. ■ , • - . - / - - * , •• ■ v - -• - ■ - 1 .. * : , - .. . ■' ' - ■■ - - _ - ' K - -r' ;■ . ’ " 1 ■ , - 1 . ... ■ : •• , . f.% ■ • . .• ' •" ' T.V - r-' . ' • • ■ . : • . • ■ ; ; ■ BRITISH MUSEUM (NATURAL HISTORY). BRITISH ANTARCTIC (“TERRA NOVA”) EXPEDITION, 1910. NATURAL HISTORY REPORTS. ZOOLOGY. VOL. IV. ECHINODERMA. PROTOCHORDATA. BIRDS. LONDON : PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM. Sold by B. Quaritch, Ltd., 11, Grafton Street, New Bond Street, London, W.l ; Dulau & Co., Ltd., 2, Stafford Street, London, W.l ; Oxford University Press, Warwick Square, London, E.C.4; AND AT THE British Museum (Natural History), Cromwell Road, London, S.W.7. 1917-1935. [All rights reserved .] ZOOLOGY. VOL. IV. CONTENTS. 1. Bell, F. Jeffrey. Echinoderma. Part I. Actinogonidiata. [ Issued July 28, 1917.] Pp. 1-10, 2 pis. 2. Ridewood, W. G. Cephalodiscus. [ Issued March 23, 1918.] Pp. 11-82, 5 pis., 1 map. 3. MacBride, E. W. Echinoderma. Part II, and Enteropneusta. [Issued April 24, 1920.] Pp. 83-94, 2 pis. 4. Norman, J. R. Rhabdopleura. \ Issued July 23, 1921.] Pp. 95-102. 5. Lowe, P. R., and Kinnear, N. B. Birds. [ Issued July 26, 1930.] Pp. 103-193, 16 pis. 6. Garstang, W. Report on the Tunicata. Part I. — Doliolida. [ Issued May 27, 1934.] Pp. 195-251. 7. Parsons, C. W. Penguin Embryos. [ Issued September 3, 1934.] Pp. 253-262. 8. Garstang, W., and Georgeson, Elizabeth. Part II. — Copelata. Report on the Tunicata. [ Issued May 25, 1935.] Pp. 263-282. BBBE7 HISTORY). n\ nvTvrmrr^ ' U - : 7 *-LVAl| >RY REPORT. .. IV, No. I. Pp. 1-10. HI ART I.-ACTINOi PV * y~ zM- BY FFREY BELL, M.A., ' of Kind .< College, University of London ; tyW&j British Museum {Natural History). gX Wvtfc i B. Mj CN.H.) ills LONDON ? THE TRUSTEES OP THE ETltTXSII MUSEUM. f Row, E.j|| ; B. QvAarrcH, U, Gwfton Street, New Bond Street, W. i ; 6 0o,P/ Lm, 37. Soho Square, % 1 ; " : 'Aim AT TUB At-.HiaTOliT), Cromwell Rosul, , London, S.W.7. mm Jfy ■-- ■ * * 1 ’ : , [Alt rigltls cwowAl ► fM fr i *■ ’ ' ,'K »»<4 *» <*S©»«5i5' •\ - rice Tvvo Shillings 'and Sixpence. iM - rmr Q0rift60 Qttueeum ((JXafuraf J)t6forg)+ This is No 3 of 25 copies of “ Terra Nova" Echinoderma , Vol. IV., Part /., printed on Special paper. 1 ECHINODERMA. PART I.-ACTINOGONIDIATA. BY F. JEFFREY BELL, M.A., Emeritus Professor and Fellow of King's College, University of London ; Assistant in the Department of Zoology, British Museum ( Natural History). Descriptions of Species — PAGE A. Antarctic ..... 1 B. New Zealand ..... 5 C. South Atlantic .... 8 A. ANTARCTIC. LIST OF STATIONS. Stations. Locality. Depth. 194 Off Oates Land, 69° 43' S., 163° 24' E 180-200 fathoms. 220 Off Cape Adare, mouth of Robertson’s Bay ..... 45-50 fathoms. 294 Ross Sea, 74° 25' S., 179° 3' E 158 fathoms. 295 „ „ 73° 51' S., 172° 57' E 190 fathoms. 314 5 miles N. of Inaccessible Island, McMurdo Sound 222-241 fathoms. 316 Off Glacier Tongue, about 8 miles jST. of Hut Point, McMurdo Sound 190-250 fathoms. 331 Off Cape Bird Peninsula, entrance to McMurdo Sound . 250 fathoms. 338 Ross Sea, 77° 13' S., 164° 18' E 207 fathoms. 339 „ „ 77° 5' S., 164° 17' E 140 fathoms. 340 „ „ 76° 56' S., 164° 12' E 160 fathoms. 341 Off Cape Bird Peninsula, entrance to McMurdo Sound . 80 metres. 348 Off Barne Glacier, McMurdo Sound ...... 200 fathoms. 355 ' Ross Sea, 77° 46' S., 166° 8' E 300 fathoms. 356 Off Granite Harbour, entrance to McMurdo Sound 50 fathoms. As I have already made two Reports on Ecliinoderms collected in the area visited by the “Terra Nova” (see “Report on the Collections of Natural History made in the Antarctic Regions during the Voyage of the ‘ Southern Cross/ ” London, 1902, pp. 214-220, which I shall refer to as Bell (l), and Vol. IV of the “Natural History” of the National Antarctic (“Discovery”) Expedition, 1901-1904, London, 1908, B VOL. IV. 2 TERRA NOVA” EXPEDITION. pp. 1-16, which I shall refer to as Bell (2)), there is no need to give in detail those references to