellington Harbour specimens were the same species as those from Lyttelton Harbour, as both are from New Zealand harbours. Should these two species ever be put into different genera the name Tvachyleberis will have to remain with the Japanese species, T. scabrocuneata, which must be based on the only available type ZOOL. Il, 1. 4 THE OSTRACOD GENUS TRACHYLEBERIS material. This is in spite of the fact that it was characters which Brady saw in the New Zealand specimens, now called 7. /ytteltonensis, which led him to erect his new genus. The present paper attempts to describe these two species. Although it is intended as a contribution to the taxonomy of the genus, we do not append a formal diagnosis. We feel that as yet too few species are known by the details of both shell and soft parts. As the specimens from Lyttelton Harbour are more complete they will be described first. TRACHYLEBERIS Brady, 18098 Type-species (by monotypy) : Cythere scabrocuneata Brady, 1880. Trachyleberis lytteltonensis sp. nov. (Text-figs. 2-19 ; Pl. i, figs. 1-4, 7; Pl. 2, figs. 1-4, 7, 8) Trachyleberis scabrocuneata, Brady, 1898, Trans. Zool. Soc. Lond. 14: 444, pl. 47, figs. 1-7, 18-25. Trachylebevis scabyocuneata, Hornibrook, 1952, New Zealand Geol. Surv., Palaeont. Bull. 18 : 32-33, Pl. 3, figs. 38, 39, 48. MaTeRIAL. About 40 specimens preserved in spirit from the Brady collection, Hancock Museum, Newcastle-on-Tyne, dredged at 1-5 fathoms from Lyttelton Harbour, New Zealand. The holotype and other figured specimens have been mounted, and through the generosity of the Hancock Museum are now kept at the British Museum (Nat. Hist.); these bear the following registered numbers: 1952.12.9.1-12. OccuRRENCE. In addition to its occurrence in the type locality 7. lytteltonensis has been reported by Hornibrook (1952, pp. 69, 71, under the name “7. scabro- cuneata’’) from four stations in the seas of the New Zealand area dredged at depths ranging from 28-67 fathoms, and from one fossil locality in the Upper Miocene (Tongaporutuan) of New Zealand. Horotype. Male, B.M. 1952.12.9.1. PARATYPES. Both sexes, B.M. 1952.12.9.2-12. DESCRIPTION. The carapace is subrectangular in lateral outline, with the dorsal and ventral borders nearly straight, and converging but slightly towards the pos- terior (Pl. 1, figs. 1-4). The anterior border is rounded, the posterior triangular, the postero-dorsal margin being straight or slightly concave, the postero-ventral margin curving evenly into the ventral margin. Sexual dimorphy is strong, the males being longer in proportion than the females (see table below). In dorsal view (Pl. 2, figs. 1-4) the carapace is seen to be widest in the region of the sub-central tubercle, and is compressed in the regions of the anterior and posterior plains. In end view (Text-fig. 2) the carapace appears approximately triangular, both valves showing a fairly abrupt ventral angulation, and the venter being almost THE OSTRACOD GENUS TRACHYLEBERIS 5 flat. The left valve is larger than the right, its selvage fitting outside that of the right throughout its course. Width Length Height (of complete carapace) Dimensions: 3 B.M. 1952.12.9.1 5 1-08 mm. 0-49 mm. 0-49 mm. © B.M. 1952.12.9.6 z 0-90 ,, 0°49 ,, 0°49 ,, Proportions: § B.M. 1952.12.9.1 2°21 if I-00 © B.M. 1952.12.9.6 1°84 I I-00 anterior Z hinge dorsal ridge posterior eye y tubercle tubercle hinge tubercle dorsal groove antero - dorsal complex postero-dorsal complex anterior plain posterior anterior marginal marginal panes spines , Osterior anterior ee inal marginal t | ah Z Os fero -ventra HM antero-ventral sub-central ventral is complex complex tubercle ilesayets Oo O.5mm. Cr 0.5mm. t 5 - . . j | TEXT-FIG. I. Diagram to illustrate the terms used in the description of the external ornament of the carapace in the Trachyleberidae. Text-Fic. 2. Tvachyleberis lytteltonensis sp.nov. Profile in end view from the posterior. B.M. 1952.12.9.6. Text-Fic. 3. Tvachyleberis lytteltonensis sp.nov. Internal lateral view of right valve. B.M. 1952.12.9.6. 6 THE OSTRACOD GENUS TRACHYLEBERIS The following analysis of the ornament (see Pl. 1, figs. 1-4) follows a scheme previously put forward (Sylvester-Bradley, 1948a) illustrated by Text-fig. 1. Eye tubercles are prominent on both valves, and run into the antero-dorsal complex, which consists of a simple swelling. The sub-central tubercle forms a prominent mammillated boss, and the antero-ventral complex consists of four tubercles. Behind the sub-central tubercle the ornament is not clearly divided into ridges and complexes, but consists of a number of tubercles, which have a tendency to bear slightly mammillated summits (this character is not well shown by the photographs reproduced in Pl. 1, which are of-specimens coated with magnesia). There are three ventral ridges, the uppermost bearing elongated tubercles. The anterior marginal spines are divided into two groups, a dorsal set of smaller rounded spines, and a ventral set of larger more pointed spines. The anterior marginal rim is composed of a double line of tubercular spines, which supplement the armature of the margin. The posterior marginal spines are prominent in the ventral region, but almost absent dorsally. The posterior marginal rim is less well developed than the anterior, and is armed with fewer tubercles. The anterior hinge tubercle lies behind and above the eye tubercle in both valves, and the posterior hinge tubercle of the left valve forms a prominent swelling near the commissure, best seen in dorsal view. The duplicature is fairly wide, extending 0:05 mm. inwards from the selvage in anterior and postero-ventral regions, and up to 0:07 mm. at the posterior extremity. The selvage is prominent in both valves, and follows a sinuous course along the ventral margin, being markedly concave to the exterior in the centre of the venter. There is no vestibule. Radial pore canals (Text-fig. 3 ; Pl. 1, fig. 7) run to each marginal spine and to each tubercle of the marginal rim. Some (presumably those serving the marginal spines) terminate along the inner margin ; others (presumably those serving the rim) terminate on the inner surface of the shell just inside the inner margin. The total number of radial pore canals is therefore rather large, and they appear crowded, and sometimes superimposed in lateral view. Most of them are inflated towards the middle. The normal pore canals are few in number, one serving each tubercle of the ornament in the same manner as the radial canals. The sub-central tubercle appears as a muscle scar pit on the inside of the valve (Text-fig. 3). The muscle scar pattern is that characteristic of so many of the superfamily, with four slightly elongated scars, lying vertically above each other, to the posterior of the pit, and a single scar, U-shaped, opening upwards, within the pit. In addition there are two other scars, one above and one below the main group. Each scar is raised as a tubercle on the inside of the valve. The antero-ventral complex forms a pit on the inner surface of the valve similar in nature to the muscle-scar pit (Text-fig. 3). The hinge of the left valve (Pl. 2, figs. 1-4, 7, 8) consists of two terminal slightly elongated sockets, unbounded on the ventral side, and a median element subdivided into an anterior rounded projecting tooth and an adjoining posterior finely crenulated bar. In the right valve the anterior tooth projects beyond a less strongly projecting THE OSTRACOD GENUS TRACHYLEBERIS y base. The posterior element is a slightly elongated lip-like tooth. The median element is a groove subdivided as in the left valve to provide a deep anterior socket, open ventrally, behind the anterior tooth. In both valves a hollow lying immediately below and in front of the anterior hinge element leads to the eye tubercle, and is for the reception of the eye of the animal. The tubercle, being highly polished and transparent, perhaps serves as a primitive lens. The first antenna (Text-fig. 4) consists of six segments and is the same in both sexes. The proportionate lengths of the segments and of the various setae, spines and sense organs are best indicated by the drawing. The first segment has a tuft of long spinules on the posterior face near the base. The second segment has a few spinules on both posterior and anterior faces, and on the postero-distal corner there is a slender, flexible, finely-haired seta. The third segment has a tapering seta on the antero-distal corner. The fourth segment has two smooth setae on the antero-distal corner, one stout and the other more slender, and from the middle of the distal edge on the medial side is a smooth, slender seta. The fifth segment has two stout setae and a slender one near the antero-distal corner, and one slender one on the distal edge on the medial side. These four are all smooth. The last segment has two stout terminal setae; and a slender one which is united at its base to a sense club (s.). The second antenna (Text-figs. 5 and 6) consists of a protopod (f.) of one segment, an endopod (en.) of three segments, and a very small exopod (ex.) of two segments, and is alike in both sexes except for the small differences in the exopod. The protopod is quite unarmed. The exopod is reduced to what appears to be a sense organ. This is best developed in the female (Fig. 6), where it is very thinly chitinized and rather broad, and is less than twice the length of the first segment of the endopod. In the male (Text-fig. 5) the exopod is shorter still and looks very like an ordinary seta. Brady (1898) describes the exopod as “a very short, falcate, urticating seta [in the female], which is absent in the male.” The first segment of the endopod bears some stiff bristles and a long slender seta with very minute hairs on the postero-distal corner. The second segment of the endopod bears a tuft of hair-like spinules on its first quarter on the inner side. About two-thirds along its length there are two smooth setae on its anterior face, and on the posterior face at about this level there is one stout, tapering seta, and one longer but more slender one, both feathered, and near them a sense club (s.). At the postero-distal corner of this segment there is one stout, feathered seta and one small naked one. The last segment has three stout, naked setae, two together half-way along its posterior margin and the other one terminal. The mandible (Text-fig. 8) is the same in both sexes. The shape of the basal biting part is shown in the figure. The biting edge contains six strong teeth each with a secondary cusp. The most anterior of these teeth is the largest, and they decrease in size gradually towards the posterior corner ; the first two take up half the biting edge. There is a bristle between the first and second and another between the second and third teeth, and on the posterior medial corner beyond the largest tooth there are two ribbon-like bifurcate setae which are about twice the length of THE OSTRACOD GENUS TRACHYLEBERIS 7 TEXT-FiGs. 4-12. Tvachyleberis lytteltonensis sp. nov. Fig. 4, first antenna, x 130. Fig. 5, second antenna of male, x 130. Fig. 6, part of second antenna of female with exopod, x 180. Fig. 7, brush-shaped organ of male, right member from behind, x 360. Fig. 8, right mandible from inner side, x 130. Fig. 9, left maxilla from outer side, x 130. Fig. 10, palp and endites of left maxilla from outer side, x 220. Fig. 11, posterior part of body of female, x 180. FFg. 12, posterior part of body and copulatory apparatus of male, x 130. b., postero-ventral corner of copulatory appendage ; ch., chamber; cop., copulatory appendage; ev., endopod ; ex., exopod; ji., flagellum; fu., furcal ramus; /y., labyrinth; od., ovi- duct; ., protopod; #e., penis; y.s., receptaculum seminis; s., sense club; ¢., terminal seta; va., vagina; v.d., vas deferens; vyr., verruca. THE OSTRACOD GENUS TRACHYLEBERIS 9 the teeth near them. Beyond them again there is a single bent seta. Between the biting edge and the palp there is a seta with a few hairs. The palp consists of the second segment of the protopod bearing a three-segmented endopod and an exopod of one segment. The exopod rises from the dorsal part of the outer face of the protopod and is directed distally. It consists of a single rounded segment narrow at its base, with five carrot-shaped, hairy setae which are thinly chitinized and of very different lengths, as shown in the figure. The ventral margin of the protopod carries two delicate setae. The one nearest the distal corner is more or less parallel to the endopod and is shorter and stouter than the other. Both are annulated and hairy. The first segment of the endopod has an annulated hairy seta on the inside of the distal dorsal corner, and on the ventral distal corner there are four setae, of which the two medial ones are small. The two on the outer side are stout and curved and are armed with long, stiff hairs which are rather wide apart. In preparations these are most evident in the outer seta, in which the hairs lie parallel to the cover slip. On the dorsal part of the distal end of the next segment is a group of seven whip-like, flexible, naked setae, and on the ventral distal corner there is a long tapering seta with hairs. On the medial side in the middle of the distal edge there is a seta which is a little longer than the terminal segment and has a few stiff hairs on the ventral side. The terminal segment of the endopod is a little longer than the preceding one, but rather less than half its diameter and slightly tapering. It has four slender terminal setae, one of which has hairs. The maxilla (Text-figs. 9, 10) is the same in both sexes. Between the long, plumose setae of the vibratory plate and the palp on the distal edge of the appendage there is a hairy swelling with a group of bristles directed parallel to the palp. The first of the 16 plumose setae arises from this swelling. This seta is constricted into two parts, both of which are rather swollen. The distal part tapers to a point. The second to the fifteenth plumose setae are longer and very much alike. They have a swollen basal part rather longer than the corresponding basal part of the first seta, and the distal part is not swollen but long and feathered. The sixteenth seta is much smaller than the others but hardly longer than the basal part of the middle 14, and it appears to be of one piece without a constriction. The palp is composed of two more or less cylindrical segments, the first segment being about twice the length and twice the width of the second. On the distal edge of the first segment and dorsal to the second segment are four slender setae differing from each other in hairiness. That next to the second segment is the shortest, being about twice the length of the second segment ; it has minute hairs all round it. The other three setae are all nearly twice this length ; one appears to be naked, the second has rather short hairs, and the other has long hairs ; these three are annulated for the distal part of their length. In addition to these setae, the distal margin of the first segment carries a seta near the outer face of the second segment. This seta has a swollen base from which arises a short secondary setule. There are three distinct endites, the first, i.e., the one nearest the palp, bears terminally eight tapering setae, many of which have swollen bases. These are all 10 THE OSTRACOD GENUS TRACHYLEBERIS naked or nearly so. The two other endites have several pointed terminal setae, the actual number being difficult to determine. In addition to these terminal setae, the last endite has a fairly long seta arising from near its base on the side facing the palp. The fifth, sixth and seventh pairs of appendages are basically similar to one another. They are all leg-like, and will be referred to here as the first, second and third pairs of legs. Each of the three pairs shows sexual dimorphism, and further- more, in the first pair of legs of the male the right leg differs from the left leg. The following description will apply to both sides and both sexes unless otherwise indicated. The first leg (Text-figs. 13-15) consists of four segments. The first segment has two annulated setae on the anterior margin and two on the antero-distal corner overhanging the knee-like joint. Near the base of the leg on the posterior side there is a soft, hairy, carrot-shaped seta, and near it along the posterior margin of the segment there are some long hairs. The postero-distal corner of the first segment has a hollow surrounded by fine spinules. The second segment broadens distally and carries one seta. This is on the antero-distal corner in the female (Text-fig. 13), and carries short hairs. In the male this seta is more hairy and placed on the anterior margin short of the corner, and is farther from the corner on the right leg (Text-fig. 14) than on the left (Text-fig. 15). Apart from a few minute spinules the right leg of the male bears nothing else on the second segment, but on the left leg of the male there is a swelling in about the middle of the anterior margin. This - carries a roughened friction pad on the medial side and a hairy boss on the outer side. In the female (Text-fig. 13) there is a group of spinules on this part of the anterior margin. The third and fourth segments of the leg are the same in both sexes and on both sides. The distal edge of each segment is armed with small spinules on both sides ; these are much longer laterally than medially. The terminal curved claw is about twice the length of the last segment and is smooth in the male and pectinate on the inside of the curve in the female. The second pair of legs (Text-figs. 16, 17) is sexually dimorphic, but in the male as well as in the female the left and right legs form a symmetrical pair. The first segment is the same in both sexes, and has similar setae to those of the first leg except that only one seta overhangs the “knee.’’ The second segment broadens gradually distally, and is nearly as long as the first segment but much narrower. It bears an annulated seta on the antero-distal corner, and in the male (Text-fig. 17) three groups of stiff, sharply-pointed bristles ; in the female (Text-fig. 16) only the most proximal of these three groups is present. The third and fourth segments are similar to those of the first leg but rather longer and more slenderly built. The claw is smooth in the male and pectinate in the female. The third pair of legs (Text-figs. 18, 19) is essentially similar to the second pair. The first segment has a normal annulated seta on the postero-proximal corner instead of the carrot-shaped seta. In the male the second segment has four groups of conspicuous, straight, sharply-pointed bristles on the anterior margin ; only the two most proximal of these groups are represented in the female, and in this sex the bristles are more flexible. The seta on the antero-distal corner is long, THE OSTRACOD GENUS TRACHYLEBERIS II and in the male, crooked. The male has a group of slender-pointed spines on the middle of the anterior margin of the last segment. The terminal claw is pectinate in both sexes. The copulatory apparatus is of the usual complex nature. Specimens were unfortunately not well enough preserved for the details of the softer parts to be made out. The structure of the less easily macerated parts is shown in Text-fig. 12. The following features were found in each of five males examined. On each side the apparatus consists of two parts, a roughly oval muscular part called by Skogs- berg (1928) and others the “penis” (fe.) and a distal triangular “ copulatory appendage” (cop.). The muscular part contains a number of chitinous struts with muscles running between them as shown. Near the base of the copulatory appendage there is a very complex labyrinth (/y.) with a rounded chamber (ch.) which has chitinous walls. A tube with spirally thickened walls, presumably the vas deferens (v.d.), runs to this chamber. The other end of the tube was found in very different positions in different specimens, having apparently become free as a result of maceration of the soft parts ; but the position figured is believed to be correct. The copulatory appendage is a thin-walled, roughly triangular organ with irregular outgrowths from its broader distal part. These outgrowths are roughly similar on both sides ; but those of the same side of two different individuals are usually more alike than the right and left ones of the same individual. The appendage on the right has a flagellum (/l.) which is absent on the left. The ‘‘ postero-ventral” part of the copulatory appendage ends with a rounded point (b.). We were unable to determine a ductus ejaculatorius. The female reproductive system and the shape of the posterior part of the body are shown in Text-fig. 11. On each side of the body near the caudal furca, which in this sex has three setae, lies the verruca (v7.). The oviduct (od.) is a sinuous tube connected with the verruca. The vagina (va.) is separate from the oviduct and opens anterior to it. Both these ducts are connected with the receptaculum seminis (7.s.). The receptaculum and the other organs are all paired. The paired brush-shaped organs of the male (Text-fig. 7) are placed on the ventral side of the body between the “ knees ”’ of the first pair of legs. The furcal rami ( fz.) are reduced to a simple swelling on each side bearing three setae in the female and two setae in the male. In both sexes the hindermost part of the body bears a single median seta (¢.). Trachyleberis scabrocuneata (Brady) (Text-figs. 20-25 ; Pl. 1, figs. 5,6, 8; Pl. 2, figs. 5, 6, 9, 10.) Cythere scabrocuneata Brady, 1880, Challenger Rep. 1: 103, pl. 17, figs. 5a-f; pl. 23, figs. 2a-C. Not Trachylebevis scabrocuneata, Brady, 1898, Trans. zool. Soc. Lond. 14: 444, pl. 47, figs. I-7, 18-25. Trachyleberis scabrocuneata, Sylvester-Bradley, 1948, J. Paleont. 22 : 794, pl. 122, figs. 13-18. Not Trachyleberis scabrocuneata, Hornibrook, 1952, New Zealand Geol. Surv., Palaeont. Bull. 18 : 32-33, pl. 3, figs. 38, 39, 48. 12 THE OSTRACOD GENUS TRACHYLEBERIS OcCURRENCE AND MaTeErRIAL. The only material that has survived is that dredged by the “ Challenger’’ Expedition from the Inland Sea of Japan at 14 fathoms. The following specimens are held in the British Museum (Nat. Hist.): LectToTyPE (here chosen) B.M. 1952.12.10.1, 2. PaRATYPES: B.M. 1948.3.10.1-5, B.M. 1952.12.10.3-9. OTHER Recorps. Brady’s other records of this species are suspected to be T. lytteltonensis or other species. TExtT-FiGs. 13-19. Tvachyleberis lytteltonensis sp. nov. Fig. 13, left first leg of female, x 180. Fig. 14, right first leg of male, x 145. Fig. 15, left first leg of male, x 145. Fig. 16, left second leg of female, x 180. Fig. 17, left second leg of male, x 145. Fig. 18, left third leg of female, x 180. Fig. 10, left third leg of male, x 145. THE OSTRACOD GENUS TRACHYLEBERIS 13 DESCRIPTION. The shell of this species has been previously described (Sylvester- Bradley, 19480) on the basis of the type material here re-examined, and only such details as serve to distinguish the species from 7. /ytteltonensis need be dealt with again. The shell of the species has been refigured on Pls. 1 and 2 to aid comparison. There are no differences to report in the general shape of the carapace, the degree of sexual dimorphy, or the size. In the following table of dimensions and propor- tions, those of the complete carapaces have been estimated from single valves : Length Height Width Dimensions: ¢ B.M. 1948.3.10.1 I-Iomm. 0-52 mm. 0-49 mm. 2 BM. 1948.3.10.5 6 0-90 ,, 0:46 ,, 0°46 ,, Proportions: ¢ B.M. 1948.3.10.1 2°13 I 0-94 2 BM. 1948.3.10.5 I-94 I I-00 The antero-dorsal complex of T. scabrocuneata (Pl. 1, figs. 5, 6, 8) consists of a sharp ridge running from the eye tubercle forwards and downwards to as far as the median line, and presents an immediate contrast to the rounded swelling which forms the antero-dorsal complex of T. lytteltonensis. Similarly the sub-central tubercle of T. scabrocuneata is surmounted by ridge-like elevations rather than the mammillae of T. lytteltonensis. The general style of ornament is similar in the two species, but the tubercles in T. scabrocwneata are smaller, more sharply defined and greater in. number than in 7. lytteltonensis. In T. scabrocuneata the marginal spines in the postero-ventral region, forming a ventral ridge, are compressed and _ blade-like, and swell out above the base (Pl. 2, figs. 9 and 10). It is interesting to note that a third species, 7. thomsoni Hornibrook, 1952, ranging in New Zealand from the Paleocene (Waipawan) to Recent, as shown by Hornibrook’s figures, is also similar in style of ornament, but the tubercles are even fewer, and are more pronounced, rising from dorsal and ventral ridges as veritable spines. The internal characters of the carapace (PI. 2, figs. 5, 6, g and Io) are essentially similar to those displayed by T. lyttellonensis. The anterior tooth of the right valve, however, projects far more strongly above its base than that of T. lytteltonensis (Pl. 2, fig. 6). The first antenna of the female is shown in Text-fig. 20. This drawing has been made with the aid of a camera lucida, but antennules from three separate females have been used in order to include as many setae as possible. In none of the speci- mens was the basal segment or the terminal seta with attached sense club present. Lightly stippled “ ghosts’ of these have been added to the drawings by analogy with the other species. As will be seen by comparing Text-fig. 20 with Text-fig. 4 the first antennae of the two species are practically identical. The second antennae also appear to be identical in the two species (compare Text-fig. 21 with Text-fig. 5). Only the three terminal segments of this appendage with the five most terminal setae are preserved, however, and we could find no exopod. The mandible (Text-fig. 22) is similar in the two species, though the palp has less than seven whip-like setae on the posterior side of the penultimate segment, but this may have been because some of the setae had been lost. The palps of two mandibles 14 THE OSTRACOD GENUS TRACHYLEBERIS were recovered, but unfortunately in both examples the first segment of the palp was rather distorted and in neither was it possible to see the exopod. A number of parts of the legs were preserved, but the only complete one is the TEXT-FIGS. 20-25. Tvachyleberis scabvocuneata Brady. Type material. Fig. 20, first antenna of female, x 165, drawn from three specimens. Fig. 21, second antenna, x 165, from two specimens. Fig. 22, mandible, x 165. Fig. 23, second leg of female, x 260. Fig. 24, three segments of the third leg of a male, x 260. Fig. 25, copulatory apparatus of male, x 260, drawn from three specimens. Segments and setae shown lightly stippled were missing in the specimens seen, and have been added by analogy with the other species. second left leg of a female shown in Text-fig. 23. Instead of the group of bristles present on the proximal third of the second segment as in 7. /yételtonensis, there are two groups of short spinules placed laterally. Otherwise there seems to be no difference of any importance. Three segments of the third right leg of the male THE OSTRACOD GENUS TRACHYLEBERIS 15 are shown in Text-fig. 24. These are very similar to the corresponding segments of T. lytteltonensis (Text-fig. 22), and the ventral margins of the second and third segments are hairy. These hairs, however, do not seem to be arranged in distinct groups as they are in T. lytteltonensis. The copulatory apparatus (Text-fig. 25), of which no less than three specimens were preserved, is quite unlike that of T. lytteltonensis (compare Text-fig. 25 with Text-fig. 12). None of the three specimens was sufficiently well-preserved to show the muscles, but all three were remarkably like one another ; two of the three were right- and left-hand members of one specimen. The apparatus in the two species is so different that it is difficult to homologize the parts. In T. scabrocuneata there is a distinctly spiral ejaculatory duct which could not be seen in T. lytteltonensis. The distal parts of the copulatory appendage in 7. scabrocuneata is a long, rather curved organ with heavily chitinized walls, as shown in the figure, and very different from the broadly expanded, laminate and delicately-built organ of T. lytteltonensis. REFERENCES Brapy, G.S. 1880. Report on the Ostracoda dredged by H.MLS. “‘ Challenger ”’ during the years 1873-1876. Rep. Sci. Res. ‘‘ Challengery’’ Zool. 1 : 1-184, pls. 1-44. 1898. On new or imperfectly known species of Ostracoda, chiefly from New Zealand. Trans. Zool. Soc. Lond. 14 : 429-452, pls. 43-47. Hornisrook, N. DE B. 1952. Tertiary and Recent Marine Ostracoda of New Zealand. Palaeont. Bull., N.Z., 18 : 1-82, pls. 1-18. Miirer, G. W. 1912. Crustacea, Ostracoda. Das Tierveich, Lief. 31:i-xxxiii, 1-434, 92 Text-figs. SKOGSBERG, T. 1928. Studies on Marine Ostracods. Part II: External morphology of the Genus Cytheveis with descriptions of twenty-one new species. Occ. Pap. Calif. Acad. Sci. 15 : 1-154, pls. 1-6, 23 Text-figs. SYLVESTER-BRADLEY, P. C. 1948a. Bathonian Ostracods from the Boueti Bed of Langton Herring, Dorset. Geol. Mag. 85 : 185-204, pls. 12-15, 7 Text-figs. 1948b. The Ostracode Genus Cythereis. J. Paleont. 22: 792-707, pl. 122, 1 Text-fig. EXPLANATION OF PLATES PEATE x All external lateral views. All figures except fig. 7 from retouched photographs of specimens coated with magnesium oxide, x 60. Fig. 7 from an unretouched photograph, x 120, of a specimen immersed in benzyl alcohol and illuminated by transmitted light. Fics. 1-4, 7. Tvachyleberis lytteltonensis sp. nov. Figs. 1, 2, B.M. 1952.12.9.1 (3); figs. 3, 4, 7, B.M. 1952.12.9.6 (9). Fics. 5, 6, 8. Tvachyleberis scabrocuneata (Brady). Fig. 5, B.M. 1948.3.10.1 (g); fig. 6, B.M. 1948.3.10.2 (3); fig. 8, B.M. 1948.3.10.5 (9). PLATE 2 All figures are from retouched photographs, x 60. Figs. 1-6, dorsal views; figs. 7-10, internal lateral views. Fics. 1-4, 7,8. Tvachyleberis lytteltonensis sp. nov. Figs.1, 2 B.M. 1952.12.9.1 (3); figs. 3, 4, 7, 8, B.M. 1952.12.9.6 (9). Fias. 5, 6, 9, 10. Tvachyleberis scabrocuneata (Brady). Figs. 5, 9, B.M. 1948.3.10.1 (3); figs. 6, 10, B.M. 1948.3.10.5 (9). Bull, B.M. (N.H.) Zool. II, r PRACHYLEBERIS TRACHYLEBERIS a ») ' . = - x 7 ‘ bal Meat . * ri Do UR 4 Ae * , 5 > ’ 5 Rae * \ ¢ led Pye . u Me 5 ¥ c PY * a. vy - \ Wha Whar i Py } chy ea vi ¥ * = Se | at) he yh » » » a ie ho 3 * ‘ - ‘ , ‘ 7 oh A ° a x ‘ gore Me A GORDON _ ‘ ; > 4 o » Y * ’ Ren) : : ‘ a é <* fon a3 ite ae 7 Grad % Mar. he URAL Wisk si ‘ \y r A a . \ Pr ‘ i rly . ¢ ‘ ; i 7 i i i “~ ® 2 pe HULERTING OF 0 . MUSEUM (NATURAL HISTORY) _ LONDON: 1953 ae ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS (PALINURIDAE) PRESENTED 1, 3 MAR 1953 BY ISABELLA GORDON, D.Sc., Ph.D. British Museum (Nat. Hist.) Pp. 17-42; 9 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 2 No. 2 LONDON: 1953 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), stituted in 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical Serves. Parts appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper ts Vol. 2, No. 2 of the Zoological series. PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued March, 1953 Price Six Shillings ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS (PALINURIDAE) By ISABELLA GORDON SYNOPSIS The paper includes a key to the determination of the adult Palinurid genera to which the known post-larvae are referable, descriptions and figures of the Puerulus stage of several species of the genus Panulivus and notes on the Puerulus of the European Spiny Lobster. In an appendix the range of variation in the adults of the species Panulirus homarus (L.) is discussed. INTRODUCTION In March, 1952, I received from Hugh Copley, Esq., Fish Warden, Nairobi, an almost complete and beautifully preserved specimen of the Puerulus stage of a Palinurid. The antennae were intact, and the most striking feature of the post- larva was the dark brown spatulate apex of the long flagellum. A very similar Puerulus has recently been described in a Japanese journal which is not yet avail- able in this country ; as the text is mostly in Japanese a description in English seems essential. A number of Pueruli from the Marquesas and Coiba were found in undetermined collections in the British Museum and they are also described in the following short paper. All these post-larvae proved to belong to the genus Panulivus, and it has been possible to refer most of them to their respective adult species. All possess a pair of long, backwardly directed spines at the posterolateral angles of the thoracic sternum (near the bases of the last pair of walking legs). All agree, moreover, in having the exopodite of maxilliped 3 vestigial, and therefore belong to species in which that exopodite is lacking in the adult. The material from the Marquesas is referable to either Panulirus homarus (L.) or the closely related P. dasypus (Latr.) and, in an appendix, I discuss my reasons for believing that these two may belong to one variable species. In the Puerulus referred to P. argus (Latr.) by Gurney (1942, p. 234, fig. 93) the posterolateral thoracic spines are replaced by blunt nodules, and the exopodite of maxilliped 3 is rather longer than the ischium of the endopodite. The postero- lateral spines are absent in Puerulus pellucidus (Ortmann), which has been referred to Panulirus japonicus (von Siebold)—see synonymy in Holthuis, 1946, p. III, and Nakazawa, 1917. At the present time the post-larva is known for only four genera of the Pali- nuridae, namely, Palinurus Fabr., Panulirus White, Jasus Parker and Justitia Holthuis. I have examined Pueruli belonging to each of the first three genera but that of Justitia is known only from a single specimen collected by the “ Blake.’’ It was first described as the “ Puerulus d’Agassiz,” referable to Palinurus longi- manus H. M.-Edw. by Bouvier (1913, pp. 82 and 87; 1914, p. 187 ; and 1925, ZOOL. II, 2. § 18 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS p. 442, pl. 8, fig. 1), and Bouvier’s figure has been reproduced by Schmitt (1935, p. 173, fig. 37). Holthuis erected the genus Justitia for J. longimana (H. M.-Edw.) and its variety mauritiana (Miers) in 1946, p. II5. In the British Museum collection are the three specimens—Phyllosoma, Phyllo- soma with the Puerulus in the act of disengaging itself, and Puerulus—of Palinurus vulgaris Latr. figured by Bouvier (1914, figs. I, 2, and 3-6). Bouvier’s sketch of the carapace of the Puerulus is not very accurate as to the anterior spines, the supraorbital spines especially being much more prominent than in the actual speci- men (reg. no. 1914.12.17.4). I take this opportunity of refiguring the anterior part of the carapace and the antennule (Fig. 1a, 6 and c) for comparison with those of Jasus (Fig. td and e) and of Justitia as figured in Bouvier, 1925, pl. 8, fig. (which appears to be accurate). Incidentally, Holthuis says that the common European “Jangouste ’’ should be known as Palinurus elephas (Fabr.), which antedates vulgaris Latr. I also give a key to the determination of the adult genera to which the post- larvae are referable ; it is more detailed than that given by Bouvier (1913, p. 87), but for the characters of the Puerulus of Justitia I have had to rely on Bouvier’s figure and descriptions. I am not sure whether the antennae are intact, but if so they do not exceed 1-5 times the body length. KEY TO THE DETERMINATION OF THE ADULT GENERA TO WHICH THE POST-LARVAE ARE REFERABLE I. Antennular flagella distinctly shorter than the peduncle, outer much more robust than the inner (Fig. 1a and d). Anterior margin of the antennular! tergum much narrower than the space between the tips of the supraorbital spines. Rostrum well formed or rudimentary (Fig. 1a and d). The “ brevicornis ’’ group. A. Exopodite of maxilliped 3 long, with multi-segmented flagellum. Posterolateral spines present on several of the somites of thoracic sternum. Rostrum rudi- mentary. 1. Supraorbital spines denticulate on upper margin; plate formed by these spines and the rostrum broad and advanced so as to conceal the eye-stalks, the corneae visible laterally. Three pairs of spinules on thoracic sternum. Distinct median carina on carapace and abdominal somites 2-5. Only a few spines on carapace near the orbits. Second segment of antennular peduncle equal to third Justitia Holthuis. 1 The antennal tergum of some authors. Fic. 1. Palinurus elephas (Fabr.) = vulgaris Latr. The Puerulus figured by Bouvier, 1914, p. 288, fig. 6. (c.l.= 7:5 mm.). a. Anterior part of carapace, in dorsal aspect, with left antennule and antennal peduncle. b. Right anterolateral spine with the two spines immediately behind it, more magnified. c. Left antennule more magnified. Jasus lalandei (H. M.-Edw.). Puerulus from “‘ Discovery’’ Stn. 101, 15/x/26 (c.l. = 9:5 mm.). d. Anterior part of carapace in dorsal aspect, with left antennule and antennal peduncle. e. Left antennule more magnified. 19 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS Fic, 1, 20 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 2. Supraorbital spines not denticulate on upper margin; plate formed by these spines and the rostrum narrower and less advanced, most of eye-stalks visible (Fig. 1a). Four pairs of spinules on thoracic sternum. No median carina on carapace or abdomen. Carapace spiny, three spines at each anterolateral angle and a longitudinal tow of spinules replacing each lateral carina. Second segment of antennular peduncle shorter than third (Fig. 1c). [Antenna not quite 1°5 times the body length and tapering distally. First pair of pereiopods more robust than the others. ] 4 . Palinurus Fabr. B. Exopodite of maxilliped 3 short, without flagellum. No posterolateral spines on thoracic sternum. Rostrum well developed (Fig. 1d). [Supraorbital spines not denticulate on upper margin; plate formed by these and the rostrum narrow, moderately advanced, but most of eye-stalks visible. One spine at anterolateral angle of carapace. No median, but two lateral carinae on posterior two-thirds of carapace. Second segment of antennular peduncle rather shorter than third (Fig. re). Antenna not quite 1°5 times the body length and Eas hare First pair of Hgpeseio pos more robust than the others.] . ‘ : - Jasus Parker. II. Antennular flagella at least as ie as the outa outer not much more robust than the inner (Figs. 2a, 5a, c). Anterior margin of antennular! tergum as wide as the space between the tips of the supraorbital spines. Rostrum absent. The “ longicornis ”’ group. [Usually a large pair of posterolateral spines on last somite of thoracic sternum and three carinae on the carapace. One (occasionally two) spine(s) at anterolateral angle of carapace. Second segment of antennular peduncle equal to third. First pair of pereiopods not more robust than the others (Fig. 7), c).] - Panulirus White. a. THE PUERULUS STAGE OR POST-LARVA OF PANULIRUS sp. (Figs. 2a ; 3; 7c) MATERIAL. Taken in trawl by M.F.V. “ Menika” on Leopold Reef ground, Malindi, Kenya Colony, in 10 to 20 fms.—one specimen, presented by H. Copley, Esq., Fish Warden, Nairobi. March, 1952. Measurements : Length of body 21-5 mm. ; Length of carapace 8 mm. ; Length of antenna (complete) 45 mm. DESCRIPTION. The specimen had been preserved in formalin and the body, though transparent in life, was apparently tinged with reddish brown. When received it was complete except that pereiopods II and V on the right side were missing. The most striking feature of the post-larva is the enormously long an- tennae, measuring just over twice the body length. The flagellum ends in a flattened oar-like expansion which is all the more conspicuous because the last 7 or 8 segments of the spatula are brown in colour (Fig. 3a ; brown portion stippled). A narrow 1 See footnote on p. 18. Fic. 2. a. Puerulus of Panulivus sp. from Kenya, in dorsal aspect, most of the antennal flagella omitted. 6. Distal portion of antenna of the last stage Phyllosoma of P. homarus (L.) from the Marquesas. c and c’. Distal segments of incomplete antennal flagella of two Pueruli of P. homarus (L.) from the Marquesas. d. Terminal segments of antennal flagellum of the Puerulus of P. avgus (Latr.) from Bermuda, ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS ZOOL. Il, 2. 21 22 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS band of the brown colour, a trifle paler, is present about one-third of the way along the flagellum (i.e., at distal end of proximal third). Fig. 2a represents the post-larva in dorsal aspect, with the greater part of the antennal flagella omitted. The carapace measures 8 mm. from the tip of the supra- orbital spine to the posterior margin ; it bears only two pairs of spinules in addition to that at each anterolateral angle. The anterior pair is situated immediately behind the eyes ; the posterior pair is some distance behind, one spinule at the anterior end of each lateral carina. If the specimen is examined carefully under a binocular microscope, three pairs of incipient spinules can be detected situated between the anterior and posterior pairs of spinules, i.e., two median (gastric) pairs a short distance in front of the median carina, and another pair posterior to the carinae on the supraorbital spines. The three longitudinal carinae on the posterior two-thirds of the carapace are not very clearly seen in dorsal aspect, but in profile they stand out clearly : Each lateral carina seems to be continued forwards to the anterior pair of spinules ; the incipient spinules appear as thickenings of the chitin, so that the median carina seems to bifurcate and to be continued forwards to the base of each rostral horn. A pair of incipient spinules is present on the antennular tergum, near the anterior margin. A thin carina, not visible in dorsal aspect, runs parallel to the lateral edge of the carapace. The cervical furrow is not very apparent at this stage. The thoracic sternum is very similar to that of the Puerulus of Panulirus regius B. Capello figured by Bouvier (1917, pl. ix, fig. g) ; in addition to the pair of large, backwardly-directed spines on the last somite, near the bases of pereiopods V, there are blunt lobules, which diminish in size anteriorly, at the postero-lateral angles of each of the three preceding somites. Bouvier’s fig. 9 does not show these, but that on the penultimate somite is visible in lateral aspect in his fig. 10. The thoracic ganglia are clearly visible through the thin integument. The abdomen and tail fan present no unusual features. No groove is present on any of the terga parallel to the posterior margin, but the anterior margin of the succeeding somite is visible through the thin chitin and is represented by a broken line in Fig. 2a. The pleura of somites 2 to 6 each possesses a sharp, backwardly directed spine which is not visible in dorsal aspect. The antennule is short, extending only a little way beyond the antennal peduncle (Fig. 2a) ; the external flagellum is the more robust, a trifle longer than the internal and beset with special sensory setae on all but the proximal 6 or 7 segments. Each flagellum has 18 or 19 segments. The details of the peduncle are shown in ventro- lateral aspect in Fig. 3b and c, of the basal segment in dorsal aspect in Fig. 3d. All the setae are very finely plumose, as are those forming what Gilchrist (1920, p. 198, fig. 13) calls the “ antennular screen.” In the Puerulus of Jasus lalandei (H. M. Edw.) Gilchrist shows these setae as heavily plumose. The basal segment Fic. 3. Puerulus of Panulirus sp. from Kenya. a. Terminal portion of antennal flagel- lum, to show the spatulate apex. 6b. Antennal and antennular peduncles, in ventro- lateral aspect. c. The latter further enlarged. d. Basal segment of antennular peduncle, in dorsal aspect. e. Pleopod on second abdominal somite, with most of the long setae omitted, and apex of appendix interna further enlarged. ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS Fic. 3. 24 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS of the peduncle is equal in length to the sum of the other two segments ; the stato- cyst is conspicuous but without any statolith. The details of the antennal peduncle are shown in Fig. 2a in dorsal and in Fig. 3b in ventrolateral aspects. The spatulate apex of the long flagellum is represented in Fig. 3a, the shaded terminal segments being, in the preserved specimen, brown in colour. The eyes are large and conspicuous, extending as far as the anterolateral spinose angles of the carapace. The external maxillipeds are widely separated basally ; when extended they reach the anterior margin of the antennular somite ; a minute trace of an exopodite is present. The mouthparts were not dissected and examined in detail as I did not wish to damage the specimen. But they appear to be very similar to those of a Puerulus from the Marquesas represented in Fig. 6a—e except that the exopodite on maxilliped 2 is rather shorter, being about three-fourths of the merus. The mandibular palp is clearly seen without dissection. Pereiopod | is appreciably shorter and somewhat more robust than the succeeding limb (Fig. 7c) ; the propodus is not quite three times as long as wide and the ratio of dactylus to propodus is I : 1-25 ; the merus is rather longer than carpus and propodus. Pereiopods II, III and IV are very similar in size and form, the propodus in each being almost exactly five times as long as wide ; the ratio of dactylus to propodus is I : I-75-I-7o ; the merus equals propodus and carpus. The shrivelled remnant of the exopodite at the base of pereiopods I—IV is very small and is best seen when the limb (ischium to the tip) is detached, leaving the rudimentary exo- podite exposed on the stump. Pereiopod V is the shortest and shows no trace of the exopodite ; the propodus is 5 times as long as wide. Pleopod 1, that attached to the second abdominal somite, is represented in Fig. 3e ; most of the long setae on the exopodite and endopodite are omitted ; the appendix interna is long, and bears a subterminal plumose seta and a number of terminal coupling hooks (see enlarged view of apex). Remarks. Unfortunately the post-larvae of the Palinuridae are seldom complete as to antennae, and this is especially true of those with long flagella. Several Puerulus stages, however, have been figured with the antennal flagellum tapering : e.g., that of Panulirus argus (Gurney, 1942, p. 234, fig. 93 ; Schmitt, 1935, p. 173, fig. 36 of the young crawfish), and I have examined the specimen figured by Gurney and find that the apex of the flagellum is as represented in Fig. 2d ; that of P. regius B. Capello (Schmitt, 1926, p. 43, fig. 67) ; of P. japonicus (Nakazawa, 1917, plate, figs. 3, 4, and 5 of the post-Peurulus). In these the antenna does not seem to exceed, if it attains, I-5 times the body length. Dr. Holthuis of the Leiden Museum kindly informed me by letter that a Puerulus with a long spatulate antennal flagellum had recently been described by Kubo in a paper written almost entirely in Japanese, but with a short summary in English (first paragraphs, p. gI—Kubo, 1950). As no copy of this paper was available in London, Dr. Holthuis sent me a photostat copy. (Later I received a reprint from the author—see p. 28.) Kubo attributes this Puerulus to Panulirus “‘ versicola”’ (Latr.)—presumably a misprint for versicolor—and finishes his English abstract ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 25 as follows : ‘“ Some descriptions in regard to the puerulus of P. versicola have been given by Ortmann (1894) [error for 1891], Calman (1909), and de Man (1916). But no one has given description on the signal feature of the antennae.”’ Kubo’s Puerulus is certainly very similar to that from Kenya ; his figures are small, and in the photostat copy some of the finer details are lost. The flagellum is just over twice the body length and ends in a spatulate expansion which is “stained with rather deep brown colour (in alcoholic specimens), and I think his figure shows the brown band about one-third of the way along the flagellum. But I do not think that his Puerulus belongs to the same species as that from East Africa described above. As a rule his specimens have two pairs of spines, one in front of the other, at the anterior end of the lateral carinae and those behind the rostral horns are well formed. The general impression one gets from Kubo’s figures is that the antennal and antennular peduncles are more robust than those of the Kenya specimen—more like those of specimens from the Marquesas described below (see p. 29 and Fig. 5a, b.). Nor do I think that Kubo is right in attributing all his post-larvae to Panulivus versicolor (Latr.) ; at any rate those figured do not agree with the specimens from Christmas Island which Dr. Calman (1909, p. 444) referred to that species (see p. 26). I have found an earlier reference to a Puerulus with a spatulate apex to the antennal flagellum in a paper by W. von Bonde (1930, p. 25). In his specimen the antenna is also about twice the body length. The following extract may be quoted from von Bonde’s paper : “‘ NotrE.—Only one of my Puerulus specimens of Panulirus has the antennae intact. Each is developed at its distal extremity into a large flattened oar-like structure. Such an antenna has been figured by Richters (12) [error for (13)] in a large Phyllosoma (Ph. longicorne Guér., length 38 mm.).”’ Richters (1873, pl. xxxi, fig. 3) shows a last stage Phyllosoma with a long antenna, the clavate apex of which is rather wider than that represented in Fig. 2b and has no terminal spinose projection. The fragment represented in Fig. 2b was found in the jar along with a number of post-larvae and young from the Marquesas (see p. 29). Thus several species of the genus Panulirus possess these spatulate antennae in the post-larval stage. Leach’s type specimens of Phyllosoma clavicorne were col- lected in West African waters ; von Bonde’s post-larvae, which he referred pro- visionally to P. biirgeri de Haan (= homarus), were from the East Coast of South Africa ; others are now known from Japan (Kubo, 1950), Kenya and the Marquesas. As von Bonde did not figure his specimens, I do not know from his description whether the antennules are like those of the Marquesas material (Fig. 50), or more slender as in the Kenya specimen (Fig. 3c). But, as they are stated to have two pairs of branchial spines on the carapace they agree in this with the Marquesas Pueruli and it is probable that they are referable to P. homarus (L.). In the absence of post-Puerulus stages it is not possible to refer the Kenya specimen to the adult species, since it is not known whether abdominal grooves are present or absent. What can be said with certainty is that it belongs to a species with longer and more slender antennules than either homarus or dasypus (if the latter is regarded as a distinct species—see p. 35). 26 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS b. THE POST-LARVA OF Panulirus versicolor (Larr.) (Fig. 4) MATERIAL. Christmas Island, Indian Ocean. Dr. C. W. Andrews coll. partly on the reef and partly in crevices in the piles of the pier at Flying Fish Cove, reg. no. 1909.5.19.248-252. Five post-larvae. (Also several later young crawfish stages, reg. no. 1909.5.19, 243-247). DESCRIPTIVE NOTE. These specimens have been identified by Calman (1909, p. 444), and his determination is fairly certain since he had older stages, up to 74 mm. in length of body, collected at the same time and place. In none of his specimens is the antennal flagellum complete, but in one it exceeds 2-5 times the body length and yet shows no trace of expansion distally—thus it must be consider- ably longer than that of either the Kenya specimen described above, or of Kubo’s post-larvae referred to ‘“‘ P. versicola”’ (Kubo, 1950, p. 94, fig. 3). De Man (1916, p. 61) says of his specimens “ external antennae 3-times as long as the body ”’ ; this I at first regarded as a misprint for ‘‘ 2-times ”’ (as Kubo may also have done), but Calman’s material suggests that de Man’s statement may be correct. If de Man’s specimens were complete the apex of the flagellum was not spatulate or he would certainly have said so. I have since examined de Man’s specimens in the Zoological Museum, Amsterdam. They certainly agree with our Christmas Island material but in none are the antennae now intact. In all the post-larvae there are only three pairs of spinules on the carapace—one at the anterior end of the lateral carinae (not two, as in the majority of Kubo’s specimens), one behind the orbits and one behind the supraorbital spines ; in addition two pairs of incipient spinules are present on the cardiac region. They are thus similar to the specimen from Kenya except that the pair behind the rostral horns is distinct. The antennal peduncle is long and slender, rather like but not identical with that of the Kenya specimen (cf. Figs. 4a, b and 2a, 3c, da). The basal segment differs in that the portion containing the statocyst is shorter while the distal part is narrow and not expanded apically. A large round statolith is visible within the sac (as in all the post-larvae examined other than the one from Kenya, in which it may not have had time to form after a recent moult). Unfortunately, in all the specimens the antennal peduncle is partially collapsed on the inner surface, but it also is relatively long and slender. The pereiopods have also collapsed in most of the specimens, but in one they are normal and I and III are represented in Fig. 4c, at the same magnification as those of the other Pueruli. They are slightly longer than those of the Kenya specimen (Fig. 7c), or of the Marquesas material described below (Fig. 78). NOTE ON THE POST-PUERULUS AND LATER STAGES As stated by Calman (1909, p. 444), the post-Pueruli are brown in colour, with a conspicuous W-shaped marking of white on the carapace and a transverse narrow white band on each abdominal somite, that on somite I on the anterior half, those on somites 2-6 parallel to the posterior margins. No abdominal grooves are present, ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS Fic. 4. Puerulus of Panulivus versicoloy (Latr.) from Christmas Island. a. Anterior portion of carapace and antennal and antennular peduncles, in dorsal aspect. b. Antennular peduncle in dorsal aspect. c. Pereiopod I (upper) and pereiopod III. 27 28 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS as in the adult, but on each somite a number of transverse rows of minute pits, from each of which a short seta emerges, are present. A rather long forwardly directed seta arises from the base of several of the median spines on the carapace. After over 40 years in alcohol the colour pattern of these post-Pueruli is still clear. In the two older stages from Christmas Island (c./. 25 and 34 mm. respectively) the colour pattern is much nearer that of the adult ; the legs are striped, the dark purplish blotches on the carapace tend to obscure the short median bands of the W. The posterior whitish band across each abdominal somite is now flanked anteriorly and posteriorly by a purplish band. What is rather surprising, however, is the presence of a broad but very shallow transverse groove, with slight median interruption, on abdominal somite 2. The pits and setae are now confined to this area, the groove being filled with them while a few continue across the median break. On somite 3 there is a much wider median interruption, beset with sparse setae, between the two setose lateral grooves. On somite 4 there are just two short widely separated lateral patches of setae. Any setae on 5 and 6 are isolated and few in number. These broad shallow grooves on somites 2 and 3, or traces of them, are present in rather larger stages from other localities in the B.M. collection. RemARKS. As already mentioned on p. 25, I do not think that all Kubo’s speci- mens belong to P. versicolor (Latr.). The antennal flagellum in the Puerulus of that species seems to be three times the body length and without a spatulate apex. Since I commenced this paper I have received a reprint from Dr. Kubo and I find that his fig. 4 0 shows the W pattern rather faintly on the carapace of a young stage (this was not apparent in the photostat copy). I now think that part of Kubo’s material only belongs to P. versicolor—namely the young with the charac- teristic pattern and perhaps the Pueruli which he lists with but one spine at the anterior end of each lateral carina (Kubo, 1950, p. 93, table 1). The other Pueruli which he figures, with two spines at anterior end of the lateral carina and the stouter antennal and antennular peduncles (Kubo, 1950, figs. 2 and 3) may belong to P. homarus (L.) like my specimens from the Marquesas. I do not know how many post-Puerulus and young stages Kubo had. But I assume that in his material no abdominal grooves were present, since that is the case in P. versicolor. In the post-Puerulus of P. homarus the abdominal grooves are present and show traces of the crenulation as well. If Kubo’s Puerulus with the spatulate antennae really has no abdominal grooves in the next stage, it might be referable to either P. ornatus (Fabr.) or P. polyphagus (Herbst). The adult of the latter species has a long multi-segmented flagellum on the exopodite of the second maxilliped as represented in Fig. 8/, and one would expect a well formed flagellum in the Puerulus also—as in that of P. inflatus (Bouvier) represented in Fig. 7a. Kubo’s Fig. 4 K seems to have no flagellum, or a vestigial one as in fig. 6d perhaps. So that P. polyphagus would appear to be excluded. On the other hand, P. ornatus does not appear to have been recorded from Japanese waters, although it goes as far north as Formosa (Holthuis, 1946, (Os 2eslre,)). ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 29 c. THE POST-LARVA OF Panulirus homarus (L.) (Figs. 2b-d ; 5a,b; 6; 7b) MATERIAL. (a) Marquesas. “St. George’’ Expedition. W. 38.27.12.24. One Puerulus. (6) Marquesas ; sand of beach, Hana Hevané. C. Crossland coll. 1925-26. Twenty four Pueruli, nine post-Pueruli and young crawfish stages up to 33 mm. in body length. DESCRIPTIVE NOTES. The 25 Puerulus stages vary in body length from 22 to 24 mm. As each specimen has a narrow band of brown pigment some distance along the antennal flagellum, I thought at first that they must be identical with the specimen from Kenya described above. The antennal flagellum in some speci- mens shows a hint of expansion of the segments at the broken distal end (Fig. 2c and c’) and the flagellum in these is nearly twice the body length. My suspicion that the apex of the flagellum must be spatulate was confirmed in a rather un- expected way. Lying in the jar amongst detached limbs was the distal end of the antennal flagellum of the last stage Phyllosoma represented in Fig. 2b. The larval flagellum is unsegmented, but within this can be seen the fully formed flagellum of the Puerulus with the segmentation faintly visible through the chitin. It is obvious that the apex of this flagellum would be spatulate. Presumably this portion of the larval antenna broke off just before or during the critical moult from larva to post-larva. The apex of the larval antenna is very similar to that of the last stage Phyllosoma clavicorne Leach, the types of which are in the British Museum collection. The antennal flagellum of these Puerulus stages, therefore, seems to agree with that of the specimen from Kenya. But on closer examination they were found to differ from the latter in a number of characters. The resemblances are many, chiefly in characters common to nearly all the Panulirus post-larvae that have been described such as the presence of the posterior pair of spines on the thoracic sternum. The chief differences are as follows : (i) In all specimens, even the smallest, four pairs of spinules are present on the carapace in addition to the spinose anterolateral angles, namely, a post- ocular pair, a postrostral pair, and two pairs one behind the other at the anterior ends of the lateral carinae. Two pairs of incipient spinules are present on the gastric region as in the Kenya specimen. (ii) The antennules are more robust ; the flagella have some 25 and 23 segments respectively ; the segments of the peduncle are in the ratio of 2 : I : I, as seems usual at this stage, but each segment is broader relatively to its length (cf. Figs. 5b and 3c, d). The first segment is 2-5 times as long as wide, the portion containing the statocyst is nearly as long as the distal portion and a conspicuous statolith is always present. In the specimen from Kenya the basal segment is 4 times as long as wide and no statolith was visible. (iii) The antennal peduncle is shorter and more robust. (iv) The pereiopods also seem to be somewhat stouter, especially as regards the merus (cf. Fig. 7b and c). The mouthparts are represented in Fig. 6a-e. 30 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 2mm. 1mm. 2mm. Fic. 5. Puerulus of Panulivus homarus (L.) from the Marquesas. a. Anterior part of carapace with left antennal and antennular peduncles, in dorsal aspect. b. Anten- nal peduncle. Puerulus of Panulivus inflatus (Bouvier) from Coiba. c. Anterior part of carapace, antennal and antennular peduncles, in dorsal aspect. d. Antennal peduncle. ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 31 Remarks. Because several later stages are represented it seems possible to refer these young specimens to the adult species. These young crawfish possess an abdominal groove on each somite, and, moreover, crenulations are present in the older specimens in the lateral portions of each groove. The furrows are slightly interrupted in the centre on the second and subsequent somites. The largest young crawfish measures only 33 mm. from tip of supraorbital spine to tip of telson, but it has a tiny anterodorsal spine on the merus, near the distal articulation, of pereiopods II and III. The soft part of the telson is almost as wide as the hard basal portion ; pereiopod III is longer than pereiopod II by half the length of the dactylus ; the antennular peduncle is now very nearly equal to the antennal peduncle, and the basal segment of the former reaches almost to the distal end of the penultimate segment of the latter. Only one pair of slender spines are present on the antennular tergum at this stage. Many spines are present on the carapace. Holthuis (1946, pp. 110-112) lists the following species as widely distributed in the Indo-Pacific region: Panulirus dasypus (Latr.), P. homarus (L.), P. japonicus (von Siebold), P. ornatus (Fabr.), P. penicillatus (Oliv.), P. polyphagus (Herbst) and P. versicolor (Latr.). Of these the species without abdominal grooves or furrows may be excluded, leaving four—dasypus, homarus, japonicus and penicillatus. The post-larva of P. japonicus has been described by Nakazawa (1917, figs. 3, 4, and 5); he figures the antennae of the Puerulus as not exceeding 1-5 times the body length and tapering distally. According to Bouvier (1913, p. 87, key) there are no posterolateral spines on the thoracic sternum in “ Puerulus pellucidus,” which is referred to P. japonicus. In any case the antennule of the adult japonicus (de Haan, 1833-49, pl. 41/42) has an unusually long slender basal segment to the peduncle, equal in length to the sum of the three segments of the antennal peduncle, and one would expect the specimens from the Marquesas to belong to a species with a relatively short robust antennule. P. pemicillatus can also be excluded because (i) the margins of the abdominal grooves are never crenulate, and (ii) there is in the adult a short but distinct exopodite on maxilliped 3 so that in the post- larva one would expect the exopodite to be larger than it is in these specimens. As the grooves on the abdominal somites are already distinctly crenulate laterally and slightly interrupted in the median line I have no hesitation in referring these young stages to either P. homarus or P. dasypus. Most authors are agreed that these two species are very closely related and, in the Appendix, I give reasons for believing that they belong to one rather variable species to which the older name P. homarus (L.) should be given. A short description, without figures, of the Puerulus and post-Puerulus stages of Panulirus dasypus from the Red Sea was given by Bouvier (1913, p. 84 ; key to the known Pueruli of the Palinuridae, p. 87). These specimens were found “ dans les crevasses du rivage, 4 Djiboutil ’’ (Bouvier, 1913, p. 88, and 1914, p. 191). These young stages appear to be inhabitants of the littoral zone. The presence of crenula- tions and slight interruptions on the abdominal grooves at this stages does not necessarily mean that the grooves and sculpturing of the adults will remain of the B-form (or dasypus form)—see p. 38. ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS Fic. 6. Puerulus of Panulivus homarus (L.) from the Marquesas. a. Mandibles, left maxillule and lower lip. b. Maxilla. c. First maxilliped. d. Second maxilliped. e. Third maxilliped (distal part of epipodite omitted). ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 33 d. THE POST-LARVA OF Panulirus inflatus (Bouvier) (Figs. 5c, d ; 7a) MaterrAL. Coiba Island. “St. George’”’ Expedition 1924-25. One specimen measuring 22 mm. in length. DESCRIPTION. This specimen has the carapace more swollen and distorted than usual. It agrees with most of the Pueruli that I have examined in having a pair of posterolateral spines on the thoracic sternum near the bases of pereiopods V. It differs from all the others in having two spines, one immediately behind the other, on each anterolateral angle of the carapace (Fig. 5c) and a backwardly directed spine at the posterior end of the median carina, i.e., on the posterior margin of the carapace. In addition to the spinules shown in Fig. 5c, namely, at the antero- lateral angles, behind the orbits and supraorbital spines respectively, and two very minute pairs on the gastric region, there is only one at the anterior end of each lateral carina. The antennae are unfortunately incomplete, so that the nature of the apex is unknown ; the longer exceeds 1-5 times the body length and shows no trace of any expansion—which does not necessarily mean that the apex is tapering. The details of the antennular and antennal peduncles are shown in Fig. 5c and d ; the former is more robust than that of the Puerulus from Kenya, or of the specimens referred to P. versicolor (Latr.). They recall those of the post-larvae from the Marquesas described above ; the region of the statocyst is rather opaque but the round statolith is visible. The mouthparts have not been examined in detail, but there is just a vestige of the exopodite on maxilliped 3 (as in Fig. 6e). The exo- podite of maxilliped 2 has a long flagellum, and in this respect differs from that of the Marquesas material (cf. Fig. 7a and 6d). Remarks. According to Holthuis (1946, pp. 110-111) there are but two species from the Pacific Coast of Central America, namely, P. inflatus (Bouvier) and P. interruptus (Randall). The latter species possesses in the adult the longest exo- podite to maxilliped 3 of all Panulirus species and therefore its post-larva should have an even longer exopodite than that of, e.g., P. argus (Latr.) figured by Gurney (1942, p. 234, fig. 93E). Since the adult of P. inflatus lacks the exopodite on maxilliped 3 and has a large multi-articulate flagellum on the exopodite of maxilliped 2, I have no hesitation in referring this Puerulus to that species. Amongst the registered post-larvae in the B.M. collection there is another speci- men from Esmeraldas, Ecuador (reg. no. 1925.12.8.4), which also belongs to P. inflatus. The carapace is normal, and again there is a double spine at each antero- lateral angle, but the posterior spine, that at the distal end of the median carina, is lacking. The antennal flagella are incomplete, and there is a rather broad band of brown pigment, a considerable distance from the peduncle, on the left side only. In both specimens there are two rows of 3-4 minute spinules on the proximal half of the soft part of the telson ; no spinules are present in the post-larva from Kenya (Fig. 2a) ; in the Pueruli of P. homarus there may be two pairs of minute spinules 34 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS in some of the specimens ; a few spinules are also present in some of the post- larvae of P. versicolor. The holotype of “‘ Puerulus inermis ’’ Pocock, said to be the post-larva of Panulirus gultatus (Latr.) (Holthuis, 1946, p. I10), has the antennular peduncle of the shorter Fic. 7. Puerulus of Panulivus inflatus (Bouvier) from Coiba. a. Right maxilliped 2. Puerulus of Panulivus homarus (L.) from the Marquesas. 6. Pereiopods I, II and III (from left to right). Puerulus of Panulivus sp. from Kenya. c. Pereiopods I (above), II and ITI. broader form found in P. homarus and P. inflatus (Fig. 5a-d). The exopodite of maxilliped 2 has a shorter flagellum than that of the Puerulus of P. inflatus (Fig. 7a) and does not reach beyond the distal third of the carpus of the endopodite. The exopodite of maxilliped 3 is vestigial, as in Fig. 6e. The antennal flagellum is ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 35 incomplete. Pocock’s specimen was dredged in about ro fms. in Water Bay, Fernando Noronha (Pocock, 1890, p. 516). APPENDIX NOTE ON VARIATION OF THE ADULTS OF Panulirus homarus (L.) (Figs. 8 ; 9) Two Indo-Pacific species of Panulirus, namely P. homarus (L.)—better known as P. btirgert (de Haan)—and P. dasypus (Latreille), are apparently very closely related. The differences between them are indicated in the following extract from Gruvel’s key (1911, p. 50) : “Exopodite des deuxiéme maxillipédes avec fouet atrophié. Sillons des tergites abdominaux toujours dentelés et nettement ininterrompus 4. P. biirgeri, de Haan. ““Exopodite des deuxiéme maxillipédes sans fouet. Sillons des tergites abdominaux dentelés et toujours au moins légérement interrompus sur la ligne médiane 5. P. dasypus, Latreille.’”’ Recently Holthuis (1946, pp. 128-136) has attempted to find clear-cut characters by which the two species may be separated ; he lists only some minor differences, and sums up by saying that more abundant material may perhaps show that the two are identical. In recent years I have had considerable correspondence with Dr. Leakey, of the Coryndon Museum, Nairobi, relating to the East African species, and from time to time he has sent me a few specimens for determination. Some of these seemed to agree with P. homarus, others with P. dasypus as far as the abdominal grooves were concerned, but, as Dr. Leakey pointed out, not always as regards the second maxilliped. For example, two specimens with the deep crenula- tion and uninterrupted furrows typical of P. homarus differed as to the exopodite of maxilliped 2 ; in one a small one-segmented flagellum was present as in Fig. 84d, whereas in the other the flagellum was absent as in Fig. 8c. The former therefore agreed with Gruvel’s definition of P. btirgeri (avec fouet atrophié), while the latter agreed with that of P. dasypus (sans fouet). Two other specimens with the crenula- tion much less marked, especially mid dorsally, differed as follows : one was of a reddish colour, with uninterrupted grooves and no flagellum on the exopodite of maxilliped 2 (as in Fig. 8c) ; the other was bluish green, with a short but distinct median interruption of the groove on somites 2, 3 and 4 respectively, whereas the exopodite of maxilliped 2 has a large flagellum. (In 1948 I did not observe that the flagellum differed on right and left sides : that on the right is represented in Fig. 8a; the flagellum can scarcely be described as “‘ atrophié,” being half the length of the basal segment of the exopodite and rather irregularly segmented ; that on the left is as represented in Fig. 8d and “‘ atrophié.’’) I have now re-examined the spirit specimens in the British Museum collection that have been referred to P. btirgeri (de Haan) and P. dasypus (Latreille) by Dr. Calman, who revised the material of Panulirus at one time (unpublished), and by 36 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 3mm. 3mm. Cc Fic. 8. Portions of the second maxilliped, in anterior aspect, to show the exopodite in various species of Panulivus. a-—f. Adults, g and h very young specimens. Panulivus homarus (L.). @. From specimen no. 3 in Table, with distal part of exopodite from the side normally facing the endopodite. 6. Distal part of exopodite of right maxilliped of specimen no. 9. c. Left maxilliped of specimen no. 4. d. From specimen no. 6. e. From specimen no. tr. g. Right maxilliped of largest young stage from the Marquesas (c./. = 13 mm.). Panulivus polyphagus (Herbst). f. Left maxilliped of a 9 (c.l. = 66 mm.) with the flagellum, in inner aspect, more enlarged. Panulirus versicolor (Latr.). h. Right maxilliped of a young specimen (c.l. = 25 mm.) from Christmas Island. ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 37 myself. Unfortunately there are only fifteen specimens, three of which were referred by Dr. Calman to dasypus and eight to btirgert, the remaining four being those already mentioned from East Africa. Table I shows briefly the variation in (a) the abdominal grooves and the sculpturing in their vicinity, and (b) the exopodite Fic. 8—e-h, 38 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS of the second maxilliped, exhibited by the twelve specimens which obviously belong to the biirgeri-dasypus complex. At first I thought that the six specimens with the TABLE I Panulivus homarus (= btirgeri)—specimens in B.M. Collection.? Abdominal 4 cl. Reg. no. oes Exopodite of second maxilliped No Sex in Br aah Locality and Flagellum Form of flagellum sculpturing ZT =. @ = 105) >. f906.5520.365). Ava & Absent R. Fig. 82. Socotra A 1b, As in Fig. 8e. 2 3 c. 90 Lo48e3r2en B+ 7 R. Alike—as in E. Africa L. } Fig. 8c. Be Gun Nem TOG meme LO4Secyi tla) /2 mre B+ c Present R. Fig. 8a. E. Africa 2535.4 (dissimilar) L. As in Fig. 8d. & 88 . Kenya,t950 . A+ ., Absent a es 3 = Gils Ue LOZ Seok oon A a R. é Off Natal Lf ee 6 2 80 . Kenya,1950 . A+ , Present R. Fig. 8d. (similar) L. As in Fig. 8d. 7 g HB. 28 /xi/08 Q B+ . Present R. Fig. 9a. ovig. No data 23: (dissimilar) L. Fig. 9a’. 8 OF =) (Gi eG258-08.60) 0. A+ : R. As in Fig. 8e. juv. Off Natal Absent IU Like R, but slightly shorter. 9. OP: 60) =) TOTO.3\20°536 = B : Present R. Fig. 80. juv. Goram Id. Fp FB (similar) L. As in Fig. 8b. LOS ee) 7 LOZS eo) Ome A é Absent R. Alike—as in juv. S. Africa Te Fig. 8c. LTS My meas eT OO, |e 80.6 5 B “ Present R. Alike—as in juv. Amboina (similar) 1p Fig. 8d. E21 eeu meas Ones 8204 B— . R. absent R. As in Fig. 8c. juv. Ceylon 3,4 I. present L. As in Fig. 8d. A ++. Crenulation and sculpturing very pronounced, of the “‘ megasculpta’’ form. A +. Intermediate between A and A ++ ; some additional sculpturing and crenulation on certain somites. A. Crenulation well marked along entire anterior margin of each uninterrupted groove. B. Crenulation moderately developed, most marked laterally, on anterior margin of groove. B +. As for B, but with traces of pitting on somites 2 and 3, or 2-4. B —. Crenulation faintly indicated laterally, fading out medially. Some specimens of the B-form show a slight median interruption of the grooves on the somites indicated below B. R and L = right and left second maxilliped respectively. c, l. is measured from tip of supraorbital spine to posterior margin; in specimens 2 and 3 the supraorbital spines are damaged. 1 Recently, in the Zoological Museums of Leiden and Amsterdam, I have examined the material referred to P. homarus and P. dasypus, and find that it is just as variable as regards mouthparts and abdominal grooves as the B,M, material, ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 39 A-form of grooves and sculpturing belonged to birgeri, the other six with grooves and sculpturing of the B-form to dasypus. In five of the former the flagellum of maxilliped 2 is absent, whereas in five of the latter it is present on both sides or (in one instance) on one side only. But the fact that in three out of the twelve specimens the exopodites are dissimilar on right and left maxillipeds (nos. 3, 7 and 12), and that the three forms of exopodite are thus combined (flagellum present and either long and 2- to 3-segmented, or short and 1-segmented, or absent), proves that this distinction is untenable. I am forced to conclude that these twelve specimens belong to a single variable species, P. biirgeri (de Haan), which according to Holthuis must now be called P. homarus (L.). In all probability many, if not all, of the references to P. dasypus in the literature refer to the less crenulate form of P. homarus (this difference does not depend on size as far as one can see). Speci- men no. I in the Table, from Socotra, is that referred to as P. dasypus in Pocock, 1903, p. 214; the sculpturing on the abdomen is very pronounced and of the “ megasculpta’’ form and I have no hesitation in referring it to biirgert. The exo- podite of maxilliped 2 is as long as the merus and yet there is no trace of a suture near the apex (Fig. 8e) ; Monod & Petit (19209, fig. 3E) figure the exopodite as only half the length of the merus in their P. dasypus, so that the range of variation here is considerable. As this is the largest specimen, the spines on the carapace are rather worn, but they are not as few in number as in Gruvel’s specimen of dasypus (19gII, pl. ii, fig. 5). The anterodorsal spine on the merus of pereiopods II and III, with one exception, have almost entirely vanished, so I do not think that the presence or absence of these spines is of specific importance. Another difference listed by Holthuis (1946, p. 135), namely that pereiopods II and III are about equal in homarus whereas III is appreciably the longer in dasypus, is, I think, capable of another explanation. In small specimens of P. homarus, irrespective of sex, the pereiopods are relatively short. In the juvenile from Goram Island, for example, pereiopod I reaches the distal end of the second segment of the antennal peduncle, II is just a trifle longer and III a trifle longer still, reaching the distal end of the third segment of the peduncle ; IV reaches the middle of the propodus of I. By the time the carapace has increased to 88 mm. in length (no. 3 in Table) all the legs are a little longer relatively to the antennal peduncle, but the difference between II and III is not marked ; I exceeds the second segment of the peduncle by its dactylus, while III exceeds I by rather more than its dactylus (by 24 mm.) and IV is now a trifle longer than I. In the largest specimen (no. 1) pereiopods I and IV are subequal and reach the distal end of segment 3 of the peduncle much as before, but II exceeds I by the dactylus and one-fourth of the propodus (33 mm.), while III, now clearly the longest, exceeds II by its dactylus (30 mm.) and I by 63 mm. This specimen thus exhibits the commencement of a feature peculiar to senile males at least. The difference gets more pronounced with age, and this may well account for the long third pereiopod in Gruvel’s specimen of P. dasypus (1911, pl. ii, fig. 5). Gruvel does not mention the sex of the specimen in the legend to the figure, which is one-fourth natural size, giving a c.l. of about 200 mm. Recently, when I had occasion to name some large dried specimens for the Zanzibar Museum, I was much impressed by an extraordinary case of allo- 40 ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS metric growth of pereiopods II and III in senile males of Panulirus ornatus (Fabr.). Perhaps the females also exhibit the same phenomenon in old age, but the speci- mens at my disposal happened to be males. Two that are now in the B.M. collec- tion show the commencement of the differential rate of growth of pereiopods II and III, and a much later stage, respectively. In the smaller the c.l. is 134 mm. Cc b Fic. 9. Panulivus homarus (L.). a and a’. Right and left maxilliped 2 of specimen no. 7 in Table. Panulivus ornatus ? (Fabr.). 6. Right maxilliped of 9 from Shanghai (c./. = 76 mm., reg. no. 62.95). c. Same of 2 from Swatow (c./. = 72 mm., reg. no. 84.2). ON THE PUERULUS STAGE OF SOME SPINY LOBSTERS 41 and greatest c.b. is 85 mm. ; pereiopods I and IV are of equal length and probably scarcely reach beyond the distal end of the second segment of the antennal peduncle? ; pereiopod II exceeds I by rather more than the dactylus (32 mm.) ; pereiopod III exceeds I by the dactylus and some two-fifths of the propodus (55 mm.). The larger male with c./. 170 mm. and greatest c.b. 1330 mm. must have been about twice the live weight of the smaller one. Pereiopods I and IV are still equal and relatively short ; pereiopod II exceeds I by nearly all its propodus and the dactylus (about 115 mm.), while III exceeds I by some 186 mm. Perhaps all species of the genus Panulivus will prove to undergo the same striking increase in length of some of the pereiopods in old age. A well-developed exopodite, with long multiarticulate flagellum, on the second maxilliped is characteristic of some species, e.g., Panulirus polyphagus (Herbst)— unless, on re-examination of many specimens, the flagellum proves to be variable. This type of exopodite is presumably primitive, and there has been a tendency to reduction and loss of the flagellum within the genus. P. homarus would appear to be in process of losing the flagellum, which would explain the variation exhibited by the few specimens available. There are some dried specimens in the collection, but they are not suitable for a study of the mouthparts. Finally, to return to the three specimens determined as P. dasypus by Dr. Calman, I am rather at a loss to know what to call them. They comprise two females from Shanghai and Swatow, China (c./. 76 and 72 mm. respectively), and a young female from Palau Satang. In the juvenile the mouthparts are damaged ; the older specimens agree in having the exopodite of the second maxilliped shorter than the merus of the endopodite, and more slender than that of P. homarus (Fig. 9), ¢ ; c.f., Fig. 8c). The carapace is quite spiny and, in all three, abdominal grooves are present on somites 2 and 3, absent or evanescent on the posterior somites. There may be a median break or interruption, but the grooves are wider and shallower than in any of the homarus specimens examined, and filled with a fine short pubes- cence. When this is rubbed off one might overlook the grooves entirely and place the specimens in the subdivision without abdominal grooves. There is no trace of the crenulation which is always referred to in connection with P. dasypus and the antennular peduncle is relatively longer and more slender than in any of the speci- mens tabulated under P. homarus. These three specimens do not seem to agree with those referred to dasypus by Gruvel (1011, pl. 11, fig. 5), Monod & Petit (1929, p- 277, fig. 3 A-E) and others, and which appear to be the B-form of the variable P. homarus discussed above. Young specimens of P. versicolor have a broad shallow groove, filled with pubescence and with a median interruption, on somites 2 and 3, and just a trace of the setae on each side of somite 4 (see p. 28). These three specimens may belong to another species with abdominal grooves absent, as a tule, in adults—such as P. ornatus (Fabr.). Incidentally, there are records of P. ornatus from Swatow and Poeloe Satang, which is doubtless the same as Palau Satang—see Holthuis, 1946, p. 141. P. ornatus attains a very large size and might retain a juvenile character like these very shallow grooves for a considerable time. 1 The pereiopods are all directed forwards and close together, while the antennae are bent back, as convenient for storing dried material, so that exact comparison is not possible. “5 40684 MAK 1999 42 ON THE PUERULUS stheaion SOME SPINY LOBSTERS These specimens are faded, but in the larger ones there are traces of the broad band of colour present on each abdominal somite of large specimens of P. ornatus. REFERENCES BonvDE, W. VON. 1932. Post-brephalus development of some South African Macrura. Rep. Fish. Mar. Biol. Surv. S. Africa, 8, Special Rept. no. 1, 1930 [1932] : I-42, 14 pls. Bouvier, E. L. 1913. Le Stade ‘“‘ Natant’’ ou ‘‘ Puerulus”’ des Palinuridés. Tvans. 2nd Intern. Congress Entomology, Oxford, 1912 [1913], 2 : 78-89. 1914. Recherches sur le développement post-embryonnaire de la Langouste commune (Palinurus vulgaris). J. Mar. Biol. Ass. Plymouth, n.s. 10 (2) : 179-193, 6 text-figs. — 1917. Crustacés décapodes (Macroures marcheurs) provenant des campagnes des yachts “Hirondelle”’ et ‘‘ Princesse-Alice’’ (1885-1915). Rés. camp. sct. Monaco, Paris, 50: 1-140, 11 plates. — 1925. Reports of the Results of Dredging ... by the ‘‘ Blake.” XLVIII: Les Macroures Marcheurs. Mem. Mus. Comp. Zool. Harvard, 47 (5) : 401-472, 11 plates. Carman, W. T. tI909. The genus Puevulus Ortmann, and the post-larval development of the Spiny Lobsters (Palinuridae). Ann. Mag. Nat. Hist. (8) 3 : 441-446. Gitcurist, J. D. F. A post-puerulus stage of Jasus lalandii (Milne Edw.), Ortmann. /. Linn. Soc. Lond. Zool. 34 : 189-201, pls. 15 and 16, 13 text-figs. GruveL, A. to911. Contribution a l'étude générale systématique et économique des Pali- nuridae. Ann. Inst. océan. Monaco, Paris, 3 (4) : 5-56, 6 pls. Gurney, R. 1942. Larvae of Decapod Crustacea. Ray Society, vol. 129 (for the year 1941), pp. viii + 306, 122 text-figs. London. Haan, W. DE. 1833-50. Crustacea. [In P. F. de Siebold’s Fauna Japonica] : 1-244, pls. 1-55 + A-O. Hottuuis, L. B. 1946. The Decapoda Macrura of the “ Snellius’’ Expedition: I. Tem- minckia, 7 : 1-178, pls. i-xi. Kuso, I. 1950. On two forms of Puerulus found in Japanese waters, with special reference to that of Panulirus versicola (Latr.). Bull. Japan Soc. sci. Fish., 16 (3) : 91-98, 4 text- figs., 2 tables. Man, J. G. DE. 1916. The Decapoda of the ‘“‘ Siboga’’ Expedition. Part III. Families Eryonidae, Palinuridae, Scyllaridae and Nephropsidae. Siboga-Expeditie. Monographie, Xxxix,”, Leiden, 122 pp., 4 pls. Monop, Tu., & Petit, G. 1929. Contribution a l’étude de la Faune de Madagascar. Crus- tacea. II. Palinuridae. Faune des Colonies Frangaises, 3 (4) : 269-291, 9 text-figs. Nakazawa, K. 1917. On the metamorphosis of Panulivus japonicus. Dobuts Z. Tokyo, 29 : 259-267, I pl. Ortmann, A. 1891. Die Decapoden-Krebse des Strassburger Museums ... III. Die Abtheilungen der Reptantia Boas: Homaridea, Loricata und Thalassinidea. Zool Jahrb., Jena, Abth. f. Syst. 6 : 1-58, 1 pl. Pocock, R. I. 1890. Crustacea. [In Notes on the Zoology of Fernando Noronha, by H. N. Ridley.] J. Linn. Soc. Lond. Zool., 20 : pp. 506-526. 1903. The Decapoda of Sokotra. [In The Natural History of Sokotra and Abd-el-Kuri, edited by Henry O. Forbes. Special Bull. Liverpool Mus. : 212-215, 1 text-fig.] RicuTers, F. 1873. Die Phyllosomen. Ein Beitrag zur Entwicklingsgeschichte der Lori- caten. Zs. wiss. Zool. Leipzig, 23 (4): 623-646, pls. 31-34. Scumitt, W. L. 1926. The Macruran, Anomuran, and Stomatopod Crustaceans collected by the American Museum Congo Expedition, 1909-1915. Bull. Amer. Mus. Nat. Hist., 53 (1): 1-67, 75 text-figs., 9 pls. 1935. Crustacea Macrura and Anomura of Porto Rico and the Virgin Islands. Sez. Surv. P. Rico (New York Acad. Sct.), 15 (2) : 125-227, 80 text-figs. PRINTED IN GORDON . Nae ) Jf s A e % “a "BULLETIN OF Maris gitPaa MUSEUM (NATURAL at m LONDON: 1953 See ete ' a ONT SIRPUS "A GENUS OF PIGMY CANCROID CRABS BY ISABELLA GORDON Pp. 43-65; 13 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 2 No. 3 LONDON : 1953 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, 1s issued in five series, corresponding to the Departments of the Museum, and an Historical Series. Parts appear at irregular intervals as they become yeady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 2, No. 3 of the Zoological Series. PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM ‘ Issued November 1953 Price Six Shillings ONUSIRPUS) cA GENUS OF PIGMY CANCROID CRABS By ISABELLA GORDON, D.Sc., Px.D. SYNOPSIS The paper contains an emended diagnosis of the genus Sivpus Gordon, a fuller description of the type species S. zaviquieyi from Spain, and a description of a new species, S. monodi, from West Africa. Both species are of very small size, though some of the specimens are sexually mature. The affinities of the genus are discussed ; while it bears some resemblance to Trachy- carcinus Faxon, it seems to be much nearer to Pivimela Leach and is referred provisionally to the family Pirimelidae. INTRODUCTION In September, 1947, I received from Dr. Ricardo Zariquiey Alvarez of Barcelona two tiny specimens of what appeared to be a species of Spider Crab from Cadaqués on the north-east coast of Spain. Although the specimens were minute it was obvious that they were females, but I did not at that time observe that the larger one has the genital openings perforate and filled with what appear to be fertilization plugs. Later Dr. Zariquiey sent me a somewhat larger ovigerous female and two males and Dr. H. Boschma of the Leiden Museum lent me a male specimen, all from the same locality. For a time I was much puzzled by this pigmy species of Brachyuran ; while at first sight it seemed to belong to the Oxyrhyncha or Spider Crabs, it proved on closer examination to belong to the Cancroid assemblage, in the wide sense employed by Rathbun, 1930. But its position in that assemblage was not easy to determine, so I consulted Dr. H. Balss, formerly of the Munich Museum, and Dr. Th. Monod of the Institut Francais d’Afrique Noire, Dakar. Shortly afterwards Dr. Monod wrote to say that he was sending me some specimens from West Africa that he had found among his undetermined material and which seemed to belong to the same genus. They proved to represent another new species which is described in this paper. Pending the arrival of Monod’s specimens I wrote a preliminary paper describing the Mediterranean species under the name of Sirpus zaviquieyi n.g. and sp., and discussed very briefly the affinities of the genus. Now that a second species is available I am able to give an emended diagnosis of the genus; more figures, a fuller description and details as to habitat of the type species, and a detailed account of my reasons for referring the genus to the family Piri- melidae are also included. Genus Sirpus Gordon Gordon, 1953 , Pp. 304. DraGnosis (emended). Carapace hexagonal, convex, areolate, the lobes bearing granular clusters or nodules. Front rather narrow, produced, trispinose, the spines or teeth diverging from a common base and thus not in one plane. Orbits large, ZOOL. Il, 3. 4] 44 ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS with forward aspect, shallow and imperfect posteroventrally ; two gaps in dorsal margin, a wide one below and one at the inner angle. Eyestalks long, cylindrical, with a bulbous base, retractile within the shallow orbits; eyes deeply pigmented and protected by the postocular or first anterolateral spine. Anterolateral and posterolateral margins about equal, the former with four spines or teeth, the third of which is the smallest. Antennal flagellum long and setose; basal segment (2 + 3) of peduncle but slightly enlarged, not quite filling hiatus at inner angle of orbit but attaining the front, subcylindrical and armed at outer distal angle ; seg- ments 4 and 5 subequal. Antennules folded rather obliquely. Anterior margin of buccal cavity distinct. Sternum long and rather narrow. Abdomen rather narrow in both sexes and terminal somite triangular, longer than wide ; all seven somites free in female, three to five coalesced in male. No appreciable gap between external maxillipeds ; a rounded lobule at antero-internal angle of merus and antero- external angle rounded ; antero-internal angle of ischium not produced. Chelipeds equal, carpus and palm of chela nodular and/or granulose. Walking legs rather short, one longitudinal dorsal ridge on carpus, two on propodus ; dactyli not com- pressed but longitudinally ridged, longer than propodus. Male openings coxal ; first pleopod a long slender style, about four times as long as the second. As I explained in the preliminary paper, I have used the less usual spelling of scirpus in deference to the wishes of my botanical colleagues because Scirpus is a large and well-known genus or plants. According to Lewis and Short, A Latin Dictionary, Oxford, 1917, scirpus, sometimes sirpus, has two meanings: (i) a rush or bulrush, and (ii) by transference “ deriving the idea of intricacy from plaited work of rushes, a riddle, enigma,’ and I use it in the latter sense. The gender is masculine. Two species are known at present, the type species Sirpus zariquieyi Gordon from Cadaqués, and Sivypus monodi n. sp. from Dakar. The affinities of the genus appear to be with Pivimela and Trachycarcinus, perhaps nearer to the former— see p. 59-63. Sirpus zariquieyi Gordon Gordon, 1953, P. 305- D1aGnosis as for the genus. In this species there is considerable sexual dimorphism, the carapace in the female being distinctly wider and more ovoid than in the male (cf. Figs. ra, 4 and 5). MateriAL. From Dr. Zariquiey: (a) Two small females collected from fisher- men’s nets while being cleaned ashore, but from 20 to 40 metres in the neighbour- hood of Cadaqués, Provincia de Gerona, Spain. Summer, 1947. (b) One ovigerous female (the holotype) and two small males from Cystosetra growth on the rusty iron plates of a sunken ship, in about 20 metres, off Cadaqués, vili. 51. . From Leiden Museum: (c) One small male, from a fishing net and caught in the vicinity of Cadaqués, 4th-16th August, 1949. ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS 45 Dr. Zariquiey has very kindly presented the holotype and one male specimen to the British Museum Collection. DESCRIPTION OF OVIGEROUS FEMALE. This specimen, the largest at my disposal, measures almost exactly 5 mm. from posterior margin of carapace to apex of rostrum and 7-I mm. in greatest width (between the last pair of anterolateral spines). Fic. 1. Sivpus zaviquieyi Gordon. A. Holotype, an ovigerous female, in dorsal aspect, pereiopods and right antenna omitted. sB. Left chela and carpus of holotype. 1. Supraorbital spine. 