other writers which have already been made. As several species come from many stations — Cycethra verrucosa, for example, having been found at Stations 194, 220, 294, 316, 331, 338, 340, 355, and 356 — and as these stations differ very little from one another except in their numbers, I do not propose to burden the text, which I hope to keep as concise as possible, with stating these numbers. Much more do I wish to emphasise a very important fact. The collection of specimens — in all nearly 500 have been registered, and some 200 put aside for exchange — is very large, although the number of species is comparatively small, but every individual specimen had to be closely scrutinised, for the variation, which I thought I was prepared for by previous experience, was quite beyond my expectation. We are, therefore, brought face to face with an altogether unexpected phenomenon, which in the plainest and most concrete form may be thus stated : — In the area dredged by the “Southern Cross,” the “Discovery,” and the “Terra Nova,” where the conditions of depth and temperature are practically uniform, there is extraordinary variety in the characters of most of the Echinoderms there collected. CRINOIDEA. 1. Promachocrinus lcerguelensis, P. H. Carp. Promachocrinus icerguelensis, P. H. Carp., Bell (2), p. 3. This species is obviously very widely distributed in the Benthos of the area with which we are dealing ; from the fact that a single specimen only was taken by the “Challenger” it was supposed to be rare, but the “Terra Nova” collections show that this view is quite a mistake. 2. Antedon antarctica, P. H. Carp. Antedon antarctica , P. H. Carp., Bell (2), p. 4. Among the examples of this now well-known form one was found to have six rays. Stations 194, 295. 3. Antedon adriani, Bell. Antedon adriani , Bell (2), p. 4. 1 have not noted any points which I think it necessary to add to the diagnosis I have already published of this species. I may be allowed to call to mind that this species got its name from the fact that I wished to associate it with the heroic Dr. Wilson, and, having already named a species A. wilsoni, I was called on to select one of his Christian names. Stations 194, 314, 348. ECHIXODERMA, I— BELL. 3 ECHINOIDEA. 4. Notocidaris mortenseni, Koehler. Goniocidaris mortenseni, Koehler, Bull. Ac. Belg. XXXII (1890), p. 816. Notocidaris mortenseni, Mortensen, Deutsch. Siid. Pol. Exp. XI (1909), p. 18. Station 294, and south of Balleny Island (69° 43' S., 163° 24' E.). 5. Ctenocidaris perrieri, Koehler. Ctenocidaris perrieri , Koehler, 2e Exp. Antarct. Fran?. (1912), p. 150. S. of Balleny Island. 6. Echinus margaritaceus , Lamk. Echinus margaritaceus, Lamk., Bell (1), p. 219. Stations 220, 338, 339, 340, 355. 7. Ahatus shackletoni, Koehler. Abatus shackletoni, Koehler, Brit. Antarct. Exp. 1907-9, Yol. II (1911), p. 51. One specimen only, from Station 339. 8. Pseudabatus nimrodi, Koehler. Pseudabatu8 nimrodi, Koehler, id., tom. cit., p. 60. Stations 294, 339, 355. This and the preceding species have been so fully described by my learned friend that it appears quite unnecessary for me to add anything to what he has said in such full detail. They are both new to the collection. ASTEROIDEA. 9. Asterias long staff i , Bell. Asterias longstajfi, Bell (2), p. 7. Station 356. 10. Asterias antarctica, Liitk. Asterias antarctica, Bell (1), p. 215. Stations 294, 295, 331, 338, 339, 340. 11. Asterias pedicellaris, Koehler. Asterias pedicellaris, Koehler, Zool. Anz. XXXII (1907), p. 145. Stations 339, 340. 12. Asterias brandti, Bell. Asterias brandti, Bell (2), p. 7. Station 339. b 2 4 TERRA NOVA” EXPEDITION. 13. Heuresaster hodgsoni, Bell. Heuresaster hodgsoni , Bell (2), p. 8. Stations 316, 338, 340, 356. This species is rather more common than I was led to suppose by the “ Discovery ” collection. A form which I think must be a young stage of this species was taken at Station 339. 