2. Infraorbital spine. The form of the carapace is represented in dorsal aspect in Fig. 1A. The front consists of three subequal spines ; the median one, or rostrum, is directed obliquely downwards, while the lateral ones are directed forwards and are slightly upturned apically. The orbits are wide, so that the distance between the postocular spines (first pair of anterolateral spines, Fig. 14) exceeds two-thirds of the maximum cara- 46 ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS pace width. Each orbit is well formed anteriorly but is shallow and incomplete ventro-posteriorly, so that the retracted eye is protected only by the postocular spine. The supraorbital and infraorbital angles are each produced to form a short triangular spine (Fig. ra, 7 and 2 respectively). The dorsal orbital margin slopes obliquely backwards and outwards to meet the base of a wide triangular spine inter- calated between it and the postocular spine. Behind the orbit the anterolateral margin comprises three large spinose projections, the anterior of which is the largest. The posterolateral margins are slightly concave and convergent; the posterior margin bears a high submarginal crest, and is only a trifle wider than the trispinose front. The dorsal surface of the carapace is distinctly areolated as represented in Fig. 1A, and the apex of each lobe is beset with some granules. The gastric region has three lobes, two protogastric and one mesogastric—the latter followed by a cardiac and a small intestinal lobe. On either side of the cardiac lobe is a dark reniform depression. The branchial region bears a median lobe (slightly behind and external to the mesogastric lobe), a triangular lobule near the base of the posterior antero- lateral spine, and two spinules and a few granular rugosities near the posterolateral margin (Fig. 14). A number of rather long plumose setae are present as indicated in the figure, but others may have been rubbed off. The anterior portion of the carapace is represented in profile in Fig. 2B. The antennae are long and setose ; the ventral spine on the outer distal angle of segment 2 + 3 of the peduncle is visible in dorsal aspect ; the fourth segment is a trifle longer than the fifth (Fig. ra, see also Fig. 2A). The flagellum is a trifle more robust than I have indicated in the figure and comprises some 14-16 segments, some of which bear long, distally plumose setae. The eyestalk is much swollen basally, cylindrical in the distal two-thirds, the corneal area is terminal. The chelipeds are equal; the distal segments of the left one are represented in Fig. rp. As the carapace is wider than in the male, the merus reaches the apex of the third anterolateral spine and only a small portion is visible dorsally—cf. Fig. 4 where more of the merus is visible beyond the anterolateral margin. The spine on the inner border of the carpus is much longer than in the male (which is of smaller size) ; on the anterior margin a blunt lobule articulates with a similar one on the chela; the distal half of the outer margin is cristate or lobulate and beset with granules ; on the dorsal surface are two inner, and two or three outer clusters of granules. The fingers of the chela are rather longer than the dorsal margin of the palm ; each bears a median crest or ridge; the cutting edges are in close contact distally, leaving a slight gap proximally. The dorsal margin of the palm bears a large distal and a smaller median granular prominence, and there are two similar prominences a short distance from the actual margin. Below these is a median series of low granular clusters. The walking legs or pereiopods II-V are for the most part bent inwards round the mass of ova and I did not attempt to straighten them out, so have omitted them from the figure. The first pair are about as long as the chelipeds. The right pereiopod V is represented in Fig. 38; the merus is about two and a half times as long as wide, the carpus bears a low dorsal ridge near which ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS 47 is a distal depression. The propodus, which is as long as the carpus, has two parallel dorsal ridges ; the dactylus is half as long again as the propodus and bears longi- tudinal ridges ; in none is it compressed. I]mm. Fic. 2. Tvrachycarcinus glaucus Alc. & And. a. Abdominal somites 3 to 7 of male syntype (c./. = 16-1, c.b. = 14:8 mm.) Sirpus zariquieyi Gordon. B. Left anterior portion of carapace of larger paratype from Lot a, in lateral aspect, to show divergence of the frontal spines. c. Abdominal somites 3 to 7 of same specimen. pb. Abdominal somites 3 to 7 of male paratype. I. Lateral frontal spine. 2. Median frontal spine. 3. Infraorbital spine. 4. Supra- orbital spine. 5. First and second anterolateral spines. ZOOL. II, 3. 3§ 48 ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS The antennules are enclosed in deeper, more developed sockets than in the male represented in Fig. 3A and the crest on the anterior margin of the basal segment of the peduncle is more pronounced. The separation between the base of the rostrum and the epistone is also more pronounced in the female, and the median notch on the anterior margin of the buccal cavity is less distinct. The longitudinal furrow on the ischium of maxilliped 3 is also more pronounced than in the male. Fic. 3. Sivpus zaviquieyi Gordon. A. Right anterior portion of carapace of larger male paratype from Lot d, in ventral aspect, to show details of orbit, epistome, buccal cavity, etc. 8B. Right pereiopod V of holotype. 1. Supraorbital spine. 2. Infraorbital spine. All seven segments of the abdomen are free ; the first segment is very short and has a transverse granular ridge ; the next three are progressively longer while the last three are of equal length, the terminal one being narrowly triangular with a rounded apex. The abdomen is comparatively narrow, since at its widest it scarcely exceeds the width of the posterior border of the carapace. Somites 3 to 7 are repre- sented in Fig. 2c. The ova are numerous, and measure approximately 0-25 to 0-30 mm. in diameter. Remarks. The larger female from Lot a measures 4:85 by 5:52 mm. and is very similar to the holotype. It appears to be sexually mature, for the genital openings are distinct and seem to be filled with dark brown fertilization plugs. The clusters of granules on the carapace are more pronounced, and from them arise ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS 49 groups of distally plumose setae. The swollen basal part of the eyestalk is also closely beset with small granules and is separated from the distal portion by a distinct depression. The smaller female from Lot @ measures 4:00 by 4:81 mm. and is | mm. Fic. 4. Sivpus zaviquieyi Gordon. The larger male paratype from Lot }, in dorsal aspect, left antenna and pereiopods II to V omitted. sexually immature ; it also is markedly granulose, and the eye-stalk has a depression distal to the bulbous base. The male paratype from Lot b represented in Fig. 4 measures 4:60 by 4-72 mm. It differs appreciably in outline from the female, since the carapace is relatively 50 ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS narrow and the anterolateral spines project but a short distance beyond the orbit. The lateral frontal spines as well as the anterolateral ones are longer and more slender and the intercalated lobe on the orbital margin is less pronounced. The areolation and the granulation of the dorsal surface of the carapace are also less conspicuous. The chelipeds are equal, and the merus projects some distance beyond the anterolateral spines (cf. Figs. 4 and 1A). The spine at the inner angle of the carpus is small; the two pairs of granulose humps on the dorsal border of the palm Fic. 5. Sivpus zaviquieyi Gordon. Carapace of male paratype from Lot c, in dorsal aspect. tend to merge, and the fingers are, as in the female, rather longer than the dorsal palmar margin. The walking legs are more slender than in the female ; the merus of pereiopod V is rather more than three times as long as wide, and the ridges on carpus, propodus and dactylus are less distinct. The details of epistome, buccal cavity, orbit, external maxilliped, etc., are shown in Fig. 34. The abdomen is narrowly triangular ; somites 3 to 5 are coalesced, and somite 7 is rather longer than the preceding one (Fig. 2p). The first pleopod, represented in Fig. 6B, is a long slender style reaching almost to the apex of the abdomen ; the second pleopod is short, about one-fourth of the first. ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS 51 The specimen received on loan from the Leiden Museum is also a male, and though measuring only 3°42 by 3°58 mm. also has the first pleopod well formed and nearly as long as the abdomen. It differs from the larger male chiefly in that the penulti- mate anterolateral spine and the intercalated spine of the orbital border are each only incipient (cf. Figs. 5 and 4). The separation of basal and distal portions of the eyestalk is also more marked by the depression mentioned as present in the females from Lot a. This specimen has, in addition to the two dark posterior reni- Fic. 6. Tvachycarcinus glaucus Alc. & And. A. First and second pleopods of male syntype. Sirpus zaviquieyi Gordon. B. First and second pleopods of largest male paratype. form areas, a similar dark patch anteriorly near each orbit. The well-formed pleo- pods of all these minute males suggests that they may be capable of impregnating females, though they are rather young. Hapitat. Dr. Zariquiey informs me that the specimens from Lot b ‘‘ were captured a short distance from the coast, at a depth of about twenty metres, amongst long Cystoseiva algae growing in great quantity on the rusty iron plates of the hulk of a sunken vessel. On examining at home the pieces of rust and algae, I found the 52 ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS Sirpus at the base of the said algae and on the pieces of rusty iron among other small algae (not determined). In the same place I found hundreds of Catapagu- voides timidus (Roux), Thoralus cranchii (Leach), Galathea bolivari Zariquiey, Por- cellana longicornis (Pennant), Alpheus dentipes Guérin and Synalpheus laevimanus (Heller).”’ . . . ‘‘ The exact point is next to Punta Oliguera facing a sector of the coast called Los Cayals, in the vicinity of Cadaqués (South of Cabo de Creus), Provincia de Gerona, Spain.” Sirpus monodi n. sp. MaTERIAL. (a) Collected by M. R. Sourie, at Dakar, 7.111.1950. See also under habitat p. 57. One ovigerous female, the holotype (c./. = 7 mm., c.b. = 7°3 mm.). (5) Collected by MM. Th. Monod and P. Budker on 10.1.41 near Dakar, dredged by the “ Cabellou,”’ station 6 bis (“ 4 l’Est de la bouée du Banc de Bel Air, 14° 42’ N., 17° 24’ W.), in about 12m. Bottom “ Sable grossier.”” One male (c./. = 4:9 mm., cio) — 57mm) (c) Collected by M. Paraiso at Gorée, near Dakar, on 1.ix.50. Entangled in net used for fishing ‘“‘ langoustes ”’ (Panulirus vegius) and set in a few metres depth only. One female with genital openings long narrow slits and so probably slightly immature (e/5—3-Qumm, c0.—.3Onmms) DESCRIPTION OF HOLOTYPE. The carapace, represented in dorsal aspect in Fig. 7A, is only a trifle wider than long ; the dorsal surface is markedly nodular in such a manner as to emphasize the areolation. The general outline is hexagonal, with the anterior and posterior margins shorter than the other sides (the orbit being con- sidered along with the anterolateral margin as one side of the hexagon). The front comprises three subequal, bluntly triangular spines, the median one situated below the level of the lateral ones and directed obliquely forwards and downwards. The orbit is oblique, rather shallow but well formed anteriorly, imperfect postero- ventrally as shown in Fig. 8A, so that the retracted eye is protected only by the postocular spine and the corneal area is visible both dorsally and ventrally. The main lobe of the dorsal orbital margin terminates anteriorly in a blunt supraorbital spine or tooth, and is separated by a narrow hiatus from the blunt lobe intercalated between it and the postocular (or first anterolateral) spine. The ventral margin of the orbit is formed partly by segments 2 + 3 of the antennal peduncle and partly by the lobe bearing the infraorbital spine (Fig. 8A, 2), behind which it gradually fades out. The anterolateral margin is divided into four spines, of which the third is the smallest and the anterior is the longest ; the last, though short, is much wider than the third. The posterolateral margins are straight or even a trifle convex, and somewhat convergent. The submarginal crest parallel to the posterior margin is low and interrupted medially. The two protogastric lobes bear each a blunt, forwardly directed spine or tooth, while the summit of the mesogastric lobe has a prominent wide cluster of granules. On the summit of the cardiac lobe is a pair of granular clusters. A high blunt hump, also beset with granules, is situated near the base of the last anterolateral spine. On either side of the shallow depression separating mesogastric and cardiac ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS 53 lobes is a prominent branchial lobe, with a prominent granular hump on its summit and a smaller one some distance behind this. In addition, there are numerous granular patches or striae on the posterior half of the carapace. Imm. Fic. 7. Sirpus monodi n. sp. Holotype, an ovigerous female. a. Carapace and left cheliped, in dorsal aspect. 8. Carpus and chela of right cheliped. The antennae are long and setose, but the setae are simple, not pluomse. The blunt spinose anteroventral angle of segments 2 + 3 of the peduncle is visible in dorsal aspect (Figs. 7A and 8a) ; segments 4 and 5 are of equal length, but the former is the more robust. 54 ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS The eyestalk is swollen and bulbous proximally, slender and cylindrical throughout most of its length ; the distinctly pigmented corneal area is rather small and ter- minal. | mm. Fic. 8. Siypus monodi n. sp. Holotype. a. Right anterior portion of carapace, in ventral aspect, to show details of orbit, epistome, buccal cavity, etc. 3B. Pereiopod V and distal segments of pereiopod IV. z. Supraorbital spine. 2. Infra-orbital spine. ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS 55 The chelipeds are equal ; the right, which is detached from the specimen, is repre- sented at a higher magnification in Fig. 7B. Part of the merus is visible in dorsal aspect beyond the two posterior anterolateral spines. The carpus and palm of the chela are both conspicuously nodular ; the fingers are rather longer than the dorsal margin of the palm, and there is a proximal tuft of setae on the upper border of the dactylus, about one-third of the way along. When closed, the cutting edges of the fingers are in contact distally, leaving a narrow proximal gap ; the outer surface of each finger bears a low ridge. The maximum height of the palm is almost equal to the length of the dorsal margin. The spine on the inner border of the carpus is conspicuous. The walking legs, pereiopods II to V, are bent inwards round the large bunch of ova and so have been omitted from Fig. 7A. The distal segments of pereiopod IV are represented in Fig. 88; there is a low median dorsal ridge or carina on the carpus, two dorsal ridges on the propodus and several longitudinal ridges on the dactylus, which is not compressed. Pereiopod V is shown also in Fig. 8B; the metus is not quite three times as long as wide, the dorsal ridge on the carpus is more prominent distally, the propodus is more compressed and the two dorsal ridges are very close together. The antennules are set in deep sockets beneath the front as represented in Fig. 8a; they are folded obliquely, and there is a conspicuous crest or carina on the swollen basal segment of the peduncle. The epzstome sends forwards a narrow prominent median carina to meet the backward projection of the front. The anterior margin of the buccal cavity is distinctly marked by a ridge, as shown in Fig. 8a. All seven somites of the narrow abdomen are free; the maximum width across somites 3 and 4 scarcely exceeds the width of the posterior margin of the carapace. Somites 5 and 7 are equal, and each a trifle longer than somite 6 ; the terminal one is narrowly triangular. The ova are numerous and small (a trifle larger than in the type species). THE MALE. The single male specimen is represented in dorsal aspect in Fig. 9. Most of the pereiopods are detached from the left side, while the last one, pereiopod V, is missing on the right side. The carapace is very similar in outline to that of the female ; the frontal and anterolateral spines are all rather blunt and massive and the supraocular lobe is broad and rounded. As the specimen is younger than the holotype the areolation of the carapace is less pronounced, but the nodular promi- nences on the principal lobes are distinct. The eyestalks are relatively more robust than those of the holotype. The crest near the posterior margin of the carapace is more distinct and continuous. The abdomen is rather damaged near its junction with the carapace, as Dr. Monod had dissected out the pleopods from one side and mounted them on a slide, so that my reconstruction of that part may not be quite correct (Fig. 9). Somites 3 to 5 are coalesced ; the distal somites are very similar to those of S. zariquieyi, represented in Fig. 2p. The first pleopod is slightly dis- torted in the micropreparation, but is a long slender style reaching nearly to the apex of the abdomen. It is very similar to that of S. zaviquieyi, as is the second pleopod. The chelipeds are equal, rather more massive than those of the holotype, 56 ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS although as yet the nodular condition is less highly developed on carpus and palm ; the cutting edges of the fingers each bear six well-formed lobes or teeth, the proximal ones being more pronounced than in the older female. REMARKS. The rather immature female is also very similar to the holotype. Fic. 9. Sivpus monodi n. sp. Male paratype, in dorsal aspect, pereiopods on left side omitted. This species undoubtedly belongs to the genus Sivpus and, like the type species, it reaches sexual maturity at a remarkably small size. S. monodi differs from S. zariquieyi chiefly in the following respects : (i) It is a much more nodular form, with shorter and much more massive spines on front, orbits and anterolateral borders. (ii) There is no marked difference between the sexes as regards the shape of the carapace. (iii) The chelipeds and walking legs are relatively more robust, and the latter rather shorter as well. The propodus of pereiopod V is rather more compressed. (iv) The setae on the antennal flagellum are simple and rather short, instead of long ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS 57 and distally plumose ; the segments of the peduncle are much more robust (cf. Figs. 8a and 3a). HasiratT. Dr. Monod informs me that the holotype was collected ‘A marée basse, 4 Dakar (Anse Bernard) au niveau de |’affleurement de tufs volcaniques (la méme station a fourni des Perimela' denticulata). M. Sourie me dit que le petit crabe, trés homochrome, étais trés difficile 4 apercevoir sur le sable.” Summing up for all three localities he writes: “‘l’Espéce semble donc, pour le moment étre littorale, depuis la zone intercotidale jusqu’a 10-15 m.” (letter dated 11.iii.53). The holotype will be deposited in the Paris Museum, the male paratype in the IFAN at Dakar, and the female paratype in the British Museum Collection. Since the MS. went to press, Dr. Monod has found an ovigerous female measuring 4°5 by 4:8 mm. amongst some undetermined material in the Paris Museum. The specimen was collected by A. Gruvel on 2.iv.08, at Pointe Cansado, C. Blanco, Morocco, much further north than the type material. It agrees well with the holotype but, being smaller, it has many more granular striations on the dorsal surface of the carapace. THE SYSTEMATIC POSITION OF THE GENUS When Dr. Zariquiey first sent me the two small female specimens comprising Lot a of S. zariquieyi he referred to them in his letter dated 16.ix.47 as “‘ dos Maiidae formas juveniles de sp. ?’’. On examining them my first impression was that they might be very young stages of Pirimela denticulata (Montagu) but this proved not to be the case, and I also was inclined to regard them as young stages of a species of Spider Crab. Now that an ovigerous female is available and I have examined them more carefully, I am convinced that they are not referable to any family of the large subtribe Oxyrhyncha. No Oxyrhynchous crab has the median frontal spine so well formed that the front looks trispinose; when present the median spine or rostrum is minute, and much more ventral in position than the well-de- veloped rostral (or pseudorostral) horns (e.g., in Maia or in Mithrax). Moreover, in the Oxyrhyncha it is unusual for the anterolateral and posterolateral borders of the carapace to be clearly separated, and the basal segment (or more correctly segments 2 + 3) of the antennal peduncle usually forms a larger part of the lower orbital border. The general form of the carapace in the ovigerous female is distinctly Cancroid, in the wide sense employed by Rathbun, 1930, apart from the front, the median spine of which slopes obliquely downwards and so lies in a different plane from the other two (Figs. 14 and 28). It would appear that my first impression— that they are closely related to Pivimela—was fairly sound. Dr. Balss, who has done so much to elucidate the classification of the Oxyrhyncha and to whom I sent some sketches, writes (6.x.52) : ‘‘ Das Bild einer jugendlichen Krabbe, das Sie mir ubersandt haben, gibt nicht eine Oxyrhynche Krabbe, sondern eine primitive Por- tunide (Carcinus maenas ? oder eine verwandte Art).” He was relying entirely on memory, being too ill to go to Munich to consult the literature. Dr. Monod, to * According to Monod (1933, Bull. Com. d’Etudes hist. sci. Afy. Occ. Fr. 15, 2-3: 52) Pivimela is etymologically incorrect. 58 ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS whom I also sent sketches and who has since found the specimens belonging to the second species S. monodi, writes (g.xii.52) : ‘‘ne devrait-on pas chercher du coté des Atelecyclidés, aux alentours de Tvachycarcinus par example?” Thus all three of us are agreed that the genus belongs to the Cancroid or Cyclometopous Crabs in the wide sense used by Rathbun, 1930. Carcinus maenas or Carcinides maenas (L.) is one of the few species of European brachyura the postlarval development of which is known in detail (see, e.g., Shen, 1935, p. 19, text-fig. 20), and the specimens from Cadaqués do not fit into this series. The nearest approach to these specimens that I have discovered so far is a Fic. 10. Two figures after Cano, at twice his magnification. a. ‘“ Maja verrucosa, stadio postlarvale "’ from 1893, taf. 34, fig. 29 (x 8). B. “‘ Fase adulta (giovane) di Pivimela”’ from 1891, 1892, tav. iv, fig. 4G (Cano’s magnification not specified). young crab stage referred to Maia verrucosa H. M.-Edw. by Cano (1893, Taf. 34, fig. 29), and which I have illustrated ( x 2) in Fig. 10a. The size of this specimen is approximately 4 x 4 mm., but I do not think that it can be the first, second or later young crab stage of M. verrucosa—Cano did not rear it from the megalopa, of course. Miss Lebour seems to have accepted Cano’s interpretation of his specimens, so it is perhaps presumptuous of me to question this stage. The development of Maia squinado (Herbst) has been described by Lebour (1927), who obtained the prezoea, first and second zoeae and megalopa from the eggs. But, as she states on p- 797, in no case did the megalopa turn intoacrab. The megalopa of Maia squinado ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS 59 from which she obtained the first three young crab stages was obtained from the plankton. What strikes one about these stages is their marked resemblance to young Portunids ; although the front is rather more pointed medially, the general resemblance of these young Maza to the corresponding stages of several species of Portunus and of Carcinides is unmistakable (cf. Lebour, 1927, Pl. III, fig. 10 with Lebour, 1928, Pl. VI, figs. 1-3 and pl. 7, figs. 1 and 3). The young crab stages of Macropodia and of Inachus, on the other hand, are unmistakably Oxyrhynchous (see Lebour, 1927, Pl. 3, figs.5 and 8). I cannot help wondering whether the mega- lopa from which these three young stages were obtained is really referable to the genus Maia. To return to Cano’s young “Maja verrucosa,” it resembles the specimens from Cadaqués as regards the front, eyestalks and orbits, and in the number of antero- lateral spines. If the supraorbital spines were omitted from Figs. 4 and 5 and the orbital margins were carried straight back from the bases of the lateral frontal spines, and the anterolateral spines on either side of the smallest one were exaggerated slightly, the result would be much as in Cano’s figure (Fig. 104). Bouvier (1942) has separated the crabs of “‘ la tribu des ‘ Corystoidea ’’’ from the rest of the Brachyrhyncha; to the Corystoidea he has referred the Euryalidae (= Corystidae restr.), Atelecyclidae and Cancridae of Rathbun (1930, pp. 10, 148 and 176) and others, placing the genera in five families (see pp. 47-48 for his scheme). The genus Trachycarcinus he places in the family Corystidae, subfamily Atele- cyclinae ; Pivimela in the monotypic family Pirimelidae. I had noted the rather striking resemblance between the males of Sivpus zariquieyi and Tvachycarcinus corallinus Faxon for example. In the British Museum Collection the genus is repre- sented by one male syntype of T. glaucus Alcock & Anderson and one male specimen of T. balssi Rathbun received from Dr. Sakai of Japan. Sivpus agrees with the diagnosis of Tvachycarcinus given by Rathbun (1930, p. 164) as regards the orbits, which are “ large, with forward aspect, imperfect,” but it differs in many respects : (i) The eyestalks are retractile within the orbits but they are not “‘ very small,” nor are the eyes “ dull and faintly pigmented.” (ii) The front is wider and, though trispinose, the spines do not lie in one plane as they do in Tvachycarcinus. (iii) The anterior margin of the buccal cavity is very distinct and the ridges of the endostome are less pronounced than in T. balssi for example. The external maxillipeds also differ in the two genera ; in Tvachycarcinus there is a considerable gap between the maxillipeds anteriorly, the anterior inner angle of the ischium is advanced and the merus is “ obliquely truncated without emargination at antero-internal angle ” (Rathbun, 1930, p. 164, pl. 72, fig. 6)). (iv) The carapace is not pentagonal with long, nearly straight, anterolateral margins, but hexagonal (more ovoid in the female of the type species S. zaviquieyi). (v) The antennules do not fold longitudinally or lengthwise as in the Atelecyclidae, but obliquely, as shown in Figs. 34 and 8a. (vi) The abdomen of the male is more narrowly triangular, with a longer apical somite and somites 3 to 5 completely coalesced (cf. Figs. 2D with 2A—somites 4 and 5 are imperfectly coalesced in Tvachycarcinus glaucus and in T. corallinus but all seven somites are free in T. spinulifer) ; and, judging from the figure given by Rathbun (1930, pl. 72, fig. 5), that of the female is also narrower, with a much longer apical ” 60 ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS somite, in Sivpus. (vil) The pleopods of the male also differ markedly in the two genera as represented in Fig. 6A and B; in Tvachycarcinus the first pleopod is rela- tively straighter, more robust and shorter, while the second pleopod is longer than the first and has a terminal lash. In all the genera of the Atelecyclinae that I have examined, namely Telmessus, Peltarion, Atelecyclus,! Trachycarcinus and Pliosoma, pleopod 2 of the male is as long as or longer than pleopod 1 except in Pliosoma ; in the latter pleopod 2 is very short, measuring only about one-sixth of pleopod 1, and the genus may be referable to a different subfamily. I should expect Evimacrus and Trichopeltarion to conform to the general pattern (i.e., to have pleopod 2 long), but I have no material at my disposal; Rathbun and Bouvier both omitted to A Fic. 11. Pivimela denticulata (Montagu). A young male from Jersey (width of carapace approx. 10-7 mm.). A. Part of carapace, in ventral aspect, to show details of orbit, epistome, buccal cavity and third maxilliped, etc. 8. First and second pleopods, with apex of each further enlarged, of young male (10-4 x I1-4 mm.). mention the form of the male pleopods. In the Corystidae (or Corystinae of Bouvier) pleopod 2 varies in relative length, being longer than pleopod 1 in Pseudocorystes and Podocactes, about the same length as pleopod 1 in Gomeza and Jonas, shorter than, but exceeding half the length of, pleopod 1 in Corystes. In the family Bellidae of Bouvier’s scheme (Acanthocyclinae in that of Rathbun) pleopod 2 reaches nearly to the apex of pleopod 1 in Bellia, but is rather less than half as long as pleopod 1 in Acanthocyclus. In the genus Cancer pleopod 2 is rather longer than pleopod 1— 1 Stephensen (1945, p. 222) is in error when he says of the subfamily Atelecyclinae, “‘ Plp. 2 short, of usual shape’; Brocchi, to whom he refers, says on p. 103 that pleopod 2 of Atelecyclus cruentatus Desm, is long, and his fig. 199 shows a long slender appendage which exceeds pleopod 1 (figs. 197-8). ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS 61 family Cancridae of Bouvier’s scheme—whereas it is short in Thia, Kraussia (family Thiidae) and in Pirimela (family Pirimelidae) (see Fig. IIB). Fig. 12 represents a small male specimen of Pirimela denticulata (Montagu) from Jersey in dorsal aspect. The areolation of the carapace is very similar to that found in both species of Sivpus. The front is trilobed, the median one being the most prominent, and all lie in the same plane. The antennae are rather long but as a tule are not setose, though a few very short plumose setae may be present in some Fic. 12. Pivimela denticulata (Montagu). Young male, in dorsal aspect, left pereiopods omitted. specimens. The orbit is closed and narrow, but the dorsal border comprises three lobes, separated by two notches ; the eystalks are rather short, but the basal portion is bulbous (Fig. 11a). The anterolateral margin bears five forwardly directed spinose lobes. Pirimela denticulata grows to a larger size and is far less precocious than either of the species of Sirpus—the first trace of the genital opening appears when the females measure 9 to Io mm. in maximum width, and the opening is not fully formed until they measure 11 to 12 mm. across the last pair of anterolateral spines. It is not so easy to ascertain when the males reach sexual maturity ; in a 62 ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS small specimen measuring 7 mm. in width the first pair of pleopods is short and obviously immature, but by the time the carapace measures 8-5 to 9 mm. in width these pleopods are well-developed long styles. Sivpus appears more closely related to Pirimela than to Tvachycarcinus in the following respects: (i) The manner in vis Q Fic. 13. A. Portumnus biguttatus Risso. 3B. Portumnus latipes (Pennant). Anterior portion of carapace, in ventral aspect, to show details of front, orbit and eyestalk, antenna and antennules, epistome and anterior margin of buccal cavity. which the antennules are folded, the distinct anterior margin to the buccal cavity, the well-pigmented eyes and the form of the external maxilliped (cf. Figs. 34 and 8a with rra). (ii) The form of the abdomen in both sexes. (iti) The form and rela- tive lengths of the pleopods of the male (cf. Figs. 68 and 118). (iv) The lobulation of the carapace—the female of S. zaviquieyi also approaches more nearly to Pirimela in general form of the carapace (cf. Figs. ra, 7A and 12). (v) The walking legs are also very similar. At present we do not know how large either of the species of ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS 63 Sirpus may become, but they would appear to be really of very small size, since even in the smallest males the pleopods are already well formed, and in all but the smallest female the genital openings are visible. It is unlikely that the females would alter appreciably with increase in size, since in each species the largest bears ova. Cano (1891, tav. iv, fig. 4¢—published in 1892) gives a figure of what he calls the “fasa adulta (giovane) di Pivimela,’’ which I have reproduced at twice his magnifi- cation in Fig. 108. As Cano does not appear to give his magnification I do not know the size of his specimen. It differs appreciably from all the stages that I have examined chiefly in having wide shallow orbits, long eyestalks and only four antero- lateral spines—in these respects his specimen recalls S. zaviquieyi even more than Pirimela, although the front is nearer to that of the latter (Cano’s specimen is repre- sented in his figure as rather asymmetrical). At present the postlarval develop- ment of Pirimela denticulata is not known ; it seems unlikely that Cano’s specimen is a young crab stage of Pirimela, but if it is the orbit must alter markedly with age and the number of anterolateral spines increase to five (cf. Figs. 12 and 104). The wide shallow orbit alone would not necessarily exclude these new species from Cadaqués and West Africa from the genus Pivimela. The two European species of the genus Portwmnus differ from each other markedly as regards orbit and eyestalk as represented in Fig. 134 and B. In Portumnus biguttatus! Risso the orbit is closed, narrow, and deep enough to conceal the cornea of the retracted eye dor- sally, though not ventrally (owing to the deep gap or hiatus in the ventral margin). The trilobed front is separated by a wide shallow emargination from the low rounded supraorbital lobe. The principal segment (2 + 3) of the antennal peduncle fills the gap between the orbit and the antennular fossa, and, in the specimen from which the figure was made, is slightly movable. In Portumnus latipes (Pennant) the front is more markedly three-pronged ; the orbit is wide and shallow, affording no con- cealment for the cornea of the retracted eye ; moreover, there is a wide gap between the infraorbital spine or angle and the antennal peduncle (Fig. 138). Sivpus and Pivimela differ from each other as regards orbit and eyestalk in much the same way as do these two Portumnus species. However, Sirpus differs sufficiently from Piri- mela as regards the form of the front, the number and form of the anterolateral spines, the more setose antennae, and in the absence of a ‘‘ lobe portunien ”’ on the endopodite of the first maxilliped (see Bouvier, 1942, p. 32, fig. 15), to justify the erection of a new genus. But these differences hardly seem sufficient to justify a new family for the reception of Sirpus. Authors do not seem to agree as to the systematic position of the genus Pivimela. Pesta (1918, pp. 386-387) places it with Cancer and Carcinides in the family Can- cridae; most authors place Carcinides in the Portunidae. Nobre (1931, p. 84; 1936, pp. 48-49) refers it to the Cancridae along with Cancer and Xantho! In the Plymouth Marine Fauna, 2nd edition, 1931 the family Cancridae includes Cancer and Atelecyclus, while Pirimela is referred to the monotypic family Pirimelidae. Bouvier (1942, pp. 47-48) includes the Cancridae and the Pirimelidae, each with only the genus from which the family name is derived, in his Corystoidea, which he removes from the Brachyrhyncha (and yet he refers to Pivimela as having a 1 Referred to the genus Portumnoides by Bohn 1902, p. 448. 64 ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS narrow “‘ lobe portunien ’’ as mentioned above). Lebour (1944, p. 15) writes: “ As far as we know from its larval stages Pirimela is probably related to Portunus and Cancer, perhaps more nearly to Cancer.’’ Enough has been said to show that the classification of the Brachyura is in need of thorough revision. In the meantime Iam inclined to place Sirvpus near to Pivimela rather than to Tvachycarcinus, although it shows some resemblance to the latter. In the shape of the abdomen in both sexes and also of the male pleopod 1 it is much nearer to Pirimela than to Carcinides and the dactylus of pereiopod V is not compressed as in the latter genus, which is a primitive Portunid (see p. 57 for Dr. Balss’ view). The two genera of the family Pirimelidae may be distinguished as follows : A. Front trilobed, the lobes in one plane; orbits narrow and complete; antero- lateral margin divided into five teeth; antennal flagellum not obviously setose ; specimens reaching a fair size (up to 25-30 mm. across carapace) Pirimela, Leach. B. Front trispinose, the spines not in one plane ; orbits wide and incomplete ventro- posteriorly ; anterolateral margin divided into four teeth or spines of which the third is the smallest ; antennal flagellum setose ; specimens of very small Sizeame c 5 ¢ - ° - : - : . Sirpus, Gordon The study of these small crabs was particularly difficult, I think, because of their extreme sexual precocity, as a result of which they have inevitably retained a number of juvenile characters. For example, the great fronto-orbital width, the projecting front, the relatively large frontal, orbital and anterolateral spines and the long setose antennae are all strongly reminiscent of the first postlarval stage of Cancer pagurus, figured by Cunningham (1898, Proc. zool. Soc. London, p. 204, pl. xxi, fig. 1). Shen has figured the outline of the carapace of the first nine young crab stages of Carcinus maenas (1935, P- 19, fig. 20), so that one can see at a glance how the shape and relative proportions vary with age. As regards size, the largest ovigerous females of the genus Sivpus would be equivalent to the sixth young crab stage of Carcinus and sexual maturity is reached even at a smaller size. The two species of the genus Sirpus may therefore be regarded as neotenous. de Beer (1951, ‘Embryos and Ancestors,’ revised edition, p. 52), applies the term neoteny to “cases where the adult animal retains larval characters’’; where a complicated metamorphosis occurs, however, it is more likely to be the postlarval characters that are retained by the adult. Acknowledgments. I wish to thank Dr. Zariquiey of Barcelona, Dr. Boschma of Leiden and Dr. Monod of Dakar for sending me these very interesting specimens for study. I am also grateful to Dr. Balss of Munich and Dr. Monod for giving me their views on the relationship of the genus, and to my colleague Mr. A. C. Townsend for assistance in finding a name for it. The authorities of the British Museum (Nat. Hist.) are much indebted to Dr. Zariquiey for presenting to their Collection the holotype and a male paratype of S. zariquieyi, and to Dr. Monod for presenting the female paratype of S. monodt. ON SIRPUS, A GENUS OF PIGMY CANCROID CRABS 65 LIST OF WORKS CITED Boun, Gk Des mécanismes respiratoires chez les Crustacés Décapodes. Bull. sci. Fr. Belg., 36: 178-451, 209 text-figs. Bouvier, E. L. 1942. Les Crabes de la Tribu des ‘“‘ Corystoidea.’”’ Mém. Acad. Sci. Inst. France, 65, 1941 [1942]: 1-52, 18 text-figs. Cano, G. 1891 [1892]. Sviluppo postembryonale dei Cancridi. Boll. Soc. ent. ital. 23: 146-158, pls. 3 and 4. 1893. Sviluppo e Morfologia degli Oxyrhynchi. Mitt. zool. Sta. Neapel. 10 (4) : 527- 583, pls. Gorpon, I. 1953. Ona new Crab from Cadaqués on the north-east Coast of Spain (Sirpus zarviquieyt n.g. and sp.). Eos, Madrid, 28 (4): 308-314, 5_text-figs. Lesour, M. V. 1927. Studies of the Plymouth Brachyura. I: The Rearing of Crabs in Captivity, with a description of the Larval Stages of Imachus dorsettensis, Macropodia longivostris and Maia squinado. J. Mar. biol. Ass. U.K., N.S. 14 (3) : 795-814, pls. 1-4. 1928. The larval Stages of the Plymouth Brachyura. Proc. zool. Soc. Lond., 1928: 473-560, pls. 1-16. 1944. The larval stages of Portwmnus (Crustacea Brachyura) with notes on some other genera. J. Mar. biol. Ass. U.K., N.S. 26 (1) : 7-15, 5 text-figs. Nopre, A. 1931. Crustaceos Decapodes e Stomat6podes marinhos de Portugal. Inst. zool. Univ. Pérto, 1-307, 2 pls., 144 text-figs. 1936. Same title. Vol. IV in Fauna Marhina de Portugal, 1-216, 61 pls. Pesta, O. 1918. Die Decapodenfauna der Adria. Versuch einer Monographie : x + 500, I map, 150 text-figs. 8°. Leipzig & Wien. Plymouth Marine Fauna. 2nd ed., 1931. Mar. biol. Ass. U.K.: 1-371 (Crustacea, pp. 150-222). RatHeurN, M. J. 1930. The Cancroid Crabs of America of the families Euryalidae, Portu- nidae, Atelecyclidae, Cancridae and Xanthidae. Bull. U.S. Nat. Mus. 152: 1-609, 230 pls., 85 text-figs. SHEN, C. J. 1935. An investigation of the post-larval development of the Shore-Crab Car- cinus maenas, with special reference to the external secondary sexual characters. Proc. zool. Soc. Lond., 1935 (1) : 1-33, 28 text-figs. STEPHENSEN, K. 1945. The Brachyura of the Iranian Gulf. Danish Scientific Investigations in Ivan, Pt. IV : 57-237, 60 text-figs. PRESENTED © _ LIMITED, bigs e 82 THE MACROTRITOPUS PROBLEM BULLETIN: OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 2 No. 4 LONDON: 1954 THE MACROTRITOPUS PROBLEM BY W. J. REES, D.Sc. Pp. 67-100 ; Pl. 3 ; 17 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 2 No.4 LONDON: 1954 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted im 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical Series. Parts appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be convpiled within one calendar year. This paper is Vol. 2, No. 4 of the Zoological series. PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued January, 1954. Price Twelve Shillings. THE MACROURITOPUS PROBLEM By W. J. REES, D.Sc. SYNOPSIS The Atlantic species of the larval octopod genus Macvotritopus (M. equivocus, M. scoypio, M. kempi and M. danae), together with some Macrotritopus-like larvae described by Degner (1925), have been re-examined, and the type material of all but the first two species has been available for study. All these forms were found to be growth stages of Scaeurgus unicivrhus (Delle Chiaje), an octopus of the continental shelf. The Macrotritopus larvae appear to be able to delay settlement if they are swept over very deep water by water movements, and to continue growing to nearly twice the normal size for metamorphosis. These large larvae (found also in other groups of animals) are now termed extended pelagic stages, and their significance in dis- persal and in maintaining the homogeneity of populations of widely distributed species is discussed. ZOOL. I, CONTENTS INTRODUCTION os Genus Macrotritopus : o 6 O (a) Octopus (Macrotritopus) pyauoeus Robson’ (b) Octopus (Macrotritopus) scorpio (Berry) (c) Octopus (Macrotritopus) kempi Robson d) Octopus (Macrotritopus) danae Joubin & Robson (e) Scaeurgus (Troschel sp. juv.) Degner. : : 0 A RE-EXAMINATION OF Macrotritopus LARVAE . (a) Macrotritopus danae Joubin & Robson ‘ : 0 (b) Macrotritopus kempi Robson and other ‘‘ Discovery ”’ larvae (c) Scaeurgus (Troschel sp. juv.) Degner C . (d) Specimens from Biscayne Bay, Florida c : 5 A COMPARISON OF THE SPECIES . 0 : 5 5 . (a) Mantle length . : : c : c . (b) Head width : 5 2 5 ; 0 : (c) Thearms . 4 é ‘i 5 , 5 . . (d) Sucker diameter . : : 5 : : (e) The radula c 5 c C o 6 . (f) Gill filaments : 5 : 2 A ° : . THE BREEDING SEASON : “ . : E c (a) The Mediterranean s 5 5 : 2 0 : (b) The Atlantic : - c 2 a : é c THE SIGNIFICANCE OF THE Macyrotvitopus LARVA IDENTITY GF Macrotritopus witH Scaeurgus . CONCLUSION . . : . . REFERENCES . . . . . : . . 3 4. Page 7o THE MACROTRITOPUS PROBLEM A. INTRODUCTION Turs study originated from a desire to carry further earlier work on the planktonic stages of Octopus vulgaris (Rees, 1950, 1952). In particular, it was hoped that late planktonic stages of this species could be obtained from a search of plankton hauls made in the tropical Atlantic by the ships of the “‘ Discovery’ Investigations. During the course of this work a number of Macrotritopus larvae were collected, and it was the remarkable similarity of the youngest Macrotritopus larvae to those of Octopus vulgaris (except in the very long third arms) which led me to re-examine them in relation to a known adult or adults. The startling agreement of the chromatophore patterns in Macrotritopus and Octopus vulgaris suggested some relationship, but further investigations showed con- clusively the Macrotritopus could not be linked with any littoral species. Final conclusions indicate that all the so-called Macrotritopus spp. of the Atlantic belong to one species, and that the adult is Scaewrgus wnicirrhus (Delle Chiaje), a benthic octopod living at moderate depths near the edge of the continental shelf. B. THE GENUS Macrotritopus Grimpe (1922) proposed the name Macrotritopus as a generic name for a juvenile pelagic octopod, Octopus gracilis Verrill, because the third arms greatly exceeded the others in length. Later, Robson (1929), p. 168, et seq.) changed Verrill’s name to equivocus because the name gracilis was preoccupied. While recognizing that Macro- tvitopus was only a convenient label for juveniles, whose adult stages might or might not be known, he referred four other forms to it, viz. : Octopus scorpio (Berry), Macrotritopus kempi Robson, Octopus bandensis Hoyle and Octopus elegans Brock. Of these, Octopus (Macrotritopus) equivocus Robson and the first three above were described from the Atlantic and the Mediterranean. The last two are Indo-Pacific species and need not concern us further here. Later, Joubin & Robson (1929) described Macrotritopus danae from the collections of the Danish Dana Expeditions. All the known Atlantic specimens of Macrotritopus are juveniles exhibiting no sings of hectocotylization. It has also been found necessary to reconsider some larvae described by Degner (1925) as Scaeurgus (Troschel sp. juv.), because on re-examination they prove to be Macrotritopus larvae. The Macrotritopus forms which have been described are so distinctive in appearance that their true identity is masked by larval characteristics. In this connection the views of earlier workers are of interest. Verrill (1884), in considering his own Octopus gracilis (i.e., equivoca), was ‘‘ very certain it is not the young of any known species.” Robson (1929b) was unable to make up his mind about the status of these forms, as is evident from his statement : ‘““ We must either conclude that they are the young forms of species hitherto undescribed or that they undergo some kind of metamorphosis in the course of subsequent development, as a result of which they assume the adult form of some described species.’’ More recently, Pickford (1945) in discussing the American forms M. equivoca and M. scorpio, expressed the opinion that ‘‘ There is no possibility that these strikingly distinct young animals with their THE MACROTRITOPUS PROBLEM 71 very long third arms could belong to a vulgaris-like species.’’ The same author (1947) added “‘ whatever they may be they are clearly not the young of any littoral species.” Voss (1951) considered the possibility of M. scorpio being the larval form of Scaeurgus but had insufficient material to reach any decision. These views reflect the need for clearing up the status of the Atlantic forms of Macrotritopus, and except for M. equivoca I have been privileged to re-examine all the original material reported on by Degner (1925), Robson (1929), Joubin & Robson (1929) and Voss (1951) together with many additional specimens from the plankton catches of the ‘‘ Discovery ’’ Investigations, 1925-1939. I have been able to do this through the generosity of Dr. Anton Brunn (for the loan of Dana specimens) ; Dr. Gilbert Voss (for the use of two specimens of Macrotritopus from Florida) ; and Dr. N. A. Mackintosh and Dr. Helen Bargmann, for ‘‘Discovery’”’ material. I also wish to thank Dr. G. R. de Beer, F.R.S., and Dr. H. W. Parker for valuable sugges- tions, Mr. G. L. Wilkins for his superb drawings (Pi. 3 and figures 1-6), and Mr. H. O. Ricketts for his great care in examining large samples of plankton for octopod larvae. (a) Octopus (Macrotritopus) equivocus Robson Octopus gracilis Verrill 1884, p. 236 non Eydoux & Souleyet 1852, p. 13. Macrotritopus equivocus Robson 19204, p. 311. Octopus (Macrotritopus) equivocus, Robson, 19298, p. 160. Octopus (Macrotritopus) equivoca, Joubin & Robson 1929, p. 93. TYPE Locality. ‘“‘Albatross’”’ St. 2084, 40° 16’ 50” N., 65° 05’ 15” W., south of Cape Sable, Nova Scotia, 1,290 fms. ; I specimen. TABLE I Measurements in mm. Length of body (? ventral mantle length) . é REL O Mantle width . ¢ . 5 c : 3 5 7-0 Interocular width . ‘ : ‘ 3 : ; 6:5 Length of head and body : : : : 5 17/20) Arm length : Stas ; 4 : 2 si 3 é - 19:0 Zintal ¢ : g 5 a 5 c 5 Gato) Bray. é : “ 5 0 - 42-0 4th (damaged) . c : : : c ———— Verrill thought that the specimen was a juvenile of a species growing to a large size and was convinced that it was not the young of any known species. He noted the extraordinarily long third arms, the second arms being only half their length. On the evidence available from the description, it agrees with other Macrotritopus larvae noted later in having (in alcohol) large purplish chromatophores on the body and the head. There are also purplish spots (i.e., chromatophores) in front of and behind each sucker, this feature being common to all Macrotritopus and to many other octopod larvae. Verrill notes that the first 3-5 suckers on the lower pairs of arms are uniserial and 72 THE MACROTRITOPUS PROBLEM that this arrangement isnot found on the dorsal arms. This is unusual, but a minor point, as the arrangement may be due to unequal contraction of the arms during fixation. Three features about this larva are noteworthy. It is the largest known specimen of Macrotritopus ; it was caught over deep water and its occurrence (off Nova Scotia) is well to the north of any other records of larvae of this genus. (b) Octopus (Macrotritopus) scorpio (Berry) Polypus scorpio Berry 1920, p. 299, pl. 16, fig. 4. Octopus (Macrotritopus) scorpio, Robson 1929), p. 169. Octopus (Macrotritopus) scorpio, Joubin & Robson, 1929, p. 93. Locatity. “Bache” St. 10204, off Biscayne Bay, Florida ; 20.11.1914 ; 75-0 m. ; I specimen. jerry gives total length as 22 mm., and 4-5 mm. as dorsal mantle length. The measurements given below have been calculated from his figure in order to provide some basis for comparison with other larvae.t TABLE II Measurements in mm. Ventral mantle length : : c ie! Dorsal mantle length 6 3 2 2 - 5°89 Mantle width . 4 3 Interocular width 3 Arm length : Ist gon 2nd ; 5 5 5 : 4 6 a fen 3rd . - < 6 A 6 - 14°9 4th c : 4:62 Mantle-arm index . 3 5 a 5 sects) Berry had some doubts about his species being specifically distinct from Verrill’s gracilis, but mentioned the inequality in the length of the first and second arms, the still greater length of the third arms and the minute papillations as distinguishing features. The chromatophore pattern is only briefly mentioned. (c) Octopus (Macrotritopus) kempi Robson. Macrotritopus kempi Robson, 19294, p. 311. Octopus (Macrotritopus) kempt, Robson, 1929), p. 170. TyPE Locality. “ Discovery ’”’ St. 276, 5° 54’ S., 11° 19’ E., off the mouth of the Congo; 5.vill.1927; N 70B, 110-0 metres; 2 syntypes (B.M. 1947.3.12.1-2). Robson regarded this species as a very distinct one on account of its narrow head, 1 Berry’s figure is not very satisfactory ; the head is said to be “‘ distinctly narrower than the body,” but as pointed out by Robson it is nearly as wide as the body in the figure; Berry’s “ dorsal mantle length ”’ of 4.5 mm. is probably the distance from the apex to the nuchal ridge, and the measurement given in the table is the distance between the apex and the point midway between the eyes. THE MACROTRITOPUS PROBLEM 73 wide mantle, the smooth surface of the skin, the size of the web and the distinctive colour pattern. He also drew attention to the proportions of the third arm, and mentioned that there were 11-12 filaments in each demibranch. Robson (1929), p. 171) drew attention to the position where this species was found over deep water and about 10 miles off the edge of the continental shelf. This species is further discussed and the specimens re-described on pp. 76-77. (d) Octopus (Macrotritopus) danae Joubin & Robson Octopus (Macrotritopus) danae, Joubin & Robson, 1929, p. 87, fig. 1. Octopus (Macrotritopus) danae, Joubin 1937, p. 33, fig. 33. TYPE LOCALITY. “Dana” St. 1152, 30° 17’ N., 20° 44’ W., 23.1x.21 (160 metres). Paratypes from the Western Mediterranean, off Guiana and off W. Cuba. Joubin and Robson were clearly puzzled about the relationship of their species with previously described forms. However, they created a new species chiefly on the grounds that although the third arms were very long (74-80% of the total length), the second arms were also particularly well developed ; the third arm exceed- ing the second by 1-7—2-1 times its length. Other differences were mentioned. The measurements given by Joubin & Robson are given below for seven specimens examined by them. TaBLeE III Measurements in mm. No. of specimen. or = a 2 I 3 4 5 6 7 Mantle length (? dorsal) . E 13 13°5 10 10°5 10 9 12 Arms: . : : : o Iba URarae ALE IRS bg Tate IS ke Aa dRe, ULE, Ost, Gs IRE Ist - : : : . 20 21 16,256) 120A, 1314 iaeia Tos — arse 28 S22 R25 iy 22) 20m 2 2 Tene 28 Ty eT eo TENG) 48 37 38 36 37 42 37 33 35 35 28 32 39 36 27 2 20) —— 18) TS TSO, 7 Oe rs eT OL Indices : Mantle width, percentage length . 57 48 60 57 55 61 — Head width, percentage length = 46 44 50 52 50 55 50 3rd arms, percentage total length . 78 74 80 78 77 78 76 3rdarm > 2ndarm_. é yi I°5 ZO 1-6 I-5 197 1-8 times The above figures have not been used in my investigations, and all the specimens which I have been able to see have been measured again as it was desirable to get additional readings. The “‘ Dana’”’ material forwarded to me, together with paratypes in the British Museum (Natural History), contains a few specimens apparently not seen by Joubin & Robson, so it has been deemed desirable to examine this material in detail in the next section. 74 THE MACROTRITOPUS PROBLEM (e) Scaeurgus (Troschel sp. juv.) Degner, 1925 “ Thor ”’ St. 184: 38° 10’ N., 22° 23’ E., Gulf of Corinth ; 17.vili.1g10 ; 65 metres of wire. Some octopod larvae from the Mediterranean in which the third arms were par- ticularly well developed and long were described by Degner (1925, p. 79) ; these are essentially Macrotritofpus-like forms which have to be considered in any discussion of the Macrotritopus problem. Degner assigned these larvae to Scaeurgus because one of his specimens was showing the beginning of a hectocotylus on the third left arm. A sinistral hectocotylus is a characteristic of Scaeuwrgus and also of Pteroctopus tetracivrus, another Mediterranean species ; the latter species was not discussed by Degner in relation to his larvae. The relationship of these forms to other described species of Macrotritobus has not been fully realized by earlier workers, and it is possible that they have been misled by Degner’s fig. 52, which portrays a larva quite unlike the usual Macvotri- topus. It has been possible to re-examine these larvae in this investigation (see p. 74). c. A RE-EXAMINATION OF Macrotritopus LARVAE All the material of these larvae which it has been possible to gather together has been re-examined. It includes the type material of Macrotritopus danae, M. kempt, Degner’s Scaeurgus, two new specimens from Florida (the type locality of M. scorpio), as well as some additional specimens brought to light from an examination of hauls made in the tropical Atlantic by the ‘‘Discovery’’ Expeditions (1925-1939). (a) Macrotritopus danae Joubin & Robson , Larvae have been available from the following stations of the Danish “ Dana’ Expeditions 1920-22 (see Schmidt, 1929) : St. 1123%%, 37° 48’ N., 2° 44’ E., Western Mediterranean, 26.ix.21, S. 200, 200 metres of wire; I specimen (8-25 mm. in ventral mantle length). St. 1124", 37° 15’ N., 2° 55’ E., Western Mediterranean 27.ix.21, S. 200, 200 metres of wire ; 2 specimens (2:5, 6°75 and g:0 mm. in ventral mantle length); a’so Joubin & Robson’s No. 2 specimen. St. rr74ill, 5° 35’ N., 51° 08’ W., off Guiana,-16.x1.1921, S. 200, 300 metres of wire ; I specimen (now dry). St. 1223'", 22° 06’ N., 84° 58’ W.., off Cuba, 1.i1.22, S. 200, 50 metres of wire ; I specimen (4-95 mm. in ventral mantle length). New measurements have been made on the six larvae at my disposal. No. 1. This specimen, in sea-water formalin, is now rather faded with faint traces of chromatophores only on the third arms. The interbrachial web is mode- rately developed ; its depth in the A sector is about 3-0 mm., and about the same in other sectors (except sector E, where it is distinctly less well developed). No. 2 is a young and very interesting larva in which the third arm is more than twice the length of the other arms; the latter and the interbrachial web are but little developed. These shorter arms have the thin whip-like tips so characteristic also THE MACROTRITOPUS PROBLEM 75 TABLE 1V.—Macrotritopus danae Measurements in mm. No. SS = + i I 2 3 4 5 6 Ventral mantle length . ¢ $+25 2°5 6°75 9:0 9°75 4°95 Dorsal mantle length . c 10:8 3:0 II*55 13°35 12:0 5°7 Mantle width a o : 7*05 2°4 6-0 7°35 8 3:6 Interocular width 6°45 2°55 4°8 6°75 6:0 4°2 Diameter of eye . ; : 2°55 0-9 2°25 2°4 2°55 1°35 Arm length Ist arm : 0 : 14°25 I°5 8-7 14°25 15°75 3:0 2nd arm é c ee 2205 1°75 15°45 23°25 27:0 4°95 3rd arm : 6 B75 4°25 27°75 2°25 33°0 12°75 4th arm . : : 18-0 I-75 11-25 21-0 21:0 4°05 Diameter largest sucker : Diam. sucker, Istarm . 0°5-0°55 0°25 0-3 0°5 0°5 0-2 bs is 2ndarm . 0:55-0-6 0-2 0-4 0:6 0°55 — Ys os 3rdarm_—._ 0+75-0°8 0°25 0-6 0°75 0°8 0°45 Indices : Mantle arm index 22°0 58-8 24°35 27°9 29°6 38-8 Sucker index, 3rd_. , 9°7 10 8-9 8-34 8-21 91 is ay ePimal 5 5 7:28 10 5°95 6:67 5°64 — 55 Se US Glam > 6:67 10 4°44 5°50 5°13 4°04 of early stages of Octopus vulgaris, and their undeveloped state is reflected in the number of suckers they carry. From the base of the arm there are 3-4 uniserial suckers, 1-2 pairs of biserial suckers followed by rudiments of others. The third arm, however, has about 7-8 well-formed, biserial suckers and rudiments of others towards the tip. All the chromatophores have disappeared. No. 3. In this specimen the right eye is torn and protruding (allowance has been made for this in measurements) and the third left arm is mutilated. Except for traces of two rows on the outer surface of the third arms the chromatophores have disappeared. No. 4. The mantle-shape in this specimen is as figured by Joubin & Robson. The interbrachial web is moderately developed, the deepest sector being D. The chro- matophore pattern has all but disappeared and there is no trace left on the mantle. On the dorsal head there are two in the position indicated for M. scorpio by Berry but the others are very faint. On the third arms there are traces of a few, opposite suckers, and two faintly indicated rows on the outer surface of these arms. No. 5. This paratype (B.M. 1929.6.29.1.) has retained far more of its colour than any of the other ‘‘Dana”’ larvae ; this is perhaps partly due to the fact that the specimen is in alcohol, and partly also because it has been kept for the greater part of the time in total darkness. However, all chromatophores of the mantle have dis- appeared except three in a row along the dorsal mantle edge and the deep-seated ones of the visceral mass. Those of the dorsal side of the head can be counted and conform to the vulgaris pattern. The chromatophores of the third arms are contracted. On the outer surface there 76 THE MACROTRITOPUS PROBLEM is, proximally, a single row, followed distally by a double row. Here and there there is a suggestion of the pattern seen in WM. kenzpt. No. 6. No trace remains of the original chromatophore pattern. (b) Macrotritopus kempi and other “ Discovery ”’ larvae In addition to the syntypes of MW. kempi, larvae have been found at the following stations of the “ Discovery ’’ Expeditions (1925~—1939) : St. 276, 5° 54’ S., 11° 19’ E., N.W. of mouth of the Congo, 5.viii.1927, TYF, 150(-0) m.; 2 specimens of 3-3 and 4-5 mm. in ventral mantle length. St. 276, N. 100 B, 110-0 m. ; 2 syntypes of MW. kempi Robson and 1 specimen of 3:75 mm. in ventral mantle length. St. 290, 3° 25’ 25” N., 16° 50’ 52” W., off West Africa. 24.viii.1927, N 100 B, 86-0 m.; sounding 5,165 m.; I specimen of 5-4 mm. in ventral mantle length. St. 1592, 09° 31’ N., 17° 37’ W., off West Africa, 17.x.1935, TYFB, 200-0 m. ; I specimen of 6:0 mm. ventral mantle length. St. 1594, 04° 15’ N., 12° 58’ W., off West Africa, 19.x.1935, TYIFB, 144-0 m. ; I specimen of 6-15 mm. in ventral mantle length. St. 2646, 05° 38’ N., 14° 03’ W., off West Africa, 19.iv.1939, TYFB, 250-0 m.; I specimen of 3°2 mm. in ventral mantle length. St. 2646, TYFB, 1,500-800 m. ; I specimen of 3°45 mm. in ventral mantle length. The standard measurements of the Discovery specimens are given in Table V. TABLE V. Measurements in mm. No. c a — 7 8 9 10 II 12 13 14 15 16 Ventral mantle length a tion 7705) 3\-3 4°5 3°75 5°4 6:00 6°15 3-2 3°45 Dorsal mantle length 5 e)R(3} OS 75) Se7b) 0745) O57 Od) 0-75. OO 3°9 4°2 Mantle width . : é laa | ists) 3°6 3°9 3°75 405, 4°95 5°85 S255 —_ Interocular width . ‘ 6-0 5°4 3°3 4°5 Beda 7 5e ae Ohe 517, Oz {oy 40fo) Diameter of eye 2 5 = pace) = Arm length : rst ViaioGye HIE) ora tds 225 5A) 525 7°38 T°5 2°40 end 222 18-0 3°75 6°75 2-7 Ole ROr25 8:25 15°0 1*65 3°0 3rd). F : a S075 — 10'95 16°5 T0+5 LOMO) else eo) 6-0 6°75 4th . “ : - 18*45 14°25 225 Ais 2°85 0°5 8:25 10°05 I°5 2°25 Diameter, largest suckers : 2ndarm . : Os 5 Ong meOne Opes ore 0-2 03 or4 — _— 3rd arm_—y é OL 55 Ors 0°35 O-4 0°35 0°45 0°55 0°55 0°3 03 Indices : Mantle-arm index . ne 25 ar5) = Z0"N5 27°25 3305 33°75 39°25 21:6 5520) sro Sucker index : 3rdarm_ . . ‘ 6°38 6-54 10:6 8-88) 9°34 8°33 oOFm7 @:05) orm 8-7 2ndarm_ . 5 c 5°50 4°58 6°06 5°55 5°33 3°42 5°0 0:5 6-82 =a Nos. 7 and 8 and the holotype and paratype of Robson's M. kempi, while Nos. 9, 10 and 11 are also from the same station. THE MACROTRITOPUS PROBLEM 77 The type is in very good condition, and so is the paratype except that in it both third arms are damaged (the left one was regenerating a new tip). Robson’s figure (19290, p. 170, fig. 60) gives a very good impression of the arrangements of chromato- phores on the arms and head. On the ventral surface of the mantle, funnel and head the pattern of chromatophores is very like that in O. vulgaris. A notable feature in this specimen is the development of the other arms (1, 2 and 4), and the degree of development of the conspicuous chromatophores on them as on the third arm. On re-examination the head is seen to be rather contracted and the mantle fully inflated in both specimens. As regards measurements, in the species M. kempi the interocular width is very much less than the mantle width—a feature of considerable importance to Robson in defining his species. These specimens were picked out on board and transferred to alcohol with consequent shrinkage of the head region ; this is evident from com- parison with other specimens from the same station which were not picked out of the plankton samples (PI. 3, figs. r and 2). After making due allowances for size (and refraining from comment at this point on the relative proportions of the arms) all the specimens from this station undoubtedly form growth stages of one species. The most constant and characteristic feature is the chromatophore pattern of the third arms. On these (and on the second arms only to a lesser degree) the double tow of chromatophores on the outer surface is exceptionally clear and distinct, especially the dorsal row. In this row each chromatophore is lengthened at right angles to the axis of the arm, giving the arm the appearance of being banded. On the sides of the third arm there is often a chromatophore at the base of each sucker but sometimes this is placed between the suckers. On the ventral face of the third arm there are also chromatophores between the bases of the suckers. (c) Scaeurgus (Troschel sp. juv.) Degner, 1925 A re-examination of the original larvae lent by the Zoologiske Museum, K¢ben- havn, reveals little resemblance between the published figure (Degner, 1925, p. 79, fig. 52) and the eight specimens available for study. In brief, the larvae are very similar in general appearance to other M acrotritopus which have been described from the Mediterranean and the Altantic (Text-figs. I-5). Table VI gives the measure- ments obtained in 1952. Degner’s specimen No. 2 (specimen H above) is the largest example in this series. It has a ventral mantle length of 6-0 mm. (given as 6-3 mm. by Degner ; the diffe- rences between his measurements and mine are probably due to shrinkage in alcohol). Measurements of this and the other specimens from St. 184 are given in Table VI. It will be seen (Text-fig. 5) that in this late larva and in the younger ones the shape and proportions of the head and mantle are quite unlike Degner’s fig. 52, and have the usual form of Macrotritopus. ZOOL. Il, 4. 5§ 78 THE MACROTRITOPUS PROBLEM TABLE VI Measurements in mm. * * Overall length (i.e., including e 5 g y 2 x & uy 3rd arm) é . = LOe5s O95.) 14-85, TO-O5) O55) | 1827 9°6 — Ventral mantle length . 3:3 3°9 4° 3°0 EOF 2°35 3a VoKo Dorsal mantle length 4°65 4°85 5°55 4:8 3°9 AO5, | 3129 7°05 Mantle width yoni) ACG; 5016) 3°3 Zr 3:0 2°55 4°5 Head width 3°3 ZO Ese 3 3°90 2°55 2:4 2:4 3°9 Diameter of eye 1°05 13} I°5 eae) Koil Ras 3 OG, 1°65 Arms : ist 1°38 20 2/555 225, 2-0 1 Ory A0(0) 4°65 2nd 2°4 Poly. | NONI | 7407, Bris OG 08} 6°75 3rd . : 6 ey Beslan PGI} Gio 3/7] 5m 420) 00252 eee 4th é ; : 5 1°8 Bon 3°0 2°55 1:8 1°75 2:°0 6-0 Diameter proximal (basal) sucker 0-2 0-2 0.225 a 0:2 0:2 0-18 org Diameter last uniserial sucker 0:25 0-25 0°25 = 0-2 0-2 0:2 0-3 Diameter rst biserial sucker . 0-2 0°25 0°25 —_ 0-2 0-2 0-2 0°35 Diameter largest sucker (3rd arm) ORS 0°25 org log 0°22 0°22 0°25 0°45 * Specimens D and G are distorted, so that the mantle length for these is estimated. The chromatophore pattern is rather faded but some features can be distinguished ; those of the dorsal mantle have disappeared, but on the ventral surface there are about 20, scattered over the surface as in O. vulgaris larvae of similar size (i.e., in mantle length). Deep-seated chromatophores are present on the dorsal head as in the pattern typical of O. vulgaris. On the arms the chromatophores occur in a single row up to the edge of the web and then become double. The web is fairly well developed and is of the order D=C >B>A=E. On the first arm it reaches to almost half its length. The body and arms are covered with the remains of Kéllikersche buschel and some papillae, scattered over the mantle are developing. There are 11 gill filaments in the outer demibranch of the left gill, The feature of greatest interest is the third left arm, which, according to Degner, was beginning to exhibit hectocotylization. I can make out no details in the struc- ture of the tip of this arm, which is constricted off to form an egg-shaped structure with a slightly-pointed tip; its size is 0-30 mm. x 0:20 mm. (Text-fig. 6). The identity of these larvae with Scaeurgus or Pteroctopus depends on our inter- pretation of this structure. Is it due to the vagaries of preservation, or can we inter- pret its symmetrical shape and delicately rounded appearances as a developing hectocotylus ? Consideration of this feature in relation to distribution is deferred to p. 96. (d) SPECIMENS FROM FLORIDA Recently G. L. Voss (1951) has mentioned some Macrotritopus larvae which he considered to be the scorpio form described by Berry (1920, p. 299). By courtesy THE MACROTRITOPUS PROBLEM 79 Fics. 1-4. Scaeurgus sp. Degner, 1925 : the smaller larvae from “‘Thor’’ St. 184, Gulf ot Corinth ; all approximately x 6. Fic.5. Scaewrgus sp. Degner 1925: the largest larva from ‘‘Thor’”’ St. 184 (specimen H), exhibiting signs of hectocotylization ; x 6. Fic. 6. Scaeurgus sp. Degner: tip of the third left arm from specimen H with rudimentary hectocotylus ; greatly enlarged. 80 THE MACROTRITOPUS PROBLEM Macrotritopus @ Degner (1925) © Dana | ® Discovery r@ Voss (1951) & Berry (1920) a i Q = Ww = ke FL < = VENTRAL MANTLE LENGTH 2 3 ZW 7h Ghee) ANG) Fic. 7. Relationship between mantle width and ventral mantle length in Macrotri- topus danae, M. kempi, M. scorpio, Scaewrgus sp. Degner and in additional specimens from the ‘“‘ Discovery ’’ Expeditions (1925-1929). The letter k denotes the type material of M. kempi. All measurements in figs. 7-17 are in millimetres. of Mr. Voss I have been able to examine two of the larval forms he mentions ; one is in formalin and the other (No. 18) in alcohol. I was unable to see much pattern on the surface of these larvae as their pigment has largely disappeared, presumably through exposure to sunlight after fixation. However, such pattern as remained indicated that there was nothing unusual about it, and that it was probably no different in this respect from other Macrotritopus larvae I had seen. THE MACROTRITOPUS PROBLEM 81 TaBLE VII.—Macrotritopus larvae from Florida Measurements in mm. No. ity] 18 Ventral mantle length 4°35 3:0 Dorsal mantle length 4:8 3:3 Mantle width 3°9 2°55 Interocular width 3°75 2°5 Arms: Ist 3°15 1-3 2nd 4°95 HOG 3rd : ; 10:2 6-0 4th . i a . 0 3/0 I°5 Diameter, basal sucke - : 0-2 0-125 second uniserial sucker 0:2 0-15 * first biserial sucker 0°25 o-2 - largest sucker, 3rd arm I'2 0'3 pv. A COMPARISON OF THE SPECIES It was recognized quite early in the investigation that some or all of the Mediter- ranean-Atlantic species of Macrotritopus might prove to belong to one or two species, and the problem was to distinguish between growth stages and specific differences. Accordingly some characters which might reveal significant points of difference were examined. Here it may be convenient to mention that ventral mantle length is used through- out in all calculations involving mantle length. It appears more reliable than the usually accepted dorsal mantle length (mid-point between the eyes to the apex of the mantle), because retraction of the head in the larva can make as big differences in measurements as distortion of the mantle margin. In the larva, too, there is a distinct ridge on the dorsal mantle which provides a useful check if by chance the ventral mantle edge is distorted. All the accompanying graphs (except Text-figs. g and 10) are plotted on Wright- man’s double logarithmic paper to demonstrate the relationships of the various parts. In using the data from this assorted collection of octopus larvae I have been very conscious of the inadequacy of the material, and for this reason statistical methods have not been exploited. It is, however, surprising in view of the assortment of material how well grouped the measurements are in relation to straight lines. It became obvious early in the investigation that material in alcohol, notably Degner’s Scaeurgus and Robson’s two syntypes of Macrotritopus kempi, were much shrunk, the shrinkage affecting different parts unequally. With age, too, shrinkage progresses, and it must be remembered that the ‘‘Thor’’ collections were made 40-50 years ago and those of the “‘ Discovery’ between 1925 and 1939. Similarly the material of M. danae and larvae from the “‘ Discovery ”’ collections, although both are in formalin, are not strictly comparable from a preservation point of view, the former being flaccid (and possibly a little swollen ?) by comparison with the latter, 82 THE MACROTRITOPUS PROBLEM Macrotritopus @ Degner (1925) °F © Dana 9 8 — @ Discovery 7 @ Voss (1951) 6 & Berry (1920) 5 4 ae = = = : ja) = [a Ww A NWUOO LENGTH OF 3rd. ARM VENTRAL MANTLE LENGTH 2 3 4.5 6 7 8 910 Fic. 13. Relationship of the third pair of arms to ventral mantle length in Macrotri- topus danae, M. kempi, M. scorpio, Scaeurgus sp. Degner, and in additional material obtained by the “ Discovery ” Expeditions (1925-1939). Fic. 14. THE MACROTRITOPUS PROBLEM 89 Macrotritopus 50 40F © Ist ARM 30F © ond ARM bk OM AN WOO | | se >= oc < {a O r ke O Zz Lu — Uin cA eas Ge TB HOULO VENTRAL MANTLE LENGTH Relationship of arms 1, 2 and 3 in relation to ventral mantle length in all the specimens examined, 90 THE MACROTRITOPUS PROBLEM Macrotritopus @ Degner (1925) © Dana ® Discovery @ Voss (1951) SUCKER DIAMETER OF 3rd ARM VENTRAL MANTLE LENGTH | 2 3 ms @en7 amon Fic. 15. Relationship of the sucker diameter of the third arm to ventral mantle length in Macrotritopus danae, M. kempi, M. scorpio, Scaeurgus sp. Degner, and in additional material obtained by the ‘ Discovery ’’ Expeditions (1925-1939). The letter k denotes the tvpe material of MM. kempzt. well formed endocone but no ectocone. The third lateral is distinctly long, slender and rather curved. The marginal plates are lozenge-shaped, a little broader than long. The radula of the holotype of Macrotritopus kempi has also been examined and is very similar to the above, although details of the seriation of the cusps are difficult THE MACROTRITOPUS PROBLEM or to interpret. Even though the preparation is a poor one, the cusps have a symme- trical arrangement and not an asymmetrical one as stated by Robson (1929, p. 312). The laterals are very similar to those of the specimen from “ Discovery ”’ St. 1592. (f) GILL FILAMENTS. The number of gill filaments per demibranch has been noted in one specimen of each of the following : Scaeurgus sp. Degner, E. Mediter- ranean, II filaments; Macrotritopus kempi Robson, West Africa, 11-12 filaments ; Macrotritopus danae Joubin & Robson, W. Mediterranean, at least 11 ; Macrotri- topus danae Joubin & Robson, off Cuba, at least 11. The third and fourth specimens listed above have deteriorated considerably in formalin, and the detailed structure of the gills, is not now very satisfactory for counting filaments. All the evidence suggests that all the Macrotritopus larvae examined belong to a single wide-ranging species. I have not seen Verrill’s Macrotritopus (M. equivocus Robson), but there appears to be no doubt about its identity with the other larvae ; it appears to be nothing more than a very large larva carried by water movements well beyond its normal range. E. THE BREEDING SEASON On the assumption that all the larvae belong to a single species it is possible to get some indication of the breeding period. We have to assume that once the eggs are laid (presumably at moderate depth) they take at least as long, if not longer, to hatch as in a littoral octopod. In Octopus vulgaris this is a period of 21-30 days, according to temperature, and it may be even longer in a deep-water species where sea temperature on its spawning grounds is likely to be lower. In order to supplement the data obtained from the material studied, I have drawn on a list of occurrences given by Joubin (1937, p. 36) in addition to my own material. In the latter (Text fig. 16) ventral mantle length is taken as a criterion of size, while in the former overall length appears to be quoted (Text-fig. 17). As breeding may be largely controlled by temperature, the Mediterranean records have to be treated separately from those of the tropical Atlantic. (a) THE MEDITERRANEAN. In this warm temperate area we have records of larvae of all sizes but only for the months of August and September. It is true that during the Third “ Dana”’ Expedition in 1921-22 the Mediterranean was visited only between 21st September and 4th October, 1921. The earlier expedition in the “Thor” (1908-1910) did, however, cover a much longer interval of time (14th December, 1908, to 21st February, 1909, and 26th June, 1910, to 7th September, 1910) but recorded larvae only in August. If, as the meagre evidence suggests, larvae are common only in August and Sep- tember, spawning must take place in June-July and possibly in August as well. By the end of September, when the largest larvae have been found, the majority are ready for settlement on the bottom, while a few late larvae may persist well into October. This is in agreement with Naef’s statement that octopod larvae at Naples 92 THE MACROTRITOPUS PROBLEM O Atlantic @ Mediterranean VENTRAL MANTLE LENGTH JAN. FEB. MAR. APR. MAY JUNE JULY AUG. SEPT. ocT. NOV. DEC. Fic. 16, Size (ventral mantle length) in relation to seasonal occurrence in the Atlantic and Mediterranean. are more common in August and September than during other times of the year and suggests summer breeding. (b) THE AtLantic. In the Atlantic only two larvae have been found to the north of 30° N. They were found at the following positions : “Dana” St. 1152, 30° 17’ N., 20° 44’ W., S.W. of Madeira, 23.x.192z ; I specimen of 18 mm. “ Dana” St. 1341, 33° 15’ N., 68° 20’ W., between Bermuda and the mainland, 14.v.1922 ; I specimen of 20 mm. All the remainder were found in tropical or sub-tropical waters off West Africa and on the American coast, mainly off islands of the Greater and Lesser Antilles. On the African coast the ships of the ‘‘ Discovery ’’ Committee took specimens in March, April, August and October during transit voyages to and from the Antarctic. There has, however, been no continuous sampling on this coast to give an accurate picture of seasonal distribution, THE MACROTRITOPUS PROBLEM 93 , In the Central American region however the “Dana” was constantly taking plankton hauls between November, Ig2r, and May, 1922. There was no month in this period when larvae were not present. This implies that the breeding period was spread over the whole of this time, and if we take the records from West Africa and Florida into consideration, June is the only month for which larvae have not been recorded, and this apparent absence seems to be due to lack of observations. LENGTH SEPT. ocT. NOV. DEC. JAN. FEB. MAR. APL. MAY JUNE @ = Mediterranean © Eastern Atlantic © Western Atlantic Fic. 17. Size (overall length) in relation to seasonal occurrence in the Atlantic and Mediterranean based on records of Macrotritopus spp. from the cruises of the ‘‘ Dana ”’ (1921-1922). It is therefore clear that breeding may be noted in tropical and sub-tropical waters throughout the year. The greatest number of larvae (31) was taken in December, but the significance of this is not known. It may be that plankton hauls during this month were made in an area where there was concentration of breeding adults, or it might mean that this was the period of maximum breeding activity. Ifthe latter it would imply maximum breeding in winter in the tropics—contrasting with summer breeding in the more temperate Mediterranean. In the Tropics, however, the factors which may exercise a big influence on temperature (which governs breeding) are local hydrographic conditions. For an animal which lives on the narrow coastal shelf (as off West Africa and the Antilles) local temperature conditions are bound to vary enormously according to movements of water masses, especially upwelling. Local breeding may therefore vary according to place and season. 94 THE MACROTRITOPUS PROBLEM r. THE SIGNIFICANCE OF THE Macrotritopus LARVA The plankton life of the larvae of Octopus vulgaris (to judge from results hitherto unpublished) may have a duration of at least four weeks, and possibly eight or ten weeks under normal hydrographic conditions. For Macrotritopus it is reasonable to suppose a planktonic life of longer duration (perhaps for three to five months) on the basis that the larvae reach a much larger size. There is, however, another factor to be considered; one of Degner’s larvae from the comparatively shallow Gulf of Corinth was developing a hectocotylus at 6-0 mm. in ventral mantle length. To judge from evidence from all sources, including Naef’s work in the Mediterranean, metamorphosis (here interpreted as the abandonment of a planktonic or pelagic phase) also means the beginning of sex differentiation and the development of gonads ; in the male the visible outward sign is the development of a hectocotylus. In the Gulf of Corinth the water is not very deep, and a larva ready to become ben- thic should find a suitable bottom at the right time, so that there would be no great delay in settlement. This implies that the ordinary Macrotritopus larva completes its planktonic phase at 6~7 mm. in ventral mantle length (allowance being made for the shrinkage of this material). Other Macrotritopus larvae, however, reach a much bigger size without developing any sign of hectocotylus or of gonads. In M. eguivocus the ventral mantle length is estimated to be 11 mm. ; in MW. kempi and M. danae respectively it is 8-1 mm. and 9775 mm. In view of the fact that I can find no other differences between these larvae and Degner’s species, some other explanation must be sought for the non- appearance of signs of sexual differentiation or maturation, and I am inclined to interpret these large larvae as extended pelagic phases. This phenomenon has been noticed before in marine invertebrates, but its signifi- cance does not seem to have been discussed in relation to distribution and dispersal. It is known from the work of Dr. Marie V. Lebour that in many molluscs a creeping- swimming stage of several days’ duration is reached during metamorphosis, which, as suggested by Wilson (1937), may enable the larva to persist in a state whereby it can be swept away by currents to a more suitable substratum. From our present knowledge of the change to a benthic stage it is evident that species with a quick metamorphosis can delay settlement for several days, thus giving them a better chance of finding a suitable bottom. When metamorphosis is gradual, settlement can be delayed for a long period, as has been demonstrated by Day & Wilson (1934) for the polychaete Scolecolepsis fulginosa. During this period of uncertainty the worm continues to grow beyond the 14-setiger stage at which these authors found metamorphosis to take place. In another polychaete, Loimia medusa, Wilson (1928) found that its planktonic exist- ence ended when the larva had reached a length of about 6-0 mm. At this size the first pair of gills is beginning to divide, the second is only a rudiment and the third has not made its appearance. Still larger larvae were referred to the species by Monro (1930 and 1931) ; of these the largest example from ‘‘ Discovery ’’ St. 102 was 15 mm. long and had three pairs of well-developed gills. If we accept the view (as is THE MACROTRITOPUS PROBLEM 95 most likely) that these larvae belong to the widely-distributed Lovmia medusa, they can only be interpreted as an extended pelagic phase.* Gurney (1942, pp. 71-75) gives details of several larvae of decapod crustacea which have been caught at sizes well above the known size for moulting into post-larvae. He was “‘ reluctant to accept any theory of abnormality ”’ for these giant larvae, and supposes that they had failed to metamorphose at the right time and continued to grow.” He adds: “ .. . ina situation like that of Bermuda, where abyssal depths are so close to the shore, it must be quite usual for larvae to find themselves over deep water at the time of metamorphosis.” Similarly when the distribution of Macrotritopus larvae is considered many of them were also taken fairly close to land, but in deep water off the edge of the con- tinental shelf. At the places where they were taken (off West Africa and in the Antilles) the continental shelf is very narrow, so that the chances of larvae being swept over deep water are great. On the West coast of Africa, particularly, this seems to occur, and Knudsen (1950) attributes the high percentage of gastropods with very short planktonic or non-pelagic development on this coast to the influence of water movements in eliminating species with planktonic stages of long duration. The Macrotritopus larva with its long third arms (which may be as useful as the squid’s tentacles in catching food), its narrow squid-like profile and its curious colour pattern recalling that of the epipelagic larvae of the Cranchiidae, seems well adapted for a pelagic life. It may be that Macrotritopus specimens of above 6-7 mm. in ventral mantle length are larvae which have grown large in the plankton because they have not been able to find a suitable bottom (or a suitable bottom at the right depth) for metamorphosis. In other words delayed settlement has resulted in an extended pelagic phase which may be of considerable significance in dispersal. What is important is that it can travel immense distances in the plankton. The M. equivocus found off Cape Sable, Nova Scotia, had reached a mantle-sac length of II mm., and must have been transported for hundreds of miles by currents (probably from the Carribean or Florida) ; it was likely to be well out of the area in which it could survive as an adult. It may, however, be pertinent to draw attention to the role of the extended pelagic phase in maintaining uniformity in a wide-ranging species. The remarkable similarity of populations of Scaeurgus from the Atlantic and Pacific was noted by Berry (1914, p. 306), who adds (im Jitt., quoted by Voss, 1951) : “if it really is all one species, then there must be some special reason for lack of divergent speciation, and I can conceive no sensible explanation for this in a bottom form except for the possession of a planktonic larval stage of some duration.”’ G. THE IDENTITY OF Macrotritopus WITH Scaeurgus When it became obvious that all the Altlantic Macrotritopus were referable to a single species, it was deemed possible to suggest a possible adult with which they * Dr. Gunnar Thorson has kindly drawn my attention to Lemche’s observations on the North Atlantic tectibranch Diaphana minuta Brown, which seem to suggest that the larva of this mollusc is able to delay metamorphosis when swept our over deep water (Thorson, 1946, p. 466 and Lemche, 1948, p. 9). In shallow water adults the embryonic shells are much smaller than those in specimens from depths exceeding 1,000 m. These larger embryonic shells are believed to indicate delayed settle- ment and resulting growth of the larval shell in specimens which have metamorphosed in deep water. 96 THE MACROTRITOPUS PROBLEM could be linked. When this paper was in early stages of preparation in 1950 there were four species with one of which they could possibly be linked, namely, Octopus vulgaris, O. macropus, Scaeurgus unicirrhus and Pteroctopus tetracirrhus. The first, O. vulgaris, was the only species which had a known distribution corre- sponding most closely to that of the larvae (Macrotritopus equivocus being con- sidered to have been carried well out of its normal distribution area). The details of the structure of the radula—asymmetrical in O. vulgavus and symmetrical in Macrotritopus—together with the occurrence of a normal larva of vulgaris in the same hauls as M. kempz (St. 276) were sufficient evidence that O. vulgaris could not be the adult. The second, O. macropus Risso, has a similar but a more restricted distribution in the Mediterranean and both sides of the tropical Atlantic. This species, too, was ruled out because the details of the radula showed specific differences, and also because larvae of O. macropus were positively identified during the investigations. Scaeurgus uniciyrhus and Pteroctopus tetracirrhus were originally included in the list of possible adults, not because of any great similarity between the known dis- tribution of the adults and that of Macrotritopus, but solely because each species had a sinistral hectocotylus. Degner’s largest larva (specimen H) had a swelling at the tip of the third left arm, and on re-examination of the species I found it difficult to imagine its delicately rounded structure being the result of any mal- formation of the tip of the tentacle, and had to conclude that it was a true rudiment of a hectocotylus. Evidence from distribution of the adults was not encouraging in 1950, especially as neither species had been found on the American side of the Atlantic and records outside the Mediterranean were very few. However, Voss (1951) gave the first records of Scaeurgus from the Western Atlantic (Florida), and Adam (1952) recorded Ptevoctopus from West Africa. Full records of the two species outside the Mediter- ranean (where both are found) are given below : Scaeurgus unicivrhus : Pacific Ocean; Hawaiian Islands (Berry, 1913 and 1914 as patagiatus) ; South of Kyushu (Sasaki, 1920 and 1929 as patagiatus) ; Indian Ocean ; Saya de Malha (Robson, 1921) ; Atlantic Ocean, Florida (Voss, 1051). Pieroctopus tetracivrhus: Atlantic Ocean; Cape Verde Islands (Fischer & Joubin, 1906 and 1907) ; Azores (Joubin, 1900) ; off the coast of Africa (Adam, 1952). Scaeurgus appears from our present knowledge of distribution to have a much wider range than Pteroctopus ; as both species live on moderately deep bottoms, it it not surprising that there are so few tropical records in view of the difficulties of operating trawls in areas where the bottom is covered with living or dead coral. Both these species have a symmetrical arrangement of cusps on the rhachidian tooth of the radula as in Macrotritopus. A closer examination of the Macrotritopus radula shows that it agrees well with Scaeurgus, and differs from Pteroctopus in having a narrower first lateral tooth, a longer, more slender and more curved third lateral, and lozenge-shaped marginal plates. In Pteroctopus (see Adam, 1952, fig. 55) the marginal plates are wider and slightly curved in outline. THE MACROTRITOPUS PROBLEM 97 All the evidence suggests that Macrotritopus is the larval stage of Scaeurgus unicivrhus, and at present there is no indication that more than one species is in- volved. Scaeurgus has been found in the warm waters of all oceans, and its wide distribution (surprising in a bottom-dwelling animal) has clearly been achieved by this striking pelagic larva. The number of gill filaments per demibranch in Scaeurgus (11-14) is exceptionally high for a benthic octopod living at considerable depths near the edge of the continental shelf, but is probably related to the pelagic require- ments of the larva. H. CONCLUSION All four Atlantic species of Macrotritopus have been demonstrated to belong to one species, also now linked with its adult, Scaeurgus unicirrhus (Delle Chiaje). Macrotritopus falls into the synonymy of Scaeurgus, and the position regarding the species is summarized as follows : Scaeurgus unicirrhus (Delle Chiaje). Octopus gracilis Verrill, 1884, p. 236 non Eydoux & Souleyet, 1852, p. 13. Macrotritopus gracilis, Grimpe, 1922. Macrotritopus equivocus Robson, 19294, p. 311. Octopus (Macrotritopus) equivocus, Robson, 1929), p. 169. Octopus (Macrotritopus) equivoca, Joubin & Robson, 1929, p. 93; Joubin, 1937, P- 34. Polypus scorpio Berry, 1920, p. 299. Octopus (Macrotritopus) scorpio, Robson, 1929), p. 169; Joubin & Robson, 1929, P- 93. Macrotritopus kempi Robson, 1929a, p. 311. Octopus (Macrotritopus) kempi, Robson, 1929), p. 170. Octopus (Macrotritopus) danae Joubin & Robson, 1929, p. 87 ; Joubin, 1937, P- 33- Macrotritopus spp. Joubin, 1937, p. 36. Scaeurgus (Troschel sp. juv.) Degner, 1925, p. 79. The great length of the third arm is a larval character only, for it is not found in the adult, in which the arms are subequal. The long third arm may have some special significance in larval life, serving perhaps for capturing food in the same way as the tentacles of the epipelagic squids. The high number of gill filaments, too, indicates an active larval existence which is in contrast with the low number found in octopods like Bathypolypus and Bentheledone, which are benthic deep-water octopods of the continental slope. The records indicate that the Macrotritopus larva lives in the upper 200 metres. Most of the hauls from deeper water were made with open nets, so that positive conclusions cannot be reached from them, but the specimen taken by the “ Dis- covery ’’ with a closing net between 1,500 and 800 metres was at a surprising depth in a species which has all the appearance of being a surface form. Its occurrence at this depth suggests that it was seeking bottom preparatory to metamorphosis. If we accept the occurrences of Macrotritopus larvae in the vicinity of land (or 98 THE MACROTRITOPUS PROBLEM near the edge of the continental slope) as evidence of the existence of benthic adults nearby, Scaeurgus has a more general distribution in the warmer waters of the Atlantic than hitherto supposed. As already mentioned, the difficulties of trawling over coral in the tropics probably accounts for the scarcity of records of the adult. REFERENCES Apam, W. 10952. Céphalopodes. Res. Sci. Evpéd. Océanog. Belge Eaux Cétiéves A fricaines de lV’ Atlantique Sud (1948-19490), 3 (3) : 1-142, pls. 1-3 and text-figs. 1-57. Berry, S.S. 1913. Some new Hawaiian Cephalopods. Proc. U.S. Nat. Mus. 45 : 563-566. 1914. The Cephalopoda of the Hawaiian Islands. Bull. Bur. Fish. Washington, 32 (1912): 257-362, 40 text-figs. and 11 pls. 1920. Preliminary diagnoses of new cephalopods from the Western Atlantic. Proc. U.S. Nat. Mus. 58 : 293-300. Day, J. H., & Witson, D. P. 1934. On the relation of the substratum to the metamorphosis of Scolecolepis fuliginosa (Claparéde). J. May. Biol. Assoc. 19 : 655-662. DeGNER, E. 1925. Cephalopoda. Danish Oceanog. Exped., 1908-1910, 2, C. 1: 1-94, 52 text-figs. Fiscuer, H., & Joupin, L. 1906. Note sur les Céphalopodes capturés au cours des expédi- tions du Travailleur et du Talisman. Bull. Mus. Hist. Nat. 1906, No. 4: 202-205. 1907. Céphalopodes. Expéd. Sci. Travailleur et du Talisman, 8 : 313-353. GrIMPE, G. 1922. Systematiche Ubersicht der Europaischen Cephalopoden. Sitzber. Naturf. Ges. Leipzig, 9: p. 36. Gurney, R. 1942. Larvae of decapod Crustacea. Ray Soc. 1942, vi + 306 pp., 122 text-figs. Jounin, L. 1900. Céphelopodes provenant des Campagnes de la Princess Alice. Res. Camp. Sct., Monaco, 17 : 1-135, 15 pls. 1937. Les octopodes de la croisiére du “‘ Dana’’ 1921-22. Dana Rep. No. 11 : 1-49, 53 text-figs. —— & Rosson, G.C. 1929. Ona new species of Macrotritopus obtained by Dr. J. Schmidt's “Dana ” Expedition, with remarks on the genus. Proc. Zool. Soc. London, Pt. 1: 89-94, I text-fig. KKnuDsEN, J. 1950. Egg capsules and development of some marine prosobranchs from tropical West Africa. ‘‘A/lantide’’ Rep. No. 1: 85-130, 31 text-figs. Lemcue, H. 1948. Northern and Arctic tectibranch gastropods. I. The larval shells. II. A revision of the cephalaspid species. MX. Danske Vidensk, Selsk. Biol. Sky. 5 (3) : 1-36, 80 text-figs. Monro, C. C. A., 1931. Note on the pelagic phase of a polychaete worm belonging to the family Terebellidae. Ann. Mag. Nat. Hist. (10), 7: 212-215. —— 1936. Polychaete Worms, II. “Discovery ’’ Rep. 12: 59-108, 34 text-figs. Narr, A. 1921-28. DieCephalopoden. Faune et Flova del Golfo di Napoli. Monographie, 35. Pickrorp, G.E, 1945. Le poulpe américain : a study of the littoral Octopoda of the Western Atlantic. Tvans. Conn. Acad. Arts Sci. 36 : 701-811, 14 pls. 1947. A review of the littoral Octopoda from Central and Western Atlantic Stations in the collections of the British Museum. Ann. Mag. Nat. Hist. (11) 13: 412-428, 1 text-fig. Rees, W. J. 1950. The distribution of Octopus vulgaris Lamarck in British Waters. J. Mar. Biol. Assoc. 29 : 361-378, 3 text-figs. and 3 pls. 1952. Octopuses in the Channel. New Biology, No. 12: 58-67, 3 text-figs. Rosson, G. C. 1921. On the Cephalopoda obtained by the Percy Sladen Trust Expedition to the Indian Ocean in 1905. Tvans. Linn. Soc. (2) Zool. 17, (4) : 429-442, 2 pls. 19294. Notes on the Cephalopoda: VII. On Macrotritopus Grimpe with a description of a new species. Ann. Mag. Nat. Hist. (10) 3 : 311-313. ——~ 1929b. A Monograph of the Recent Cephalopoda. Part1: Octopodinae. Brit. Mus. (Nat. Hist.), London, xi-+236 pp., 89 text-figs. and pls. 1-7. THE MACROTRITOPUS PROBLEM 99 Sasaki, M. 1920. Report on Cephalopods collected during 1906 by the United States Bureau of Fisheries steamer ‘‘ Albatross’ in the North-western Pacific. Pyvoc. U.S. Nat. Mus. 57 : 163-203, pls. 23-26. 1929. A monograph of the dibranchiate cephalopods of Japanese and adjacent waters. J. Coll. Agri. Hokkaido Univ. 20, Supplementary Number, 1-357, 30 pls. and 159 text-figs. Scumipt, J. 1929. Introduction to the Oceanographical Reports, including list of the stations and hydrographical observations. Danish ‘‘ Dana’’ Expeditions 1920-22, No. 1: 1-87, 18 text-figs. and 6 pls. THorson, G. 1946. Reproduction and larval development of Danish marine bottom inverte- brates, with special reference to the planktonic larvae in the Sound (¢resund). Medd. Komm. Danmarks Fisk. Havundersog. Ser. Plankton. 4 (1) : 1-523, 199 text-figures. VERRILL, A. E. 1884. Second Catalogue of Mollusca recently added to the fauna of the New England Coast and the adjacent parts of the Atlantic, consisting mostly of Deep-sea species, with notes on others previously recorded. Trans. Conn. Acad. Sci. 6 : 139-294, pls. 28-32. Voss, G. L. 1951. A first record of the cephalopod, Scaeurgus unicivrhus, from the Western Atlantic. Bull. Mar. Sci. Gulf Caribbean, Coral Gables I (1): 64-71, 2 text-figs. Witson, D. P. 1928. The post-larval development of Loimia medusa Sav. J. Mar. Biol. Assoc. 15 : 129-146, 2 pls. and 7 text-figs. 1937. The influence of the substratum on the metamorphosis of Nofomastus larvae. Ibid, 22 : 227-243. 1952. The influence of the nature of the substratum on the metamorphosis of the larvae of marine animals, especially the larvae of Ophelia bicornis Savigny. Ann. Inst. Océanog. Paris, 27 (2) : 49-156, 2 text-figs. PLATE 3. Fies. 1 and 2. Macyrotritopus larva of 3:45 mm. in ventral mantle length from ‘‘ Discovery” St. 2646. Figs. 3 and 4. Macrotritopus larva of 3-75 mm. in ventral mantle length from “‘ Discovery ’ St. 276. This specimen was taken in the same haul as Robson's syntypes of M. kempt. PLATE 3 Bull. B.M. (N.H.) Zoology IT, 4. St oe SR CiS MACROTRITOPUS LARVAE. a i] w s 4 H 4 4 a H < i) 4 oO Z Lol Aa a is a 4 % & LIMITED, SON, AND ADLARD DORKING BARTHOLOMEW PRESS, HEARING IN CETACEANS pees FRASER: AND ee bE PURVES BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 2 No. 5 LONDON : 1954 HEARING IN CETACEANS BY F, €. FRASER AND Paks PURVES Pp. 101-114 ; Pls. 4-5 BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) Vol. 2 No. 5 ZOOLOGY LONDON: 1954 _ THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY). instituted in 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical Series. Parts appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 2, No. 5 of the Zoological series. PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued January, 1954. Price Four Shillings. HEARING IN CETACEANS By F. C. FRASER anp P. E. PURVES. SYNOPSIS Essential anatomical features of the region of the ear of cetaceans are described and figured. Earlier hypotheses relating to the mode of hearing in cetaceans are briefly reviewed and criticized. Reasons are given for regarding the method of hearing in cetaceans as being essentially similar to that of terrestrial mammals with certain anatomical modifications required for receiving water-borne vibrations. A simple experiment is described demonstrating the superiority of the walls of the meatus over other parts of the body as a conductor of longitudinal vibrations. At the present time the sounds emitted by cetaceans and consequently their sense of hearing are the subject of a considerable amount of investigation. The remarks that follow are in anticipation of a more detailed account of the cetacean ear which the writers have in preparation. In order to clarify the hypothesis arrived at a brief summary of the anatomical features involved is necessary (in connection with which reference to Plates 1 and 2 should be made). The external auditory meatus is a continuous narrow tube in the toothed cetaceans (with the exception of Physeter catodon, see Clark, 1948). In the whalebone whales it is usually closed for some part of its length immediately internal to the blubber layer. The meatus is lined by a pigmented extension of the epidermis. Surrounding the lining layer is a fibro-elastic sheath, the fibres of which run predominantly along the length of the meatus. Surrounding this sheath again is a fibro-cellular structure in which a thin stratum of circular constrictor muscle has been observed. Associated with the tube are incompletely investing cartilages into which auricular muscles are inserted. Of these one of the most important is the m. auricularis- occipitalis-profundus which Bonninghaus (1904) described and figured in detail in connection with the dissection of a Common Porpoise. Bonninghaus’ figure shows the acute dorsally- and slightly caudally-directed bend into which the meatus is thrown in its course. The muscle just referred to originates at the top of the skull and extends downwards to its insertion at the apex of the bend in the meatus. Another equally important muscle is Bonninghaus’ m. zygomatico-auricularis which has its origin on the zygomatic process of the squamosal and converges to be inserted into the cartilage at the inner end of the meatus. In both toothed and whalebone whales the meatus widens out mesially to ter- minate at the tympanic membrane. The latter is not as in terrestrial mammals a thin translucent membrane. It consists of two parts, a fibrous region containing ZOOL, 2, 5. 104 HEARING IN CETACEANS radial and concentric fibres and a fibreless region similar to the pars flaccida of land mammals. The fibrous portion consists of a broad flattened triangular “ ligament ” with one edge of its base attached to part of the tympanic ring, the base thus forming part of the external surface of the ‘‘membrane.” The attenuated apex of the “ligament ’’ formed by the convergent radial fibres is directed mesially and is attached to a small process on the manubrium of the malleus. It is this ligament? which, in the opinion of the writers, is the true homologue of the fibrous portion of the tympanic membrane of land mammals. The non-fibrous portion bridges the space between the fibrous portion and the remainder of the tympanic ring. In Phocaena phocoena (fide Bonninghaus) the non-fibrous portion merely forms a few islets in the fibrous portion ; it is single and more extensive in area in the Globi- cephala melaena specimen examined by the writers and in their opinion the “ glove finger” of the whalebone whales, described by Beauregard (1891) and Lillie (1910) as well as by earlier anatomists, is the same structure very greatly enlarged. The three auditory ossicles commonly found in the Mammalia are also represented in the Cetacea. In the latter the malleus is fixed by the processus gracilis to the margin of the tympanic annulus at the position of the Glaserian cleft. This con- dition is not peculiar to cetaceans, Hyrtl (1845) and Bonninghaus (1904) having described it in monkeys, carnivores and insectivores. The hammer-shape is main- tained but it is the processus gracilis which resembles the handle. The head, in addition to the portion recognized as such in other mammals, includes the manu- brium, which is very much reduced in length to forma short tubercle. The processus gracilis is longer and stouter than that of other mammals and in structure somewhat resembles a short length of channel girder. At the junction of the malleus with the Glaserian cleft the tympanic ring is developed into a strong wing-like lateral projection buttressed to the processus gracilis. This projection known as the sigmoid process is peculiar to the Cetacea, and plays an important part in the mechanics of the cetaceans’ hearing. The head of the malleus is deeply notched by two large facets making a re-entrant angle on its posterior aspect. Both of the facets have smooth, shallowly convex surfaces covered with articular cartilage which with the corresponding facets on the incus form part of a synovial joint. The radii of the convexities, as well as that of the arcuate junction between the two facets, lie approximately at right angles to the long axis of the tympanic ligament. The tensor tympani muscle, which appears to be largely tendinous, is attached to the ventral wall of the periotic near that part of the tympanic cavity associated with the outlet of the Eustachian tube. Although in its course to the malleus it is directed meso-laterally approximately in line with the long axis of the tympanic ligament, the attachments of the muscle and tympanic ligament to the malleus are displaced dorso-ventrally from one another by a small distance. The ligament and muscle are not antagonistic to each other but form a mechanical couple. The form of the incus is comparable with that of mammals generally. The facets for articulation with the malleus are shallowly concave, furnished with articular 1 It will be referred to as the tympanic ligament in this paper. HEARING IN CETACEANS 105 cartilages and meet at an angle which fits into the re-entrant angle of the malleolar facets. The processus brevis is a short conical projection directed anteriorly. The ligament attaching the incus to the periotic bone, the axis of the processus brevis and the lateral margin of the processus gracilis make a common axis. The stapes departs so little from the normal that its description is unnecessary, but it should be mentioned that, contrary to the statements which have been made that the foot is fused to the periotic at the fenestra ovalis, no evidence of this could be found by the present writers. It is probable that the impression of fusion is due to the perfect fit of the foot in the fenestra ovalis. The tympanic bulla, the general form of which is sufficiently well known not to require detailed description, may be said to have a mesial and a lateral portion which meet ventrally in a roughly semi-circular ridge and are separated dorsally by the cavity of the bulla. The mesial part is characterized by its stout petrous nature (the bone in rorquals being up to 3-4 cm. thick), and by the smoothness of most of its surface, both of these features are associated with its contiguity with the air sinuses. The lateral portion is much thinner, and has a roughened outer surface, these features being associated with its contiguity with the surrounding fibrous layer, which is about 12 cm. thick in a large rorqual and strongly adherent to the lateral surface of the bone. The bulla in the whalebone whales is attached by two thin, flat pedicles placed respectively anteriorly and posteriorly and having their planes approximately at right angles. The cavity of the bulla is continuous with that of the middle ear and its associated air sinuses. In the Odontocetes the anterior pedicle is absent and the posterior support does not involve bony fusion of the tympano-periotic junction. In all cetaceans the periotic is separated from the rest of the skull at the squamo- mastoid boundary. In the rorquals the much attenuated mastoid process is loosely wedged between the squamosal and basi-occipital bones—being maintained in position by fibrous tissue. The mastoid process of the beaked whales and some River Dolphins is less attenuated but much convoluted—the convolutions inter- digitating with corresponding cavities on the postero-ventral tip of the squamosal. There is no fusion of the bones. In the Delphinidae the periotic is neither wedged into, nor interdigitated with, the squamosal, but is separated from the bones of the cranium by an appreciable gap. Between the mastoid process—which is very short—and the basi-occipital there is a fibrous ligament which appears to be the only well defined point of attach- ment to the skull. The cochlea and semi-circular canals occupy only a relatively small part of the petrous bone as a whole. It has been stated that in cetaceans the laminae spiralis primaria and secondaria of the cochlea make a continuous bony septum across the spiral labyrinth. The authors’ dissections show that this is not so and that whilst the gap between the two laminae is very narrow at the entrance to the scalae it gradually widens towards their apex. There is a basilar membrane and organ of Corti as in other mammals. In all but absolute size the cochlea conforms with the usual pattern. The cavity of the middle ear communicates with the Eustachian tube and with . 106 HEARING IN CETACEANS a system of pneumatic sinuses which in volume, extent and structure are peculiar to the cetaceans. In the region of the auditory organ accessory air sinuses occupy the space between the mesial aspect of the periotic and the surrounding bones of the skull as well as separating the periotic from soft parts on its lateral aspect. Anterior to the tympanic bulla a very large air sinus is developed which in the rorquals occupies nearly the whole of the pterygoid fossa. In the beaked whales and dolphins this sinus is even more extensive and covers the greater part of the base of the cranium in addition to the pterygoid fossa, which is very much enlarged. In the Common Dolphin this sinus has an anterior extension which passes forwards on the ventral surface of the rostrum for approximately two-thirds of the latter’s length. The mesial walls of the pneumatic sinus are closely applied to the bones of the skull. Laterally and externally the sinuses are closed by a tough, fibrous membrane resembling periosteum to which some of the muscles of the lower jaw and palate are attached. A comparative study of this membrane shows that in the more primitive species it overlies thin sheets of bone and it has been suggested that the enclosed sinuses are the homologues of the pneumatic sinuses found in the skulls of terrestrial mammals (Monro, 1785). Within the fibrous closing membrane there is a fibro-venous plexus which extends over the whole of the inner surface of the air sinuses. The plexus is made up pre- dominantly of large vessels which appear flattened when the sinuses are injected with air or other media but become turgid when injected themselves under pressure. In this state they obliterate the cavity of the air sinuses. Lining the latter and covering the inner aspect of the fibro-venous plexus is a thick mucous membrane which is continuous with that of the middle ear and Eustachian tube and is remark- able for the richness of its glands and ducts. The openings of the ducts, which cover the entire inner surface of the air sinuses, are less than -I mm. apart and lead into a maze of smaller racemose channels and crypts lined with columnar epithelium and goblet cells. At the entrance to the ducts and on the exposed surface of the mucous membrane is a layer of ciliated epithelium. The writers’ observations, and those of Brazier Howell (1930) on freshly killed specimens reveal that the sinuses are entirely filled with an albuminous foam. Whether this is so in living animals cannot be ascer- tained, but the anatomical evidence suggests that it is.* The Eustachian tube diverges from the general air cavity at a point just anterior to the tympanic bulla—and passes forwards for a short distance along the mesial aspect of that part of the sinus which goes to occupy the pterygoid fossa: thereafter it passes upwards and gradually narrowing, opens into the nasopharynx. Its closing walls are made up of fibrous tissue and a fibro-venous plexus similar to, and continuous with, that of the air sinuses. The lining mucous membrane is deeply indented with valvular pockets and folds which are directed towards the choanae. * Additional evidence of foam in the air sinuses has been received from Mr. D. E. Sergeant, Newfound- land Fisheries Research Station, St. John’s, Newfoundland, who writes (10-8-53): ‘‘Globicephala melaena, female, 9’ 6” in length, dead about 1} hours. Anterior region of sinus within pterygoid exposed by means of saw cuts; seen to be filled with foam”’. HEARING IN CETACEANS 107 The fibro-venous plexus referred to above is supplied by an arterial plexus which emerges from the external maxillary artery immediately anterior to the tympanic bulla, and by small arterial branches which emerge from the same artery as it passes forwards across the ventral surface of the cranium. The plexus is drained by three distinct paths : (1) By large vessels which communicate with the transverse and cavernous sinuses of the cranium and drain eventually into the spinal-meningeal veins. (2) By large vessels which join the external jugular vein via the vena pterygoidea. (3) By an intricate plexus of small veins which penetrates the fibrous covering of the sinus at the angle formed by the lateral pterygoid and the tensor palati muscles. This plexus is very dense and ramifies throughout the mass of fatty tissue which lies on the mesial aspect of the lower jaw— eventually coalescing into a single vessel which joins the external jugular vein. The aquatic environment of whales and the extreme narrowness of their external auditory meatus have led most students of cetacean anatomy since Camper (1767) to search for some mode of hearing other than that which is normal for terrestrial mammals. Camper’s hypothesis was to the effect that cetaceans could only hear when lying quietly at the surface with the external aperture of the meatus out of water. Hyrtl (1845), who recognized that whales must be able to hear well under water, held the view that while the external meatus was unsuitable for the reception of air-borne sound waves, it was suitable for the reception of ‘‘ waves of compression ”’ such as are transmitted in water. Buchanan’s (1828) belief was that sound was conducted to the tympanic mem- brane by way of the Eustachian tubes. This view was criticized by Claudius (1858) on the ground that the Eustachian tubes are normally closed while the ceta- cean is submerged. Claudius’ own view was that the water-borne sound vibrations were taken up by the bones of the skull, transmitted to the pneumatic sinuses, thence to the auditory ossicles but principally to the membrane of the fenestra rotundum which he regarded as the main oscillator of the cochlear fluid. Turner (1913) agreed with Claudius’ interpretation. Bonninghaus (1902) invoked the pinna- like shape of the anterior portion of the tympanic bulla as a collector of sound waves ; vibrations which he terms “ molecular’’ are transmitted by way of the processus gracilis through the ossicles to the fenestra ovalis. Lillie (1910) says ‘“‘the whale probably receives sound vibrations by means of vibrating bony surfaces, after the manner of fishes. The tympanic bulla is a relatively dense and heavy sounding-box, fastened to the periotic bone by two thin pedicles, so that it could easily be set in vibration. The bulla is connected with the fenestra ovalis by the chain of ossicles, the auditory apparatus being thus independent of the tympanic membrane.”’ Kellogg (1938) concurred with this view. In the opinion of the present writers Hyrtl’s interpretation is the most acceptable one. The anatomical features of the cetacean ear, far from indicating that cetaceans hear by abnormal means, support the view that the mode of hearing is essentially 108 HEARING IN CETACEANS as in land mammals with precisely those modifications of structure which are required for receiving water-borne vibrations. Any working hypothesis relating to hearing in cetaceans must be based on the assumption that these animals have the same requirements as terrestrial mammals. These requirements are that hearing should be directional, discriminative, selective and anticipative. The accepted hypotheses for directionality in hearing involve an appreciation of an intensity and/or phase difference in the sounds received by the two essential organs of hearing respectively. On the face of it it is difficult to see how this can be achieved by bone conduction. In the binaural reception of sound necessary for directionality the advantages of the cochlea being situated at or near the extreme lateral limits of the skull are offset by the likelihood of mechanical damage and extreme limitation of voluntary control- ability of the sounds entering the external opening of the meatus. The muscularly controlled pinnae of land mammals enables them, in addition to securing direction- ality, to control the volume of sound received. In the Cetacea the anatomical evidence indicates that binaural hearing is achieved by placing the two receptors remote from the surface of the body, by isolating them acoustically from the rest of the body and each other, and by connecting them to the surface of the body laterally by an apparently acoustically efficient muscularly-controllable vibration conduit. The acoustic isolation of the periotic is achieved by (1) its dissociation from the adjacent bones of the skull, (2) the provision and maintenance of an air gap between it and the rest of the skull at all hydrostatic pressures, (3) the foam which, on the evidence of freshly-killed animals, appears to occupy the air sacs and middle ear of the living animal. The first of these has already been described. The second is achieved by the mechanism of the accessory air sinuses. As the hydrostatic pressure increases with depth the blood is forced into the venous plexus lying below the fibrous outer covering of the sacs. The volume of the air sacs is correspondingly reduced and the compressed gas is contained in the rigid tympanic cavity and in the space between the periotic and skull. With diminishing depth the process is reversed. This interpretation of the function of the air sacs agrees with that put forward by Beauregard (1891) except that the role of the venous plexus does not coincide with his view. Owing to the flattening of the large lining veins when the air sinsues are inflated the blood drainage from the venous plexus and mucous membrane must be pre- dominantly through the rich plexus of small vessels which passes through the mass of adipose tissue occupying mesial concavities of the lower jaw. The very close association of the plexus with the fatty tissue and the known high nitrogen absorp- tion capacity of the latter compared with that of blood (6: 1) suggest that the fat may play a part in the absorption of nitrogen contained in the pneumatic sinuses. That the mucous membrane of the sinuses may have a gas-secreting function is a possibility which requires further investigation. In considering the theory expressed by Claudius and implied by Bonninghaus that sound waves are conveyed to the tympanic bulla through the air in the pneumatic sinuses it must be pointed out that films of air surrounded by media of high acoustic resistance like water do not transmit sound well. For example, a film of -I cm. HEARING IN CETACEANS 109 of air surrounded by water reflects about 93% of the incident energy of a sound wave of 1000 c.p.s. (Wood, 1941). If the air is surrounded by solid rock or presumably by a substance equally hard, for example, petrous bone, the reflected energy is about 99:93%. If in addition the air cavity is filled with foam the reduction in energy would be even greater. The efficiency of the bubbles in damping vibrations increases rapidly both as their diameters and distance apart diminishes, thus in the cetacean as the animal dives the acoustic isolation could be maintained in spite of the reduction in volume of the foam-filled cavity. In view of this apparent acoustic isolation of the essential organs of hearing it was decided to examine the external meatus as a possible vibration conduit. Claudius states that “All true cetaceans toothed as well as whalebone possess absolutely no functioning ear passage’ and subsequently after a very incomplete description of the meatus states “The cord is then, neither homogeneous nor taut, and its conducting power very slight, and by no means any greater than that of the sur- rounding fat layer.’ The anatomical and experimental investigations carried out by the writers do not support Claudius’ conclusions, but point to a set of conditions which are exactly the reverse. The lining walls of the meatus appear extremely homogeneous (vide supra) and the arrangement of the two most conspicuous auricular muscles would appear to be precisely that necessary for maintaining either or both auditory conduits in any state of tension required. In order to obtain some indication of the sound conductivity of the meatus relative to that of adjacent tissues a simple test was made. For the mechanical vibrations required the output of a variable frequency oscillator (30—10,000 c/s) was connected to the input of an amplifier of a loud speaker. A loop of dental floss was attached to the centre of the loud speaker and the two ends attached to two small pieces of perforated zinc which had been stitched to the wall of the meatus and the adjacent fibrous tissue respectively. The specimen used consisted of a large piece of the squamo-occipital region of the head of a Fin Whale which, deep frozen while fresh, had been defrosted and dissected to expose the tympanic cavity. For indicating the relative strength of received vibration signals a gramophone pick-up was connected to an amplifier, the output of which was connected to a rectifying millivoltmeter. With a fixed output from the oscillator and with the pick-up needle placed at the external end of the meatus, the input potentiometer of the amplifier in the pick-up circuit was adjusted to give a reading of 0-4 mv. at the millivoltmeter. To obtain the same reading with the pick-up needle placed at the malleolar end of the tympanic ligament the input potentiometer of the amplifier had to be adjusted. This adjust- ment corresponded to a voltage ratio of 10:1 so that assuming that the voltage reading is directly proportional to the mechanical vibration picked up, the signal received at the malleolar end was about 1/10 that applied at the outer end. When attempts were made to pick up signals on the periotic bone, on adjacent skull bones, on the muscular or on the fibrous tissue in the neighbourhood of the tympanic cavity the millivoltmeter reading fell to zero. With the apparatus available for this simple 110 HEARING IN CETACEANS test the limits set by the oscillator output and the amplification in the pick-up circuit, prevented a quantitative comparison being made between these last tests and the first one. The indication is, however, that the signal in any case is less than a quarter of that obtained at the malleolar end of the tympanic ligament. The results confirmed earlier preliminary tests made with a stethoscope. Because of the mechanical difficulty of placing the stethoscope on the malleus the vibration was applied to the point of attachment of the tympanic ligament and the signal picked up at the external end of the meatus. With this arrangement good signals were received, but with the stethoscope applied to surrounding tissues results were negative. To satisfy the discriminative condition it must be assumed that the cetacean cochlea is as selective as that of terrestrial mammals. Certainly its general anatomy suggests that thisisso. The physical properties of water-borne vibrations, however, differ markedly from those which are airborne. The pressure-amplitude for the same intensity and frequency of water and air-borne sound is in the ratio 61: TI, and the displacement amplitude 1:61. The mathematical treatment of the subject is not within the scope of this paper but it can be shown that adjustments of ampli- tude and pressure to values normally experienced in the cochlea by terrestrial mammals are probably achieved in cetaceans by the modifications of the malleus and incus. Before describing the functioning of the cetacean ossicles it is necessary to repeat the often quoted modus operandi of the human ear. The following is from Gray’s Anatomy (1946) : “The handle of the malleus follows all the movements of the tympanic membrane, while the malleus and incus rotate together about an axis which runs through the short process of the incus and the anterior ligament (and therefore the processus gracilis) of the malleus. When the tympanic membrane and handle of the malleus move inwards the long process of the incus also moves in the same direction and pushes the base of the stapes towards the labyrinth.” A mechanical study of their ossicles shows that the foregoing description is true for cetaceans with one important difference. In man and other terrestrial mammals the long axis of the malleus is parallel with the plane of the tympanic membrane and the manubrium is firmly attached to the latter as far as its centre. Thus the amount of movement of the centre of the tympanic membrane is identical with that of the tip of the manubrium. In cetaceans the membrane (i.c., the tympanic ligament) is attached only to a point at the tip of the manubrium of the malleus. The attachment to the tip is so situated that a line through the centre of the tympanic ligament is parallel with the long axis of the manubrium. This means that any small movement of the meatal end of the tympanic ligament produces a relatively large movement of the tip of the manubrium and therefore of the tip of the long process of the incus. In order that this amplification be produced the effective movement of the malleus must be solely rotational and this is achieved by the stout channel-girder construction of the processus gracilis and its firm fusion to the tym- panic ring. In addition, its lateral edge is strongly buttressed by the sigmoid process of the tympanic ring. This construction ensures that lateral movement of the malleus is obviated and that the processus gracilis is capable of torsional vibration only. The angular articulation of the malleus with the incus produces HEARING IN CETACEANS Ill a condition in which there is positive oscillatory drive by rotation only. The smooth curved surfaces of the facets tend to slip over one another in one or both directions when the oscillation is transverse or longitudinal to the processus gracilis. This factor must be of great importance in the exclusion of extraneous vibrations. In this last connection it is necessary to examine the hypothesis of: Lillie and Kellogg (v. supra) in more detail and as a preliminary it will be useful to recall Beauregard’s (1894) conclusions on the subject. “The tympanic bulla cannot be regarded as an apparatus for resonance as has been asserted by various anatomists and physiologists (Hunter, Home, Esser, etc.). The massiveness of the walls of the bulla and its shape recalling more or less that of a shell have evidently given birth to this idea, but the facts do not justify it. These so-called resonators are more or less completely enveloped by soft parts which do not favour their action. The shape is extraordinarily variable in reality and the massiveness of the walls is necessary to obtain a rapid equilibrium of pressure between the air which fills the bulla and that of the exterior. Savart was therefore correct in saying that the tympanic cavity has the function of retaining near the entrance of the labyrinth and on the internal aspect of the tympanic membrane an atmosphere of which the physical properties are constant.” In agreeing with this view the writers would draw attention to additional anatomical details which support it. (1) the great thickness and strength of attach- ment of the fibrous capsule which envelops the ventro-lateral aspect of the bulla, this being the part which would be most easily thrown into vibration ; (2) the extreme density and thickness of that part of the bulla which is exposed to air on both surfaces, the internal damping of such a mass must be considerable ; (3) the angular set of the two thin pedicles is such that the bulla would not easily be thrown into vibration, but assuming it were it would be expected to vibrate about an axis running through the pedicles. Examination of the positioning of the articular facets of the malleus in relation to this axis indicates that no drive to the incus could be achieved (the connection between the malleus and incus has already been described). The majority of sounds in water can be regarded as being conveyed from a point source by spherical pressure waves. When such waves impinge on the cetacean body, and assuming the least favourable acoustic conditions, i.e., that the general shape and various tissues of the body are non-reflecting, the body mass would form no barrier to the progress of the waves except in the neighbourhood of the ear. Here the sound energy, except that conveyed by the external meatus, must undergo almost complete reflection because of the gaseous content of the tympanic cavity and accessory air sinuses. The walls of the external meatus would be thrown into longitudinal vibration whether or not a continuous free lumen existed. Such vibrations would be conveyed to a point on the mesial end of the manubrium mallei by the convergent fibres of the tympanic membrane. The oscillations of the malleus thus initiated would be conveyed to the cochlear fluid over the bridge formed by the auditory ossicles. In terrestrial mammals the intensity of sounds received can be governed by orientation of the head, but principally by the use of auricular muscles. In the Cetacea the orientation of the head in relation to the rest of the body is limited by the absence of a well-defined neck. If, however, it is accepted that the sounds 112 HEARING IN CETACEANS received at the cochlea of the cetacean are conducted along the external auditory meatus by longitudinal waves the degree of tension of the walls of the meatus must be a controlling factor in sound intensity. The m. auricularis occipitalis profundus and the m. zygomatico-auricularis of the Porpoise described by Bonninghaus seem particularly well suited in position and structure for the function of adjusting the tension. The meatus, however, is attached to the tympanic annulus and an addi- tional mechanism is required to maintain the tension of the fibrous portion of the tympanic membrane lying within the tympanic cavity. The positioning of the non-fibrous portion of the membrane in relation to the fibrous portion is such that under the influence of the gaseous pressure within the tympanic cavity the necessary tension could be provided. The final expression of this function can be seen in the “ slove-finger ’’ of the rorquals. In them the non-fibrous portion is enlarged in correspondence with the greatly elongated tympanic ligament. The arrangements which exist in terrestrial mammals for making small temporary adjustments to the tension of the tympanic membrane are also present in the Cetacea, but in view of the coarse adjustment required it is doubtful whether the tensor tympani muscle is as effective in this respect as it is in terrestrial mammals. The palatal and lateral pterygoid muscles are however very well developed in cetaceans and their relationships to the Eustachian tube and pneumatic sinuses suggest that they may play an important part in the adjustment of the gaseous pressure within the tympanic cavity and consequently of the tension on the tympanic ligament. BIBLIOGRAPHY BEAUREGARD, H. 1894. Recherches sur l'appareil auditif chez les Mammiferes. J. Anat, Paris, 1894: 367-413, 3 pls. BonninGHaus, G. 1904. Das Ohr des Zahnwales, zugleich ein Beitrag zur Theorie der Schalleitung. Zool. Jb., 19: 189-360, 28 figs., 3 pls. BucuHanan, T. 1828. Physiological Illustrations of the Organ of Heaving. London. CampPER, P. 1767. Verhandeling over het Gehoor van den Cachelot, of Pot-Walvisch. Verh. Holland. Maatsch., 9 (3) : 193-2209, figs. Carte, E., and Macatister, A. 1868. On the Anatomy of Balaenoptera rostrata. Phil. Trans., 158 : 201-261, 5 pls. Craupius, M. 1858. Physiologische Bemerkungen tiber das Gehdororgan der Cetaceen und das Labyrinth dey Sdugethiere. Kiel. Home, E. 1812. An Account of some Peculiarities in the Structure of the Organ of Hearing in Balaena mysticetus of Linnaeus. Phil. Tvans., 102 : 83-88, 2 pls. Howe tt, A. B. 10930. Aquatic Mammals : their Adaptations to Life in the Water. Pp. vii; 338. Springfield, Illinois. Hunter, J. 1787. Observations on the Structure and Oeconomy of Whales. Phil. Traus., 77 : 371-450, 8 pls. Hyrt1, J. 1845. Vergleichend-anatomische Untersuchungen iiber das innere Gehororgan des Menschen wnd dey Saéugethiere. Wp. viii, 139, 9 pls. Prag. Jounston, T. B., and Wuittis, J. (Editors), 1946. Gvray’s Anatomy, 29th ed. 1-1597, figs. London . Keiioce, R. 1938. Adaptation of Structure to Function in Whales. Publ. Carneg. Instn., 501 : 649-682. Litiiz, D. G. 1910. Observations on the Anatomy and General Biology of some members of the Larger Cetacea. Proc. Zool. Soc. Lond., 1910 : 769-792, 1 pl., Io figs. Monro, A. 1785. The Structuve and Physiology of Fishes. 1-128, 44 pls. Edinburgh. HEARING IN CETACEANS 113 TuRNER, W. 1914. Observations on the Auditory Organ in the Cetacea. Pyroc. Roy. Soc. Edinb., 34 : 10-26, 4 figs. Woop, A. B. 1941. A Text-book of Sound, being an account of the Physics of Vibrations with special reference to recent theoretical and technical Developments. 1-578, figs., tables and plates. London. Munesato Yamada’s “ Contribution to the Anatomy of the Organ of Hearing of Whales,” reprinted from the Scientific Reports of the Whales Research Institute, No. 8, 1953, was received after completion of the foregoing preliminary account. It will be considered in the more detailed paper which the authors have in preparation. EXPLANATION OF PLATES PLATE 4 Fic. 1.—Dissection to expose the ventral aspect of the left ear of a Pilot Whale, Globicephala melaena. viations used in this and the following figures :— Cut surfaces of bone hatched and the corpus cavernosum shown reflected. Abbre- —Manubrium mallei. —Occipital condyle. —FPeriotic. —Processus gracilis. —Posterior pedicle. pts—Pterygoid sinus. AAS —Accessory air sinus. MM ac —Auricular cartilage. oc At —Anterior ligament. P ap —Anterior pedicle. PG B —Blubber. PP c —Cochlea. cc —Corpus cavernosum. s ErAM—External auditory meatus. sc Fo —Fenestra ovalis. SM FR —Fenestra rotundum. sP Fvp —Fibro-venous plexus. TA GL —Glove finger. TB 1 —Incus. TL Lt —Ligamentum incudis. 1T mM -—Malleus. Fic. 2.—Detail of figure 1. PLATE 5 —Stapes. —Semicircular canal. —Stapedial muscle. —Sigmoid process. —Tympanic annulus. —Tympanic bulla. —Tympanic ligament. —Tensor tympani muscle. Fic. 1.—Dissection to expose the ventral aspect of the left ear of a foetal Lesser Rorqual, Balaenoptera acutorostrata. (Abbreviations as for Plate 4, figure 1.) Corpus cavernosum shown reflected. Fic. 2.—Right tympanic annulus and auditory ossicles of a Humpback Whale, Megaptera novaeangliae. as for Plate 4, figure 1.) The dotted line indicates the axis of the tympanic ligament. (Abbreviations Bull. B.M. (N.H.) Zool. 2, 5. PLATE 4. Fic. 1. Bull. B.M, (N.H.) Zool. 2, 5. PLATE 5. GE A \ Fic. 1 (se 1: ) ~~ a ; na Ly “yin FO SC - ‘ ¥ © a — 7 : ’ * he 7 ‘ PRIN TED IN GREAT BRIT ALN BE -ADLARD AND SON, LIMITED, OMEW PRESS, DORKIN ‘ * ‘ is “7 octi9s4 | a _ PRESENTED x THE “ROSAURA EXPEDITION THE “ROSAURA” EXPEDITION Pp. 115-239; Pls. 6-9; 40 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 2 No. 6 LONDON : 1954 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) instituted in 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical Series. Parts appear at trvregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 2, No. 6 of the Zoological series. PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued September, 1954 Price Thirty Shillings THE “ROSAURA” EXPEDITION 1937-1938 CONTENTS Page 1. Gear, Narrative and Station List. J. S. CoLman : ELS 2. Under-Water Illumination and Ecology in Tropical Estuaries, a. Ss. Cotman and L.H.N.Cooper. : : 5 G lighe 3. The Echinodermata. D.D. Joun and A. M. Gaia 5 . 139 4. Fishes. Part I: Families Carcharhiniidae, Torpedinidae, " Rosauridae (nov.), Salmonidae, Alepocephalidae, Searsidae, Clupeidae. DENys W. TucKER : : : c : c ' c LOS 5. Sponges. M. BurTOoN . - 5 a : b : 2 . 215 FOREWORD The collections made by Mr. J. S. Colman during the cruise of Lord Moyne’s yacht “‘ Rosaura ”’ in the north and central Atlantic and the Caribbean region were received at the British Museum (Natural History) only a short time before the outbreak of war. In consequence little work was done on them at the time, and during the post-war years there have been many tasks of greater urgency. In some groups the material collected adds little to existing knowledge and no special report on it, as a whole, is warranted. In such instances new knowledge has been, or will be published elsewhere ; published papers dealing with ‘‘ Rosaura ’”’ material are listed by Mr. Colman in the bibliography attached to his narrative of the cruise (p. 130). The present Bulletin contains a random collection of reports at present available. Others may be expected from time to time in this Bulletin. es ae Th ee , AMO Mie Wiha tive i PP ey a a x ae ; ; Papa “ ii se , ea ily 3 a a | ee es ee fee, B04 8 cel — by th +0 ia) ee Re es ‘2 eae mata huis a ar ‘ aa er ae o ee eG ee ae 2D fie > a ‘ .*) dab a) ey ‘ay Re Re gee “~ : Meas 7 il pigsty AA auiaaein aay - KL aes, eo TAD i a al ke nett ie aa rey Ociant OE, cA bai Foe | along TAR hy (el Yin Fang aw ayes ; a ei) | fs igi (eaten ; amet f pidagh 3 Vas bak aun Uk ehh rats al Al BBW a inal ee en ee cvs 1 here ' i j ‘ f ; f pen Ww D i? |: Pd Puatoe y swt ; = By LT “We Tny Wel Pemtinl, jie U ne A se st ei i baggie ee wt pa. tt ae Lata , ote Bret an’ TE, BY Poy, EP ye out re, it et) hjie man Ge Pate —— oP keel ewe os) ay ghia Spee af Parag per ae a he inn, drag tap eg oT) fate Et a) SEAL ii i cassapeeee a anaeion ee =a” eens Lone NS Tn 8 aan $e HOP aie or ; 5 3 F . Several Crustacea: Decapoda: Sergestidae, : . , - Penaeus sp., young. c 6 - Many Zoeas, j ‘ 5 - 6 FI iF Other larvae, 5 : : ‘ Pa Amphipoda : in tubes, . : : 3 5 . Several Isopoda: . : . : - 5 - : 2 Copepoda: including Labidocera fluviatilis Dahl Acartia lilljeborgi Giesbr. Many Acartia giesbrechti Dahl. Chaetognatha : Sagitta sp., small : ; : Many Coelenterata: Scyphomedusae : Stomolophus meleagris, young, . 3 * Such numbers of animals as are given in this paper must be regarded only as indications of relative abundance. 134 THE “ROSAURA” EXPEDITION The bulk of the catch consisted of zoeas, copepods and Sagittae. The presence of the first two of these groups suggests that some planktonic plant-food must have been present in spite of the low level of under-water illumination. Station 39: At the junction of the Barima and Arouka Rivers, 2 miles above Station 38, 10.xii.37 The salinity on the ebb was 8-3%,. The river was remarkably full of fish, many of which must have weighed several pounds. The plankton was here sampled with a 2-m. stramin net (about 16 meshes to the linear inch), streamed from the ship while at anchor. The catch was large but contained no zoeas, copepods or Sagittae ; these would all have passed through the meshes of the net. The following forms were caught : Pisces: Engraulidae: . ; . >80 Ariidae : Tachysurus paris (Ganthes) 301 mm. . : I Sciaenidae : Stellifer stellifer (Bloch), 10-35 mm. . c 15 Gobiidae: post-larval,27 mm. . ; : 5 ; I Soleidae: Achirus sp.,23 mm. . : c 0 I Cynoglossidae : Symphurus sp., 33 mm. : : ; I Mugilidae: Mugilsp.,40mm. . : vi Tetrodontidae : Colomesus psittacus (Schn. Jee zomm. ~ )s-L00 Also several larval fish, unidentified. Crustacea: Decapoda: Xiphopenaeus kroyert (Heller), small. 2 Penaeus sp., very small : : . Many Sergestidae, young. 6 2 ; 2 Palaemon sp... : : . Several Prawns, not dereeonnable ‘ ‘ : 9s Amphipoda: Pseudoceradocus lutzi Schoemaker 5 I Isopoda: . - : 2 ; : 5 Coelenterata : Scyphomedusae: Stomolophus meleagris, adults . 5 Station 40: Demerara River at Georgetown, B.G., 13.xi1.37 The river is here about a mile wide and the ship lay only one mile from the open, though very shallow, sea. (There is a depth of only about 8 metres 15 kilometres off shore). The tide ebbs and flows at about 5-6 knots, and on the ebb the salinity was 13:9%,. The water was so turbid that if one’s hand was immersed to the wrist the fingers were invisible. The tidal current in 7 metres of water no doubt keeps the water well stirred right down to the bottom. This is suggested also if we compare the extinction coefficients in the Demerara river with those in the neighbouring Barima. Whereas the Barima is strongly coloured brown and shows much the greater absorption in the blue and violet, the Demerara shows a more THE “ROSAURA” EXPEDITION 135 uniform absorption through the spectrum, suggesting that a non-selective absorption due to suspended detritus is super-imposed on selective absorption due to brown pigments in true or colloidal solution in the water. (The tidal current in the Barima is less than half that in the Demerara.) The very rapid extinction of under-water illumination must confine active photosynthesis to a very shallow layer. Jenkin (1937) found that the compensation point for Coscinodiscus excentricus lay at about 360 lux. This is a mean figure for 24 hours. For the hours of daylight it would be considerably higher. If we take 140 kilolux as a round figure for midday tropical daylight (Atkins, Ball & Poole, 1937) the daylight factor for Coscinodiscus would exceed °-3° =0:260%. The value 1% for the daylight factor over 24 hours is therefore about as near as we can get at present. Retlection losses at the surface may also reduce the incident light by 4 to 15% according to the degree of roughness. Moreover land plants, and no doubt marine algal also, differ greatly in their response to light. Even so, under the most favourable conditions the depth of the layer of active photosynthesis could not have exceeded one-third of a metre, but nevertheless, the river was full of algae in globular masses up to about an inch in diameter and occurring every few inches; they were covered with brown detritus but appeared to be healthy, and were certainly not confined to the surface. The plankton was sampled with the silk net, which soon became completely clogged with these algae so that the catch was small, though it was probably fairly Tepresentative qualitatively. It contained : Pisces: Engraulidae . : : : ‘ < ‘ ‘ 2 Unidentified . i : : : ‘i ; : I Crustacea: Decapoda: Penaeus sp., young . : : . A few Cumacea . ; 2 : : 3 : : 2 Amphipoda: small : c c : : : I Copepoda : including Acartia lilljeborgi Giesbr. . Many Chaetognatha: Sagitta sp., small c : : F : 8 Coelenterata : Scyphomedusae : Stomolophus meleagris, young . 3 Ctenophora. : : : ; ; , T Here again the presence of copepods seems to indicate unicellular plants on which they feed, this in spite of the great opacity of the water. The two cumaceans must represent a bottom fauna living in perpetual darkness. The occurrence of a ctenophore in water with a salinity as low as 13-9%, is perhaps worth noting. The abundance of living organisms in these Guiana estuaries is much greater than would be anticipated from the conditions of illumination, and a quantitative ecological study should prove of considerable interest. Many marine animals grow to a large size in the ocean abyss far below the region of active photosynthesis. They depend on the rain of food from the illuminated waters above. Similarly many of the animals described above from tropical rivers must have sought their food in the vegetable and animal detritus brought down from the rain forest and from more transparent inland waters. 136 THE “ROSAURA” EXPEDITION 3B. GAMBIA RIVER, WEST AFRICA Station 47: Nianimaru, 2.1.38 Nianimaru, where the river is nearly a kilometre wide, lies about 200 km. from the mouth of the Gambia river, which remains tidal for yet another 210 km. Even so the surface water was perfectly fresh, with a salinity of 0-00%, just after high- water. The plankton, sampled with the silk net on the ebb, was of a fresh-water type, containing many reproducing cladocera, Bosmina coregoni, but nothing else. The 2-foot dredge, shot in g metres, came up full of mud devoid of anything living. The spectral absorption of the water was uniform and in no way remarkable. The depth of 1% illumination lay at 94 cm. so that the conditions for photosynthesis were much more favourable than in the Guiana rivers investigated, where the corresponding depths ranged from 22 to 40 cm. (see Table III). TABLE III.—Depth in metres at which total visible radiation was reduced to 1% of that incident on the surface. No correction for surface loss Station Depth No. Place m, 38B < Morowhana, Barima River A 0-40 39 o Jct. of Barima and Arouka Rivers : 0-31 40 4 Demerara River, Georgetown a 0:22 47B iz Nianimaru, Gambia River : 0-94 Neal Point Buoy . 0°3-2°2* * The lower figure applies only for short periods near low water during wet weather in winter. For much of the year at this turbid station the depth of 1% illumination is likely to exceed 1-5 m. TABLE IV.—Depth in metres at which daylight factors of 80% and 60% would be found (corresponding to the zonal boundaries of Cooper & Milne (1939, revision of Table V, p. 395) Zonal boundary . 4 : : ns Il 5 II-III Daylight factor . : 5 : : 80% : 60% Guiana rivers : Morowhana, Barima R. . 2 c 0-0155 0+035 Jct. of Barima and Arouka . : oO-orl 5 0:°027 Demerara, Georgetown . 5 A 0-010 : 0:024 Gambia River, Nianimaru . : ° 0-040 a 0-100 English estuary : Plymouth Sound . 4 5 : 0-180 6 0+356 Hamoaze Buoys Nos. 4, 7 & 11 0-08—0-10 © 0+ 18-0-25 Station 48: Tendeba, 3.1.38. Tendeba is some 120 km. below Nianimaru and about 80 km. from the mouth of the river. The surface salinity was 2:63%, one hour after high-water. The plankton was sampled with the stramin net and appeared to be not markedly THE “ROSAURA” EXPEDITION 137 more numerous than and definitely not so varied as in the much darker waters of the Barima river. The catch contained : Pisces: Clupeidae: Pellonula afzeliusi Johnels . . : >200 3 Ilisha africana (Bloch) 115-130 mm. : 2 Sciaenidae : Johnius elongatus (Bowdich) 58-65 mm. . 2 Atherinidae: Larvae, 25 Crustacea: Decapoda: Callinectes latimanus Rathbun, - eae I Penaeus sp.,immature. . . Many Crangonidae, immature . 5 5 ne Mysidacea i 4 ‘ : ‘ , . Abundant Isopoda . ; é : 6 ‘ : : it Nematoda é : : : : : c . . I 1 Several other crabs, presumably of this species and up to Io cm. across, were seen at the surface drifting with the tide. No copepods or Sagittae were caught, but they may have passed through the meshes of the net. The plankton at this station was remarkably colourless. We are indebted for the names of species to the following specialists: Dr. I. Gordon (Decapods and Amphipods), Dr. J. Harding (Copepods), Lieut.-Col. W. P. C. Tenison, D.S.O., and Mr. D. W. Tucker, (Fish) and Capt. A. K. Totton (Coelenterates) ; also to Dr. W. R. G. Atkins, F. R. S., for discussion of the conditions of illumination and photosynthesis. SUMMARY Waters from the tidal estuaries of the Rivers Barima and Demerara in British Guiana show extremely heavy absorption of light of all wave-lengths; never- theless, in spite of the unfavourable conditions for photosynthesis, the zooplankton is both varied and abundant. These observations in rivers of British Guiana are compared with similar ones made in the Gambia River, West Africa. REFERENCES Atkins, W. R. G., Batt, N. G., & Poort, H. H. 1937. The photo-electric measurement of the diurnal variations in daylight in temperate and tropical regions. Proc. Roy. Soc., ser. A, 160, No. 903, pp. 526-539. Cooper, L. H. N., & Mitne, A. 1938. The ecology of the Tamar Estuary. II. Under- water illumination. Journ. Mar. Biol. Assoc. 22 : 509-528. 1939. The ecology of the Tamar Estuary. V. Under-water illumination. Revi- sion of data for red light. Ibid. 23 : 391-396. Jenxin, P. M. 1937. Oxygen production by the diatom Coscinodiscus excentricus Ebr. in telation to submarine illumination in the English Channel. Ibid. 22 : 301-343. Moore, J. G. 1947. The determination of the depths and extinction coefficients of shallow water by air photography using colour filters. Phil. Trans. Roy. Soc., ser. A, 240 : 163- 217. ca - _ : . ‘ / aa a e > ‘ < c 7 4 a i a | 5 ey _ = 1 ae 1 ie _ eG A ee = ie 3. THE ECHINODERMATA By D. DILWYN JOHN anp AILSA M. CLARK With Plate 6 and Text-figs. 1-12. THE collection of Echinoderms taken by the “‘ Rosaura’”’ includes specimens from off Greenland and north-east America, as well as from the West Indies and one station off the north-west coast of Africa. Some of the Echinoids were identified by the late Dr. Th. Mortensen before the war. The work was continued by one of us (D. D. J.), who described those species which are new and commented on various other species; the remarks about the new species and the remaining specimens, the preparation of the figures and the final compilation were done by the other (A. M. C.). The species taken are as follows (those marked with an asterisk being mentioned in the text) : Number Name Station of Asteroidea specimens Ctenodiscus crispatus (Retzius) . . 11 (Newfoundland) and 6 (S.W. Green- 54 land) Astropecten marginatus Gray c - 36 (British Guiana) . cC 5 3 Astropecten antillensis Littken . + 35 (off R. Orinoco) 2 : C c 2 Astropecten articulatus (Say) i Aei5' e ES : 6 3 6 I *Tethyaster vestitus (Say) 5 c cera 5 i FP I *Luidia rosauraesp.n. . 5 2 35) pS we ¢ 3 *Luidia barimae sp. n. : 35) Te as 2 Luidia senegalensis (Lamarck) . - 36 (British Guiana) 5 6 I Luidia clathrata (Say) c 6 . 22 (British Honduras) and 37 (Brit. 2 Guiana) Anthenoides piercei Perrier 6 - 35 (off R. Orinoco) and 36 (Brit. Guiana) 4 Nymphaster arenatus (Perrier) . . 26 (Brit. Honduras) and 49 (off N.W. 3 Africa) Goniaster cuspidatus Gray . : . 26 (Brit. Honduras) and 35 (off R. 2 Orinoco) *Ceramastey granularis forma balteatus 49 (off N.W. Africa) . oO 0 . 9 (Sladen) Echinaster sentus (Say) 5 35 (off R. Orinoco) : c : I Henricia sanguinolenta (O. F. Miller) . to (South Labrador) . 0 : * I Crossaster papposus (Linnaeus) . eel - re and 11 (Newfound- land) *Brisingella coronata (G. O. Sars) . . 49 (off N.W. Africa) . 0 : - I *Brisingidae—fragment . 26 (British Honduras) . 4 - lIarm Zoroaster ackleyi Perrier . i a me : - e P ii *Stephanasterias albula (Stimpson) . 6 (S.W. Greenland) 6 fs S : 2 Leptasterias polaris (Miller & Troschel) 9 (South Labrador) . 3 é A I 140 THE Name Ophiuroidea Astrophyton muricatum (Lamarck) Astroporpa annulata Liitken Ophiacantha pentacrinus Liitken . Ophiacantha metallacta H. L. Clark Ophiacantha bidentata (Retzius) . Ophiomitra valida Lyman . *Ophioplinthaca grenadensis a n. Ophioplinthaca incisa (Lyman) Amphiura sundevalli (M. & Tr.) . *Amphipholis gracillima (Stimpson) Ophiactis savignyt (M. & Tr.) *Ophiactis abyssicola (Sars) . Ophiopholis aculeata (Linnaeus) . *Ophionereis dolabriformis sp. n. Ophiocoma echinata (Lamarck) Ophioderma brevicaudum Liitken *Amphiophiura metabula H. L. Clark Ophiocten sericeum (Forbes) Stegophiura nodosa (Liitken) *Ophiomusium validum Ljungman Ophiolepis elegans Liitken . Ophiura sarsi Liitken *Ophiernus adspersum Lyman Echinoidea Tretocidaris bartletti (A. Agassiz) Salenocidaris varispina A. Ag. Plesiodiadema antillarum (A. Ag.) Spevosoma grimaldii Koehler Calveriosoma hystrix (W. Thomson) Phormosoma placenta W. Thomson) Tripneustes ventricosus (Lamarck) Strongylocentrotus drébachiensis (O. F. Miiller) Holothuroidea Holothuria mexicana Ludwig Mesothuria gargantua Deichmann Phyllophorus pellucidus (Fleming) Chiridota laevis (Fabricius) Crinoidea Tropiometra carinata (Lamarck) . “ROSAURA” 34 34 I 22 a EXPEDITION Station (off R. Orinoco) ” ” (Grenada) (Newfoundland) (Grenada) A and 26 (Brit. Honduras) (Greenland) : 2 (British Honduras) 1 (Mosquito Bank, W.I.) oe N.W. Africa). 4, 6 (Greenland) and 10 (s. Dapeaded) 5 (off R. Orinoco) 16 37 34 1 Io 26 31 (Grand Cayman Id.) (British Guiana) (Grenada) and 26 (Brit. Honduras) (Greenland) 4 6 (S. Labrador) (British Honduras) (Mosquito Bank, W.I.) —. : - Greenland) and 11 (Newfoundland) . ( 4 (Grenada) and 26 (British Honduras) . 5 (off R. Orinoco) 6 (Brit. Honduras) and 49 (off N.W. 26 49 49 49 Africa) (Brit. Honduras) and 49 (off N.W. Africa) (off N.W. Africa) ” ” ” ” 8 (Honduras) 16 26 6 ye 37 5 (Greenland), 9, 10 (South Labrador) and 12 (Nova Scotia) (Grand Cayman Id.) (British Honduras) (Greenland) and 6 (Greenland) (British Guiana) 5 5 5 - Number of specimens HOHMNHWDN OHA iy wre - Oo fo) DAunwnt AON WN Hw a nN NU bHAO c. 30 10 THE “ROSAURA” EXPEDITION 141 Class ASTEROIDEA Family ASTROPECTINIDAE Tethyaster vestitus (Say) Asterias vestita Say, 1825 : 143. Sideriaster ? vestitus, Verrill, 1915 : 193. St. 35. 9° 25’ N., 59° 52’ W. (off the mouth of the river Orinoco) ; otter trawl ; 86m. One specimen. R= 70 mm., r = 16 mm. R/r = 4:3/1. This name was eventually arrived at after a great deal of correspondence with Mr. Austin Clark of Washington. He has recently received a huge Astropectinid from off North Carolina which answers very well to Say’s description of A sterias vestita as far as that goes. This specimen has R = 250 mm., with a diameter of about 18 inches, probably as, compared with one of 14 inches for Say’s lost type from Cape May, New Jersey, and it almost certainly represents the same species. A notable feature of this form is the development of large spatular ventro-lateral and infero-marginal spines, similar to, but less well-developed than, those of Archaster magnificus Bell from St. Helena in the south-east Atlantic. One of Bell’s types with R = 215 mm., has these spines so large as to overlap, producing a very “‘shaggy”’ appearance, each spine being about 6 mm. long, whereas in the even larger specimen from North Carolina these spines are only about 4mm. long. There seems to be no other significant difference between the two forms. Caso (1947 : 225) appears to think that magnificus is synonymous with Sideriaster grandis Verrill from the Gulf of Mexico. This opinion was shared by Mr. Clark and Dr. W. K. Fisher, but on the evidence of a number of specimens from the type locality Mr. Clark now concludes that grandis also is distinct from both magnificus and vestitus, being differentiated by the pointed (rather than truncated) form of the ventro-lateral and infero-marginal spines, developed only at a late stage (when R is more than 150 mm.) as in the latter. This is coupled with other features such as the relatively shorter dorsal paxillae in grandis. The specimen taken by the “‘ Rosaura’”’ is too young for the ventro-lateral spines to have developed, so that it is not absolutely certain whether it represents vestitus or grandis. However, Mr. Clark has a large specimen with R = 160 mm. from Porto Rico which has small blunt ventro-lateral spines like vestitus. He therefore believes that vestitus occurs in the West Indies as well as up the East coast of the United States as far as New Jersey while grandis is limited to the Gulf of Mexico. This is similar to the distribution of A stropecten articulatus (Say), which extends from North Carolina (? New Jersey) south to Uruguay with a subspecies valencienni in the Gulf of Mexico, as pointed out by Mr. Clark. Whether the two forms of Tethy- aster are specifically or subspecifically distinct is controversial. As for the generic name used here, it was finally decided that Tethyaster Sladen with genotype Astervias subinermis Philippi embraces all the species formerly inclu- ded in Sideriaster Verrill and Moiraster Sladen as well as Anthosticte Fisher. This 142 THE “ROSAURA” EXPEDITION amalgamation will be discussed in more detail in a separate paper on Tethyaster by A. H. Clark and A. M. Clark. It is sufficient here to list the species included in the genus. subinermis (Philippi) ceevEe) . Mediterranean, Bay of Biscay to Mauritania. vestitus (Say) : 5 - New Jersey to Venezuela. grandis (Verrill) . 5 : . St. Helena and Ascension. magnifica (Bell) . : i . Gulf of Mexico. aulophora (Fisher) ¢ : . Philippines. canaliculatus (A. H. Clark) . . Gulf of California. gigas (Caso) (possibly only the Gulf of California. large form of canaliculatus) pacei (Mortensen) - C - South Africa. Family LurpmDar Luidia rosaurae sp. n. (Pl. 6 fig. 1; Text-figs. 1 and 2) St. 35. 9° 25’ N., 59° 52’ W. (off the mouth of the river Orinoco) ; otter trawl ; 86m. Three specimens. Dracnosis. A species of Luidia with five rays belonging to the Alternata group of Déderlein ; no enlarged dorsal spines present ; the paxillae of the two outermost rows larger than those of the supero-marginal series ; pedicellariae present on many of the dorsal paxillae particularly the lateral ones, as well as on the infero-marginal plates ; four adambulacral spines on each plate, the two outermost arranged in a line parallel to the furrow ; three- or four-valved pedicellariae on the ventro-lateral plates of many segments, especially proximally ; one large marginal spine on each infero-marginal plate with small appressed ones on the ventral surface of the plate. DESCRIPTION. The three specimens are all rather broken. Each has five rays. In the only one which has a complete and original (i.e., not regenerated) arm, R is about 53 mm. and r is 6-5 mm. R/r = 8/t. The three outer rows of paxillae on the arms are larger than the remainder; the paxillae of the second and third rows are larger than those of the first. The paxillae of the first (ie, supero-marginal) series, are rectangular, being slightly elongated. Those in the proximal part of the arm consist of 25-30 spinules, five to eight of which are in the centre, the remainder around the edge. In the centre of most there is also a small bivalved pedicellaria. The paxillae of the second and third rows are rectangular, being slightly longer in a transverse than a longitudinal direction ; those shown in fig. 2 are foreshortened. At the base of the arm they consist of about 30 to 38 spinules, of which 14 to 16 are in the centre. The central spinules are heavier and blunter than the peripheral ones. A small number of these paxillae have two- valved pedicellariae among the central spinules. The paxillae occupying the centre of the arm are smaller and roughly circular. At the base of the arm there are eight or nine rows of them, of which those of the An THE “ROSAURA” EXPEDITION 143 outer row on either side are slightly larger than the remainder. They consist of a central group of 14 to 16 spinules, of which those nearest the centre are heavy and blunt, and a peripheral circle of about the same number of spinules. In the paxillae in the middle of the arms the contrast between the central and peripheral spinules is more marked. There are up to eight central spinules, of which some or all are considerably heavier than the peripheral ones, of which there are about ten. These paxillae are very variable. They do not form regular rows. Those at the centre of the disc are of the same nature but smaller and more crowded. The innermost adambulacral spine arises within the furrow. It is long and com- pressed and strongly recurved. A second and similar spine, not quite so recurved, Fic. 1. Lwuidia rosaurae sp.n. Type. Ventral view of one side of two adjacent joints, that on the left having been denuded with sodium hypochlorite. (The spines are displaced sideways to avoid foreshortening as much as possible; the arrow points towards the mouth.) occurs just outside it. External to this are two slightly shorter spines standing side by side ; they are straight and compressed and taper from a wide base to a narrow tip. The outer part of the adambulacral plate is partly overlapped by the ventro- lateral plate which frequently carries a large and conspicuous three-or four-valved pedicellaria as well as several spinules. The pedicellariae are most numerous near the interbrachial angles. The underside of each infero-marginal plate is raised into a strong ridge, the adjacent ridges being separated by deep grooves. The outer surfaces of the ridges are occupied by flattened spinelets of medium but varying lengths; fine spinules arise from the edges of the ridges. Near the outer edge there are often one or two pedicellariae, smaller than those of the ventro-lateral plates, of two, or more rarely three, valves. One strong, somewhat flattened, tapering spine nearly 2 mm. long 144 THE “ROSAURA” EXPEDITION arises from near the upper edge of each plate forming a regular border to the arm. There is a paxilla-like group of spinules with one or two bi-valved pedicellariae above each spine, at least at the base of the arm. The outer parts of the arms are rather badly damaged. When freshly preserved one of the specimens was straw coloured, with a brown patch in the centre of the disc continuous with broad brown bars running a short distance down the middle of each arm; there were also brown bands across the outer parts of the arms at intervals. The other two specimens were brown, but with deeper colour in areas corresponding to the brown parts of the first specimen. (After 15 years all colour has been lost.) 2mn Fic. 2. Dorsal view of the top of two infero-marginal plates (seen in profile because of distortion in preservation causing an abrupt bend just above these plates) and the adjacent paxillae. REMARKS. This species is obviously most closely related to Luzdia scotti Bell (1917 : 8), which is described fully in the third note on Asteroids in the British Museum by A. M. Clark (1953 : 383). The types of L. scottt were taken off Rio de Janiero in 40 fathoms (73 metres). They differ from L. rosaurae chiefly in having much thicker, more stumpy spines and spinules on the ventral plates, a second enlarged marginal spine above the ambital one on the infero-marginals, and in having fewer ventro-lateral and infero-marginal pedicellariae and no dorsal ones. This last character—the abundance of pedicellariae—is probably not of sufficient value to be used alone for specific distinction but the three points taken together are enough, on the evidence of the material available, to distinguish the two forms as separate species. The second point is interesting since in the types of Lwidia doello-juradoi Bernas- coni (a synonym of L. scofti in the opinion of A.M.C.) from off Buenos Aires (Bernas- coni, 1941 and 1943) the two large marginal spines on each infero-marginal plate are said to be nearly the same size, the upper being only slightly smaller. In the types THE “ROSAURA” EXPEDITION 145 of L. scotti the upper spine averages only one-third of the length of the lower one, although occasionally it is two-thirds as long. Luidia armata Ludwig (1905: 85) from the Gulf of Panama also has much in common with Luidia rosawrae but perhaps even more with L. scofti. It has ventro- lateral pedicellariae only in the interbrachial angles, but also has dorsal pedicellariae on the supero-marginal paxillae. Figures of this and of the other related species of Luidia described from Central and South America will probably show that they cannot all be maintained as valid species. This may be true of Luidia armata Ludwig, L. /udwigi Fisher (1906 : 122 and 1911 : 113), L. scott? Bell, L. doello-juradot Bernasconi and L. rosaurae, which could represent local forms of a single species occurring on both sides of South and Central America. However, without Pacific specimens for comparison this suspicion cannot be verified. Luidia barbadensis Perrier (1881: 29) also appears to be related to L. vosaurae. It is apparently usually six-rayed, whereas the three specimens of L. rosaurae are all five-rayed. The only pedicellariae described in L. barbadensis are said to be on the adambulacral plates (probably these are actually on the ventro-lateral plates as in related species) and in the interradial areas (Verrill, 1915 : 205-7). In Luidia rosaurae they also occur on the outer paxillae of the arms dorsally and on the outer and lower surfaces of the infero-marginal plates. The infero-marginal plates of L. barbadensis bear two large spines on the margin of the ray. In L. rosaurae there is only one large marginal spine. There are also differences in the arrangement and proportions of the adambulacral spines of the two species. In the type of L. barba- densis with R = 125 mm. there are only two spines behind the furrow spine not three as in L. rosaurae, although Verrill (1915 : 207) describes an ‘“‘ Albatross ”’ specimen of L. barbadensis with four adambulacral spines arranged as in this new species. A good series of specimens of Lwidia barbadensis or detailed figures of the types are needed to define more clearly the relationships of the species. Perrier’s other six-rayed species Luidia convexiuscula (1881 : 30), is based on a specimen with R only 28mm. There appears to be no significant difference by which it can be distinguished from L. barbadensis, taking into account the very considerable size discrepancy between the types of the two species. The presence of three-valved pedicellariae only in the arm angles, only one enlarged marginal spine and the poor development of the adambulacral spines would be expected in such a young speci- men. Luidia barimae sp. n. (Pl. 6, fig. 2; Text-figs. 3 and 4) St. 35. 9° 25’ N., 59° 52’ W. (off the river Orinoco) ; otter trawl; 86 m. Two specimens. Diacnosis. A ten-rayed species of Luidia belonging to the Quinaria group of Déderlein, with all the dorsal paxillae irregularly arranged, some of them with one or more pedicellariae ; no spino-paxillae ; three adambulacral spines in a row at 146 THE “ROSAURA” EXPEDITION right angles to the furrow ; one to four ventro-lateral pedicellariae ; no furrow pedi- cellariae except on the mouth plates; infero-marginal plates extending on to the dorsal side, with four (sometimes three) large spines which alternate in position on adjacent plates. DESCRIPTION. The two specimens are ten-rayed. Most of the arms are broken off near the disc and lost. The longest arm remaining attached to a specimen is 65 mm. long ; its tip is regenerated so that its length cannot be taken as a measure- ment of R. The radius of-the disc is about 13 mm. The arms are constricted to a width of 9°5 mm. where they join the disc ; farther out they are 13 mm. wide. None of the paxillae of the arms are arranged in regular rows. Those near the margin are somewhat bigger than, but otherwise similar to, those in the centre. Fic. 3. Luidia barimae sp.n. Type. Ventral view of one side of two joints, that on the left having been denuded with sodium hypochlorite. (The arrow points towards the mouth.) Each consists of a peripheral circle of eight to twelve widely spreading spinelets and a central spinelet, a little stouter but no longer than the peripheral ones. One of the peripheral spinelets is frequently replaced by a small bi-valved pedicellaria. There are no spino-paxillae. The paxillae in the centre of the disc are small and much more crowded than those of the arms. In many the central spinelet is missing. Nearly every paxilla is sup- plied with a pedicellaria, while many have two. The pedicellaria sometimes arises from near the centre of the paxilla. The madreporite is 2-5 mm. in diameter and may be trefoil-shaped ; it arises from very near the edge of the disc. A few paxillae appear to arise from it, but in reality project from between the lobes, not from the madreporite itself. The bigger elements in the armature of the adambulacral, ventro-lateral and infero-marginal plates form one continuous transverse line from the furrow to the abactinal surface. The adambvlacral armature consists of spines and pedicellariae. THE “ROSAURA” EXPEDITION 147 There is a strongly curved furrow spine, about 3 mm. long; it is compressed to a blade-like form and its end is blunt. It is followed by another blunt compressed spine which is broader and may be longer, up to 4:5 mm. long ; it may be bent at its base but is thereafter straight. There is a third spine which is nearly as long but much narrower, being only slightly compressed. A few spinelets usually arise from near the base of this spine on the adoral side. There may be a fourth and smaller spine borne on the inner end of the ventro-lateral plate, but it is usual for these plates to carry two or three large bi-valved pedicellariae, of which the outermost, about r mm. long, is larger than the inner ones. Small spinules may arise around _the base of the pedicellariae, especially, in the proximal part of the arm, where there may even be four rather than three pedicellariae on each plate ; in the distal part Fic. 4. Dorso-lateral view of three infero-marginal plates and the adjacent paxillae. of the arm there is only one. There are no pedicellariae in the furrow as are found in L. avicularia and L. integra. The infero-marginal plates encroach strongly on the abactinal surfaces of the arms. They are raised into strong ridges. In the proximal part of the arm there are four large spines on each plate. Three are on the lateral edge, one above the other, the lowermost shorter than the second and third which may be 6 mm. long. The fourth and longest spine, up to 8 mm. long, arises from or near the inner edge of the plate on the abactinal surface. The spines of adjacent plates arise at different levels as a rule, so that the uppermost spine of one plate is based on the extreme adradial edge, while that of the next plate is somewhat inset, the lower spines being correspondingly displaced on every second plate. On each row of infero-marginal plates there are thus eight series of spines, those of any one series all arising at the same level and usually only occurring on alternate plates. On those plates where the uppermost spine is not on the extreme inner edge of the plate there may be only three spines in the distal part of the arm. Numerous spinules occur in a thick row on the proximal and distal edges of the infero-marginal plates; they point into the furrows between the plates. One or two small bi-valved pedicellariae occur on ZOOL. 2, 6. 9 148 THE “ROSAURA” EXPEDITION most of the plates, often on the innermost abactinal edge when there is no spine in that position. Pedicellariae are numerous on the jaws and in the small interradial areas. The colour of the spirit specimens is greyish-brown above, light below. REMARKS. This species appears to be most nearly related to Luidia heterozona Fisher from West Africa (of which L. mortenseni Cadenat is a synonym), although that species is included by Fisher (1940 : 268) in the Czlzaris group since its supero- marginal plates are markedly larger than the adjacent lateral paxillae. Déderlein’s genealogical tree (1920 : 223) indicates that the Czliaris group is very closely related to the subgenus Integraster of the Quinaria group (in which L. barimae seems to belong), and indeed Luidia heterozona is as much, if not more, in accord with the ten-rayed Quinaria group species such as L. integra Koehler, L. avicularia Fisher and L. moroisoana Goto as with the other species of the Ciliaris group, none of which has more than eight arms. In the subgenus /ntegraster the supero-marginal paxillae are larger and fewer in number than the adjacent rows of paxillae, even it not so conspicuously different as in L. heterozona. In L. barimae the lateral paxillae are much more irregularly arranged. The only other notable difference between L. heterozona and this new West Indian species is in the number of infero-marginal spines, of which there are commonly four on each plate in L. barimae but only two or three in L. heterozona, although the conspicuous alternation in the position of these spines is similar. The number of pedicellariae between the outermost adambulacral spine and the lowermost infero- marginal one may be as many as three in some specimens of L. heterozona according to Madsen (1950: 204), although the number is usually one. On the whole it seems better to regard Luidia heterozona as also belonging to the Quinaria group, it being the east Atlantic counterpart of L. barimae. Whether the differences between the two are less than specific, as with the east American Luzdia alternata and its subspecies numidica from West Africa, remains to be seen from a greater range of West Atlantic material, showing the variation particularly of the armature of the infero-marginal and ventro-lateral plates. The Indo-Pacific species Luidia integra, L. avicularia and L. moroisoana are easily distinguished from L. bavimae by the possession of spino-paxillae, and in having a long-jawed pedicellaria within the furrow on the adambulacral plates. Family GONIASTERIDAE Ceramaster granularis forma balteatus (Sladen) Pentagonaster balteatus Sladen, 1891: 688, pl. 25, figs. 1-5. Pentagonaster hystricis von Marenzeller, 1893 : 4, pl. 1, fig. 2, pl. 2, fig. 2; Ludwig, 1897: 179, pl. 8, fig. 2. Pentagonaster gosselini Perrier, 1894 : 399, pl. 26, fig. 4; 1896: 45; Koehler, 1909: 84, pl. 1, fig. 9. Pentagonastery kergroheni Koehler, 1896 : 63, pl. 2, figs. 8-10. Ceramaster balteatus, Mortensen, 1927 : 82, text-fig. 45. ? Pentagonaster haesitans Perrier, 1894 : 397, pl. 23, fig. 7, pl. 25, fig. 2. THE “ROSAURA” EXPEDITION 149 St.49. 28° 25’ N., 13° 34’ W. (between Fuerteventura Island and Africa) ; agassiz trawl; c.1300m. Nine small specimens. The typical north-European form of Ceramaster granularis (Retzius), from depths as shallow as 20 metres, is nearly pentagonal in outline, with the short triangular rays projecting abruptly. The R/r ratio equals c. 1-3 to 1°5/1, rarely more. The marginal plates of both series each have a bare patch, which is relatively larger in the young sea-star. There are usually three or four short, thick furrow spines and two, sometimes three, shorter spines in the second row backed by several rows of granules. The variations have been studied in detail by Grieg (1907 : 22-32, text- figs. 1 and 2). It has been reported from off Morocco by Perrier and from the east coast of North America by Sladen and Verrill. From south-west of Ireland to north-west Africa and the Mediterranean, at depths of more than 1,000 metres, is found a relatively longer-armed form with R/r averaging 1-7/1 and with the interbrachial arcs regularly curved, not angular. The marginal plates are more or less completely covered with granules and the furrow spines number from four to six on each adambulacral plate, with three spines in the second series backed by granules. Several names have been given to such specimens, of which Pentagonaster bal- teatus Sladen has priority. Sladen’s type was dredged in 750 fathoms (1,372 metres) in 51° or’ N., 11° 50’ W. (south-west of Ireland). It has all the marginals completely covered with granules, R/r = 1-7/1, and there are six furrow spines on each plate, the adoral one being inset and shorter than the others. Perrier originally described Pentagonaster gosselini as having only three or four furrow spines, but later (1896) mentions specimens with four to six adambulacral spines. The supero-marginal plates may be covered with granules or have bare patches according to Koehler (1909) ; Perrier says that their granulation is “ fugace.”’ “ Pentagonaster’’ kergroheni Koehler from the Bay of Biscay in 1,710 metres has only the distal supero-marginals with a small bare patch free of granules. R/r=1°7/ I and there are five furrow spines. Ludwig (1897) declares that P. kergroheni is a synonym of P. hystricis von Marenzeller from the Mediterranean, which may also have small bare patches on the marginal plates. This seems very likely, although the Mediterranean form apparently has the granulation more spaced out than in Atlantic specimens. However, Ludwig is mistaken in thinking that P. concinnus Sladen and P. greeni Bell might also be synonyms of P. /ystricis, for the former having only marginal granules on the dorsal plates is synonymous with Plinthaster perrieri (Sladen) as other authors have pointed out, while P. greeni does not have tabulate dorsal plates and like Ceramaster placenta (Miiller & Troschel) might be teferred to Peltaster Verrill. Ludwig also says that older specimens of P. hystricis tend to have more numerous and larger bare patches on the marginals than young ones, which is the converse of what occurs in typical granularis. However the “ Rosaura’’ specimens, the largest of which has R only 15 mm., all have bare patches on the supero-marginals and five of them on all the infero-marginals also. The other four, including the largest one, have the proximal infero-marginals completely covered with granules, and only the last two or three have a small bare patch. The R/r ratio varies between 150 THE “ROSAURA” EXPEDITION I-45 and 1-7/1, averaging 1-6. The width of the marginals also seems to vary with the length of the arms, as they are broader when the shape is more pentagonal. As for North American specimens of Ceramaster granularis, those collected by the “ Challenger’ south of Halifax, Nova Scotia, in 156 metres, differ little from the north European form except in having relatively longer arms, so that R/r averages 1-7/1. “‘ Pentagonaster”’ eximius Verrill (1895) is a synonym of Ceramaster granu- lavis, for Verrill’s conception of the latter obviously approximates more to balteatus than to the typical form, as can be seen from his comparison. It seems quite probable that a longer-rayed form somewhat similar to balteatus occurs beside the typical one on the east coast of North America, just as typical granularis is found off North Africa according to Perrier, as well as the longer-armed form. Family BRISINGIDAE St. 26. 17° 53’ N., 87° 44’ W. (off British Honduras) ; agassiz trawl; c. 900m. One damaged arm fragment consisting of 46 joints, 95 mm. long. At one end of the fragment the first and second adambulacral ossicles have, lying above and fused to them, two marginal plates, showing that this is the proximal part ofanarm. The first marginal plate on one side is the same size as that on the other. The first two pairs of ambulacral plates are shorter than the remainder, which are each about 2:25 mm. long. The soft tissue is stripped off the first seven joints, but some remains on the dorsal side between the eighth and the sixteenth joints. Five costae, consisting of light ossicles with no spines, occur in that interval, that is, one to about every two joints. Beyond the sixteenth joint the arm consists of ambulacral and adambulacral ossicles, spines and shreds of tissue. Each of the adambulacral ossicles carried a moderate-sized spine on its lower edge, a little nearer to the distal than the proximal end. Larger spines than these arise from the distal end of the outer edge of the adambulacral ossicles beyond the seventh ; they appear to occur, with slight irregularities, on alternate ossicles. None is complete, the longest remaining being about 4:5 mm. long. The skin coverings of most are torn off but they are covered with numerous small crossed pedicellariae. We cannot with confidence assign this fragment to any of the three species of the Brisingidae known from the Caribbean region: Hymenodiscus agassizi Perrier, Odinia antillensis A. H. Clark and Freyella mexicana A. H. Clark. What remains of the adambulacral spines are unlike those described for O. antillensis and F. mexi- cana. They are not unlike those of H. agassizi, and the structure of the arm is very like what Perrier describes for that species (1884 : 189, pls. 1 and 2) : the ambul- acral ossicles have an exactly similar process for articulation with the adambulacral ossicles. Fisher (1918 : 104, figs. r and 2) describes how the first marginals of H. agassizi which meet in an interbrachium are unequal in size. The first marginals of this single arm are of equal size; it may be that each met in its interbrachium a smaller first marginal. Fisher found that the abactinal integument of the rays of H. agassizi contained holothuroid plates. There are none in this specimen ; there are on the other hand a number of light costae. THE “ROSAURA” EXPEDITION 151 Brisingella coronata (G. O Sars) Brisinga coronata G. O. Sars, 1871: 5; 1875: 1. St. 49. 28° 25’ N., 13° 34’ W. (between Fuerteventura Island and Africa) ; agassiz trawl; c.1300m. One disc. The disc is of a young specimen, for it is only 12 mm. in diameter and the spines on its abactinal surface are fine and fairly far apart. There were eight arms. Family ASTERIIDAE Stephanasterias albula (Stimpson) Asteracanthion albulus Stimpson, 1853: 16, pl. I, fig. 5. Stephanasterias albula Heding, 1935 : 38. St. 6. 60° 06’ N., 45° 25’ W. (off S.W. Greenland) ; three-feet dredge ; 110 m. Two specimens. At first sight the larger specimen with R = 11 mm. appears to be five-rayed, but one of the larger arms has a very short arm coalesced with its proximal portion. The smaller specimen is six-rayed. Class OPHIUROIDEA Family OPHIACANTHIDAE Ophioplinthaca grenadensis sp. n. (Text-figs. 5-7.) St.34. 12°05’N., 61° 49’ W. (off St. George, Grenada) ; agassiz trawl; 720-800 m. One specimen. Diacnosis. A species of Ophioplinthaca with oval, medium-sized, bare, radial shields ; long, tapering, rugous disc-spines not concealing the scales (at least in a dry specimen) ; the dorsal arm-plates triangular, the first two barely in contact (when the disc diameter is 5 mm.), the rest separate ; arm-spines flattened serrulate, five in number ; five acute oral papillae each side of each mouth angle; adoral shields barely meeting interradially, but with a distal lobe separating the first lateral arm plate from the oral shield ; one large, oval, tentacle-scale throughout. DESCRIPTION. The single specimen is dried and is rather twisted. Being dry the outlines of the plates show up more than ina spirit specimen. The disc diameter Js 5 mm.; there are five arms about 27 mm. long. The disc appears to be slightly indented in the interbrachial spaces. It is covered with large scales from most, if not all, of which arise strong glassy spines about 0-5 mm. long; they are thick at the base but rapidly taper, sending off smaller spines on all sides and ending in two or three thorns. The radial shields are moderately large, naked, convex and about Imm. long. They are separated from one another by a wedge of scales bearing spines ; the wedge narrows from three scales within to one without. 152 THE “ROSAURA” EXPEDITION The first two dorsal arm-plates are just in contact with one another, but beyond the second one the plates become more and more widely separated from one another. They are of a simple triangular shape, about as wide as long, with a sharp angle within, quite straight lateral edges and a gently curved distal edge. The side-plates are large and raised into strong spine-ridges. Those of the distal segments are very long. There are five glassy arm-spines, increasing in length from the lowermost, which is as long as one segment, to the uppermost, which may be as long as two segments. They are of a characteristic shape, being flattened with each of the two blade-like edges produced into a row of teeth. The dorsal ones taper to a blunt end but the ventral spines have the end truncated, each corner being produced into a tooth. Fic. 5. Ophioplinthaca grenadensis sp.n. Type. Dorsal view of the base of an arm and part of the disk. The proximal ventral arm-plates are just separated from one another; those at the end of the arm are very widely separated. They are five-sided. The two straight or slightly concave proximal sides form an angle; the lateral sides are deeply con- cave; the distal edge is widely rounded. The proximal plates are slightly broader than long; those further out on the arm are longer than broad. The tentacle-scale is single and very large. It is leaf-like with a rounded free end on the proximal segments, a more pointed end on the distal segments. The interbrachial spaces on the ventral side are covered with large, overlapping scales without spines. The oral shields are shaped like a broad spear-head, about as wide as long. The proximal sides are slightly convex, the distal concave ; the distal angle is much more broadly rounded than the proximal. The adoral shields are tri-lobed. The long inner lobe narrows to a point which may be just in contact with its fellow. The outer lobes separate the oral shield from the first lateral arm- plate. THE “ROSAURA” EXPEDITION 153 There are five or six similar triangular oral papillae on each side of the jaw and one much larger, below the teeth, at the apex. The colour of the preserved specimen is white. REMARKS. Having the radial shields not bar-like but oval and naked, this species cannot be included in the genus Ophiacantha in the restricted sense. It seems to fall within the group of genera Ophiomitrella, Ophiophthalmus and Opmoplinthaca by virtue of the armature of the mouth-plates and the more or less distinct disc scales. It is distinguished from Ophiomitrella Verrill (1899a : 332) by the presence of distal lobes to the adoral shields interposed between the first lateral arm-plates and the oral shields. Koehler (1922 : 124) has separated off the two species which show this character from the rest of the genus Ophiomitrella by creating a new genus Ophiomelina. However, these two species have such insignificant radial shields that Fic. 6. Ventral view of an interbrachial angle and the bases of the two adjacent arms. (Some of the tentacles are drawn in.) they are obviously not congeneric with this new species. Ophiophthalmus Matsumoto has been limited and redefined by Koehler (1922 : 121), who relegated some of the species to Ophiomitrella and others to Ophiomelina. None of the species left in Ophiophthalmus has a distal lobe to the adoral shields, which instead meet widely interradially although Ophiophthalmus hylacantha (H. L. Clark 1911 : 227) does have thorny tapering disc-spines, arm-spines, radial shields, ventral arm-plates and tentacle-scales very like this new species. Ophioplinthaca Verrill (1899: 351), which grades into Ophiomitrella, includes forms with oral papillae in a simple series, a single apical papilla, relatively large bare radial shields and the disc notched interradially usually with specialized marginal scales. Except for the last character, which is indistinct in the immature specimen of O. grenadensis, it conforms quite well to the other species of Ophioplinthaca. How- ever most of these, like Ophiomitrella, have the adoral shields lacking a distal lobe between the first lateral arm-plate and the oral shield. Ophioplinthaca partita (Liitken & Mortensen 1899: 179), from the west coast of Mexico, included in this genus by H. L. Clark in 1915, differs from the rest in this respect. It seems to be 154 THE “ROSAURA” EXPEDITION congeneric with this new species which is accordingly placed in the genus Ophioplinthaca. Within the genus, Ophioplinthaca rudis (Koehler) (1922 : 142, pl. 96, fig. 1) has disc-spines which are just like those of O. grenadensis, although their occurrence is very variable and they may be completely absent, as illustrated by Koehler. From Ophioplinthaca incisa (Lyman 1883), also taken at station 34, this new species differs most obviously in having relatively smaller adoral and larger oral shields, more numerous and spaced oral papillae, single tentacle-scales on the first pores and much finer thorns on the arm-spines. Fic. 7. a, Two spinelets from the disk, and b, a short blunt lower arm-spine and a long upper one. The genus Ophioplinthaca is also characterized by the presence of broad dorsal arm-plates which are in contact in the proximal part of the arm of the adult. How- ever, the relatively small plates of O. grenadensis may also be shown by young speci- mens of the other species of comparable size. This is illustrated in Koehler’s photographs of Ophioplinthaca globata (1922), from the Philippine area. The saw-like arm-spines are similar to those of Ophiacantha hirsuta Lyman (1875) as well as to the spines of Verrill’s genus Ophiopristis although that genus is also characterized by the complex oral armature with several apical papillae. In the present state of the family Ophiocanthidae the generic positions of many species are open to some doubt and a thorough revision is needed to clear up the limits of the genera. Family OPHIACTIDAE Ophiactis abyssicola (Sars) Amphiuva abyssicola Sars, 1861: 18, pl. 2, figs. 7-12. Ophiactis abyssicola, Ljungman, 1867 : 324. St. 49. 28° 25’ N., 13° 34’ W. (between Fuerteventura Island and Africa) ; agassiz trawl; c. 1,300 m. Ten specimens. A note by Mr. Colman on the original label describes the arms as scarlet and the body slaty-blue in life. Most of the specimens were on sponges, but it is not known whether this is a natural association or one which came about in the trawl. The disc of the largest is 7 mm. in diameter. THE “ROSAURA” EXPEDITION 155 Family AMPHIURIDAE Amphipholis gracillima (Stimpson) (Text-fig. 8). Ophiolepis gracillima Stimpson, 1852 : 224. Amphipholis gracillima, Ljungman, 1867 : 314. St. 22. 17° 28’ N., 88° 11’ W. (Belize Harbour, British Honduras) ; two-feet dredge; 6m. Six specimens. Most of Stimpson’s original specimens had lost their discs, as three of the present specimens have done. None has complete arms but many long fragments remain, and where they are fractured they show no sign of tapering to an end. al 2mm Fic. 8. Amphipholis gracillima (Stimpson). Ventral view of a middle arm segment. The species does not appear to have been described as possessing tentacle-scales. The present specimens have two, a long narrow inner scale reaching slightly further than the segment to which it belongs, and a small outer scale. One specimen sent to the late Dr. H. L. Clark of the Museum of Comparative Zoology, was compared by him with a specimen of the same size from Tobago and with one of the types of Amphipholis gracillima. He thought there was no doubt that the “‘ Rosaura’’ specimens belong to this species, but their tentacle-scales, particularly the inner, are bigger than in the type and the Tobago specimen. They are accordingly figured here. The specimens retain in spirit bars of a rich reddish colour at the distal end of each arm-segment on the underside. Family OPHIOCHITONIDAE Ophionereis dolabriformis' sp. n.* (Text-figs. 9-11). St. 35. 9° 25’ N., 59° 52’ W. (off the mouth of the river Orinoco) ; otter trawl ; 86m. One specimen. Diacnosis. A species of Ophionereis with very fine disc-scales, becoming a little larger around the radial shields ; supplementary dorsal arm-plates very small 1 dolabriformis = axe-shaped, referring to the ventral arm-plates. 2 Owing to delay in publication of this paper a brief description of this species has already appeared in print. (A. M. Clark, 1953, A Revision of the genus Ophionereis. Proc. Zool. Soc. Lond. 123: 65-94.) 156 THE “ROSAURA” EXPEDITION and limited to the distal half of the joint, even on the proximal segments ; three long, tapering arm-spines which vary irregularly in relative length, the uppermost usually the longest, sometimes exceeding one-and-a-half times the length of the corresponding joint, the middle one sometimes the shortest but often all three about the same length ; adoral shields barely meeting interradially and also narrow outwardly, being overlapped by the oral shield which, at its widest part just reaches the first lateral arm-plate on each side; disc coloured with very compact dark reticulations. DESCRIPTION. The specimen has been dried and the disc has perforated centrally so that the teeth are visible from the dorsal side through the gap. It has contracted Fic. 9. Ophionereis dolabriformis sp.n. Type. Dorsal view of the base of an arm and part of the disk. considerably in the interradial areas so that the diameter is hard to measure ; it is about 8 mm. All the arms are broken, the longest remaining being about 40 mm. in length and still quite stout at the end of that distance, the total length being probably in the region of 70 mm. The disc scaling is fine but a little coarser around the radial shields, a few of the mid-radial scales being enlarged also. On the ventral side it is fairly uniform up to the genital slits, which lack papillae. The scaling extends out on to the dorsal side of the arm-bases for several segments. The dorsal arm-plates are hexagonal in shape, the widest part being about the middle. The two latero-proximal sides are in contact with the lateral arm-plates, while the latero-distal sides (which usually curve round into the distal side) are THE “ROSAURA” EXPEDITION 157 bordered by the short supplementary arm-plates, which are triangular with a curved distal edge. The lateral arm-plates bear three long, needle-like, tapering spines of varying relative length; they may all be equal, or the uppermost may be longest and the middle one shortest, or the uppermost may be the shortest and the other two about equal. Their maximum length exceeds one-and-a-half times the length of the corresponding joint. 