14. Cycethra verrucosa, Philippi. (PI. I, figs. 1-6.) Cycethra verrucosa, Philippi, Bell (1), p. 215; id. (2), p. 10. If a mystic wanted a type of human life he might well take this species, so astounding are its variations and its modifications, some of which are shown in Plate I. Here, again, we have a warning as to coming to conclusions as to the rarity of species ; described many years ago by Philippi from a single specimen, of which a good many years later I also got a single example, it is now known to be exceedingly common and very widely spread. Numerous Antarctic stations ( see p. 2.). In Plate I are given six photographs of Cycethra verrucosa to show the variations in form, proportion, and ornamentation. Further light will be found in Prof. Perrier’s report on the Cape Horn specimens. Sladen’s three sets of two figures each are (i) of natural size, (ii) magnified 1^ diameter, and (iii) magnified 2 diameters, and are, therefore, almost useless for purposes of comparison, though admirable for the establishment of distinct species. 15. Solaster lorioli, Koehler. Solaster lorioli, Koehler, Zool. Anz. XXXII (1907), p. 144. Numerous Antarctic stations. 16. Solaster octoradiatus, Ludwig. Solaster octoradiatus, Ludwig, Bell (2), p. 11. This species wras only taken from south of Balleny Island. 17. Cryaster antarcticus , Koehler. Cryaster antarcticus, Koehler, Exp. Antarct. Francj. 1903-5 (1906), p. 24. Station 339. OPHIUROIDEA. 18. Ophionotus victoriae, Bell. Ophionotus victoriae, Bell (1), p. 219. Station 341. ECHINODERMA, I— BELL. 5 19. Ophiosteira antarctica, Bell. (PI. II, figs. 1-3). Ophiosteira antarctica, Bell (1). p. 218. This species, in its wide distribution and its great variability, is own brother to Cycethra verrucosa. Plate II shows three photographs of the remarkable differences in the plates of the disc. Numerous Antarctic stations. 20. Ophiacantha vivipara, Ljungman. Ophiacantlia vivipara, Ljungman, Bell (2), p. 13. Station 220, and off Inaccessible Island (Station 314). 21. Astrotoma agassizi, Lyman. Astrotoma agassizi, Lyman, Bell (2), p. 15. This also proves to be an exceedingly common form. Stations 339, 341, 348. B. NEW ZEALAND. The basis of our knowledge of the Echinoderm fauna of New Zealand is an essay with that title published in the Proceedings of the Linnean Society of New South Wales, 1898, by Mr. H. Farquhar. Eleven years later Prof. AY. B. Benham, F.R.S., published a report on the Echinoderms collected by the New Zealand Government trawling Expedition in the Records of the Canterbury Museum, Arol. 1. Among the specimens collected by the “ Terra Nova ” which were unknown to me were two Astrophiurids,* which I thought possibly might be recognised by Prof. Benham. This was not so, but he was kind enough to interest himself with the specimens, and to make on them some observations which have been of great value to me. I have hardly any doubt that they are now described for the first time ; the delay consequent on communicating with New Zealand has allowed Mr. IT. L. Clark to describe a new species of Ophiomyxa, wTiich was represented in this collection by a single specimen that might easily have been mistaken for 0. australis. The discovery of two specimens of Ophiothrix off the coast of New Zealand solves a problem of some importance, for, widely distributed as this genus is, it has not until now been known as a member of the New Zealand fauna. No Astrophiurid was known from the waters of New Zealand till the year 1900, Cf. Bell, P. Z. S. 1892, p. 180. 6 TERRA NOVA” EXPEDITION. when Mr. Farquhar* described a form which he called Ophiocreas const rictus ; this Doderleinf has removed to the genus Astrochema. The only other Astrophiurid known from New Zealand was described by Prof. Benham in 1 909 J under the name of Astrotoma waitei. In describing, therefore, three new species of this division I more than double the number of species recorded from New Zealand waters. I. ECHINOIDEA. 