2mm Fic. 10. Ventral view of the base of an arm and the adjacent interradial areas. The ventral arm-plates are shaped like a battle-axe head, with the distal edge very convex and a little wider at the distal end of the tentacle-scales than proximal to them. (This form might also be interpreted as “ bell-shaped”’). After the first few the plates are a little longer than wide. There is one large, oval tentacle-scale completely covering each pore. The oral shields are broadly spear-head-shaped, the widest part being towards the distal end and the two nearly straight proximal sides meet at an angle, while the shorter distal sides are each slightly concave across the head of the genital slit but form a blunt distal angle. The adoral shields are not very large and are com- pletely overlapped by the oral shields at their widest point. Internally they just meet inside all the oral shields except the madreporite, which is a little larger and is 158 THE “ROSAURA” EXPEDITION swollen. There are four or five oral papillae on each side, the outermost being the largest when they are four; on some angles there is a small apical papilla, but on others the lowest tooth is clearly visible. Beyond the outermost papilla is an oral tentacle-scale. The disc is marked with very close dark reticulations, much more compact than those of O. reticulata. On the arms at intervals of three or four segments there are brownish-purple bands, each extending for one-and-a-half segments. Besides Fic. 11. Lateral view of part of an arm. these there are fainter markings elsewhere on the dorsal sides of the arms, including an ill-defined longitudinal line down the centre of the arm. REMARKS. The other three West Indian species of Ophionereis are O. reticulata (Say), which is very common, QO. squamulosa Koehler, known from St Thomas, Parahiba in Brazil and common at Tobago and the Tortugas according to Dr. H. L. Clark, and finally O. olivacea H. L. Clark, known only from two specimens, from Porto Rico and Florida. Dr. Clark has given a key to the identification of these (1933 : 39). Ophionereis dolabriformis differs from the first two named above by its small supplementary arm-plates and from all three by the needle-like arm-spines and the coloration of the disc. Family OPHIOLEPIDAE Amphiophiura metabula H. L. Clark Amphiophiura metabula H. L. Clark, 1915 : 311, pl. 17, figs. 1-3. St. 26. 17° 53’ N. 87° 44’ W. (off British Honduras) ; agassiz trawl; c. 900 m. Four specimens. St. 34. 12° 05’ N., 61° 49’ W. (off St. George, Grenada) ; agassiz trawl , 720-800 m. Two specimens. The largest have discs 8 mm. in diameter. They agree in every way with Clark’s photographs except that the arm-spines are only four, or on the proximal segments five, in number, and that they are widely separated and of minute size—less conspicuous than in Clark’s picture. Where there are four, the lowermost is close against the tentacle-scales, the uppermost close to the dorsal plate ; the other two are so placed that the distance between all four spines is equal, or so that the distance THE “ROSAURA” EXPEDITION 159 between the middle two is slightly greater than that between the first and second and between the third and fourth. Clark’s fig. 3 shows at least six arm-spines, close against one another and occupying the whole distal edge of the lateral plate. There are two specimens in the British Museum collection which were labelled Ophtoglypha variabilis from “ Blake” station 227, 573 fathoms (1,048 metres) off St. Vincent. They of the same size as the “ Rosaura’’ specimens and agree with them in every way, including the arm-spines. ; Ophiomusium validum Ljungman Ophiomusium validum Ljungman, 1871 : 618. St. 26. 17° 53’ N., 87° 44’ W. (off British Honduras) ; agassiz trawl; c. goo m. Sixty-five specimens. The diameter of the disc is up to 10 mm., which is greater than Lyman gives. There is a far larger specimen, diameter 17 mm. among three in the British Museum from “ Blake’ station 238, 127 fathoms (232 metres) in the Grenadines. Lyman (1878, pl. 1, fig. 9) shows four equally spaced arm-spines. Most of the present specimens and some of the “ Challenger”’ ones in the British Museum collection have only three spines, the third (the highest) a greater distance from the second than the second is from the first. Some have four spines at the base of the arms, as the “ Blake” specimens do; the two lowermost spines are closer together than the second and third and the third and fourth. Minute ventral arm-plates are present for a short distance beyond the third seg- ment ; they persist a long way out in the large ‘“‘ Blake ”’ specimens. The radial shields are usually contiguous without, but they are sometimes com- pletely separated as in the large ‘‘ Blake’ specimens. Family OPHIOLEUCIDAE Ophiernus adspersus Lyman (Text-fig. 12) Ophiernus adspersus Lyman, 1883 : 236, pl. 3, figs. 19-21. St. 26. 17° 53’ N., 87° 44’ W. (off British Honduras) ; agassiz trawl; c. 900 m. One specimen. St. 34. 12°05’N., 61° 49’ W. (off St. George, Grenada) ; agassiz trawl, 720-800 m. Five specimens. The six rather broken specimens from this collection and two co-types in the British Museum from “ Blake” station 185, Dominica, 333 fathoms (610 metres), all show a character which has not been described in this species or in any other of the genus. A single closely placed row of fine glassy bristles arises from the distal 160 THE “ROSAURA” EXPEDITION Fic. 12. Ophiernus adspersus Lyman. Dorsal view of part of an arm to show the bristles on the lateral arm-plates. edges of the side plates on the dorsal side of the arm. In the proximal part of the arm they are slightly longer than an arm joint ; farther out on the arm they are longer. REFERENCES BELL, F. J. 1881. Description of a new species of the Genus Archastey from St. Helena. Ann. Mag. Nat. Hist. (5) 8: 440-441. —— 1917. Brit. Antarctic (‘‘ Terra Nova”’) Exped., 1910, Nat. Hist. Rep. Zool., Echinoderma, 4: 1-10, pls. 1, 2. BERNASCONI, I. 1941. Dos nuevas especies argentinas de “‘ Luidia.”” Physis. B. Ares, 19 : 117-118. 1943. Los Asteroideos sudamericanos de la familia “ Luidiidae.” An. Mus. argent. B. Aives, 41: 1-20, pls. 1-5. Caso, M. E. 1947. Descripcion de una nueva especie del genero Moiraster de Santa Rosalia, Golfo de California. An. Inst, Biol. Univ. Mex. 18 : 225-231, figs. I—5. CLaRK, A.M. Notes on Asteroids in the British Museum (Natural History) III: Luidia. Bull. Brit. Mus. (Nat. Hist.), 1 (12) : 379-396. Cuark, H. L. 1911. North Pacific Ophiurans in the collection of the United States National Museum. Bull. U.S. nat. Mus. 75: xvi, 1-302, text-figs. 1-144. 1915. Catalogue of recent Ophiurans. Mem. Mus. comp. Zool. Harv. 30: 165-376, pls. 1-20. 1917. Reports on the scientific results of the expedition to the Tropical Pacific on the “ Albatross,’ 1899-1900, ... . 18. Reports on the scientific results of the expedition to the Eastern Tropical Pacific ... by the ... ‘‘ Albatross,” 1904- 1905, ... . 30. Ophiuroidea. Bull. Mus. comp. Zool. Harv. 61 : 429-453, pls. 1-5. — 1933. A handbook of the Littoral Echinoderms of Porto Rico and the other West Indian islands. Sci. Surv. P. Rico, 16 (1) : 1-147, I text-fig., 7 pls. D6pDERLEIN, L. 1920. Die Asteriden der Siboga-Expedition. 2. Die Gattung Luidia und ihre Stammesgeschichte. Siboga-Exped. 46b : 193-203, text-figs. 1-5, pls. 18-20. FisHer, W. K. 1906. New Starfishes from the Pacific Coast of North America. Proc. Wash. Acad. Sci. 8 : 111-139. — 1g11. Asteroidea of the North Pacific and adjacent waters. 1. Bull, U.S. nat. Mus. 76 : vi, 1-419, pls. 1-122. 1918. Notes on Asteroidea. 2. Ann. Mag. Nat. Hist. (9) 11: 103-111. 1919. Starfishes of the Philippine Seas and adjacent waters. Bull. U.S. nat. Mus. 100 (3) : xi, I-712, pls. 1-156. — 1940. Asteroidea. ‘‘ Discovery’ Rep. 20 : 69-306, figs. A—M, pls. I-23. GrizG, J. A. 1907. Echinodermen von dem norwegischen Fishereidampfer ‘‘ Michael Sars” in den Jahren 1900-1903 gesammelt. 3. Asteroidea. Bergens Mus. Aarb. 1906 (13): 1-87, text-figs. 1-10, pls. 1, 2. THE “ROSAURA” EXPEDITION 161 Hepinc, S. 1935. The Scoresby Sound Committee’s 2nd East Greenland Expedition in 1932 to King Christian IX’s Land. Echinoderms. Medd. Gronland. 104(13) : 1-68, text-figs. I-28, maps I-3, pls. 1, 2. KoEHLER, R. 18096. Résultats scientifiques de la Campagne du “ Caudan” dans le Golfe de Gascogne. Echinodermes. Ann. Univ. Lyon, 26 : 33-127, pls. 1-4. 1908. Astéries, Ophiures et Echinides de l’Expedition Antarctique Nationale Ecossaise. Trans. roy. Soc. Edinb. 46 : 529-649, pls. 1-16. — 1909. Echinodermes provenant de campagnes du yacht “ Princesse-Alice,” Résult. Camp. sci. Monaco, 34 : 1-317, pls. 1-32. — 1922. Ophiurans of the Philippine Seas and adjacent waters. Bull. U.S. nat. Mus. 100(5) : x, 1-486, pls. 1-103. Lyuncman, A. V. 1867. Ophiuroidea viventia huc usque cognita enumerat. Ofvers. VetenskAkad. Forh. Stockh. 23 : 303-336. 1871. Férteck Gfver uti Vestindien af Dr. A. Goes samt under korvetten Josefinas Exped. i Atlantiska Oceanen samlade Ophiurider. Ofvers. VetenksAkad. Forh. Stockh. 28 : 615-658. Lupwic, H. 1897. Die Seesterne des Mittelmeeres. Fauna u. Flora Neapel, 24: 1-496, pls. 1-12. 1905. Reports on an exploration off the West Coasts of Mexico, Central and South America, and off the Galapagos Islands, ... , by the ... “ Albatross,”’ 1891, ... . 35. Reports on the scientific results of the Expedition to the Tropical Pacific, $ , on the ... ‘“ Albatross,’’ 1399-1900, ... . 7. Asteroidea. Mem. Mus. comp. Zool. Harv. 32 : ix, 1-292, pls. 1-36. Lirxen, C. F., & MorTENSEN, TH. 1899. Reports on an exploration off the West Coasts of Mexico, Central and South America and off the Galapagos Islands, ... , bythe . “‘ Albatross,” 1891, ... . 25. The Ophiuridae. Mem. Mus. comp. Zool. Harv. 23: 97-208, pls. 1-23. Lyman, T. 1875. Zoological results of the Hassler Expedition. 2. Ophiuridae and Astro- phytidae. Illus. Cat. Mus. comp. Zool. Harv. 8 : 1-34, text-figs. 85-88, pls. 1-5. 1878. Ophiuridae and Astrophytidae of the exploring voyage of H.M.S. “ Challenger ”’ 1. Bull. Mus. comp. Zool. Harv. 5 : 65-168, pls. 1-10. 1883. Reports on the results of dredging, ... , in the Caribbean Sea, 1878-79, and along the Atlantic Coast of the United States, 1880, by the ... ‘Blake.’ 20. Ophiuroidea. Bull. Mus. comp. Zool. Harv, 10 : 227-287, pls. 1-8. Macan, T. T. 1938. Asteroidea. Sci. Rep. John Murray Exped. 4: 323-435, text-figs. I-12, pls. 1-6. Mapsen, F. J. 1950. The Echinoderms collected by the Atlantide-Expedition 1945-46. 1. Asteroidea. A. F. Bruun: “ Atlantide’’ Rep. 1 : 167-222, text-figs. I-11, pls. 14-16. MARENZELLER, E. von. 1893. Veroffentlichungen der Commission fiir Enforschung des Ostlichen Mittelmeeres. S.B. Akad. Wiss. Wien. Math.-Naturwiss. Kl. 102: 1-5. MortTENSEN, TH. 1927. Handbook of the Echinoderms of the British Isles. London. ix + 471 pp., text-figs. 1-269. 1933. The Echinoderms of St. Helena. Papers from Dr. Th. Mortensen’s Pacific Expedition 1914-16. 66. Vidensk. Medd. naturh. Foren. Kbh. 93: 401-473, text-figs. I-29, pls. 20-22. PERRIER, E. 1881. Reports on the results of dredging ... in the Gulf of Mexico, 1877-78, by the ... “ Blake,” and in the Caribbean Sea, 1878-79. 14. Description sommaire des espéces nouvelles d’Astéries. Bull. Mus. comp. Zool. Harv. 9 : 1-31. 1884. Les Etoiles de Mer recueilles dans la mer des Antilles et du Golfe du Mexique. Nouv. Arch. Mus. Hist. nat. Paris, (2)6 : 127-276, pls. 1-10. 1894. Expeditions Scientifiques du ‘ Travailleur’’ et du ‘“‘ Talisman” pendant 1880-83 Echinodermes. Paris. 1-431 pp., pls. 1-26. 1896. Contribution a l'étude des Stellérides de l’Atlantique Nord. (Golfe de Gascogne, Agores, Terre-Neuve). Résult Camp. sci. Monaco, 11: 1-57, pls. 1-4. 162 THE “ROSAURA” EXPEDITION Sars, G. O. 1871. Nye Echinodermer fra den norske Kyst. Forh. Vidensk-Selsk. Krist. 1871 : 1-31. 1875. On some remarkable forms of animal life from the great deeps off the Norwegian coast. 2. Researches on the structure and affinity of the genus Brisinga. Christiania. iv + Ik2 pp., pls. 1-7. Sars, M. 1861. Oversigt af Norges Echinodermer. Christiania. 1-160 pp., pls. 1-16. Say, T. 1825. On the species of the Linnaean Genus Astevias inhabiting the coast of the United States. J. Acad. nat. Sci. Philad. (1)5: 141-154. SLADEN, W. P. 1889. Asteroidea. Rep. Sci. Res. ‘‘ Challengey’’ Zool. 30: 1-935, pls. I-I19. 1891. Report on the collection of Echinodermata from the S.W. coast of Ireland dredged in 1888 by a committee appointed by the Royal Irish Academy. Proc. R. Irish Acad. (3)1(5) : 687-704, pls. 25-29. Stimpson, W. 1852. Two new species of Ophiolepis (Amphiura) from the southern coasts of the United States. Proc. Boston Soc. nat. Hist. 4: 224. 1853. Synopsis of the Marine Invertebrata of Grand Manan. Smithson. Contr. Knowl. 6: 16-17. VERRILL, A. E. 1895. Distribution of the Echinoderms of North-eastern America. Amer. J. Sct. 49 : 127-141, 197-212. 1890. Revision of certain Genera and Species of Starfishes with descriptions of new forms. Tvans. Conn. Acad. Arts Sci. 10 : 145-234, pls. 24-30. 18994. North American Ophiuroidea. 1. Revision of certain families and genera of West Indian Ophiurans. 2. A faunal catalogue of the known species of West Indian Ophiurans. Tvans. Conn. Acad. Arts. Sci. 10 : 301-386, pls. 42, 43. —— 1915. Report on the Starfishes of the West Indies, Florida and Brazil, including those obtained by the Bahama expedition from the University of Iowa in 1893. Bull. Labs. nat. Hist. Univ. Ia. 7: 1-232, pls. 1-29. Bull. B.M. (N.H.) Zool. 2, 6. PLATE 6, gue WR Sree 43 Fic, 2 Plate 6, fig. 1. Luidia yvosauvae sp.n. Type. Dorsal view. Fig. 2. Luidia 8 YI 5 bavimae sp.n. Type. Dorsal view. Both natural size. 24 mereeEPORT ON THE FISHES ‘COLLECTED BY me ROSAURA~ IN THE NORTH AND CENTRAL “ATLANTIC, 1937-38 PART 1. FAMILIES CARCHARHINIDAE, TORPEDINIDAE, ROSAURIDAE (NOV.), SALMONIDAE, ALEPOCEPHALIDAE, SEARSIDAE, CLUPEIDAE By DENYS W. TUCKER, B.Sc., British Museum (Natural History) [Received 25th April, 1953] With Plates 7 and 8 and Text-figs. 1-19 SYNOPSIS Specimens of Carcharhinus leucas (Miller & Henle); Diplobatis pictus Palmer; Rosaura votunda nov, ; Salvelinus alpinus (L); Leptoderma macrops Vaillant; Searsia koefoedt Parr ; Persparsia taaningi Parr; Pellonula afzeliusi Johnels; Havengula pensacolae G. & B.; and Ilisha africana (Bloch) are reported upon. The family Rosauridae is proposed for Rosaura votunda nov., an aberrant new oceanic Isospondyl. The paper presents new observations and discussion of tumours in Salvelinus ; of the structure and functions of the shoulder-organ of Searsidae ; of the occurrence of lateral line papillae in Alepocephalidae; and of the general taxonomic significance of photophore systems. INTRODUCTION As soon as the “‘ Rosaura’’ fishes were received at the British Museum (Natural History) in 1938, a preliminary rough-sorting and assessment were carried out by Lieut.-Col. W. P. C. Tenison, D.S.O., who also made illustrations of six of the more bizarre deep-sea species which were included in a popular account by J. S. Colman published as an appendix to the late Lord Moyne’s Atlantic Circle in the same year. (Another version of the same material is included in Colman, 1950, The Sea and its Mysteries.) With the approach of war the collection was set aside and remained unstudied until 1951, when I began my apprenticeship in ichthyology with a very laboured consideration of it. Many of the larvae and smaller specimens were still unnamed, and for a variety of reasons it was sometimes necessary to modify Lieut.-Col. Tenison’s determinations ; but his work had materially shortened the initial task of identification and is gratefully acknowledged. The collection is especially rich in well-preserved bathypelagic species from the little-worked Caribbean and equatorial Atlantic, and the present report will lay greater emphasis upon these than upon the shore fishes. A general list of species taken at each station will appear with the concluding instalment of the report, together with any minor addenda and corrigenda and general conclusions suggested by the work. Photographs used are acknowledged as applicable ; all other figures are from drawings by the author. Except where specially stated, standard lengths (S.L.) of fishes are given, and denote the measure- ment from the tip of the snout to the base of the caudal fin. ZOOL. 2, 6 10 164 THE “ROSAURA” EXPEDITION My best thanks are offered to those specialists and museum curators in other institutions whose assistance is mentioned in the text. I am especially grateful for the interest, advice and guidance so gladly given by my colleagues Dr. Ethelwynn Trewavas and Mr. N. B. Marshall, and for the services rendered by the technical staffs of the Fish Section and Photographic Studio of the British Museum (Natural History). Iam indebted to Mr. J. S. Colman (now Director of the Marine Biological Station, Port Erin, I.o.M.), who made this very fine collection and has permitted the reproduction of extracts from his field notes which appear as quotations in the text. Order PLEUROTREMI Family CARCHARHINIDAE Carcharhinus sp., probably C. leucas (Miiller & Henle), alternatively C. longimanus (Poey) or C. obscurus (Lesueur). (Pl. 7) (1) Carcharhinus (Prionodon) leucas (Miller & Henle). Carchavias (Prionodon) leucas Miiller & Henle, 1841, Syst. Besch. Plagiostomen : 42. Berlin. Carcharhinus leucas Bigelow & Schroeder, 1948, Fishes of the Western North Atlantic, 1, Sharks : 337; Myers, 1952, Copeta : 268. Has. Western Atlantic, New York to Southern Brazil; (?) Peruvian Amazon. (2) Carcharhinus longimanus (Poey). Squalus longimanus Poey, 1861, Memorias hist. nat. Cuba, 2 : 338. Carcharhinus longimanus Bigelow & Schroeder, 1948, t.c. : 354. Has. Tropical and subtropical Atlantic ; Mediterranean. (3) Carcharhinus obscurus (Lesueur). Squalus obscurus Lesueur, 1818, J. Acad. nat. Sci. Philad., 1: 223. Carcharhinus obscurus Bigelow & Schroeder, 1948, t.c. : 382. Has. Western Atlantic, New York to Florida & Bermuda; doubtfully south- wards to Brazil. St. 41. 18.xii.37. 0° 38'S., 43° 42’ W. 2-metre stramin net, 900(-0) metres. One specimen (a female), 211 cm. total length. Notes and photographs taken ; heart only preserved. Reg. No. 1953.3.6.1. The following data are extracted from notes by J. S. Colman : “Shark. Female. Carcharhinus sp. (acc. to Norman). Colour—grey dorsally, white ventrally. Two rows of teeth. No spines in front of fins. Ducti endolymphaticus present. 5 gill-slits each side. Stomach quite empty. Ovaries very immature ; no eggs developing. I could find no parasites. Four Remoras accompanied it.’ THE “ROSAURA” EXPEDITION 165 Dimensions : Snout to end of caudal fin (¢otal length) . 211cm. 5 mouth Sa ra of upper eal : EAD, (COHGY Gate.) mE; eyes . : E55 ( 71% ” ) 43 Ist eill-slit : : 3 7 40\ 3, + (183095 %, ) ~ BD : : aS (eee Ea a) 3 base of pectoral fa : P : Ba) 558 (24-65, ) a , Ist dorsal fin . : : 66r 57 (er-295) ua) 35 cloaca 5 3 5 3H (GBSOYA, re base of 2nd anise ia, ‘i 5 Bee (ORIG a )) = a) anallfin: : - TS 40 (O85, m ean) , caudal fin ? 5 SOL Ep OASYS” Gy) Girth at Ist gill-slit ; : : cA Max. girth (at base of Ist feel fin) . . SSiss Min. girth (at base of caudal fin) : Dil Spread of pectoral fins. 2 5 TES) On the following characters the specimen may be assigned to the family Carcha- thinidae in the sub-order Galeoidea: A normally shaped head, not laterally expanded ; five gill-slits, the last situated behind the origin of the pectoral fins ; two series of teeth functional ; two dorsal fins, which are not preceded by spines ; first dorsal much shorter at base than caudal and terminating well before the origin of the pelvics ; an anal fin present ; caudal fin not lunate and occupying less than half the total length. Indirect methods of identification, taking into account the absence of a spiracle, the short snout, and the relative sizes, shapes and insertions of the fins, allow the elimination of Galeocerdo, Paragaleus, Prionace, Negaprion, Aprionodon, Scoliodon and Hypoprion. There remains only Carcharhinus, with fourteen Western Atlantic species, of which one, C. micaraguensis, is confined to fresh water, and others have notably longer snouts as well as differences in the fins which are unmistakable. Of three species one “ possible,’’ C. /eucas, seems by far the most likely. C. leucas has the reputation, consonant with a capture in a plankton net, of being a heavy, slow-swimming species. Against an almost complete correspondence with the smaller female described by Bigelow & Schroeder, the present specimen offers only very small differences in the more posterior insertions of the pectoral and first and second dorsals, and these seem negligible when considered against small differences likely from methods of measurement and probable ontogenetic and other variations. C. longimanus agrees in the short snout and long, slender pectoral, and in the relative insertions of the fins, but has broader 1st dorsal and ventral fins and longer posterior tips to the 2nd dorsal and anal. C. obscurus belongs to the group of species having a median dorsal ridge, the presence or absence of which cannot certainly be determined from the photograph. 166 THE “ROSAURA” EXPEDITION It has a short snout and other characters in common with the present specimen, but appears to differ considerably in the shorter pectoral and lower second dorsal. The Remoras mentioned by J. S. Colman are four examples of Remora remora (L), 102-195 mm. S.L. In view of a very interesting paper by Szidat & Nani (1951, Rev. Inst. Ciencias nat. y Mus. Argentino, 2 : 385) presenting evidence that Remoras feed upon the copepod ectoparasites of sharks and other large fishes, I made a point of examining these specimens. No recognizable organisms were found in the very small amount of food present in their mouths and stomachs ; the result, though disappointing, is in accord with the field-note on the absence of shark parasites. Order HyPoTrEMI Family TORPEDINIDAE Diplobatis. pictus Palmer (Pl. 8) Palmer, 1950, Ann. Mag. nat. Hist. (12) 3: 480. St. 35. I.xii.37. 9° 25’ N., 59° 52’ W.; otter trawl; 86(—o) metres. One specimen (a female), 131 mm. total length. Reg. No. 1953.3.6.2. St. 36. 2.xii.37. 7° 11’ N., 57° 59’ W.; otter trawl, 20(—o) metres. Two specimens (females), t10-140 mm. total length. Reg. No. 1953.3.6.3-4. Has. Guiana Coast; mouth of R. Orinoco. This species was previously represented in the Museum’s collections by the unique holotype, Reg. No. 1950.5.15.4. The specimen from St. 35 is badly deformed, extensive injuries to the body and pectorals having been followed by irregular healing and regeneration. Those from St. 36 (Pl. 8) agree adequately with the holotype, save that a rounder form replaces the slight narrowing of the body behind the eyes, the bodies are better nourished and the hinder edges of the pelvics have broader points. The last observa- tion led to some conjecture whether the holotype might not be an immature male, but there is nothing else to support that view, and by analogy with other Torpe- dinidae recognizable claspers would be differentiated at a much younger stage. Palmer’s fig. 3, it must be noted, is very diagrammatic: the edge of the pelvic is not smooth as shown but includes three or four shallow undulations each beset with several small points, and anterior to the free hind tip there is a marked zone of fusion between the inner edge of the pelvic and the ventral surface of the tail. Reproduction of Palmer’s fig. I rather over-emphasizes the contrast in tone between the dark spots surrounding the light markings on the back and that of the remaining dark blotches on the body. In the smallest specimen now available the light spots on either side the back have an almost pearly whiteness which evi- dently becomes much clouded in ontogeny; also the tips of the dorsal fins are rounder in the young stage, as in the Pacific D. ommata, and become pointed with THE “ROSAURA” EXPEDITION 167 age. The remaining differences described between D. pictus and D. ommata (the genotype and only other known species) are sustained by the present material. Bigelow & Schroeder (1948, J. Mar. Res. New Haven, 7: 43) give two main characters separating Diplobatis from other Torpedinidae: the subdivision of the nostril into two about the middle of its length by a bridge of stiff tissue, and a relationship of tooth bands to the thick, fleshy lips such that the teeth are entirely concealed when the mouth is retracted and closed. In consideration of the possibility of ontogenetic variation in the Torpedinidae I have examined a long series of Narcine brasiliensis (Olfers) down to embryos with yolk-sacs attached, and find that at all stages the nostrils are undivided, so that the condition in Diplobatis may be regarded as distinctive enough. Never- theless the character is of doubtful use to the taxonomist, since only in the smallest of the present specimens is the bridge intact, and even then it can hardly be described as being of “stiff tissue” throughout ; rather does there seem to be a close ap- proximation of a concave upper and convex lower surface with a little tenuous tissue between the two. In the remaining material it would be difficult to decide from superficial appearances whether a bridge ever existed. In the holotype (of D. pictus) it remains as figured only on the right side, and Mr. Palmer informs me that on the left side it was ruptured during study subsequent to his paper. Order IsosPponDYLI Family ROSAURIDAE nov. Rosaura rotunda gen. et Sp. Nov. (Text-figs. 1-8) St. 42. 21.xii.37. 5° 51'S., 34° 38’ W.; 2-metre stramin net, 1200(-0) metres. One specimen (holotype), 8-4mm.S.L. Reg. No. 1953-3.6.11. Has. Deep Atlantic, N.E. from Brazil. Material and methods This solitary specimen, though but a post-larva, is sufficiently advanced in ossification and general development to be recognized as substantially different from all recent and fossil Isospondyli yet known. Its transparency aroused hopes of attaining a complete knowledge of the anatcmy of the untreated fish, but these could not be realized fully without resort to artificial aids. I therefore made accurate notes of all structures visible in the material and then proceeded to clear and stain, controlling all stages under the microscope and employing, instead of the usual more elaborate methods (Hollister, 1934, Zoologica, 12: 89-101; Raitt, 1935, J. Cons. Explor. Mer, 10 : 75-80) a simplification used in the British Museum (Natural History) for many years: THE “ROSAURA” EXPEDITION 168 [esiop Jo woroaford [eurozuy ueutoeds ayy T's ‘wur +.g « *petddys umoys Aj1aevo-Apoq ozur sarawoAur “BUIUTe}S pue SuLea] 03 Jord woyrpuos s}r ur uMoys SI adAjojoyy =*aou ‘ds yo ‘ua DPUNJOA DANDSOY “I *OIA-LXA, THE “ROSAURA” EXPEDITION 169 (1) Preserved material in 70% spirit transferred to distilled water—15 minutes. (2) Partly cleared in 0-5% KOH solution—3o0 min. (3) Stained in an old deep purple solution of alizarin in distilled water—16} hours. (4) 0°5% KOH solution—7o minutes. At this stage the specimen was still understained, so staining was repeated with a freshly-prepared solution. (5) Solution of about } c.c. of alizarin powder in 200 c.c. distilled water—z hours. (6) Solution of 2 volumes of 0-5 KOH to 1 volume of pure glycerine. During the next 24 hours glycerine was added a little at a time until the specimen could safely be transferred to pure glycerine containing a thymol crystal as preservative. The untreated specimen had sustained compound fractures of the cleithra and wanted parts of two teeth. Processing was accomplished without further damage, but there were several breakages of caudal and branchiostegal rays during subse- TEXT-FIG. 2. Rosauva votunda. Anterior view of head from below. quent study and manipulation, which do not seem very serious in proportion to the additional information gained. The text-figures were drawn freehand with the assistance of a 5 X 5 eyepiece graticule, calibrated with a stage micrometer. 170 THE “ROSAURA” EXPEDITION General appearance (Text-figs. I and 2) Radial formula D.16; A.1z2+; C.23+-; P.12-+; V.5(L), 2(R). Branchiostegal rays 10. Gills 4. Myomeres of body 37. At first glance reminiscent of the larval Ceratioids, a minute fish with an overall length of 10-3 mm., transparent or translucent, scaleless and colourless save for a few isolated melanophores atop the head (Text-fig. 3) and on the caudal peduncle. TExT-FIG. 3. Rosauva rotunda. Disposition of melanophores on dorsal surface of head, shown in relation to the brain and frontal bones. The greatest height is in the broad otic region. From this projects obliquely forwards and downwards the narrow ethmoid cartilage. The suspensorium is vertical, forming with the cranial axis and the jaws an approximately equilateral triangle. Behind this the vertebral column with its myomeres descends obliquely to the posterior end of the almost spherical body-cavity and is continued as a short caudal region. The body-cavity is bounded in front by an almost vertically placed pectoral girdle, lightly curved forwards below, with the pectoral fins inserted about half-way in the height. The origin of the dorsal fin is substantially in advance of that of the much larger anal. An adipose finis present. The caudal fin is large and fan-like. A prominent anal papilla is flanked by a pair of minute abdominal pelvic fins. The gill-openings are wide and bordered mainly by the branchiostegal membranes, which are continuous below (free from the isthmus), and supported by the much- reduced opercular and subopercular and by ten conspicuous, long, slender branchio- stegal rays. Pre- and inter-opercular bones are absent. The large mouth is bordered by toothed premaxillaries and dentaries with prominent caniniform teeth. The maxillaries are toothless and probably enter the gape only to a very limited extent. The eye is small for an oceanic fish and is widely separated from the jaws. No division can be seen in either of the single pair of minute nostrils. THE “ROSAURA” EXPEDITION 171 Z (tin : FR 4, 2 a ; ) A v re br I. bel. TEXT-FIG. 4. Rosaura rotunda. Osteology of head and pectoral girdle in a cleared and stained specimen, with optical dissections of the maxillary, palato-pterygoid and hyomandibular-symplectic-quadrate. Cartilage shown stippled. For key to abbre- viations used see p. 181. 172 THE “ROSAURA” EXPEDITION Cranium (Text-figs. 4 and 7) A great part of the cranium is still unossified, and its description limited by the difficulty of interpreting a single small and delicate specimen which may not be dissected. The antorbital portion of the cranium is a substantial part of the whole structure. The largest element is the ethmoid cartilage, which is arched above, flat below, a long cartilage extending half the length of the head and of uncertain lateral extent. Anteriorly it is overlain by the rostral processes of the premaxillaries ; mesially it is exposed between the widely separated anterior limbs of the frontals, and posteriorly it sends thin supraorbital blades beneath the lateral processes of the frontals towards the auditory capsule. The vomer and parasphenoid are applied to its ventral surface. The orbit is placed high on the head, laterally directed and remote from the jaws. The postorbital portion of the cranium is an unossified chondrocranium, globose and including considerable auditory capsules. In dorsal view it is possible to see, through the frontals, edges of cartilage which run parallel and close to the optic lobes of the brain; whether these are in fact the anterior margins of fonta- nelles cannot be ascertained since posteriorly they become lost in the complications of the auditory region, but it seems very likely that they are. The only bones present on the dorsal surface of the head are the frontals, which are widely separated in the middle line. They have slender anterior extensions reaching to the tips of the premaxillary rostral processes, short posterior extensions part over the auditory capsules, and broader lateral processes contributing to the roof of the orbit. Separate parietals are not present. The vomer (seen only in lateral view) is a narrow sliver of bone extending to beneath the anterior part of the orbit, where it overlies the front end of the para- sphenoid. The latter cannot be seen distinctly and is therefore not included in any of the figures. There is no mesethmoid, lateral ethmoid, or prefrontal ossified, nor are any bones of the otic or occipital series. Jaws (Text-figs. 4, 5 and 7) The toothed premaxillaries form the greater part of the upper boundary of the mouth. Each is V-shaped, the inner and longer arm of the V being a pointed rostral process which overlaps the anterior tip of the frontal of the same side, the outer and shorter arm being the dentigerous margin of the jaw. The premaxillaries are separated by a narrow median fissure which broadens anteriorly and posteriorly into wide notches. The toothless maxillaries are broadly elliptical, much like those of the Sterno- ptychidae, and can enter but little into the gape. Each maxillary narrows anteriorly to form a twisted process inserted into the ethmoid beneath the posterior angle of the premaxillary. The squarish posterior end of the maxillary has a broad shallow notch. SS THE “ROSAURA” EXPEDITION 173 Meckel’s cartilage persists throughout the length of the lower jaw, though almost entirely overlain by membrane bones. The large toothed dentary is an elongated oval with a broad notch in the wide posterior end receiving, and in part overlying, the much smaller angular. There is a marked tuberosity at the mandibular symphysis. The body of the angular is roughly an equilateral triangle with the articulating facet of the quadrate above its hinder corner, and with a short narrow limb posterior to this partly overlying the end of Meckel’s cartilage. The hinder- most tip of Meckel’s cartilage is ossified as a truncated conical retro-articular. There is no other articular ossification. TEXT-FIG. 5. Rosauva votunda. Dentition, viewed from left and slightly below. The broken line shows the position of the maxillary. Dentition (Text-figs. 5 and 6) The dentition is highly raptorial. The teeth on the premaxillary are in two series : an external row of seven minute teeth downwardly and outwardly directed from the edge of the premaxillary, the most anterior of the series being weakly caniniform ; and on the inner face of the premaxillary a series of four stronger, inwardly-directed and recurved teeth. The maxillary is toothless. In the lower jaw the dentition is dominated by a pair of very large and strongly recurved caniniform teeth, set inside and a little behind the mandibular symphysis. Each dentary further carries five moderately recurved caniniform teeth of oddly- assorted sizes, and posterior to these a graded series of six to eight laterally com- 174 THE “ROSAURA” EXPEDITION pressed, shearing teeth. So far as may be determined without manipulating the jaws there are no vomerine teeth. (The palatines are not ossified.) Accessory structures present in the mouth may be related to the need for re- moving prey impaled on the enlarged mandibular teeth. The tip of the basihyal (Text-fig. 6) bears a similar pair of enlarged and recurved teeth (one with a replace- ment tooth behind it), evidently capable of some degree of motion between and behind the great teeth of the mandible. Above there is a median fleshy down- growth depending from the maxillary valve between these two pairs of teeth which may bear some functional relation to the lower dentition. Palatine arch (Text-fig. 4 and detail) The palatine end of the palato-pterygoid cartilage is unossified, but is not believed to be synchondrous with the lower end of the ethmoid cartilage. A broad 7-shaped endopterygoid, rounded at its anterior end and tapering to a point at the posterior, sheaths it on its inner side. The metapterygoid is a large rhomboidal bone covering much of the cheek between the orbit and the upper limb of the endopterygoid ; its posterior end bears a suborbital prominence and is inserted mesiad to the hyo- mandibular-symplectic cartilage A zone of cartilage lies exposed to the surface at the periphery of the metapterygoid. Hyoid arch (Text-fig. 4, detail and 6) The hyomandibular is ossified only in its middle portion, the head and lower end being cartilaginous. The head comprises two dorsal protuberances (unossified) and one ventral (ossified), the latter bearing a small condyle for the articulation of the minute opercular. The symplectic, likewise ossified in the middle portion only, is rod-shaped. Its upper end is synchondrous with the hyomandibular. Part of the ossified portion, and the whole of the lower cartilaginous end, are overlain by the quadrate. The quadrate is a short triangular ossification, the posterior edge of which is extended upwards as a long tapering rod. The articular condyle remains carti- laginous. The interhyal is a short cylindrical cartilage attached at its upper end to the cartilage between hyomandibular and symplectic and at its lower end to the lower limb of the hyoid arch. It is not possible to distinguish with certainty between epi-, cerato- and hypo-hyal elements, and Text-fig. 6 shows as much as can be discerned under the limitations imposed by viewing through the suspensorlum and unopened jaws. The basihyal element is more readily distinguished and bears a pair of enlarged teeth. The branchiostegal rays are long, slender, tapered towards their extremities, and elegantly curved, giving a globose contour to the anterior body of the fish. The last four describe a somewhat S-shaped curve in traversing a vertical fold of the gill-cover; this curve is strongest in the uppermost (posterior) rays. The EE THE “ROSAURA” EXPEDITION 175 second to sixth rays approach the hyoid arch in a smooth simple curve. The first (lowest) branchiostegal ray is twice as broad as the rest, slightly twisted, and passes well between the rami of the mandible. The curvature and apparent grouping of the branchiostegal rays suggest, by analogy with other fishes, that there may be a ceratohyal bearing six rays and an epihyal with four. br. TEXxT-FIG. 6. Rosaura rotunda. Hyoid arch. The ten branchiostegal rays are accu- rately represented ; the basihyal and interhyal are also correct, but the remaining cartilages (epi-, cerato- and hypo-hyal) cannot be shown properly owing to difficulty of observation. The positions of premaxillary and dentary and of some of their teeth are shown by broken lines. Operculay apparatus (Text-fig. 4) The gill openings extend very nearly the full depth of the fish and are bordered by voluminous branchiostegal membranes which are continuous below (free from the isthmus). The branchiostegal membranes are supported by the opercular and sub-opercular bones and the ten branchiostegal rays; the pointed tips of the sub- opercular and last nine branchiostegal rays project beyond the hinder edge of the branchiostegal membrane like the ribs of an umbrella; the tips of the opercular and first branchiostegal ray do not reach to the margin of the membrane. The opercular is a rudimentary, ventro-laterally directed, twisted, leaf-shaped bone articulating with a small condyle on the head of the hyomandibular. The sub-opercular is a narrow, pointed bone about twice as long as the preceding, con- verging with the tenth branchiostegal ray. Gills and gill-apparatus Viewed as they have been through the suspensorium, the gill-arches have yielded little information. There appear to be four only, bearing double rows of long and 176 THE “ROSAURA” EXPEDITION not very numerous gill-filaments. The lower limbs of the arches bear minute conical denticles or rakers, about a dozen to each limb. Pectoral girdle and fin (Text-figs. 4 and 7) The post-temporal (Text-fig. 7) is a flat, spatulate bone which mesially comes near to meeting its fellow over the otic region of the cranium, and passes almost transversely and horizontally before the dorsal somites to overly the upper tip of the supracleithrum. The supracleithrum, so far as may be ascertained in its damaged state, is a flattened, oval bone, substantially smaller than the post-temporal; its broad postero-ventral end overlies the cleithrum a little below the latter’s upper tip. The cleithrum is very large and boomerang-shaped, and with its partner comes near to forming a complete bony ring round the body. The upper limb is nearly vertical ;. the lower is rather slenderer and curves forward and downward. The pectoral fin is inserted behind the posterior angle of the cleithrum at about the middle of the body-height. The coracoid cartilage is perforated by a minute coracoid foramen and bears a prominent, posteriorly directed ventral spine, but has no trace of ossification nor of differentiation. Its anterior edge is inserted inside the hind edge of the cleithrum, and there is a small antero-ventral process which follows the edge of the cleithrum downward. The pectoral fin of each side is much crumpled and the structure difficult to make out. Each appears to be a vertically-inserted reniform cartilaginous lobe, having no actinosts ossified as yet. Twelve adult rays have stained on the upper edge of the fin, and doubtless the number would have been increased later by the gradual replacement of the very numerous unossified larval rays. Pelvic girdle and fin (Text-figs. 1 and 8) There is a very slender and apparently unpaired pelvic cartilage lying transversely in the ventral body-wall before the anal papilla. The tapered ends of the pelvic cartilage turn downwards and backwards into the fin-bases ; these are backwardly directed papilliform structures from which the short close-set rays project upwards and backwards. The fins are obviously in the early stages of development ; that of the right side is placed higher and more anteriorly, and has but two rays com- pared with five on the opposite side. Axial skeleton (Text-fig. 1) The vertebral column is as yet entirely unossified. The surface of the stout notochord has the usual reticulate appearance ; its posterior end is curved upwards and backwards and the slender tip projects between the bases of the 3rd and 4th upper caudal rays. There is a slight constriction in the notochord at the point —— eee THE “ROSAURA” EXPEDITION 177 where it breaks the edge of the fin-lobe. The first two pairs of basidorsal elements at least are laid down in cartilage (Text-fig. 7). The dorsal fin is supported by 16 cartilaginous basalia, each of which bears one tay. The fin is smoothly rounded; the middle rays are the longest, but do not quite reach to the adipose fin when depressed. There is a wisp or two of larval ray at the hinder end, but the full adult complement seems to be present. The fin base is slightly elevated. The adipose fin is rather large, slightly hooked, and with a fimbriated edge. The anal fin, though probably incompletely developed, is still very much larger than the dorsal. Anteriorly there is a fringed adipose blade in which further rays might have been differentiated ; there are 12 cartilaginous basalia thus far, each with aray. The fin is rounded, the middle rays are the longest, and the depressed fin reaches to the anterior rays of the caudal. The first anal ray corresponds to the same myocomma as the eleventh dorsal ray, while the anal fin base extends beyond the posterior end of the adipose. Discussion of verticals through fin origins would be unprofitable, as may be observed from Text-fig. I. The caudal fin is supported by the tip of the notochord, and by two epurals and three hypurals, all cartilaginous. Between the distal extremities of the hypurals are minute triangular supplementary cartilages. There are three epaxial and twenty hypaxial rays, all elegantly curved and tapered at their extremities; the broadest part of each ray lies a short distance from its insertion, as in a quill. The dorsal rays are a little more slender and closely set than the ventral; the rays are evenly gradated and form a considerable caudal fan. Anterior to the caudal rays on both dorsal and ventral edges is a fringed adipose blade in which it is evident that a substantial number of procurrent caudal rays would have been differentiated. Branching is incipient at the ends of most of the fin-rays. Visceral anatomy (Text-figs. 1 and 8) The body-cavity is spacious and near-spherical. The funnel-shaped pharynx narrows to a short oesophagus. The stomach is small with a median constriction. A single pyloric caecum bears subtle transverse surface undulations suggesting subdivision into three or four at a later stage. The intestine leaves the stomach to descend round the left side of the body, narrows as it runs transversely upon the lower body wall, ascends some distance up the right side, and then turns and runs obliquely downwards and backwards from right to centre to open on the anal papilla. There is some differentiation of texture and thickness in the regions of the gut, but no certain indication of a spiral valve nor of any rectal glands. The pronephros was remarkably distinct even in the untreated material, a pear- shaped structure upon whose surface tubules are apparent, attached near the anterior somites and the upper end of the cleithrum. The liver comprises two major lobes, narrowing as they run dorsally to left and right. A problematic strip of similar-seeming tissue may be pancreas; one cannot say. 178 THE “ROSAURA” EXPEDITION i AES Ness | fil) ee Xe all Gini ne We Ti 1 scl. Text-Fic. 7. Rosaura votunda. Dorsal anatomy of the head: the lateral scale slightly exaggerated due to drawing with a twin-objective microscope. Cartilage coarsely stippled. The eye-muscles and auditory organ of the left side are“omitted to"avoid over-complication. For key to abbreviations used see p. 181. cl. THE “ROSAURA” EXPEDITION 179 Of the heart may be seen a large spherical auricle, a smaller thick-walled ventricle, and a structure which must almost certainly be the sinu-auricular valve, though the sinus itself is too transparent to be distinguished. A iarge vessel runs in the mesentery to the middle of the intestine. There is no trace of gonad development. Musculature, etc. The total number of myomeres developed is 37. Of these 16 are in front of the dorsal fin; the dorsal fin-base occupies 64; there are 24 between dorsal and adipose fins ; the adipose fin-base occupies 3, and g remain to the end of the body. There are 22 myomeres before the anal fin, and the anal fin-base occupies 93. The dorsal muscle somites are shallowest and forwardly extended where they are ap- proximated to the posterior end of the cranium, and become progressively deeper towards the origin of the tail. The ventral body musculature forms a thin con- tinuous balloon-like surface, much crinkled in preservation, and with only here and there a line to indicate a myocomma. The m. adductor mandibulae has a superficial component inserted on the inside of the angular in advance of the jaw-articulation; another component inserted nearby on the maxillary, and a deep component which turns sharply forwards and inwards to an adjacent insertion on the inside of the dentary. The long tendon of the m. adductor mandibulae runs along the anterior edge of the hyomandibular- symplectic to an insertion on the head of the hyomandibular or somewhere on the skull close by. A very strong ligament runs from the hindermost tip of the lower jaw along the hinder edge of the suspensorium to the region of the opercular. As in Chatliodus (Tchernavin, 1953, The Feeding Mechanisms of a Deep-sea Fish (B.M. [N.H.}), however, the opercular is small and the m. levator operculi, if present, consequently likely to be weak and relatively useless in opposition to the m. adductor mandibulae. I cannot see any ligamentous connection between the hind end of the mandible and the posterior end of the epihyal, and therefore assume that the main work of opening the jaws is accomplished by the m. sterno-hyoideus, which is well developed in common with the other muscles in the floor of the mouth. There is a patch of fibrous tissue binding the lower posterior end of the maxillary to the dentary. The latero-ventral boundary of the buccal cavity is formed by a thin sheet of tissue, probably smooth muscle, whose lateral edge extends as an arc from the level of the middle of the symplectic to the tip of the basihyal. The abductor pinnae pectoralis is a thick muscle between the coracoid cartilage and the anterior face of the pectoral fin; its strong development adds another obstruction to the observation of the radialia. Traces of muscles are visible upon the hypurals of the caudal fan, together with a loose aggregation of myoblasts behind the ultimate somite. The eye muscles are long and slender (Text-fig. 7). The m.m. recti and m. obliquus ZOOL. 2, 6. 11 180 THE “ROSAURA” EXPEDITION inferior require no special comment, but the m. obliquus superior is unusual in that it follows a very independent course to an insertion well forward on the ethmoid. pely. cart. Text-Fic. 8. Rosauvavotunda. Abdominal viscera, viewed from a point behind and below the vent. For key to abbreviations used see p. 181. Nervous system and sense organs (Text-fig. 7) Differentiation of the brain is well advanced, though there are no structures that call for very special notice. There are olfactory lobes with olfactory tracts running forward and outward through the ethmoid cartilage to the olfactory organs ; behind these lie corpora striata, and between them an epiphysis or pineal organ. The optic lobes are very large, as also is the cerebellum, and the third and fourth ventricles are readily distinguished. The swollen origin of the spinal cord is exposed between the anterior somites. THE “ROSAURA” EXPEDITION 181 Key to abbreviations used in text-figures ang, angular. oes, oesophagus. an.pap, anal papilla. o.l, optic lobe. aur auricle. olf.l, olfactory lobe. bas.h, basihyal. olf.o, olfactory organ. bd, basidorsal. op, opercular. brst, branchiostegal ray. pect, pectoral fin. cer.h, ceratohyal. pelv, pelvic fin. ath cleithrum. pelv.cart, pelvic cartilage. cor, coracoid cartilage. pin, pineal body. GS} corpus striatum. pmx, premaxilla. den, dentary. p.neph, pronephros. echt, ectopterygoid. p.tem, post-temporal. enpt, endopterygoid. pyl.c, pyloric caecum. eth, ethmoid cartilage. qd quadrate. fr, frontal. Y.a, retro-articular. hm, hyomandibular. ? S.Q.U, ? sinu-auricular valve. int, intestine. s.cl, supra-cleithrum. int.h, interhyal. sym, symplectic. 1, liver. S.op, sub-opercular. LLF, opening of lateral line canal. st, stomach. mpt, metapterygoid. ven, ventricle. mx, maxilla. vom, vomer. my, myomere. The eyes are laterally directed, set high on the head, and rather small for an oceanic fish. They are loosely suspended in orbits too big for them, but there is nothing else to evidence a post-Stylophthalmus condition. The line of suture of the choroidal fissure is quite distinct (Text-fig. 4). The olfactory organs are minute and extremely difficult to interpret; they are undivided, and seem to have the structure of very shallow thick-walled funnels whose openings are occluded to elliptical chinks. There is a wisp of tissue de- pending from the edge of one which is of dubious significance. The auditory labyrinth is very well developed; the three semicircular canals with their ampullae are present, and the cristae acousticae of the ampullae clearly visible (Text-fig. 7). Part of the otolith can be seen in the bulla in dorsal view, but its shape cannot be distinguished through the complicated light-refracting surfaces of the auditory capsule. Very little of the lateralis system was visible in the untreated specimen, and that has since disappeared in clearing. There are pits with sense organs on the frontals (Text-fig. 7) and on the dentaries (Text-figs. 2 and 4), and one or two other organs of the suborbital and preopercular-mandibular series were faintly seen. There is no trace of a lateralis system on the body. Portions of the IInd, IIIrd, IVth, VIth and VIIth cranial nerves have been noticed, but the observations were typical and too fragmentary to merit comment. 182 THE “ROSAURA” EXPEDITION The surface of the body has many minute pilose projections. It is unlikely that these indicate anything more than muscle-fibre-bundles, but since damage to this specimen would have been of an all-or-nothing character, and in view of the discussion of lateral line papillae in other deep-sea fishes elsewhere in this paper, it may be worth while to put the observation on record. Diagnosis and relationships of the Family Rosauridae Isospondylous fishes in which the premaxillaries have long rostral processes ; the maxillaries are toothless but enter into the gape ; there are no supramaxillaries ; the opercular apparatus is much reduced, the opercular being minute and the sub-opercular but slightly less so; the branchiostegal rays are long, slender and relatively numerous (10 in Rosauwva) and support voluminous membranes; the eyes are laterally directed ; the metapterygoid is large ; an adipose fin is present ; there is an unpaired pelvic cartilage (bone ?), and there is no post-cleithrum. One known Genus Rosaura nov. with the characters of the Family, and with one species R. rotunda nov. based on a single immature specimen of 8-4 mm. S.L. This definition is adequate to sustain the new family as distinct without involving characters likely to be modified in ontogeny. Additional characters present in the young Rosaura rotunda which will become significant if it can be shown that they are carried through to the adult stage are: the short, stout body; the non- telescopic eyes; the undivided nostrils; the absence of pre- and inter-opercular, pre-frontals and parietals. In view of certain convergent similarities it may be as well to emphasize that the Rosauridae are readily removed from the Iniomi by their possession of pre- maxillaries which do not exclude the maxillaries from the gape; an unpaired ethmoid ; an unforked post-temporal; and a reduced opercular apparatus. Within the Isospondyli the affinities of Rosauva are difficult to trace, and it may indeed, like Macristium, represent one of those anomalous offshoots which are not likely to be related to any of the main branches. Perhaps its most likely relationship is to the Stomiatoids, which group have a moderate tendency towards the development of rostral processes, though less pronounced ; a raptorial dentition (though usually with teeth on the maxillary, and supramaxillaries present); a tendency towards loss of the parietals and post-cleithrum, and a reduced opercular apparatus ; and a high number of branchiostegal rays (19 in Chauliodus). The general body-form of Rosaura is not unlike that of the young stages of the Sterno- ptychidae. But none of these resemblances is very fundamental and the differences are formidable. During a recent hurried visit Dr. Anton Fr. Bruun glanced at the specimen and drawings, and expressed the belief that he had seen this little fish in some numbers in the “Dana” Collections, and thought that a comparison with Scopelengys tristis Alcock might be worth making. I followed up his suggestion, but from a consultation of the literature and an examination of a specimen taken by the “ John Murray” Expedition (Reg. No. 1939.5.24.463) find that the differences are very considerable. THE “ROSAURA” EXPEDITION 183 Family SALMONIDAE Salvelinus alpinus (Linnaeus) Salmo alpinus Linnaeus, 1758, Systema Naturae : 300. Salvelinus alpinus BEPY, 1932, PbIBbI IIPECHbIX BO], CCCP M4 COMPENEJIBHbIX CTPAH : 170 (Berg, Les poissons des eaux douces del’U.R.S.S., Leningrad. Text in Russian, full synonymy) ; Oliva, 1951, Copeia : o1, recent teferences). St. 7. 9.ix.37. 60° 16’ N., 44° 41’ W.; trammel net at mouth of stream entering Tasermuit Fjord, S. Greenland. (Pathological material.) ? Do. Three specimens, 222-330 mm. S.L. (No label.) Reg. No’s. 1953. 3.6.8-10. Haz. Arctic Boreal. (Anadromous and lacustrine races.) If not collected on the same occasion as the pathological material, the entire fishes must certainly be from one of the S. Greenland stations. Concerning a tumour- bearing lower jaw J. S. Colman notes : “When these tumours occur on the Char the fish are poor and thin and the Greenlanders will not eat them. The authorities at Julianehaab, S.W. Greenland, are very anxious to find out their nature and cause.” This specimen has been submitted to Prof. Alexander Haddow, Director of the Chester Beatty Research Institute, Royal Cancer Hospital, who forwards the following report by Dr. E. S. Horning : “ Microscopical Examination: Subcutaneous Fibroma. “ Growth consists of fibroblasts and bundles of interlacing fibrous tissue, together with what appear to be, with this fixation,” (Bouin) “ elastic fibres. There is also a rather delicate stroma of connective tissue, in which run nutrient vessels lined by an irregular epithelium. This tumour has the appearance of a hard fibroma, as some areas show a tendency to undergo hyaline degeneration. It is definitely not a myxomatous type of growth, as there is no evidence of fatty degeneration.” Another case of tumours in Salvelinus has been described by Hoshina (Jap. J. Ichthyol. 2 (1952) : 81-88. Family ALEPOCEPHALIDAE (For a recent generic revision and summary of literature, see Parr, 1951, Amer. Mus. Novitates, No. 1531 : I-21.) Leptoderma macrops Vaillant (Text-figs. 9-13) Lepicderma macrops Vaillant, 1888, Poissons Expéd. Sci. “ Travailleur” et “ Talisman”: 166, pl. 13, fig. 2; Goode & Bean, 1895, Oceanic Ichthyology : 49 (Washington) ; Koehler, 1896, Rés. Camp. Sci. “ Caudan”’: 523; Roule & Angel, 1933, Rés. Camp. Sci. Monaco, 86:8; Fowler, 1936, Bull. Amer. Mus. nat. Hist. 70 : 193; Bertin,! 1940, Bull. Mus. Hist. nat. Paris (2), 12: 275; Rey, 1947, Ictiologia Iberica, 2: 92 (Madrid) ; (?) Alcock, 1892, Ann. Mag. nat. Hist. (6) 10 : 361. (?) Leptoderma affinis Alcock, 1899, Cat. Deep-sea Fishes “ Investigator’: 182. Calcutta; id., 1900, IJlust. Zool. ‘ Investigator,”’ Fishes : pl. 32, fig. 3; id., 1902, A Naturalist in Indian Seas : 237, fig. 35 (London). (?) Leptoderma retropinna Fowler, 1943, Bull. U.S. Nat. Mus. 100 Vol. 13, Pt. 2: 55, fig. 5. Leptoderma Springeri Mead & Béhlke, 1953, Texas Journ. Sci. 5, No. 2: 265 (received too late for inclusion), 184 THE “ROSAURA” EXPEDITION St. 49. 1.ii1.1938. 28° 25’ N., 13° 34’ W.; Agassiz trawl, c. 1,300(-o) metres. Three specimens, 85-150 mm. S.L. Reg. No’s. 1953.3.6.5-7. Has. Deep Atlantic off N.W. Africa; Gulf of Mexico; Gulf of Gascony ; ? Indian Ocean, Philippines. 650-2,300 metres. The British Museum (Natural History) formerly possessed only a paratype (from the same station as the holotype, fide Prof. L. Bertin! 2m litt.) of this species, Reg. No. 1890.6.16.44, still bearing the usual Paris tie-on label punched 85 * 223. In all four specimens now available for study the delicate black skin has peeled forwards, leaving most of the body naked; as with Vaillant’s other fifty-eight, therefore, “aucun des exemplaires n’est dans un état tout a fait satisfaisant.” Koehler (1896) likewise complains: “ le tronc est completement pelé.” Roule & Angel (1933) fared no better: ‘‘ Un exemplaire, en état défectueux The accounts of Goode & Bean (1895), Fowler (1936) and Rey (1947) appear to be dependent upon Vaillant’s original description (1888) ; there is no evidence that these authors handled material themselves and all perpetuate Vaillant’s errors. Koehler (1896) describes one specimen from the Gulf of Gascony. Roule & Angel (1933) merely add a locality record. Bertin (1940) lists Vaillant’s material remaining in the Paris Museum as part of a general catalogue of types. Neglecting Alcock’s very dubious specimen (1892, 1899, 1900, Ig02) and Fowler’s (1943) it appears that this species has been described only three times, and a new account may be of value. GENERAL DESCRIPTION (BASED MAINLY ON THE LARGEST “ROSAURA” SPECIMEN) Radial formula D.66; A.86; P.8; V.5-6; C.16. Branchiostegal rays 6. Gills 4 + pseudobranch. Body naked, elongate, tapered, laterally compressed, especially in the long caudal region, where it becomes almost filiform (Text-fig. 9). Head small, its length equal to the trunk and 5-55 in the length from tip of snout to caudal. Snout blunt, short, I-27 times the interocular width and 1-40 in the diameter of the eye; it bears a short, blunt, laterally directed spinule on each 1 Further consideration of Bertin’s ‘‘ Catalogue ’’ reveals an unfortunate misapplication of the term ““paratopotype,”’ which affects not only the types of Leptoderma macrops but a great many others as well. Bertin states (Bull. Mus. Hist. nat. Paris, (2) 11 : 64, 1939) : “Les autres sont des paratypes s'ils proviennent de la méme localité que l’holotype ou des para- topotypes s'ils ont été pris dans d’autres régions.” The term “ paratype’ need not necessarily imply any such geographic restriction, while “ para- topotype ”’ has hitherto denoted the exact opposite to Bertin’s definition—c.f Alexander, 1916 (Proc. Acad. nat. Sci. Philadelphia, 68 : 496), who introduces the term earlier in the same paper and on the page cited defines it by implication ; also Frizell, 1933 (Amer. Midland Naturalist, 14 : 659) : “ Paratopotype’’ (= paratype + topotype)—a paratype from the same locality as the holo- type.” Bertin’s two categories are already covered by the terms “ paratype omotype’’ and “‘ paratype allo- type "’ proposed by Silvestri (in Holland, 1929, Tvans. 4th Inter. Congr. Entom. Ithaca : 693) ; still-born expressions and unlamented by those who consider systematic zoology overburdened with redundant jargon. Itisa pity that ‘‘ paratopotype’”’ is as well known as it is, for few users of Bertin’s invaluable “ Catalogue ’’ will feel any need to refer back to his definitions and many will thus be misled. 185 EXPEDITION THE “ROSAURA” SOINJONIYS 9} ‘IO]TLUIS oY ‘oerided o3se} Atqeqord ore ‘aseo yora ur sisAydurAs oy} Jo 39] ey} 0} ‘Areyuop pue Areyprxeuroid oy} wor; 4300} Jo sdnorsy *“sdosovur vuiaapoiqaT ‘(-J'g ‘wu o£1 Jo ueurtoads) ‘II ‘DId-LXaJ, “Areprxeureidns ‘wus ‘ Arveyrxeur ‘ww { Avepixeuoid ‘xmd ‘mel rsddn ayy jo souog 9 ‘sdosonm vuaapojgay ‘Ol ‘DIs-1XT], ‘WW S {i eee ‘xud XLS “XU ‘uaultoads sty} Wor Udy} ore EI pue ZI ‘or ‘sSy-3xaq, “(umoys you ovpided Arosuas) “J's ‘wu o$1 jo uaumoedg ‘sdossvm wuaapojgay “6 ‘OId-1XAL, ‘W2S wee yyy, 186 THE “ROSAURA” EXPEDITION side arising from the anterior end of the maxillary. Interocular space very slightly convex, almost flat, I-79 in the diameter of the eye. Eyes enormous, projecting considerably on either side of the head, 2-78 in the head length. Nostrils large, their mid-point (in the shrunken condition of the specimen) approximately midway between the eye and the tip of the snout, the anterior smaller than and slightly dorsal to the posterior. Mouth small, terminal to inferior. Upper jaw 1-20 times as long as the snout, barely reaching to anterior rim of orbit when the mouth is closed ; lower jaw 1-36 times as long as snout, reaching to below anterior eyeball. Premaxillary slender, toothed, elongate, almost excluding maxillary from gape. Maxillary long, toothless,! curved, a small blunt spinule on its clubbed anterior end and a large @-shaped supramaxillary at the posterior (Text-fig. 10). The space between the upper ends of the maxillary and supramaxillary is spanned by membrane, and the whole upper jaw well shielded in a sac of thin, tough skin forming a sound functional unit. The upper jaw rotates about its anterior end when the mouth opens and discloses a lightly pigmented recess on the cheek. The moving maxillary is steadied against a broad coronoid process of the dentary. Teeth! in a single row in each jaw, about 14-20 on the premaxillary, seldom TABLE I.—Leptoderma macrops: Table of Measurements and Comparative Proportions It should be noted that Vaillant writes variously ‘“‘ Longueur " and “ Longueur totale.” In the present case he gives the length of the caudal fin as 3 (?) mm.; the likely error in a false assumption of standard length is therefore of the order of o-1-0-4% underestima- tion in the figures quoted, and may be neglected. Rosaura specimen Holotype Total length . 4 , : 0 5 152 mm. 3 “ Longueur Standard length : = 5 150 164 mm.” Percentages of standard length Head length . + 6 3 27 mm, 18:0% . 176% Tip of snout to pectoral fin c : 2 21°3% ay n. pelvic ie : A A 49 32°6% 5 — ” » vent . : . : 55 36°6% + 6a. 37°5% (3/8) ql . analfin . 5 5 5 57 38:0% ° — 9 dorsal fin . “ 6 : 75 500% — Greatest depth of head. c qj 12-4 8-2% _— i ,, body behind head . 5‘ II 7°3% 6-7% Pr thickness of body behind head. 8 573% 4:2% Depth at vent a 0'7 6-4% _ Length of pectoral fin ; : 5 ; I2°2 8-1% — Longest ray of dorsal fin . : 5 3 8 —_— 3 a anal oe. . 9 — — Least depth of caudal peduncle 0O-7 0°4% = —_ Percentages of head length Snout 69mm, 25°5% 27°5% Eye : : : . : 9°7 35°9% 37°9% Interorbital width E 5 6 Z 6 2°4 8-8% r7°2% Interocular ,, ci 5°4 20:0% Upper jaw : ‘ : “ . 83 30°7% — Lower ,, 3 : : : 974 34°8% _— Greatest depth of head . 2 « - 12+4 45°9% —= ns breadth ,, (across eyes) . é 14°6 54°0% _ + I am unable to find the “single row of minute conical teeth at the anterior end of the maxillary, described by Mead & Béhlke (1953). But theirs is a larger specimen (190: 150 mm.). THE “ROSAURA” EXPEDITION 187 extending far to the side, and about 12-15 on the dentary. The teeth in each jaw are simple, slightly incurved, readily depressible fanglets (Text-fig. 11), and do not exceed 0-2 mm. in length in the largest specimen available. There is a danger of being misled into believing that the dentition is multiseriate, because rows of white structures (believed to be taste-buds rather than photophores) stand imme- diately behind the teeth, as well as being lavishly scattered over the lips (Text-fig. 11), and buccal cavity. The teeth may be largely concealed behind the lips; this condition varies in different specimens. There seem to be no palatine or vomerine teeth. Gill-opening wide, though only extending upward to the level of the middle of the eye; membranes voluminous, posteriorly separate and free from the isthmus. The six branchiostegal rays and the opercular bones which together form the gill- cover are all concealed under heavily-pigmented skin, so that only the more promi- nent ridges of the bones show white at the surface and their boundaries are largely concealed. The preopercular has a long elbowed ridge extending from a point midway between the eye and the top of the gill-opening to the articulation of the jaw. The interopercular accounts for another ridge arising in the angle between pre- and sub-opercular and running some way with the former. The opercular has about a dozen ridges, some of which terminate in small denticulations at the hinder margin. The subopercular has one principal ridge which supports a small lobe at the lower posterior corner. The branchiostegal rays are slender and increase in length posteriorly ; the sixth appears (without dissection) to be a rather stouter structure of T-shaped cross section (Hubbs’ ‘“‘ branchioperculum ”’). Gills 4; a pseudobranch of half-a-dozen filaments present on the inside of the gill-cover. Gill-rakers 12 on lower limb of first arch ; lanceolate with a few minute bristles. Gill-filaments rather short, leaf-like, shorter than rakers. Pores of lateralis system of head few and large; 3 in the supra-temporal series, 3 or more in the supra-orbital, 4 in the infra-orbital, 5 in the preoperculo-mandibular (Text-fig. 12A). There may be others on the snout. The thin skin of the body has been stripped forward and hangs in tatters about the head ; it is possible, however, to paint a few centimetres of it back into position with a fine brush, and to decide pretty certainly that there were no lateral line pores on the body. Except that the skin tends to fold into a groove of uncertain significance neither is there any sign of a lateral-line canal upon the skin, nor upon the horizontal septum of the body. There are, however, upon the head and along the skin of the body large numbers of “‘ tag organs’’ or lateral line papillae, in tracts corresponding to the theoretical lateralis distribution although quite possibly independently innervated (Text-fig. 12A, B). These papillae are compressed, closed, and of much the same shape and variability of form as the common littoral sponge Grantia compressa. The papillae of the body (Text-fig. 12c) and operculum are longest, and approach 0-3 mm.; they are heavily pigmented save at the tip and in a very narrow marginal zone, and have a core of connective tissue which is in some cases seen to be penetrated by a short branch of a nerve running along the papilla-tract. It is presumed that sensory cells were located in the tips of the papilla, but surface magnification reveals none, and since the fishes have taken a 188 THE “ROSAURA” EXPEDITION B TEXT-FIG. 12. Leptoderma macrops. a. Diagram showing disposition of lateral line pores and sensory papillae over the head and trunk. 8. Three papillae of the sub- orbital series, and a papilla and a pore of the preoperculo-mandibular series. (The area enclosed in the square on A, more highly magnified.) c. A single sensory papilla from the body. (That arrowed in a, more highly magnified). great deal of punishment in the trawl the chances of their preservation are con- sidered too slight to justify sectioning for the present. Upon the head the papillae become shorter and squarer, until on the tip of the snout and about the lips they become very small, lose their pigment, and become superficially identical with the presumed taste-organs inside the mouth. (See further discussion on p. 193.) The vent opens on a slight prominence exactly three-eighths the standard length from the tip of the snout. Pectoral fins long, narrow, obliquely inserted about one-fifth the body-height —— THE “ROSAURA” EXPEDITION 189 from the ventral surface. All the rays except one are deeply forked, nearly to their bases. Pelvics ventral, abdominal, separated by the width of one base, reaching to the anal. Dorsal and anal fins very long, extending almost to the caudal, and each erected fin much taller than the body depth at the same level. The anal commences immediately behind the vent ; the dorsal much further back at the level of the 15th anal ray. Caudal with a distinct and very slender peduncle, ending in a rather prominent little fan-like expansion of hypurals. The caudal rays are much damaged, and both the length and the count likely to be underestimated. Colour (in spirit) deep purple-black all over the body, inside the buccal cavity and gill-chamber (and probably inside the body-cavity) ; considerably lighter on the snout tip and eyeballs. There is no evidence of organized photophores, though this does not preclude general luminescence of the epidermal mucus. The largest specimen to hand is a female, with opaque white eggs of varying diameters (0-4-1-2 mm.) extruding from the vent. The extrusion is probably a decompression phenomenon, although Vaillant has already commented upon the great variation in the sizes of the eggs in this species. COMPARISON WITH VAILLANT’S DESCRIPTION (1888) Vaillant’s description is comprehensive, detailed, and for the most part more accurate than most work of his period. Despite the few marked discrepancies now to be noticed there could be no reasonable doubt but that the present material should be referred to Vaillant’s L. macrops, even without the advantage of a para- type for comparison. The new account is more an expansion of Vaillant’s than a correction, although paucity of material leaves his account of the internal anatomy unchecked. Vaillant gives D.48, A.71 +, and his figure agrees, showing these fins extending nearly to the caudal. The paratype now to hand, a specimen of 123 mm. S.L., gives practically these counts in median fins ending abruptly more than a centimetre before the caudal, and beyond their terminations there remains little more than vertebral column and a few tattered caudal rays. Having regard to the spacing of the posterior rays, and assuming that the fins once reached nearly to the caudal, as they do in Vaillant’s figure and in the “ Rosaura’”’ material (which, it must be noted, was collected not far from the type-locality and at a similar depth), I estimate D.46 + ca. 20, A.74-+ ca. 12, a precise correspondence with the ‘“‘ Rosaura”’ specimens. It is possible that Vaillant’s artist reconstructed the dorsal, anal and caudal in their correct relations, perhaps from a better specimen, and fitted in the Tays in accordance with counts supplied by Vaillant himself from a different fish. Later in his text Vaillant gives P.15 (uncountable from the figure). Though at first I agreed, counting from the projection of the fin through a hole in the loose skin, I later found that I had merely repeated his error in counting the branches of rays which bifurcate close to their bases. Of the ventrals Vaillant remarks only 190 THE “ROSAURA” EXPEDITION that they are “‘ si rapprochées qu’elles semblent se confondre’”’; the figure appears to show V.5 or 6 and I count V.5 on the specimen. The measurements available and shown in the comparative table (p. 186) show close agreement, except at one point where it seems evident that Vaillant’s “ inter- orbital ’’ corresponds with my “‘interocular.’’ He gives B.V + instead of B.VI; he slightly undervalues the length of the maxillary ; he mentions the dorsal fin originating over the 23rd anal ray instead of the 15th—a likely variation and one common in other elongated fishes, such as Notacanthus for example; and he fails to notice the sensory papillae, which, with the condition of his material and the probable magnification used in examining his specimens, is understandable enough. (Vaillant seems to have been impressed by the minuteness of the teeth—“ assez peu developpées, pour n’étre visibles qu’a un assez fort grossissement ’’—and the papillae of the lips and snout are smaller still, while, having decided that the lateral line was absent, Vaillant might well have felt no need to examine the skin of the body.) COMPARISON WITH LEPTODERMA AFFINIS ALCOCK, 1899 Alcock (1892) described as L. macrops a single fish of 8? inches (222 mm.) S.L., trawled by the “ Investigator ”’ at 753 fathoms (1,357 metres) in the Bay of Bengal ; later (1899, 1900, 1902) he published a fuller description and figured it as L. affinis. Mr. A. E. Parr (Director of the American Museum of Natural History, at present engaged on a monograph of the Alepocephalidae) tells me that on inquiry of the Indian Museum, Calcutta, recently he learned that Alcock’s unique holotype had been lost. No other records of Indian Ocean material have been given and Alcock’s papers therefore provide the only evidence for consideration. The invariably damaged condition of L. macrops in collections has been empha- sized ; L. affinis stands in striking contrast, although the conditions of its capture were not particularly auspicious. Alcock makes no mention of the skin peeling from the body and is able to describe the lateral line in its entirety ; thus (1899) : “ a row of pores extends from the occiput to the caudal. In spirit the colour is purple, the contracted opaline epidermis forming a sort of bloom.” Again, the popular account (1902) : a it is quite black and has no special phosphorescent glands, yet its entire skin is enveloped in a thick, opalescent epidermis, like a luminous ‘bloom.’ The only specimen captured glimmered like a ghost as it lay dead at the bottom of a pail of turbid seawater.” We can be satisfied as to the perfect condition of Alcock’s specimen, and wonder indeed whether its apparent resistance to abrasion in the trawl may not indicate a specific difference from the Atlantic form. Alcock (1899) concludes : “This species seems to differ from Leptoderma macrops Vaillant in having the body less elongate, the lateral line very distinct, and the rays of the dorsal and anal fins more numerous. It agrees fairly well with the figure but not with the description of that species, and is probably identical with it.” DHE “ROSAURA? EXPEDITION I9I The revised fin-ray counts for L. macrops show a remarkable similarity to those for L. afinis. Adding pectoral and caudal counts taken from Alcock’s figure (1900) to the incomplete formula published in 1899, we have : L. afinis D.ca.66; A.ca.85; P.8; V.5; C.17 L. macrops D.66 ; A.86; P.8; V.5-6; C.16 Despite this similarity there remain several serious discrepancies between the two species. (1) Body proportions—From Alcock’s publications it may be deduced that in L. affinis the body-height measures 12:6% and the head-length 25-7% in a standard length of 222 mm. (Alcock does not state whether “standard” or “ total ’’ length is intended in this case, but his frequent mention of “length without caudal ”’ elsewhere in his report of 1899 sufficiently indicates his custom. The proportions given, being taken from his figure, are unaffected by this consideration). Com- parative figures are 6-7%, 17°6%, 164 mm. for Vaillant’s holotype, and 7-3%, 18%, 150 mm. for the largest “ Rosaura’”’ specimen of L. macrops. It would be hard to attribute these wide differences to the effects of ontogenetic changes, geographic variation and the fortunes of preservation. There is in fact evidence that the relative size of the head decreases with age in L. macrops, whereas the identification of L. affinis with L. macrops would require a contrary tendency. L. macrops af a _essésiC:*“: S'afvnsiss “ Rosaura’”’ Paratype “ Rosaura ”” Holotype (Alcock) Standard length (mm.) . 85 eet 23) 5 T3008 oc 150 A 164 . 222 Headas%S.L. . e Sen Oia oct U Tal beady SEAS ug TOPOL eak ETO ania One further explanation may be considered. The caudal peduncle in Alcock’s figure is short and deep; despite the disarming elegance of the caudal fork is it possible that the figure represents a truncated and regenerated tail? If so, then estimates of the length of the missing portion based on computations of the length necessary to reconcile the proportions of L. affinis with those of L. macrops should show reasonable agreement. Estimated L. macrops L. affinis truncation. Standard length A c 5 150 mm. f 222 mm. . — Head length . 6 a : 18-0% 4 Bir op) c 94 mm. Snout to anal 9 : 38-0% - TOS: Pos (oo Ee Snout to dorsal 4 : « 50:0% A I29 5, . {0 fy This hypothesis is clearly untenable also. (2) Lateral line——Alcock comments upon the apparent difference between his species and Vaillant’s in this respect ; it was a point which he would undoubtedly have verified. I have examined the skin and bodies of four specimens of L. macrops and find no pores save on the head ; it seems hardly likely that the stripping of the skin should destroy them so thoroughly as to defy microscopic examination. Yet is Alcock truly representing the case when he speaks of “a line of pores’? His figure suggests, rather, a line of sense organs lying flush with the skin. 192 THE “ROSAURA” EXPEDITION (3) Sensory papillae——Nineteenth-century workers! did not often use very high magnifications when examining fishes, and Alcock may have overlooked papillae for this reason, as did Vaillant. (4) Supramaxillary.—I find one supramaxillary in L. macrops, as also did Vaillant —‘un sus-maxillaire distinct,’ “un petit osselet en serpe.’’ Alcock (1899) on the other hand speaks of a maxilla which “is very broad and consists of three pieces.”” It is unthinkable that he should have included the entire premaxillary in this trio, but very likely, as his figure seems to indicate, that he mistakenly dissociated the posterior process of the premaxillary from that bone as a second supramaxillary. (5) Pectoral fin—The pectoral of L. macrops is low on the body and almost horizontally inserted. Alcock’s figure shows that of L. affinis much higher and practically vertically inserted. The evidence seems at the very least to support the retention of Leptoderma affinis Alcock as a distinct species. If Alcock reported correctly the presence of two supramaxillaries, moreover, we are faced with a distinction of generic importance, and one which Parr (1951) uses as the basis of the first dichotomy of his key to the genera of Alepocephalidae. Conversely, with the retention of L. affinis in Leptoderma, the key drafted by Parr will require modification of the stated body- proportions, as well as of the fin-ray counts corrected by the present work. BIONOMICS AND ADAPTATIONS OF LEPTODERMA MACROPS the ‘ Michael Sars’ at Station 48, between the Canaries and the Azores, brought up an Alepocephalus in the large trawl towed at the bottom in 5,000 metres, just as these fishes have been captured by most deep sea expeditions; on the trawl-rope a small tow-net was fixed in such a way that it was towed about 1,000 metres above the bottom and in this net an Alepocephalus was also captured. Such facts warn us against hasty conclusions. Murray & Hyjort, 1912, The Depths of the Ocean: 412. London. Leptoderma macrops has (on published records) been taken in ten trawl hauls ; by the ‘‘ Talisman ”’ (6), ‘“ Caudan”’ (x), ‘‘ Princesse Alice II” (1), ‘Oregon’ (x), and now the “ Rosaura ”’ (1), in depths ranging from 650-2,330 metres. Of a total of 65 speci- mens 59 were collected by the “ Talisman’’, 47 at St. 93 on the Bane d’Arguin. This paucity of records is remarkable when one considers the number of bottom hauls made by, say, the “ Michael Sars,” within the known range of the species and without results. Unless the “‘ Talisman” encountered a breeding concentration the fish seems singularly capricious in its occurrence. The association of predominantly bottom fauna with L. macrops is indicated in the reports of Vaillant (1888) and of Roule & Angel (1933). At “ Rosaura’”’ St. 49 it was accompanied by 2 Bathypterois dubius Vaillant (with minute eyes and tactile pectorals) ; 5 Gadomus longifilis (G. & B.), 1 Gadomus arcuatus (G. & B.), 5 Bathy- gadus vaillantt Roule & Angel, 1 Halosaurus owent Johnson, 1 H. johnsonianus Vaillant and 29 Synaphobranchus pinnatus Gronoyv. In anticipation of the in- vertebrate papers, and by courtesy of Dr. M. Burton, Mr. N. Tebble, Dr. I. Gordon, 1 Charles Darwin, for example, had a great preference for the simple microscope, and had no com- pound microscope with him on his “‘ Beagle ’’ voyage (Life and Letters, ed. F. Darwin, 1: 145. London, 1888). deed aoe en M THE “ROSAURA” EXPEDITION 193 Dr. W. J. Rees and Miss A. M. Clark, it is possible to report associated bottom fauna in the groups Porifera, Polychaeta, Decapoda, Mollusca (dead pteropod shells) and Echinodermata. Preliminary assessments indicate that no mid-water species were taken in any group. All the available evidence therefore points to Leptoderma macrops being a fish living on the bottom beyond the Continental Shelf. That it is of the select group of deep-sea fishes which perform considerable vertical migrations is unproven and unlikely : its general appearance makes this seem much more improbable than in the case of Alepocephalus macrolepus Koefoed (1910, Sci. Rep. ‘‘ Michael Sars,” 4: 44), the fish whose exploit is recorded in the cautionary text at the head of this section. Little can be learned from the gill-rakers. The filaments suggest that cutaneous Tespiration has permitted a reduction in their size. The small gape, puny dentition and unathletic form point to a fish living upon detritus and/or small and feeble prey. The jaws are not protractile, though the maxillaries are practically vertical when the mouth is fully opened, so that in one way the following comparison is invalid, but the great number of taste-organs in the mouth may point to the fish living like a deep-sea carp, sucking-detritus from the bottom and sampling it for nutritious particles. (I have not gutted the specimens, for obvious reasons, but a radiograph of the largest shows objects in the stomach which resemble a small lamellibranch shell and a couple of large sponge spicules.) The system of lateral line papillae, obviously useful in warning of enemies, may be used offensively against small prey such as polychaetes and crustaceans, a vibration in the water being followed by a reflex snap and a leisurely assessment of the catch by the taste organs. (See also p. 194ff.) In either case Walls’ dictum holds: ‘‘ Go far enough along the bottom (if you’re a fish), and you’re bound to bump into something good to eat.”’ The excellent summary of the eyes of deep-sea fishes by Walls (1942, “‘ The Vertebrate Eye,” Bull. Cranbrook Inst. Sci. 19 : 391-405) is the starting point for the discussion which follows. Vaillant (1888) has admitted that his published figures do not adequately represent the lateral protrusion of the eyes of Leptoderma. The diagrammatic reconstruction presented in Text-fig. 13 attempts to do this, and, it is believed, with considerable accuracy. The largest ‘‘ Rosaura’”’ specimen was used for it, the left eye being undamaged, and in an apparently normal position in the orbit, and the shrunken cornea was gently pulled out to its proper convexity with fine forceps. The lens is not shown ; in dorsal view it is concealed under the pigmented portion of the eyeball, and whether this condition is normal or not it is certainly very large, much larger than the diameter of the pupil. These eyes are enlarged, but quite unlike those of the fishes of the twilight zone, such as Epigonus and Aphanopus. If Walls is correct in stating that the eyes of Bathytroctes and Platytroctes are ‘“‘ apparently in a half-way stage of evolution toward an eventual forward-aimed, tubular organ”’ like those of Gigantura and Winteria, then it must be admitted that Leptoderma is a great deal more advanced towards the same type. The parallel tubular eyes already cited are, like the upwardly directed ones of Argyropelecus and Opisthoproctus, most likely concerned with space-perception, assessing the distance of the infrequent flashes from the 194 THE “ROSAURA” EXPEDITION photophores of other organisms. With a light penetration limited to 400 metres depth it is most unlikely that Leptoderma can ever aspire even to the twilight zone ; the structure of its eyes makes it seem most unlikely that it can form a distinct image under whatsoever lighting conditions, and yet it is clear from Text-fig. 13 that if light-sensitivity be admitted as “ vision’’ Leptoderma has, by moderate nasad asymmetry and considerable protuberance of the eyes, achieved a very fair degree of stereoscopic vision, if not of binocularity. If its eyes have largely lost their effectiveness as cameras through structural modification and conditions of intense darkness (and it will be noted that the small ocular photophores concerned with neuro-facilitation in fishes which rely on normal vision below the photic zone Text-Fic. 13. Leptoderma macyops. Diagrammatic reconstruction of dorsal surface of head, showing lateral projection of the eyes. are not developed in this genus), it is still obvious that they have a high degree of potential usefulness as range-finder photometers, capable of locating animal and bacterial luminescence and enabling assessment of its potentialities as food or foe. The associated bottom fauna from St. 49 included numerous specimens of Poly- cheles spp. and Hetevocarpus sp., crustacean genera whose Indian Ocean species at least are known to be luminous, and also the Asteroid Brisingella coronata which belongs to a family whose luminescence is in dispute. It is also interesting to note that in this one haul the enlarged eyes of Leptoderma macrops are paralleled in a polychaete Robertianella and a crustacean Hetevocarpus, while the reduced eyes of Bathypterois dubius are parallelled in the crustacea Nephropsis atlantica and Poly- cheles spp. OCCURRENCE AND SIGNIFICANCE OF SENSORY PAPILLAE IN DEEP-SEA FISHES Papilliform lateral line organs have been reported in several quite unrelated groups of deep-sea fishes. Zugmayer (1911, Résult. Camp. Sci. Monaco, 35: 92, pl. iv, fig. 3) records them among the Lyomeri in Gastrostomus (= Eurypharynx). Pfiiller (1914, Z. Naturw. Jena, 52 : I) in an intensive study of the lateral line organs THE “ROSAURA” EXPEDITION 195 of the Macruridae gives accounts and figures sections of papillae in Coelorhynchus and Macrurus. Roule & Bertin (1929, Dana Reps. No. 4:12) mention the co- existence of lateral line pores and papillae in the deep-sea eel Nemichthys, and the same authors (op. cit.:53) and Trewavas (1933, Proc. zool. Soc. London : 601) describe a variant of the same condition in the nearly-related Cyema. The most comprehensive demonstration of papilla variation is given by Regan & Trewavas (1932, Dana Reps. No. 2:23), who describe and figure five distinct types in the Ceratioid Angler-fishes, and show that the several families may be classified into five corresponding groups. Their Dolopichthys papilla approaches most nearly to that of Leptoderma. Among the Alepocephalidae, however, no very precise demonstration seems to have been given before the present. In the Alepocephalidae sensu Jato (including Parr’s Searsidae), Brauer (1906, Wiss. Evgeb. Deutsch. Tiefsee-Exped. “ Valdivia,” 15: 17 et seg.) describes “ papillen”’ in several species, with recognizable figures, but makes no functional distinction between photophores general and special, the “shoulder organs”’ of Searsidae, and true lateral line papillae. The species mentioned by Brauer, in none of which seems a Leptoderma condition to be clearly demonstrated, are: ” Fam. Alepocephalidae Bathytroctes longifilis (= Nemabathytroctes longifilis (Brauer)). Aleposomus lividus (= Rouleina lividus (Brauer)). Aleposomus nudus (= Rouleina nudus (Brauer)). Fam. Searsidae Bathytroctes rostratus (nec Giinther) (= Searsia koefoedi Parr). Beebe (1933, Zoologica, N.Y. 16:17) describes a pattern of tubercles in his Anomalopterus megalops, but which can hardly be reconciled with a lateralis dis- tribution (nor, for that matter, Beebe’s species with Anomalopterus Vaillant (now Anomalopterichthys Whitley, the name being several times preoccupied). Beebe (op. cit. : 83) also describes small stalked structures in Photostylus which he regards as photophores, but which may well be primarily sensory papillae; these also are not regularly arranged. Unfortunately many genera of Alepocephalidae are not yet represented in the collections of the British Museum (Natural History) but a rapid search for other forms bearing lateral line papillae has been rewarded. The holotype of Xeno- dermichthys nodulosus Giinther (1887, Sci. Rep. ‘‘ Challenger’’ Zool. 22 : 230, pl. 58, fig. c) is much as Giinther described it, with a normal tubular lateral line, and a tough skin scattered with minute white scale-like structures and small, nodular and undoubted photophores. Somewhat surprisingly Giinther failed to notice that almost the whole of the head (including the isthmus) is unpigmented (save for the photophore rims), intact, and sharply distinguished from the body; a few years later and he must surely have noted an appearance strongly recalling the all- luminous head of Azlastomatomorpha. He also missed the lateral line papillae : minute, club-shaped, with slender colourless attachments and black tips (Text- ZOOL. 2, 6. 12 196 THE “ROSAURA” EXPEDITION figs. 14 and 15). They are arranged in single and double linear tracts upon the head, corresponding to the /ateralis distribution, and are upon the body rather sparsely placed immediately above the lateral line tube and upon the tube itself, either between or above the lateral line pores. Down the upper limit of the gill- cover, along the boundary between pigmented and unpigmented integument, the papillae are more elongate and their stalks whiter; they resemble photographic 5 MM. TExT-FIG. 14. Xenodermichthys nodulosus Giinther. (Holotype, 180 mm. S.L.). Portion of skin from the left flank. The lateral line, LL, crosses the centre of the figure. po, lateral line pore; pa, papillae; s, rudimentary scale, three of which may be seen supporting the outer wall of the lateral line canal from within; pH, large photophores, arranged in alternating segmental rows so as to mark out rhomboidal areas on the body ; MP, very numerous white spots, possibly luminous (mucus glands according to von Lendenfeld), arranged in close longitudinal rows. negatives of the fungus Xylaria hypoxylon. Von Lendenfeld (1887, “‘ Challenger” Reps. Zool. 22, Appendix B. : 307, pl. 53, figs. 49-53) has described the histology of the photophores and mucus glands, but again does not mention sensory papillae. It is likely that he was given a piece of tissue from the flank (missing in the speci- men) which would not have borne any papillae, and that he did not receive the entire fish for examination. Rauther (1927, Bronns Tierreich, 6, Abt. 1, Book 2: 125) in his discussion of light organs in fishes regards the first-mentioned struc- tures as sensory organs. Convergent evolution of lateral line papillae in several widely separated groups of deep-sea fishes has not yet received an explanation, despite its clear adaptive THE “ROSAURA” EXPEDITION 197 value. All the fishes concerned are forms of inconsiderable locomotive powers, not subject to violent water movements consequent upon their own swimming, and living at depths beyond the range of wave action and of strong currents. In such conditions extreme exposure of lateral line sense organs upon papillae results in no redundant irritation to the fish; on the contrary, it must materially assist the apprehension of the slightest local disturbance in the water, and be a vital asset in the collection of information from an environment that can impart so 0:5 MM. B TEXT-FIG. 15. Xenodermichthys nodulosus. Three types of sensory papillae. a. Stout papilla from occipital region, at the boundary between pigmented and unpigmented integument. The papilla bears two large photophores towards its base, and has a number of minute, intense white spots scattered over its surface, probably further photophores. B. Clavate papilla standing upon the supra-orbital canal. The papilla is traversed by a nerve, and has presumed minute photophores over its surface. This is a type which punctuates rows of similar papillae of about half the height in the lateralis system of the head. A similar size and type of papilla to this stands upon and above the lateral line canal of the body. c. Slender-stalked papilla, type confined to upper margin of operculum, along boundary between pigmented and unpigmented integument. (Text-figs. A, B and c are drawn to the same scale.) little so reluctantly to the more orthodox sense organs. It is probable that the lateralis system in these fishes allows precise location and capture of small prey without any visual assistance. Experimental evidence in support of this theory may be derived from the experiments of Dykgraaf (1932, Z. vergl. Physiol. 17 : 802), who showed that the freshwater minnow Phoxinus can be trained to snap at small bodies gently agitated in the vicinity of its caudal lateral line papillae; in this case, however, the sensory cells are situated at the bases of the tubular papillae, instead of at the tips as in the deep-sea fishes so far-investigated. For a general introduction to this subject see Pfiiller (ante) and Wunder, 1936 (in Demoll & Maier, Handbuch der Binnenfischeret Mitteleuropas, 2B: 49. Stuttgart). 198 THE “ROSAURA” EXPEDITION Family SEARSIDAE. Parr, 1951, American Museum Novitates, No. 1531 : 1-21. ALEPOCEPHALIDAE (part) Giinther, 1887, ‘‘ Challenger ’’ Reps. Zool. 22 : 227-8. Brauer, 1906, Wiss. Evgeb. deutsch. Tiefsee-Exped. ‘‘ Valdivia,” 15: 17-18. Beebe, 1933, Zoologica N.Y. 16 : 36-56. Parr, 1937, Bull. Bingham Oceanogr. Coll. 3 : 12-19. And numerous other authors. NOTES ON TAXONOMIC AND STRUCTURAL FEATURES (1) Shoulder organ The Searsidae have been separated off from the Alepocephalidae by Parr (1951), and are characterized by the possession of a peculiar structure which Parr, in this, his latest paper, calls the “ shoulder organ,’ which name is the least unsuitable ~~ lateral line Supractaviculare process TExT-FIG. 16. Diagram of shoulder organ (‘‘ supraclavicular process’’) in a young Searsid, showing its relation to pectoral girdle and lateral line. pto, post-temporal ; pel, supracleithrum ; cl, cleithrum. (From Beebe, 1933.) of those so far applied. The new family forms a quite distinct group, and its recogni- tion seems surprisingly overdue, although one regrets its immediate proliferation of new subfamilies, genera, and subgenera. The shoulder organ, @/ias “‘ supraclavicular organ/process ’’’ and “ postclavicular organ,” lies neither upon nor behind the shoulder-girdle, but within it (Text-fig. 16). THE “ROSAURA” EXPEDITION 199 Overlooked by nineteenth century authors and their artists (as well as several later ones) it seems first to have been noted by Brauer (1906) : “Dorsal von der Brustflosse liegt eine gréssere nach hinten gerichtete Papille.” Beebe (1933) notes its gross structure, provides the figure here copied, and mentions ontogenetic changes : “In cleared specimens the supraclavicular process is seen to be a retort- shaped sac, with a posteriorly directed neck terminating in a single pore. This seems to be the only opening, external or internal, to the organ, which arises just interior to the pectoral girdle at the junction of the supracleithrum and cleithrum, well below the lateral line. The posterior part of the sac and the entire neck are exterior to the body wall. It is possible that, instead of being greatly reduced in relative size in larger specimens, as is generally believed, the organ merely has a smaller part of its surface projecting exteriorly.”’ To this account I can add that in the smaller “‘ Rosaura ’”’ specimen now available (Text-fig. 17) there is evident a complication in the neck of the shoulder-organ. Its opening is oblique, like that of a hypodermic needle, and its duct is supported by a tubular reinforcement, a modified scale or thin bone (? rudimentary post- cleithrum). This skeletal structure is flexible, and penetrates deeply towards the sac of the shoulder organ. Its presence in small Searsids in which the squamation is undeveloped rather points to its being a specialized bone. I have not proceeded further with this investigation for the present, since most of the British Museum material is taxonomically significant as well as scanty, and haphazard dissection undesirable. Part (1937) first attached systematic importance to the shoulder organ in erecting the genus Searsia. In his latest publication (1951) he omits mention of reinforce- ment of the external tube, but adds certain information to Beebe’s earlier account : “The sac is lined with black integument and traversed from outer to inner wall by irregular strands or columns of soft tissues. There can be no doubt that this represents a secretory organ, and it seems quite likely that its function may be to secrete a luminous mucus. The sac discharges through a tube, also lined with black tissue ... (this tube) is apparently quite strong since it seems to withstand considerable abrasion and the loss of all scales.” Is it not possible to go further than this, on the anatomical evidence now available ? Everything points to the shoulder organ being a structure for the reception of luminous particles which are effectively blacked-out by the densest pigment in the body until required, and then expelled in a jet to form a luminous protective cloud, as in the squid Heteroteuthis dispar and in several deep-sea crustacea—A canthephyra pellucida, Heterocarpus alfonsi, Pandalus alcocki, Aristaeus coruscans, etc. (For encyclopaedic bibliography see Newton Harvey, 1952, Bioluminescence. New York). It is true that a ductus ejaculatorius or muscles in the wall of the sac have yet to be demon- 200 THE “ROSAURA” EXPEDITION strated, but even in the absence of these a mesiad movement of the cleithrum could readily induce ejaculation: very light pressure on the pectoral girdle of a preserved specimen produces a strong current of preservative fluid. If this case is established it is the first instance of a luminous cloud-thrower to be found among fishes. Ventral saccular luminous organs are known in many Macrourids and in Physiculus among the Gadids (references in Newton Harvey, ante), but anatomical evidence and field observation in these cases point to incom- plete concealment of the luminous contents of the glands, and to a gradual emission of luminous slime onto the body. (2) Light organs Light organs in animals are known or believed to subserve a variety of functions : attraction of prey; diversion and confusion of pursuing predators ; local environ- ment illumination as an aid to vision; neuro-facilitation of the optic nerve ; self- indication of the state of the whole photophore system by means of “‘ pilot lights ”’ ; intra-group (species) recognition; sex recognition, advertisement and maturity indication. So much is generally known, but it does not seem to be as generally realized that the relative evolutionary potentialities and consequent taxonomic significance of each type of photophore system must vary very widely in relation to the function performed. In many cases selection-pressure will favour lighting systems having greater efficiency, and consequently bring about the convergent evolution of similar arrange- ments in unrelated groups. Analogous luminous cloud-projection devices in squids, crustacea and fishes have been discussed in the previous section (p. 199) ; special modification of homologous structures may also be replicated, as for example the terminal photophores on the first dorsal rays of Chauliodus and the Ceratioid Angler-fishes and the minute light organs on the eyeballs of Chauliodus, Persparsia and certain new Myctophids in the “ Rosaura ”’ Collection. Once a luminous organ or organ-complex has attained a certain measure of efficiency, selection-pressure may, while maintaining that standard, permit such variation in detail as does not detract from the usefulness of the whole. Thus it happens that in the illicium (esca) of the Ceratioids and in the mental barbel of the higher Stomiatoids there is considerable inter-specific (and, to an unknown extent, intra-specific) variation. In the latter group it has sometimes been possible, without affecting the attractiveness of the mental barbel to prey, to superimpose a sexual dimorphism and thus fulfil two functions with the same structure. Only in the case of recognition marks is there a strong tendency towards differen- tiation of photophore patterns, and that in definite steps with a rigorous disciplining of the range of variation at each stage. Selection-pressure in favour of such differentiation is least intense under conditions of allopatric speciation (which in the marine habitat has great possibilities in the third dimension, e.g., bathymetric stratification of species of Cyclothone) and strongest under conditions of sympatric speciation. The tendency is well demonstrated by the wide divergences between the fundamental photophore patterns of the deep-sea squids, the several families THE “ROSAURA” EXPEDITION 201 of Stomiatoid fishes and the Myctophids, and by the wonderful diversity within each of these groups. It follows, therefore, that in cases where there is reason to believe that intra- species recognition is the primary function of a photophore system the variations of that system may be expected to provide the most sensitive index of speciation. Moreover, variations in recognition patterns, like variations in physiological tolerances, breeding behaviour, etc., themselves constitute isolating factors, and may thus precede or even initiate further genetic and therefore morphological differentiation of populations. Conversely, if, after making due allowance for the disparity between the taxono- mist’s and the animals’ assessment of a variation, that variation does not seem likely to be clearly characteristic at a distance, then one should be reluctant to apply it as an index of speciation in sympatric populations. Having stated these general principles we may now consider the light organs in the two species introduced by Parr (1937) as Searsia koefoedi and S. polycoeca. Modest and probably relatively inconspicuous as photophore systems go, clearly distinguishable from the basic patterns of the Myctophidae and Stomiatoidea by their very economy, localized at a few points along the ventral surface, there can be little doubt that their primary function is intra-specific recognition. Yet in these two species the fundamental patterns are, in lateral view, absolutely identical (assuming no difference of colour): five lights placed at equivalent points along a straight line. Only in ventral view does the relatively subtle difference in disposition and relative sizes of the subventrals and postventrals appear (cf. Parr’s figs. 4 and .5), and although the eyes of Searsidae have probably a wide field of vision (particu- larly in the young stages; my Text-fig. 19), there is no such obvious connection between upward binocular vision and downwardly directed photophores, as can be demonstrated in the Sternoptychidae. The differences between the lateral views of the photophore systems in these two species and in Persparsia, another genus of the same family (fost, p. 209) support the case for lateral distant recognition. Beebe, (1926, The Arcturus Adventure, N.Y.: 216, pl. 5) has very beautifully illustrated the sexual dimorphism in Myctophum coccoi ; the male has conspicuous luminous scales along the upper side of the tail and the female along the lower side. To me the divergence between Searsia koefoedi and S. polycoeca for long seemed cf much the same order and likely significance; a divergence of far less use than that in Myctophum as a means of recognition at a distance, but of likely utility in identifying the sexes to one another in a group of individuals of one species brought together by the general species recognition pattern of the remaining photo- phores. On that hypothesis I made a considerable effort (Table II) to justify the treatment of S. polycoeca as a synonym of S. koefoedi, but after the examination of material of each I am driven to accept them as distinct species. It will be interesting to see whether the promised ‘ Dana ’ Report on the Searsidae will fulfil the prediction now made that the Atlantic populations of these two species will prove to be geo- graphically and/or bathymetrically isolated. Although this hypothesis of sexual dimorphism in Searsia has proved unfounded, the general similarity of the photophore patterns remains a powerful argument 202 THE #ROSAURA’ EXPEDITION against the removal of S. polycoeca to another genus—Holtbyrnia—and within that genus to a sub-genus Mentodus containing one other species, H. (M.) vostratus (Giinther), said to be known only from a unique holotype devoid of photophores. The last-mentioned specimen has been badly cut about by successive investigators, but there is no doubt that it originally possessed no trace of light organs. The ventral squamation is complete and intact except at those points where exploratory borings have been made, and experiments on specimens of the more plentiful Persparsia taaningi indicate that abrasion of the photophores without damage to the body is highly unlikely. All the specimens identified as “‘ Bathytroctes rostratus Giinther ”’ by other authors and examined by Parr have proved to be either S. koefoedt or P. taaningi. Parr (1937) supports his argument against ontogenetic atrophy of the light organs in B. rvostratus by citing from the literature a series of Searsids of ascending size in which those structures are present, including one (unfortunately not figured by its original author) of equivalent size to Giinther’s holotype. The critical specimen is still 10 mm. S.L. shorter than Giinther’s, and this may, in terms of age, be rather more significant in fishes apparently near their limit and presumably growing at a correspondingly slower rate. However, Krefft (1953, see my footnote, p. 206) in a paper received while the present was in the press, describes a photo- phore system constant in number and arrangement through a series from 87-174 raabon, SHES 7 A specimen in our collections from “‘ Discovery ’’ St. 2074 (21.ix.37. 10° 10’ N., 21° 13’ W.; closing net, oblique haul; 875-400 metres, sounding 5,148 metres) bears a label “ Holtbyrnia polycoeca Id. A. E. Parr.” The salient characters of this fish are given in Table II, where it will be noted that it gives good agreement with the holotype of that species in all characters except the head length, which runs contrary to the expected allometry. There are, however, discrepancies in the light organs. The supraventrals and supra-anals are present and typical, but the posterior anals are marked only by small patches of dark pigment, the median subcaudal is very faint and has only a trace of pigment, the thoracic organ is very thin if it is present at all, and the subventrals and postventrals are altogether wanting. Since the specimen is in excellent condition I am inclined to regard the photophore system as undergoing degeneration. (3) Notes on taxonomy of Searsids Beebe (1933) has interpreted Giinther’s (1887) figure of the snout of Bathytroctes vostratus in terms of the typical Searsid condition with forwardly directed tusks, and this interpretation was at first disputed by Parr (1937). On examining the specimen I find the premaxillaries thin and somewhat eroded, but the terminal lip-like projections are still present and prove to be hollow and capable of admitting a bristle for a fair distance. I regard these structures therefore as representing the broken stumps of tusks with their related pulp-cavities; alternatively and less probably, as the premaxillary sockets of tusks which have been shed (see also p. 208). The increased number of accessory teeth is a function of age and size, while the elaborate crenellation of the edge of the jaw is of a type common on the Tae rma pea = SE = THE “ROSAURA” EXPEDITION 203 _ edges of thin bones which have completed their major growth, comparable, for example, with the elaboration of the skull crests in Melamphées. Parr has also complained that Beebe (1933) shows a shoulder organ in a figure stated to be based on Giinther’s when the original and description do not include it. Perhaps Beebe saw the specimen at some time and erred only in the legend to his figure, otherwise it must appear very odd that both Beebe and Parr are correct ; there is a shoulder organ present in the specimen, though not shown in the original figure. Only the tip of the tube, with its internal support, projects from the body, but the sac must still be quite large, for a fair amount of preservative fluid and air bubbles can be extruded from the tube by exerting moderate pressure on the shoulder girdle. Parr (1951) modifies his earlier standpoint on this species by placing B. rostratus fairly among the Searsidae as Holtbyrnia (Mentodus) rostratus. Though he does not elaborate his reasons in what, after all, only purports to be a preliminary re- vision, we can fairly assume that he has made the above observations during one of his vists to the British Museum (Natural History). The principal characters of three species of Searsids are tabulated in Table II. A number of identical or nearly identical features are omitted for brevity’s sake, but all the salient points have been included. Principal specific differences relate to the light organs (already discussed p. 201), the proportions of the head and upper jaw to standard length, the gill-raker counts, presence or absence of a dermal pit behind the shoulder, and the number and nature of the pyloric caeca. Having tested and rejected the hypothesis that S. koefoedi and S. polycoeca were sexes of the same species, I hoped at least to find support for a theory that S. polycoeca was a young stage of B. rostvatus. Unfortunately this idea must be rejected also, although I am satisfied that B. vostvatus and S. polycoeca are very closely related, and confidently expect that the young stages of B. vostratus when found will prove to possess photophores. I would advocate the inclusion of all three of these species in the genus Searsia as the expression of this close relationship. The importance to be attached to the pyloric caeca may be questioned. Con- sidering Parr’s figures of the pyloric caeca in S. koefoedi and S. polycoeca (1937, figs. ra & B), it may be seen that there is a fundamental similarity between them, for both have five primary caeca along the right side of the stomach and a further _ group of three along the anterior end adjoining these. Even the condition in S. rostratus which has been described as having “14 simple, straight pyloric ap- pendages ’’’ can be reduced to this same basic pattern. Although the immediate impression is of fourteen simple caeca (11 + 3) packed side by side like the fingers of some polydactyl hand, closer inspection shows that there is one group of five primary caeca, of which the anterior one forks twice, the second, third and fourth once each and the last is simple, like the remaining group of three. The forking takes place very close to the stomach, and the sum of the terminal diverticula, I.I.1.4.2.2.2.1, gives the total of 14. It may fairly be asked how much variation is due to individual differences and to ontogenetic changes? The new specimen of S. polycoeca presents a much simpler picture than the smaller holotype, while the é quesqy quesaig snorouinyy snosourn Ny (1661) yuasoig (4€61) Zu uoTjUSWI ON quasolg quasolg 06 “v2 quosqy v4 S$ $x/1/9 S1/1/9 = HB . o. Lal Qa a 8 mye A a ‘ « ie] QuAL Oza] eA {€-ez-g . PeIewo Tez eli ya 02a = 8-9 Ly =| o-Zt g:SI a 0-0Z LLY wo z-9f o0-S¢ fo) z-So c.to ~ o0-Sz 0-61 > L.6 +.6 ia] 6.7 C.1z i ¢.0z g-0z H b.o1 b.o1 1-9 o-Lz zl gt *M £1 (Iz *“M ,6z .F9 Rr On Or SN he7e “blot “4S gf 4S ,, AIBAONSIC, ,, «, DOUME ,, (¢) (ed430]0H) FH vaa0rkjod visavas qoursiao sapurmmad { (rS6r 204 quasqy quasorg quosqy 6/1/0 «, PANBSOY ,, = quesaig ee es quesqy ¢ edAjoj0y sy quasoig quesoig og ‘79 “ g1/1/8 ¢ adAjofoy sy He. . LAK SOs foz-y f1z'q ¢edAjojoysy gb * z-Li ¢.gr 0-02 1-07 9-gb o.1¢ zZ-09 0.49 Z-91 o-Z1 ¢C.L o-Z 6-g1 5a I-91 6.1 9-6 Q-o1 €.1€ g-6z C.ob Le “M ,2% (81 *M ,60 £2 NT Ge get nT SS gee PEAS) 64S ,, AIBAOISICT ,, «, BOUME ,, (adAzereg) quesarq 9 guecorgy < * quesaid ma,t quesqy - quasqy quosaig 2 quesoig (1961) (u10}4 OO! 0} 0g di} yepneo) (Z€61) Sgynoqy * ozi—o1l ‘v9 (1S61) S€—-6z (Z€61) 61/1/Z2 - (qq) 1z/1/or (J) 0z/1/8 oe . 6r a . 8 qeAS ee aT 6'A !L1/gt-gd HOC gLerGien wigieay, mOGaG b.+ 0 $.¢ o-Lt p ¢.o1 ¢.g1 : 0-61 0.18 : L.¢S ¢.to : Q:z 0-61 . O-1z 0-8 0 2:8 ZeOL <9): 6.0% C.F 5 z-O1 0.01 L.g F.of oth 69 gh “MIF 94 "M ,0€ FE “N OE £2 “§ ,£E 08 phe OS) *OzI “4S « DOUMEZ ,, ,, JesueT[eyo ,, (ed430[0}7) (edAj0[0}7) —— wpaofaoy visavas SNIDAJSOA DISADAS + eyided jexuesoursy, yerodu19}-ysod e105 -oq soi0d duly [e10ye'T a[piis 1appnoys puryeq yd jewsed uvsi0 Japynoys + sayeos aul] [e10eT yore ysIy «UO sIOHeI-T[ID, (ydei8orpes) oe1qQ9719 A, * sAei [eSo}sorpoueig P[NUIOF [EIPeY. * JeHeI-TITD e * aseq jeuy aseq [esiod + ye1yuaA—jnous S yes1op—jnous . . > yydeq o * ynous > mel aMoT ° + Areyexeyy * eh Pee :°7]'S Jo sosezuso10g “utd yy Sue] prepuezys ‘LE61) avg mor papduos evovoohjod *s puv ipaojaoy *S fo saddy sof vwq—T] ATaVL 205 EXPEDITION THE “ROSAURA” P[NOTFAVATP yeuruiis} £1 pur seyouriq Arepuodas ZI Yam ‘eoaRd Arewid 9 I (4 s}USWIPNI Maq snorownyy pepore “x b jussolg sox ¢ Iejn00 ON (Qurey) srepneoqns z (qure;) sjeue-o10}Ssod z sjeue-eidns z petydosze syerjueaysod poerydoijze sjeaqueaqns s]e1qUaA -eidns z oroei0yy é (1961 ‘11eq) «/S2]91}I9ATP yeururie, o1 uey} JoMoy ynq ‘payoursq aulos ‘g-9 ,, (Z€6r “3y wo e}eq) *R[NOIFIOAIp yeururia} Sz pure soyoueiq Arepuooos QI ZIM ‘eo0R9 Areuid g zt z 9}yNUIW ‘M97 é Sz-oz + quosolg sox é repno0 oN (‘2€6r ‘reg ur sjreyaq) “sjepneoqns z ‘speue-o10}sod z ‘seue-eidns z ‘(adze]) sjeajuoaysod z ‘(andea) sjeiyuaaqns z ‘sperjuoAeidns z ‘gToRI04} qeoul] 1 P[NOTAIVAIP Tews Auepy “UOT}DOSSIP ul poseueq if @ s}uour -Ipni + z snosouin yy +or ¢ I pajdnis yon ON pedozeaop auoN, P[NIIZAVATP yeuturi9s} £1 pure soyoueiq Arepuoosas z1 YIM “eo9eo Azewtid ¢ oO I syusuIIpny sno1ouin Ny ob—of ynoqy c-b zt (apts Ja] Uo z) quosoig ON é repnoo ]es10p-o10}Ue 1 QOUdIOYIP-Xes ¢ ~paosaeoo aaey 0} Weas sjejueaqns ayy ynq ‘adAjojoy oy} Ul Sy “en “ “ (LM) —P|N9T}IVAIP yeurursz} Fr pure soyouriq (Z€61 “84 Axrepuoosas ZI wor ee) JIM “eo0v9 *R[NOTPIOATp Areutrid g yeuturio} 61 pure soyouriq (1$61 ‘11eq) Arepuosas ,, Sosepuodde Cr yyIM “eo0v9 ors0yAd yysreiys Areutid g SS Gey(obesii yay) o * eoaevo o110]Ag ° y I-o 9 Y}09} ourzEleT z P Zz * —- ja} oUTIOUIO A auou IO solias JOYS * Buoys ‘6-9 ~* sarties 1a3nO snoJoUuIn yy * o9-0S jnoqy * — Ssatioes ure : Yyoo} Arezu0q oS-ob a o$ ynoqy * - yjve} Areyprxeypy 4309} o-+ E C1-hr Axeyixeuiid ry ysnj yora z S Oe sie MOA 1999} [BITFIOA (1-1) quosory * pays 10 ueyoIg; sysn} AreyIxeweig + WOr4T UE ¢ 31q10 jO ulsieur r011aysod ONG ‘ sox 0} soyorer Areyrxep (Z€61) ‘osap jou é “Sy repnoo 1 (-2€6r ‘eg ur sjte}9q) “sjepneoqns z ‘sjeue-o1aysod z ‘speue-eidns z ‘speayuoaysod z ‘speiqusAqns Z ‘sje1juoAeidns z “‘Q1oe10YZ qeoulyr * _ ser0ydozoug 206 THE “ROSAURA” EXPEDITION larger new S. koefoedi is markedly different from the holotype of that species. It may well be that the primary forks remain as they are, while the total of tertiary divercula becomes reduced by subsequent growth at the secondary forks. The British Museum (Natural History) now possesses the following material of this family : I Searsia rostratus (Ginther) Holotype. 2 Searsia koefoedi Parr. I Searsia polycoeca Parr. 5, Persparsia taaningi Parr. 1 Normichthys operosa Parr. I Pellisolus facilis Parr. As regards Parr’s draft classification of 1951, I would modify it as already indi- cated to show the supposed affinities of the three species which I place in Searsia. There arises another problem, common enough when reticulate evolution has to be expressed in a two-dimensional classification, of indicating the affinity which ap- pears to exist between Normichthys operosa and S. polycoeca alone among the Searsia group. Normichthys has a row of six large pits behind the shoulder, each containing three opaque white bodies (? photophores or sense-organs ?), as well as a smaller pore in every scale of the body. The single dermal pit behind the shoulder of S. polycoeca seems to be the precursor of such a condition. The grouping adopted in the key according to relative size of head and length of upper jaw is useful enough for taxonomic purposes, but does not impress me as a natural phyletic classification, for the reasons which have already been indicated. Searsia koefoedi Parr! (Text-fig. 17) Parr, 1937, Bull. Bingham Oceanogr. Coll. 3:16, Fig. 4; Maul, 1948, Bol. Mus. Municipal Funchal, No. 3, Art. 5: 12, Fig. 3; Parr, 1951, American Mus. Novitates, No. 1531 : 17. Bathytroctes (Talismania) homopterus (non Vaillant) (part) Norman, 1930, ‘‘ Discovery’ Reps. 2 : 269 (not fig.). St. 46. 30.xli.37. 7°27’ N., 23° 08’ W. 2-metre Stramin Net, 1oo0o(-o) metres. I specimen, 15-5 mm.S.L. Reg. No. 1953.3.6.12. Has. North and South Atlantic; Caribbean. A young fish and contorted, its body bent downwards from behind the nape and its tail curved forwards along the left flank. The following information expands and supplements that given in Table II. D.20+; A.17+; P.20+; V.7; C.7+:XXII:5 + Branchiostegal rays 7. Gill-rakers on first arch 0/1/9. Proportions in per cent of length without caudal (S.L.): length of head, 34-1 ; snout, 8-3; orbit, 11-6; interocular width, 7-0; upper jaw, 16:7; premaxillary, 4°5; posterior supramaxillary, 5-8; lower jaw, 21-9; snout to ventrals, 60-0 ; snout to dorsal, 66-4; snout to anal, 70:9; greatest depth of body, 18-0; least depth of caudal peduncle, 6-4; dorsal fin base, 16-1; anal fin base, 12-9. 1 A paper by Krefft (1953, Zool. Anz. 151 : 259-266) received while the present work was in the press, describes as new Searsia schnakenbecki from Iceland. 207 THE “ROSAURA” EXPEDITION ‘oF 4S ,, PINeSOY ,, WoL] "TS “wur $.Sr “2pa0faoy visavag sunox ‘Wd | SS ere LI Dla -LKA] 208 THE “ROSAURA” EXPEDITION Least depth of caudal peduncle 35-7 per cent of greatest depth of trunk. 0:0625 S.L. = 0:97; 0:0375 S.L. = 0-58 (data for Parr’s formulae). DENTITION. Only the extreme tip of the forwardly directed tusk on the pre- maxillary is yet erupted. One downwardly directed fang at the front of each pre- maxillary, followed by five slightly smaller teeth which are fairly regularly spaced. Ten similar teeth on the maxillary, with indications of smaller teeth to be erupted between and within these. One pair of fangs on vomer, the largest teeth yet de- veloped. One slender tooth on each palatine. On the dentary a single row of numerous very fine teeth, with a few (4-5) larger vertically directed ones outside these and probably of the same series. Of the horizontally directed series two teeth are erupted at the front end of each dentary and others unerupted may be seen opaque white in the jaw. A tiny projection below the dentary symphysis. Eyes. The orbits are slightly longer than high. The eyes project some way beyond the general contour of the head and must have wide and partly coincident fields of vision. The lens is pale green, large but slightly smaller than the pupil as fixed. There is a greyish-white retinal tapetum lucidum, which may be luminous. CoLour. Dorsally light brown, ventrally dusky. Outer integument over shoulder organ granular dark brown, with darker internal pigment of tube showing through. A darker triangular patch between the shoulder organ and the pectoral fin. A dark line along the lateral line. Fin bases pale. There are no certain indications of light organs. The fins are well advanced, but the full complement of anterior dorsal rays, procurrent caudal rays, and possibly of anterior anal rays, is not complete. Gill-rakers moderately long and widely spaced; their complement also incomplete. Lateral line pores before ascending limb of post-temporal. This specimen agrees very well with Searsia koefoedi in all general proportions. The relative distances from snout to dorsal and anal differ by 4-3% (“less than 4% ’’)—a trivial difference against the likely error in measuring a difficult specimen. Lower counts for teeth, fin-rays and gill-rakers reflect the relatively small size of the fish, and in some respects show an approach to S. polycoeca. Of the three specimens referred to Bathytroctes (Talismania) homopterus Vaillant by Norman (1930), the largest, from ‘‘ Discovery ’’ St. 269, proves to be Normichthys operosa Parr, as already noted in Parr (1951). The second specimen, from the same station, seems no longer to be in this Museum. The third, from an un- numbered “‘ Discovery ’’ station, is a fine Searsia koefoedi whose characters are givenin Table II. It is of interest to note that in this specimen the left premaxillary tusk is laterally duplicated, which may be an indication of periodic replacement and so provides a possible explanation of the tusk-less condition of the holotype of Searsia rostratus. Persparsia taaningi Parr (Text-figs. 18 and 19) Parr, 1951, American Museum Novitates No. 1531 : 18. Bathytroctes vostvatus (nec Giinther) Norman, 1930, “ Discovery’ Reps. 2: 268, Fig. 1; Pl. 2, Fig. 3; Beebe, 1933, Zoologica, N.Y. 16: 43, Figs. 8c, 8p. Searsia n. sp. Parr, 1937, Bull. Bingham Oceanogr. Coll. 3: 16. THE “ROSAURA” EXPEDITION 209 St. 33. 22.xi.37. 11° 00’ N., 75° 43’ W.; 2-metre stramin net; c. I,200(-0) metres. I specimen, 23 mm.S.L. Reg. No. 10953.3.6.13. Has. North and South Atlantic; Caribbean. A young fish and badly contorted, the posterior half of the body twisted and bent forward to lie along the left cheek. D.21; A.14; P.2o+; V.8; C.3+4+:XXII:2+. Branchiostegal rays 7. Gill-rakers on first arch 4/1/12. Proportions in per cent of length without caudal (S.L.): length of head, 40-0 ; snout, 10-8; orbit, 13-9; interocular width, 6-0; upper jaw, 20-8; premaxillary, 6:5; posterior supramaxillary, 7°38; lower jaw, 25:2; snout to ventrals, 62-1; snout to dorsal 63-4; snout to anal, 71-3; greatest depth of body, 16:5; least depth of caudal peduncle, 6-5; dorsal fin base, 20-4; anal fin base, 12-1. Least depth of caudal peduncle 39-4 per cent of greatest depth of trunk. 00625 S.L. = 1-437; 0:0375 S.L. = 0-862 (data for Parr’s formulae). DentiTIon. The forwardly directed tusks are present but not erupted. The premaxillaries are very thin and their edges damaged; teeth that remain at the hinder ends are like those of the maxillary. The maxillary has about 20 teeth with indications of very fine ones between these. Four fangs on vomer, the largest teeth yet developed. Two slender teeth on each palatine. The dentary has a small upwardly directed projection at the symphysis and many acicular teeth in a single row. Eyes. The eyes and orbits are elongated. The eyes project some way beyond the general contour of the head, and must have wide and partly coincident fields of vision, especially in the upward direction (Text-fig. 19). The lens is pale green and as large as the pupil as fixed. There is a greyish-white retinal tapetum lucidum. Cotour. Dorsally light brown, ventrally dusky. Integument over shoulder organ dark brown. A dark line of pigment along the lateral line. Fin bases pale. There are no certain indications of light organs. Gill-rakers long, the longest 0-9 mm. (3:9% S.L.). Gill-filaments numerous, short and stumpy. No lateral line pores before ascending limb of post-temporal. This fish shows a number of small divergences from Parr’s diagnosis (1951), especially in the slightly longer head, longer lower jaw, and more posterior insertion of the ventrals. In view of the youth and twisted condition of the specimen it seems better not to attach much importance to these differences in face of the general agreement with P. taaningi. Norman’s (1930) figures from “ Discovery’ specimens which I identify with this species do not do full justice to the photophore system. In addition to those shown in his drawings and plate there are: two light organs on the eyeball,1 in the upper and lower anterior corners; one on the skin over the lower posterior edge of the orbit ; one near the hinder edge of the operculum, above the level of the pectoral fin base ; in the largest specimen (47 mm. S.L.) a median one at the lower ends of the cleithra, and in the same specimen an additional median photophore between the supraventrals (not to be confused with the curious two-way light _ 3 Searsia schnakenbechi Krefft (1953)has a photophore on the eyeball. Parr (1937) seems to figure one in S. koefoedi though he makes no mention of it in his text. EXPEDITION THE “ROSAURA” 210 “gi ‘Sig ul uMOYS wuruyy) visavdsaaq Sunod Jo peoy JO MIA [esioq, “OI ‘O1d-LXA]L, "QI ‘Old-LXaL ‘WO | \\" \\ ANY AAA THE “ROSAURA” EXPEDITION 211 organ already noted by Norman). In addition to these special organs there are likely indications of a strip of luminous tissue extending from the occiput to the dorsal, and elsewhere about the fin bases. The few specimens available are small, and, while they provide indications that the light organs are not developed simul- taneously, they are inadequate for a study of possible sexual dimorphism. The light organs on the eyeball are quite inconspicuous in the largest specimen avail- able, but there is otherwise no evidence to support ontogenetic atrophy; on the contrary. Doubtless Parr will be dealing with this matter in full in his forthcoming monograph, Family CLUPEIDAE Pellonula afzeliusi Johnels Johnels, 1954, Arkiv. f. zool. Stockholm (2) 6: 352. Pellonula vorax (non Giinther) Svensson, 1933, K. Svensk. Vetensk. Handl. (3) 12, No. 3 ; 47, fig. 16 (young stages). St. 48. 3.1.38. 13° 27’ N., 15° 47’ W.; 2-metre stramin net, 3(-o) metres. 526 specimens, 4-81 mm. S.L. Reg. No’s. 1953.3.6.14-54I. Has. Gambia River. Apart from the other characters noted by Johnels, this species differs from P. vorax in the possession of much more trenchant keeled scutes between the isthmus and the ventrals. Harengula pensacolae Goode & Bean Clupea humevalis (part) Giinther, 1868, Cat. Fish. Brit. Mus., 7: 422; Harengula pensacolae Goode & Bean, 1879, Proc. U.S. Nat. Mus. 2:152; Regan, 1917, Ann. Mag. nat. Hist., (8) 19 : 389. St. (—) 28.x.37. 17° 28’ N., 88° 11’ W.; dip-net, at surface. 5 specimens, 60-98 mm. S.L. Reg. No’s. 1953.3.6.542-6. Has. Gulf of Mexico; Caribbean. Ilisha africana (Bloch) Clupea africana Bloch, 1795, Naturgesch. Ausland. Fische, 9: 45, pl. 407. (Type loc. Accra, Gold Coast.) Clupanodon africanus Lacépéde, 1803, Hist. Nat. Poissons, 5: 469, 471. Pellona iserti Cuvier & Valenciennes, 1847, Hist. Nat. Poissons, 20 : 307. Pellona gabonica Duméril, 1858, Avch. Mus. Paris, 10: 259, pl. 23, figs. 3, 3a. Pellona africana Bleeker, 1863, Natur. Verh. Holl. Maatsch. Wetensch. (2) 18: 122, pl. 27, fig. 1. Pristigastey dolloi Boulenger, 1902, Proc. zool. Soc. Lond. : 271, pl. 30, fig. 3. Ilisha melanota Derscheid, 1924, Rev. Zool. Afr. 12: 278; Irvine, 1947, The Fishes & Fisheries of the Gold Coast, London : 111, fig. 39. Ilisha dolloi Giltay, 1935, Bull. Mus. roy. Hist. nat. Belg. 11, fasc. 36: 3, figs. 1 and 2. Isha africana Fowler, 1936, Bull. Amer. Mus. nat. Hist. 70: 178, fig. 71. Delsman, 1941, Mem. Mus. roy. Hist. nat. Belg. (2) Fasc. 21 : 50. Pristigaster martii (non Agassiz) Fowler, 1936, op. cit. : 180, fig. 72. (non Pellona africana Giinther, 1868, Cat. Fish. Brit. Mus., 7: 455.] [non Ilisha africana Norman, 1922, Ann. Mag. nat. Hist. (9) 11: 4.] ZOOL, 2, 6, 212 THE “ROSAURA” EXPEDITION St. 48. 3.1.38. 13° 27’ N., 15° 47’ W. 2-metre stramin net, 3(—o) metres. 2 specimens, 84-102 mm. S.L. Reg. No’s. 1953.3.6.547-8. Has. Coastal waters and estuaries of W. Africa, Senegal to Angola. Norman (1922) includes J/isha africana in the first section of his key to the genus Ilisha : ““T, Less than 46 scales in a longitudinal series ; 16-20 scutes before insertion of pelvics.”’ Bloch (1795) does not supply this information, but counts from his figure give 46 and 24 + respectively, and therefore rather better agreement with the second section of Norman’s key : “TT. More than 46 scales in a longitudinal series; 20-28 scutes before insertion of pelvics.”’ At the time when Norman wrote, the British Museum (Natural History) did not, in fact, possess any material of Bloch’s species, nor indeed any belonging to the genus J/isha and of undoubted African origin. Norman followed Giinther (1868) in mistakenly identifying with Bloch’s fish a single specimen, B.M. (N.H.) No. 1852.9.13.175, which came originally, and without any locality data what- soever, from the collections of the Zoological Society. This specimen should be assigned to I. indica (Swainson) or I. brachysoma (Bleeker), two doubtfully distinct species from India and the East Indies, from which region it very probably came. Its ventral scutes (20 + 8) give a lower pre-pelvic count than in any African Idisha known, and are much less salient than those in the copious African material of J. africana (Bloch), (including specimens from Accra, the type locality), received at this Museum in recent years. Norman’s error has effectively bedevilled every subsequent author except Fowler (1936), who providentially overlooked his paper in the main body of The Fishes of West Africa and later mentions it without comment in an “ Appendix” to this work. Derscheid (1924) leaned heavily upon Norman’s revision and advice and (Bloch’s book being extremely rare) seems not to have checked the original description and figure of I. africana. Taking very literally Norman’s statement that J. africana (non Bloch) is “scarcely distinct from J. brachysoma,” Derscheid has apparently deduced a pseudo-description of the genuine J. africana by so modifying Norman’s detailed account of J. brachysoma as to reconcile it with the few minor differences noted by Norman in his brief comments on a supposed J. africana which was, as we have seen, not Bloch’s species. Derscheid then proceeds to detail the differences between his African material from the mouth of the Congo and this pseudo- I. africana whose antecedents, if they may be said to have any reality at all, are Asiatic. Naturally enough he fails to recognize his own material as the genuine I. africana (Bloch) and so I. melanota joins the lengthening list of synonyms. Ex Africa semper aliquid novum ! Giltay (1935), having examined the types of I. melanota Derscheid and of Pristi- gaster dolloi Boulenger (1902), makes I. melanota a synonym of I. dolloi (Boulenger). THE “ROSAURA” EXPEDITION 213 This view will be supported in greater detail presently ; for the moment it is merely emphasized that Giltay did not carry his synonymy back any further than Boulenger (tg02), and merely includes “‘ Ilisha africana Norman ... (pro parte)” as an allusion to Norman’s tentative synonymy of P. dollot Boulenger with J. africana (Bloch). Fowler (1936) overlooked Derscheid’s paper as well as Norman’s, and conse- quently makes no contribution to the brachysoma-africana-melanota tangle. In- directly, however, he adds to the confusion by making P. dollo: Boulenger a synonym of Pristigaster martii Agassiz (Spix & Agassiz, 1829, Pisc. Brazil: 55, pl. 24a) and presenting a re-drawn version of Boulenger’s figure of 1902 as “ Pristigaster martit, modified from Boulenger.”’ Ludicrous as it may seem after a comparison with the original figures, the error is liable to prove as dangerous as Norman’s, for Spix & Agassiz is almost as rare a work as Bloch. Delsman (1941) has questioned the validity of J. melanota, but upon the same false premises as led Derscheid to erect his species. Delsman concentrates upon the ventral scutes, which Derscheid gives as 18-20 + 7-8 in his pseudo-I. africana based on Norman’s account of I. brachysoma and 25-27 + 7-8 in I. melanota, and endeavours to close the imaginary gap by citing counts of 22-24 + 7 from new material. Bloch’s original figure (1795) shows ca. 24 + 5; if there is any problem it is presented by the post-pelvic rather than the pre-pelvic count, and if we allow the artist a small error in the pelvic fin insertion we can reconcile this difference without compromising the opening argument in the present article, for it is possible to postulate a correction from ca. 24 + 5 to ca. 22 + 7 without even invoking the probable scutes concealed by the operculum in Bloch’s figure. Norman (1924), Giltay (1935) and Delsman (1941) have all, from their several concepts of J. africana, queried the distinctness of Pristigaster dollot Boulenger (1902). Boulenger introduced this species with a brief description and figure, and no discussion. Derscheid (1924), having compared the type of P. dolloz with his types of J. melanota, is unable to discover any other important difference than that the former lacks even the rudiments of pelvic fins. Giltay, working also from original material, gives an intensive comparison, together with a theoretical disserta- tion on the body-forms of fishes which, curiously enough, fails to make the obvious comparison with the Gastropelecidae. Giltay’s very valuable contribution is to cite the work of Pellegrin (1926, C.R. Acad. Sc. Paris, 183 : 1301) who examined I3 specimens of the African Clariid Clariallabes variabilis Plgr. and found 6 with both ventrals present, 2 with one only, and 5 with none. Hildebrand (1943, Bull. Bingham Oceanogr. Coll. 3 Art. 2: 140) reports I specimen lacking ventrals among 20 Pterengraulis atherinoides (L.) examined by him. Considering such precedents, the minute pelvic fins in J/isha, and the close correspondence between Boulenger’s unique holotype of P. dolloit and the many African specimens of I/isha now examined by several authors, there exist no reasonable grounds for retaining Boulenger’s species as distinct. The considerable African material of I. africana (Bloch) received at the British Museum since Norman’s time may properly be referred to [isha Gray & Richardson as generally understood, since it possesses a maxillary not adherent to the pre- 214 THE “ROSAURA” EXPEDITION maxillary, and rounded at the posterior end ; a ligament rather than a bone between the maxillary and premaxillary (but see Myers, Copeza, 1950 : 63), and no canine teeth. A new world-revision of J/isha, desirable as it appears, cannot be undertaken in the middle of a general report. I state therefore only a general impression that, on the material which I have handled for comparative purposes, there may be a case for recognizing the distinctness of Ilisha africana from the Indo-Pacific forms by the creation of a new sub-genus. Compared with J. brachysoma, I. indica, I. kampent, I, sladeni, etc., it shows the following differences: the ventral scutes are far more salient ; the lower jaw projects further, so that the mouth opening is carried further dorsad ; the lower part of the body (ventral to the lower jaw) is more developed and keel-like, so that the overall body depth is greater and the pectoral fin, though maintaining the same relations to the head, is inserted pro- portionally higher; the pectoral fin, though obliquely inserted, tends to twist so as to project in a horizontal plane perpendicular to the surface of the body. In these same characters it approaches the related genera and species of the Atlantic coasts and rivers of South America. Perhaps the most interesting feature of I/isha africana is that in the development of the deep ventral keel, in the carriage of the pectoral fins, elongation of the anal, exaggeration of the lower lobe of the caudal, and incipient loss of the pelvics, (though not in the normal lateral septum), it foreshadows the form and probably also the habit of the Gastropelecidae—South American Ostariophysi which “ fly ” by “a taxi-ing movement, with the large thoracic keel cleaving the water and the pectoral fins beating the surface ”’ (Fraser-Brunner, 1950, Ann. Mag. nat. Hist. (12) 3: 959). These characteristics of J. africana attain a full convergent evolution of the Gasteropelecid condition in the related South American Pyistigaster cayanus Cuvier (1817, Régne Animal, 4, pl. io, fig. 3). (P. martit Agassiz is a synonym of this species.) Parallel evolution of the same condition has occurred in the Indo- Pacific Opisthopterus, in which the correspondence to the Gasteropelecid form is increased by the smooth ventral profile and the more posterior insertion of the dorsal, which in this genus commences behind the anal origin. (uewyo9 ‘Ss ‘[ oJoYq) “wo T1z YASueT [e}0} ‘usuTOads alTeUey W “IF u0ljRAS ,, PINeSOY ,, }B U93e} ‘(9TUSFY 2 IoT[NY) svonay -g Ajqeqoid “ds snurysvyoany “yeYS “4 ALVId 9 ‘2 "007 ((H'N) “Wd “11d Bull. BM. (NE) Zookhz;6. 7 _ BD PLATE 8. Electric Rays, Diplobatis pictus Palmer. Two female specimens, actual size, from ‘‘ Rosaura’”’ Station 36. The outlines have been retouched. (Photo British Museum (Natural History).) 5. SPONGES By MAURICE BURTON With Plate 9 and Text-figs. 1-9 SYNOPSIS The 42 specimens, representing 35 species, of which 8 are new, were collected from widely separated points on the Atlantic seaboards, providing useful fauna records. Two specimens are of special interest, since they make possible re-descriptions of species long known but too inadequately described to make recognition possible. The species are Siryphnus pachastrel- loides (Schmidt) and Callyspongia tenervima (Duchassaing and Michelotti). I. LIST OF SPECIES MENTIONED IN THIS REPORT Order HEXACTINELLIDA. Sub-order AMPHIDISCOPHORA. Family HYALONEMATIDAE. Pheronema carpenteri (Thomson). Sub-order HEXASTEROPHORA. Family FARREIDAE. Farrea occa Bowerbank. Family MELETTIONIDAE. Aphrocallistes beatrix Gray. Family ROSSELLIDAE. Rossella mortensent Burton. Order CALCAREA. Family HOMOCOELIDAE. Leucosolenia botryoides (Ellis and Solander). Leucosolenia canariensis (Michlucho-Maclay). Order TETRAXONIDA. Sub-order STREPTASTROSCLEROPHORA. Family THENEIDAE. Thenea fenestrata (Schmidt). Thenea wyvilli Sollas. 216 THE “ROSAURA” EXPEDITION Sub-order ASTROSCLEROPHORA. Family STELLETTIDAE. Stelletta anancora (Sollas). Stryphnus pachastrelloides (Schmidt). Family CLAVULIDAE. Radiella sol Schmidt. Family TETILLIDAE. Tetilla cranium (Miller). Sub-order SIGMATOSCLEROPHORA. Family HAPLOSCLERIDAE. Haliclona rubens (Pallas). » evina de Laubenfels. , Sptculosa (Dendy). », calcinea sp. n. » tenerrima sp. 0. Callyspongia tenerrima Duchassaing and Michelotti. Family DESMACIDONIDAE. Desmacella annexa (Schmidt). » tmornata (Bowerbank). Myxilla distorta sp. n. Phorbas amaranthus (Duchassaing and Michelotti). Inflatella viridis (Topsent). Plocanionida topsenti sp. n. Tedania anhelans (Lieberkiihn). Family AXINELLIDAE. Axinella ramosa sp. n. Hymeniacidon assimulis (Levinsen). ie glabrata sp. n. Halichondria panicea (Pallas). A osculum Lundbeck. ie bowerbanki Burton. 3 cornuloides sp. n. Family RASPELIIDAE. Higginsia strigilata (Lamarck). Dragmatyle topsenti sp. n. Order KERATOSA. Family SPONGIIDAE. Hircinia variabilis Schmidt. THE “ROSAURA” EXPEDITION 27, II. SYSTEMATIC NOTES Order HEXACTINELLIDA. Sub-order AMPHIDISCOPHORA. Family HYALONEMATIDAE. Genus Pheronema Leidy. Pheronema carpenteri (Thomson). Holtenia carpenteri Thomson, 1869, p. 120, woodcut; Thomson, 1869, p. 210; Thomson, 1869, p. 702, pls. Ixvii-Ixxi; Pheronema carpenteri, Kent, 1870, p. 243, pl. Ixiii, fig. 1; Gray, 1870, p. 210; Holtenia carpenteri, Bocage, 1871, p. 69 ; Pheronema carpenteri, Marshall, 1876, p. 130; Schulze, 1886, p. 64; Schulze 1887, p. 241, pl. liii; Schulze, 1893, p. 562; Schulze, 1904, p. 50, pl. xv, figs. 1-5 ; Burton, 1928, p. 14. OccuRRENCE. St. 49, between Fuerteventura Island (Canaries) and Africa (28° 25’ N., 13° 34’ W.), 1st February, 1938. One specimen. DISTRIBUTION. Iceland; Faroes; Lewis; Portugal; Brazil; Zanzibar; 823- 2928 m., on mud. Remarks. In spite of the wide geographical range, and the number of references to this species in the literature, it has not been recorded on more than five occasions. Yet from the numbers of individuals referred to in those references the species would appear to be abundant at suitable depths. This is borne out by the account Bocage (l.c.) gives. He refers to it as well-known to the “ pécheurs des squales ”’ in the deep waters off Setubal. In dealing with this species in 1928, I gave a reference to ‘‘ Carpenter and Thomson 1869 ”’ both in the synonymy list and in the list of literature. This was taken from Lendenfeld’s bibliography in his work of 1886 (p. 598). There are two papers listed by Lendenfeld under this supposed joint authorship, both for the year 1869, and both were written, in fact, by Thomson alone. Another correction is necessary, this time from Marshal (1876). This author lists the genus “‘ Holtenia Schmidt” on p. 126, for H. pourtalesti and, on p. 130, gives ‘‘ Holtenia Schmidt p.p.’’ under Pheronema. The only valid authorship of Holtenia is, of course, of Thomson, 1869. Sub-order HEXASTEROPHORA Family FARREIDAE Genus Farrea Bowerbank Farrea occa Bowerbank (for discussion see Ijima, 1927) OccURRENCE. St. 49, between Fuerteventura Island (Canaries) and Africa (28° 25’ N., 13° 34’ W.), 1st February, 1938, 1300 m. One specimen. DisTRIBUTION. World-wide, from 204-1901 m. 218 THE “ROSAURA” EXPEDITION Remarks. I find it difficult to believe that the many subspecies, into which Ijima divided this species, represent anything more than simple fluctuating variations, or, at most, ecological varieties. Family MELETTIONIDAE Genus Aphrocallistes Gray Aphrocallistes beatrix Gray Aphrocallistes beatrix Gray, 1858, p. 114, pl. xi; Ijima, 1927, p. 286, pl. xxiv, figs. 20-30, pl. xxv, figs. I-25, text-fig. 35. OCCURRENCE. St. 34, off St. George, Grenada, West Indies (12° 25’ N., 61° 49’ W.), 27th November, 1937, 720-800 m. One specimen. DISTRIBUTION. South-west of Ireland; Azores; Madeira; Canaries; Cape Verde Islands ; Ascension Island; Bermuda; Florida; West Indies ; Indian Ocean (south of Bombay, Bay of Bengal, Andaman Islands) ; Malay Area; Philippines ; Japan ; 105-1633 m., on gravel, sand, mud and ooze. Family ROSSELLIDAE Genus Rossella Carter Rossella mortenseni Burton Rossella mortenseni Burton, 1928, p. 9, figs. 3-9. OCCURRENCE. St. 49, between Fuerteventura Island (Canaries) and Africa (28° 25’ N., 13° 34’ W.), 1st February, 1938, 1300 m. One specimen. DIsTRIBUTION. Iceland; Faroes; 479-957 m. Order CALCAREA Family HOMOCOELIDAE Genus Leucosolenia Bowerbank Leucosolenia botryoides (Ellis & Solander) Spongia botryoides Ellis and Solander, 1786, p. 190, pl. lviii, figs. 1-4 ; Leucosolenia botryordes, Arndt, 1939, p. 4, fig. 1. OccurRENCE. St. 10, NW. Arm, Horn Bay, Newfoundland (52° 13’ N., 55° 47’ W.), 23rd September, 1937, I5 m. One specimen. DIsTRIBUTION. Arctic; Atlantic coasts of Europe; Mediterranean; West Africa ; South Africa; Red Sea; Australia (west and east coasts) ; Atlantic coast of North America; California; Chile; Sandwich Isles; mainly littoral, but occurring down to 860 m., growing on rocks, stones and seaweeds. THE “ROSAURA” EXPEDITION 219 Leucosolenia canariensis (Michlucho-Maclay) Nardoa canariensis Michlucho-Maclay, 1868, p. 230; N. sulphurea Michlucho-Maclay, 1868, p. 230; N. rubra Michlucho-Maclay, 1868, p. 230; Tarroma canariense, Haeckel, 1870, p. 244; T. rubrum, Haeckel, 1870, p. 245; T. sulphureum, Haeckel, 1870, p. 245; Ascaltis canariensis, Haeckel, 1872, p. 52, pl. ix, figs. 1-3, pl. x, fig. 1 ; Awloplegma canariensis, Haeckel, 1872, p. 52; Ascuris arvredifae Haeckel, 1872, p. 52; A. papillata Haeckel, 1872, p. 52; Ascetta compacta Schufiner, 1877, p. 404, pl. xxv, fig. 9; A. coriacea var., Fristedt, 1885, p. 8; Ascuris canariensis, Lackschewitsch, 1886, p. 338; Leucosolenia nausicae Lackschew- itsch, 1886, p. 300, pl. vii, fig. 1; L. nanseni Breitfuss, 1896, p. 427; Breitfuss, 1898, p. 106, pl. xii, figs. 1-9; Breitfuss, 1898, p. 13; L. coriacea, Arnesen, 1901, p. 10; Arnesen, 1901, p. 67; L. canariensis (pars), Thacker, 1908, p. 762, pl. xl. fig. 3, text-figs. 157-160 ; Clathrina canariensis var. compacta Row, 1909, p. 184; Leucosolinia nanseni Breitfuss, 1911, p. 311 ; L. canaviensis, Dendy and Row, 1913, p. 724; L. nanseni, Derjugin, 1915, p. 289; L. canari- ensis, Hentschel, 1916, p. 4; Hozawa, 1918, p. 528, pl. Ixxxiv, fig. 2; Breitfuss, 1932, p. 240; Hozawa, 1933, p. 2, pl. i, fig. 1; Topsent, 1934, p. 7. OccURRENCE. St. 10, NW. Arm, Horn Bay, Newfoundland (52° 13’ N., 55° 47’ W.), 23rd September, 1937, 15 m. Two specimens. DISTRIBUTION. Spitzbergen; Kola Fiord; Norway; Sweden; Mediterranean ; Canaries ; Cape Verde Islands; Mauritius; Suez; Japan; on rock, sand, stones, mud, shells ; littoral to 165 m. REMARKS. The similarity between the external form of this species and that of the better-known L. coriacea tempts the suggestion that it is a recurrent mutant of the latter. Both occupy a similar range, have the same colour varieties and a closely similar external form. If, as we have reason to believe, the presence or absence of oxea in calcareous sponges has no specific value, then the affinity between L. canariensis and L. coriacea is even closer, and the recognition of the former as a mutant of the latter depends mainly on the presence of quadriradiates. Order TETRAXONIDA Sub-order STREPTASTROSCLEROPHORA Family THENEIDAE Genus Thenea Gray Thenea fesestrata (Schmidt) Tisiphonia fenestrata Schmidt, 1880, p. 71, pl. x, fig. 2; Thenea fenestrata, Sollas, 1886, p. 185 ; T. wrightii Sollas, 1886, p. 185 ; Sollas, 1888, p. 63, pl. viii, figs. 11-20; T. fenestrata, Sollas, 1888, p. 71, pl. viii, figs. 1-8 ; Ancorina (Thenea) fenestrata, Lendenfeld, 1903, p. 55; A. (T.) wrightii, Lendenfeld, 1903, p. 58; T. fenestrata, Wilson, 1904, p. 88, pl. xiii, figs. 2-4, 6-7, 9; T. echinata Wilson, 1904, p. 91, pl. xii, figs. 1-9; T. lamelliformis Wilson, 1904, p. 95, pl. xii, figs. 10-13, pl. xiii, fig. 1. OccurRENCE. St. 26, north of Turneffe Island, West Indies (17° 53’ N., 84° 44’ W.), 7th November, 1937, 900 m. One specimen. DistRiBuTION. Galapagos; Chile; West Indies; Sierra Leone; 1745-4114 m., on mud and ooze. 220 THE “ROSAURA” EXPEDITION Thenea wyvillii Sollas Thenea wyvillii Sollas, 1886, p. 184; Sollas, 1888, p. 74, pl. vi, figs. 1-9; T. calyx Thiele, 1898, p. 24, pl. v, figs. 9-10; Ancovina (Thenea) wyvillit, Lendenfeld, 1903, p. 56; A. (T.) grayi calyx, Lendenfeld, 1903, p. 57. OccuRRENCE. St. 49, between Fuerteventura Island and Africa (28° 25’ N., 13° 34’ W.), 1st February, 1938, 1300 m. One specimen. DIsTRIBUTION. Off Zebu, Philippines, 174 m., on blue mud. REMARKS. The species evidently ranges throughout tropical waters. In addition to the present specimen from the tropical Atlantic and the holotype from the Philippines, there is a third specimen in the John Murray Collection, in the British Museum, from the Gulf of Aden. Sub-order ASTROSCLEROPHORA Family STELLETTIDAE Genus Stelletta Schmidt Stelletta anancora (Sollas) Pilochvota anancora Sollas, 1886, p. 189; P. gigas Sollas, 1886, p. 190; P. tenuispicula Sollas, 1886, p. 190; P. crassispicula Sollas, 1886, p. 190; P. gigas Sollas, 1888, p. 124, pl. xx, figs. 1-13; P. tenuispicula Sollas, 1888, p. 127, pl. xv, figs. 28-32; P. crassispicula Sollas, 1888, p. 128, pl. xiv, figs. 9-15 ; P. anancora Sollas, 1888, p. 132, pl. xiv, figs. 16-22 ; Stelletta crassi- spicula, Lendenfeld, 1903, p. 42; S. tenuispicula, Lendenfeld, 1903, p. 42 ; S. anancora, Lenden- feld, 1903, p. 43; S. gigas, Lendenfeld, 1903, p. 43; S. crassiclada Lendenfeld, 1906, p. 281, pl. xxxi, figs. 3-12; S. crassispicula, Topsent, 1922, p. 1; Topsent, 1928, p. 123, pl. 1, fig. 25, pl. v, fig. 13. OCCURRENCE. St. 44, St. Paul’s Rock, 28th December, 1937, 50-60 m. One specimen. DISTRIBUTION. Bermuda; Cape Verde Islands; St. Paul’s Rock; Bahia; on rock and calcareous algae ; 12-91 m. REMARKS. Topsent (1922) has given an account of the variations in this species, and on the basis of this has suggested the identity of S. crassiclada and S. crasst- spicula. It is difficult to understand why he failed at the same time to recognize Sollas’ Pilochrota anancora, P. gigas and P. tenwispicula as synonyms also ; the species name is, by priority, Stelletta anancora. Genus Strvyphnus Sollas Stryphnus pachastrelloides (Schmidt) (Text-fig. 1) Ancorina pachastrelloides Schmidt, 1870, p. 68; Sollas, 1888, p. 205; Lendenfeld, 1903, p. 66. OccuRRENCE. St. 49, between Fuerteventura Island and Africa (28° 25’ N., 13° 34’ W.), 1300 m., Ist February, 1938. One specimen. THE “ROSAURA” EXPEDITION 221 REMARKS. Schmidt’s original description of this species, alone, was inadequate for purposes of identification. With the few details he gives, together with a poor spicule preparation made by him and now in the British Museum (Reg. No. 70.5.3-48), it is possible to recognize the species when these two things can be compared with fresh material. The present specimen, although collected in the eastern Atlantic, whereas the holotype was taken off Florida, is massive, with a ). We e TEXT-FIG 1. Stryphnus pachastrelloides (Schmidt). @, oxeote, x 5°; b, dichotriaene, X 50; ¢, microxeote, x 500; d, microstrongyle, x 500; e, amphiaster, x 500. 222 THE “ROSAURA” EXPEDITION cylindrical outgrowth. There is one oscule at the summit of the main mass and one at the end of the outgrowth, each leading into a deep cloaca. The colour, in spirit, is a pale yellowish-brown. Schmidt described the sponge as “ unregelmassige Knollen mit unregelmassig cylindrischen Fortsitzen.” The only other guide to the external appearance is given: ‘‘ Die Rindenschicht wird lediglich durch ein braéunliches Pigment vertreten.’’ The skeleton of the “ Rosaura’”’ specimen consists of large radially-arranged oxea, short-shafted dichotriaenes, with microxea, roughened microstrongyla and amphiasters. There appear to be present, also, oxyasters, but these, I believe, are no more than the amphiasters seen in end-view. The spicule-preparation (70.5.3.48) contains these same categories, and they are of similar dimensions in each case. There are, however, a number of smooth strongyla, about the same length as the microxea but much thicker. At first sight, the strongyla seem proper to the sponge but the discovery of a number of other more obviously extraneous spicules in this preparation suggests that the strongyla may also be foreign to it. Schmidt’s preparations, of which the British Museum has many, are notorious for either the poverty of their proper spicules or the richness of those spicules accidentally included. The identity of Schmidt’s holotype, based on B.M. 70.5.3.48, and the present specimen is made the more certain by the presence, in both, of a faint annulus at the centre of so many of the microxea. The species may be re-described : Irregularly massive with cylindrical outgrowths ; surface uneven, hispid ; oscules large, apical ; texture firm, incompressible ; colour, in spirit, light brown ; skeleton of radially-arranged oxea and short-shafted dichotriaenes, with microxea, micro- strongyla and amphiasters for microscleres : Spicules: Oxea, slightly curved, 3:2 by 0-06 mm., dichotriaenes, cladome 0-6 mm. across, rhabdome 1:5 by 0-048 mm., microxea, 0-14 by 0-004 mm., microstrongyla 0-024 by 0-004 mm., amphiasters 0-028 mm. across. Family CLAVULIDAE Genus Radiella Schmidt Radiella sol Schmidt Trichostemma hemisphaericum Sars, 1869, p. 259 [nom. nud.]; Radiella sol Schmidt, 1870, p. 48, pl. iv, fig. 6; Trichostemma hemisphaericum Sars, 1872, p. 62, pl. vi, figs. I-15; Radiella sol, Burton, 1930, p. 510; Burton, 1934, p. 15. OccuRRENCE. St. 26, north of Turneffe Island, West Indies (17° 53’ N., 87° 44’ W.), 900 m. One specimen. DISTRIBUTION. Greenland ; between Norway and Bear Island; Barents Sea ; Kara Sea; Franz Josef Land; Norway (Lofoten) ; West Indies; 97-1167 m., on clay and mud. THE “ROSAURA” EXPEDITION 223 Family TETILLIDAE Genus Tetilla Schmidt Tetilla cranium (Miller) Alcyonium cranium Miiller, 1776, p. 255; Tethya cranium, Lamarck, 1815, p. 71; Cvraniella cranium, Sollas, 1888, p. 51; Tethya cranium, Topsent, 1913, p. 12, pl. iii, fig. 3, pl. v, fig. 12. OccURRENCE. St. 34, off St. George, Grenada, West Indies (12° 25’ N., 61° 49’), 27th November, 1937, 720-800 m. One specimen. DIsTRIBUTION. Arctic; Atlantic coast of Europe; Azores; West Indies ; Ceylon ; Japan; Pacific coast of North America; 9-1829 m., on rock, gravel, sand, mud and ooze. Family HAPLOSCLERIDAE Genus Haliclona Grant Haticlona rubens (Pallas) Spongia vubens Pallas, 1766, p. 389; Duchassaing and Michelotti, 1864, p. 41, pl. x, fig. 1; Pachychalina rubens, Schmidt, 1870, p. 37; Chalina rubens, Carter, 1882, p. 276; Chalinopsis vubens, Lendenfeld, 1887, p. 744; Ceraochalina rubiginosa Lendenfeld 1887, p. 779; Pachy- chalina vubens, Wilson, 1902, p. 392; Cladochalina rubens, Burton, 1927, p. 511; Haliclona vyubens, de Laubenfels, 1932, p. 59; 1936, p. 42, pl. vu, fig. 2, pl. viii, fig. 1; de Laubenfels, 1949, P. 9. OccURRENCE. St. 31, Gorda Cay, Mosquito Bank, West Indies (15° 54’ N., 82° 13’ W.), 16th November, 1937, 34 m. One specimen. DISTRIBUTION. Dry Tortugas; Bahamas; Long Key Island; Guadeloupe ; Viecques ; St. Domingo; St. Thomas; Cuba; Porto Rico; I-17 m. Haliclona erina de Laubenfels Haliclona erina de Laubenfels, 1936, p. 457. OccuRRENCE. St. 10, NW. Arm, Horn Bay, Newfoundland (52° 13’ N., 55° 47’ W.), 23rd September, 1937, I5 m. Two specimens. DISTRIBUTION. Panama (Caribbean). REMARKS. These specimens, although agreeing with the holotype in structure and appearance, were obtained well away from the type locality. It is therefore worth recalling de Laubenfels’ comment, in describing the species: ‘‘ It must be admitted that this may be a remarkable modification of some previously described Haliclona, due perhaps to unusual environment conditions, but it is impossible to say which Haliclona has been so modified.’’ This expresses the situation for so many specimens of Haliclona from all parts of the world. 224 THE “ROSAURA” EXPEDITION Haliclona spiculosa (Dendy) Siphonochalina spiculosa Dendy, 1887, p. 505; Dendy, 1890, p. 354, pl. lviii, fig. 2, pl. Lx, fig. 3; Wilson, 1902, p. 394. OccuRRENCE. St. 24, Turneffe Island, West Indies (17° 15’ N., 87° 49’ W.), ist November, 1937, 2m. One specimen. DIsTRIBUTION. Turk’s Island, Bahamas; St. Thomas; 2-42 m. REMARKS. The species is doubtfully referred to the genus Haliclona, principally on the absence of a special tangential skeleton at the surface. Haliclona calcinea sp. n. (Pl. 9, fig. 1; Text-fig. 2) HoLotyre. B.M. 1938.6.30.31. OccuURRENCE. St. 34, off St. George, Grenada, West Indies (12° 25’ N., 61° 49’ W.), 27th October, 1937, 720-800 m. One specimen. DESCRIPTION. Sponge massive, rounded ; surface even, minutely hispid ; oscules small, scattered; texture firm, friable; colour, in spirit, dark brown; skeleton subhalichondroid, mainly unispicular; megascleres oxea, 0-5 by o-o16 mm., TEXT-FIG. 2. Haliclona calcinea sp. n. a, oxeote, x 200; 6b, toxa and sigma, X 500. Tse THE “ROSAURA” EXPEDITION 225 microscleres centrangulated sigmata, 0-024 mm. chord, and toxa, 0:04 to 0:08 mm. chord. Remarks. Although the colour is given as dark brown, this applies strictly to the ectosome, whether of the outer surface or of the linings of the large canals. The choanosome is a dull brownish-yellow. The surface, moreover, is largely coated with a pale greyish yellow, in places up to 3 mm. thick. This apparent incrustation has the same spicules as the main body and must be presumed to represent new growth. The species differs, in its external form, as well as in the detailed measurements of the microscleres, from other species of Haliclona having such long oxea combined with the presence of microscleres. Haliclona tenerrima sp. n. (Pl. 9, fig. 2; Text-fig. 3) HototyrPe. B.M. 1938.6.30.42. OccuRRENCE. St. 24, Turneffe Island, British Honduras (17° 16’ N., 87° 50’ W.), ist November, 1937, 3} m. One specimen. DESCRIPTION. Sponge a fragment only; surface uneven, minutely hispid ; oscules not present; texture soft, compressible, elastic; colour, in spirit, pale Text-Fic. 3. Haliclona tenerrima sp. n. a, oxeote, xX 200; 6, toxon, xX 500. yellow; skeleton irregularly sub-isodictyal (almost halichondroid) with triangular mesh, and with occasional fibres of 2 to 3 spicules width running to surface ; megascleres oxea, 0°28 by 0007 mm. ; microscleres toxa, 0°03 to o°I mm. chord. Callyspongia tenerrima Duchassaing & Michelotti Calyspongia tenerrima Duchassaing & Michelotti, 1864, p. 57, pl. x, figs. 3-4. OccuRRENCE. St. 31, Gorda Cay, Mosquito Bank, West Indies (15° 54’ N., 82° 13’ W.), 16th November, 1937, 34 m. One specimen. DIsTRIBUTION. West Indies (St. Thomas, Viecques). 226 THE “ROSAURA” EXPEDITION D1aGnosis. Sponge branching, stipitate, slender, dichotomous; surface even ; oscules small, in linear series along branches ; texture soft, compressible ; colour, in life and in spirit, yellow tinted with rose ; main skeleton an irregular network of mainly quadrangular mesh, with recognizable primary, secondary and tertiary fibres ; special ectosomal skeleton, of similar structure but with meshes mainly triangular ; fibres 0-03 to 0-08 mm. diameter, cored by 1-4 spicules; megascleres strongyla, 0:1 by 0-002 to 0-003 mm. ; microscleres absent. Remarks. The holotype has been lost, as was ascertained some years ago by de Laubenfels during his visit to Turin. The original description, inadequate though it is, leaves, no doubt of the identity of the present specimen. The holotype was described as ‘‘ phytoide, dichotome, gréle ... les rameaux ... cylindriques jaune un peu rosé ’’ in life with small oscules in series along the branches. Every one of these features is found in the present specimen. Above all, the peculiar appearance of the surface, shown in fig. 4 (Duchassaing & Michelotti, l.c.) puts the identification beyond reasonable doubt. Family DESMACIDONIDAE Section MyCcaLEAE Genus Desmacella Schmidt Desmacella annexa (Schmidt) Desmacella annexa Schmidt, 1870, p. 53; Tylodesma annexa, Burton, 1930, p. 525. OccCURRENCE. St. 34, off St. George, Grenada, West Indies (12° 25’ N., 61° 49’ W.), 27th November, 1937, 720-800. Two specimens. DISTRIBUTION. Iceland; Norway; south-west Ireland; Atlantic coast of France; Mediterranean; West Indies; Florida; Indian Ocean; 85-1331 m., on rock, mud and ooze. Desmacella inornata (Bowerbank) Halichondvia inornata Bowerbank, 1866, p. 271; Tylodesma inoynata, Burton, 1930, p. 527. OccuRRENCE. St. 34, off St. George, Grenada, West Indies (12° 25’ N., 61° 44’ W.), 27th November, 1937, 420-800 m. One specimen. DISTRIBUTION. British Isles; France; Spain; Mediterranean; Azores ; 200- 1025 m., on mud, sand, gravel, broken shells, rocks. Section MyXILLEAE Genus Myxilla Schmidt Myxilla distorta sp. n. (Pl. 9, fig. 3, Text-fig. 4) HOLOTYPE. 1938.6.30.30. OccuRRENCE. St. 34, off St. George, Grenada, West Indies (12° 25’ N., 61° 49’ W.), 27th November, 1937, 720-800 m. One specimen. THE “ROSAURA” EXPEDITION 227 DESCRIPTION. Sponge erect, lobate; surface uneven; harsh to touch ; oscules at apices of lobes; texture firm, incompressible, friable; colour, in spirit, pale yellow; main skeleton an irregular reticulation of bluntly-pointed curved styli, 1-0 by 0-024 mm. ; ectosomal skeleton a tangential reticulation of irregular mesh formed of tornota with both ends tylote and measuring 0-72 by 0-016 mm. ; micro- a 0 TEXT-FIG. 4. Myzxilla distorta sp. n. 4, style, x 300; 6, tornote, x 300; c, isochela with teeth meeting in middle line (front view), x 600; d, normal large isochela, x 600 ; é, small isochela, x 600. ZOOL. 2, 6, 14 228 THE “ROSAURA” EXPEDITION scleres isochelae spatuliferae, larger frequently distorted by unusual length of teeth, ranging from 0-024 to 0-08 mm. chord. REMARKS. The species is peculiar in the presence of the distorted chelae, but is also separable from other known species in the remaining details of its spiculation. Genus Phorbas Duchassaing & Michelotti Phorbas amaranthus Duchassaing & Michelotti Phorbas amavanthus Duchassaing & Michelotti, 1864, p. 92, pl. xxi, fig. 1; Cribrella hospitalis Schmidt, 1870, p. 56, pl. iv, fig. 12; ? Anchinoé fictitius, Topsent, 1928, p. 284; Phorbas amaranthus, de Laubenfels, 1936, p. 63. OccuRRENCE. St. 31, off Gorda Bay, Mosquito Bank, West Indies (15° 54’ N., 82° 13’ W.), 16th November, 1937, 34 m. One specimen. DISTRIBUTION. Florida; West Indies; ? Azores; ?? Morocco and Spain ; 27 m. REMARKS. Since this species has never been adequately described, a description based on the holotype is given : Sponge encrusting or massive and irregularly digitate; surface villous, parti- cularly in larger individuals ; oscules not apparent ; pores in sieves ; texture firm ; colour, alive, maroon, in spirit, almost white ; skeleton of ascending fibres of serially- arranged tornota, echinated by acanthostyli of two sizes, and with a tangential ectosomal layer of tornota similar to those of main skeleton ; tornota, with oxeate to hastate ends, 0-2 to 0-36 by 0:004 to 0-005 mm.; large acanthostyli, 0-2 by 0-006 mm., small acanthostyli, 0-1 by 0-006 mm., microscleres isochelae arcuatae, 0-026 mm. chord. In a very brief redescription of the species, de Laubenfels (1936) describes the microscleres as chelae and rare sigmata. I have found no sigmata in my preparations from the type. Topsent (1928) records specimens of Achinoé fictitius, from the Azores, Morocco and Spain, having a spiculation similar to that of Phorbas amaranthus. In addition, he includes Cribrella hospitalis Schmidt as a synonym of his supposed Anchinoé fictitius, but Schmidt’s species is identical with Phorbas amaranthus. Whether, indeed, this West Indian species extends to the Azores is problematic, but it is highly doubtful that it should also occur off Morocco and Santander. Re-examination of Topsent’s material is needed to settle this. Genus Inflatella Schmidt Inflatella viridis (Topsent) Joyeuxia viridis Topsent, 1890, p. 29; Topsent, 1892, p. 94, pl. il, fig. 8, pl. x, fig. 19; Inflatella viridis, Lundbeck, 1910, p. 20, pl. ii, figs. 11-12, pl. iv, fig. 7; Hentschel, 1929, p. 968; nec Topsent, 1904, p. 205. OccCURRENCE. St. 34, St. George, Grenada, West Indies (12° 05’ N., 61° 49’ W.), 27th November, 1937, 720-800 m. One specimen. DISTRIBUTION. Denmark Strait; Azores; 136-1768 m. THE “ROSAURA” EXPEDITION 229 Genus Plocamionida Topsent Plocamionida topsenti sp. n. (Text-fig. 5) HOoLoTyPE. 1938.6.3.29. OccURRENCE. St. 34, off St. George, Grenada, West Indies (12° 25’ N., 61° 49’ W.), 27th November, 1937, 720-800 m. One specimen. DESCRIPTION. Sponge thinly encrusting ; surface hispid ; oscules not apparent ; colour, in spirit, pale greyish-yellow ; skeleton of styli, 1-2 by 0-04 mm., and acan- thostyli, 0-24 by 0-008 mm., with occasional intermediates ; ectosomal tornota, with ends strongylote or faintly subtylote, 0-36 by 0-006 mm., acanthostrongyla, 0-18 by 0-014 mm.; microscleres isochelae arcuatae, 0-06 to 0-068 mm. chord. REMARKS. The species differs from all others in the genus in the much larger sochelae, as well as in other details of the spiculation. Section TEDANIEAE Genus Tedania Schmidt Tedania anhelans (Lieberkihn) Halichondria anhelans Lieberkihn, 1859, p. 521, pl. xi, fig. 6; Tedania nigrescens, Burton, 1932, p- 346, fig. 44; Xytopsibis asperus de Laubenfels, 1936, p. 61. OccURRENCE. St. 24, Turneffe Island (17° 15’ N., 87° 49’ W.), 1st November, 1937, 2m. One specimen. DISTRIBUTION. Mediterranean ; Cape Verde Islands; Bermuda; West Indies ; West Africa; Red Sea; Indian Ocean; Indonesia; Cochin-China; Australia ; littoral to 120 m., on rock, stones, shells, mud or sand. RemARKS. The species has been fully discussed (Burton, 1932) under Tedaria nigrescens (Schmidt). This name is, however, pre-empted by Halichondria anhelens Liebekitihn which, as shown by the original description and by Schmidt’s (1862, p- 74) finding of tornota in preparations of the type, is identical with the widespread species known so long under Schmidt’s name. Family AXINELLIDAE Genus Axinella Schmidt Axinella ramosa sp. n. (Pl. 9, fig. 4; Text-fig. 6) HOoLotyPe. 1938.6.30.37. OccurRENCE. St. 31, off Gorda Bay, Mosquito Bank, West Indies (15° 54’ N., 82° 13’ W.), 16th November 1937. Three specimens. 230 THE “ROSAURA” EXPEDITION TEXT-FIG. 5. Plocamionida topsenti sp. n. a, style, x 300: b, acanthostyle, x 300; c, acanthostrongyle, x 300; d, tornote, x 300; e, isochela, x 600. THE “ROSAURA” EXPEDITION 231 DESCRIPTION. Sponge branching; surface slightly uneven, minutely hispid ; oscules not apparent ; texture firm; colour, in spirit, pale yellow; skeleton sub- isodictyal, ascending fibres somewhat plumose, connectives mainly unispicular ; megascleres styli, 0-4 by 0-016 mm.; microscleres absent. TExt-FIG. 6. Axinella ramosa,sp.n. a, Style, xX I00. REMARKS. There are three specimens, the largest, the holotype, is 65 mm. high with its branches 4 mm. in diameter at most. The other two are considerably smaller. Genus Hymeniacidon Bowerbank Hymeniacidon assimilis (Levinsen) Halichondria assimilis Levinsen, 1886, p. 352, pl. xxx, fig. 5 ; Hymeniacidon assimilis, Burton, 1935, P- 74- OccuRRENCE. St. 5, Julianshaab, Greenland (60° 43’ N., 46° 02’ W.), 4th September, 1937, 80 m. One specimen. DISTRIBUTION. Kara Sea; Sea of Japan; Sea of Okhotsk ; littoral to 124 m., on seaweeds. Hymeniacidon glabrata sp. n. (Pl. 9, fig. 5; Text-fig. 7) HOLOTYPE. 10938.6.30.35. OccURRENCE. St. 22, Belize Harbour, West Indies (17° 28’ N., 88° 11’ W.), 29th October, 1937,6_m. One specimen. DESCRIPTION. Sponge encrusting, agglutinating ; surface even, minutely hispid ; oscules at ends of cloacae running parallel to and just under surface; texture soft, compressible ; colour, in spirit, pale greyish-yellow; main skeleton hali- chondroid, with tendency to multispicular ascending fibres; ectosomal skeleton a loose network of more or less quadrangular mesh; megascleres styli, thickest at centre (? incompletely differentiated into two sizes), ranging from 0-24 to 0-8 by 0-004 to 0:024 mm. Remarks. The holotype is a thin and flattened, somewhat irregular plate, with a number of small bivalve shells incorporated in its lower surface. Being no more 232 THE “ROSAURA” EXPEDITION than a fragment, the interpretation of its mode of growth is difficult. In general appearance it recalls H. tubulosa (Ridley & Dendy) from the mouth of the River Plate, which has, presumably, a more erect growth ; and also larger spicules. Text-FIG. 7. Avinella glabrata sp.n. Large and small styli, x Ioo. Genus Halichondria Fleming Halichondria panicea (Pallas) Spongia panicea Pallas, 1766, p. 388 ; Halichondria panicea, Arndt, 1928, p. 52. OccuRRENCE. St. 1, Angmagssalik Harbour, Greenland (65° 35’ N., 37° 20’ W.), 27th August, 1937, 25-50 m. One specimen. DisTRIBUTION. Arctic; Atlantic coasts of Europe (including Baltic Sea) ; Mediterranean ; Black Sea; Atlantic coast of Africa to Saldanha Bay; St. Paul; Patagonia; Chile; Atlantic coasts of North America, south to Newfoundland ; Pacific coast of Asia, south to Japan; Indian Ocean (Ceylon) ; Kerguelen; New Zealand ; littoral to 183 m. Halichondria tenera (Marenzeller) Isodictya tenera Marenzeller, 1877, p. 364, pl. i, fig. 2; Burton, 1930, p. 516; Halichondria tenera, Burton, 1935, p. 76. OccuRRENCE. St. 1, Angmagssalik Harbour, Greenland (65° 35’ N., 37° 20’ W.), 27th August, 1937, 25-50 m. One specimen. THE “ROSAURA” EXPEDITION 233 DIsTRIBUTION. Franz Josef Land; Sea of Japan; Sea of Okhotsk; 2-178 m., mud. REMARKS. This sponge, which is massive, or may be massively branched, with uneven surface, small scattered oscules and a soft friable texture, has for main skeleton a loose isodictyal reticulation. The tangential ectosomal skeleton is ill- developed ; the sole spicules are oxea 0-3 to 0-4 by 0-003 to 0-01 mm. The species is not, therefore, a typical Halichondria. It may even have affinities with the species of Adocia, except that the oxea are more like those typical for Halichondria. Halichondria osculum Lundbeck Halichondria osculum Lundbeck, 1902, p. 23, pl. iii, figs. 3-7, pl. ix, figs. 7-9; Hentschel, 1929, Pp. 991. OccurRENCE. St. 10, NW. Arm, Horn Bay, Newfoundland (52° 13’ N., 55° 47’ W.), 23rd September, 1937, I5 m. One specimen. DISTRIBUTION. Greenland, 18-718 m. REMARKS. The specimen forms a small irregular mass, Io mm. across, growing on a Nullipore. There is an encrusting body with processes arising from it and it is these which determine its identification with Lundbeck’s species. Although much smaller than Lundbeck’s original specimens, each process bears a strong resemblance to them. On the other hand, the spicules are somewhat smaller, none exceeding 0-6 mm., or slightly more, by 0-012 mm., against a maximum of 0-7 by 0-17 mm.., in the types. On first examination I had taken this to be the type of a new species, but on further consideration accept it as an atypical specimen of Halichondria osculum. Halichondria bowerbanki Burton Spongia coalita Lamouroux, 1816, p. 80; Halichondria coalita, Topsent, 1911, pp. i-xv; H. bowerbanki Burton, 1930, p. 489. OccurRENCE. St. 10, NW. Arm, Horn Bay, Newfoundland (17° 53’ N., 87° 44’ W.), 23rd November, 1937, I5 m. One specimen. Remarks. Halichondria bowerbanki and H. panicea resemble each other closely in spiculation and in external appearance. In addition, both are extremely variable in form. The consequence is that, although Topsent (I9g1I) has shown marked differences in their larvae, it is rarely possible to be certain of the identification of a given specimen. The present sample is of a loose-textured, massive sponge, about I00 mm, across and some 50 mm. high. Its surface is covered with low digitate processes, 3 to 4mm.high. The identification given here is based on the translucent appearance and the fact that specimens having the cavernous quality of the inner tissues as well as the digitate processes of the surface have been collected by me off the coast of Devon. Even so, it is given with hesitation. 234 THE “ROSAURA” EXPEDITION Halichondria cornuloides sp. n. (Pl. 9, fig. 6; Text-fig. 8) Hototyre. B.M. 1938.6.30.41. OccuRRENCE. St. 1, Angmagssalik Harbour, Greenland (65° 35’, 37° 20’ W.), 27th August, 1937, 25-50 m. One specimen. DESCRIPTION. Sponge irregularly massive, lobular; surface smooth, even, translucent ; oscules (and pores?) in circular cribriform areas at summits of Text-Fic. 8. Halichondria cornuloides sp.n. Oxea, X 200. low secondary lobes or apical on main lobes ; texture firm, compressible owing to cavernous structure of inner tissues ; colour, in spirit, yellowish ; megascleres oxea, 0-2 to 0:5 by 0-008 to 0-012 mm.; microscleres absent. Remarks. The species resembles in external appearance the various forms in- cluded by Dendy in his family Coelosphaeridae, and especially those of the genus Cornulum. So strong is this resemblance that I found it difficult to be convinced that it is a true Halichondvia. Nevertheless, the oxea have the typical form and variable length, and the structure of both main and tangential ectosomal skeletons leave little doubt that it is a Halichondria. THE “ROSAURA” EXPEDITION 235 Family RASPAILIIDAE Genus Higginsia Higgin Higginsia sirigilata (Lamarck) Spongia strigilata Lamarck, 1813, p. 450; Lamarck, 1816, p. 377; Higginsia coralloides Higgin, 1877, p. 291, pl. xiv, figs. I-5. OccuRRENCE. St. 31, off Gorda Bay, Mosquito Bank, West Indies (15° 54’ N., 82° 13’ W.), 16th November, 1937, 34 m. One specimen. DISTRIBUTION. Ireland; west coast of Africa ; West Indies ; Natal; Amboina; Australia ; littoral (?) to 20 m., on rock or coral rock. REMARKS. Topsent (1933, p. 112) has shown that Higgin’s Higginsia coralloides is the same as Lamarck’s Spongia strigilata. Genus Dragmatyle Topsent Dragmatyle topsenti sp. n. (Pl. 9, fig. 7; Text-fig. 9) OccURRENCE. St. 26, north of Turneffe Islands, West Indies (17° 53’ N., 84° 44’ W.), 7th November, 1937, 900 m. Several fragments (belonging to one specimen ?) TEXT-FIG. 9. Dvagmatyle topsenti sp. n. Section at right angles to surface, showing axial core of tylostyli; surface brushes of styloids, with occasional tylostyli set at right angles to surface ; and trichodragmata. DESCRIPTION. Sponge comprising slender branches; surface uneven, hirsute ; oscules not apparent ; texture soft, compressible; colour, in spirit, light brown ; skeleton an axial core of tylostyli, with occasional tylostyli set at right angles and projecting beyond surface, with ectosomal brushes of oxeote styloids and tricho- dragmata for microscleres ; tylostyli 1-6 by 0-014 mm., styloids 1-2 by 0-008 mm. trichodragmata, 0-06 mm. long. 236 THE “ROSAURA” EXPEDITION Order KERATOSA Genus Hircinia Nardo Hircinia variabilis Schmidt Hircinia variabilis Schmidt, 1862, p. 34, pl. ili, fig. 17; Lendenfeld, 1889, p. 557, pl. xxxvi, figs. II-14. OccCURRENCE. St. 31, off Gorda Cay, Mosquito Bank, West Indies (15° 54 N., 37° 20’ W.), 16th November, 1937, 34 m. One specimen. DISTRIBUTION. Mediterranean; West Indies and Florida; Indian Ocean ; Australia ; “‘ Pacific Oceanic Islands’ ; 6-75 m. Remarks. If it be possible to recognize varieties in this extremely variable species, then the present specimen should be assigned to var. dendrovdes. LIST OF LITERATURE Arnpt, W. 1928. Porifera, Schwamme, Spongien [in] Tierwelt Dtsch. 4: 89, 110 figs. 1935. Porifera [in] Grimpe, Tierwelt N.u. Ostsee, Teil 3a, Lfg. 27: 1-140, 239 figs. ARNESEN, E. rgor. Spongien fra den norske Kyst. i, Calcarea. Bergens Mus. Aarbo. 5: 1- 46, pl. i. —— 1go1. Calcarea Meeresfauna, pp. 65-72. BocaGeE, J. V. B. pu. 1871. Sur l’existence de la Holtenia carpentert Wyville-Thompson, dans les cétes du Portugal. J. Sci. Math. Phys. nat. Lisboa, 3 : 69-70. BowERBANK, J.S. 1866. A Monograph of the British Spongiadae, 2. London (Ray Society). BreitFuss, L. 1896. Kalkschwamme der Bremer Expedition nach Ost-Spitzbergen. Zool. Anz. 19 : 426-432. 1898. Note sur la faune des Calcaires de l’Océan Arctique. Annu. Mus. Zool. Acad. St. Pétersb. 8 : 12-38. 1898. Die Kalkschwamme von Spitzbergen. Zool. Jb. Syst. 11 : 103-120. 1911. Zur Kenntnis der Spongio-Fauna des Kola-Fjords. Tvav. Soc. Nat. St. Pétersb. 42 : 223-226. 1932. Die Kalkschwammfauna des arktischen Gebietes. Fauna Arct. Jena, 6: 235- 252. Burton, M. 1927. A revision of the genera and species contained in Lendenfeld’s “ Die Chalineen des australischen Gebietes.’’ Ann. Mag. nat. Hist. (9) 19: 289-296, 502-512. — 1928. Hexactinellida. Dan. Ingolf-Exped. 6 (4) : 1-18, 9 pls. 1930. Norwegian sponges from the Norman Collection. Proc. zool. Soc. Lond. : 487- 546, 2 pls., 8 figs. 1932. Sponges. ‘‘ Discovery’ Rep. 6 : 327-392, 20 pls., 56 figs. — 1934. Sponges. Further zool. Res. Swed. Antarct. Exp. 3 (2) : 1-58. 8 pls., 16 figs. —— 1935. Some sponges from the Okhotsk Sea and the Sea of Japan. Explor. Mers russes, 22 : 61-79, 6 figs. CartER, H. J. 1882. Some sponges from the West Indies and Acapulco in the Liverpool Free Museum, Ann. Mag. nat. Hist. (5) 9 : 266-301, 346-368, 2 pls. Dr LauBENFELS, M. W. 1932. New sponges from the Puerto Rico deep. Smithson. misc. Coll, 91 (17) : 1-28. 1936. A discussion of the sponge fauna of the Dry Tortugas. Pap. Tortugas Lab. 30: 1-225, 22 pls. THE “ROSAURA” EXPEDITION 237 — 1949. Sponges of Woods Hole and adjacent waters. Bull. Mus. comp. Zool. Harv. 103 : 1-55, 3 pls. Denpy, A. 1887. West Indian Chalininae. Proc. zool. Soc. Lond. : 503-507. 1890. Observations on the West Indian sponges. Tvrans. zool. Soc. Lond. 12: 349- 368. — & Row, R. W.H. to11. The classification and phylogeny of the calcareous sponges. Proc. zool. Soc. Lond. : 704-813. Derjucin, K.M. 1915. La faune du golfe de Kola. Mém. Acad. Sci. St. Pétersb. (8) 35 (1) : 1-929. Ducnassainc, P., & MicHELotT1, G. 1864. Spongiaires de la mer Caraibe. Nat. Verh. Holland Maats. Haarlem, 21: 119 pp., 25 pls. Ettis, J., & SoLanpDER, D. 1786. Natural History of Many Curious and Uncommon Zoophytes. pp. xii, + 208, 63 pls. London. Fristept, K. 1885. Bidrag till Kannedomen om de vid Sveriges vestra Kust lefvande Spongiae. K. svenska Vetensk-Akad. Handl. (6) 21: 1-56, pls. i-iv. Gray, J. E. 1858. On Aphrocallistes, a new genus of Spongiadae from Malacca. Proc. zool. Soc. Lond. : 114-115. 1870. Notes on anchoring sponges (Pheronema, Holtenia). Ann. Mag. nat. Hist. (4) 6 : 309-312. HarEckeEL, E. 1870. Prodromus eines Systems der Kalkschwamme. Jena. Z. Naturw. 5: 236-254. 1872. Die Kalkschwémme. 2 vols. & Atlas, 60 pls. Berlin HENTSCHEL, E. 1916. Die Spongien des Eisfjords. K. svenska. Vetensk-Akad. Handl. 54 (3) : 1-18. — 1929. Die Kiesel- und Hornschwamme des Nordlichen Meeres. Fauna arct. Jena, 5 : 859-1042, 4 pls. Hicein, T. 1877. Description of some sponges obtained during a cruise of the steam yacht “ Argo’ in the Caribbean and neighbouring seas. Ann. Mag. nat. Hist. (4) 19 : 291-199, I pl. Hozawa, S. 1918. Report on the calcareous sponges collected during 1906 by the United States Fisheries Steamer ‘‘ Albatross’’ in the Northwestern Pacific. Pyoc. U.S. nat. Mus. 54 : 525-556, 2 pls. 1933. Report on the calcareous sponges obtained by the survey of the Continental shelf bordering on Japan. Sci. Rep. Téhoku Univ. 8 (2) : 1-20, 1 pl. Tjima, I. 1927. The Hexactinellida of the Siboga Expedition. Siboga Exped. 40 (6) : 383 pp., 26 pls. Kent, W. S. 1870. On the Hexactinellidae. Mon. Microsc. J. 4: 241-252. LackscHEWwIt1z, P. 1886. Ueber die Kalkschwamme Menorcas. S. B. naturf. Ges. Dorpat, 7 : 336-341. 1886. Ueber die Kalkschwamme Menorcas. Zool. Jb. 1: 279-310, I pl. Lamarck, J.B. P.pEM. 1813. Sur les Polypiers empatés. Ann. Mus. Paris, 20 : 294-312, 370-386, 432-458. —— 1815. Suite des Polypiers empatés. Mém. Mus. Paris, 1: 69-80, 162-168, 331-340. —— 1816. Histoive naturelle des animaux sans vertébres. 2 Paris, 568 pp. Lamovurovux, J. V. F. 1816. Histoives des Polypiers covalligénes flexibles. Caen, pp. lxxxiv + 558, 19 pls. LENDENFELD, R. von. 1887. Die Chalineen des Australischen Gebietes. Zool. Jb. 2: 723- 828, ro pls. — 1889. A Monograph of the Horny Sponges. London. 936 pp. 30 pls. 1903. Tetraxonia [in] Tierveich, 19: 1-168, 44 figs. —— 1907. Die Tetraxonia. Wiss. Evgebn. ‘“‘ Valdivia,’’ 11 : 57-373, 46 pls. Levinsen, G.M.R. 1886. Kara-Havets Svampe (Porifera), pp. 84, 3 pls. LuBerktun, N. 1859. Neue Beitrage zur Anatomie der Spongien. Arch. Anat. Physiol. Lpz. 30 : 293-420, pls. ix—xi. 238 THE “ROSAURA” EXPEDITION LunpBEck, W. 1902. Homorrhaphidae and Heterorrhaphidae. Dan. Ingolf-Exped. 6 (1): 108 pp., 19 pls., 1 map. —— 1906. Porifera (Part III) Desmacidonidae (Pars). Jbid. 6 (3): 124 pp., 11 pls. MaRENZELLER, E. von. 1877. Die Coelenteraten, Echinodermen und Wiirmer der K. K. Oesterreichisch-Ungarischen Nordpol-Expedition. Denkschr. Akad. Wiss. Wien. 35: 357-398, pls. i-iv. MarsHALL, W. 1876. Ideen iiber die Verwandschaftsverhdltnisse der Hexactinelliden. Z. wiss. Zool, 27 : 113-156. MicutucHo-Macray, N. bE. 1868. Beitrage zur Kenntniss der Spongien. I. Ueber Guancha blanca, einen neuen Kalkschwamm. Jena Z. Naturw. 4: 221-240. Miter, O. F. 1776. Zoologiae Danicae, etc. Havniae. xxxii + 282 pp. Patras, P. S. 1766. Elenchus zoophytorum, etc. Hagae-Comitis. pp. 28, 451. Row, R. W. H. 1909. Reports on the marine biology of the Sudanese Red Sea. XIII. Report on the sponges collected by Mr, Cyril Crossland in 1904-5. Part I. Calcarea. J. Linn. Soc. Zool. 31 : 182-214, 2 pls., 8 text-figs. Sars, G. O. 1869. On the deep sea fauna of the Norwegian coast. Vidensk. Selsk. For- handl. : 250. —— 1872. On Some Remarkable Forms of Animal Life from the Great Deeps off the Norwegian Coast. Christiania. 82 pp., 6 pls. Scumipt, E.O. 1862. Die Spongien des adriatischen Meeves. Leipzig. 44 pp., 5 pls. —— 1870. Grundziige einer Spongien-Fauna des atlantischen Gebietes. Leipzig. 88 pp., 6 pls. 1880. Die Spongien des Meerbusens von Mexico und des Caraibischen Meeres. Heft 2. Jena. 33-90, pls. v—x. ScHUFFNER, O. 1877. Beschreibung einiger neuer Kalkschwamme. Jena. Z. Naturw. 11 : 403-433, pls. xxiv—xxvi. Scuuize, F. E. 1886. Ueber den Bau und das System der Hexactinelliden. Abh. preuss. Akad. Wiss. : 1-97. . 1887. Report on the Hexactinellida collected by H.M.S. “ Challenger.” Rep. Sct. Res. “‘ Challenger,’ Zool. 21 (53): 514 pp., 104 pls. I map. 1893. Revision des Systemes der Hyalonematiden. S.B. Preuss. Akad. Wiss. 30: 541-589. 1904. Hexactinellida. Wiss. Ergebn. ‘‘ Valdivia,” 4: viii + 226 pp., 52 pls. Sotias, W. J. 1886. Preliminary account of the Tetractinellia sponges dredged by H.M.S. “Challenger.” Sct. Proc. R. Dublin Soc. 5 : 177-199. 1888. Report on the Tetractinellida. Rep. Sci. Res. Voy. ‘‘ Challenger,” Zool. 25: clxvi, + 458, 54 pls. TuackerR, A.G. 1908. On collections of the Cape Verde Islands fauna made by Cyril Cross- land. Proc. zool. Soc. Lond. : 757-782, 1 pl., 12 text-figs. THIELE, J. 1898. Studien iiber pazifischen Spongien. Zoologica, Stuttgart, 24: 1-72, 8 pls. Tuomson, C. W. 1869. On Holtenia, a genus of vitreous sponges. Ann. Mag. nat. Hist. (4) 4: 284-287. Topsent, E. 1890. Notice préliminaire sur les Spongiaires receuillis durant les campagnes de l’Hirondelle. Bull. Soc. zool. Fr. 15 : 26-32, 65-71. 1892. Contributions a l’étude des Spongiaires de l’Atlantique Nord. Résult. Camp. sci. Monaco, 2: 165 pp., 11 pls. 1904. Spongiaires des Acores. Résult. Camp. sci. Monaco, 25: 280 pp., 18 pls. —— 1g1t. Sur les affinités des Halichondria et la classification des Halichondrines d’aprés leurs formes larvaires. Arch. Zool. exp. gén. (5) 7: 1-xv, 3 figs. 1913. Spongiaires provenant des campagnes scientifiques de la “‘ Princess Alice’ dans les Mers du Nord. Résult. Camp. sci. Monaco, 45 : 1-67, 5 pls. —— 1922. Sur Stelletta crassispicula Sollas et son synonyme Stelletta crassiclada Lendenfeld. Bull. Inst. océanogry. Monaco, No. 416: pp. 1-3. THE “ROSAURA” EXPEDITION 239 1928. Spongiaires de 1l’Atlantique et de la Mediterranée. Résult. Camp. sci. Monaco. 74 : 376 pp., I pls. 1934. Eponges observées dans les parages de Monaco. Bull. Inst. océanogr. Monaco, 650 : 1-42, 3 figs. Witson, H. V. 1902. The Sponges collected in Porto Rico. Bull. U.S. Fish. Comm. 2: 375-411, 30 figs. —— 1904. Reports on an exploration off the West Coasts of Mexico, Central and South America, and off the Galapagos Islands. Mem. Harv. Mus. comp. Zool, 30 (1) : 1-164, 26 pls. Fic. Fic. Fig. Fic. Fic. Fic. Fic. al By eS es as GGA select o rPmcoocini GH PLATE 9 Haliclona calcinea sp.n. X 1/1. Haliclona tenerrima sp.n. xX 3/2 Myxilla distortasp.n. X 1/1. Axinella vamosa sp.n. X 1/I. Hymeniacidon glabrata sp.n. X 1I/t. Halichondria cornuloides sp.n. X 1/1. Dragmatyle topsenti sp.n. X 3/2. Bull. B.M. (N.H.) Zool. 2, 6. ye “\ ae i “oe ~ PRINTED IN GREAT BRITAIN BY _ ADLARD AND SON, LIMITED, -BARTHOLOMEW PRES CS suo - iy oy i “a / i) yi > ~ ‘ ‘ A ~e . 2 i . indy che + ; a Sa ‘THE POLYZOAN GENUS CREPIDACANTHA -LEVINSEN =9 NOV1954 PRESENTED. D. A. BROWN sia BULEE TIN, OF SH MUSEUM WEES HISTORY) Let hee Voly2 "No 7 ~ LONDON: 1954 THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN BY DAVID ALEXANDER BROWN (Geology Department, University of Otago, Dunedin, New Zealand) Php. 241-263 ; 4 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 2 No. 7 LONDON : 1954 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, 1s issued in five series, corresponding to the Departments of the Museum, and an Historical Series. Parts appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 2, No. 7 of the Zoological Series. PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued October, 1954 Price Six Shillings ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN By DAVID ALEXANDER BROWN SYNOPSIS The family Crepidacanthidae is considered to include only the genus Crepidacantha Levinsen. The known species of that genus, including two new ones, are described, and their geographic distribution discussed. INTRODUCTION Durinc a study of the New Zealand Tertiary Cheilostomatous Polyzoa (Brown, 1952), in which the family Crepidacanthidae is moderately well represented, I have made many comparisons with material from other parts of the world. The results and conclusions concerning that family, which are beyond the scope of the Catalogue, are now put forward. Canu & Bassler (1929 : 409-412 ; 1930 : 32, 33) and Marcus (1938 : 233) have already dealt briefly with the species of Crepidacantha, but as their results, so far as the Indo-Pacific fauna was concerned, were probably not based on actual material, a good deal of confusion has occurred. This, I think, may now be rectified after an examination of the material in the British Museum (Natural History). ACKNOWLEDGMENTS My thanks are due to the Keepers of Zoology and Geology in the British Museum (Natural History) for facilities provided for my research, to the Director of Universi- tetets Zoologiske Museum, Copenhagen, for the loan of specimens from the Levinsen Collection, and to Dr. Anna B. Hastings and Dr. H. Dighton Thomas of the British Museum (Natural History) for their helpful criticisms of this paper. FAMILY CREPIDACANTHIDAE Levinsen’s original definition of this family (1909 : 266) is quite adequate, though it is clear that the words “‘ distal half’ should be substituted for “ proximal half ”’ in regard to the position of the “ 9-12 very long marginal spines,” and that ‘“‘ hetero- zooecia’’ is a better term than “ vibracula”’ in view of the large chambers occurring in C. zelanica Canu & Bassler which often appear to be avicularian in nature. Genus Crepidacantha Levinsen Crepidacantha Levinsen, 1909, Morph. Syst. Stud. Cheil. Bry.: 266; Brown, 1952 : 359 (cum syn.). 244 ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN TyPE SPECIES (by monotypy).! Cvepidacantha poissonii (Audouin) var. crimispina Levinsen, 1909 : 266, 267, text-figs. I, 2, 3, 5,6. Recent: Thailand. The orifice of this genus has a wide poster, and the recumbent? ovicell has an ectooecium with an uncalcified frontal area of variable size in which the entooecium, which usually has pores, is visible. Marginal spines occur along the distal-lateral edges of the zooecia alternating with the pore-chambers which are often marked by slit-like areolae. Paired heterozooecia’, usually vibracular in nature, with long setiform flagella, are present, either placed as tubular chambers on each side of the orifice or as rounded chambers on the frontal wall proximally to the orifice. Bassler (1935 : 81) records the genus as ranging from Cretaceous to Recent, but I have not been able to find any records of species earlier than the Waiauan [Middle Miocene] stage of New Zealand. However, as it is even then a clearly distinct genus, it will probably be found in beds of earlier age. [From the evidence available, it appears to be mainly a tropical and sub-tropical encrusting genus ranging from 0-250 fathoms, though specimens have been found in New Zealand waters as far south as Kaka Point (169° 50’ E., 46° 25’ S.). The orifice of Crepidacantha is evidently similar in construction to that of the family Hippoporinidae though rather peculiar in the breadth of the shallow poster. The latter may vary considerably in shape, with a convex proximal lip as in C. solea Canu & Bassler (Fig. 1H) or a very concave one as in C. setigera Smitt sp. (Fig. IE). However, a separate family appears to be justified in view of the nature of the orifice and of the constant occurrence of paired, symmetrically placed heterozooecia. Canu & Bassler (1929 : 408) apparently considered the presence of a recumbent ovicell as a family character for they grouped the genera Mastigophora Hincks and Schizobathysella Canu & Bassler with Crepidacantha on this account (see also Canu & Bassler, 1927: 21, where Pachykraspedon Koschinsky, Nimba Jullien and Nimbella Jullien are also included in the Crepidacanthidae ; and Canu & Lecointre, 1930: 110). An examination of these genera, however, shows clearly (from the nature of the orifice, which is, in my opinion, of far greater importance from the point of view of classification than the presence of a recumbent ovicell) that they can be grouped neither with Crepidacantha nor together. Mastigophora Hincks (= Escharina Milne-Edwards) and Schizobathysella Canu & Bassler belong to the Schizoporellidae. Nimba Jullien and Nimbella Jullien appear to be closely related, and, though Jullien & Calvet (1903) placed the former in the Schizoporellidae and the latter in the Sertellidae (Reteporidae), they should probably be placed together 1 It may be argued that Levinsen also included the species C. potssonit itself as a genosyntype. If this be accepted, then the selection of type must be Flustva poissonit Audouin, 1826, p. 240, pl. 10, figs. 51, 52, by Edwards (1910: 21). 2 This term appears to have been used first by Canu & Bassler (1917 : 66) and later amplified in their 1920 monograph (pp. 54, 55, text-fig. ro, ‘“‘ Independent (recumbent) aneucleithrian hyperstomial ovicell’’). The peculiarity of this type of ovicell may be seen at the edge of acolony. Here, the ovicell is observed to be attached to the distal wall of the fertile zooecium and is fully formed before any trace of distal zooecia appears. The latter then grow round the ovicell so that it appears more or less embedded in them. In the case of the ordinary hyperstomial ovicell, the frontal wall of the distal zooecium is in process of growth at the same time as the ovicell of the proximal zooecium, the rudiment of the ovicell making its appearance on the proximal part of the frontal wall of the developing distal zooecium (see Levinsen, 1909, pl. 19, fig. 4a). 3 The ancestrula on a specimen of C. altivostris Canu & Bassler (= C. crinispina Levinsen sp.) appeared to possess a single vibracular chamber (see under C. crinispina). ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN 245 in a separate family. Pachykraspedon Koschinsky probably belongs also to the Schizoporellidae as its primary orifice possesses a deep, often sub-circular sinus flanked by a pair of stout condyles ; there is a thick raised peristome and a coarsely perforate frontal wall. I therefore regard the Crepidacanthidae as including only one genus, as did Levinsen. a SS ee EES ee ee Oo A 0.25mm. (e) Boy O25mm. Fie. 1. Species of Crepidacantha showing orifice, ovicell and position of heterozooecia. A. Crepidacantha poissonii (Audouin) var. teres Hincks. Oblique view showing structure of ectooecium. B.M., 11.10.1.783. Recent: Madeira. 3B. The same, frontal view. c. C. crinispina (Levinsen). Lectotype. Recent: Bangkok, Thai- land. pb. C. cvinispina (Levinsen) var. pavvipora (Canu & Bassler). B.M., D.36804. Waiauan [Middle Miocene]: Weka Pass, New Zealand. E. C. setigera (Smitt). B.M., 32.3.7-27. Recent: Straits of Florida. ¥F. C. longiseta Canu & Bassler. B.M., 99-7-1.1722. Recent: John Adams Bank, Brazil. G. C.gvandis Canu & Bassler. B.M., 31.12.30.158. Recent: Philippines. u. C. solea Canu & Bassler. Lectotype, B.M., 89.8.21.19. Recent: Tizard Reef, China Sea. J. C. zelanica Canu & Bassler. B.M., D.36897. Nukumaruan [Middle Pliocene]: Petane, New Zealand. ZOOL. Il, 7. 15§ 246 ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN Vigneaux’s contention (1949 : 1g) that the family Crepidacanthidae is superfluous and that Crepidacantha, along with a number of other genera, should be placed in the family Phylactellidae! appears to be quite without foundation. In so far as Crepidacantha is concerned, the only point of resemblance to Phylactella Hincks, 1879 [genolectotype (chosen by Canu & Bassler, 1917 : 67) : Lepralia labrosa Busk, 1854: 65, 82, pl. 92, figs 1, 2, ? 3] is the possession of a recumbent ovicell. Livingstone (1929 : 94) has stated that the position of the vibracula is of little importance in classifying the species of Crepidacantha. However, the orientation of the flagella (ie., laterally or proximally) does appear to be constant within a species, and it has been found, with one exception, viz., C. crimispina (Levinsen) var. parvipora Canu & Bassler var., that where the vibracular chambers are placed proximally to the orifice, the flagella are directed laterally inwards, whereas those species with vibracular chambers alongside the orifice have their flagella directed proximally. It is found also that these differences in the orientation of the flagella are accompanied by other important distinguishing characters such as the structure of the ovicell and the shape of the zooecial orifice. KEY TO THE SPECIES OF Crepidacantha DESCRIBED HERE 1. Ovicell with uncalcified area of ectooecium occupying most of the frontal area 2 Ovicell with uncalcified area of ectooecium much less extensive 3 : 5 . LO, . Heterozooecia placed alongside orifice. Flagella directed proximally 7 Heterozooecia placed proximally to orifice. Flagella usually directed laterally inwards. 5. 3. Ovicell large. Poster very deep, concave . é ¢ . E. C. setigera (Smitt). Ovicell not large. Poster shallow, convex . : * . ¢ 0 : . 4. 4. Proximal lip of orifice square and plate-like. Calcified ectooecium marginal c. C. crinispina (Levinsen). Proximal lip of orifice rounded, convex. Calcified ectooecium wide Gc. C. grandis Canu & Bassler. 5. Heterozooecia large. Ovicell coarsely perforate . . Jj. OC. zelanica Canu & Bassler. Heterozooecia small . : : : 6. 6. Median frontal umbo much enlarged. Flagella directed proximally D. C. crinispina (Levinsen) var. parvipora (Canu & Bassler). nN Median tubercle small or absent. Flagella directed inwards . é 2 3 ° Te 7. Entooecium with longitudinal radiating ridges c : 7 : 8. Entooecium without radiating ridges, but may have median c carina c 9. 8, Proximal lip of orifice convex, rounded 3 : 6 AEC. poissonii (Audouin). Proximal lip of orifice convex, square, plate-like . : H. C. solea Canu & Bassler. g. Ovicell deeply immersed. Calcified ectooecium nataer Proximal lip of orifice convex, square, plate-like . : ; L. C. kirkpatricki, sp. nov. Ovicell globular. Calcified ectooeciant covering proximo- -lateral portions of ovicell. Proximal lip of orifice nearly straight . 6 . B. C. poissonii (Audouin) var. teres Hincks. ro, Entooecium exposed in depressed transverse oval area . F. C. longiseta Canu & Bassler. Entooecium exposed in raised longitudinal ovalarea_.. K. C. bracebridgei, sp. nov. A. Crepidacantha poissonii (Audouin) Flustva poissonii Audouin, 1826, Description de l’ Egypte :240; Savigny, 1826, pl. ro, figs. 51, 52; Audouin, 1828 : 68. Reptescharellina poissonii d’Orbigny, 1852, Paléontologie Frangaise. Terr. Crét. 5 : 454. Crepidacantha poissoni, Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 409, text-fig. 1604. 1 See also Canu & Bassler, 1923: v. ' ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN 247 Diacnosis (from Savigny’s figure). Crepidacantha with entooecium ornamented frontally with longitudinal radiating bars. Vibracula placed proximally to orifice. Proximal lip of orifice convex, rounded. Remarks. Audouin (1826: 240; 1828: 68) in his notes on the species figured in Savigny’s plate ro, gives no locality for these, and they may, therefore, be either from the Red Sea or the Egyptian coast of the Mediterranean. I have not seen any specimens from either of these localities? and I am unable to decide whether this species is the same as the variety from Madeira (see below). Savigny has, however, clearly figured radiating longitudinal ridges on the frontal surface of the entooecium and these do not appear on any of the specimens examined from Madeira. The proximal lip of the orifice of Audouin’s species is also much more convex than that of the Madeiran material, which I have recognized, therefore, as a distinct variety, teres Hincks. It is unfortunate that Savigny did not indicate in his figure the orientation of the vibracular setae, but it is most probable that, as the vibracular chambers are placed proximally to the orifice, the flagella were directed laterally inwards. DISTRIBUTION!, ? Eastern Mediterranean ; ? Red Sea. B. Crepidacantha poissonii (Audouin) var. teres Hincks (Fig. IA, B) Lepralia kivchenpaueri Heller var. teres Hincks, 1880, Ann. Mag. nat. Hist. (5) 6: 77, pl. 9, figs. 7, 74; Hincks, 1891 : 88, 89. Lepralia teres Hincks, 1895, Index Mar. Polyzoa : ii, note. Lepralia poissonii Audouin, Waters, 1899, J.R. micr. Soc.: 16; Norman, 1909: 307, pl. 41, figs. 7,8; L. poissoni, Calvet, 1907: 409. Crepidacantha poissoni, Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 409, text-figs. 1608, c; C. poissonii, Osburn, 1940: 451; ? Crepidacantha poissonii, Canu & Bassler, 19284 : 136, pl. 34, fig.3. MATERIAL EXAMINED :— B.M., 11.10.1.783% (Figs. 1a, B). Recent: Madeira. Specimen encrust- ing small Pecten. Labelled ‘‘Lepralia Poissonii Aud. Madeira, 1897. A.M.N.”’ Norman Collection. B.M., 11.10.1.784. Recent: Madeira. Labelled “ Lepralia Potssonii Audouin. Salvages. Senor D. Noronha.” Norman Collection. B.M., 11.10.1.785. Recent : Madeira. Labelled “Lepralia Poissonii Aud. Porto Santo. Senor de Noronha. 1908.” Norman Collection. Dracnosis. Crepidacantha with the calcified ectooecium confined to the lateral Margins of the ovicell, the entooecium with a transverse porous ridge. Vibracular flagella directed laterally. Proximal lip of orifice straight or slightly convex. Marginal spines slender, about twelve in number. * D'Orbigny (1852 : 454) cites the Red Sea for C. poissonit. See also Postscript. ® Monsieur E. Buge of the Laboratoire de Paléontologie, Muséum National d’Histoire Naturelle, Paris, states (in litt.) that the Savigny & Audouin Collection is considered lost. * Specimens in the British Museum (Natural History) are referred to by their Register Numbers, 248 ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN Remarks. The chief differences between this variety and Audouin’s species have been noted above. Another important character is the proximo-lateral extension of the calcified ectooecium which covers the entooecium near the orifice. The ridge with the row of pores traversing the frontal wall of the ovicell does not mark the edge of the calcareous ectooecium, as I thought at first (Brown, 1952 : 361), but is an ornamentation of the entooecium. DISTRIBUTION. Recent: Madeira ; Cape Verde Islands, 110-180 metres (Calvet) ; Porto Rico, 6 fathoms (Osburn) ; Bermuda (Osburn). Fossil: ? Pliocene: Bocas Island, NW. Panama (Canu & Bassler). c. Crepidacantha crinispina (Levinsen) (Fig. 1c) Lepralia poissonii Audouin, Waters, 1887, Q. J. geol. Soc. Lond. 43: 42, 59 (part.—non pl. 8, fig. 37 = C. zelanica Canu & Bassler); Philipps, 1899: 440, 446; ? Lepralia poissonii, Hincks, 1881: 122; ? Lepralia poissonii (second form), Hincks, 1885 : 256. Lepralia setigera Smitt, MacGillivray, 1883, Trans. roy. Soc. Vict. 19: 133, pl. I, figs. 2, 3; MacGillivray, 1887 : 212. ? Crepidacantha poissoni Livingstone, 1929, Vidensk. Medd. dansk naturh. Foren. Kbh. 87 : 93. Crepidacantha poissonii Audouin var. cyinispina Levinsen, 1909, Morph. Syst. Stud. Cheil. Bry. : 266, text-figs. 1, 2, 3, 5, 6; C. poissoni var. crinispina, Canu & Bassler, 1923: 174, text-figs. 33A-—C, E, F. Crepidacantha setifera Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 409; Canu & Bassler, 1930 : 32, 33. Crepidacantha papulifera Canu & Bassler, 1929, Bull. U.S. nat. Mus, 100 (9) : 410, text-figs. 160K, 1618, pl. 57, fig. 8. Crepidacantha crinispina (Levinsen), Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 409, text-figs. I60D-F, H, 1; Brown, 1952 : 359, text-fig. 283 ; Brown, 1954: 433. Crepidacantha altivostvis Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 411, pl. 57, fig. 9. LEcTOTYPE (chosen by Brown, 1952: 359). Universitetets Zoologiske Museum, Copenhagen, specimen encrusting fragment of large lamellibranch (Fig. 1c). Recent: Koh Kram, Bangkok, Thailand, 30 fathoms. Levinsen Collection. OTHER MATERIAL EXAMINED :— B.M., 97.5.1.832, 835. Recent: Port Phillip Heads, Victoria. Two encrusting specimens. Labelled “‘Lepralia setigera.’’ Bracebridge Wilson Collection. B.M., 88.11.14.307. Recent: Port Phillip, Victoria. Labelled “ Lepralia poissontt Aud.” by Kirkpatrick. Bracebridge Wilson Collection. B.M., 99.5.1.825. Recent: Wellington, New Zealand. Labelled “ Lepralia Poissontt Audouin” by Miss Jelly. Hincks Collection. B.M., 31.12.30.157. Recent: Albatross Stn. 5179, 37 fathoms off Romblon Light, Romblon, Philippines. Paratype of C. altivostris Canu & Bassler. Presented by the U.S. National Museum. B.M., D.36893. Castlecliffian CU3 [Upper Pliocene]: Castlecliff, New Zealand. Sent by Mr. C. A. Fleming. Waitotaran [Middle Pliocene]: submarine limestone off Three Kings Islands, New Zealand. ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN 249 Diacnosis. Cvepidacantha with the calcified ectooecium confined to the distal- lateral margins of the ovicell. Entooecium with a longitudinal median carina. Vibracula placed alongside the opercular condyles, the flagella directed proximally. Proximal lip of orifice convex, often plate-like. Frontal wall of zooecium often bearing a prominent sub-oral tubercle. Marginal spines ten in number. Remarks. In addition to the comments already made on this species by Brown (1952 : 360) the following remarks are apposite. This species is clearly distinct from both C. poissonii (Audouin) and from C. poissonti var. teres in the different construction of its ovicell and in the position and orientation of its vibracular setae. In the lectotype the ovicells have marked eversion of the transverse ridge, which tends to be masked in the Australian and New Zealand specimens by thickening of the ectooecium. Although MacGillivray (1883 : 133, pl. 1, fig. 2) does not show the median carina on the ovicell of his Lepralia setigera Smitt, there can be little doubt that this is the species described by him, as it differs completely from the other species of Crept- dacantha from Port Phillip, Victoria, namely, C. bracebridgei, sp. n., and C. kirkpatrickt, sp. n. This identification is substantiated by two specimens from Port Phillip Heads sent by Mr. J. Bracebridge Wilson and labelled by him “Leprala setigera”’ (B.M., 97.5.1.832, 835). Canu & Bassler’s name, C. setifera (1929 : 409), proposed for Lepralia setigera MacGillivray, must therefore lapse in favour of Levinsen’s name, C. crinispina. The specimen (B.M., 88.11.14.307) from Port Phillip, Victoria, has a very marked sub-oral umbo resembling that of C. crinispina (Levinsen) var. parvipora Canu & Bassler var. from New Zealand, but the plate-like proximal lip of the orifice and the rather distally placed vibracula distinguish it from that variety. In the Recent specimen (B.M., 99.5.1.825) from Wellington, New Zealand, the vibracular chambers are rather larger than those of the lectotype and the sub-oral tubercle is often absent : it is probably C. crinispina. Canu & Bassler’s reasons for separating C. papulifera (1929 : 400, text-fig. 160K ; 410, 411, text-fig. 1618, pl. 57, fig. 8) from C. crimispina appear to be without foundation. C. altivostris Canu & Bassler (1929 : 411, pl. 57, fig. 9) is represented in the British Museum (Natural History) by a paratype (B.M., 31.12.30.157) and appears to be conspecific with C. crinispina. The vibracula are sometimes placed near the distal end of the zooecium as stated by Canu & Bassler, but this feature is also seen in the lectotype of Levinsen’s species. The only difference to be seen in the specimen of C. altirostris is the rather more marked carination of the entooecium which shows no eversion of the transverse ridge. The ancestrula is preserved in this specimen. It is small and appears to have a single vibraculum placed proximally and to one side of the orifice. DIstRIBUTION. Recent: Thailand (30 fathoms) ; Philippines (21-105 fathoms) ; Port Phillip, Victoria; Loyalty Islands, New Caledonia (35 fathoms) (Philipps) ; ? Bass Strait (Hincks) ; New Zealand (3-10 fathoms). Fossil; Castlecliffian [Upper Pliocene]: Castlecliff; ? Shakespeare Cliff, New 250 ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN Zealand. ? Nukumaruan [Middle Phocene]: Waipukurau, New Zealand (Waters). Waitotaran [Middle Pliocene]: submarine limestone off Three Kings Islands, New Zealand. Db. Crepidacantha crinispina (Levinsen) var. parvipora (Canu & Bassler) (Fig. 1D) Lepralia poissonit Audouin (third form), Hincks, 1885, Ann. Mag. nat. Hist. (5) 15: 256; L. poissonii, Waters, 1887 : 59 (part.—non pl. 8, fig. 37 = C. zelanica Canu & Bassler). Crepidacantha parvipora Canu & Bassler, 1930, Proc. U.S. nat. Mus. 76 (13) : 32, 33- Crepidacantha crinispina (Levinsen) var. parvipora, Brown, 1952, Tert. Cheil. Polyzoa N. Zealand : 361, text-fig. 284. NEOTYPE (chosen by Brown, 1952 : 361). B.M., D.37042. [Pliocene]: [Napier], New Zealand. Labelled ‘“‘ Lepralia Poissonii var. R. Tert. N. Zealand’’ by Miss Jelly. Hincks Collection. OTHER MATERIAL EXAMINED :— B.M., D.36777.. Nukumaruan [Middle Pliocene]: Waipukurau Gorge. Encrusting specimen. Hincks Collection. B.M., D.32531. Nukumaruan [Middle Pliocene}: Waipukurau Gorge. Encrusting specimen. Labelled “Lepralia Poissonii Aud.” by Miss Jelly. Hincks Collection. B.M., D.36894 (Fig. 1p)-36896. Waiauan [Middle Miocene], Base of Uppermost Mt. Brown “ E”’ Limestone : Junction of Weka Creek and Weka Pass Stream, Waipara, S.D. Three encrusting specimens. Sent by Pro- fessor B. H. Mason. DiaGnosis. C. crinispina with a stout median ridge on the frontal wall ending distally in a sub-oral umbo. Vibracula placed on either side of the umbo, the flagella directed proximally. REMARKS. This variety has been fully discussed by Brown (1952 : 361). Unlike the other known species of Crepidacantha in which the vibracula are placed proxi- mally to the orifice, this variety has the flagella directed proximally. The presence of the marked sub-oral umbo would inhibit or prevent the inward turning of the vibracula. DistRIBUTION. Fossil: Nukumaruan [Middle Pliocene]: [Napier]; Waipu- kurau Gorge, New Zealand. Waiauan [Middle Miocene] : Weka Pass, New Zealand. E. Crepidacantha setigera (Smitt) (Fig. IE) Escharella setigera Smitt, 1873, K.svensk. Vetensk.—Akad. Hand. 11 (4) : 58, 75, 82, pl. ro, fig. 206. Crepidacantha setigera, Canu & Bassler, 19284, Proc. U.S. nat. Mus. 72 (14) : 135, pl. 21, fig. 10; Canu & Bassler, 1929: 409; Osburn, 1940: 452; Osburn, 1952: 479, pl. 58, fig. 1. MATERIAL EXAMINED :— B.M., 32.3.7.27 (Fig. IE). Recent: Albatross Stn. 2639, Straits of Florida. Two encrusting specimens. Presented by the U.S. National Museum. Dracnosis. Crepidacantha with large ovicell not clearly separated from the zooecium, the entooecium with a curved row of pores, the ectooecium wide. ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN 251 Vibracula placed just above the level of the opercular condyles, the flagella directed proximally. Poster very deep and concave. REMARKS. This species is very clearly distinguished by the shape of its orifice and its large ovicell. Smitt (1873:57, 82) regarded it as a variety of Lepraha depressa Busk (1854 : 75, pl. 91, figs. 3, 4), to which, however, it does not seem to be related. Osburn’s record (1952: 479) of this species from the Galapagos Islands is interesting. Whether the distribution of this and other Polyzoan species, now living on both sides of the Isthmus of Panama, indicates a geologically recent connection or is due to transport of larvae and colonies on, for instance, the bottoms of ships traversing the canal, will probably never be known. It is likely, however, that the fresh-water Gatun Lake would be an efficient barrier to such a passage. DistRiBuTION. Recent: Gulf of Mexico (50-60 fathoms) : Galapagos Islands (Osburn). F. Crepidacantha longiseta Canu & Bassler (Fig. IF) Lepralia poissonii Audouin, Kirkpatrick, 1888a, Ann. Mag. nat. Hist. (6) 1: 78, pl. 8, fig. 1; ? L. poissonti, Thornely, 1912: 150; Waters, 1914 : 856. Crepidacantha longiseta Canu & Bassler, 19284, Proc. U.S. nat. Mus. 72 (14) : 135, pl. 21, figs. 3,4; Osburn, 1940: 452. Crepidacantha levinseni Marcus, 1938, Vidensk. Medd. dansk naturh. Foren. Kbh. 101 : 231 (part.), text-figs. 28a, B. Hototyre. U.S. National Museum No. 7826. Recent: Cuba. MATERIAL EXAMINED :— B.M., 88.1.25.31. Recent: Mauritius. Labelled “Lepralia Potssonit Audouin.” V. Robillard Collection. B.M., 34.10.6.19. Recent: Mauritius. B.M., 99.7.1.1722 (Fig. 1F). Recent: John Adams Bank, South Atlantic, off Brazil. Labelled “ Lepralia Poissonit Aud.” Busk Collection. Diacnosis. Crepidacantha with porous entooecium visible through a transverse, sub-circular or oval fenestra in the ectooecium. Vibracular chambers tubular, placed at the level of the opercular condyles, the flagella very long and directed proximally. Proximal lip of orifice straight or slightly concave. Marginal spines ten in number. Remarks. I have not examined any West Indian material of C. longiseta, but a specimen (B.M., 99.7.1.1722, Busk Coll., collected by H.M.S. “ Herald’ from the John Adams Bank off the coast of Brazil), undoubtedly represents that species. This specimen also agrees closely with the Jectotype (here chosen) of Marcus’s C. levinseni from St. Helena (viz., the specimen figured by him, 1938, text-fig. 28a on p- 232), except that the porous area (entooecium) in the ectooecium of the Brazilian specimen appears to be quite separated from the orifice (Fig. rr). As shown later, the species from Port Phillip, Victoria, with which Marcus correlated his C. levinsent, is new, namely, C. kirkpatricki. 252 ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN The original material from Mauritius described by Kirkpatrick as Lepralia Poissonti Audouin (B.M., 88.1.25.31), and another specimen from the same locality (B.M., 34.10.6.19) are conspecific. They differ from the Brazilian material only in having a slightly longer orifice and in having a rather coarsely granular frontal wall. Marcus (1938 : 233-234 key and distribution statements) working, as he explained, from the literature without material for examination, described the ovicell area of C. longiseta as without pores, thus distinguishing it from C. levinsent. He did, however, refer the Mauritius form to C. longiseta, but also regarded the form from the China Sea (C. solea) as synonymous. Osburn (1940 : 452; 1952: 479) thought that C. longiseta might be synonymous with C. setigera (Smitt), but the ovicell and the orifice of the latter species are clearly distinctive. DISTRIBUTION. Recent: Cuba (67-201 fathoms) (Canu & Bassler); John Adams Bank, Brazil ; St. Helena (36-69 metres) (Marcus) ; Mauritius ; ? Zanzibar (Waters) ; ? Providence, NNE. Madagascar (50~—78 fathoms) (Thornely). c. Crepidacantha grandis Canu & Bassler (Fig. 1G) Crepidacantha grandis Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 411, text-figs. 160L, 161A, pl. 57, figs. 4-7. SynTyPEs. U.S. National Museum No. 8220. Recent: Philippines. MATERIAL EXAMINED :— B.M., 31.12.30.158 (Fig. 1G). Recent: Albatross Stn. 5217, 105 fathoms, Ragay Gulf, off North Burias, Philippines, in coarse, grey sand. DiaGnosis. Crepidacantha with large zooecia, the calcified ectooecitum confined almost entirely to the distal-lateral edges of the ovicell. Vibracula placed at the level of the large opercular condyles, the flagella directed proximally. Proximal lip of orifice very convex and rounded. RemMARKS. This species appears to be quite distinct. Canu & Bassler state (1929 : 412) that on the “‘ nonmarginal zooecia the setiform spines [spines arising from the marginal portions of the zooecia] are short (= 0-15 — 0-20 mm.) and all equal; they are very long on the contrary (= 0-60 mm.) on the free side of the marginal zooecia.”” In a number of zooecia in the British Museum specimen (B.M., 31.12.30. 158) the frontal wall is punctured by a large rounded hole through which, presumably, the polypide was extracted for food by some predatory borer. DIsTRIBUTION. Recent: Philippines (105 fathoms). H. Crepidacantha solea Canu & Bassler (Fig. TH) Lepralia poissonii Audouin (first form), Hincks, 1885, Ann. Mag. nat. Hist. (5) 15 : 256 (part.) ; L. poissonit, Kirkpatrick, 1890 : 16. } ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN 253 Crepidacantha poissonit, Canu & Bassler, 1928b, Bull. Soc. Sci. nat. méd. Seine-et-Oise (2) 8 (7) suppl. : 37, pl. 8, fig. 7; Canu & Bassler, 1930: 33, pl. 5, fig. 5; C. potssont, Osburn, 1952: 478, pl. 58, fig. 2. Crepidacantha solea Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 409; Canu & Bassler, 1930 : 32. LEcTOTYPE (here chosen). B.M., 89.8.21.19 (Fig. tH). Recent: Tizard Reef, China Sea, 27 fathoms. Specimen encrusting lamellibranch. Labelled “‘ Lepralia poisson Aud.” by Kirkpatrick. H.M.S. ‘‘ Rambler ”’ Collection. OTHER MATERIAL EXAMINED :— B.M., 97.5.1.828. Recent: Tahiti. Labelled “Lepralia Poissoniit Aud.” by Miss Jelly. Bracebridge Wilson Collection. B.M., 99.5.1.824. Recent: Tahiti. Labelled “ Lepralia Poissonit Aud. 7gt. Avicularia horizontal ’’ by Miss Jelly. Hincks Collection. Note. Another slide, B.M., 99.5.1.823, Hincks Collection, similarly labelled, has no specimen. DiaGnosis. Cyrepidacantha with a prominent transverse ridge across the entooecium which also has a faint median carina and traces of longitudinal markings. Calcified ectooecium confined to the distal-lateral margins. Vibracula large, placed proximally to the orifice, the flagella directed laterally inwards. Proximal lip of orifice square, plate-like. Remarks. Canu & Bassler (1929 : 409) introduced the name Crepidacantha solea for ‘‘ Lepralia poissonni |sic], Kirkpatrick, 1888, from Mauritius, Sea of China and Australia.” Kirkpatrick’s material from these localities is preserved in the British Museum collections and proves to belong to three different species. Later (1930 : 32), however, Canu & Bassler restricted the name C. solea to Kirkpatrick’s material from the China Sea (1890 : 16) and the species is here understood in that sense. C. solea is, indeed, very closely related to the original Flustra poissonii Audouin, especially in the appearance of faint longitudinal markings! on the frontal part of the entooecium. It is distinguished by its large vibracula placed more proximally than usual and by the square plate-like lip of the orifice. Hincks (1885 : 256) recorded two forms of “Lepralia Poissonit, Audouin’”’ from Tahiti. The first, in which ““ ... the vibracula are situated below the orifice and are placed horizontally,” is undoubtedly the present species as shown by Hincks’s material (B.M., 97.5.1.828 and B.M., 99.5.1.824). The second (B.M., 99.5.1.825), with vertical vibracula placed alongside the orifice, is probably C. crinispina (Levinsen). It is evident that the material from the Hawaiian Islands and the Galapagos Islands, identified by Canu & Bassler (1928): 37; 1930: 32, 33) as C. poissonit, can also be placed here, as well as Osburn’s material from the American Pacific coast (1952: 478, pl. 58, fig. 2). DistripuTion. Recent: Tizard Bank, China Sea (27 fathoms); Tahiti; Hawaii (130-250 fathoms) (Canu & Bassler) ; Galapagos Islands (33-40 fathoms) (Canu & Bassler) ; American Pacific coast (Osburn). 1 See Postscript. 254 ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN J. Crepidacantha zelanica Canu & Bassler (Fig. 1J) Lepralia poissonit Audouin (first form), Hincks, 1885, Ann. Mag. nat. Hist. (5) 15 : 256 (part.) ; L. poissonii, Waters, 1887 : 59 (part.), pl. 8, fig. 37. Crepidacantha poissont (?) Waters, Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 409, text-fig. 160J. Crepidacantha zelanica Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 410 ; Canu & Bassler, 1930 : 32, 33; Brown, 1952 : 362, text-figs. 285, 286. NeEotyPe (chosen by Brown, 1952: 362). B.M., 99.5.1.826. Recent: Napier, New Zealand. Specimen encrusting pebble. Labelled ‘‘ Lepralia Poisson, Aud.” by Miss Jelly. Hincks Collection. OTHER MATERIAL EXAMINED :— Recent : Waitarere Beach, Wellington, 25 fathoms. Specimen encrusting concretion. Geology Department, Otago University. Recent: Kaka Point, Southland, New Zealand. Specimen encrusting Atrina zelandica (Gray). Geology Department, Otago University. B.M., D.32531. Nukumaruan [Middle Pliocene}: Waipukurau Gorge. Encrusting specimen. Labelled ‘“‘Lepralia Poissonii Aud.” by Miss Jelly. Hincks Collection. B.M., D.36901. Nukumaruan [Middle Pliocene]: Napier Harbour. En- crusting specimen. Hincks Collection. B.M., D.1430, D.1431. Nukumaruan [Middle Pliocene]: Petane. Two encrusting specimens. Labelled ‘‘ Lepralia Poissonit Aud.” by Miss Jelly. Vine Collection. B.M., D.36897 (Fig. 1j)—36900. Nukumaruan [Middle Pliocene]: Petane. Four encrusting specimens. Blake Collection. Nukumaruan [Middle Pliocene]: Waipukurau. Encrusting specimen. Labelled “Lepralia Poissonii”” by Waters. Waters Collection, Manchester Museum. Nukumaruan [Middle Pliocene]: Petane. Encrusting specimen. Labelled “Lepralia Poissonit”’ by Waters. Waters Collection, Manchester Museum. Diacnosis. Cvrepidacantha with uncalcified area of ectooecium occupying most of the frontal surface of the globular ovicell, exposing coarsely perforate entooecium. Heterozooecia large, prominent, placed proximally to the orifice, the flagella directed inwards. Orifice with slightly convex proximal lip. ReMARKS. This species has been fully discussed by Brown (1952 : 362). In the Recent specimen from Kaka Point, Southland, the heterozooecia, although fairly large in size, are still vibracular in nature and have pronounced setiform flagella. The first form of “Lepralia Poissonii, Audouin’’ recorded from New Zealand by Hincks (1885 : 256) is almost certainly this species though no material labelled by Hincks is available. DISTRIBUTION. Recent: Napier; Wellington; Kaka Point, New Zealand. Fossil: Nukumaruan [Middle Pliocene]: Petane; Waipukurau Gorge; Napier Harbour. ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN 255 kK. Crepidacantha bracebridgei! sp. nov. (Fig. 2) HoLotyre. B.M., 88.11.14.224 (Fig. 2). Recent: Port Phillip Heads, Victoria. Specimen encrusting lamellibranch shell. Slide labelled ‘‘Lepralia poissonii Audn.”’ by Kirkpatrick. Bracebridge Wilson Collection. OTHER MATERIAL EXAMINED :— B.M., 88.11.14.434. Paratype. Recent: Port Phillip Heads. Encrust- ing specimen. Labelled “Lepralia. Jan. 1887’ by Kirkpatrick. Brace- bridge Wilson Collection. B.M., 97.5.1.833. Paratype. Recent: Port Phillip Heads. Encrusting specimen. Labelled ‘“‘Lepralia setigera. 1883” by J. Bracebridge Wilson. Bracebridge Wilson Collection. B.M., 88.11.14.116. Paratype. Recent: Port Phillip Heads. Encrusting specimen. Labelled “‘ Lepralia setigera MacG. 1884” by J. Bracebridge Wilson. Bracebridge Wilson Collection. Diacnosis. Cvrepidacantha with the porous entooecium visible in a narrow, oval, longitudinal fenestra in the ectooecium. Vibracula placed on the distal side of the opercular condyles, the flagella very long and directed proximally. A prominent sub-oral tubercle usually present. Orifice long and narrow with a convex proximal lip. ee es en es See (eo) 0.5mm. Fic. 2. C. bracebridget, sp. nov. Holotype, B.M., 88.11. 14.224. Recent : Port Phillip Heads, Victoria. Description. Zoarium encrusting. Zooecia ovate or hexagonal, arranged quincuncially in radiating rows, distinctly separated by deep furrows. Orifice elongate, divided into an oval anter and a shallow poster, with a convex or squared proximal lip, the opercular condyles strong. Peristome thin and raised distally into a hood on non-ovicelled zooecia. Frontal 1 After Mr. J. Bracebridge Wilson, the Australian naturalist. 256 ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN wall ventricose, finely granular and porous, with a prominent sub-oral tubercle. Marginal spines about twelve in number. Vibracula paired, placed above the level of the opercular condyles, the flagella directed proximally and reaching beyond the proximal margins of the zooecium. Ovicells globular, hyperstomial, longer than wide, with the uncalcified area of the ectooecium small, longitudinally-oval, placed towards the distal end of the ovicell, somewhat raised. Often a distinct umbo at its proximal end. Exposed entooecium porous. MEASUREMENTS. L, = 0-50-0°55 mm.; /, = 0:29-0-33 mm. h, = 0:08-0:09 mm. ; J, = 0:07-0:08 mm. REMARKS. The ovicell of this species is very distinctive and differs from that of C. longiseta in that the porous area is always raised and not depressed or surrounded by salient walls as it is in the latter species. The area is, moreover, usually placed transversely in C. longiseta. The presence of a sub-oral tubercle in the Australian species is probably of no great significance as its size appears to vary with the age of the zooecia, but the hooded peristome of the non-ovicelled zooecia may be an important character. Although two of the specimens mentioned above in the Bracebridge Wilson Collection (B.M., 97.5.1.833 and B.M., 88.11.14.116) are labelled “ Lepralia setigera,’ they do not belong to the species described by MacGillivray as Lepralia setigera Smitt (1883 : 133, pl. I, figs. 2, 3) as his description and figure of the ovicell are clearly those of C. crinispina Levinsen sp. DistriBuTION. Recent: Port Phillip Heads, Victoria. L. Crepidacantha kirkpatricki’ sp. nov. (Fig. 3) Lepralia poissonii Audouin, Kirkpatrick, 1888b, Ann. Mag. nat. Hist. (6) 2: 14. Lepralia poissonii Audouin var. Waters, 1889, J. R. micr. Soc. : 14, pl. 2, fig. 17. Crepidacantha poissoni Audouin var. Levinsen, 1909, Morph. Syst. Stud. Cheil. Bry. : 268, text-fig. 4; Canu & Bassler, 1923 : 174, text-fig. 33D. Crepidacantha crinispina Levinsen var. Canu & Bassler, 1929, Bull. U.S. nat. Mus. 100 (9) : 410, text-fig. 160G. Crepidacantha levinseni Marcus, 1938, Vidensk. Medd. dansk naturh. Foren. Kbh. 101 : 231 (part.— non text-figs. 28A, B). Hototyre. B.M., 88.11.14.426 (Fig. 3). Recent: Port Phillip Heads, Victoria. Specimen encrusting Fucus. Labelled “ Lepralia poissonii Audn.” by Kirkpatrick and “‘ Jan. 1887’ by J. Bracebridge Wilson. Bracebridge Wilson Collection. OTHER MATERIAL EXAMINED :— B.M., 97.5.1.834. Paratype. Recent: Port Phillip Heads. Labelled “Lepralia setigera. 1885” by J. Bracebridge Wilson. Bracebridge Wilson Collection. Paratype: Universitetets Zoologiske Museum, Copenhagen No. 1/8/90. Recent: Port Phillip Heads. Levinsen Collection. 1 After the late Dr. R. Kirkpatrick, formerly of the Department of Zoology, British Museum (Natural History). ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN 257 Dracnosis. Cyrepidacantha with the ovicell flattened and deeply immersed in the distal zooecium, the uncalcified area of the ectooecium occupying almost the whole of the exposed frontal surface. Vibracula placed proximally to the orifice, the flagella short and directed laterally inwards. A small sub-oral tubercle present. Proximal lip of orifice squared, plate-like. Marginal spines of mature zooecia short and thick, of marginal zooecia long and flexible. DESCRIPTION. Zoarium encrusting. Zooecia ovate, arranged somewhat irregularly in radiating rows, scarcely separated by shallow furrows. Orifice sub-circular distally with a shallow poster separated by inconspicuous condyles, the proximal lip square and plate-like. Pevistome thin, scarcely raised. Frontal wall usually flattened, occasionally ventricose, finely perforate and covered with small papillae, a small, sub-oral tubercle often present. Marginal spines of mature zooecia short and thick, eight in number, those of the young zooecia long and flexible. (ee ee (@) O.5mm. Fic. 3. C. kirkpatricki, sp. nov. Holotype, B.M., 88.11.14.426. Recent: Port Phillip Heads, Victoria. Vibracula paired, placed proximally to and on each side of the orifice, the short flagella directed obliquely laterally inwards. Ovicells large, sub-circular in outline, deeply immersed in mature zooecia, the exposed frontal surface flattened and occupied almost entirely by a porous entooecium, the calcified ectooecium marginal. MEASUREMENTS. L, = 0:35-0:40 mm. ; /, = 0:25-0:28 mm. h, = 0:09-0:10 mm. ; /, = 0:07-0:08 mm. Remarks. The holotype is part of the material described by Kirkpatrick (18880 : 14) as Lepralia Poissonti, Audouin. I have also examined the specimen described by Levinsen (1909 : 268) as Crepidacantha poissoni var., which was kindly lent by Dr. P. L. Kramp of Universitetets Zoologiske Museum, Copenhagen, and this proves to be a small fragment, also encrusting Fucus, which was sent to Levinsen 1 See Postscript for further information about this species. 258 ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN by Miss Jelly in 890. It is unquestionably from the same material, if not from the same colony, as the holotype. This is also the species described by Waters (1889 : 14) from Green Point, New South Wales, in which the ovicells are “ immersed, showing, however, a round ovicellular area on a level with the wall of the zooecium.’’ Canu & Bassler (1929 : 410) thought that this probably belonged to another genus. This species is closely related to C. solea Canu & Bassler, but differs from it in having a large, flattened ovicell and eight stout marginal spines, the spines of C. solea being slender and 12 to 14 in number. In C. kirkpatricki the surface of the zoarium is fairly even and the zooecia are not usually separated by furrows but are indistinctly merged with one another. The proximal lip of the orifice is a distinct median plate inclined to turn upwards and outwards from the orifice, the lateral indentations appearing as small sub-circular openings. Marcus (1938 : 231), in proposing the new name C. /evinsent for Levinsen’s variety of C. poissonit (t909 : 268) and for his own material from St. Helena, stated that he had not seen the Port Phillip Heads material. As shown above, the St. Helena species is most probably C. longiseta Canu & Bassler. DISTRIBUTION. Recent: Port Phillip Heads, Victoria; Green Point, Port Jackson, N.S.W. (Waters). M. Crepidacantha odontostoma (Reuss) Lepralia odontostoma Reuss, 1874, Denkschr. Akad. Wiss. Wien. 33 : 156, pl. 4, fig. 8. Crepidacantha (Lepralia) odontostoma, Canu & Bassler, 1930, Proc. U.S. nat. Mus. 76 (13) : 32. REMARKS. Without seeing Reuss’s material, which comes from the Miocene of “ Rauchstallbrunngraben bei Baden,” it is difficult to decide on the generic affinities of this species. The orifice has a straight or slightly concave proximal lip and the paired heterozooecia are placed at the level of the opercular condyles. If this is a species of Crepidacantha and if the preservation is as good as Reuss’s figure suggests, then one would expect to see some traces of marginal areolae. The absence of ovicells makes identification even more difficult. DOUBTFUL REFERENCES TO SPECIES OF Crepidacantha Thornely has recorded Lepralia poissonii Audouin from Ceylon (1905 : 119) and the Andamans! (1907: 190). It is impossible to identify the species, but, from comparison with the world distribution map (Fig. 4), it is possible that C. longiseta Canu & Bassler or C. cyinispina Levinsen sp. is represented here. Waters (1914 : 832, 856) recorded Lepralia poissonit Audouin from the Atlantic [2 = C. potssonti var. teres], Indian Ocean [? = Thornely’s record, probably C. longiseta], Australia, and Japan. I did not see Waters’s material from these localities in the collections at Manchester Museum, and, therefore, I cannot suggest to which species the Australian and Japanese records refer. 1 A specimen (B.M., 84.3.25.1, Wilmer Collection) from the Andaman Islands, labelled “‘ Lepralia poissonti Aud.” by Kirkpatrick, is not a Crepidacantha. ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN 259 pee | el [Shale oS faa Fic. 4. Recent distribution of species of Crepidacantha. Species are indicated by the index letters used throughout the text. 260 ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN CONCLUSIONS The known species of Crepidacantha Levinsen are 11 or 12 in number and, until recently, had been recorded only from tropical and sub-tropical waters. They have recently been found, however, as far south as Kaka Point, Southland, New Zealand (169° 50’ E., 46° 25’ S.), on a coast-line swept by a northerly drift. This is in accordance with their presence in rocks of Nukumaruan [Middle Pliocene] age in Hawkes Bay. This region (about 40° S.) has been shown by Fleming (1944) from molluscan evidence to be the northern limit of subantarctic waters in Nukumaruan times. Species of Crepidacantha are also plentiful in the Waiauan [Middle Miocene} rocks of North Canterbury, New Zealand, rocks deposited when New Zealand was experiencing a tropical climate. In the northern hemisphere Crepidacantha has not been recorded north of Madeira (33° N.), though Waters’s record from Japan (1914 : 832), if correct, may extend this range. The first species described, Flustra poissonii Audouin, is not at all cosmopolitan in distribution, and, in fact, appears to be restricted to the Eastern Mediterranean or Red Sea area.1 The accompanying map (Fig. 4) shows the known distribution of the various species. POSMSG Rr Par Dr. Anna B. Hastings, in litt., writes : “In his forthcoming, posthumous ‘Siboga’ Report, Sir Sidney Harmer remarks that a revision of the genus Crepidacantha is urgently needed. “ Pending such a revision he has recognized only two species in the Indo-Malayan, region—C. poissonii and C. crinispina—and has given tentative synonymies referring much of the previously recorded material to one or other of these species. You have now provided the necessary revision discriminating more species in this material and superseding (although ante-dating) Harmer’s synonymies, which, but for the war of 1939-1945, might have been published ten years ago. “Certain material which, as you mention, was not available when you did your work was considered by Harmer and has since been presented by him to the British Museum (Natural History). I have examined the specimens and you may like to include some notes about them : “1. Two specimens from Ghardaqa on the Egyptian coast of the Red Sea (Crossland, 4.P on slides 4.G1 and 4.G® of Avachnopusia spathulata). The material ‘agrees closely with Savigny’s fig. 52 (Pl. X) except in not showing radial markings on the ovicell ’ (Harmer MS.). “In the absence of any record of Crepidacantha from the Mediterranean, the presence on the Red Sea coast of Egypt of a species agreeing closely with Savigny’s figure strongly suggests that the material of Savigny & Audouin came from the Red Sea. I am prepared to accept the two specimens from Ghardaqa as repre- senting true C. potssonit. 1 See Postscript. 2 The information given in this Postscript is incorporated in the distribution-map. ; ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN 261 “C. poissonii (as represented by the Ghardaqa specimens) and C. solea are, as you have suggested, very closely related. They may perhaps be distinguished by the more proximal position of the heterozooecia in C. solea (which are directed laterally inwards in both species), but the Ghardaqa material shows some variation in this character. The ovicells of both show the proximo-lateral calcareous layer described by you in C. poissonii var. teres, where it is sometimes conspicuous for its opaque whiteness. They also agree with that variety in having a porous, transverse ridge on the ovicell. The pores are, however, more or less obscured by a membranous layer (membranous part of the ectooecium). The appearance depends very much on the direction of the light. Illuminated from the proximal end the membrane covering the frontal surface shines and, being stretched in drying, may give an impression of some radial striation, and the pores are hidden. Illuminated from the distal end the membranous layer is inconspicuous and the transverse row of pores may be observed. Such effects as this probably account for the absence of any appearance of the row of pores on the entooecium in your figures of C. crini- spina var. parvipora both here (Fig. rp) and in the Catalogue (Brown, 1952 : 361, text-fig. 284) and of typical C. crinispina (Catalogue, text-fig. 283). “2. Thornely’s Ceylon*specimen (Thornely Coll., 1906, no. 46). This specimen agrees very closely with the Ghardaqa specimens of C. poissonit. It has one ovicell, which shows the membrane, the row of pores, and the proximo-lateral calcareous ectooecium as described above. “3, Thornely’s Andaman specimen (Thornely Coll., 1935, no. 118). This belongs to C. crinispina as limited by you. “4. A specimen from Victoria (Cambridge Museum, E. C. Jelly, May 24, 1895) is also of interest. Harmer (MS.) mentions it as probably representing a new species. It has no ovicells, but in other respects agrees with the type-specimen of C. kirkpatricki. Two parallel, longitudinal, oval bosses on the opercula, mentioned by Harmer, are also present in your type-specimen. They are blister-like prominences on the outer surface of the operculum, and are conspicuous when the specimens (which are dry) are examined under a binocular microscope.” REFERENCES Aupoutn, V., & Savicny, M. J. C. L. de. 1826. Description de l’Egypte. Histoire Naturelle, T, 4, Explication sommairve des Planches dont les dessins ont été fournis pay J.-C. Savigny, pp. 225-244, pls. r-14. Paris. 1828. Description de l’Egypte. 2nd. Ed., XXIII, Histoive Naturelle, Zoologie, Explication sommaire des Planches de Polypes del’ Egypte et de la Syrie, Publiées par Jules- Césav Savigny. ... , pp. 40-78. Paris. Basster, R. S. 1935. Fossilium Catalogus I: Animalia, 67: Bryozoa. ’s-Gravenhage. Brown, D. A. 1952. The Tertiary Cheilostomatous Polyzoa of New Zealand. Cat. Brit. Mus. (Nat. Hist.), pp. 1-405. — 1954. Polyzoa from a Submerged Limestone off the Three Kings Islands, New Zealand. Ann. Mag. nat. Hist. (12) 7: 415-437- Busk, G. 1854. Catalogue of Marine Polyzoa in the British Museum, Part II, pp. 55-120, pls. 69-124. London. 262 ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN Catvet, L. 1907. Bryozoaires in Expéd. Sct. ‘‘ Tvavaillewr’’ et ‘‘ Talisman,” pp. 355-495, pls. 26-30. Paris. Canu, F., & Basster, R.S. 1917. A Synopsis of American Early Tertiary Cheilostome Bryozoa. Bull. U.S. nat. Mus. 96 : 1-87, pls. 1-6. 1920. North American Early Tertiary Bryozoa. Ibid., 106: 1~xx, 1-879, pls. 1-162. 1923. North American Later Tertiary and Quaternary Bryozoa. Ibid. 125: i-vii, 1-302, pls. 1-47. 1927. Classification of the Cheilostomatous Bryozoa. Proc. U.S. nat. Mus. 69 (14) : 1-42, pl. 1. 1928a. Fossil and Recent Bryozoa of the Gulf of Mexico Region. Ibid. 72 (14): 1-199, pls. 1-34. 1928b. Bryozoaires des Iles Hawai. Bull. Soc. Sci. nat. méd. Seine-et-Oise, (2) 8, 7, suppl. : 33-66, pls. 6-11. 1929. Bryozoa of the Philippine Region. Bull. U.S. nat. Mus. 100 (9) : 1-685, pls. I-94. 1930. The Bryozoan Fauna of the Galapagos Islands. Proc. U.S. nat. Mus. 76 (13) : 1-78, pls. 1-14. —— & Lecorntre, G. 1930. Les Bryozoaires cheilostomes des faluns de Touraine et d’Anjou. Mém. Soc. géol. Fr. (N.S.) 4: 83-130, pls. 16-25. Epwarps, F. W. 1910. Bryozoa in Zool. Rec. (1909) 46 : 14-26. Fremine, C. A. 1944. Molluscan Evidence of Pliocene Climatic Change in New Zealand. Trans. roy. Soc. N.Z, 74 : 207-220. Hincxs, T. 1880. Contributions towards a General History of the Marine Polyzoa. Part I. Ann. Mag. nat. Hist. (5) 6 : 69-92, pls. 9-11. 1881. Contributions towards a General History of the Marine Polyzoa. Ibid. (5) 8 : 122— 136, pls. 1-5. 1885. Contributions towards a General History of the Marine Polyzoa. Ibid. (5) 15 : 244- 257, pls. 7-9. —— 1891. Contributions towards a General History of the Marine Polyzoa.—Appendix. Ibid. (6) 8 : 86-93. 1895. Index to ‘‘ Marine Polyzoa: Contributions towards a General History,’ pp. i-vi. London. Juiiien, J., & Catvet, L. 1903. Bryozoaires provenant des campagnes de 1’ Hirondelle. Result. Camp. sci. Monaco, 23 : 1-188, pls. 1-18. KirkKPATRICK, R. 1888a. Polyzoa of Mauritius. Ann. Mag. nat. Hist. (6) 1: 72-85, pls. 7-10. —— 1888). Polyzoa from Port Phillip. Ibid. (6) 2: 12-21, pl. 2. —— 1890. Report upon the Hydrozoa and Polyzoa collected ... by H.M.S. ‘‘ Rambler.” Ibid. (6) 5: 11-24, pls. 3-5. Levinsen, G. M. R. 1909. Morphological and Systematic Studies on the Chetlostomatous Bryozoa, vii + 431 pp., 27 pls. Copenhagen. Livinestone, A. A. 1929. Bryozoa Cheilostomata from New Zealand. Vidensk. Medd. dansk naturh. Foren. Kbh. 87: 45-104, pls. 1, 2. MacGILiivray, P.H. 1883. Descriptions of New, or Little-Known, Polyzoa. Part II. Trans. voy. Soc. Vict. 19 : 130-138, pls. 1-3. 1887. A Catalogue of the Marine Polyzoa of Victoria. Ibid. 23 : 187-224. Marcus, E. 1938. Bryozoen von St. Helena. Vidensk. Medd. dansk naturh. Foren. Kbh. 101 ; 183-252. Norman, A. M. 1909. The Polyzoa of Madeira and neighbouring Islands. J. Linn. Soc. Lond. (Zool.) 30 : 275-314, pls. 33-42. pD’OrpiGNy, A. D. 1851-1854. Paléontologie Francaise. Terrains Crétacés. 5. Bryozoaires : I-I192, pls. 600-800. Paris. ON THE POLYZOAN GENUS CREPIDACANTHA LEVINSEN 263 OspurN, R. C. 1940. Bryozoa of Porto Rico with a Résumé of the West Indian Bryozoan Fauna. Sct. Surv. Porto Rico and Virgin Islands, 16 (3) : 321-486, pls. 1-9. 1952. Bryozoa of the Pacific Coast of America. Part 2. Cheilostomata—Ascophora. Allan Hancock Pacif. Exped. 14 (2) : 269-611, pls. 30-64. Puiwiprs, E. G. 1899. Report on the Polyzoa collected by Dr. Willey from the Loyalty Isles, New Guinea and New Britain. A. Willey’s Zoological Results, 4: 439-450, pls. 42, 3. Reuss, A. E. 1874. Die fossilen Bryozoen des Gsterreichisch-ungarischen Miocans. I. Abtheilung. Denkschr. Akad. Wiss. Wien, 33 : 141-190, pls. I-12. Smitt, F. A. 1873. Floridan Bryozoa. Part II. K. svenska Vetensk.-Akad. Handl. 11 (4): 1-83, pls. 1-13. THORNELY, L. R. 1905. Report on the Polyzoa ... Rept. Govt. Ceylon Pearl Oyster Fisheries Gulf Manaar, Roy. Soc. London, 4 Suppl. Rept. 26 : 107-130, pl. 1. —— 1907. Report on the Marine Polyzoa in the collection of the Indian Museum. Rec. Indian Mus. 1 : 179-196. 1g12. The Marine Polyzoa of the Indian Ocean, from H.M.S. “ Sealark.’”” Tyvans. Linn. Soc. Lond. (Zool.) (2) 15 : 137-157, pl. 8. VicNEAux, M. 1949. Révision des Bryozoaires néogénes du Bassin d’Aquitaine et essai de classification. Mém. Soc. géol. Fr. 60: 1-155, pls. I-11. Waters, A. W. 1887. On Tertiary Chilostomatous Bryozoa from New Zealand. Quart. J. geol. Soc. Lond. 43 : 40-72, pls. 6-8. 1889. Bryozoa from New South Wales. Ann. Mag. nat. Hist. (6) 4: 1-24, pls. 1-3. 1899. Bryozoa from Madeira. J. R. micy. Soc. 1899 : 6-16, pl. 3. 1914. The Marine Fauna of British East Africa and Zanzibar, ... Bryozoa—Cyclo- stomata, Ctenostomata and Endoprocta. Proc. zool. Soc. Lond. 1914 : 831-858, pls. 1-4. 9 OVS ba PRESENT ED ; BY PRINTED IN GREAT BRITAIN LIMITED, SON, AND ADLARD DORKING BARTHOLOMEW PRESS, GEN US. PTILINOPUS (AVES, COLUMBAE) =9 NOV1954 PRESENTED A. J. CAIN ee BULLETIN OF | SH MUSEUM (NATURAL HISTORY) pi | Vol. 2 No. 8 =... “LONDON: 1954 _ SUBDIVISIONS OF THE GENUS PTILINOPUS 2 (AVES, COLUMBAE) ARTHUR JAMES CAIN (Department of Zoology and Comparative Anatomy, University Museum, Oxford) Pp. 265-284 ; 2 Plates BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 2 No. 8 LONDON : 1954 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL . HISTORY). «stituted in 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical Series. Parts appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be compieted within one calendar year. This paper is Vol. 2, No. 8 of the Zoological series. PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued October, 1954 Price Eight Shillings ——. = ae aiid SUBDIVISIONS OF THE GENUS PTILINOPUS (AVES, COLUMBAE) By ARTHUR JAMES CAIN CONTENTS Page INTRODUCTION . : F . : : ‘ ¢ : : ae2zG7, PRINCIPLES ; ‘ : : . : ‘ ZS (z) Recognition of natural groups : : : : : : -, 268 @)iSve : é : : : . 269 (3) Keys and mapa Sesstierifiene Sa : : : ; . 269 (4) Use of species-groups and subgenera 2 : > 6 : 270 (5) The superspecies . c : : - - c 70 CLASSIFICATION . = 9 5 6 : 5 : a be iit (x) Brief list : 2 2 3 : E - o Ae (2) Detailed classification : : : 2 c 7 c e273} Notes ON THE CLASSIFICATION . : : J ‘ 7. : 5) ye (1) Generic and subgeneric limits . : - : 3 2 3 Dis (2) Leucotreron and Ramphiculus . ; 5 : : : =) 280 (3) Ptilinopus . : : : : : : a Afaht (4) The purpuratus eens: etoup : c ‘ é : é 7 282 SUMMARY . 0 : : : ; : : ; : : 5 Berl REFERENCES. : ; : ; 5 “ ; : : 204! SYNOPSIS Previous arrangements of the species of Prilinopus result in unnatural groupings. When one abandons the attempt to provide diagnoses for all groups, and gives due weight to the facts of geographical distribution, a far more natural classification can be constructed. INTRODUCTION Tue genus Ptilinopus contains a large number of brightly coloured fruit-pigeons distributed from the Malay Peninsula, Philippines and Marianas to New South Wales, the Marquesas, and Henderson Island. Even within obviously natural groups of species of this genus so much interspecific variation occurs that it is almost impossible to find diagnostic characters. The last comprehensive review with diagnoses is that by Salvadori (1893) which, as was usual at that period, ascribes exaggerated importance to so-called anatomical characters. Some species have been reviewed by Rensch (1929), Mayr (1940), Ripley and Birckhead (1942) and Amadon (1943), but no general survey has appeared, except that Peters (1938) reviewed the limits of the genera of fruit-pigeons and gave diagnoses for those ZOOL, I, 8, 16§ 268 SUBDIVISIONS OF THE GENUS PTILINOPUS genera he recognized, together with a very brief survey of the interrelations of species within each genus. The checklist given by Peters (1937) is good for sub- specific ranges and synonymy, but the arrangement of the species is unfortunate, several closely related forms being separated widely. Also, no indication is given of the limits of superspecies or species-groups. The purpose of this paper is to present a classification of these pigeons, based on a study of the collections in the British Museum (Natural History). I am deeply indebted to Mr. J. D. Macdonald, who has given me every facility in examining the specimens, and to Mr. R. E. Moreau for advice and criticism. PRINCIPLES (1) Recognition of natural groups If each of the members of a group of species is more closely related to some other members than to any species outside it, the group is natural; but it may well happen that every character that is confined to the group and therefore might be proposed as diagnostic is lost or modified in one or more species. If those species are clearly related by all their other characters to the rest of the group, they cannot be excluded from it. This is the state of affairs found in the genus Ptilinopus. Consequently, the descriptions given below must be understood as indicating general characteristics, not as rigidly defining each subdivision. As Peters (1938) remarked of the subgenus Ptilinopus, it is very difficult to draw up a diagnosis that applies to all the species. I would add that it is unnecessary. When a natural group can be recognized but not diagnosed as such, it is sufficiently indicated by a description and a list of the contained species, as in the present paper. When two species are clearly very closely allied but differ in one conspicuous character, it is obvious that in respect to these two species this character is only of specific importance. In another group, its distribution may coincide with that of so many other characters that it can be cited as characteristic of a genus, family, or group of even higher rank. The test of taxonomic importance of any character in a particular group is its distribution in that group with respect to all other characters. This statement is true whatever the nature of the character—anatomical, physio- logical, embryological, genetical or ecological. The remark of Manuel (1936b) that “ The subfamily Ptilopodinae is an artificial group for the reason that there are no trenchant structural characters peculiar to it’ is consequently unacceptable. In some natural groups of birds the beak, feet, or structure of the feathers may show extraordinary variability while the colour pattern remains very constant ; in others, the reverse is true. Consequently there is no need to assume @ priori that in any group of birds structure is always more important taxonomically than colour- pattern. This was very nearly the assumption made by Salvadori (1893) as of course by most authors of that time. Consequently his keys almost always begin with an “ anatomical ’’ character (emargination of the first primary, length of the tail and wing, or strength of the beak). Unfortunately these characters are not invariably of the highest importance in Ptilinopus and its allies, and Peters (1938), SUBDIVISIONS OF THE GENUS PTILINOPUS 269 although he implies (p. 376) that external structural characters are a sine qua non for erecting a subfamily (which may be true in practice but cannot be made into an absolute rule), remarks of Ptilinopus and its allies that ‘“‘ On the whole, colour in this group seems just as important, if not more so than structure.’’ A survey of all the species, originally made without reference to Peter’s paper, confirms his opinion. (2) Size The importance of any character, as was shown above, is determined only by its distribution in any particular group with respect to all other characters, and no one class of characters can be regarded on principle as always of greater taxonomic importance than any other. Equally, no one class of characters can be considered on principle as always useless. Size tends to vary greatly in birds both between and within species, and is therefore not usually of much taxonomic value in groups above the level of the species. However, in the genus Ptilinopus size tends to remain comparatively constant in the species of some groups. The subgenera Leucotreron and Ramphi- culus contain species which, for the genus, are rather or very large (wing-length 150-170 mm.). In the subgenus Ptilinopus all the members of the ornatus species- group ((10)-(14)) are rather large for this subgenus (wing-length 135-160 mm.), while those of the purpuratus species-group ((15)-(30)) with a solitary exception (Pt. huttoni (26)) are rather small (wing-length 95-145 mm.). Size is therefore included as a character in the descriptions of the groups. Within the genus as a whole wing-lengths of 90-139 mm. are considered small, 140-159 medium, and 160-170 large. (3) Keys and natural classifications Keys are constructed for the rapid identification of specimens. The characters employed are therefore those that are readily discernible, present in the largest possible number of individuals (not confined, for example, to a fleeting stage of development), least variable in their expression, and most easily defined verbally. “Natural” classifications, on the other hand are constructed to display the general affinities and presumed evolutionary interrelationships of different forms. For this purpose characters totally different from those used in a key to the same group may be chosen, since the conspicuousness of a character bears no direct relationship to its taxonomic and phylogenetic importance. Yet monographs are still often burdened with “keys” which are designed to act both as guides to identification and as succinct accounts of a natural classification. In such “ keys ” it is common to find not only that the characters employed are often of little practical use for identification, but also that the necessity for defining each group by one or two characters may actually prevent the recognition of such natural groups as those described immediately above, which have no diagnostic characters. The Teason why such ‘‘ keys” are constructed is, of course, that by using them a new and undescribed form can usually be put into its natural group, whereas in a key made solely for identification it may very well be associated with widely differing 270 SUBDIVISIONS OF THE GENUS PTILINOPUS species with which it happens to have some unimportant but conspicuous character in common. (For example, the character “ bifid breast feathers ’’ would bring together Pt. porphyrea (s.g. Leucotreron) and Ducula goliath, and most of the purpuratus species-group of Ptilinopus.) Nevertheless, the practice of combining both functions in the same key is not good. Together with the assumption that structure is always more important than colour, it was responsible for some of Salvadori’s least successful groupings in Ptilinopus, as can be readily verified by anyone who will work through his classification. However, it is a striking tribute to Salvadori’s genius that although he worked under these unnecessary limitations, his classification was by far the best produced until then, and most of his groups can still be regarded as natural. (4) Use of spectes-groups and subgenera Mayr (1942 : 290) has pointed out the great taxonomic advantages of the species- group. It is an informal natural group of closely related species, below the rank of a subgenus, the recognition of which involves no complication of the scientific names of the species contained in it. Moreover, having (unlike the subgenus) no formal name, it is not readily raised to the rank of a genus with a consequent change of the generic name throughout its component forms. In addition, in Ptilinopus, its use permits the recognition of natural groups of species without a general promotion of the subgenera to genera, and the genus perhaps to a subfamily. By using species-groups, the generic name can be retained throughout (which indicates the close relationship of the included species) while unmanageably large groups of species can nevertheless be broken up. As the species-group is burdened by no formal rules there is no necessity to use the oldest valid name within a group as the name of that group. Consequently the group containing as its oldest named form the very atypical Pt. tannensis (14) can be called the ornatus group after a much more typical member (12). I have not changed the name of the purpuratus species-group (Ripley & Birckhead, 1942) because one should avoid nomenclatural disturbances wherever possible, and Pt. purpuratus (27), although a very simplified form, is certainly a member of the species-group that bears its name. Although the subgenus has the disadvantage that it lengthens the names of its component forms in full citations, and is a standing temptation to reviewers to raise it to generic rank, it is nevertheless useful when, as in the genus Péilinopus, many species-groups must be recognized but they fall clearly into two or more groups within the genus. Consequently it is used here to distinguish the three very natural groups which the genus Ptilinopus comprises. In smaller or less heterogeneous genera the species-group should be sufficient, without the use of subgenera. (5) The superspecies The superspecies (Mayr, 1942: 169) is a monophyletic group of forms which replace each other geographically but are too diverse for all to be ranked as sub- SUBDIVISIONS OF THE GENUS PTILINOPUS 275 species of a single species. The limit of permissible diversity cannot of course be specified, since in some groups good species are very similar, in others extremely diverse, so that in the former the diversity permissible among subspecies of a single species will be much less than in the latter, and forms with striking characters will be unhesitatingly separated as distinct species. Mayr (1942) gives examples of monophyletic arrays of geographical representatives in which some forms have even been segregated into a separate genus. Consequently a single superspecies may include only a few species of a single species-group, as in the occipitalis species- group ((4)-(7)), or all of them as in the lechlancheri species-group ((8)—(9)), or all of one species-group and one or more of another, as, for example, in the genus Halcyon (Mayr, 1942). Consequently, the superspecies cannot strictly be used as a rank in the taxonomic hierarchy. However, when a series of geographically replacing forms is very heterogeneous there is usually some doubt as to whether it is really monophyletic, and it would be unwise, therefore, to list it as a superspecies. If there is no doubt, then most workers would consider that, since geographical variation is considerable in this series, the limits of subgeneric and generic variation in the natural group containing it should be wide. Consequently it would be given a low rank in the hierarchy. As a result it is rare to find a superspecies that transgresses the limits of a species-group, and in practice the superspecies can be regarded as a rank between the species and the species-group. The hierarchy of ranks used in the following classification is therefore : Genus, Subgenus, Species-group, Subgroup (of species-group), Superspecies, Species. CLASSIFICATION The proposed classification is illustrated by the diagrams (Figs. 1 and 2) in which are represented all the species recognized by me. These diagrams are intended to demonstrate only the plumage pattern of each species. Relative size is not given and in each species the most highly ornamented subspecies and sex are shown (except in Pé. rivoli (32) and solomonensis (31) where there is very confusing con- vergence). Each species is given the same number in the diagrams as in the classi- fication. Cross-hatching is used to represent shades of red, from orange to purple. All other colours are represented by stippling of appropriate darkness. Except for Pt. merrilli (5), specimens of all species and nearly all subspecies have been examined. The diagram of P?. merrilli is based on the coloured plate of P. m. faustinoi given by Manuel (19362), and on the descriptions by Delacour and Mayr (1946) and McGregor (t916). Full references to original descriptions of all forms will be found in Peters (1937) and Salvadori (1893). 272 SUBDIVISIONS OF THE GENUS PTILINOPUS (x) Brief list Genus Ptilinopus Subgenus Leucotreron Subgenus Ramphiculus (a) Occipitalis species-group (i) Marchei subgroup (ii) Occipitalis subgroup (b) Leclancheri species-group Subgenus Ptilinopus (a) Ornatus species-group (i) Wallacii subgroup (ii) Ornatus subgroup (iii) Tannensis subgroup (b) Purpuratus species-group (i) Superbus subgroup (ii) Purpuratus subgroup Pt. cincta superspecies Pi. Pt. Pi. PE: Pi. porphyrea (x) dohertyt (2) cincta (3) marchet (4) merrill (5) Pt. occipitalis superspecies Pt. occipitalis (6) Pt. fischeri (7) Pt. leclancheri swperspecies Pt. leclancheri (8) Pt. subgularis (9) Pt. wallacw (x0) Pt. auvantitfrons (11) IPG Pt JEh; . ornatus (12) . perlatus (13) . tannensts (14) Pt. superbus swperspecies Pt Pi Pi. Pt. IPE. Pi. Pt. Pt. Piz Pt . superbus (15) . perousit (16) monacha (17) coronulatus (18) pulchellus (19) regina (20) voseicapilla (21) greyut (22) vichardsit (23) . porphyraceus (24) SUBDIVISIONS OF THE GENUS PTILINOPUS 273 Pt. rarotongensis (25) Pt. huttont (26) Pt. purpuratus (27) Pt. insularis (28) Pt. mercieri (29) Pt. dupetithouarsu (30) (c) Viridis species-group (i) Rivoli subgroup Pt. solomonensts (31) Pt. rivolt (32) (ii) Viridis subgroup Pt. viridis (33) (d) Hyogastra species-group (i) Iozonus subgroup Pt. 1oz0nus superspecies Pt. tozonus (34) Pt. insolitus (35) (ii) Hyogastra subgroup Pt. hyogastra superspecies Pt. hyogastra (36) Pt. granulifrons (37) Pt. naina (38) Dubious member of this super- species Pt. melanospila (39) (e) Jambu species-group Pt. jambu (40) (f) Lutovirens species-group Pt. luteovirens superspecies Pt. victor (41) Pt. luteovirens (42) Pt. layardi (43) (2) Detailed classification Genus Ptilinopus Swainson, 1825 Medium to small fruit-pigeons (wing-length 170 to 90 mm.) almost always with a plumage pattern including well-defined patches of colour forming conspicuous ornaments on the head and underparts, without a conspicuous yellow stripe on the wing, and with the first primary usually more or less clearly emarginate (falcate) at the distal end. Subgenus Leucotreron Bonaparte, 1854 (Fig. 1). Rather large Ptilinopids (wing-length about 160 mm.), with the head, neck, mantle and breast concolorous or simply patterned in bright colours and sharply marked off 274 SUBDIVISIONS OF THE GENUS PTILINOPUS from the rest of the body by a narrow pale line followed on the ventral surface by a dark transverse bar. No tendency to ornamentation of the wings. First primary emarginate. Greater and Lesser Sunda Isles, Australia. (1) Pt. porphyrea (Temminck, 1823). Sumatra, Java, Bali. (2) Pt. dohertyt (Rothschild, 1896). Sumba. (3) Pt. cincta (Temminck, 1810). Lesser Sunda Isles except Sumba, from Bali to Teun, Damar and Babar, and Northern Territory of Australia. These three species are very closely allied, and if the very small overlap of Pt. porphyrea and cincta on Bali may be disregarded, they can be grouped as the Pz. cincta superspecies. Subgenus Ramphiculus Bonaparte, 1854 (Fig. 1) Medium-sized to large Ptilinopids (wing-length 150 to 170 mm.) with the most deeply coloured patches of the head-ornamentation (excepting the chin-stripe) lateral, sometimes meeting on the hind neck to form a ring. (A red cap occurs in Pt. mervilli faustinoi.) No tendency to ornamentation of the wings, nor to a pale line bounding the whole of the anterior parts as in Leucotreron. First primary emarginate, usually very obviously. Philippines and Celebes. (a) Occipitalis species-group Rather highly ornamented forms (plus one with a simplified pattern, Pt. merrill2), with a complex pattern on the head in which a chin stripe is not a well-marked feature, with breast and belly usually of different colours separated by a dark transverse abdominal line or band, and with spotted under tail coverts. (i) Marchei subgroup With a red cap and black auriculars, or only a cap, or no ornamentation on the head. Outer web of the secondaries composed of short widely separated red barbs. (4) Pt. marchei (Oustalet, 1880). Philippines ; Luzon and Polillo. (5) Pt. merrilli (McGregor, 1916). Philippines ; Luzon and Polillo. (ii) Occipitalis subgroup. With a red or black band joining the red auriculars across the nape. One super- species. (6) Pt. occipitalis G. R. Gray, 1844. Philippines. (7) Pi. fischeri Briiggemann, 1876. Celebes. (0) Leclancheri species-group Rather plain species with a well marked dark chin-stripe, and plain (unspotted) brown under tail coverts. One superspecies. (8) Pt. leclanchert Bonaparte, 1855. Philippines. (9) Pt. subgularis Meyer and Wiglesworth, 1896. Celebes, Peling, Banggai and Sula Mangoli. SUBDIVISIONS OF THE GENUS PTILINOPUS 275 Subgenus Piilinopus Small to medium Ptilinopids (wing-length 90 to 160 mm. only in Pt. huttont (26) 170 mm.), with strong tendencies to ornamentation of the wing coverts and scapulars and of the underparts, either a pectoral or abdominal patch or both being almost always present. The most deeply coloured patches on the head (excepting the chin stripe) are dorsal and median, forming acap. First primary often clearly emarginate, sometimes obscurely so or merely tapering to the tip. Principally the New Guinea region and islands to the eastwards. (a) Ornatus species-group Forms rather large for this subgenus (wing length 135-160 mm.), highly orna- mented (except for Pt. tannensis (14), an isolated form with a simplified pattern). Wing coverts spotted with pink, grey or white. There is a strong tendency to grey or olive yellow on the neck and breast. Abdomen plain green or with only a pale abdominal patch. Abdominal spot or bar (dividing the abdomen from the breast) always absent. Under tail coverts spotted. First primary not or only very slightly emarginate. (i) Wallacii subgroup With a red or orange cap, grey-spotted wing coverts, white chin and cheeks, and grey breast. (10) Pt. wallacii (G. R. Gray, 1858). Babar, Timorlaut, Kei and Aru Isles. (11) Pt. aurantiifrons (G. R. Gray, 1858). New Guinea, Western Papuan Islands, and Aru Isles. (ii) Ornatus subgroup With an olive-yellow cap (red in one subspecies), a grey chin, and olive-yellow breast. Wing coverts spotted with grey or bright pink. (12) Pt. ornatus (Schlegel, 1871). New Guinea. (13) Pt. perlatus (Temminck, 1835). New Guinea. (iii) Tannensis subgroup Head olive-yellow, most of the rest of the body plain green. Scapular spots white. Related to the ornatus subgroup. (14) Pt. tannensis (Latham, 1790). New Hebrides and Banks Islands. _ (b) Purpuratus species-group Small to medium species (wing-length 95 to 145, 170 in Pt. huttoni (26)) witha bright red, purple or blue cap bordered behind with yellow (or with vestiges of such a cap), and with clearly or obscurely bifid breast-feathers. Scapular and wing- covert spots present, dark blue, pink, pale purple, or emerald green, often not clearly marked. Abdomen ornamented with a large patch and a darker transverse bar or spot, reduced or absent in a few forms. Under tail coverts plain red, orange or 276 SUBDIVISIONS OF THE GENUS PTILINOPUS yellow (spotted only in Pt. superbus). First primary emarginate, almost always very clearly. (i) Superbus subgroup Lower neck and upper back with a more or less extensive bright brownish red or dark red band. Wide transverse abdominal band present. One superspecies. (15) Pt. swperbus (Temminck, 1810). Celebes and the Sulu Archipelago to the Solomons and eastern Australia. (16) Pt. perousit Peale, 1848. Fiji, Tonga and Samoa. (ii) Purpuratus subgroup With no red on the lower neck and upper back, and no distinct humeral patch on the wing. With a narrow abdominal band transversely elongated (Pt. pulchellus (19)) or, much more frequently, a rounded or longitudinally elongated abdominal spot. One superspecies with two doublets (see p. 283). (17) Pt. monacha (Temminck, 1824). North Moluccas (Halmahera, Ternate Batjan). (18) Pt. coronulatus (G. R. Gray, 1858). New Guinea, Japen, Salawati and Aru Isles. (19) Pt. pulchellus (Temminck, 1835). New Guinea and Western Papuan Islands. (20) Pt. regina Swainson, 1825. Eastern and northern Australia, eastern Lesser Sunda Isles, Banda and Kei Isles. (21) Pt. roseicapilla Lesson, 1831. Marianas. (22) Pt. greyt Bonaparte, 1857. New Caledonia to Santa Cruz Islands and Gower Island. (23) Pt. richardsii (Ramsay, 1882). Ugi, Santa Anna, and Rennell Island (Solomons). (24) Pt. porphyraceus (Temminck, 1821). Fiji, Tonga, Samoa, Caroline and Palau Islands. (25) Pt. varotongensis Hartlaub and Finsch, 1871. Rarotonga. (26) Pt. huttont (Finsch, 1874). Rapa Island. (27) Pt. purpuratus (Gmelin, 1789). Society Isles and Tuamotu or Lau Archi- pelago. (28) Pt. imsularis (North, 1908). Henderson Island. (29) Pt. merciert (Des Murs and Prévost, 1849). Nukuhiva and Hivaoa (Mar- quesas). (30) Pt. dupetithouarsit (Neboux, 1840). Marquesas, widespread. (c) Viridis species-group Forms medium sized for this subgenus (wing-length 115-135 mm.). Pectoral patch large, sharply defined and coloured white, yellow, or deep red. Abdominal patch reduced and dark purple, or absent. Wing-covert spots present, dark blue or grey. First primary not or only very slightly emarginate. SUBDIVISIONS OF THE GENUS PTILINOPUS 277 (i) Rivoli subgroup Wing spots dark blue. Abdominal patch present. Red or purple cap present, sometimes reduced to a pair of supraloral spots. Under tail coverts yellow. (31) Pt. solomonensis (G. R. Gray, 1870). Solomon Islands, Bismarck Archi- pelago and some islands in Geelvink Bay. (32) Pt. rivoli (Prévost, 1843). South Moluccas, Western Papuan Islands, north-west New Guinea, islands in Geelvink Bay, some islands off south-eastern New Guinea, Solomon Islands. (ii) Viridis subgroup Wing spots grey. Abdominal patch and red cap absent. Head shades of green and grey (rarely white). Size of pectoral patch varies greatly. Under tail coverts spotted. (33) Pt. viridis (Linnaeus, 1766). South Moluccas, Western Papuan Islands, north-west New Guinea, islands in Geelvink Bay, some islands off south-eastern New Guinea, Solomon Islands. (d) Hyogastra species group Forms medium-sized to small for this subgenus (wing-length go-130 mm.). Breast, neck and back plain green, unornamented. Abdomen with an orange or ‘Violet patch (absent in Pt. melanospila (39)). Head plain green or grey, unorna- mented or with a chin stripe and nuchal spot. First primary indistinctly emar- ginate or merely tapering. (i) Iozonus subgroup Head green with greyish ill-defined chin stripe. Abdominal patch large, orange. Grey patch on bend of wing, grey spots on wing-coverts. Under tail coverts spotted. One superspecies. (34) Pt. iozonus (G. R. Gray, 1858). New Guinea, Western Papuan Isles and Aru Isles. (35) Pt. insolitus (Schlegel, 1863). Bismarck Archipelago. (ii) Hyogastra subgroup Head grey (green in Pt. naina). Abdominal patch violet (absent in Pt. melano- spila). Under tail coverts yellow, or yellow grading to red, not spotted. One superspecies. (Possibly Pt. melanospila should be kept separated.) (36) Pt. hyogastra (Temminck, 1824). Halmahera and Batjan (North Moluccas). (37) Pt. granulifrons Hartert, 1898. Obi Major. (38) Pt. naina (Temminck, 1835). New Guinea and Western Papuan Isles. (39) Pt. melanospila (Salvadori, 1875). Philippines, Celebes, Java, Lesser Sunda Isles (Bali to Alor) and Ceram. 278 SUBDIVISIONS OF THE GENUS PTILINOPUS (e) Jambu species group Head red with black chin stripe. Upper parts green, underparts white with a pink flush on the upper breast and brown under tail coverts. First primary emar- ginate. (40) Pt. jambu (Gmelin, 1789). Malay Peninsula, Sumatra, Borneo and islands between. (f) Luteovirens species group Rather small forms (wing-length about 115 mm.). Head more or less olive- yellow, rest of body almost uniform green, orange or yellow, with no colour-orna- ments. Some contour feathers lax and hairy or bifid, or long and thickened. First primary not emarginate. One superspecies, confined to Fiji. (41) Pt. victor (Gould, 1872). Vanua Levu, Taviuni, Kio Rambi, Ngamea, Lauthala. (42) Pt. luteovirens (Hombron and Jacquinot, 1841). Viti Levu and nearby islands. (43) Pt. layardi (Elliot, 1878). Kandavu and Ono. NOTES ON THE CLASSIFICATION (1) Generic and subgeneric limits Peters (1938) divides the species considered in the present paper into three genera, Leucotreron (i.e., the subgenera Leucotreron and Ramphiculus), Ptilinopus (the sub- genus Ptilinopus without the luteovirens species-group), and Chrysoena (the luteo- virens species-group), which is divided into two subgenera. In this he differs from Salvadori only in promoting Leucotreron from a subgenus of Ptilinopus, and removing Pt. jambu (40) from it, in using no subgenera of Ptilinopus (Salvadori recognized twelve), and in accepting Wetmore’s subdivision of Chrysoena (1925 : 833). Amadon (1943) has shown conclusively that Chrysoena must be ranked as a single super- species (41-3), which requires no subdivision, of Ptilinopus. The separation of Leucotreyon is justifiable since it differs sharply from the other groups recognized as subgenera of Ptilinopus by Salvadori, but when these are ranked only as species- groups, there is no good reason why Leucotreron should be generically separate. Moreover, as the classification given above will show, Leucotreron itself requires division. A glance at the diagrams indicates that the major division of the genus is into three, not two, groups, which have very different colour patterns and distri- butions ; accordingly, all three are recognized here as subgenera. This arrangement is the more satisfactory since the colour pattern of the subgenus Leucotreron, as recognized here, is so like that seen in both Drepanoftila (as Peters recognized) and in some of the most ornamented species of Ducula, that a complete revision of all the fruit pigeons may possibly show that Leucotreron should be separated generically from Ramphiculus and Ptilinopus. SUBDIVISIONS OF THE GENUS PTILINOPUS 279 The characters by which both Peters and Salvadori define their group Leucotreron (i.e., Leucotreron plus Ramphiculus) are the relatively long tail and the absence of a sharply-defined red or violet cap. Peters adds that the tarsus is feathered for more than half its length, there are no spots on the scapulars, tertials or wing coverts, and a humeral patch is absent, and refers Pt. jambu (40) not to Leucotreron but to Ptilinopus. In all these forms, the length of the tail tends to be associated directly with body size, which can vary considerably within a single species-group. The feathering of the tarsus is variable, but it is true that it is on the whole more exten- sive in both Ramphiculus and Leucotreron than in Ptilinopus. All the other characters are negative. Consequently, what Ramphiculus and Leucotreron share is mainly medium to large size and the absence of the distinctive characters of Ptilinopus. On such grounds as these it is difficult to see why Ptilinopus and Leucotreron should not be combined instead, in opposition to Ramphiculus, an arrangement which has never yet been proposed ; but even a cursory inspection of the positive characters of Leucotreron and Ramphiculus shows that all these subgenera must be separated. The only species difficult to place is Pt. jambu (40), which shows an extraordinary mixture of characters. Because of its brown under tail coverts and black chin stripe Salvadori associated it with Pt. leclancheri (8) and subgularis (9), its emar- ginate first primary, lack of a cap, and rather long tail being sufficient to include it in his Leucotreron. Peters merely remarks that by his definitions it is a Ptilinopus (1938 : 378), but it is not easy to understand this remark, since his diagnosis of Ptilinopus is “‘ tail less than seven-tenths of wing, often less than six-tenths, usually a trifle over six-tenths ; tarsus never feathered for more than three-quarters of its length, seldom over one-half ; a sharply-defined red or violet cap ; spots on scapu- lars, wing-coverts or tertials; bend of wing often differently coloured from the back,” while for his Leucotreron he gives “ tail more than seven-tenths length of wing (usually more than 75 per cent.) ; tarsus feathered for more than half its length (usually from three-quarters to completely); no sharply-defined red or violet cap ; no spot on scapulars, wing coverts or tertials ; bend of wing concolour with back (no humeral patch).’’ In fact, the curious distribution of red on the head of Pt. jambu could be derived either by extension of the red cap (and sometimes malar spots) of some Ptilinopus or by reduction of the pattern found in either Leucotreron or Pt. marchei (4). The chin stripe, unicolorous under tail coverts, pale underside, and lack of wing-orna- ments are very reminiscent of Ramphiculus (the leclancheri species-group) but also, except for the pale underside, of Pt. melanospila (39) (Ptilinopus, hyogastra species- group). The distribution could be the result of a westward invasion from Celebes by either of these species-groups, or indeed of an extension from Borneo or Malaya by a geographical representative of Pt. porphyrea, producing a double invasion of Sumatra. Garrod (1874) states that Pt. jambu (40), Pt. peroustt mariae (purpuratus species- group (16)) and Pt. melanospila melanauchen agree in the structure of the gizzard and differ from Tveron calva. Cadow (1933) who corrects and extends Garrod’s observations, compared Pt. cincta (3), porphyrea (1) and doherty: (2) (all Leucotreron) and Pt. jambu (40) with species of Megaloprepia, Ducula, Treron, Columba, and 280 SUBDIVISIONS OF THE GENUS PTILINOPUS Didunculus. He concluded that three main types of gizzard could be distinguished in the fruit pigeons, and that Pt. jambu agreed in this respect with the sub-genus Leucotreron. It seems, therefore, that the structure of the gizzard is much the same within the genus Ptilinopus, in the subgenera Leucotreron and Ptilinopus, and in Pt. jambu, and consequently gives no information about the position of Pt. jambu. On the whole, the characters of Pt. jambu seem to me to suggest an association with Pt. melanospila (39) and thence with the rest of the subgenus Ptilinopus, but I have placed it in a species-group of this subgenus only with the greatest hesitation. It is most remarkable that the species situated geographically at the point of con- vergence of the ranges of the subgenera of Ptilinopus should show such a mixture of the characters of all three. (2) Leucotreron and Ramphiculus In discussing the species of these two subgenera Peters (1938 : 378) was misled by the bifid breast feathers of Pt. porphyrea (1) and the rich red on its head and breast into considering it most closely allied to the subgenus Ptilinopus, and to Pt. occipitalis (6) and Pt. marchei (4), the most ornamented members of Ramphiculus. Consequently he proposed an artificial arrangement beginning with Pt. porphyrea (1), followed by the occipitalis species-group (4-7), then the leclancheri species-group (8-9), and finally Pt. cincta (3) and doherty: (2) which he regarded as highly specialized. “The more I study cinctus and doherty1,”’ he writes, “ the more apparent it becomes that these two species are the most specialized members of the genus [his Lewco- treron|; it is also evident that in spite of their superficial dissimilarity to each other in colour, they are certainly derived from the same ancestral stock, the densely feathered tarsi, proportion of wing to tail, modification of the inner primaries and the pale anterior part of the body sharply defined from the dark posterior, all point to some common ancestor.’’ His failure to associate them with Pt. porphyrea in spite of these very apposite remarks is in agreement with the fact that in the Checklist (1937) he arranged the members of the purpuratus species-group in a very similar and artificial way, bringing together the most highly ornamented forms with no regard for the geographical evidence. Bifid breast-feathers occur independently (as Peters points out, p. 388) in several groups of pigeons, including (to take examples only from the fruit-pigeons) species of the purpuratus species-group and luteovirens species-group in the subgenus Ptilinopus, and Ducula goliath. They vary much in degree of development even among closely related forms ; even if they do represent the retention of an ancestral character, it is evident that the common ancestor possessing them must be a long way back in the lineage of the pigeons. Certainly those species that possess them to-day are not closely related. The whole pattern of ornamentation of Pé#. por- phyrea links it not with the subgenera Ramphiculus or Ptilinopus, but with Pt. doherty: and Pt. cincta. Even if richness of ornamentation can be considered as one “ character,” it is not a useful character in this genus since it has evidently been lost (or perhaps gained) independently several times, as shown in the following table : “ ” SUBDIVISIONS OF THE GENUS PTILINOPUS 281 Ornamentation in Closely Allied Species Species-group. Complex. Simple. Occipitalis . i Pt. marchei (4) : Pt. merrilli (5) Pt. occipitalis (6) A Pt. fischeri (7) Pt. leclancheri (8) 4 Pt. subgularis (9) Ornatus : 3 Pt. ornatus (12) and 6 Pt. tannensis (14) perlatus (13) Purpuratus . 5 Pt. superbus (15) 2 Pt. perousii (16) (see Cain, 1954) Pt. regina (20) 5 Pt. purpuratus (27) and geographically adjacent forms. The luteovirens species-group (41-3) also consists entirely of a group of long- isolated forms with extremely simple patterns derived from the subgenus Prilinopus (Amadon, 1943) which contains principally highly ornamented forms. Variation between a complex and a more simple pattern has therefore occurred at least seven times independently in the genus (and certainly involves a simplification in three examples). But, more important, “richness of ornamentation” as a character is clearly applicable to species which agree only in that they are richly ornamented and differ profoundly in their patterns of ornamentation. Pt. porphyrea (1), marchei (4) and swperbus (15), for example, are all richly ornamented, but their affinities are not with each other but with Pt. doherty: (2), merrilli (5) and the purpuratus species- group (15-28) respectively ; to class them together on the basis of this very super- ficial “‘ character ’’ is to ignore entirely the wealth of evidence on their real affinities provided by their pattern of ornamentation. Colours often vary greatly in intensity between subspecies, still more between species, but the pattern tends to remain constant in most species-groups. There is no doubt that the pattern of Pt. porphyrea (z) links it with Pt. dohertyi (2) and Pt. cincta (3). The three species of Leucotreron are geographical representatives except for a single overlap on Bali, and can be regarded as a single superspecies. The loss of bright red pigment in the Lesser Sunda Isles form (Pt. cincta (3)) is paralleled by the reduction of red and yellow in the subspecies of Pt. regina (20) (purpuratus species-group) in the same area. No other species of the genus Ptilinopus is found, in these islands except Pt. melanospila (39), which is obviously a very recent arrival from Celebes. The species of Ramphiculus fall naturally into two ecological groups, the marchei subgroup (4-5) being confined to mountain forest in the Philippines (Delacour and Mayr, 1946), while Pt. occipitalis (6) and lechlancheri (8) are lowland-forest species both of which have colonized Celebes and produced there simplified forms with no ornamentation on the breast and belly. (3) Subgenus Ptilinopus This subgenus can be readily subdivided into five species-groups, characterized by the accentuation of different elements of the pattern. In the ornatus group 282 SUBDIVISIONS OF THE GENUS PTILINOPUS there is considerable ornamentation of the wing-coverts and head, but very little on the lower breast and belly. In the purpuratus group the cap, breast-patch, abdominal band or spot, and abdominal patch are conspicuous. In the viridis group the breast-patch is greatly accentuated, and in the hyogastra group the abdominal patch. The luteovirens species-group (41-3), like Pt. perousii (16) (Cain, 1954) shows a simplified pattern, as is usual in members of reduced avifaunas. Strong sexual dimorphism is found sporadically in the subgenus Ptilinopus (but in the other subgenera only in Pt. leclancheri (8)) and requires discussion in relation to specific limits. It occurs in the ornatus group only in Pt. tannensis (14) (Amadon, 1943). In the purpuratus group it is present in both members of the superbus subgroup, and in Pt. monacha (17), which is otherwise very closely related to Pt. coronulatus (18). In the hyogastra group dimorphism is slight or absent except in Pt. melanospila (39) in which the female has a green head. It is also seen in Pt. jambu (40) in which the red of the head is much duller in the female and there is some green on the breast. In the viridis group strong dimorphism is usual, but it varies greatly in the various subspecies of P#. viridis (33). In this species there is a red breast-patch, which is large in Pé. v. viridis (southern Moluccas) and present in both sexes although perhaps very slightly smaller in the females. In Pt. v. pectoralis (western Papuan Islands and north-west New Guinea) it is very small in the males and absent in the females. In Pt. v. salvadorii (Japen and the adjacent part of northern New Guinea) it is rather larger and is present, though reduced, in the females. In fact it appears that in these subspecies the patch is always smaller in the females than in the males and when it is very reduced in the males it is neces- sarily absent in the females. But in Pt. v. geelvinkianus (some islands in Geelvink Bay) it is large in the males and quite absent in the females. In Pt. v. vicinus (D’Entrecasteaux Archipelago and Trobriand Isles), Pé. v. Jewisii (Lihir Islands and most of the Solomons), and Pt. v. eugeniae (San Christobal) it is quite large in both sexes. Because of this variation in dimorphism, Peters divides Pt. viridis, as understood here, into four species, Pt. viridis, eugeniae (including lJewisii and vicinus), geel- vinkiana, and pectoralis (including salvadorit). This is not necessary. All these forms are very closely related and all are geographical representatives, with no overlap or contiguity of ranges. Consequently, since they never meet in the wild, it is impossible to say whether they are species or subspecies ; when there is reason- able doubt it is much more convenient to list such forms as subspecies, so that their close relationship is immediately obvious from their names. Sexual dimorphism is not necessarily a specific character. It is an individual character varying within single populations in some of the Geospizinae (Lack, 1947). (4) The purpuratus species-group This species-group is the largest and the most difficult to subdivide in the genus. It has been revised recently by Ripley and Birckhead (1942) who have introduced many notable improvements in the arrangement of the various forms. A further consideration of it (Cain, in press) shows that certain alterations to Ripley and SUBDIVISIONS OF THE GENUS PTILINOPUS 283 Birckhead’s subdivisions are necessary. They omit P41. pulchellus (19) and Pt. superbus (15) which also belong to the group (Cain, 1954), and divide the forms into four subgroups, (i) the “‘ old stock,” the forms in Australia, New Guinea the North Moluccas, and the Marianas (17, 18, 20, 21), (i) Subgroup A, the forms from Raro- tonga eastward and Pt. perousii, (16, 25-28), (iii) Subgroup B, those in the Solomons, New Hebrides, Fiji (Pt. porphyraceus), Carolines and Palau Islands (22-24) and (iv) the two Marquesan forms (29, 30) which they think should probably be placed in Subgroup B. The characters given by them as distinctive of Subgroups A and B are inconstant. They were unfortunately unable to see specimens of Pt. raro- tongensis (25). An examination of specimens in the British Museum (Natural History) shows that it is in every way intermediate in pattern as well as geographi- cally between Pt. porphyraceus (24)) (Subgroup B) and P+. purpuratus (27) (Subgroup A), and that the proposed distinction between the two subgroups cannot be upheld. On the basis of this distinction, Ripley and Birckhead suggest that the “‘ old stock ” has given rise to two eastward expansions, one producing Subgroup A including Pt. perousii (16), the other Subgroup B, to the west of A, which has since spread into Fiji (Pt. porphyraceus (24)) and there overlaps with Subgroup A without inter- breeding. Subgroup B, they suggest, has also colonized the Marquesas twice (producing Pt. dupetithouarsii (30) and Pt. mercieri (29)), leaping over the enormous range of Subgroup A to do so. The abolition of the distinction between Subgroups A and B allows us to recognize that, with the exception of Pt. perousii (16) and the Marquesan species (29, 30) all the forms from Pt. regina (20) eastward are geographically representative and so closely allied that there is no reason to believe them to be more than the results of a single vast eastward expansion (Cain, in press). Pt. perousii (16) does not belong to this superspecies, but forms with Pt. superbus (15) a distinct subgroup within the species-group (Cain, 1954). The closest allies of the Marquesan species are Pt. insularis (28) on Henderson Island (regarded by Ripley and Birckhead as a subspecies of Pt. purpuratus (27)) which appears to be closely related to Pt. mercieri (29), and Pt. purpuratus (27) which is allied to Pt. dupetithouarsti (30). There is no doubt (Mayr, 1940) that the Marquesan species are the result of a double invasion by the same stock. Both, consequently, are geographical representatives of their closest allies, but it is not possible to choose one rather than the other to add to the Superspecies, leaving one outside it because of their overlap. Consequently in the classification given above they are placed next to their closest allies and bracketed together as a doublet. The same procedure is used for Pt. pulchellus (19) and Pt. coronulatus (18) for the same reason. The whole of the purpuratus subgroup can then be accurately described as one superspecies with two doublets. The Henderson Island form, Pt. insularis (28) although a geographical representa- tive of all the others in the subgroup, is probably rather more closely allied to a member of the Marquesan doublet (namely Pt. mercieri (29)) than to any of the Non-overlapping forms. But its relationships are so complex that it is retained here as a single species, since it could probably be considered almost equally well as a subspecies of either Pt. mercieri (27) or Pt. purpuratus (27). Whatever its detailed Telationships, it is certainly a member of the superspecies. ; a 284 SUBDIVISIONS OF THE GENUS PTILINOPUS SUMMARY I. Proposed classifications of the large genus Ptilinopus are unsatisfactory, because either they are constructed on the principle that anatomical characters are invariably more important than colour-pattern, or they are really keys. A classification considered to be free from these defects is given. 2. Because of considerable variation in specific characters, it may be impossible to diagnose a very natural group of species. The group is then sufficiently defined by a description of its principal trends of variation, and a list of its contents. 3. Examples are given of the use of the species-group as a convenient informal taxonomic rank, indicating relationships without causing nomenclatorial upheavals. 4. The use of brackets for indicating multiple invasions by closely related stocks is exemplified. REFERENCES Amapon, D. 1943. Birds collected during the Whitney South Sea expedition. 52. Notes on some non-passerine genera, 3. Amer. Mus. Novit. : 1237. Capow, G. 1933. Magen und Darm der Fruchttauben. J. Oyn. 81 : 236-252. Cain, A. J. 1954. Affinities of the fruit pigeon Ptilinopus perousii Peale. Ibis 96 : 104-110, Detacour, J., & Mayr, E. 1946. Birds of the Philippines. xv + 309 pp., 69 text-figs. New York. Garrop, A. H. 1874. On some points in the anatomy of the Columbae. Proc. Zool. Soc. London, 1874 : 249-259. Hartert, E. 1896. An account of the collection of birds made by Mr. William Doherty in the Eastern Archipelago. VII. The birds of Sumba. WNovit. Zool. 3: 576-590 and pls. Ir and 12. Lack, D. 1947. Darwin’s Finches. x + 208 pp., 8 pls., 27 text-figs. Cambridge. ManueEL, C. G. 1936a. New Philippine fruit pigeons. Philippine J. Sci. 59: 307-310, 1 col. pl. —— 1936b. A Review of Philippine pigeons. II. Subfamily Ptilinopodinae. Jbid., 327-336. Mayr, E. 1940. Speciation phenomena in birds. Amer. Nat. 74 : 249-278. 1942. Systematics and the origin of species. xiv + 334 pp., 29 text-figs. New York. McGrecor, R. C. 1916. New or noteworthy Philippine birds 1. Philippine J. Sci. Section D, 11 : 269-277. Peters, J. L. 1937. Checklist of Birds of the World, 3. xiii + 311 pp. Cambridge (Mass.). 1938. Generic limits of some fruit pigeons. Proc. 8th int. orn. Congr. Oxford (1934). 371-391. RenscuH, B. 1929. Die Berechtigung der Ornithologischen systematischen Prinzipien in der Gesamtzoologie. Pyvoc. 