1. A stropyga radiata, Leske. Three specimens from Station 96, 70 fathoms, 7 miles E. of North Cape. 2. Laganum, sp. Two young specimens, which it is impossible to determine with accuracy, were taken at Station 96, 70 fathoms, off North Cape. II. ASTEROIDEA. 3. Aster ina regularis, Verrill. Off New Zealand. 4. Astropecten polyacanthus, M. Tr. Bay of Islands, and off North Cape. III. OPHIUROIDEA. 5. Pectinura, sp., juv. Station 96, E. of North Cape. 6. Ophiothrix, sp. The fact that this genus has not been recorded from the neighbourhood of New Zealand is now one of the commonplaces of geographical distribution, but the genus is spread so widely through the seas that it was only natural to suppose that it would be discovered within this area, and it remained for the naturalists of the “ Terra Nova” to find a single small specimen at a depth of 70 fathoms east of North Cape (Station 96), in company with a very small specimen of Astroporpa wilsoni, and two others from Station 144 (near Cape Maria van Diemen). * Trans. N. Z. Inst. XXXII, p. 405 (1900). f Beitrage z. Naturg. Ostasiens, p. 113 (1911). ± Rec. Canterbury Museum, I, p. 101 (1909). ECHINODERMA, I— BELL. 7 I do not recognise the species, which is probably new, and will add to the growing list of undescribed species, which can only be satisfactorily diagnosed when, a complete revision is made of this extensive genus. 7. Ophiomyxa brevirima, H. L. Clark. Ojpltiomyxa brevirima, H. L. Clark, Mem. Mus. Comp. Zool. XXV, no. 4 (December 1915), p. 169. There is a single representative of this species, quite recently described by Mr. H. L. Clark from some specimens obtained at Otago Harbour, N.Z. Station 96, east of North Cape : 70 fathoms. Mr. Clark appears to doubt whether the specimens from New Zealand, which have been attributed to 0. australis, have been correctly determined. As there is not in the collections of the Museum any specimen from any locality nearer than 39° 32' S., 171° 48' E., I am unable to discuss this question ; but if the records are correct, 0. australis is easily to be obtained from the shallow waters of New Zealand. 8. Ophiocreas constriction, Farquhar. Station 91, 25 miles off Three Kings Islands : 300 fathoms. 9. Astroporpa wilsoni, sp. n. Station 96, E. of North Cape : 70 fathoms. The upper surface of the disc has a more or less distinct central area covered with tubercles smaller than those of the periphery. These are well developed, and set in very regular rows, separated from one another by bare bands, and continuing the same character and arrangement on the upper surface of the arms. The lower interbrachial areas of the disc are shagreen-like, very delicate and elegant ; the arm-spines are numerous, very delicate, with minutely roughened surfaces. The tips of the arms form a more than usually close coil. There is a total absence of ornamentation from the plates, both of the arms and disc. Colour, creamy white. Diameter of disc, 13 ; 19 mm. Length of arms (about) 50 ; 60 mm. This genus was not known before from New Zealand. 10. Astroschema elegans, sp. n. Station 96, E. of North Cape : 70 fathoms. A somewhat variable species, but one that can easily be recognised by its well- marked coloration ; preserved in spirit, it exhibits transverse reddish-brown bands separated by white spaces, which are often, though by no means always, wider than the red bands. 