6th int. orn. Congr. Copenhagen (1926), 228-242. Rretey, S. D., & BrrckHEapD, H. 10942. Birds collected during the Whitney South Sea Expedition. 51. On the fruit pigeons of the Prilinopus purpuratus group. Amer. Mus. Novit. : 1192. RotTuHscHILD, W. 1896. [A new pigeon from Sumba Island.] Bull. Brit. Ovn. Club, 5: xlvi. Satvapor!I, T. 1893. Columbae. Catalogue of the Birds in the British Museum, 21. Wetmore, A. 1925. In Wood, C, A., and Wetmore, A. A Collection of birds from the Fiji Islands. Ibis (12) 1: 814-855. ee - er IP Se sass RAN ERI Key to Colouring —_— Black or Blue-Black White Orange, Red, Violet | Intensity proportional to depth of colour. Other Colours ) LEUCOTRERON TetGeete LECLANCHERI OCCIPITALIS RAMPHICULUS ORNATUS PTILINOPUS HYOGASTRA VIRIDIS Fic. 2. PRINTED IN GREAT BRITAIN BY ADLARD AND SON, LIMITED, . BARTHOLOMEW PRESS, DORKING _ BULLETIN OF SH MUSEUM (NATURAL HISTORY) : bs Vol. 2 No. 9 LONDON: 1955 | A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA BY G. OWEN EVANS Pp. 285-307; 42 Text-figures i } BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 2 No. 9 LONDON: 1955 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series corresponding to the Departments of the Museum, and an H istorical Series. Parts appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 2, No. 9 of the Zoological series. PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued January, 1955 Price Seven Shillings A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA By G. OWEN EVANS SYNOPSIS One new genus and six new species are described ; the classification of the Epicriina is dis- cussed and a new family ARCTACARIDAE is erected; the Arctic species (females only) of the genus Arctoseius Thor are keyed. INVESTIGATIONS on the acarine fauna of the Arctic prior to 1904 have been sum- marized by Tragardh (1904). Since this date a number of papers on the systematics and ecology of Arctic mites has appeared particularly on the Oribatei. The most important contributions to the study of the Mesostigmatid mite fauna have been made by Tragardh (1910), Thor (1930), Sellnick (1940) and Haarlov (1942). The present work deals with part of a collection of mites from Alaska deposited in the collections of the British Museum (Nat. Hist.) by the collector, N. A. Weber, in 1950. A preliminary account of the fauna and descriptions of the habitats from which collections were made has been published by Weber (1950). Of the seven species of Mesostigmatid mites present in the collection six are described as new to science including one new genus and one new family. The type material is deposited in the British Museum (Nat. Hist.). MESOSTIGMATA—GAMASINA Family LAELAPTIDAE Berl., 1892. Haemogamasus alaskensis Ewing, 1925 1925. Haemogamasus alaskensis Ewing, H. E., Proc. biol. Soc. Washington 38 : 138-139. 1933. Haemogamasus sternalis Ewing, H. E., Proc. U.S. Nat. Mus. 82 : 3. The type specimen, a single female, was collected on a Microtus sp. at Crater Mountain, Ophir, Alaska. It is also known from a number of localities in United States of America (Keegan, 1951). The collection under study contained numerous females and nymphs from Point Barrow. Family PHYTOSEIIDAE Berl., 1913 Genus Arctoseius Thor 1930. Arctoseius Sig. Thor, Skr. Svalbard Ishavet Oslo, 27 : 112. 1948. Tvistomus Hughes, A. M., Mites associated with stored food products, H.M.S.O., London : 138-139, syn. nov. ZOOL, II, 9. 17 288 A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA The genus Arctosezus was erected by Thor (1930) for Arctoseius laterincisus Thor (1930), a mite collected under stones on the island of Spitzbergen. The original description and photographs of the genotype are unfortunately inadequate for its precise identification. Haarlov (1942) published a redescription, with figures, of what he considered to be A. daterincisus from material collected in Greenland. Thor and Haarlov’s species are undoubtedly congeneric but without examining the type material of the former it is not possible to be certain that they are conspecific. For the present, however, the writer accepts Haarlov’s interpretation of Thor’s species. In 1948, Hughes erected a new genus, Tvistomus, which is closely related to Arcto- seius in the structure of the venter of the female, the epistome and gnathosoma. Recently the writer remounted and examined the genotype, Tvistomus butleri Hughes, and found that the dorsal shield in both sexes has lateral incisions charac- teristic of Arctoseius ; the species should be removed to that genus. The most comprehensive study of the genus Arctoseius to date was made by Willmann (1949) who added seven species to the genus ; six of which were new to science. Willmann also proposed the division of the genus into four subgenera, namely, Arctoseius s. str., Arctosevulus, Arctotarseius and Arctoseiodes. The sub- genus Arctoseiodes (type Arctoseius (Arctoseiodes) ibericus Willm. (1949) is charac- terized by females having a ventri-anal shield and a genital shield with a pair of setae. It is sufficiently distinct from the Arctose:us group to warrant generic status. The subgenera Avctoseiulus (type Laelaps (Iphis) semiscissus Berl.) and Arctotarseius (type A. (Arctotarseius) austviacus Willm.) are separated from Arctosezus s. str. by the occurrence of a pair of jugular plates in the former and the structure of tarsus I in the latter. Vitzthum (1941) placed Arctoseius in the Hypoaspidinae, but as Willmann (1949) has pointed out, the genus is more closely related to the Lastoseus-group. The sub- family Podocininae into which Willmann (loc. cit.) transferred the genus contains as its type genus Podocinium Berl., a typical Macrochelid. Arctose:us should be placed in the Phytoseiinae (family Phytoseiidae). The following is an emended definition of the genus Arctoseius : Small free-living mites with the entire dorsal shield incised laterally. Dorsal shield with 31 pairs of simple setae, lateral interscutal membrane with ro pairs of setae. Ventral surface of female with a narrow genital shield without setae and pores. Sternal shield with two or three pairs of setae. Metasternal setae may or may not be situated on small platelets. Anal shield well separated from the genital. Peri- trematal plate usually fused with dorsal shield anteriorly but may be free in those species showing a reduction in the peritreme. In the male the sterniti-genital shield is distinct from the large ventri-anal shield. The gnathosoma is normal for the family. The chaetotaxy of the palptrochanter, femur and genu is (2-5-6). The epistome is bi- or tridentate. All the legs are provided with a pulvillus and two claws. Genotype, Arctoseius laterincisus Thor (1930). The Arctic species (females only) of Arctoseius s. str. may be separated as follows : 1. Tarsus I with a conspicuous sensory organ laterally Arctoseius laterincisus (Thor) Haarlov, 1942. —. Tarsus I without such sensory organ laterally. 0 a : : . 2. A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA 289 . Epistome bidentate . 4 3 0 5 Arctoseius cetvatus (Sell), 1940. . Epistome tridentate . d c c 0 ¢ : : : ° 3- . Peritreme extending beyond the level of coxal . . Arctoseius oynatus sp. nov. . Peritreme reduced, not reaching coxa I 2 = c C ° 4 : . Movable digits of chelicerae about 3 times as long as the corniculi, para-anal setae about $ length of the post-analseta . . Arctoseius webert sp. nov. . Movable digits of chelicerae about 14 times as long as the corniculi, para-anal setae approximately equal in length to the post-anal seta Arctoseius multidentatus sp. nov. A. laterincisus has been recorded from Spitzbergen (Thor, 1930) and Greenland (Haarlov, 1942). The type locality of A. cetratus (syn. Arctoseius bispinatus Weis Fogh, 1947) is in Iceland. This species has also been recorded from Denmark and England under A. bispinatus. 4. - |wtlnr I Arctoseius ornatus sp. nov. Female. The dorsal shield is strongly ornamented and bears 31 pairs of simple setae distributed as in Fig. 1. The lateral incisions of the shield are distinct. The interscutal membrane surrounding the shield is provided with ten setae laterally. Fics. 1-2. Arctoseius ornatus sp. nov., female. Fig. 1, dorsal view. Fig. 2, ventral view. ZOOL. 2, 9. 17§ 290 A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA Ventrally, the heavily ornamented sternal shield has the normal three pairs of setae and their associated pores (Fig. 2). The anterior margin of the shield in the paratype is indentated between the first pair of sternal setae. The posterior margin is truncated. The third pair of sternal pores normally associated with the meta- sternal setae are situated on the posterior margin of the sternal shield. The meta- sternal setae are not borne on platelets. The genital shield is narrow with its pos- terior margin slightly convex. The endopodal plates in the region of coxae III and IV are poorly developed. The genital setae and associated pores are situated lateral toit. The post-epigynial plates are four in number. The region between the genital Fics. 3-6. Arctoseius ornatus sp. nov., female. Fig. 3, lateral view. Fig. 4, Gnathosoma ventral. Fig. 5, tectum. Fig. 6, chelicera. and the anal shield is furnished with ten setae and four small platelets. The meta- podals are elongate. The remaining four ventral setae are placed lateral to the anal shield—two on each side. The anal shield is almost circular in contour (102 x ggu) and bears the normal three setae. The para-anal setae are less than one-half the length of the post-anal seta. The stigmata are situated ventro-laterally in the region of coxa IV. The peri- treme is well developed and reaches beyond coxa I (Fig. 3). The peritrematal plate is relatively large for the genus and is fused anteriorly with the dorsal shield. Pos- teriorly it extends a short distance around the posterior margin of coxa IV. The exopodal plates are fragmentary and not fused with the peritrematal. The tritosternum is normal for the genus. The ventral surface of the gnatho- soma is provided with four pairs of setae distributed as in Fig. 4. The ventral pee’ A) ——— A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA 291 groove is provided with seven rows of denticles. The corniculi are short and do not extend beyond the anterior margin of the palptrochanter. The pedipalps are normal and the chaetotactic formula for the first three free segments is (2-5-6). The specialized seta on the palptarsus is two-pronged. The tectum (epistome) is tri- dentate (Fig. 5). The chelicerae are strongly formed (Fig. 6). The movable digit is bidentate and the fixed quadridentate with a short pilis dentilis. All legs are normal for the genus. Dimensions. Length 490-528u, breadth 280-285. Locatiry. Two females from Point Barrow, Alaska (Coll. A. Weber, 1950) ; holotype (1954.3-19.8) and paratype (1954.3.19.9). Arctoseius weberi sp. nov. FEMALE. The dorsal shield is faintly reticulated. The chaetotaxy of the shield and the lateral interscutal membrane is normal for the genus. Ventrally, the anterior and posterior margins of the ornamented sternal shield are indentated (Fig. 7). The shield bears three pairs of setae and three pairs of pores ; the third pair of sternal pores being situated in its posterior margin. The prae- endopodal shields, although weakly sclerotized, are relatively large. The meta- sternal setae lie posterior to the shield. The endopodal plates are weak. The flask- shaped genital shield is faintly sculptured and covered with minute punctures. The genital setae and pores lie off the shield. The post-epigynial and endopodal shields are well-formed. The eight pairs of ventral setae are distributed as in the figure. The anal shield is broader than long (80 x 98u). The para-anal setae are about one- half the length of the post-anal setae. The stigma is situated ventro-laterally between coxae III andIV. The peritreme is reduced and reaches just beyond the anterior border of coxa II (Fig. 8). The peritrematal plate is strongly formed but not fused with the dorsal shield as in the preceding species. The exopodals are fragmentary and not fused with the peri- trematal plate. The tritosternum is normal for the genus and the gnathosoma is basically similar to that of the preceding species (Fig. 9). The corniculi are approximately one-third the length of the fixed digit of the chelicerae. The pedipalps are normal. The specialized seta on the palptarsus has two prongs. The tectum (Fig. Io) is tri- dentate with the median prong divided into three short processes distally. The movable digit of the chelicera is bidentate and the fixed multidentate with a short pilis dentilis (Fig. 11). All legs normal for the genus. Dimensions. Length 540u, breadth 286u. Locatity. A single female (the holotype 1954.3.19.10) from Point Barrow, Alaska (Coll. A. Weber, 1950). Arctoseius multidentatus sp. nov. The form and chaetotaxy of the dorsal shield is similar to the preceding species. The lateral interscutal membrane bears the normal ten pairs of setae. 292 A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA Ventrally, the sternal shield is lightly ornamented; the ornamentation being strongest in the median region of the shield (Fig. 12). The three pairs of sternal setae are normal in position as are the two anterior pairs of pores. The third pair of pores is unusual, perhaps aberrant, in that one is situated on the sternal shield while the other is located on the interscutal membrane between the posterior-lateral Fics. 7-11. Arctoseius weberi sp. nov., female. Fig. 7, ventral view. Fig. 8, lateral view. Fig. 9, gnathosoma ventral. Fig. 10, tectum. Fig. 11, chelicera. A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA 293 margin of the dorsal shield and the metasternal seta. The region anterior to the sternal shield is lightly sclerotized but not differentiated into prae-endopodal shields as in the preceding species. The large genital shield is wedge-shaped, its posterior margin being slightly convex. The genital setae and pores lie off the shield. The post-epigynial shields are four in number. The eight pairs of ventral setae are dis- Fics. 12-16. Arctoseius multidentatus sp. nov., female. Fig. 12, ventral view. Fig. 13, lateral view. Fig. 14, gnathosoma ventral. Fig. 15, tectum. Fig. 16, chelicera. 204 A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA tributed as in the figure. The metapodals are elongate. The anal shield is irregular in contour. The anal setae are all approximately equal in length. The stigma is situated ventro-laterally between coxae III and IV (Fig. 13). The peritreme is reduced and extends as far as the middle of coxa II. The peritrematal plate is well-developed being fused into the dorsal shield anteriorly and externally around part of coxa IV posteriorly. The exopodal plates are fragmentary. The gnathosoma has four pairs of ventral setae (Fig. 14). The corniculi are rela- tively long and extend beyond the anterior margin of the pedipalp trochanter by nearly half their length. The fixed digit of the chelicera is only about one and a half times as long as the corniculi. The tectum is tridentate (Fig. 15). The movable digit of the chelicera is bidentate, the fixed multidentate and with a short pilis dentilis (Fig. 16). All the legs are normal for the genus. Dimensions. Length 605u, breadth 275n. Locatity. A single female (the holotype, 1954.3.19.7) from Point Barrow, Alaska (Coll. A. Weber). MESOSTIGMATA—EPICRIINA At present there are two conflicting views on the classification of the EPICRIINA depending on the relationship between the Epicriidae and Zerconidae. Vitzthum (1941) considers these families to be closely related and includes both in the EPpi- CRIINA, which he defines as follows : “Gnathosoma von oben meist nicht sichtbar. Mannliche Genitaléffnung inmit- ten des Sternale. Rumpf meist etwas flachgedriickt, so dass Ritickenflache und Bauchflache durch eine mehr oder weniger scharfe Kante geschieden sind. Mann- liche Cheliceren ohne Spermatophorentrager. Weibliche Genitaléffnung ein Quer- spalt vor einem Genitiventrale, das zwei oder mehr Haarepaare tragt.—Bisher ist nur bei Tviangulozercon ein dem der Gamasides gleichendes Herz nachgewiesen.”’ Tragardh, on the other hand, disagrees with Vitzthum and in his important work on the classification of the Mesostigmata, based primarily on the structure of the sterniti-genital region in the female, places the Zerconidae in a distinct division, or cohort, the ZERconINA. In Tragirdh (1938) the Epicriidae are included in the Sejidae (= Liroaspidae) a family of the SEJINA (= LrroaspiNa) on the basis of the females possessing a primitive genital shield. This relationship between the Epi- criids and Liroaspids was retained by Tragiardh (1946a) in a revised classification of the Mesostigmata, but later (Tragardh, 19460) the Epicriidae were removed from the LiroasPINA because of the differences in structure of the genital region of the male. This resulted in the EpicriNa being grouped with the ZERCONINA since both had the male “‘ aperture closed by a biarticulated plate attached at the anterior margin.” Tragardh’s separation of the Epicr1INA and ZERCONINA appears therefore to be based on a difference in the structure of the sterniti-genital region of the females, which he considered to be fundamental. Recently the writer has compared the structure of the genital plate in a number of species of Epicrius s. lat. and Zercon. In both genera the function of the plate is basically the same as in the Laelaptoidea, the only difference being in the reduction of the hyaline epigynial portion of the plate which A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA 295 is so well-developed in the free-living Laelaptoidea and Macrocheloidea. In Epi- crius and Zercon the epigynial portion of the plate is reduced to a narrow chitinized rim so that the egg is not extruded along a “ chute” asin the Laelaptoidea. Further, the males of both Epicrius and Zercon agree in a number of characters, for example, in the position of the genital orifice, in the presence of a pair of hairs on the anterior plate covering the genital orifice, and in the absence of a spermatophoral process on the chelicerae. In view of the apparent close relationship between the families the writer follows Vitzthum’s concept of the EpicriiNa and proposes the following classi- fication : Division EpicrtIna: Mites with the epigynial portion of the genital plate in the female reduced to a narrow chitinised rim overlapping the genital orifice. Male genital opening situated in the sternal shield in the region of coxae II and III and closed by two plates, the anterior of which bears a pair of hairs. Chelicerae dentate in both sexes, but without spermatophoral process in the male. Pedipalps with five free segments, specialized seta on palptarsus two or three pronged. The EpicriIna may be divided into the following superfamilies based on the orna- mentation of the dorsal shield and the structure of the stigmata : 1. Peritreme and peritrematal plate absent in both sexes ; the stigmata enclosed in a dorsal shield richly provided with small elevations forming a polygonal network. Jugular plates well-developed in both sexes superfam. Epicrioidea nov. —. Peritreme and peritrematal plates present, dorsal shield without characteristic ornamentation. Jugular plates absent in the female superfam. Zerconoidea nov. The Epicrioidea contains one family only, the Epicriidae, represented by the genera Epicrius (with the subgenus Diepicrius Berl.) and Berlesiana Turk. The Zerco- noidea contains the following families : 1. Dorsal plate divided into a notocephale and notogaster of approximately equal size. Body triangular in outline with serrated lateral margins. All legs with claws. Peritreme reduced inlength. Female genital plate with a pair of setae. Specialized seta on palptarsus two-pronged fam. Zerconidae Berl. —. Dorsal plate entire. Peritreme normal. Leg I in both sexes without claws. Strong sexual dimorphism in the structure of the dorsal shield. Female genital plate without setae. Specialized seta on palptarsus three-pronged fam. Arctacavidae nov. The family Zerconidae comprises three genera, Zevcon Koch (syn. Triangulozercon Jacot), Pavazercon Tragardh (syn. Trizerconoides Jacot) and Prozecon Sellnick. The family Arctacaridae is erected for Arctacarus rostratus gen. nov. et sp. nov. described below. The genus Seiodes Berl. 1887 (genotype Sezodes ursinus Berl. 1887) also belongs to the Epicriina, but whether it is referable to either of the two superfamilies here diagnosed is uncertain pending a re-examination of the types. The genotype species, S. ursinus, has the following characters: Dorsal shield entire, ventri-anal shield occupying the greater part of the region posterior to coxae IV in both sexes. All legs with well-developed claws. Femur II in the male spurred. Male genital orifice in sternal shield between coxae III. Chelicerae of the male without a spermatophoral process. The species named S. hystricinus by Berlese (1892) may not be congeneric with ursinus. 296 A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA The collection from Alaska contained two species belonging to the Eprcriina and these are described below. Family ZERCONIDAE The only representative of this family is a female of the genus Zercon. To date four species of the genus have been recorded from the Arctic, namely, Zercon curiosus Tragardh, 1910, Zercon arctuatus Tragardh, 1931, Zercon triangularis Koch, 1836, and aeons Ce 7 Lee eR LFS ey AES. i ae Ax ae Oh ae J S\N oo {¥ == "_, SY pp— ea NaS Bag te -0ey ; (Riel ity TES a ee Oe 22 wee a, 4 Y WA me") 39” 23 } See ecu j Vo. ee Fe Jaq S \e Boy orate ra Pos. ; h\ noe Be J59 oT f 1 \ RotooNe . yA zs Fic. 17. Zercon fenestralis sp. nov., female. Fig. 17, dorsal view. Zercon solenites Haarlov, 1942. Z. curiosus is recorded from a number of localities in Swedish Lappland, Z. avctuatus was first collected in an old orchard at Thorshavn in the Faroes (Tragaérdh, 1931) and has recently been recorded from S. England by Evans (1953), Z. trviangularis, reported from a number of localities in Spitzbergen by Thor (1930), is a widely distributed species in Europe. These three species are described and keyed in Sellnick (1944). The fourth species, Z. solenites, is known only from the type locality in Northern Greenland (Haarlov, 1942) and differs from the other known species of Zercon in having a well-developed peritreme reaching to coxa I. A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA 297 The species collected in Alaska appears to be new to science on the basis of the chaetotaxy of the dorsal shields. The nomenclature for the chaetotaxy of the noto- gaster follows that of Sellnick (1944). Zercon fenestralis sp. nov. FEMALE. The dorsum is entirely covered by two shields of approximately equal size. The anterior shield, the notocephale, bears 15 pairs of simple dorsal setae dis- tributed as in Fig. 17. The remainder of the chaetotaxy comprises stout coarsely serrated setae. These number six pairs, including the verticals. The lateral margins Fics. 18-19. Zercon fenestralis sp. nov., female. Fig. 18, sterniti-genital region. Fig. 19, peritreme and peritrematal plate. of the notocephalic shield are serrated. Its surface is strongly reticulated and is provided with pores. The notogaster bears fifteen pairs of setae excluding a lateral series of seven. All the lateral setae are simple. Setae of the J series are well- developed. Jx1—J4 are short, simple, and of about equal length. The stout ser- tated setae, J5 and J6, are more than three times the length of Ji-J4. Series Z comprises six setae of which Zr and Zz are similar to Jr and J2 in size and form. Z3 and Z4, however, resemble J5 and J6 in structure. Z3 is shorter than Z4. Z5, situated in close proximity to J6, is simple. Series S consists of four setae on either side. Sz is simple and the same length as Jr and Zr. Setae S2-S4 are stout and serrated. Sr is closer to Zr than S2, but S2, S3 and S4 are equidistant from each other. Pore 3 is situated a short distance in front of seta Z4 and inside the line connecting Z3 and Z4. The distribution of the remaining pores is shown in the 298 A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA Fics. 20-23. Arctacayus rostratus gen. et sp. nov. Fig. 20, dorsal view. Fig. 21, dorsal seta. Fig. 22, ventral view. Fig. 23, tectum. A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA 299 figure. The anterior half of the notogaster is strongly reticulated and the posterior half coarsely punctured. The four posterior depressions are conspicuous. Ventrally, the sternal shield bears three pairs of simple setae and three pairs of pores (Fig. 18). The ornamentation is strong except for an oval area in the region of sternal setae II. The endopodal plate in the region of coxae II is fused with the lateral margin of the sternal shield. The metasternal setae are situated on the inter- scutal membrane posterior to the sternal shield and between them lie a pair of weakly sclerotized plates. The genital shield, truncated posteriorly, has a pair of genital setae. The epigynial portion of the plate is strongly sclerotized. The head of a club-shaped vaginal sclerite (?) protrudes from under the epigynial shield. The endo- podal plate in the region of coxae III and IV is interrupted opposite the middle of coxae IV. The large ventri-anal shield is of the form normally found in the genus and bears eight pairs of simple setae (excluding the paranals and the post-anal setae). The stigma and its short peritreme are enclosed in a large peritrematal plate trun- cated posteriorly and fused anteriorly with the notocephalic shield. The internal margin of the peritrematal is fused with the exopodal (Fig. 19). The peritrematal plate has a stout serrated seta and a large pore. The legs and gnathosoma are normal for the genus. Dimensions. Length 465, breadth 375u. Locauity. A single female, the holotype (1954.3.19.27) from Point Barrow, Alaska (Coll. A. Weber). Family ARCTACARIDAE fam. nov. Diacnosis. Dorsal shield heavily sclerotized in both sexes completely covering the dorsum of the female. Sterniti-genital region of female normal for the Epi- criina. In the male the sterniti-genital plate is fragmented in the region of coxae III-IV and separated from the ventri-anal plate. Male genital orifice closed by two plates, the anterior of which is provided with a pair of stout setae. Peritrematal plate fused with the dorsal shield anteriorly in the female but fused along its entire length in the male. Coxa I, femur II and IV spurred in male. Tarsus I in both sexes without pulvillus and claws. Tarsi II-IV with pulvillus and claws. Chelicerae dentate. Specialized seta on palptarsus three-pronged. Genus Arctacarus gen. nov. With the characters of the family. Type: Arctacarus rostratus sp. nov. Arctacarus rostratus sp. nov. FEMALE. The dorsal shield, 820-8304 x 515-528, is chestnut brown in colour and does not completely cover the dorsum of the mite (Fig. 20). The surface of the shield is finely striated and the dorsal setae are setose distally (Fig. 21). The shield has a large pair of lyriform pores postero-lateral to the vertical setae and a number of smaller pores distributed as in the figure. The posterior region of the dorsum is 300 A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA Fics. 24-29. Arctacarus rostratus gen. et. sp. nov. Fig. 24, lateral view. Fig. 25, Gnathosoma ventral. Fig. 26, specialized seta of palptarsus. Fig. 27, chelicera show- ing variation in the dentition of the movable digit. Fig. 28, Distal end of tarsus I. Fig. 29, tarsus IV. A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA 301 not sclerotized and bears nine pairs of dorsal setae of similar form to those on the dorsal shield. Anterior to the posterior dorsal setae lies a small, weakly sclerotized plate incompletely divided into two equal-sized platelets. It is interesting to note that a pair of platelets is also present in this position in some species of Vezgaia which have a larger dorsal shield (or shields) than Arctacarus. The tritosternum is well-developed with the lacinae long and setose. The orna- mented sternal shield bearing three pairs of sternal setae and three pairs of pores extends from the posterior edge of coxae I to the anterior border of coxae III (Fig. 22). It is produced into a broad sclerotized band between coxae I and II and is fused along its lateral margin with the endopodal plates. The region between the anterior margin of the sternal shield and the tritosternum is occupied by the frag- mented prae-endopodal plates. The general structure of the sternal region in Arcta- carus is not unlike that in the females of Vezgaia. The metasternal setae lie on the interscutal membrane between the sternal shield and the anterior lip of the genital orifice. In between the metasternals lies a pair of small lightly sclerotized platelets (cf. Zercon). The endopodal plate in the region of coxae III and IV is well-developed and extends around the posterior border of coxae IV. The genital shield is flask- shaped, strongly ornamented and without setae ; the genital setae and pores lying lateral to the shield. The structure of the anterior (epigynial) portion of the shield is similar to that in Zercon. The vaginal sclerite (?), however, is not nearly as well developed. A further comparison with Zercon can be made in the presence of a small “ pore ’’- bearing plate situated posterior to coxae IV. The chaetotaxy of the region between the genital and anal shields is shown in the figure. The anal shield is pear- shaped and bears three setae—the para-anals and the post-anal setae. The shield is conspicuously ornamented. The chaetotaxy of coxae I-IV, is, respectively, 2-2-2-1. The stigma lies ventro-laterally in line with the third intercoxal space (Fig. 24). The peritreme is well-developed and extends beyond the anterior margin of coxa II. The peritrematal plate, fused at its distal end with the dorsal shield, is broad in the region of the first intercoxal space but narrows markedly lateral to coxae II-IV. It is not fused with any part of the exopodal plate. The exopodal plate in the region of coxa IV is well-developed but in the region of coxae II and III it is fragmented. The interscutal membrane between the dorsal shield and the peritrematal plate has three setae and two weakly sclerotized platelets. Ventrally the gnathosoma bears four pairs of stout setae (Fig. 25). The external posterior rostral seta is situated midway between the anterior rostral and internal posterior rostral setae and external to the line connecting them. The floor of the ventral groove is provided with ten rows of denticles. The corniculi are pointed distally and extend a short distance beyond the middle of the palptrochanter. The chaetotaxy of the palptrochanter, femur and genu is 2-5-6. A number of the inter- nal setae on the palpgenu, trochanter and femur are serrated distally and the two internal setae on the palp spatulate. The specialized seta on the palp tarsus is three-lobed (Fig. 26). The tectum (epistome) is large and triangular in form with five distinct projections (Fig. 23). The dentition of the movable digit is variable. Fig. 27 shows a unidentate and multidentate form; the latter having nine small teeth. The fixed digit has two 302 A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA strong teeth and five to six smaller ones. The fils dentilis is of an unusual form (see figure). Leg I, varying between 1030n and 1045 in length, is approximately equal in length to the body of the mite. Tarsus I is without pulvillus and claws (Fig. 28). The majority of the dorsal setae on leg IV are stout and spiculate distally (Fig. 29). One of the females examined contained an egg measuring 400% X 320u. Fics. 30-33. Arctacarus rostrvatus gen. et sp. nov., male. Fig. 30, dorsal view. Fig. 31, ventral view. Fig. 32, leg I. Fig. 33, leg IV. A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA 303 Mate. The contour of the body differs markedly from the female. The dorsal shield covers the entire dorsum of the male. The shield is broadest at the shoulders and tapers strongly towards the posterior end of the body. The chaetotactic pattern and the distribution of pores is shown in Fig. 30. The dorsal setae are serrated. The tritosternum is basically the same as in the female (Fig. 31). The sterniti- genital region is provided with a number of plates. The prae-endopodals are elongate and distinctly separate from the sternal shield. Sternal setae I are situated on the heavily ornamented anterior portion of the plate and are separated from it by a narrow strip of less strongly chitinized cuticle. Sternal setae II and III and pores I and II are placed in the weakly ornamented portion of the sternal shield extending from the middle of the first intercoxal space to the posterior margins of coxae II. Fics. 34-36. Arctacarus rostratus gen. et sp. noy., male. Fig. 34, lateral view. Fig. 35, tectum. Fig. 36, chelicera. The shield is not produced between coxae I and II as in the female. The genital orifice is situated in the posterior third of the shield and is closed by two plates ; the anterior of which bears two stout setae. The shield in the region between coxae II and III is fragmented and variable in form. The lateral margin of the shield is fused with the endopodal plate. In the type specimen the shield is divided longitudinally into two shields of unequal size and each shield bears two setae and the metasternal or third sternal pore. The shields are punctured. The ventri-anal shield has four pairs of pre-anal setae in addition to those setae associated with the anal region, namely, the para-anals and the post-anal setae. 304 A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA The peritrematal plate is fused along its entire length with the lateral margin of the dorsal shield (Fig. 34). The stigma is situated laterally in the region of the third intercoxal space. The peritreme is strong and extends beyond the anterior margin of coxa II. The exopodal plate of coxa IV is well-developed and entire. The remainder of the exopodal plate is fragmented. Fics. 37-39. Dinychus micropunctatus sp. nov., female. Fig. 37, dorsal view. Fig. 38, gnathosoma ventral. Fig. 39, tectum. The gnathosome and palps are the same as in the female. The epistome (tectum) on the other hand is considerably more strongly developed than in the female and is beak-like in form (Fig. 35). The fixed digit is bidentate and bears a short spine-like pilus dentilis (Fig. 36). The movable digit, without spermatophoral process, is unidentate. A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA 305 Leg 1 is long and slender as in the female and is without pulvillus and claws. Coxal is provided with a strong spur (Fig. 31). Legs II-IV terminate in a pulvillus and two claws. The femur of legs II and IV are spurred (Figs. 32 and 33). Drvensions. Female: Length 1050—-1080p, breadth 515-528. Male: Length 925-940, breadth 575-590u. Locatity. Two females and four males from Point Barrow, Alaska (Coll. A. Weber, 1950). Holotype female (1954.3.19.21), Allotype male (1954.3.19.23) and paratypes (1954.3.19.22, 24-26). MESOSTIGMATA—UROPODINA Family PRODINYCHIDAE Dinychus micropunctatus sp. nov. FEMALE. The dorsal shield is oval in contour and densely covered with minute punctations. The chaetotaxy and distribution of pores is shown in Fig. 37. The vertex is short with the vertical setae long and stout. The two pairs of long setae situated posteriorly on the dorsal shield are setose. The marginal shield is of the Fics. 40-42. Dinychus micropunctatus sp. nov. Fig. 40, sterniti-genital region of female. Fig. 41, lateral view of female showing peritreme. Fig. 42, sterniti-genital region of male. 306 A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA same structure as the dorsal shield. The posterior marginal shield bears two pairs of short, stout serrated setae and two pairs of simple setae. The latter lie on either side of the serrated setae. The tritosternum is short and of the form shown in Fig. 40. The sternal region is covered with minute punctuations and a faint polygonal network. The anterior margin of the sternal shield is straight. Sternal setae I and II are situated between coxae II; on setae I and II on either side are place on a sclerotised ridge (Fig. 40). Sternal setae III and IV (according to Tragirdh, 1943, seta IV is the pseudosternal) lie lateral to the large epigynial shield which extends from the middle of coxae IV to the posterior margin of coxae II. The ornamentation of the epigynial shield is similar to that of the sternal shield. Four pairs of pores are present in the sterniti- genital region. The remainder of the venter of the female is typical for the genus. The peritreme is long and convoluted. The stigma is situated ventro-lateral of coxa III (Fig. 41). The peritreme continues a short distance posterior to the stigma. Anterior to the stigma the peritreme is looped in the region of coxae II. A pore is present anterior and posterior to the loop. The four pairs of ventral setae are distributed as in Fig. 38. The capitular setae are short and setose distally. The external posterior rostrals are long and smooth and the internal posterior rostrals, situated anterior to the exterior posterior ros- trals, short and simple. The anterior rostrals are long, about equal in length to the exterior rostrals, and setose. The corniculi are short and extend to the anterior margin of the palpfemur. The chaetotaxy of the palptrochanter, femur and genu is 2-5-5). The setae on the trochanter are stout and setose. The tectum (epistome) is produced into a long, narrow process, bifid distally (Fig. 39). The shaft is covered with small spines. All legs terminate in a pulvillus and two claws and are normal for the genus. Mate. The structure of the dorsal shield and peritreme is similar to that of the female. The sterniti-genital region is well differentiated (Fig. 42). The genital orifice is situated between coxae IV. The distribution of setae and pores is shown in the figure. The complete venter of the male is finely punctated. Dimensions. Female: Length 670-693u, breadth 370-375u. Male: Length 682-715, breadth 380-385 w Locatiry. Two females and four males from Point Barrow, Alaska (Coll. Weber.). Holotype female (1954.3.19.1), Allotype male (1954.3.19.2) and paratypes (1954 3-19.3-6). Dinychus micropunctatus sp. nov. is closely related to Dinychus sublaevis (Tragardh) 1943, in the chaetotaxy of the posterior dorsal region of both sexes. It differs from D. sublaevis in the ornamentation of the dorsal and ventral shields, and the structure of the sternal region of the female. SUMMARY 1. A small collection of Mesostigmatid mites from Alaska comprises the following seven species : Haemogamasus alaska Ewing. Arctose1us ornatus sp. nov. A COLLECTION OF MESOSTIGMATID MITES FROM ALASKA 307 Arctoseius weberi sp. nov. Arctoseius multidentatus sp. nov. Zercon fenestralis sp. nov. Arctacarus rostratus gen. et sp. nov. Dinychus micropunctatus sp. nov. 2. The Arctic species (females only) of the genus Arctosezus Sig Thor s. str. are keyed. 3. Tristomus Hughes 1948 is made a synonym of Arctosezus. 4. The classification of the EpicriINA is discussed and a new family ARCTACARIDAE is erected for Arctacarus gen. nov. Acknowledgments Tam grateful to Dr. Max Sellnick, Stockholm, for comparing specimens of Dinychus mucropunctatus with the type of Dinychus sublaevis and to Dr. H. W. Parker for his criticisms of the manuscript. REFERENCES BERLESE, A. (1887). Acari, Myriopoda et Scorpiones ... Fasc. 38, Tav.9. Evans, G.O. (1954). Some new and rare species of Acarina. Proc. zool. Soc. Lond. 123: 793-811. Haari9v, N. (1942). A morphologic-systematic-ecological investigation of Acarina, etc. Medd. Gronland, 128 (1) : 1-71. Hucues, A.M. (1948). Mites associated with stored food products. H.M.S.O. London : 138- 139. Keercan, H. L. (1951). The mites of the subfamily Haemogamasinae (Acari: Laelaptidae). Proc. U.S. Nat. Mus. 101 : 213-218. SELLNIcK, M. (1940). Die Milbenfauna Islands. Géteborgs Vetensk. Samh. Handl. (5) 6B : 1- 129. (1944). Zercon C. L. Koch. Acari-Blattey fur Milbenkunde Konigsberg. 5 : 30-41. Tuor, Sig. (1930). Beitrage zur Kenntnis der Invertebraten Fauna von Svalbard. Skr. Sval- bard Ishavet Oslo, 27 : 1-156. Tracirou, I. (1904). Monographie der arktischen Acariden. Jena: 1-78. —— (1910). Acariden aus dem Sarekgebirge. Naturw. Untersuch. Sarekgeb. 4 (4) : 375-586. (1931). Terrestrial Acarina. Zool. Faroes, 2 (49) : 1-69. (1938). Further contributions towards the comparative morphology and classification of the Mesostigmata. Ent. Tidsky. 59 : 123-158. —— (1943). Zur Kenntnis der Prodinychidae (Acarina). Ark. Zool., 34 : 1-29. (1946a). Outlines of a new classification of the Mesostigmata based on comparative morphological data. Acta. Univ. Lund. N.F. (2) 42 (4) : 1-37. —— (1946b). Contributions towards the comparative morphology of the Mesostigmata (Aca- rina) VII. Ent. Tidsky. 67 : 88-108. VitzTHuM, H.G. (1941). Acarina. Byonns Tierreich, Bd. 5, Abt. 4, Buch 5, Lief. 5 : 641-800. Weser, N. A. (1950). A survey of insects and related arthropods of Arctic Alaska. Tvans. ent. Soc. Amer. 76 : 147-206. Wittmann, C. (1949). Das genus Arctoseius Sig Thor, 1930 (Acari). Abh. naturw. Ver. Bremen, 32 : 349-358. = 6 JAN I955 % ve maine We a a a ¥ ax SG ae &F => : Sar. ae F r ay a STATE: t é ne > 3 Mabe 0 i 2 ‘ 4 . har % ‘ id y rr, ¢- SY a i a - . f ane: | " ‘ * ; : : f ' - tay. >. i “ * . PRINTED IN GREAT BRITAIN BY ADLARD AND SON, LIMITED, BARTHOLOMEW PRESS, DORKING fe TOPARIID) SEALS “OF ‘THE, PACIFIC COAST OF AMERICA BY JUDITH E. KING Pp. 309-337 + Pls. 10-11 ; 3 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 2 No. 10 LONDON: 1954 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series, corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 2, No. 10 of the Zoological series. PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued November, 1954 Price Eight Shillings THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA By JUDITH E. KING From Alaska to Cape Horn along the Pacific coast of the Americas several species of Otariid seals have been recognized. These include both fur seals and hair seals which have been exploited by commercial sealers for over two centuries ; but in spite of this there has always been some doubt about the exact identities of some of the animals concerned. SPECIFIC IDENTITY OF ARCTOCEPHALUS GALAPAGOENSIS HELLER 1. Heller's type specimen During 1898-99 the Hopkins Stanford Galapagos Expedition collected some seal skulls from Wenman Island in the Galapagos group. The skull of an adult male animal was described by Heller (1904) as the type of Arctocephalus galapagoensis. This skull is now in Stanford University, California, and although Heller gave its number as 2480, Mayer (1949) gives the correct number as 2812. Measurements of two other skulls are given by Heller: No. 2481, male, nearly adult ; occipital crests low; sutures indistinct. No. 2482, female, same as 2481 in age. Mayer (1949) gives the correct numbers of these as 4442 and 4446 respec- tively. The specific characters of the skull are given by Heller as, “ Distinguishable from its nearest ally A. philippi of Juan Fernandez by its wider skull, both the zygomatic and mastoid measurements being considerably greater, and by its longer snout and mandible.” Heller gives no indication of having compared this skull with that of any other seal except A. philippi. Photographs of the type skull of A. galapagoensis (Pls. 10A and c; 11A) have been compared with skulls of A. philippi and A. australis, the ZOOL, 2, 10. 18 312 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA two members of the genus Arctocephalus known to occur along the western coast of America. As Heller has stated, the skull of A. galapagoensis is markedly different from that of A. Philippi but it differs not only in the characters he mentions but also in having a wider and less excavated palate, deep zygomatic arches and short broad nasals. These are just the characters in which skulls of A. australis differ so markedly from those of A. piilippi and a comparison of the photographs of the skull of A. galapagoensis with skulls of A. australis shows a remarkable similarity (Pls. 10 and 11). It should perhaps be mentioned that Heller does not indicate whether the total length of the animal from which the type skull was derived was taken from the tip of the snout to the tip of the tail or to the end of the longest digit of the hind flippers. An indication that the measurement was to the latter point is gained from dimensions which Turner (1888, Rep. Sci. Res. H.M.S. “ Challenger’ Zool. 26 : 39) gives of a specimen of an adult male A. australis. They are as follows : From snout to tip of longest digit of pes: 5 ft. 104 in. (1,790 mm.). From snout to tip of tail in straight line: 4 ft. rr in. (1,498-6 mm.). Extreme condylo-premaxillary length of skull: 233 mm. The greatest length of the type skull is 214 mm. and the total length of the animal is given by Heller as 1,675 mm. (5 ft. 6 in.). Thus by comparison with Turner’s measurements, the A. galapagoensis type skull length is in proportion to body length if the latter is taken to the end of the hind flippers and not otherwise. Measurements and proportions of the three Galapagos skulls have been compared with those of all the A. australis skulls in the British Museum collection (Table I). This includes 67 skulls from the Falkland Islands, four from the Straits of Magellan, one from Lobos Island, Uruguay, two from Messier Channel, Chile, and one from Tierra del Fuego. Graphs of the measurements of the various components of the skulls expressed as percentages of their condylo-basal lengths gave no indication of allometric growth. Length ratios plotted against the condylo-basal length showed no geographical differentiation, but width ratios showed that there was some difference between the skulls from the Falkland Islands and those from the “ mainland ”’ population, including the Galapagos Islands. Taking the zygomatic width as an example and using a graphical method the zygomatic width expressed as a percentage of the condylo-basal length in the Falkland sample has a mean value of 57:2%, a standard deviation of 2-5, and a standard error of the mean of 0-29. The correspond- ing figures for the mainland sample including those from the Galapagos, are 61, 3:2, and 0-96. The difference between these means is 3:8 and the standard error of the difference of the means is 1. The ratio = is therefore 3:8. A difference of this magnitude would occur twice in 10,000 trials, but only once in 10,000 in the direction observed. This is statistically significant and indicates that the A. australis population on the Falkland Islands is slightly different from that inhabiting the South American coast and adjacent islands, but there are not enough skulls in the collection to see whether the individual mainland populations are in any way distinct from one another. The zygomatic width proportions of the Galapagos skulls are within the standard deviation of the mainland sample and slightly above that of the Falkland Islands sample. * Posterior to supraorbital processes. THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA | 313 TABLE I.—Measurements and proportions of skulls of A. australis (See notes on measurements used) FALKLAND ISLANDS Registered No. and 1949.3.17.1 1949.3-.17-2 1949.3-17-3 1949.3-17-4 sex, if known Male aT GT: Tk Ga aa mites 9h jaobaay | WK: mm. % ror, | OF Greatest length . 0 ZA 247 248 243 Condylo-basal length . . 24I 100 239 100 246 100 239 100 Zygomatic breadth . - 155 64°3 143 59'8 138 9056-1 146 61-1 Snout length r 5 5 73 «-30°3 73. 30-5 75 30°5 70) | 29°38 Snout width at canines a Gp eatkeXs) Bp > ats) 58 23:6 54 22:6 Snout width at level of 2nd cheek tooth . . p40) nto -9 46 19:2 45 18-3 400) 50-2 Anterior breadth nasals ee Vi esi toar 32 413°4 35) 4:2 ZO) Leia Greatest length nasals sO 49 AL L712 36 §=614°6 38 8615-9 Least interorbital width* . 32 13°3 35 14:6 29 «11°8 20) eer Mastoid breadth 3 - 142 58-9 134 560°1 126 51:2 129 53°9 Length upper tooth - row iim. 6. F 3 92 38:2 88 39 368 96 39°0 89) 37-2 Registered No. and 1949.3.17.5 1949.3.17.6 1949.3-17-7 1949.3-17.8 sex, if known Male (a) SS) a mm. % mm. % mm 19% mm % Greatest length . 230) 249 237 248 Condylo-basal ee . 235 100 242 100 233 100 244 100 Zygomatic breadth . ns 2. 50r2 140 57°9 138) 86 59:'2 145 59°4 Snout length é ce}, SW ioR 7B 30-2 72 30°9 TA 3083 Snout width at canines ay Geb er SAME 22 55 23°6 59 24-2 Snout width at level of 2nd cheek tooth . © 45)" Lo: 45 18-6 46 19°7 5I 20-9 Anterior breadth nesals 4300) 2:8 32.) engh2 34 14:6 Boers Greatest length nasals Be Bee SEX) 35 14°5 40 17°2 40 16-4 Least interorbital width* . 25 10°6 28 11-6 29 12°4 Xo ae 9 aXe) Mastoid breadth c 2 One 5310) 137 -50°6 129 55°4 134 54:9 Length upper tooth - row i—m. 6. . ee Somes 74! 92 38-0 92 39°5 92 37°7 Registered No. and 1949.3.17.9 I1949.3-17.I10 1949.3.17.1I 1949.3.17.12 sex, if known mm % rear MM 5 mm. % Greatest length . ‘ . 240 239 239 255 Condylo-basal length . : 233 100 234 100 233 100 247 100 Zygomatic breadth . Cet 38m E50). 2 139 59°4 I28 54:9 146 59°! Snout length . : oO HE BOO) 7O 29°9 68 29:2 76 30:8 - Snout width at canines co, Or Bue 48 = 20°5 Gi Zino) 72a Snout width at level of 2nd cheek tooth . @ By ANTS T7116 At l7 5 44 18-9 49 19:8 Anterior breadth nasals 20 eL2-4 300 23) 30 12:9 BY autor} Greatest length nasals gon Los: Bou lo 35 15:0 285 5a Least interorbital width* . 31 13:3 27. 1-5 gfe) 10790) 2 Seamer Mastoid breadth ‘ » 129 55°4 126 §653:°8 128 54:9 Misi) cbyey75} Length upper tooth - row i—m. 6. K ‘ bo ie, acion 88 37-6 87 37-3 95 «38°5 314 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA FALKLAND ISLANDS—continued Registered No. and 1949.3.17-13 1949.3-17-I14 1949.3.17.15 1949.3-17.17 sex, if known |) Ga) a reebok, GF pore, GG Travedk WE bobs Oh Greatest length . 0 ez32 c P5He) 2) 235 a 8235 Condylo-basal length . - 225 100 72335100 . 225 I00 =| 220 Too Zygomatic breadth . 3470) 159-0 135 57°9 144 64:0 . 124 54:9 Snout length . 6 OT enzo 70 30:0 68 30-2) - Fo Zo-9 Snout width at canines = » 46) N20=4 50 821°5 52 230 AQE 20a7) Snout width at level of 2nd cheek tooth . ‘ S| 40: E78 42 18-0 44 I9°6 42 18-6 Anterior breadth nasals 3) eat ZO Le Ae eo Tees 28) 124 Greatest length nasals 2 30) 670 BQ Ae | 350) SiO 34 %I5:0 Least interorbital width* . 30 13-3 30) eL25O). | 208 259 25 I'l Mastoid breadth 5 a) T2347 126 54°1 132 58-7 II5 50°9 Length upper tooth - row 1-1. 1675 - - oS ipl} 87" 63723) fa 188 9 300 sce ciEG! Registered No. and 1949.3-17-18 1949.3.17-19 1949.3.17.20 1949.3.17.21 sex, if known Male SO roberts) A mime 0/5 rao OE mm os Greatest length . 5 = 233 225 - 224 Go HS Condylo-basal length . a) 227) LOO ue 223 wTOO! ne? LOM LOO) = 220) 4t00 Zygomatic breadth . 7 223°) 5452 eo te8 Mas7ica, aS 2A) 250-6) Ga OOnG Snout length A 2) 07 6 2055 a2 65). 2020, 7) 164. 20-2) lob mana Snout width at canines a) 404 32023) 40 50) 224) 75 45) 20-5 eS Ome zeed Snout width at level of 2nd cheek tooth . : sz, eeLOcs 41 18-4 38° 174 42 19:1 Anterior breadth nasals 20) LDS 20ers 0 = = 32 «14°'5 Greatest length nasals SAS ting! 33) 48 — = 33 +%15°0 Least interorbital width* . 30 13:2 29) ) 130 32 14-6 29° 32 Mastoid breadth g - 113 49°8 I2I 54°3 TI4 52+] 6 (125) |) 56-8 Length upper tooth - row i—m. 6. ; : Se S7 38h 3tet, SSO” a8 65) he i SOUNESO sis mmo Aim cree Registered No. and 1949.3.17-22 1949.3-17-23 1949.3.17-24 1949.3.17.25 sex, if known a SF OO SO oa rantooy, mm. 95 joobanl, WA oie, Greatest length . 2 zZo 5 (eese) 220 OFT Condylo-basal length . =| 222 100 - 216 tI00 = 204) “100 - 2II 100 Zygomatic breadth I3I 59:0 128 §6559°3 124, 57°9 3 TE esha Snout length . : oy OO Mn Scena OMo EO 65 30°4 . 60) 2sen Snout width at canines Ci 472 eA A ne Osc 40 25 42 19°9 Snout width at level of 2nd cheek tooth . . 2 42S: 9) 36 16-7 30) 18:2 . 376 Sa Anterior breadth nasals oun) SOMELSE De sen 20) 2-9 26 2-1 27 i258 Greatest length nasals eS ORME On 2 3 One TO); 33. «15:4 37. 17°5 Least interorbital width* . 31 13:9 33) 1553 33. «154 32, hee Mastoid breadth a Ree ey 27) II5 53°2 D7 | E547, 108 51'2 Length upper tooth - row i.—m. 6. 9 . PSS) 3096) 183% 38a San a s8i-s. ie, Siemans sie * Posterior to supraorbital processes. THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA Registered No. and sex, if known Greatest length . < Condylo-basal length . Zygomatic breadth Snout length Snout width at canines Snout width at level of 2nd cheek tooth Anterior breadth nasals Greatest length nasals Least interorbital width* Mastoid breadth . Length upper tooth - row i—m. 6. a0 6 6 Registered No. and sex, if known Greatest length . 3 Condylo-basal length . Zygomatic breadth Snout length Snout width at canines : Snout width at level of 2nd cheek tooth 5 Anterior breadth nasals Greatest length nasals Least interorbital width* Mastoid breadth a Length upper tooth - row i—m. 6. : 4 a Registered No. and sex, if known Greatest length . ; Condylo-basal length . Zygomatic breadth Snout length Snout width at canines 5 Snout width at level of 2nd cheek tooth 7 Anterior breadth nasals Greatest length nasals Least interorbital width* Mastoid breadth 6 Length upper tooth - row i—m. 6. S < FALKLAND ISLANDS—continued 315 I949.3-17.26 1949.3.17-27 1949.3.17.28 1949.3.17.29 ——— oe marten, | WA enuesl, | OE mm % mm. % 213 214 215 219 209 100 210 100 210 100 214 100 118 56°5 117 55°7 114 5473 127 59°3 61g 20)-2 59 «628-1 6I 29-0 63 29°4 43 20-6 44 20°9 42 20:0 48 2274 36 4«=«17°2 S770 34) 516-2 390—S «18-2 27 LZ 9) Zone hss 2 ALLA 29 ©=©13°6 3r ) 14-8 3m 14°8 3r «414°8 3I =«14°5 31148 28 =13°3 27 12:9 34 «159 IIo 65526 II3 53°8 I0o6 =50°5 II5 53°7 80 38-3 83 39°5 80 38-1 81 37-9 1949-3-17-30 1949.3-17.31 1949.3-17-32 10949.3-17-33 Male SS HS. oO OO mm % mm % mm % mm % 222 218 214 215 218 100 212 100 212 100 209 100 122 55°9 115 54°2 116 54°7 II5 55°90 66 30°73 63 29°7 59 27°8 59 28-2 Azim 21-0 42) 19-8 BO) apsievil 40 19-1 4r 18-8 Sy AYES) 358 10-5 Bon a7 29 861373 28 13-2 2One L233 24 I1°5 3E «54-2 30 «6914-2 28° T3\2 3r 14°8 29 «13°3 35 165 26 12-3 29 © =13°9 Io8 §=49°5 Io2 48-1 102 48:1 Ioz2 48:8 85 38-9 84 39°6 83. 39°2 ap 1949-3-17-34 1949.3-17-35 10949.3-17.36 1949.3.17.37 ale Male a GH aa Tk Gao mm 6 mm We mm. % renters | OK 213 205 197 197 209 100 203 100 I9gI 100 193 I00 I1g 56-9 108 532 104 54°5 107 5574 61 29-2 59 «=029°1 54 28°3 5I 26-4 43 20-6 Bis) R89 37. 19°4 36 «18-7 Sum ned, 34 -16°7 sis E753} 33. 7 Et 25 i1:9 2A tT 8 23) 9520 23 tits) 520533 BPE) H5IS5) 28 14°7 29 15°0 27210 30 «=—-«14°8 3 eelon2 35), Tore 107 51:2 ror 49-8 gr 47-6 96 =49°7 8x = 38-8 78 38:4 76 39°8 7 aSOn8 * Posterior to supraorbital processes. 316 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA FALKLAND ISLANDS—continued Registered No. and 1949.3.17-38 1949.3-17-39 1949.3.17-40 1949.3.17.41 sex, if known t Male 7). ee : mm Oh mm a mm oF mm i Greatest length . A 2 Ley + 290 7) 187 . 180 Condylo-basal length . 5) ase aifofo) 7 87 00 . 184 100 = LG eeOO Zygomatic breadth . = 98) “Sask &) To2 } 545 100) 3 54°3 IOl 57:1 Snout length . See On 27 Oe AONE 2D 50.) 27-2) 48 9 -27°T Snout width at canines ESO) FEQ=-0) 9-3 Caen LOO) 34 «18°5 33. «18-6 Snout width at level of 2nd cheek tooth . a By aya 27, = TAA) 30) log 29 16°4 Anterior breadth nasals Ce PRE Vaan a Red srinditts) 23.) (12-5 22) eZ) Greatest length nasals Br 20nd 4 ee Se eeTONG. 28 815-2 27 1533 Least interorbital width* . 34 18:8 28 14°9 35 19-0 37 + 20°9 Mastoid breadth 5 - 86 47°5 93 49°7 88 47:8 86 = 48-6 Length upper tooth - row i—m. 6. E : mole Ve SOez cn) 170 3764 ne 720 soe OMEG OS Registered No. and 1949.3-17-42 10949.3.17.44 1949.4.17.46 1949.3.17.5I sex, if known Male a OR IND yo mm. % avin, 96 mimes eng Greatest length . ae ez9) ZA + 245 3 BB Condylo-basal Eanin o : - 225 100 5 as{o) atlefo) . 239 100 2) (207) SLOO Zygomatic breadth . oe) L245 5h Wy LAO) SONS. MAT! (W500) eel creme Snout length 5 mo 505) 280) ee 7 30-1 ee 725 30h a. sOS mmm OEO Snout width at canines dO e2OtA ee S222 Oley, | 5 2-year ee Snout width at level of 2nd cheek tooth . Be) ()06) 44 18:6 45 18°8 38) =«17'5 Anterior breadth nasals RZ) 902 4) eo 33 tot Ole eS hemeACo 27 2:4 Greatest length nasals ss. 5 oem SO) ESS. P sonl aLoeo 34 «15°7 Least interorbital width* . 24 10-7 23 9°7 30 12-6 26 I1-9 Mastoid breadth c =) SEO) (5 E60 126 53°4 136 56-9 10g 50°2 Length upper tooth - row m6, Z c ne S7a) 38570) (80) | 30n4y ee Ole 35-0 eos Sem Registered No. and 1949.3-17-53 1949.3-17-57 1949.3.17-58 1949.3.17-59 sex, if known ‘(a —\— =) (== a =) — oot OF | TMM ae mm. % TMG mm. % Greatest length . ‘A = g2 a LS e223 = kg Condylo-basal length . . 188 100 - 172 oo ; 220° 00 » 267 roo Zygomatic breadth . 2 og 7 1553). oz!) 50%3 9° a8) 536). oom Snout length . : a) ESA 28) ego) e2Ss) 165 205) amen Snout width at canines ~ BOnoeT By e2T5. 42) Oe 2 Own Snout width at level of 2nd cheek tooth . eye aks} ae Bue atsyOfy 5 YS a B)/7t 28 86168 Anterior breadth nasals 2) 24 L228 = —= Yo) aI OL 22) rare Greatest length nasals 5 3 ayIEe) 2, L597 32) Tals 25 14:9 Least interorbital width* . 35 18-6 35) 8 20:8, 5 32) TAs = — Mastoid breadth < = | O64, 5080 87 50:6) = 112 509 84 50-3 Length upper tooth - row i—m. 6. 0 5 Tn 73) Bors te 1667 38-44 Sqn ors. 160) esaer * Posterior to supraorbital processes. THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 317 FALKLAND ISLANDS—continued Registered No. and 1949.3-17.60 1949.3-17-62 1949.3-17-63 1949.3-17-64 sex, if known -—_ mm | YG mm % mm % mm. % Greatest length . 208 220 213 167 Condylo-basal ene 5 204 100 216 100 209 100 164 100 Zygomatic breadth 114 55'9 I23 56-9 120) 6057°4 04 «57°3 Snout length 560 «27°5 G5p gos 60 28-7 AS ezOs2 Snout width at canines 40 19°6 46 9 21°3 45 21°5 29 «4«17°7 Snout width at level of 2nd cheek tooth BA) tO 7, 39«s18-1 40 I9°1 27 16-5 Anterior breadth nasals 25 12:3 28 12-9 27, 120 17 10°74 Greatest length nasals 27) 03\ 2 36 «616-7 Bg 15-5 24 14°6 Least interorbital width* BAO, Ek ee ie} 200 L7e2 32 ©19°5 Mastoid breadth . 102 50:0 II2 51:9 107 51-2 85 51-8 Length upper tooth - row im. 6. : 82 40:2 86 39°8 82 39:2 64 39°2 Registered No. and 1949.3-17-65 1949.3-17.66 1949.3-17-67 1949.3. 17.68 sex, if known SS —.——_, - ———~ ;—— —_.——_,, mm. % LILI yo) mes mm. % Greatest length . 231 188 180 236 Condylo-basal pant : 226 100 184 100 177 100 230 100 Zygomatic breadth Tig 52°7 O7 52:7 tor 57:1 136 0659"! Snout length 66 29°2 Gp PENCE 47 26-6 Tes 338%} Snout width at canines 43 19:0 34 18-5 sy Ug) 5 8 Snout width at level of 2nd cheek tooth 36 «=«15'9 27) LAT, 28 15:8 42 18-3 Anterior breadth nasals = = 23) eZ 5 — = — = Greatest length nasals B5 me hi): 25 13:6 — = — = Least interorbital width* 200) i223 37 ZOLk 30 416-9 28 12-2 Mastoid breadth Lov 47-3 85 46-2 92 51°9 TrQ) 51-7 Length upper tooth - row im. 6. 5 89 6394 70 38:0 66 37°3 90 «39:1 Registered No. and 1949.3-17-69 1949.3-17-71 1949-3-17- 72 1949.3-17-73 sex, if known ————S$——— co A > mm 6 mm. % mm. % mm. % Greatest length . A 221 210 211 214 Condylo-basal length . 215 100 206 100 206 100 213 100 Zygomatic breadth 114 53°'0 117-508 115 55°8 TI9g 55°9 Snout length 65 30:2 59 28-6 60 291 64 30°0 Snout width at canines 48 22-3 41 19°9 40 19°4 45 21°! Snout width at level of 2nd cheek tooth ; 39 «18+ 35 16:9 33. «+16'0 37. +1774 Anterior breadth nasals 24 1-2 22 10-7 26 12-6 — —_ Greatest length nasals 2 13°5 31 15:0 3310 L050 31 14:6 Least interorbital width* Fy aX) 20m 1472 28 13-6 26 12:2 Mastoid breadth > = — 103. 50°00 107 51°9 II2 52:6 Length upper tooth - row im. 6. . 83 38-6 Hi Bae 77 37°4 86 40°4 * Posterior to supraorbital processes. ZOOL, 2, 10 18§ 318 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA FALKLAND ISLANDS—continued Registered No. and 1949.3-17-75 1949.3.17.80 1949.3.17.8I1 1950.11.14.1 sex, if known ? Male (ee SS mm % mm. % nita mm Th Greatest length . fs 5 aus) Aye PRIG/ LOT a les Condylo-basal length . aziz) LOO 5 Hing} ao\) - 194 100 > 218). 100 Zygomatic breadth . oe KEES 5420 ue TEES 554) tog 56:2 . 132 (60:6 Snout length . ; a) OZE 20r2) oN WO yi 235O00r | 5A 27 Oi |. Om Onr, Snout width at canines 7) 840 HSMON. (94g) 2052 36 =—«18-6 49 22°5 Snout width at level of 2nd cheek tooth . : 2 S57 eS LOSS ee 950M LOO) 30 «=15°5 41 18:8 Anterior breadth nasals i —— a. 420 VI2*2e 00) (25. (2°0), este le Greatest length nasals 5 = Se (38) et oesime. 27rd 3°01. SOME Least interorbital width* . 38 17:9 . 29 13:6 27 E39) 25) Sates Mastoid breadth 6 > £08) —50%0) = lro7, 502 IOI 52°1 125 57°3 Length upper tooth - row im. 6. 6 a LO ZnS Oe77ea en O53 OO mem OllmNS Ol 8 72 Registered No. and I1950.11.14.2 1855.12.26.167 1o13e sex, if known oF "7 o_O 7 paten, O mm % mm. §% Greatest length . ° 5 FAI) ez30) - 143 Condylo-basal length . =» 203, £00 5 ai) aKele) 2 L140 00 Zygomatic breadth . Be Os} 45) 6ne7 9. sS0r 56°77 Snout length 6 7 64, 30%0 Gf. | 3400, gh Ks ADOC) Snout width at canines BLK S009 B77 2a 20) ZO Snout width at level of 2nd cheek tooth . o - 38 17°8 48 20°4 27, TOs: Anterior breadth nasals a ehh) Bhieaae 30, 12-8 18 128 Greatest length nasals SS Se) AOW 720) (20a rAe Least interorbital width* . 28 13:1 . 32 13:6 32 «24-1 Mastoid breadth : eb — 4. 935 57°4 73 «-«51°8 Length upper tooth - row im. Gn - 6 = 5 teeth se.) 188) 37a) San 57 40ra! MESSIER CHANNEL, CHILE STRAITS OF MAGELLAN oF —_ uc es Registered No. and 1950.1I.14.4 I950.11.14.3 1879.8.21.5 1880.7.28.12 sex, if known Male Female Male aa De a Ga Cae mm vs Tam. 9/5 mm a mm Greatest length . 5 RK) =) 20% + 205 5 24740) Condylo-basal length . . 225 Too =) LO6) yr00 - 201 100 - 220 100 Zygomatic breadth . ) G2) SSL ieee ROMER OC 8 BIL2 S/5h 70 ta naa Snout length . o ] (66 920-20 nOOW sono 9 5605) 27-0) es "Go Snout width at canines 3) 55) etede sO Morg (Sa Pred eo: Snout width at level of 2nd cheek tooth . J ci 46 20:4 . 35 17°9 42. “ERRG 41 Anterior breadth nasals Se ASLAM USTs 3 hey) teioe 22 10°9 — Greatest length nasals oe siGhen Stcyolotias 30) er5ies 29— 4-4 — Least interorbital width* . 32 4:2 . 26 13:3 2 13°4 — Mastoid breadth 4 F pl28) 9 56-90. Tob) S40 I0oo)§= 49:8 124 Length upper tooth - row i—m. 6. 5 : OTe S57.) Wes SOO/a) -40Gim we 79% 83053) 85 * Posterior to supraorbital processes. THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 319 LOBOS TIERRA STRAITS OF MAGELLAN— ISLANDS, DEL continued URUGUAY FUEGO oo ——— —— Registered No. and 1880.7.28.7 I9Q19.7-7- 1947-7-16.4 1880.7.28.17 sex, if known 10090 Female — oF oOEO— oO (SS mm ye mm Ves mms 5 mien, IG Greatest length . é . 226 = L283 > 250 2 Sr Condylo-basal length . E223) £00 - 126 100 207.) (L00 . 129 100 Zygomatic breadth . 5 WO. Gees gp R\ OxjoGj! al gpa” Weike) 74 57°4 Snout length . 5 Pe Ones t-4ee me s4n 20-08 1 OA BORO 360 -27°9 Snout width at canines a eB iS) 26 20:6 48 23:2 2S alee Snout width at level of 2nd cheek tooth . 2 6 Zi | ZOVEO 2 19°8 42 20°73 ZONNZOR2 Anterior breadth nasal: ore sine PT SON 7a eLS5) en Ze ELD SD) 7a Sz Greatest length nasals A ESC LA Om peeLon 1 L493) Lier 32) PELE 5 20 15°5 Least interorbital widtht . 34 15:2 36 «6-28 -6 he 1/126) Sey y/eat Mastoid breadth A = 225) 50-2 66 52-4 113-546 67 51-9 Length upper tooth - row i.—m. 6. a : 5 Gi BORO a OR 7) ED BP GY cece) GALAPAGOS ISLANDS Gaz ie = Least interorbital width* . 2 11:8 28 14-0 2S 7, Mastoid breadth ; 2 126 59-7 _— _- —_ Length upper tooth - row i—m. 6. 9 2 86 40°8 . 80 40:0 . 75 38-1 * Posterior to supraorbital processes. The length measurements show that the mainland skulls tend to be rather smaller than Falkland skulls of approximately the same age. Photographs of the Galapagos skull show that it is considerably more adult in appearance than its small size would suggest. Skulls of similar condylo-basal length from the Falkland Islands are obviously from young animals (Pl. 11). They have no sagittal crest, practically no occipital crest and the zygomatic arch has not yet attained the great depth that it does in the old animal. Heller describes the Galapagos type skull as “‘ old adult ; sutures largely obsolete ; occipital and parietal crests high,” and the photographs show these characters of age, as well as the deep zygomatic arch. 320 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA The fact that there is tending to be a smaller mainland race of A. australis and a larger Falkland Island race is interesting in the light of a recent paper by Sivertsen (1953) on a new species Zalophus wollebaeki from the Galapagos Islands. Measure- ments of 20 skulls of this new species were compared with those of 21 Zalophus skulls from California. The mean condylo-basal length of the Galapagos skulls was 264 mm., with a maximum of 276 mm., while the corresponding figures for the Californian skulls were 292 mm. and 323 mm. The type of Z. wollebaeki, an old male, had a condylo-basal length of 267 mm., with a high sagittal crest and very worn teeth, while the only Californian skull of a similar age had a condylo-basal length of 323 mm. Thus while there can be no doubt that the fur seal on the Galapagos Islands is A. australis, the comparative measurements given here (Table I) make it seem probable that there is a race of smaller animals living on the islands. There is not at the moment sufficient material in the collection to be able to distinguish the Galapagos population from other communities living on or close to the mainland of South America, and it is for this reason that subspecific rank cannot be given to the Gala- pagos fur seals. The type locality for A. australis Zimmermann is the Falkland Islands, so that if this population is considered subspecifically distinct from that on the mainland, the name should be A. australis australis. Since the individual populations on the mainland cannot be separated subspecifically, the first available name for the whole is A. australis gracilis Nehring 1887, and should the Galapagos population eventually prove to be distinct A. australis galapagoensis is available. As the main purpose of the present paper does not principally concern the genus Zalophus, and pending the publication of Sivertsen’s more comprehensive paper on the Otariidae, Z. wollebaeki may be regarded as bearing the same relation to Z. californianus that the Galapagos Arctocephalus bears to A. australis. 2. Townsend's specimens During 1932 and 1933 several living fur seals were captured at the Galapagos Islands and sent to the San Diego Zoological Gardens. Three of these animals (an adult male, an adult female and a juvenile) died, and their skins and skulls form the subject of a paper by Townsend (1934) who referred them to Arctocephalus galapagoensis Heller. As Heller’s desciption of the type of this species, particularly of the skull is not very detailed, and he published no photograph or drawing of the skull, Townsend seems to have based his identification mainly on the locality. He says that he compared his skulls with those of Arctocephalus collected “ in the Straits of Magellan and on the beaches at the Galapagos and from Guadalupe,” but gives no details of this comparison. A few measurements of the Arctocephalus from the Straits of Magellan are inserted in a list of measurements of the male skull from the Galapagos, but no reference is made to these in the text. As the animals described by Townsend are fur seals of the genus Arctocephalus the photograph of the skull of one of them has been compared with those of A. australis and A. philippi, the only fur seals known to occur off the coast of South America. It is at once obvious from the general similarity, and in particular from THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA | 321 the broad snout, wide palate, short nasals and deep zygomatics that the animals received by Townsend from the Galapagos are Southern fur seals, Arctocephalus australis. Townsend gives the length of the male carcase as 4 ft. 6 in. from the tip of the nose to the tip of the tail, and the greatest basal length of the skull as 212 mm. Townsend’s measurements of the male skull are given together with those of a skull of A. australis (1880.7.28.7) in the British Museum collection (Table II). It is evident from the table that there is a close similarity between the two skulls. The two measurements that show most divergence are those involving interorbital width. With regard to these it may be said that the British Museum collection includes specimens of length comparable to that of Townsend’s with smaller inter- orbital width, and the variability of this part of the skull is striking. TABLE I. (See notes on measurements used) Measurements of A. australis male skull given by Townsend 1880.7.28.7 % % Greatest length 6 ‘ : 7 : 212 ¢ 226 Condylo-basal length 5 4 é 0 208* 100 A 223 100 Basal length (gnathion-basion) . 5 “ 202 97°1 214 95°9 Basilar length (basion-henselion) : : 198 95°2 209 93°7 Palatal length (gnathion-palation) . 0 99 47°6 103 46-2 Zygomatic breadth . : 4 132 63°5 146 65°5 Canine to last upper molar (inclusive) : 68 32°7 70 31-4 Distance between upper canines (internally) 26 I2°5 25 II-2 Distance between 3rd upper molars (in- ternally) . : 27 12-9 33 14°8 Interorbital width (anterior to supraorbital process) 27 12-9 5 40 17'9 Interorbital width (posterior to supraorbital process) c : 26°5 12-7 34 15:2 Width of sapracrbital processes 3 6 46 22:1 58 est. 26-0 Greatest length nasals_. : : o 29 13-9 33 14°8 Ant. breadth nasals . ‘ 5 25°5 12:3 31 13:9 Breadth rostrum at 2nd molar . a 2 44 21-2 , 46 20-6 Mastoid breadth . 3 5 c : 117 56°3 : 125 56-1 * Condylo-basal length estimated as it is not given by Townsend. SPECIFIC IDENTITY OF ARCTOCEPHALUS TOWNSENDI MERRIAM The other fur seal of the genus Arctocephalus from the American coast, A. philippi, was described in 1866 by Peters from the skull of an adult (but not old) male animal collected on Juan Fernandez in December, 1864, by Dr. Philippi. Full size drawings 322 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA EE as oe iO cMS, Fic. 1. Type skull of Arctocephalus philippi from Peters, Monatsb. Akad. Berlin, 1866, pl. 11. THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 323 of the skull were given (Fig. 1) from which measurements have been taken for comparison. The most important skull characters for distinguishing the species are given by Peters as follows : I. The palate is deep in front and shallow behind, and is twice as broad between the last molars as between the canines. 2. The posterior ends of the maxillae are prolonged into small hooks.} 3. The maxillary root of the zygomatic arch is very broad antero-posteriorly. 4. The tympanic bullae are flat. 5. The mastoid process is long, but does not jut out ventrally. 6. The zygomatic arch is narrow. 7. The lower jaw has no pronounced angle ; the coronoids are rounded and project backwards. 8. The peculiar pointed form and the lack or weak development of accessory cusps on the molar teeth, which are separated from each other by relatively large spaces. All these characters are visible both in Peters’ drawings of the type skull and in a skull of this species from Juan Fernandez in the British Museum collection (Reg. No. 1883.11.8.1.). A. philippi, the type locality of which is Juan Fernandez, is a fur seal, as is evident from Peters’ description of the long overfur and the thick underfur. The only other animal with which it may be confused is the Southern fur seal, A. australis, which ranges from Uruquay, round the Straits of Magellan and along the west coast of South America. Externally A. philippi may be distinguished from A. australis chiefly by its long tapering snout which is very unlike the short, rather upturned snout of A. australis. When seen side by side skulls of the two species show many differences. The skull of A. philippi (Fig. 1) is more finely built than that of A. australis (Fig. 2). In general shape it is long and slender, as are also the nasal bones ; and the zygomatic arches are narrow in lateral view. The palate is also narrow and is very concave between molars 1-3. The skull of A. australis is much more robust, the nasals are short and broad, the palate is not particularly narrow or concave and the zygomatic arches are strong and deep. These differences can easily be seen in the drawings of the two skulls (Figs. 1 and 2). In 1870 the Museum at Santiago was presented with a male, a female and a young seal from the island of Masafuera. These were considered to be a new species and formed the subject of a paper in 1871 in which they were called Otaria argentata (Philippi in Peters, 1871). Philippi gives a drawing of the skull of this new species and a list of characters in which it differs from A. philippi. The skin of O. argentata is said to differ chiefly by its lighter colour and by the guard hairs of the neck being shorter than those of A. philippi. The measurements given of the stuffed male show it to be a young animal as it is only 3 ft. 9 in. from nose to tail, and Philippi himself says that the skull is not fully grown as the crests are not developed. The length of the neck hairs and many of the characters of the skull used by Philippi to distin- guish the species may be explained by the youth of the specimen, and the drawing 1 Not confined to this species 324 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA : lo cms, Fic. 2. Orthographic projection of skull of Arctocephalus australis, 1880.7.28.7. THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 325 of the skull gives proof that the animal is another specimen of A. philippi. Peters, in his discussion of Philippi’s letter, says that if it had not been for Philippi’s insistence that O. argentata and A. philippi were different, he “‘ would have thought twice about describing them as two different species.”’ In 1892 an expedition sailed from San Diego to Guadalupe for the purpose of identifying the fur seals present on the island. Seven seals were seen and one was shot, but sank. Four skulls were picked up on the beach and one of these forms the type of a new species Arctocephalus townsendi which was described by Merriam in 1897. Theskull is that of an adult male No. 83617 U. S. National Museum. Merriam compares this type “ with skulls of Avctocephalus (australis or philippi) from the Galapagos Islands,” and from the characters that he gives it is evident that he com- pared his skull with A. australis and had not seen a skull of A. philippi. Fic. 3. Type skull of Arctocephalus townsendi, U.S. Mus. No. 83,617. Taken from Allen, 1905, Patag. Exp. III; pl. xviii. Merriam lists the most important characters for identifying the skull of A. townsendi as follows : 1. The exceedingly narrow and excavated palate. 2. Flat tympanic bullae. 3. Short and thick ascending arm of premaxilla. 4. Broadly expanded zygomatic root of maxilla, forming a floor under the anterior half of the orbit. 5. The 5th molar mainly posterior to the plane of the anterior root of the zygoma. 326 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA When these characters are compared with those given by Peters for the type skull of A. philippi it will be seen that Merriam’s Nos. 1, 2 and 4—important diag- nostic features—correspond exactly with Nos. 1, 4 and 3 respectively of Peters’ description. Merriam’s character No. 3 is not visible in the ventral view of the skull which he gives. His fifth character, while it is good for the specimen of A. philippi in the British Museum collection, is not distinctive. When a further comparison is made between the measurements and proportions of the type skull of A. Philippi, the skull of this species in the British Museum, and the type skull of A. townsendi (Table III) it will be seen that these three skulls are so similar that it is probable that they all belong to the same species of Arctocephalus. Comparison of the drawings (Figs. 1 and 3) of these skulls also confirms this view, and it is thus possible to add Arctocephalus townsendi Merriam to the synonymy of Arctocephalus philippi Peters. Allen (1905) described skulls from the Galapagos which he called A. philippi, and combined a very “ free translation ” of the characters given by Peters with some which are obviously only from the skulls he was describing. His figured skulls show the typical high sagittal crest of the male Zalophus and it is clear that he has mistakenly identified these skulls as A. Philippi. Osgood (1943) says that Remington Kellogg and G. M. Miller examined Allen’s skulls from the Galapagos and found that they were undoubtedly Zalophus californianus. TABLE III. (See notes on measurements used) A. townsendi A. philippi A. philippi type type 1883.11.8.1 mm, OG mm. Of mm. 263 Greatest length : ; 3 250 230 - 255 Condylo-basal length 5 0 oS 100 =) 5233 100 7 248 100 Zygomatic breadth . c : . I51 59-9) 0) 229 55°4 139 56:0 Snout length . é : : a yiele) 2A OS 27°0 73 29°4 Snout width at canines . 153 ZT OW nA 5 19°3 45 18-1 Canine to last upper molar (inclusive) 88 3409) 175 32°2 he 30-7 Distance between 3rd upper molars (internally) . : 22°5 Gio@) ne 970. 19 U7 Antero-posterior width of zygomatic root of the maxilla 5 21 Sige 20. SeOl ee eeee) 8-9 Interorbital width (posterior to eceee orbital process) . . . 730 PCI) 2 30 12'9 31 I2°5 Mastoid breadth . : c (PLES) 52°8 115 49°4 125 50°4 Greatest length nasals : = — 41 17°6 45 18-1 Anterior breadth nasals_ . 5 : — = ns 26 Il-2 25 Io'l * Estimated. t Measurements taken from 1/1 drawing of type skull in Allen (1905). Other measurements as given by Merriam (1897). { Peters’ type has only 5 upper molars. Measurements of skull taken from 1/1 drawing in Peters (1866). THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 327 DISTRIBUTION AND PARTIAL RECORDS OF A. PHILIPPI ON THE PACIFIC COAST OF AMERICA The records of the occurrence of A. philippi which are listed below are based partly on the identification of actual specimens, and partly on records of “‘ fur seals - from the Californian Islands which are, according to Starks (1922) A. townsend: = A. philippi. 1. Juan Fernandez (a) 1864. Type specimen killed (Peters, 1866). (6) — “Some years ago,” possibly in 1864, a few dozen skins taken by Philippi (Philippi in Peters, 1871). (c) 1883. Skull of specimen in British Museum collection. Presented by Chilean Government. (d@) — Skull in Santiago Museum. No locality given (Philippi in Peters, 1871). 2. Masafuera (a) 1870. 3 seals collected. One of these described as type of O. argentata (Philippi in Peters, 1871). 3. Chonos Islands (a) 1871. Askin taken by Philippi’s assistant (Philippi in Peters, 1871). 4. San Benito Islands (a) 1806. 8,338 skins taken by the “ Port au Prince ” (Townsend, 1924, 1931). 5. Guadalupe (a) 1878. A few hundred skins taken from Guadalupe and Santa Barbara (Starks, 1922). (b) 1880. Fur seals tightly packed (Merriam 1897). c) 1883. 2,000 killed, but commercially extinct (Merriam, 1897). d) 1891. Only a few seen (Merriam, 1897). (e) 1892. 7 seals seen and 4 skulls obtained from which type of A. townsendi described (Merriam, 1897). (f) 1922. None seen (Anthony, 1925). (g) 1922. 2 year old female sent to San Diego Zoo (Starks, 1922). (h) 1928. 2 males caught and sent to San Diego Zoo, and 60 seals estimated to be on the island (Townsend, 1931). (i) 1929. Some fur seals seen in May (Townsend, 1930). (j) 1950. None seen (Bartholomew and Hubbs, 1952). 328 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 6. Santa Barbara Islands (a) 1878. A few hundred skins taken from Guadalupe and Santa Barbara (Starks, 1922). (b) 1890. 5 seals taken off San Miguel (Starks, 1922). (c) 1901. 3 seals killed on Santa Cruz (Starks, 1922). (d) 1949. Single adult male seen (Bartholomew and Hubbs, 1952). 7. Ventura, California (a) Eighteenth to nineteenth centuries. Abundant remains found in shell mound (Lyon, 1937). 8. Farallone Islands (a) 1810-1812. 73,402 skins taken by Americans (Starks, 1922). (b) 1812-1840. Russians took 1,200-1,500 skins annually (Starks, 1922). (c) 1818. Seals diminishing rapidly (Starks, 1922). (d) 1824. Seals profitable again, over 1,000 skins taken, but decreased to 54 skins in 2 years (Starks, 1922). ~— The first record of a relatively recent occurrence of the fur seal is the finding of skull remains from the eighteenth to nineteenth centuries in a shell mound at Point Mugu, Ventura County, California, 1,557 identifiable bones of this seal were found and it was by far the most abundant animal in the collection. No more records are available until 1806 when over 8,000 skins were taken from the San Benito Islands. The seal was abundant on all the islands off the Californian coast and was taken for its fur in great numbers until about 1883 when it was considered extinct commercially. From that time until the present day there have been only a few sporadic noted occurrences, the last being an adult male seen on San Nicolas Island in 1949. As there are so few records of A. philippi being seen on Juan Fernandez and Masafuera it seems probable that the normal habitat of this seal is the islands off the Californian coast. When it is known to have been so plentiful in the northern part of its range it is noteworthy that there have been so few records of its occurrence on Juan Fernandez, although this may be attributed partly to the remoteness of this island compared with those off the Californian coast. It is probable that Juan Fernandez represents the most southerly extent of its distribution and that there has never at any time been a great number of A. Philippi on the island. The presence of a small number of A. philippi would be masked by the large numbers of A. australis which undoubtedly occurred on the island, and the differences would not be noticed by the sealers who would class them all as “ fur seals.’ It was only scientifically in- terested people like Dr. Philippi who collected the type which was named after him, and the chemist Fr. Leyboldt who collected the type of O. argentata who noticed the difference between the two species. The present state of A. Philippi is unknown. It may be extinct, but since an adult male has been seen as recently as 1949 it seems possible that there may still be a small colony left. THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 329 Map to show seal-inhabited islands off the Pacific American coast. 1. Chonos Islands. 8. Santa Barbara Islands. 2. Juan Fernandez. San Nicolas. 3. Masafuera. Santa Cruz. 4. St. Felix, St. Ambrose, St. Mary’s Islands. Santa Rosa. 5. Galapagos Islands. San Miguel. 6. San Benito Islands. g. Ventura. 7. Guadalupe. 10. Farallone Islands. 330 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA PARTIAL RECORDS OF ‘‘ FUR SEALS”’ OFF THE COAST OF 1. Masafuera (a) 1563. (b) 1792. (c) 1797- (4) 1798. 1807. 1824. 18o1. ~~ ~ pa US NT SOUTH AMERICA Island discovered. Swarming with fur seals (Allen, 1899). Capt. Stewart of the “ Eliza’’ took 38,000 skins to Canton and sold them for 16,000 dollars (Allen, 1899). Americans came about this time. 2-3 million seals on island. 3 mil- lion taken to Canton in 7 years. People from 14 ships on island at the same time, all killing seals (Allen, 1899). Capt. Fanning of the “‘ Betsy’’ took 100,000 skins to Canton and estimated that when he left there were still 500,000-700,000 seals on the island (Allen, 1899). Capt. Morrell—business scarcely worth following (Allen, 1899). Capt. Morrell—island almost entirely abandoned by seals (Allen, 1899). Capt. Gaffney saw 300-400 and took 19 (Allen 1899). 2. Juan Fernandez (a) 1683. Dampier—seals thick about the island (Allen, 1899). (6) 1687-90. Capt. Davies of the “ Bachelors Delight ’’ left some men on the (c) 1741. (d) 1800. (e) 1891. (f) 1931. island to cure seal-skins (Albes, 1914). Anson—many seals and sea-lions (elephant seals) (Anson, 1744). Capt. Delano—human population of 3,000 on island, so no seals left (Allen, 1899). Capt. Gaffney saw a few fur seals in December (Allen, 1899). Nybelin got a parasite from an A. australis (Nybelin, 1931). 3. St. Felix, St. Ambrose, St. Mary’s Islands (a) 1792. (b) 1801. (c) 1816. (d) 1891. 13,000 skins taken during August and September by American ship “ Jefferson ” (Howay, 1930). Islands were visited by sealers and there must have been large numbers of seals (Allen, 1899). Capt. Fanning took 14,000 skins at St. Mary’s (Allen, 1899). No seals worth mentioning (Allen, 1899). 4. Galapagos Islands (a) 1535. (b) 1800. (c) 1816. (d) 1825. (e) 1872. Seals mentioned by discoverer of Galapagos (Baur, 1897). Large numbers of fur and hair seals (Allen, 1899). Capt. Fanning took 8,000 fur seals and 2,000 hair seals (Allen, 1899). Capt. Morrell took a few fur seals from Albemarle (Allen, 1899). Capt. Reed took 3,000 fur seals and about as many more during three subsequent voyages between 1872 and 1880 (Allen, 1899). THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 331 ) 1872. Hassler Expedition. Specimens of A. australis collected (Allen, 1880). (g) 1885. Capt. Gaffney took 1,000 fur seals (Allen, 1899). (h) 1898-99. Capt. Noyes—seals not very numerous, took 224 skins (Allen, 1899). Heller collected type of A. galapagoensis on this voyage (Heller, 1904). ) 1923. Beebe saw only two fur seals (Beebe, 1924) (j) 1931. Sea-lions and seals numerous and quite tame (Korwin, 1931). (k) 1932-33. Capt. Hancock caught several fur seals which were presented to San Diego Zoo (Townsend, 1934). As even the incomplete list above shows, there is quite a number of records of sealers taking fur seals from the islands off the South American coast from the Galapagos Islands southwards. The story of these fur seals is the usual one of uncontrolled exploitation. Millions of skins were taken until the middle of the nineteenth century when the numbers dropped rapidly and for about the last 40 years not only are there no records of any commercial sealing, but there are very few referring to seals at all, and even these indicate that there cannot be many fur seals on the islands. It appears from the available records that the sealing expeditions worked north- wards along the west coast of South America, as seals at Masafuera and Juan Fer- nandez were abundant until about 1798, at St. Mary’s until 1816, while there were still considerable numbers at the Galapagos as late as 1885. There is still some doubt as to the identity of these “ fur seals.” Townsend (in Allen, 1899) refers to the Galapagos seal as A. philippi, but does not give any reasons for his identification. The only other records where specific names are used are those of Allen (1880), Heller (1904), Nybelin (1931) and Townsend (1934) for seals on the Galapagos and Juan Fernandez and in all four instances A. australis is the animal named. Both A. australis and A. philippi have been reliably recorded from Juan Fernandez, and A. australis from the Galapagos, so it is not unlikely, though not proven, that A. philippi also occurs on the Galapagos. The bulk of the records of “‘ fur seals”” from these South American islands probably refer to A. australis although the scarcity of records makes it impossible to be certain. For the same reason it is not known what the present status of this seal is along the western South American coast. Other Otariids frequent the shores and islands of the Pacific coast of America. Of these the best known is the Californian sea-lion Zalophus californianus. The most southerly point from which it has been recorded is the Galapagos Islands. Beebe (1924) went to these islands in 1923 and makes several references to the sea- lions that he saw there. He identifies them as “‘ Southern sea-lion, Otaria jubata,”’ but the photographs he gives show the short pointed nose of Zalophus and not the heavy upturned snout of O. byronia. Wollebaek (1927) went to these islands in 1925, and although he does not mention the seals by name he gives a photograph which is undoubtedly of a Zalophus. Skulls of Zalophus from the Galapagos which are in the U.S. National Museum are figured by Allen (1905) although he wrongly calls them A. philippi. Zalophus is well known on the islands off the Californian coast and probably extends northwards as far as the bay of San Francisco. 332 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA Steller’s sea-lion, Eumetopias jubatus, is known to extend its range as far south as San Nicolas Island, although there are no records of it breeding any further south than Santa Rosa Island, 42 miles north of San Nicolas. The Northern fur seal Callorhinus ursinus, although it breeds mainly o on the Pribilof Islands, spends the winter and early spring at sea and may migrate southwards to the latitude of California, although it rarely comes ashore there. The distribution of the Southern sea-lion, Otaria byronia, corresponds with that of A. australis. Both Heller (1904) and Allen (1880) note its presence on the Gala- pagos Islands and it is known on Juan Fernandez, round Cape Horn and along the eastern coast of South America as far as Lobos Island, Uruguay. SUMMARY OF THE MAIN CHARACTERS AND DISTRIBUTION OF THE OTARIIDS OF THE PACIFIC COAST OF AMERICA Otaria byronia Blainville, 1820. Southern sea-lion. SizE. Length of adult male from nose to end of tail 6-7 ft., and of female 5-6 ft. Weight of adult male over 1,500 Ib. DEscRIPTION. Males and females similar in colour, back usually dark brown, mane of male and neck of female, and the belly dark yellow. Face dark and hair of flippers reddish. Bare parts of skin black. Whiskers long, reaching 12 in. in adult male, about 30 on each side, straw coloured and forming a pale moustache. Variations in colour frequent, and many paler animals found. Newly-born pups practically black, soon fading to chocolate, and after first moult at a few months old, a dark grey. Male with extremely thick and heavy neck thrown into folds, this together with the lighter and thicker hair on the neck gives the appearance of a mane. Female slighter in build. Head short and muzzle deep and upturned. DistRiBuTION. From Lobos Island, Uruquay to Straits of Magellan, Falkland Islands, Galapagos Islands. SKULL. Condylo-basal length up to 358 mm. in adult males, and 267 mm. in adult females. Easily distinguishable from skulls of all other seals by the greatly elongated palate extending backwards as far as the pterygoids, and becoming progressively more concave posteriorly. Adult male skull with well-developed sagittal and parietal crests, and various processes on the parietal for the attachment of muscles. : I 6 DENTITION. oe c.- m.-. 2 I 5 Eumetopias jubatus (Schreber 1776). Steller’s sea-lion. Size. Length of adult male from nose to end of tail 11-13 ft. and of adult female 8-9 ft. Weight of adult male 1,000-1,200 lb., of adult female 400-500 Ib. DeEscRIPTION. Both males and females a light reddish brown ; slightly darker on the belly. Colour varying with age and season and coat lighter when just moulted. Naked parts of skin black. Whiskers long and slender, the longest about 20 in., white or brownish white. Adult males with mane on neck. The largest of the eared ¥ . THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 333 seals. Has a rather bear-like head with a short straight nose, not upturned as in O. byronia. Muzzle and neck very heavy. DisTRIBUTION. Shores of the north Pacific from Behring Straits southward to Santa Barbara Islands, California, and Japan. SKULL. Condylo-basal length up to 380 mm. in adult males. Skull distinguished by the anteriorly placed quadrate supraorbital processes and, in all except young skulls, a large gap between the upper 4th and 5th molars. Palate with hinder end contracted and truncate. I DENTITION. is C.- m2 2 I 5 Arctocephalus philippi (Peters 1866). “ Juan Fernandez ”’ fur seal. SizE. Length of adult male from nose to end of tail ca. 5 ft. Adult female probably slightly smaller. Description. Above blackish grey, more yellowish grey on head and neck ; brownish black below, proximal part of limbs, lips and chin rusty brown. Overhair rusty brown with black tips, thick underfur rust red. Whiskers in six rows, some black, some white, and some white with black tips. Animal distinguished by slender tapering snout. DIsTRIBUTION. Juan Fernandez, Guadalupe, Santa Barbara Islands, Farallone Islands. SKULL: 1. Skull narrow, the mastoid breadth being 48-50% of condylo-basal length. 2. Palate narrow and very deep in front. . Maxillary root of zygoma very wide. Flat tympanic bullae. Narrow zygomatic arch. Snout narrow. . Posterior prolongations of palatines not thickened. . Teeth usually without accessory cusps and widely spaced. Measurements of few authentic specimens known but condylo-basal length of adult male skull ca. 256 mm. OWI AAR Y I molar teeth of Arctocephalus rather variable). DENTITION. id Ae me. (Peters’ type skull has only m.2 but number of Arctocephalus australis (Zimmermann 1783). Southern fur seal. S1zE. Length of adult male from nose to end of tail 5 ft. 6 in., and of adult female 4 ft. DESCRIPTION. Overhair black, tipped with grey except on belly where blackish brown. Hairs 1-2 in. long on back of neck but shorter over rest of body. Underfur reddish brown. Dorsal surface of manus and pes covered up to nails with black hairs. Tip of nose and naked skin of limbs black. Whiskers white, though some of the smaller are greyish black, about 20 on each side. Very few photographs 334 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA available, but appears to be rather like Callorhinus in the external shape of the head, except that the nose is rather longer and slightly upturned. DISTRIBUTION. Shores and islands of South America from Lobos Islands, Uru- guay, to the Galapagos Islands, Straits of Magellan, Falkland Islands, S. Georgia. SKULL. Condylo-basal length up to about 250 mm. in adult male. Skull squarish in general shape, heavily built, with thick zygomatic arches and short broad nasals. Bt shun 6 CoS ie DENTITION. i= 7. 2 I 5 Callorhinus ursinus (L.). Northern fur seal. Size. Length of adult male from nose to the end of tail 7-8 ft. and of adult female 4ft. Weight of adult male 500-600 lb., of adult female 80-100 |b. DESCRIPTION. Males nearly black on the back and brownish ventrally, neck and shoulders greyish. Limbs reddish brown. Naked areas of skin black. Females lighter than males. being grey dorsally, but otherwise like males. Whiskers white, or with brownish tips. Black in young animals. Distinguished externally by the high forehead and extremely short pointed nose. DISTRIBUTION. From Pribilof Islands south to California and Japan. SKuLL. Condylo-basal length up to ca. 250 mm. in adult males. Interorbital region long, facial region very short and high, descending abruptly. Dentition weak. : I 6 DENTITION. He c.- m-. 2 I 5 Zalophus californianus (Lesson 1828). Californian sea-lion. S1zE. Length of adult male from nose to end of tail 8 ft., and of adult female 6 ft. Weight of adult male 500-600 lb. DESCRIPTION. Not so heavily built as Otaria or Eumetopias with not such a thick neck in adult males. DISTRIBUTION. Galapagos (but see Sivertsen’s (1953) account of Z. wollebaekt), shores and islands of California north to Farallone Islands. SKULL. Condylo-basal length up to 320 mm. in adult males. Skull of adult males easily distinguishable by the very high sagittal crest. Skulls of females and young males slender with elongated nasals and facial regions. 4 I 6 DENTITION. i2 c.- m.-. 2 I 5 KEY FOR DISTINGUISHING SKULLS OF PACIFIC AMERICAN OTARIIDAE 1. Skull with supraorbital processes and alisphenoid canal . c : . Otariidae 2. Posterior end of floor of palate very concave and extending areas as far as the pterygoids a ¢ . O. byronia 2a. Posterior end of floor of palate aoe peerings backwards as far as the pterygoids 3. 3. Supraorbital processes quadrate. Molars 5/5 with We gap between upper m. 4 and m. 5 in adult skulls : 4 . £. jubatus i THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 335 3a. Supraorbital processes triangular. Molars 6/5 with no large gap between upper m. 4 and m. 5 in adult skulls 4. 