8 “TERRA NOVA” EXPEDITION. The disc, which is small, is marked by five pairs of rows of prominent plates, which might at a superficial view be taken for radial shields. These plates vary slightly in size and number, and are more irregular in larger than in smaller specimens. The arms are long and coiled somewhat at their ends ; there are no continuous arm-plates, but the calcareous plates deposited on the upper surface are arranged with great regularity, are oblong in form, and extend on to the sides of the arm, where they are less well developed ; the spines on the lower surface are set in a single row on either side of the median furrow. The mouth is guarded by five prominent spines. Diameter of disc, 9 ; 11; 12 mm. Length of arms (about) 75 ; 90 ; 130 mm. 1 1 . Astrotoma benhctmi, sp. n. Station 91, 25 miles off Three Kings Islands: 300 fathoms. A single specimen. The disc is completely covered with rather large coarse tubercles, not all of the same size, but the interbrachial spaces below are almost bare ; the mouth-spines form simple clusters, and the papillae are encircled by well-marked granules which become spiniform towards the periphery. The arm-joints are closely covered by a pavement of granules, some of which are larger than others ; there are two or three arm-spines on the sides of each joint. The arms are, in some specimens, closely coiled and creamy yellow in colour. Diameter of disc, 11 mm. Length of arm, 40 mm. (ca.). C. SOUTH ATLANTIC. The collection from the South Atlantic is not of particular interest, except for the discovery of a new species, the description of which will be found below ; from Station 36 (South Trinidad) come Echinometra lucunter * of Linnaeus and Echino- cardium australe of Gray; from Station 38 (52° 23' S., 63° 50' W., near the Falkland Islands) were taken a considerable number of specimens of Ophiaccuitlia vivipara, Ljungman, some with seven rays and some with nine, and some bearing young. From Station 42 (22° 56' S., 41° 34' W., near Kio de Janeiro) were taken a number of Astropecten pontoporaeus , Slaclen ; from Station 42 also came the new species. Luidia scotti, sp. n. This species belongs to the group of L. clathrata, as defined by Mr. Sladen, but it is at once distinguished from that widely distributed form by greater delicacy of structure. * For the synonomy of this form, see Loven “ On the species of Echinoidea described by Linnaeus ” in Yet. Akad. Hdlgr. 13, IY, no. 5 (1887). ECHIXODERMA, I— BELL. 9 The fire arms are marked with irregular patches of brown ; they are rather delicate and tapering ; their marginal plates are square, very regular, and equal in the upper and lower rows. Inside the supero -marginals is a regular row of plates exactly like them, and within these are a number of rows of plates, which gradually decrease in size, and become more irregular in arrangement. The arms are quite flat and taper regularly ; there are two arm-spines, the lower of which is much the longer ; the inner part of each arm-plate is crowded with spiniform granules of various sizes, and irregular in their distribution. All the spines are pure white, while the podia are yellow. The characters of this species, of which several specimens were collected, appear to be pretty constant. R. r. 58 mm. 65 mm. 69 mm. 11 9 10 Station 42, near Rio de Janeiro. VOL. IV. c 10 “TERRA NOVA” EXPEDITION. Abatus shackletoni, 3. adriani, Antedon, 2. agassizi, Astrotoma, 5. antarctica, Antedon, 2. ,, Asterias, 3. „ Ophiosteira, 5. antarcticus, Cryaster, 4. Antedon adriani, 2. ,, antarctica, 2. Asterias antarctica, 3. „ brandti, 3. ,, longstaffi, 3. „ pedicellaris, 3. Asterina regularis, 6. Astropecten polyacanthus, 6. ,, pontoporaeus, 8. Astroporpa wilsoni, 7. Astropyga radiata, 6. Astroschema elegans, 7. Astrotoma agassizi, 5. ,, benhami, 8. „ waitei, 6. australe, Echinocardium, 8. australis, Ophiomyxa, 5, 7. benhami, Astrotoma, 8. brandti, Asterias, 3. brevirima, Ophiomyxa, 7. constriction, Ophiocreas, 7. Cryaster antarcticus, 4. Ctenocidaris perrieri, 3. Cycethra verrucosa, 2, 4. Echinocardium australe, 8. Echinometra lucunter, 8. Echinus margaritaceus, 3. elegans, Astroschema, 7. Goniocidaris mortenseni, 3. Heuresaster hodgsoni, 4. INDEX. hodgsoni, Heuresaster, 4. Laganum sp., 6. longstaffi, Asterias, 3. lorioli, Solaster, 4. lucunter, Echinometra, 8. Luidia scotti, 8. kerguelensis, Promachocrinus, 2. margaritaceus, Echinus, 3. mortenseni, Goniocidaris, 3. ,, Notocidaris, 3. nimrodi, Pseudabatus, 3. Notocidaris mortenseni, 3. octoradiatus, Solaster, 4. Ophiacantha vivipara, 5, S. Ophiocreas constrictum, 7. Ophiomyxa australis, 5, 7. ,, brevirima, 7. Ophionotus victoriae, 4. Ophiosteira antarctica, 5. Ophiothrix sp., 6. Pectinura sp., 6. pedicellaris, Asterias, 3. perrieri, Ctenocidaris, 3. 