4. Nasals long and slender, combined widths at anterior Bade ca. yo-so%% of fener 5. 4@. Nasals short and broad, combined widths at anterior ends ca. 80-90% of length 6. 5. Palate very narrow, the internal distance between m. 1-3 being the same. Floor of palate very concave between these molars. Tympanic bullae smooth and rounded. No very high sagittal crest in old males . A. philippi 5a. Palate wider, the distance between the molars increasing gradually. Floor of palate only very slightly concave. Surface of tympanic bullae irregular. Very high sagittal crestin old males . 2 . Z. californianus 6. Interorbital region long in adult animals, ca. 20% of condylo-basal length. Snout, at level of m. 2, practically as high as it is long ; . . C. ursinus 6a. Interorbital region less than 20% of condylo-basal length in adult animals. Snout much longer than itishigh . A. australis t@. Skull without supraorbital processes or alisphenoid canal (not ponaderedi in this key) Phocidae Note: The above characters apply mainly to adult skulls. SUMMARY After comparison with skulls of fur seals known to occur along the Pacific coast of America, the type skull of Arctocephalus galapagoensis, which was collected on Wenman Island in the Galapagos group in 1898-99, is shown to be similar to skulls of A. australis. Skulls of fur seals brought back from the Galapagos Islands in 1932-33 are also identified as A. australis. A statistical treatment of proportions of skull measurements of A. australis from the Falkland Islands on the one hand, and the mainland of South America and the Galapagos Islands on the other, makes it seem probable that the Falkland Islands animals belong to a larger race. The shortage of specimens makes it impossible at the moment to distinguish the Galapagos skulls from those of other mainland popu- lations. It is‘shown by a comparison of skull drawings, measurements and proportions that the Guadalupe fur seal A. townsendi is the same as the Juan Fernandez fur seal A. philippi. A chronological list is given of the recorded occurrences of this seal and it is suggested that its normal habitat is the islands off the coast of Cali- fornia. The few records from Juan Fernandez represent animals taken at the most southerly point of the range, but its abundance tere was probably masked by the greater numbers of A. australis. A list is given of records of un-named “ fur seals’ from the Galapagos Islands southwards. It is probable that most of these records refer to A. australis. A brief account is given of the distribution of Z. californianus, E. jubatus, C. ursinus and O. byronia, the other Otariids which occur along the coast. A summary is given of all the Otariids occurring along the Pacific coast of America, and also a key for distinguishing their skulls. ACKNOWLEDGMENTS. I wish to express my indebtedness to Professor G. S. Myers and Mr. J. M. Savage of Stanford University, California, for the assistance they have given me by supplying 336 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA information and measurements relating to the type skull of A. galapagoensis, and also to Dr. Antenor L. de Carvalho, Museu Nacional, Rio de Janeiro, who took the photographs of this skull reproduced in the text, while working in Stanford University. NOTES ON THE MEASUREMENTS USED 1. Greatest length—from the posterior surface of the occipital condyles to the anterior surface of the knob formed by the premaxillae, above the incisors. 2. Condylo-basal length—from the posterior surface of the occipital condyles to the anterior surface of the premaxillae at the level of the incisors. 3. Snout length—from the most anterior part of the edge of the orbit to the anterior surface of the premaxillae at the level of the incisors. 4. Antero-posterior width of the zygomatic root of the maxilla—between the inferior lip of the antorbital foramen and the orbit. 5. Gnathion—is defined by Thomas (1905) as the most anterior point of the pre- maxillae, on or near the mid line. In Arctocephalus the most anterior point of the premaxillae is a projecting knob of very variable size, so the gnathion is taken here as the most anterior surface of the premaxillae at the level of the incisors, and does not include the knob. 6. Basion—as defined by Thomas (1905), a point in the middle line of the hinder edge of the basioccipital margin of the foramen magnum. 7. Henselion—as defined by Thomas (1905), the back of the aveolus of either of the median incisors. 8. Palation—as defined by Thomas (1905), the most anterior point of the hinder edge of the bony palate, whether in the middle line or on either side of a median spine. REFERENCES AuBEs, E. 1914. The Island of Juan Fernandez. Bull. Pan American Union, 39 : 201-216. ALLEN, J. A. 1880. History of North American Pinnipeds. U.S. Geol. and Geog. Survey of Territories. Misc. Pub. No. 12. 1899. in Jordan, D. S., The Fur Seals and Fur Seal Islands of the North Pacific Ocean. Washington, pt. III. 1905. The Mammalia of Southern Patagonia. Rep. Princeton Univ. Exp. to Patagonia, I pts L: Anson, G. 1744. An authentic and genuine journal of Commodore Anson's Expedition. Pub- lished from the journals of several of the principal officers who attended him throughout the whole Expedition. AntHony, A. W. 1925. Expedition to Guadalupe Island, Mexico in 1922. The Birds and Mammals. Proc. Calif. Acad. Sci. 4th ser., 14 : 277-320. BarRTHOLOMEW, G. A. 1950. A male Guadalupe Fur Seal on San Nicolas Island, California. Journ, Mamm. 31; no. 2; 175-180. BarTHoLomew, G. A., and Hupss, C. L. 1952. Winter Population of Pinnipeds about Guadalupe, San Benito and Cedros Islands, Baja California. Journ. Mamm. 33: no. 2, 160-171. Baur, G. 18097. Science, N.Y. 5: 956-957. BEEBE, W. 1924. Galapagos. World’s End. xxi-+ 443, 82 figs. 8 pls. New York: Putnam, 3 THE OTARIID SEALS OF THE PACIFIC COAST OF AMERICA 337 HELLER, E. 1904. Mammals of the Galapagos Archipelago, exclusive of the Cetacea. Papers from the Hopkins Stanford Galapagos Expedition, 1898-99. Proc. Calif. Acad. Sci. 3rd ser., 3: no. 7, 233-250. Howay, F. W. 1930. A list of trading vessels in maritime fur trade, 1785-1794. Tvans. Roy. Soc. Canada, Toronto, sect. 2, ser. 3, 24: I1I—134. Korwin, H. 1931. The Galapagos Islands, or Colon Archipelago. Bull. Pan American Union, 65 : 1140-1145. Lyon, G.M. 1937. Pinnipeds and a Sea Otter from the Point Mugu Shell Mound of California. Pub. Univ. Calif. Los Angeles Biol. Sct. 1: no. 8, 133-168. Mayer, W. V. I949. Catalogue of type specimens of Mammals in the Natural History Museum of Stanford University. Proc. Calif. Zool. Club, 1: no. 6, 29-32. Merriam, C. H. 1897. A new fur seal or sea-bear (Arctocephalus townsendi) from Guadalupe Island, off Lower California. Pvoc. Biol. Soc. Washington, 11 : 175-178. Neurinc, A. 1887. Ueber eine Pelzrobben-Art von der Kiiste Siid-Brasiliens. Arch. fir Naturges. 1 : 75-94. NYBELIN, O. 1931. Saugetier- und Vogelcestoden von Juan Fernandez. The Nat. Hist. of Juan Fernandez and Easter Island. Ed. by Dr. C. Skottsberg. III: Zoology, pt. IV, 493-523. Oscoop, W. H. 1943. The Mammals of Chile. Pub. Field Mus. Nat. Hist. Chicago, Zool. Ser. 30. PeTERS, W. 1866. Uber die Ohrenrobben (Seelowen und Seebaren), Otariae, insbesondere liber die in den Sammlungen zu Berlin befindlichen Arten. Monatsb. Akad. Berlin, 261-281. 1871. Uber eine fiir Chile neue Art von Ofavia und schloss daran einige Bemerkungen iiber die Verschiedenheit der in dem atlantischen und stillen Ocean vorkommenden Pelz- tobben. Monatsb. Akad. Berlin, 558-566. SIVERTSEN, E. 1953. A new species of Sea-Lion, Zalophus wollebaeki, from the Galapagos Islands. Det. Kong. Norske Videnskabers Selskabs Forhandlinger, 26: no. 1, 1-3. Starks, E.C. 1922. Records of the capture of Fur Seals on land in California. Calif. Fish and Game, 8: no. 3, 155-160. TownsEND, C. H. 1924. The Northern Elephant Seal and the Guadalupe Fur Seal. Nat. Hist. N.Y. 24: no. 5, 567-578. 1930. Guadalupe Fur Seal in 1929. Bull. N.Y. Zool. Soc. 33: 32. 1931. The Fur-seal of the Californian Islands. Zoologica, N.Y. 9: 443-450. 1934. The Fur-Seal of the Galapagos Islands. Jbid. 18: no. 2, 43-56. Tuomas, O. 1905. Suggestions for the nomenclature of the cranial length measurements and of the cheek teeth of Mammals. Pyvoc. Biol. Soc. Washington, 18 : 191-196. TURNER, W. 1888. Report on the Seals. Zoology of the Voyage of H.M.S. “ Challenger,” pt. LXVIII. WOLLEBAEK, A. 1927. El Archipiélago de Colon. (Galapagos-Qene) 11. Dyveliv. M.V. Norsk Geog. Tidss. 1 : 476-508. -§ DEC 1954 PRESENTED PLATE 10 ; A. Ventral view of type skull of Arctocephalus galapagoensis, No. 2812, Stanford Univers California. (Photograph by Dr. Antenor L. de Carvalho.) B. Ventral view of skull of - australis, 1880.7.28.7, B.M.(N.H.). c. Lateral view of type skull of A. galapagoensis, 2812. " PLATE 10 It PLATE i= \ $ 7; STANFORD 2812 ADULT MALE WENMAN Ip. HOLOTTPE OF Pretoceghalur galapgaqoens, pe PLATE 11 Dorsal views of: a. Type skull of A. galapagoensis, No. 2812. B. Skull of A. australis, 1880.7.28.7, B.M.(N.H.). c. Skull of A. australis, 1949-3.17.19, B.M.(N.H.). 3B and c are skulls of A. australis of the same condylo-basal length from the Straits of Magellan and the Falkland Islands respectively. PRINTED IN GREAT BRITAIN BY See} Myisie = aq % Ea a te ae are’ a A. Peas pare re alee mh 7 A <5 ° Ale Stas “ou inl i] a oy “URAL WIS Pp. 339-346, Pl. 12; 1 Text-figure BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY. Vol. 2 No. 11 LONDON: 1955 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical Series. Parts appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be compiled within one calendar year. This paper is Vol. 2, No. 11 of the Zoological series. PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued January, 1955. Price Five Shillings. THE SOUTHERN RIGHT WHALE DOLPHIN, LISSODELPHIS PERONI (LACEPEDE) EXTERNAL CHARACTERS AND DISTRIBUTION By F, C. FRASER EarLy in 1952 three dolphins became stranded on the beach at Onekaka, Golden Bay, which is at the northern end of South Island, New Zealand. Mrs. G. J. Goulter of Golden Bay took an 8 mm. colour film of the animals, all of which were alive. They were eventually returned to the sea and swam away apparently none the worse for their temporary sojourn on dry land. I wish to express my thanks to Mrs. Goulter for her permission to let prints be made of three of the frames of the film and also to her brother, Captain R. E. Washbourn, R.N., who brought the film to the British Museum and helped in the subsequent arrangements for pictures from it to be used in the present note. It was clear as soon as the film was projected that the dolphins belonged to the species Lissodelphis peroni, the Southern Right Whale Dolphin, and although this species has been recognized for a century and a half I am not aware that any photo- graphic record of it has previously been published. The photographs (Pl. 12) show the characteristic external features of this species, the most noteworthy of course being the absence of a dorsal fin and the spectacular coloration of the body. The pictures of the three animals resting on the sand (Pl. 12) suggest that they unlike most dolphins, are broader than deep in the thoracic region. While this appearance may only be due to the shape assumed by the body when not water- borne, it is possible that it may be natural and is perhaps a stabilizing factor in an ocean-going dolphin which lacks a dorsal fin. The slender tail stock can be seen and, although none of the pictures reproduced here gives a good impression of the pectoral fins, inspection of the whole film indicated that their shape is, as in very many of the Delphinidae, with shallowly convex anterior border and with the hinder margin convex proximally and concave distally. The head is seen to taper to a snout which, laterally, is well defined from the “ forehead” by an angular depression shallowing as it approaches the middle plane of the head, so that in the profile of the head the “‘ forehead’ merges into the rostrum with only the slightest indication of a break in the line. The tip of the lower jaw protrudes to a noticeable distance beyond that of the upper. The dorsal surface of the body is of a bluish-black colour which is sharply demar- cated from the glistening white ventral region. Snout, forehead and lower jaws are ZOOL, 2, II. 19 342 THE SOUTHERN RIGHT WHALE DOLPHIN white ; the blowhole and eyes come within the pigmented area. From the top of the head, where the pigment extends anteriorly in a little peak, the boundary between black and white sweeps low down on the lateral surface of the thorax, reaching its lowest limit about the insertion of the flippers, which have their origin in the unpig- mented region. The boundary then takes an upward direction until about half the depth of the body is reached in the lumbar region. Thence it continues caudad about midway between the dorsal and ventral margins of the body. A close inspec- tion of the film was made to decide about the dorsal surface of the tail flukes. Some indication of the amount of pigment on the tail of one of the specimens can be obtained from Pl. 12, fig. 3, but allowance has to be made for the high-light which gives an impression of a lightness of pigmentation not present in the animal itself. While in the specimen just referred to the colour was grey-blue, in one at least of the remaining two the dorsal surface of the flukes was of exactly the same bluish-black colour as the back. In addition it was seen in the film that the posterior margin of the white ventral surface of the flukes was bordered by a thin band of dark pigmentation in one specimen, while in another the band was appreciably wider, about an inch at its widest. There was definite evidence in part of the film that, although at first sight the flippers appeared altogether white, at least one of the animals had a narrow elongated darkly-pigmented patch filling the apex of the proximal convexity on the posterior border of the limb. The first published description of L. peroni, that of Lacépéde (1804) includes comment on the contrast of the dark colour of the back with the brilliant white of the sides and belly and with that at the end of the tail, the extremity of the snout and on the flippers. Lesson (1827) gives a more detailed account of the external appearance of this species of dolphin. “ Arrondi dans ses contours, gracieux dans ses formes, lisse dans toutes ses parties, ce cétacé est d’autant plus remarquable, qu'il semble recou- vert d’un camail noir. Son museau jusqu’a l’oeil est d’un blanc soyeux ou argentin. Il en est de méme des cétés du corps, des nageoires pectorales, du ventre et d’une partie de la queue. Un large scapulaire, d’un bleu-noir foncé, prenant naissance aux yeux ou le blanc décrit un croissant, se dessine et se recourbe sur les flancs, pour recouvrir seulement la partie supérieure du dos. Le bord antérieur des nageoires pectorales et caudales est brun. Le museau est allongé, séparé du crane par un sillon profond.’’ The figure of the specimen (Lesson’s PI. 9, fig. 1) does not convey the elegance of form expressed in the description and in particular the shape of the attenuated snout is quite unlike that of the animals photographed in New Zealand. Bennett (1840) says: “‘ The upper and hind part of the head, the back, and flukes are of a uniform deep-black colour which about the lower third of the body, termi- nates by a straight and abrupt line, leaving the entire abdomen, and inferior portions of the sides, of a pure and dead white. The snout, and anterior third of the head, are entirely white ; as also are the swimming paws, with the exception of a broad black spot on the upper surface and posterior margin of each.” Bennett’s figure is of an animal with a disproportionately large head and a body deep in comparison with its length. Gray (1846) quotes Cuvier’s criticism of the snout shape as depicted by Lesson and THE SOUTHERN RIGHT WHALE DOLPHIN 343 on his own account draws attention to the restriction in the latter’s figure of the pigmented dorsal part of the body. He reproduces drawings made at sea which in these essentials resemble the animals photographed in New Zealand. In Gray’s figures the upper surface of the tail-flukes is of the same dark colour as the dorsal surface of the body. His lateral view of the dolphin shows the profile of the anterior end of the lower jaw continuing the line of the profile of the snout. This is not substantiated in the rounded lower jaw tip of the New Zealand animals. D’Orbigny and Gervais (1847) have the briefest description of a specimen har- pooned near Cape Horn. The drawing in their account shews an animal somewhat slender in form, with the conventional pattern of pigmentation on the dorsal surface of the body and having the dorsal surface of the flukes of the same black colour. In addition the dorsal surface of the snout has pigment on it which does not, however, extend to the margin of the upper jaw and stops short of the pigmentation on the top of the head. The flipper is shown bearing a large irregular black spot about the middle of the posterior edge, and on the tail stock, just anterior to the insertion of the flukes, four short oblique flecks of dark colour project ventro-caudally from the boundary which separates the black of the back from the white of the under surface. Philippi (1893) describes two specimens caught off the east of Patagonia about 41°S. The figure which he gives of one of these indicates that, in general, the animal was like those described by earlier writers. Philippi says that the flippers were pure white above and below, only the rearmost corner and tip being black, and the tail black above, white below with black posterior border. Lillie (1915) refers briefly to L. peroni having been seen twice, a pair of animals on each occasion. He states that all four were exactly alike and agreed with Gray’s figure except that the tail flukes were quite white above and below. Summing up the foregoing descriptions, it would appear likely that only one species of Lissodelphis is involved for which the name L. peroni has priority. The colour pattern is constant so far as it concerns the amount of black on the back and white on the belly but varies in the amount of pigmentation on the snout, flippers and on the dorsal surface and ventral margin of the tail flukes. This colour pattern is clearly distinguishable from that of the northern form L. borealis (Peale). Excellent photographs of this northern species are to be found in the account of Scheffer and Slipp (1948). These show an animal which is much more extensively pigmented than its southern congener. In L. borealis the dark pigmentation covers the greater part of the body. The rostrum and anterior part of the “ forehead ’’ appear light grey in one of the photographs in the paper just referred to and the end of the lower jaw is without pigment. A roughly lozenge- shaped area of white lies between the flippers on the ventral surface of the body which, anteriorly, extends on to the throat, where it ends in a sharp angle. Pos- teriorly it is drawn out to become a thin white line in the umbilical region and thence to the tail it widens out again very gradually. The dorsal surface of the tail is black and the ventral surface white with a black band on the middle line and posterior border. In the consideration of the geographical distribution of L. peronz it is convenient 344 THE SOUTHERN RIGHT WHALE DOLPHIN to start with the type locality and continue eastwards round the globe. Peron (1807) himself says that the species which bears his name was based on a specimen caught off the south coast of Tasmania ; on the day following the capture the ship’s position was 44°S., 141° 27’ E. Flower and Garson (1884) record a skull which was sent to the Royal College of Surgeons from Tasmania. Lillie saw the dolphin on two occasions in positions 42° 51’ S., 153° 56’ E. and 47° 04’ S., 171° 33’ E. respectively, the former being eastward of Tasmania, the latter off the southern end of the east coast of South Island, New Zealand. The three specimens which are the subject of the present paper stranded on the north end of South Island. The specimen figured by Gray (1846) was seen in position 46° 48’ S., 142° W. that is roughly halfway between New Zealand and the southern end of South America. In the neighbourhood of Cape Horn, Malm’s (1871) record of a skull from “‘ Southern Chile ” is not by itself very indicative as to exact locality but it may legitimately be mentioned in association with the other reports from the same region. Lesson (1827) saw the dolphin near Pillar Cape at the western end of the Straits of Magellan and says that on the 12th January, 1823, several hundred surrounded the “ Coquille’, the ship then being three days south of the Chonos Archipelago, Chile, that is about 45° S. Bennett (1840) first saw specimens of L. peroni in 40° S., 50° W., south eastwards from the mouth of the Rio de la Plata. It was afterwards frequently seen by him during the passage round the Horn as far south as 54° S., but was not observed in lower latitudes than 40° S. on the western side of Cape Horn, nor indeed during any subsequent part of the voyage. Philippi’s (1893) specimens, a male and a female, were caught off the east coast of Patagonia about 41° S. D’Orbigny and Gervais (1847) say it was observed around Cape Horn from 58° to64°S. Mr. R. M. Laws (St. Catherine’s College, Cambridge) to whom I am indebted for the record, has informed me that while on a passage from Monte Video to Stanley, Falkland Islands, he saw dolphins which were unquestionably L. peroni. The extract from his diary reads: “ Jan. 22nd, 1948, 16°50 hours. A school of about 60 dolphins approached the ship from a S.E. direction. Some accompanied us for a few minutes. ... There were two types, one greyish on top and white below with no definite line of demarcation and with a very pointed beak. It also had two parallel white lines along the flanks. (Probably Lagenorhynchus obscurus. F.C.F.). The other species was black and white clear cut, with no dorsal fin and having a white beak ... a fix at 19.40 hours gave our position as 42° 2’ S., 56° 6’ W.” Eastwards across the South Atlantic at the southern end of the African continent two records are known. Schlegel (1841) refers to a skull which came from the Cape of Good Hope, but True (1889) is doubtful about the identity of this specimen. The other record is that of Dr. J. E. Hamilton, who in 1927 recorded and made a drawing of a dolphin which is certainly L. peront, seen in position 38° 34’ S., 8° 06’ E. The only remaining acceptable record, and it is not a very precise one, is that of Lesson (loc. cit.), who says that the dolphin “ was seen in 45° S. in circumnavigating Australia.” Quoy and Gaimard (1824) refer to a dolphin with long white beak seen in 2° S. near New Guinea. This was distinguished as “‘ delphinus Peroni de Lacépéde ”’. THE SOUTHERN RIGHT WHALE DOLPHIN 345 Neither this brief description nor the place of occurrence justifies acceptance of the record. The well-substantiated places of occurrence of Lissodelphis peroni indicate that it probably ranges round the world in the southern hemisphere. Although it is not entirely restricted to the West Wind Drift it appears to have some predilection for it, because, as shown in Fig. 4, those records to the north of the Sub-tropical Converg- ence are for the most part close to that boundary, and except for one sight record there is no evidence that it penetrates into the Antarctic Ocean. Complete circles indicate precise positions, dotted circles and lines indicate less well defined places of occurrence. 346 THE SOUTHERN RIGHT WHALE DOLPHIN BIBLIOGRAPHY BENNETT, F.D. 1840. A Whaling Voyage vound the Globe. 2: 1-395, figuresin text. London. D’Orsieny, A., & GERVAIS, P. 1847. Voyage dans l’'Amérique Méridionale. 4: Mammiféres. 1-32, Atlas, Pl. XXI. FiLower, W. H., & Garson, J.G. 1884. Catalogue of the Specimens ... in the Royal College of Surgeons of England. Pt. 2. Mammalia other than Man. London (J. & A. Churchill). Gray, J. E. 1846. The Zoology of the Voyage of H.M.S. ‘‘ Evebus and Terror,’ 1839-43. Vol. 1. Mammalia, Birds. 13-53, 37 pl. LackpEDE, B. 1804. Histoive Naturelle des Cétacés. xliv + 329, 16 pl. Paris, Lesson, R. P. 1827. Voyage autour du Monde ... suv... ‘‘La Coquille.’’ Zool. 1: iv + 743, 50 pl. inatlas. Paris (Arthur Bertrand). Lituiz, D. G. 1915. Cetacea. “ Terra Nova”’ Reports Zool. 1: No. 3, 85-124, 14 figs, 8 pl. London. Trustees of Brit. Mus. Matm, A. W. 1871. Hvaldjur i Sveriges Museer, Ar 1869. K. Svenska. Vetensk. Akad. Handl. 9: no. 2, 1-104, 6. pl. PERON, F. 1807. Voyage de Découvertes aux Terres Australes, 1: xv + 498. Paris. Puiippl, R. A. 1893. Die Delphine an der Siidspitze Sudamerikas. Ann. Mus. Nac. Chile : 1-16, 5 pl. Quoy, J. R. C., & Garmarp, P. 1824. Voyage autour du Monde ... ‘‘Ovanie’’ et ‘‘ Physi- cienne’’. Zool. : 1-671. Paris (Pillet). ScHEFFER, V. B., & Stipp, J. W. 1948. The Whales and Dolphins of Washington State with a Key to the Cetaceans of the West Coast of North America. Amer. Mid. Nat. 39: No. 2, 257-337) 50 figs. SCHLEGEL, H. 1841. Abhandlungen aus dem Gebiete dey Zoologie und vergleichenden Anatomie. Heft. 1, 1-44, 6 pl. Leiden. TruE, F. W. 1889. Contributions to the Natural History of the Cetaceans. A Review of the Family Delphinidae. Bull. U. S. Nat. Mus. 36: 1-101, 47 pl. EXPLANATION OF PLATE PLATE 12 Fics. 1-3. Lissodelphis peroni. Figs. 1 and 2. The dolphins on the beach at Onekaka, South Island, New Zealand. Fig. 3. One of the animals having been returned to the sea but not water-borne. (Photo : Mrs. G. L. Goulter). Bull, B.M. (N.H.) Zool, II, 11. Nae AS Dave te! LISSODELPHIS PERONI PRINTED IN GREAT BRITAIN BY | ADLARD AND SON, LIMITED, _BARTHOLOMEW PRESS, DORKING F caer 4 * f \ in b*- + . . Ms % re} f é ; 4 . F 1 : ig} oe i. KE JIPE AND THE _PANGANI RIVER, EAST AFRICA | Voli 2 No; 33 ~LONDON : 1955 NEW SPECIES OF TILAPIA (PISCES, CICHLIDAE) FROM LAKE JIPE AND THE PANGANI RIVER, BAST APRICA BY ROSEMARY H. LOWE (East African Fisheries Research Laboratory, Jinja, Uganda) Pp. 347-368 ; Pls. 13-17 ; 4 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 2 No. 12 LONDON : 1955 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), ‘stituted in 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 2, No. 12 of the Zoological series. PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued June, 1955 Price Ten Shillings NEW SPECIES OF TILAPIA (PISCES, CICHLIDAE) FROM LAKE JIPE AND THE PANGANI RIVER, EAST AFRICA ROSEMARY H. LOWE CONTENTS I. INTRODUCTION ui “ = 2 : é ci : styl II. SysTEMATIC DIAGNosIs" 3 2 é : 5 . 5 352 III. FreLpD AND PoND OBSERVATIONS . 5 6 - 0 : - 356 A. T. girigan sp.n and T. jipe sp. n. c é > 2 350: B. T. pangani sp. n 5 5 o ¢ c 5 {ou c. T. mossambica kovogwe subsp. ls 6 5 5 : : ESOL p. Lake Chala Tilapia, T. hunteri Giinther - ° - - 362 IV. Lake Victorta Species oF Tilapia Grown in Ponds o é es O2 V. DIscusSION . o 2 ° “ 5 J : 2 0 = 303) VI. SUMMARY . 5 : : 2 5 3 2 - 365 ACKNOWLEDGEMENTS : é : 2 c - 5 é - 366 REFERENCES 5 0 a : 3 2 2 - 366 SYNOPSIS Three new species and one new subspecies of Tilapia are described from the Pangani River system. Field and pond observations are given, and also observations on the growth of T. esculenta Graham and T. vaviabilis Boulenger, collected as fry from Lake Victoria and reared in ponds. INTRODUCTION In recent years recognition of the need for more protein in the diet of Africans has led to investigations of the growth of African fish in ponds. Certain species of Cichlid fishes of the genus Tilapia have proved very amenable to pond culture. This paper is concerned with an investigation, made in January, 1951, of the Tilapia species of the Pangani River system, certain of which were already being grown in fish ponds. The Pangani River flows through Tanganyika Territory in a south-easterly direction from Lake Jipe, which lies on the Kenya border just south-east of Kil- manjaro, to the Indian Ocean (Text-fig. 1). Samples of Tilapia were collected from Lake Jipe, from ponds at Taveta near Lake Jipe, and from the Tanganyika Government’s experimental fish farm at Korogwe. The Taveta ponds were stocked with fry from Lake Jipe. The Korogwe ponds were supplied with water by a channel leading from the Pangani River, and separate ponds were stocked with ZOOL. Il, 12, 20 SEO) Pangani River Tilapia and Lake Jipe Tilapia; Lake Victoria and Belgian Congo species foreign to the Pangani system had been introduced into other ponds at Korogwe. Lake Jipe is connected with the Ruvu or Pangani River in wet seasons through In 1951 the lake was about twelve miles long by one and-a-half miles wide, but was said to be silting up and spreading ; it was a shallow lake, only a few feet deep over much of its area, with large patches of water plants, Potamogeton a series of swamps. NEW SPECIES OF TILAPIA FROM EAST AFRICA and Najas species, rising to the surface, and frequented by numerous water birds. Sonalleioai anni lanunanisasanneaiaeupaaanch necameatsnmtennedtetatian) KENYA \ SN marae rama nnn, » poaven ] | i fsa THE PANGANI RIVER ean | | | TANGANYIKA TERRITORY | i) } Sence ' erin aS REEFS mies i { INDIAN OCEAN |} ‘ soe 3W’oo' E. sae Text-Fic. 1, The Pangani River system, Tanganyika Territory. NEW SPECIES OF TILAPIA FROM EAST AFRICA 351 The Tilapia collected were examined at the British Museum (Natural History) together with other specimens from Lake Jipe collected previously by Mr. H. Copley, Fish Warden, Kenya, and from the Pangani River collected by Major R. E. Gould, Fisheries Officer i/c Fish Farming in Tanganyika Territory. The following Tilapias considered to be new and described below were found: T. jipe sp. n. and T. girigan sp. n. in Lake Jipe ; T. pangani sp. n. from the Pangani river at Korogwe ; and T. mossambica korogwe subsp. n. from Korogwe ponds fed by the Pangani River. Although, at first, resemblances to T. mossambica Peters and T. nigra Giinther suggested that the Jipe and Pangani forms might find a place with these in one supraspecies, closer examination has emphasized the differences and, except for T. mossambica korogwe subsp. n., they are here described as distinct species, with some reservation as to the rank of the allopatric T. pangani and T. girigan relative to each other. T. 77pe though superficially like T. gzvigan has very distinctive pharyn- geal teeth quite unlike T. mossambica and more like T. galilaea (Linn.). Further knowledge of the ecology of all these species is required before the interspecific relationships can be understood. These Tilapia are, however, already being used on fish farms and transported round the countryside. It is therefore very desirable that those responsible for the fish should be able to name them to record movements, quite apart from the desira- bility of recording as much as possible about natural distribution before species from different river systems are inevitably mixed. For example, further study has shown that fish of supposed “T. nilotica’’ grown in the Tanganyika Government’s experi- mental fish farm at Korogwe really belong to the new species T. pangani, and appear to be more nearly related to T. mossambica than T. nilotica ; this explains the differences in behaviour between these fish and the genuine T. milotica which have been tried in dams in Uganda and the Belgian Congo. The situation has been further complicated by the introduction of four distinct species foreign to the Pangani system into the Korogwe ponds, T. esculenta Graham and T. variabilis Blgr. from Lake Victoria, and T. macrochir Blgr. and T. melanopleura A. Dum., from the Belgian Congo. There are also reports that 7. migra Giinther was once introduced into streams flowing south from Kilimanjaro, streams which also find their way into the Pangani. T. mossambica korogwe which was not intentionally stocked in the Korogwe ponds was nevertheless found in them and presumably gained access from the Pangani water supply together with some small Haplochromis which also appeared in these ponds. These accidental entries into ponds indicate the difficulty of making really fishproof screens and the need to examine the indigenous Tilapias in these rivers before escapes from ponds, and possibly hybridization with the local species, occur. A summary of the diagnostic characters of the new forms is given in Table I, and some of their characters are compared with those of T, mossambica, T. nigra and T, nilotica in Text-fig. 2, NEW SPECIES OF TILAPIA FROM EAST AFRICA 352 €-z 0 € (gx) 41-91 : oz-gI z£-0€ a SE-rE (€€) b1-11 Pee c orbr{ 1 = 11-6 ; €1-z1 6z-Sz gI-or sf-of { OT Li1-+1 : Iz-61 c-+ : g-b C-r o g-S _ : a} PIPw19} Uy — * repnsuery 1eyyeus I0-I-16-0 : go-1-26-0 £6.0-69-0 : 0-1I-g-0 ¢.zz-So-61 g €.€z-b.1z 0-9£-0-1€ s z-g€—z- bE £ 0-ob-g-o€ (6 F-bE) ° +.ob-6.4€ P S.bb-1-9€ (6 #-FE) - 1.L€-z.z€ —_ . 6.6€-L.1€ L.GE-€.zE 4 €.F€-g.0€ L.ob—z-S€ 2 ¢.€b-L.1€ QI-II JO zI . 6z-zz Jo 8 ‘u'ds ‘qns amsos0y “uM * 7 ‘u'ds zuvdund * fT - €-z : oz-Lr aD (9€) S€-FE (€€) + gI-F1 eee - FI-zI 8-08 { (77 . oz—gi (21) 9-b o gb * paie}}e9s [[aM ‘asIeOD qe[nsuersy osreT : 00-I-Ig-0 ‘ o00-1-64-0 S L.1Z-1-g1 : G.gf-+- FE C L.zb-6-1€ : 9-g-g-4z ; +.L€-0.0€ s z-g9€-9-1€ . +.ob-z.b€ If-9QI jo 61 turds unsaid * Zz . oz-L1 0 CE-FE (z£) G zI-Ol : € 10+ . €1-11 : 61-L1 ‘ giI-hi . 1€-6z * (zz) 1z-0z (61) (g) 2-§ : Q g-S . *papmoio Ajasuap ‘outy* a poedeys-jreayy . © zz-1-6-0 : Go.1-0g-0 0 O +¥.zz-L.gt g : o0.ob-€ .€€ : £.6€-L.6z ‘ : g-€€-g-Sz : . w-zb—-z.GE x z-g9€-0.z€ . : z-ob-0-9€ : Lz-ZI JO ZI 9821S suds adil “7 yeeyo uo (94a mojaq) saties Jo 1aquinyy ) 4 + gjounped jepnes punory . * — series our yerezeT, sei 4jOs Jo Ioquinyny soutds jo 1aquinyy sAex }JOSs JO IaquINyy soutds Jo 1oquinN soTeoS, ug [euy uy [esiog. yore Joriozue Jo zed IaMOy UO siOYeITIS Jo Ioquinyy : dasel, SaMOT solias JO Joquinu :1499, * — mef soddn rage doa Lean eg * 43903 * Boe Pet} \ su0q peasudzeyd Jomoy 5 > WPI yqsug] : . . mda, ounped fepne 43809] pounped Tepney) see pistol /PEoH = 00% * dap peqqi0arg, ; : oor x 48001 PeeH YIPIM [e}IqQI010;UT chee te oor x Wi8uel PeeH ynous : F oor x W28Uel PeeH mel 1aMoT ; : oor x Sue] Apog prepueys Sug] UY [e10}00q ‘ ; oor x U#8ueI Apog prepueys y3Su9] peep ; ‘ oor x Uuel Apog prepueys uydap Apog "WD YySUeT [e}0}—oSueT !OPeU s}USWIOINsvaM popIejop YOM wo ‘ON vidojry fo savwads mau fo sdajovavys I4sousvig— |] AAV 353 ‘soul, WoHOIG 4q uMoys sosuex jo syed yuonbery sso] (2TIN "M) “T vayopiw * pure (tyyV “A) “739 vas *z ‘(spremyynos pur tsequreZ) NEW SPECIES OF TILAPIA FROM EAST AFRICA s1ajag vaiquepssow * 7, yO se_dures yo 9soy} YIM satoads paqriosap AT Mou OY} JO sIopORIeYO snoreA jo uostreduio9 W—'z “S14 IXAL = el Se Sees ° « U 094joj1U “EE ened ee] een — SS ° s + pase "I = 11a ee 00 === ESS 00 eee 0 * parquipssou *T, aE al a) om a . * amdodoy "Wt * I, = oD =z meno ——a 0 : quvsuvd “I Sj — =a 00 eos 1 eames tt . : + unsias *T re] Ce] ead CU 3 > 3 aqul “I, Gi P 6 | AYA MATERIALS 6 ; 5 : 6 6 3 ‘5 ee o| SWe ACKNOWLEDGMENTS . 2 z 5 c : ‘ i , A | ye HIsToRICAL NoTE F , 5 5 5 4 5 0 Bye DESCRIPTION. é 5 z . 2 c c : 5208) 1. Rhamphotheca 4 : : : C 5 0 c 375 2. Skull ‘i - = f 5 . c 3 7 : maT: 3. Myology of the Jaw : 0 c ¢ : ‘ a | Soke) CoMPARATIVE NoTE . 4 c é : - 5 5 . » 385 Discussion : 3 9 C 5 z 5 5 : 6 7 389 CONCLUSIONS . 3 : 4 : , i c 4 c - 390 APPENDIX : 0 “ 5 A 3 6 é = - 2) 30 REFERENCES 6 5 z f 6 3 6 : 5 a Byep. SYNOPSIS The ability of the Hawfinch, Coccothraustes coccothraustes, to crack open the stones of cherries, damsons and olives in order to feed on the kernels is generally well known, but the modifications which are present in the structure of the head and enable this comparatively small bird to perform such a feat have apparently received little attention. The method employed by C. coccothraustes to crack open the fruit stones is noteworthy because each stone is held between the mandibles and is broken solely by the force applied by the jaw muscles, and not by any “‘ artifice ’’ such as that employed by the thrush to smash snail shells, or the woodpecker to split almond stones. The behaviour of the bird when feeding on fruit stone kernels appears to follow a definite pattern. The bird usually selects fallen fruit apparently discarding the soft parts. Observations and photo- graphs show that a cherry stone is positioned in the mouth by the combined movements of the head and tongue until it is held lengthways between the mandibles at the back of the horny palate of the mouth with the suture of the stone lying in the median sagittal plane of the head. A quick snap of the jaw and the shell is neatly cracked along the suture or ‘‘ seam ”’. The halves of the shell are rejected and the kernel is swallowed whole, usually without being crushed. Experiments on breaking open cherry and olive stones (given in detail in the appendix) ~ show that pressures in the region of about 100 lb. are required to perform this feat | ZOOL. 2, 13. 21 372 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH The purpose of this paper is to show how the head of C. coccothraustes is adapted to apply and withstand such forces by describing some aspects of the morphology of the head, namely, of the horny bill, or rhamphotheca, the skull and the jaw muscles. MATERIALS Tus study is based on specimens of C. coccothraustes collected in Great Britain. The material consists of two adults and two juvenile male specimens preserved in spirit in the National Collection, one head, sex unknown, supplied by Mr. G. R. Mountfort, and several skeletons in the National Collection. In most of the osteo- logical material the rhamphothecae, which can only be removed with difficulty, were intact. The heads of two spirit specimens were dissected completely during the course of the investigation. The head of a spirit specimen of a Brambling, Fringilla montifringilla, was also dissected, and reference was made to the series of skulls of the Chaffinch, Fvingilla coelebs, in the National Collection. ACKNOWLEDGMENTS I am indebted to Mr. Guy Mountfort for suggesting this investigation. The work was originally undertaken in order to answer some of the points raised by him in the preparation of his monograph of the Hawfinch. The scope of this study was subse- quently enlarged. Iam obliged to Mr. H. L. Cox and his colleagues of the National Physical Laboratory, and Mr. D. Welbourn of Cambridge, for providing me with the interesting data which form the appendix to this paper; and to Dr. J. Wahrman of the Hebrew University, Jerusalem, for a supply of fresh olives for the crushing tests. I also wish to thank my colleagues in the British Museum (Natural History) for reading the MS. and for providing many helpful criticisms. HISTORICAL NOTE References to the general anatomy of the head of C. coccothraustes are remarkably few and are confined mainly to the consideration of characters of the skull as an aid to classification, or less frequently to its architecture. The skull with its high degree of ossification particularly impressed W. K. Parker (1879) by what he called “ its ridgy strength ”’ when he described some aspects of the anatomy of the skull in his systematic work on the structure of the palates of passerine birds. At the beginning of the section on C. coccothraustes he pointed out that most of the differences between the skull of this species and those of other members of the Fringillidae were of little phylogenetic importance because the modifications in the skull of C. coccothraustes were for “‘ mechanical purposes’. Nevertheless, Bowdler Sharpe (1888) used many of these modifications ‘‘ for convenience’ to separate what he termed the Cocco- thraustinae from the buntings and true finches. The peculiarities of the horny bill or rhamphotheca were noted very briefly by Pycraft (1905), and his observations were occasionally referred to by subsequent workers. Nothing new was recorded on the horny bill until Sushkin (1925) described similar modifications in the bill of the Evening Grosbeak, Hesperiphona vespertina. THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH 373 He concluded that they were essentially the same although purely adaptive charac- teristics. He agreed with W. K. Parker that most of the peculiarities of the skull were adaptive features, but believed that only one character of the bony palate was of systematic significance, and on that he regarded the Evening Grosbeak as a member of a sub-family containing the Old World hawfinches. Later references are confined almost entirely to remarks in papers dealing with the -mechanical structure of bird skulls, for example those by N. G. von Lebedinsky (1921) and von Kripp (19334). The latter author showed that the skull was struc- turally stronger than that of a harrier. The literature appears to contain few references to the myology of C. coccothraustes, so an account of the musculature of the jaw is given in this present paper. The nomenclature adopted is that proposed by Lakjer (1926) in his comparative work on the Sauropsida. DESCRIPTION 1. Rhamphotheca The rhamphotheca of birds is derived from the malpighian layer of the epidermis. In most birds it is seldom more than a thin sheath which is readily detached from a dried skull, but in C. coccothraustes parts of it inside the mouth are greatly thickened and enlarged. In the palatal region (Text-fig. 1a) it forms a longitudinally striated thick pad partly divided by a longitudinal median depression. This thickened area extends from the anterior border of the palatine bones over the posterior third of the premaxillae. The anterior two-thirds of the premaxillae are covered by the remainder of the horny palate which is strengthened by one median and a pair of lateral ridges. The grooves formed between the lateral ridges and the edges of the upper mandible accommodate the edges of the lower mandible. On the lower jaw (Text-fig. 1b) the horny sheath is thickened to form two large bosses which lie B Fic. 1. Coccothraustes coccothraustes. The oral surfaces of the rhamphotheca showing the striated pads. a, Upper jaw; B, Lower jaw with the anterior of the tongue in situ. ZOOL. 2, 13. 21§ 374 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH immediately over the posterior parts of the dentary bones which are themselves specially thickened in this area. The bosses are striated, but the striations being curved more or less follow the contours. The depression between the two bosses accommodates the tongue. The tip of the tongue is broadly cuneiform in shape and lies flush with the anterior surfaces of the lateral bosses. 2. Skull The principle modifications of the skull which appear to be connected with the ability of C. coccothraustes to exert a great force with its jaw muscles can be summarised as follows : In the first place the whole skull is strengthened by a greater ossification than is usually found in most fringilline species. An indication of the extent of ossification as determined by weight, is given in a later section. Von Kripp (19334) described how the architecture of the skull is suited to withstand stress and strain. Viewed laterally the outline of the skull approximates to one horn of a crescent, the dorsal profile being convex and the central one concave (Text-fig. 2a). The concave ventral surface is strengthened by a system of struts formed by a massive pyterygoid-quadrate-zygoma system. The anterior part of the head is further strengthened by the heavily ossified nasal and interorbital septa. These elements play an important role in the rigidity of the skull for their function is analogous to that of the vertical component of an I-shaped girder. In the skulls of many birds the whole of the upper mandible is hinged to the cranium, but in C. coccothraustes it is rigidly fixed. The suture between the frontal and nasal bones, where the hinge is usually located, is obliterated ; the nasal and the interorbital septa form what is functionally a continuous vertical wall ; and the palatines are ankylosed with the rostrum and vomer which form most of the base of that wall. The fixity of the upper mandible also may be associated with the nature of the relationship between the cranial head of the quadrate and the surrounding structures in the otic region of the skull. The quadrate is joined to the palatine by the ptery- goid and to the upper mandible by the zygoma, and owing to the immobility of the upper mandible the quadrate is held firmly by these bones. The absence of move- ment together with the great crushing strain imposed on the quadrate by the lower jaw has resulted in the squamosal and opisthotic facets in the otic region extending partly around and in close contact with the small squamosal and opisthotic heads of the quadrate. The extension of the area of contact between the quadrate and the cranium is functionally possible only because the quadrate is held firmly by the pterygoid and the zygoma. The increased area of contact provides a strong base for the quadrate which in its capacity as the fulcrum of the lower jaw experi- ences great pressures when hard food, such as fruit stones, are cracked. The great pressures have also influenced the nature of the quadrate which is very massive. Moreover, a powerful muscle, M. Quadrato-mandibularis, originates from its orbital process, and the inclusion of the muscle as an important member of the THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH 375 Adductor Group has no doubt contributed towards the massive development of this bone. The post-orbital process of the squamosal is unusually large and this feature may be correlated with the powerful muscle, the pars medius of the M. mandibulae externus, which has a tendinous origin on it. The pars medius is inserted into the post. orb. proc. proc. cnd. post. pal. proc. Pt. proc. md. int. pars ‘ant. pars post. for. md. Fic. 2. Coccothraustes coccothraustes. a, cranium and upper jaw, lateral view from the right side. 3B, Lower jaw. Ant. pal. proc., anterior process of the palatine ; for. md., foramen mandibularis ; pays. ant., anterior part of the lower jaw (dentary) ; pars. post., posterior part of the lower jaw ; post. orb. proc., post-orbital process of the squamosal; post. pal. pyroc., posterior process of the palatine ; proc. cnd., processus coronoideus ; proc. md. int., processus mandibulae internus (articular) ; pf., pterygoid ; quad., quadrate ; zy., zygoma (quadrato-jugal and jugal). lateral surface of the processtis coronoideus, but the modifications associated with this will be discussed later. Dislocation of the lower jaw from the quadrate during the contraction of the power- ful jaw muscles is prevented by a ligament which extends from the zygoma to the posterior surface of the processus mandibulae internus, passing over the posterior 376 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH surface of the distal end of the quadrate. This ligament has two centres of ossifica- tion ; one forms a large sesamoid which lies close to, and posterior to, the articula- tion of the jaw; the other is smaller and lies more laterally near the attachment of the ligament to the zygoma. The lower jaw, like the remainder of the skull, is massive and ossified to the extent that the foramen mandibularis, usually a large opening, is reduced to a size that permits only the passage of a branch of the trigeminal nerve (Text fig. 20). The anterior part of the lower jaw is strengthened by an inward expansion and union of the paired dentaries, the long symphysis forming a bony shelf. Postero- mesially the dentaries are greatly thickened to serve as strong foundations for the heavy bosses of the rhamphotheca. 3. Myology of the Jaw The muscles involved in the movement of the jaws of birds form four functional groups : ) The Adductor Group. ) The Constrictor Group. ) The Protractor Group. ) The Retractor Group. The Adductor Group raises the lower jaw, the Constrictor Group depresses it ; the Protractor Group raises the upper mandible and the Retractor Group lowers it. In C. coccothraustes the upper mandible is not hinged on the cranium, as already mentioned, and the muscles of the Protractor Group which persist are functionless, while the muscles of the Retractor Group assist in the elevation of the lower jaw and are dealt with here as members of the Adductor Group. The Constrictor Group and the Protractor Group in C. coccothraustes each consists of only one paired muscle and therefore neither im sensu stricto can be considered as a “‘group”’. Nevertheless, the term has been used here for convenience in both cases. I Tue Appuctor Group The Adductor Group, which raises the lower jaw, is the largest of the three groups. This assemblage in C. coccothraustes may be further divided into two clearly marked functional units, one acting from the top and back of the cranium and the other acting from the orbital walls and their vicinity. The muscles of the former are inserted in the ramus of the lower jaw near the superior (dorsal) and the posterior margin of the coronoid elevation. Functionally, this is the more important unit and consists of the Ms. Adductor mandibulae and the M. Quadrato-mandibularis. The other muscles of the Adductor Group, the M. Pterygoideus and the M. Ethmo- mandibularis, are inserted near the inferior (ventral) margin of the ramus of the lower jaw. a i ie THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH 377 DESCRIPTION OF THE ADDUCTOR Group M. ADUCTOR MANDIBULAE EXTERNUS (Text-fig. 3) (r) PARS SUPERFICIALIS The rostral part is fan-like and arises from the cranium over most of the anterior half of the frontal. The area of attachment is confined anteriorally and mesially by the frontal crest, and laterally by the flattened rim of the orbit. The fibres pass forwards and downwards between the post-orbital process of the squamosal Fic. 3. Coccothraustes coccothraustes. Diagrammatic view of the right side of the head showing the external jaw muscles. p.m., p-p., p.s., partes medius, profundus, superficialis of the M. Adductor mandibulae externus; Dp. md., M. Depressor mandibulae. and the rim of the orbit. The fibres converge becoming more tendinous and are inserted at the apex of the processus coronoideus of the lower jaw. The main part extends over the posterior half of the frontal; it is bounded by the rostral part in front and the transverse post-frontal crest behind, and extends inwards to the meso-frontal crest. The converging fibres are directed more forwards than those of the previous part and pass over its aponeurosis and the post-orbital process of the squamosal. The head is inserted in the pars posterior of the lower jaw laterally to the head of the rostral part at the apex of the processus coronoideus. The caudal part arises from that area of the cranium bounded by the transverse post-frontal crest in front and the transverse crest on the posterior part of the Squamosal and the parietal. (Mesially the two prominences converge and form a triangular area of attachment). The fibres passing more forwards than downwards _ become more tendinous as they converge; they are inserted somewhat laterally in the posterior margin of the processus coronoideus. This part is similar in size to 378 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH the other two but it differs in that the converging fibres pass under both the distal part of the post-orbital process of the squamosal and the ascending fibres of the pars medius (see below). (2) PARS PROFUNDA The pars profunda arises from the cranium on the anterior surface of the supra- meatal ridge of the squamosal and in the concavity in the otic region between the cranium and the quadrate, also to the neck of the quadrate. This muscle is very tendinous throughout although it is inclined to be more fleshy in the part occupying the concavity in the otic region. It is inserted in the ramus of the lower jaw ventrally to the site of insertion of the pars superficialis, that is, about half-way along the outer edge of the posterior margin of the processus coronoideus. Some of the fibres and the dense fascia surrounding the muscle coalesce with those of the pars superficialis. (3) PARS MEDIUS The pars medius is a fan-like muscle with a tendinous origin on the ventral surface of the post-orbital process of the squamosal. The fibres pass downwards and become more fleshy as they diverge. The muscle is inserted over most of the lateral surface of the processus coronoideus of the lower jaw. M. ADDUCTOR MANDIBULAE POSTERIOR (Text-fig. 4) The M. adductor mandibulae posterior arises from the posterior wall of the orbit, namely, from the alisphenoid outside the latero-ventral margin of the fossa in the orbital wall. Mesially there is a thin bony partition which separates the muscle from the M. Pseudotemporalis (see below), and laterally another thin bony promi- nence extends down from the rim of the orbit and separates the muscle from the descending fibres of the pars superficialis. The muscle is short and very tendinous. It is inserted in the mesial surface of the processus coronoideus near the upper half of the posterior margin. There are a few tough ligamentous strands extend- ing the whole length of the muscle and these are inserted in the small dorsally directed conical process of the mesial surface of the processus coronoideus near the posterior margin. M. ADDUCTOR MANDIBULAE INTERNUS (M. PSEUDOTEM PORALIS) (Text-fig. 4) The M. Pseudotemporalis arises on the posterior wall of the orbit from the alis- phenoid below the fossa in the orbital wall. The site of attachment is separated laterally from the M. Adductor mandibulae posterior by a thin bony partition (see below), and mesially from the optic foramen by a similar partition. The muscle is tendinous throughout and it is attached to the ramus of the lower jaw in the mesial surface of the processus coronoideus. The site of attachment is a little below the antero-dorsal margin of the ramus anterior to the M. Adductor mandibulae posterior, THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH 379 Pt. d. m. Tic. 4. Coccothraustes coccothraustes. A, Diagrammatic view of the right side of the head, eye and eye muscles removed to expose the jaw muscles in the orbit. 8B, Dia- grammatic dorsal view of the right orbit. A.m.p., M. Adductor mandibulae posterior ; E. md., M. Ethmo-mandibularis ; Pst., M. Pseudotemporalis (M. Adductor mandibulae internus) ; Pt.d.l., M. Pterygoideus dorsalis lateralis; Pt.d.m., M. Pterygoideus dorsalis medialis ; Qmd., M. Quadrato-mandibularis ; 7, Optic nerve. M. QUADRATO-MANDIBULARIS (Text-fig. 4) The M. Quadrato-mandibularis forms a thin sheet of muscle, almost devoid of tendinous tissue, underlying the Ms. Pseudotemporalis and Adductor mandibulae posterior. It extends from the quadrate to the mesial surface of the processus coronoideus. The muscle arises from both surfaces of the orbital process and as far as the neck of the quadrate where part of the pars profunda of the M. Adductor 380 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH mandibulae externus originates. The muscle is inserted obliquely in the ramus of the lower jaw along a line tending to pass from below the site of insertion of the M. Pseudotemporalis towards the point of articulation with the quadrate. M. ETHMO-MANDIBULARIS (Text-fig. 5) The M. Ethmo-mandibularis arises from the anterior part of the interorbital septum above the palatine and partly from the anterior wall of the orbit, that is, from the mesethmoid. Some of the dorsal fibres pass horizontally across the anterior angle of the orbit and assist in the support of the main fibres which pass obliquely downwards and outwards back to that part of the lower jaw behind the gape. The fibres become tendinous a little before being inserted into the mesial surface of the ramus of the lower jaw above the ventral margin. M. PTERYGOIDEUS (Text-fig. 5) M. PTERYGOIDEUS DORSALIS M. Pterygoideus dorsalis medialis The M. Pterygoideus dorsalis medialis arises from the interorbital septum at the antero-dorsal angle of the orbital wall above the attachment of the posterior part of the M. Ethmo-mandibularis. The muscle has numerous tendinous fibres near its site of origin, and these increase in number until the muscle is completely tendinous a little before the site of insertion. The muscle passes backwards and downwards between the ventral surface of the orbital process of the quadrate and the dorsal surface of the pterygoid (beneath the M. Protractor pterygoidei, see below). The muscle is attached by a very strong tendon to the ramus of the lower jaw at the head of the processus mandibulae internus. The tendon lies over the head of the inflexion and is inserted into the upper part of the posterior surface. M. Pterygoideus dorsalis lateralis The M. Pterygoideus dorsalis lateralis arises principally from the palatine and from the flattened area at the union of the pterygoid and the palatine. The thick fleshy fibres pass obliquely downwards and backwards and outwards to the site of insertion which is towards the ventral margin of the mesial surface of the ramus of the lower jaw, posterior to the M. Ethmo-mandibularis. The only tendinous tissue in the muscle is in the nature of a thin sheet extending throughout the entire muscle. M. PTERYGOIDEUS VENTRALIS M. Pterygoideus ventralis medialis The M. Pterygoideus ventralis medialis arises from near the top of the mesial surface of the processus mandibulae internus. It is a slender muscle and the fibres pass forwards and upwards becoming more tendinous. The tendon is inserted in the palatine along the outer (ventral) surface of the posterior process where it curves as if to meet the other member of the pair. er ee eee THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH II Pt. d. m. Pr. pt. pars ant. ant. pal. proc. Pt. v. I. a. pars See | Pt. v. I. p. post. pal. proc. proc. md. int. Pt. v. m. I'1c. 5. Coccothraustes coccothraustes. a, Diagrammatic view of the right side of the head dissected to expose the ventral jaw muscles. B, Diagrammatic ventral view of the ventral jaw muscles. Pr.pt., M. Protractor pterygoidei; Pt.v.l.a., anterior portion of the M. Pterygoideus ventralis lateralis ; Pt.v.l.p., posterior portion of the M. Ptergoideus ventralis lateralis; Pt.v.m., M. Pterygoideus ventralis medialis. 381 382 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH M. Pterygoideus ventralis lateralis Anterior Portion. The anterior portion of the M. Pterygoideus ventralis lateralis is a thin fan-like muscle arising mainly from the ventral margin and partly from the mesial surface of the ramus of the lower jaw. The site of attachment is mid-way along the ventral margin of the pars posterior of the ramus. The fibres pass upwards mesially to other muscles, then they become more tendinous and are inserted along the outer member of the bifurcation at the distal end of the anterior palatine process. Posterior Portion. The posterior portion of the M. Pterygoideus ventralis lateralis is a stouter muscle than the anterior portion. It arises along most of the mesial surface of the processus mandibulae internus of the lower jaw, that is, all of the mesial surface except for the small area occupied by the M. Pterygoideus ventralis medialis. It is also attached along a little of the ventral margin and mesial surface of the main body of the ramus of the lower jaw. The fibres are markedly tendinous and ultimately form a tough tendon which in inserted along the inner branch of the bifurcation at the distal end of the anterior palatine process (the anterior portion is attached to the outer branch). REMARKS ON THE ADDUCTOR GROUP The relatively large area of attachment of the muscles of the Adductor Group indicates that the total stress of the combined contraction is evenly distributed over most of the cranium. The M. Adductor mandibulae externus arises over the lateral, dorsal, and posterior surfaces of the cranium and the Ms. Adductor mandibulae posterior and internus (M. Pseudotemporalis) arise over the anterior of the cranium, that is, the posterior wall of the orbit. The Ms. Pterygoideus and Ethmo-mandibularis are attached to the mesethmoid, palatine and the pterygoid and the stress of their contraction is relayed to the antero-ventral surface of the cranium by the interorbital septum. Functionally, it is not strictly correct to include all the elements of the M. Ptery- goideus in the Adductor Group of muscles, for when both the mesial members of the M. Pterygoideus dorsalis and ventralis contract they seem to assist in the depression of the lower jaw; they draw the processus mandibulae internus of the articular forwards, and so rotate the anterior of the lower jaw downwards through the same angle (Text fig. 6). However, on closer examination it becomes apparent that these muscles and the posterior portion of the M. Pterygoideus ventralis lateralis hold the articulatory surfaces of the lower jaw and the quadrate in contact. Thus they strengthen the hinge mechanism and help prevent the dislocation of the jaw when the powerful crushing muscles contract. If the upper mandible of C. coccothraustes were hinged and the palatines were free and formed part of this mechanism then those members of the Ms. Pterygoideus dorsalis and ventralis which are inserted in the palatines would act as retractor muscles. These would draw the palatines backwards so lowering the upper mandible and holding it fast. However, it seems proper to include all the members of the Ms. Pterygoideus dorsalis and ventralis in the Adductor Group for by their contrac- tions they ensure the most efficient functioning of the main ‘“‘ adducting ” muscles. 383 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH group because of Not only is it valid to include these muscles in the same functional their necessary simultaneous contractions but because the same branch of the trigeminal nerve. y are innervated by the ‘OB FANS [PISOUL OY} IOAO sofosnu Mel oY} JO UOIJESUT IO UISIIO Jo sozIs ay} Surmoys Mel IaMoT oy7 Jo Fey [eHWTSes yYstr 9y7 JO yo] oy} WoIy motA oFeUUTeISeIG *S2ISNDAYIOIIOD SasisnvsYI0IIOD *9 “OI 1p ld 384 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH II THE CONSTRICTOR ‘“‘ GROUP ”’ The depression of-the lower jaw, as mentioned above, is effected by only one paired muscle. DESCRIPTION OF THE CONSTRICTOR ‘‘ GROUP ”’ M. DEPRESSOR MANDIBULAE (Text-figs. 3 and 7) The M. Depressor mandibulae arises from the greater part of the exoccipital, especially the part forming the posterior wall of the bony meatus. The fibres of the muscle pass downwards without converging, or becoming very tendinous, to where they are inserted into the post-articular part of the mandible. Tendinous Dp. md. Fic. 7. Coccothraustes coccothraustes. Diagrammatic posterior view of the head showing the external jaw muscles. insertions are made in the angular behind the lateral process, ventrally in the posterior of the angular, and in the posterior of the processus mandibulae internus. The fibres form a flat muscle which completely covers the mandibular articulation from behind and also forms a fleshy extension to the posterior wall of the meatus. REMARKS ON THE CONSTRICTOR “‘ GROUP”’ The M. Depressor mandibulae is a feeble muscle in comparison with the muscles of the Adductor Group. Its action does not necessitate any rapid contraction, nor is it required to exert any great force to depress the lower jaw. The chief function of the muscle is to overcome the effect of the relaxed, but unextended, muscles of the Adductor Group that would otherwise tend to oppose the depression of the lower jaw. THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH 385 III THE PRoTRACTOR “ GROUP” The upper mandible does not articulate on the cranium in C. coccothraustes but although this mechanism is absent one functionless paired muscle, as already mentioned, persists as a vestigial structure. DESCRIPTION OF THE PROTRACTOR “‘ GROUP ” M. PROTRACTOR PTERYGOIDEI (Text-fig. 5) The M. Protractor pterygoidei is a slender muscle arising from the alisphenoid in the posterior wall of the orbit a little below the optic foramen. The fibres pass downwards and outwards beneath the orbital process of the quadrate ; they converge and becoming tendinous are attached to a small dorsally situated spine-like process near the quadratal end of the pterygoid. REMARKS ON THE PROTRACTOR ‘‘ GROUP” In species where the upper mandible is hinged on the cranium the function of the M. Protractor pterygoidei is to draw the distal (quadratal) end of the pterygoid upwards. The quadrate is attached to this and when the pterygoid moves upwards the quadrate is rotated forward, thrusting the pterygoid forward at the same time. This action pushes forward the palatine under the interorbital septum. The upper mandible is hinged to the cranium on the dorsal side so that forward movement on the ventral side communicated by the palatines has the effect of lifting the bill onits hinge. The zygoma is also attached to the quadrate and the forward rotation of the quadrate causes the zygoma to be moved forward, thus assisting in the eleva- tion of the upper mandible since the zygoma is joined to the maxilla. Although the muscle is functionless in C. coccothraustes is is interesting to note that its fibres are not completely lacking in tendinous strands and generally the muscle closely resembles the form of the muscles of the Adductor Group. COMPARATIVE NOTE For comparative purposes the Chaffinch, Fringilla coelebs, and the Brambling, F. montifringilla, have been selected as more generalized fringilline species. In comparison with C. coccothraustes the oral surfaces of the rhamphotheca of F. coelebs are smooth and lack any dilations similar to the striated pads, or bosses (Text-fig. 8b). It would appear (Sushkin, 1925) that this specialization is not confined to C. coccothraustes for apart from Hesperiphona of North America the occurrence of specializations of this nature are found in a few other Old World genera, namely, Eophona, Perissospiza and Mycerobas. In comparison with the skull of C. coccothraustes that of F. coelebs is a fragile structure and this difference is illustrated by comparing the sizes and the weights of the skulls of the two species. The average maximum dimensions of skulls 386 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH of C. coccothraustes are 40 X 20 x 20 mm. and F. coelebs 30 x 14 X 15 mm.; if the cube of each of the linear proportions of the skull of C. coccothraustes is compared with those of a hypothetical skull of the same average dimensions as F. coelebs, but of the same design as that of C. coccothraustes, it is found that the dimensions of the skull of C. coccothraustes average 250%, “‘ larger’. Yet the average weights of the skulls of C. coccothraustes and F. coelebs are 3:235 gm. and 0-654 gm. respectively, that is, the skull of C. coccothraustes contains nearly 400°% more bony material than that of F. coelebs. Therefore, the skull of C. coccothraustes is very massive for its size. The outline of the skull of C. coccothraustes has been likened to one horn of a crescent, but the skull of F. coelebs differs by being more cuneiform, and its dorsal profile is nearly straight (Text-fig. 8a). The difference is attributable mainly to the post. orb. proc. for. md. Fic. 8. Fringilla coelebs. a, Cranium and upper jaw lateral view from the right side (compare Text-fig. 2a). 8B, Lower jaw (compare Text-fig. 20). greater angle between the basi-cranial and basi-maxillary axes! in F. coelebs which is 150° but only 120° in C. coccothraustes. Van der Klaauw and Duijm consider that a difference of this nature can be a functional adaptation and Duijm stated, ‘‘. . . the bill and the cerebral capsule behave mainly as independent functional elements.”. In the present instance the downward rotation of the bill relative to the cranium in C. coccothraustes reduces some of the stresses set up by the application of powerful forces to crack fruit stones. This is shown diagrammatically in fig. 9. Here BC and AB represent the basi- maxillary axis and a line parallel to the basi-cranial axis in two skulls of the same overall length. In (a) the subtended angle between the two is 120° as in C. cocco- 1 Basi-maxillary axis: The line of intersection between the vertical longitudinal plane of the head and the mean ventral surface of the premaxillae. THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH 387 thraustes, and in (B) it is 150° as in F. coelebs. Assuming that the same forces were applied to the lower jaw to crack a nut held at the same distance from the angle of the jaw, then the magnitude and direction of the force, relative to the upper jaw, can be represented as the line XY. The direction of the force is tangential to a circle whose centre is at the articulation of the lower jaw, A, and periphery at D, where the lower mandible presses against the stone. The force XY is resolvable into two Bi Q Fic. 9. Diagram showing the main forces imposed on the upper jaw when food is crushed between the mandibles. a, The subtended angle, ABC, between the basi- maxillary and basi-cranial axes is 120°. 3B, The subtended angle is 150°. components at right angles to one another, XP and XQ. The former exerts a compressional and the latter a bending strain on the upper jaw. A bending strain is clearly more likely to fracture the bill than a compressional strain along its length and the bending component is approximately 6% less in (A) where the angle between bill and cranium is similar to that of C. coccothraustes. Therefore, the downward rotation of the bill in C. coccothraustes may be regarded as an adaptation to feeding so that hard stones are cracked with a smaller bending strain being experienced by the bill; that is, there is less chance of fracture than in a skull of the design of F. coelebs where the subtended angle between the bill and cranium is more obtuse. 388 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH The bending strain which is experienced by the bill of C. coccothraustes when fruit stones are cracked must nevertheless still be relatively enormous when forces in the region of 100 lb. have to be applied to crack them, but its effect is largely nullified by the highly ossified septa. These are markedly different in F. coelebs where the interorbital septum exhibits a more normal passerine condition forming only a thin bony partition between the orbits and the nasal septum appears to be mainly unossified. Again, a high degree of ossification is found in the pterygoid-quadrate-zygoma system, for in comparison the pterygoids and zygomae of F. coelebs are long and slender, while the quadrates of C. coccothraustes are about twice the size of those of F. coelebs.1 An examination of the skulls reveals that not only has the skull of C. coccothraustes become modified to withstand greater stress, but also to provide a greater area of the attachment and insertion of the jaw muscles. This latter point is well illustrated by the difference observed between the lower jaw of C. coccothraustes and F. coelebs. In the former the coronoid is elevated in correlation with the large size of the pars medius of the M. Adductor mandibulae externus which is inserted into it, moreover, the foramen mandibularis is very small while it is relatively large in F. coelebs (Text-fig. Io). The disparity in the sizes of the partes medii in the two birds is Fic. 10. Fringilla montifringilla. Diagrammatic view of the right side of the head showing the external jaw muscles (compare Text-fig. 3). also reflected in the degree of development of the respective post-orbital processes of the squamosals from which they originate. Generally the lower jaw of F. coelebs appears to be more primitive in character than that of C. coccothraustes which, in addition to the adaptations just referred to, is modified to carry the heavy bosses of the bill, as described above. 1 Quadrate of C. coccothraustes ; mass 0:0375 gm., 9 mm. high X 7 mm. long (orb. proc.). Quadrate of F. coelebs : mass 0-0041 gm., 4°5 mm. high x 4°5 mm (orb. proc.). THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH 389 The jaw muscles of C. coccothraustes in comparison with those of the Brambling, Fringilla montifringilla, are exceedingly tendinous. In the latter they are composed of fleshy fibres almost devoid of tendons, except at the site of insertion, and even then the tendons are not as evident as those in C. coccothraustes. The jaw muscles of F. montifringilla are generally more bulky and appear to contain much non-con- tractile tissue. One gains the impression that the flaccid muscles of F. montifringilla are incapable of the rapid powerful contractions of the muscles of C. coccothraustes. The difference in the form of the M. Adductor mandibulae internus (M. Pseudo- temporalis) in the two species illustrates the comparatively compact nature of the muscles of C. coccothraustes. In this bird the M. pseudotemporalis is squarish in cross-section and extends from the posterior wall of the orbit to the lower jaw. In F. montifringilla, on the other hand, the M. pseudotemporalis although similarly attached and inserted forms a fleshy floor to the orbit filling the spaces between the Ms. Adductor mandibulae posterior, Pterygoideus dorsalis and Ethmo-mandibularis. Yet, despite its size, the muscle probably does not contract with the same forces as its smaller counterpart in C. coccothraustes. The nature of the jaw muscles in C. coccothraustes is not the only factor contributing to a powerful musculature, but their size and distribution are also important. A comparison of the relative sizes and the areas of attachment of the various portions of the M. Adductor mandibular externus illustrates this point. It can be seen (Text-figs. 3 and Io) that the area of attachment of this muscle in C. coccothraustes is disproportionately greater than in F. montifringila. In C. coccothraustes the muscle arises from over most of the external surface of the cranium whereas in F. montifringilla the areas of attachment are confined to the lateral and postero- lateral surfaces. The essential difference between the myology of the jaw of C. coccothraustes and that of F. montifringilla appears to be one of degree, that is, area of attachment and tendinosity. However, what I have termed the “anterior portion” of the M. Pterygoideus ventralis lateralis appears to be absent in FP. montifringilla. In this species the M. Ethmo-mandibularis is more prominent and its site of origin extends over the comparable area of origin of the anterior portion of the M. Pterygoideus ventralis lateralis in C. coccothraustes. It would appear from this that the M. Ethmo- mandibularis is divided into two parts in C. coccothraustes. However, since in C. coccothraustes the anterior portion is inserted into the palatine along with the posterior portion of the M. Pterygoideus ventralis lateralis I am of the opinion that the anterior portion should be considered as a part of the M. Pterygoideus ventralis lateralis, and this seems justifiable since the portions are also functionally complementary. DISCUSSION The purpose of this paper has been to indicate some of the modifications in the morphology of the head of C. coccothraustes which appear to be related to its ability to crack open various fruit stones, such as those of the cherry, damson and olive. There are, however, several points requiring further investigation. 390 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH There is, for example, a point of interest in the development of the horny pads of the rhamphotheca. It seems that these pads, or bosses, do not appear until the bird reaches maturity. In the limited sample examined they were not found in nestlings nor in birds in immature plumage. Even if pads were present in young birds they would be non-functional because from an examination of the stomach contents it appears that the adults are primarily insectivorous during the breeding season (Mountfort, in press) which indicates that the nestlings are fed mainly on insects. Similarly, the pads are not required by immature birds because they have been ob- served to feed on caterpillars, especially the Hornbeam caterpillar, Carpinus betulus. The appearance of the horny pads late in the development of the individual could be interpreted as evidence that the structures have been acquired recently in the history of the species. However attractive this argument may appear to be at first sight it is misleading because the pads should not be considered apart from the asso- ciated osteological and myological modifications ; and it is possible that if the pads were present in an immature bird they would even be a danger to it! One of the chief characteristics of young passerine birds is the absence of osseous material in most of the cranium and the slow rate at which ossification occurs as the young bird matures. The absence of osseous material makes the skull extremely fragile, the more so since the sutures between the bones remain open almost until maturity. The skull of a young bird is, therefore, not strong enough either to wishstand the stresses of cracking open fruit-stone kernals or to accomodate the powerful muscu- lature capable of closing the jaws with a force in the region of too pounds, that is if the muscles could be precociously developed to contract with that force. It seems, therefore, that the late development of the pads cannot be regarded as indicative of their appearance in the phylogeny of the bird. CONCLUSIONS The structural modifications which have occurred in the head of C. coccothraustes and enable the bird to apply and withstand a force in the region of 100 lb. are pro- found. It appears that no part of the head, which has been considered, has escaped specialization. The oral surfaces of the horny bill are modified and equipped with striated pads between which the food is gripped. The skull, particularly the upper and lower mandibles, is strengthened to withstand the stress of cracking a hard fruit stone and it is modified to accommodate the powerful musculature of the jaw as well as to withstand the force of the contraction of the muscles. The musculature of the head is highly developed with large areas of attachment, and the individual muscles are tendinous in nature and powerful in action. The possession of these modifications are undoubtedly a selective advantage to the bird particularly during the period of great increases in population which partly coincides with the “ soft-fruit ”’ season. C. coccothraustes comes less into competition with other seed-eaters and its own young at an important time of the year by utilizing fruit-stones and similar large seeds, berries and, during the breeding season, large insects. THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH 391 APPENDIX Crushing Tests on Cherry and Olive Stones (a) The National Physical Laboratory carried out a series of crushing tests and made the following report : The apparatus used for the tests was designed to simulate, as far as possible, the mandibles and pressure pads of this bird. It consisted of a steel compression rig with two }-in. diameter rods which fitted into two deep grooves in the lower block of the apparatus, and a flat serrated cross piece attached to the upper block. The bare fruit stone was placed between these loading points and cracked; the load being increased by moving a jockey weight along the lever of the testing machine. The time taken to crack each stone was approximately 20 seconds. In order to represent the influence of the direction of the seam or suture of the stone in the bird’s beak, the tests were made with the seam facing several directions. The results of the tests are given in Tables I and II. (5) Mr. D. B. Welbourne, of the Department of Engineering, Cambridge, working independently obtained comparable values when he crushed cherry stones in a Housefield Tensometer. In correspondence, he writes, that in his tests he effected different rates of loading and found that more rapid loading resulted in higher failure loads, that is, a greater pressure is required to break a stone quickly than to break it slowly. Therefore, the crushing loads given in Tables I and II should not be regarded as the maximum pressures applied by C. coccothraustes when cracking fruit stones; for the values, determined by the tests, were obtained in each case by applying the pressure for approximately 20 sec. which is a greater time than it takes C. cocco- thraustes to crack a fruit stone. TABLE I.—Crushing tests on Cherry Stones Test Direction of Type of Cherry Crushing load seam (Ib.) Sideways . Dark 3 64 ” 62 ” : 5 : 70 Upwards 6 ” . 7s) ” é Downwards - fr 4 61 CO OMAN AUARWNH H Sideways Light F 95 NOH Upwards . » : 12, ” . ” . 7o on Ww 392 THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH TABLE II.—Crushing tests on fresh Olive Stones Test Direction of Type of stone Crushing load seam (Ib.) Sideways 3 Olive : III ”» bp 3 110 a “ 6 . 106 ” . re c 114 Upwards " * . 125 Rb : 8 ; 137 t : - é 147 os - *D : 143 Downwards . A “ 136 » : rp 107 x 4 : 138 » : *p F 159 On ANBRW NH Woe OH NHOW) Physic’s Division. National Physical Laboratory. REFERENCES BEECHER, W. J. 1953. A phylogeny of the Oscines. Awk, 70 : 270-333. Duiym, M. 1951. On the head posture in birds, and its relation to some anatomical features. Pt.l and Il. Proc. Kon. Ned. Akad. Wentensch. Ser. C, 54: 202-211; 260-271. DE Beer, G. R. 1951. Embryos and Ancestors. Oxford. EpGEwortH, F. H. 1907. The development of the Head-muscles in Gallus domesticus and the Morphology of the Head-muscles in the Sauropsida. Quart. Jour. Micro. Sci. N.S. 51: 511-556. 1935. The Cranial Muscles of Vertebrates. Cambridge. FIEDLER, W. 1951. Beitrage zur Morphologie der Kiefermuskulatur der Oscines. Zool. Jahrb. (Anat.), Jena, 71: 235-288. Gapow, H. 1891. Vogel: Aves. in Bronn’s Klass n. Ord. des Thier-Reichs. 6. Hate, J. R. 10945. Great-spotted Woodpecker breaking open almond nuts. Brit. Bds. 38 : 17. Horer, H. 1945. Untersuchungen iiber den Bau des Vogelschadels, besonders iiber den der Spechte und Steisshiihner. Zool. Jahrb. (Anat.) Jena, 69 : 1-158. 1949. Die Gaumenlucken der Végel. Acta. Zool. Stockholm, 30 : 209-248. 1950. Zur Morphologie der Kiefermuskulatur der Végel. Zool. Jahrb. (Anat.) Jena, 70 : 427-550. IncrRAM, C. 1941. Hawfinches and Cherries. Brit. Bds. 34 : 220. James, T. O. 10945. Great-spotted Woodpecker breaking open almond nuts. Brit. Bds. 38 : 274-275. Kirkman, F. B. 1911. The British Bird Book, 1:155. T.C. & E. C. Jack, London. Kraauw,C. J., VANDER. 1948-1952. Thesize and the position of the functional components of the skull. A contribution to the knowledge of the architecture of the skull. Extrait des Arch. Neerl. Zool. 9 : 1-4. Kripp, D. von. 19334. Der Oberschnabel-Mechanismus der Vogel. (Nach den Methoden der graphischen Statik bearbeitet). Gegenbaur, Morph. Jahrb., Leipzig, 71 : 469-554. 19330. Der Spezialisationsreiher der St6rche, Reiher und Kormorane von konstruktiven und biotechnischen Standpunkt. Gegenbaur. Morph. Jahrb., Leipzig, 72 : 60-92. 1933c. Beitrage zur mechanischen Analyse des Schnabel-Mechanismus. Gegenbaur. Morph. Jahrb., Leipzig, 72 : 541-556. THE MORPHOLOGY OF THE HEAD OF THE HAWFINCH 393 Laxjer, T. 1926. Studien tiber die Trigeminus-versorgte Kaumuskulatur der Sauropsiden, Kopenhagen, C. A. Reitzelg. LEBEDINSKY, J. 1921. Zur Syndesmology der Végel. Anat. Anz. Jena, 54 : 8-15. LEBEDINSKY, N. G. von. 1921. Der Unterkiefer der Vogel. Acta. Univers. Latviensis, 1: 12- 44- Motter, W. 1930-1931. Uber die Schnabel- und Zungenmechanik bliitenbesuchender Vogel. Tund II. Ein Beitrag zur Biologie des Blumenvogels. Biologia Generalis. 6 : 651-726; 7 : 99-154. Mountrort, G.R. The Hawfinch (in press). London, Collins. Owen, R. 1866. The Anatomy of Vertebrates. London, Longmans Green & Co. Parker, W. K. 1877-1879. On Aegithognathous Birds. Tyvans. Zool. Soc. Lond. 9 : 289-352; 10 : 251-314. Pycrart, W.P. 1905. Bull. Brit. Ornith. Club, 15 : 37. SHarPE, R. B. 1888. Cat. Bds. Brit. Mus. 12 : I-3. SHUFELDT, R. W. 1890. The Myology of the Raven, MacMillan, London. SUSHKIN, P. 1925. The Evening Grosbeak (Hesperiphona), the only American genus of a Palaearctic group. Auk, 42 : 256-261. STRESEMAN, E. 1927-1934. Aves. In Handb. d. Zool. herausgegeb. von Th. Krumbach, deseo: Watson, M. 1883. Report on the anatomy of the Spehniscidae collected. ‘Challenger’ Report. Sci. Results. Challenger, Zool. 7, H.M.S.O., London. WITHERBY, H. F., et al. 1939. The Handbook of British Birds, 1: 51-54, Witherby, London. WitHersy, H. F. & Ticrnurst, N. F. 1907. On the more important additions to our know- ledge of British Birds since 1899. Brit. Bds.1:1 52. nN nN ZOOL. 2, 13. i ill ar Ow Ee ee ee eee ey Tee ee ES es # Ay. Aa Wome ast oo ee 1 t i - \ 7h i ¥ : -2 SEP 1955 PRESENTED The page numbers of the principal references are printed in Clarendon type. INDEX TO VOE. H New taxonomic names are printed in bold type. abyssicola, Ophiactis Acanthephyra Acanthocyclus Acartia Achirus f ackleyi, Zoroaster . aculeata, Ophiopholis adspersum, Ophiernus adspersus, Ophiernus africana, Ilisha afzeliusi, Pellonula alaskensis, Haemogamasus albula, Stephanasterias alpinus, Salvelinus amaranthus, Phorbas Amphiophiura Amphipholis . Amphiura anancora, Stelletta anhelans, Tedania . annexa, Desmacella annulata, Astroporpo Anthenoides . antillarum, Plesiodiadema antillensis, Astropecten . Aphrocallistes Arctacaridae . Arctacarus Arctocephalus Arctoseius 5 arctuatus, Zercon . arenatus, Nymphaster argentata, Otaria argus, Puerulus : articulatus, Astropecten . assimilis, Hymeniacidon . Astropecten . Astrophyton . Astroporpa Atelecyclus aurantiifrons, Ptilinopus australis, Arctocephalus . Axinella balssi, Trachycarcinus balteatus, Ceramaster ecanalanis bandensis, Octopus barimae, Luidia 137, 137, 311, 163, 211 287 139, 151 163, 183 216, 228 140, 158 140, 155 - 140 216, 220 216, 229 216, 226 140 139 140 139 , 215, 218 287-307 299 311-336 287 296 139 323 17, 33 cae &X°) 216, 231 139 124,140 140 60 272, 275 324, 333 216, 229 a Se) 139, 148 | 9) 139, 145 bartletti, Tretocidaris Bathothauma Bathypolypus beatrix, Aphrocallistes Bentheledone Berlesiana 0 bidentata, Ophiacantha 2 biguttatus, Portumnus borealis, Lissodelphis Bosmina botryoides, encase bowerbanki, Halichondria bracebridgei Crepidacantha brevicaudum, Ophioderma Brisingella biirgeri, Panulirus . byronia, Otaria caecus, Phoberus calcinea, Haliclona californianus, Zalophus Callinectes Calliteuthis Callorhinus Callyspongia . Calveriosoma. canariensis, encocolenian Cancer . Capromys Carcharinus . Carcinides carinata, Tropiometra carpenteri, Pheronema catodon, Physeter . Caulerpa Ceramaster cetratus, Arctoseius Chaetodipterus Chiridota Chrysoena . 6 c cincta, Ptilinopus . clathrata, Luidia clavicorne, Phyllosoma coccothraustes, Gaccoth estes! coelebs, Fringilla Coelorhynchus Colomesus Columba 125, 140 125 + 97 215, 218 215, 218 216, 288 255 . 140 139, 151 25, 35 331, 332 1 (124 216, 224 334 137 125 - 334 216, 225 130, 140 215, 219 60, 64 2 (§24 163, 164 59 140 215, 217 103 2 go 139, 148 289 133 140 278 272, 274 139 25, 29 371-391 372 130 133 279 396 corallinus, Trachycarcinus coregoni, Bosmina . cornuloides, Halichondria coronata, Brisingella coronulatus, Ptilinopus Corystes Cosinodiscus . cranium, Tetilla Crepidacantha crinispina, Crepidacantha crispatus, Ctenodiscus Crossaster Ctenodiscus curiosus, Zercon cuspidatus, Goniaster Cystoseira Cystisoma Cythere danae, Macrotritopus dasypus, Panulirus denticulata, Perimela Desmacella diadema, Stephanotrochus Didunculus Dinychus Diplobatis distorta, Myxilla dohertyi, Ptilinopus dolabriformis, Ophionereis Dragmatyle . . Drepanoptila drobachiensis, Strongylocentrotus Ducula dupetithouarsii, Ptilinopus Echinaster echinata, Ophiocoma elegans, Octopus elegans, Ophiolepis elephas, Palinurus . elongatus, Johnius Eophona Epicrioidea Epicrius equivocus, Octopus Erimacrus erina, Haliclona esculenta, Tilapia . Eumetopias . excentricus, Coscinodiscus faber, Chaetodipterus facilis, Pellisolus Farrea é fenestralis, Zercon . fenestrata, Thenea INDEX : 59 e350) 216, 234 139, 151 272, 270 s 60 oy 245) 216, 223 243-201 248 139 139 139 206 139 44,49 : 7°-97 17, 25, 31, 35 57, 60-63 216, 226 123 280 » 305 163, 166 216, 226 272, 274 140, 155 216, 235 278 "rao 270, 278 273, 276 133 206 215, 217 e207 215, 219 fischeri, Ptilinopus fluviatilis, Labidocera Fringilla Fungia galapagoensis, Arctocephalus . galilaea, Tilapia gargantua, Mesothuria giesbrechti, Acartia girigan, Tilapia - glabrata, Hymeniacidon . glaucus, Trachycarcinus . Globicephala goliath, Ducula Goniaster gracilis, Octopus gracillima, Amphipolis granadensis, Ophioplinthaca grandis, Crepidacantha granularis, Ceramaster granulifrons, Ptilinopus . greyii, Ptilinopus grimaldii, Spherosoma guttatus, Panulirus Haemogamasus Halichondria Haliclona Halimeda Haplochromis Harengula Henricia Hesperiphona Higginsia Hircinia Holothuria homarus, Panulirus hoylei, Calliteuthis hunteri, Tilapia huttoni, Ptilinopus Hymeniacidon - hyogastra, Ptilinopus hystrix, Calveriosoma Tlisha Inachus F Incisa, GeWiopliatiaes inermis, Puerulus infernalis, Vampyroteuthis Inflatella o inflatus, Panulirus . inornata, Desmacella insolitus, Ptilinopus insularis, Ptilinopus iozonus, Ptilinopus jambu, Ptilinopus japonicus, Panulirus Jasus 269, 273, 275, 276 sae. 133 372 123 311 351 140 Stas 351, 355 216, 231 47, 59 216, 232 216, 223 130 >) 9350 163, 211 : : | 5139 372, 385 216,285 216, 236 140 17-41 125 364 216, 281 273, 277 130, 140 137, 163, 241 enk24 216, 228 28, 33, 34 216, 226 273, 277 273, 276 273, 277 273, 278 17, 31 17, 18, 20 jipe, Tilapia . ° é Johnius 3 jubatus, Eumetopias : Justitia 6 D : kempi, Macrotritopus_ . kirkpatricki, Crepidacantha koefoedi, Searsia korogwe, Tilapia mossambica kroyeri, Xiphopenaeus labiatus, Coelorhynchus Labidocera Laelaptidae f laevis, Chiridota . c lalandei, Jasus laterincisus, Actoseius latimanus, Callinectes latipes, Portumnus layardi, Ptilinopus leclancheri, Ptilinopus Leptasterias . Leptoderma . 5 leucas, Carcharhinus Leucosolenia . Leucotreron . a lilljeborgi, Acartia . Lissodelphis . longimana, Justitia longimanus, Carcharhinus longiseta, Crepidacantha Luidia : 5 luteovirens, Ptilinopus lutzi, Pseudoceradocus lyromma, Bathothauma . lytteltonensis, Trachyleberis macrochir, Tilapia . Macropodia macrops, Leptoderma Macrotritopus a . maenas, Carcinus marchei, Ptilinopus Marginatus, Astropecten Megaloprepia melaena, Globicephala melanopleura, Tilapia. melanospila, Ptilinopus meleagris, Stomolophus . mercieri, Ptilinopus merrilli, Ptilinopus Mesothuria metabula, Amphiophiura metallacta, Ophiacantha Mexicana, Holothuria micropunctatus, Dinychus monacha, Ptilinopus INDEX 351, 354 137 + 332 17, 18 1297 256 163, 201, 206 133, 351, 856 272, 274 - 139 163, 183 163, 164 215, 218 , 278, 278 133, 135 341-345 : 18 164 + 251 139, 142 273, 278 134 125 + 3-15 « 351 . 59 163, 183 69-97 397 Monodi, Sirpus “ : 44, 52-56 montifringilla, Fringilla . 372 mortenseni, Rossella 215, 218 mossambica, Tilapia 351 Mugil - 134 multidentatus, Aretoseius 289, 291 muricatum, Astrophyton 124, 140 Mycerobas a - 385 Myxilla 7 216, 226 naina, Ptilinopus 273, 277 natalensis, Tilapia . - 358 nigra, Tilapia A Shr nilotica, Tilapia 351 nodosa, Stegophiura 140 nodulosus, Xenodermichthys 196 Normichthys 206 nuchalis, Tachysurus 134 Nymphaster 139 occa, Farrea 215, 217 obscurus, Carcharinus A 164 occipitalis, Ptilinopus 272, 274, 281 Octopus 69-097 odontostoma, Crepidacantha : 258 operosa, Normichthys 206 Ophiacantha . : 40 Ophiactis 140, 154 Ophiernus 140, 159 Ophiocoma - 140 Ophiocten 140 Ophioderma 140 Ophiolepis 140 Ophiomitra + 140 Ophiomusium 140, 159 Ophionereis 140, 155 Ophiopholis . - 140 Ophioplinthaca 140, 151 Ophiura 140 ornatus, Arctoseius 289 ornatus, Panulirus F 28, 31, 41 ornatus, Ptilinopus 272, 275, 281 osculum, Halichondria 216, 233 Otaria . 323 Otariidae 311-336 pachastrelloides, Severs 216, 220 Pagurus, Cancer - 64 Palaemon fi 134 panicea, Halichondria 216, 232 Papposus, Crossaster 139 Palinuridae . < 17-42 Pangani, Tilapia 351, 356 Panulirus 17-42 Parazercon 295 parvipora, Crepidacantha crinispina 250 Pellisolus A A - 206 Pellonula : : 137, 163, 211 398 pellucidus, Phyllophorus pellucidus, Puerulus Peltarion Penaeus : penicillatus, Panulirus pensacolae, Harengula pentacrinus, Ophicantha Perissospiza . perlatus, Ptilinopus peroni, Lissodelphis perousii, Ptilniopus Persparsia Pheronema ° philippi, Aretocephalus Phoberus Phocaena phocoena, Phocaena Phorbas . Phormosoma Phronimae Phyllophorus Physeter Phytoseiidae pictus, Diplobatis . piercei, Anthenoides pinnatus, Synaphobranchus Pirimela placenta, Phormosoma Plesiodiadema Plocamionida Podocactes poissonil, Crepidacantha | polaris, Leptasterias polycoeca, Searsia polyphagus, Panulirus porphyraceus, Ptilinopus porphyrea, Pee Portunus Prodinychidae Prozecon Pseudoceradocus Pseudocorystes psittacus, Colomesus Ptilinopus Puerulus 3 pulchellus, Ptilinopus purpuratus, Ptilinopus Radiella ramosa, Axinella Ramphiculus rarispina, Salenocidaris rarotongensis, Ptilinopus regina, Ptilinopus . regius, Panulirus richardsil, Ptilinopus rivoli, Ptilinopus Rosaura 163 311 269, 272 271 270, 273, INDEX 140 17 60 133, 134 . 3r 163, 211 140 385 272, 275 341-345 272, 276 206 ,208 215, 217 322, 333 124 104 3 ~6FO4 216, 228 140 125 140 133, 134 267-284 ue kL 272, 276 276, 281 216, 222 216, 229 274, 278 273, 277 163, 167 rosaurae, Luidia rosaurae, Sphaerodactylus roseicapilla, Ptilinopus Rossella 4 rostratus, Aneeacane rostratus, Searsia rotunda, Rosaura tubens, Haliclona . Sagitta Salenocidaris Salvelinus js sanguinolenta, Henricia sarsi, Ophiura savignyi, Ophiactis scabrocuneata, Cythere . scabrocuneata, Trachyleberis . Scaeurgus scorpio, Macrotritopus Searsia . Seiodes senegalensis, Luidia sentus, Echinaster sericeum, Ophiocten setigera, Crepidacantha . Sirpus . : sol, Radiella . solea, Crepidacantha solenites, Zercon solomonensis, Ptilinopus. Sperosoma Sphaerodactylus spiculosa, Haliclona spinnulifer, Trachycarcinus Spirula squinado, Maia Stegophiura . Stelletta Stellifer stellifer, Stellifer Stenella Stephanasterias Stephanotrochus Stomolophus strigilata, Higginsia Strongylocentrotus Stryphnus subgularis, Pailinapus) sublaevis, Dinychus sundevalli, Amphiura superbus, Ptilinopus Symphurus - Synaphobranchus . taaningi, Persparsia Tachysurus tannensis, Ptilinopus Tedania Telmessus 163, 199, 163, 270, 139, 142 123 272, 276 215, 218 299-305 - 206 163, 167 216, 223 133, 135 ~ 140 163, 183 139 140 140 74-97 7°-97 201, 206 295 139 139 140 250 43-64 216, 222 « ese 20906 tenera, Halichondria - tenerrima, Callyspongia . tenerrima, Haliclona teres, Crepidacantha poissonii . Tethyaster Tetilla tetracirrhus, Pteroctopus Thenea Thia thoracatus, Capromys Tilapia topsenti, Dragmatyle topsenti, Plocamionida townsendi, Arctocephalus Trachycarcinus Trachyleberis - Treron . 6 Tretocidaris . triangularis, Zercon Trichopeltarion Tripneustes Tropiometra . unicirrhus, Scaeurgus ursinus, Callorhinus ursinus, Seiodes valida, Ophiomitra validum, Ophiomusium . INDEX ao HR 216, 225 216, 225 - 247 139, 141 216, 223 96, 97 215, 219 . 61 124 349-368 216, 235 216, 229 321 44-64 - 3-15 279 140 296 60 140 140 TOOT, 334 295 = 140 140, 159 Vampyroteuthis variabilis, Hircinia variabilis, Tilapia . ventricosus, Tripneustes verrucosa, Maia versicolor, Panulirus vesitus, Tethyaster vespertina, Hesperiphona victor, Ptilinopus viridis, Inflatella viridis, Ptilinopus vulgaris, Octopus vulgaris, Palinurus wallacii, Ptilinopus weberi, Arctoseius wollebaeki, Zalophus wyvilli, Thenea Xenodermichthys . Xiphopenaeus Zalophus : zariquieyi, Sirpus zelanica, Crepidacantha . Zercon . A Zerconoidea . Zoroaster 399 eT 216, 236 351, 364 140 58, 59 25, 26, 31 139, 144 372 273, 278 216, 228 273, 277 5 Ge) 18 272, 275 289, 291 = 320 215, 220 - 196 134 320, 334 +43-51 - 254 294, 296 295, 296 139 PRINTED IN GREAT BRITAIN BY ADLARD AND SON, LIMITED, ~ rs BARTHOLOMEW PRESS, DORKING. _ = - ~ Be s ty ‘