1 polyacanthus, Astropecten, 6. pontoporaeus, Astropecten, 8. Promachocrinus kerguelensis, 2. i .. Pseudabatus nimrodi, 3. radiata, Astropyga, 6. regularis, Asterina, 6. scotti, Luidia, 8. shackletoni, Abatus, 3. Solaster lorioli, 4. ,, octoradiatus, 4. verrucosa, Cycethra, 2, 4. victoriae, Ophionotus, 4. vivipara, Ophiacantha, 5, 8. wilsoni, Astroporpa, 7. LONDON : PRINTED BY WILLIAM CLOWES AND SONS, LIMITED, DUKE STREET, STAMFORD STREET, S.E. 1, AND GREAT WINDMILL STREET, W. 1. Echinoderma, Part I, PI. I. PLATE I. Cycethra verrucosa, Phil. Figs. 1-6.- — Six photographs to show the differences in the appearance of specimens due to the differences in the proportions of II to r, the varying bluntness of the arms, the form of the interbrachial curve, and the differences in the delicacy or coarseness of the granulation. All from Station 314. * Brit. Mus. (Nat. Hist.) Brit. Antarctic (Terra Nova) Exped.1910. Z oology, Vol. TT. Echinoderma, Parti, P1I. J. Green., Photo, del. Bemrose. Collo., Derby Echinoderma, Part I, Pi. II. PLATE II. Ophiosteira antarctica, Bell. Figs. 1-3. — Three photographs to show three very different stages in the size of the radial shields, and in the formation of the arm-joints. All from Station 316. ' s ' Brit. Mas. (Nat. Hist.) Brit. Antarctic (Terra Nova) Exped.1910. Z o olo gy, Yol. 1Y. Echino derma Part I, PI II. Q5rift60 (gtuseum (QTafurdf J^tsforg) This is No. Q °f 2 5 copies °f “ Terra Nova ” Cephalodiscus , printed on Special paper. It CEPHALODISCUS. BY W. G. RIDEWOOD, D.Sc WITH TWELVE FIGURES IN THE TEXT, FIVE PLATES AND A MAP. CONTENTS. PAGE Material ........... 12 Localities ........... 12 Historical Introduction ........ 13 Cephalodiscus and Rhabdopleura . . . 14 Sub-genera of Cephalodiscus ........ 1G Zooids of Cephalodiscus . . . . 20 Cephalodiscus ( Idiothecia ) evansi, n. sp. • . 26 Cephalodiscus ( Idiothecia ) nigrescens, Lankester . . . 31 Material collected ......... 31 Zooids .......... 33 Cephalodiscus ( Orthoecus ) densus, Andersson ..... 37 Material collected ......... 37 Zooids .......... 43 Buds ........... 48 Cephalodiscus ( Demiothecia ) hodgsoni, Ridewood . 48 Material collected . . . . . . . ... 48 Forms A and B of C. hodgsoni . . . . . . 51 Species of the Sub-genus Demiothecia . . . . 57 Empty Coenoecia of Cephalodiscus . . . . 60 Zooids of C. hodgsoni ........ 61 Foreign Ova found on the (Joenoecium . . . . . 65 Synopsis of the Species of Cephalodiscus at present known, and List of all Recorded Specimens ........ 66 Sub-genus Demiothecia, Ridewood . . . . . 66 Sub-genus Idiothecia, Ridewood . . . . . . 70 Sub-genus Orthoecus, Andersson . . . . . . 74 Summary ........... 77 References .......... 78 Index ........... 81 vol. iv. n 12 “TERRA NOVA” EXPEDITION. MATERIAL. The material of Cephalodiscus collected on the “Terra Nova” Expedition is of considerable hulk, and arrived at the British Museum (Natural History) on February 5th, 1914, in sixty-six bottles. Half of the bottles are of a capacity of litres, the others are one-litre bottles, or of smaller size. Almost the whole of the material is preserved in 4 or 5 per cent, formalin solution, a small quantity of it in 70 per cent, alcohol. A few zooids of Cephalodiscus hodgsoni, C. nigrescens and C. densus had been specially fixed in corrosive sublimate solution, and sub- sequently transferred to 70 per cent, alcohol. The material obtained from New Zealand proves to belong to a new species, for which the name Cephalodiscus evansi is proposed (p. 26). Unfortunately, only two small pieces of colony were obtained (plate 2, figs. 1 and 2). The rest of the material collected is referable to the three species C. nigrescens (p. 31), C. densus (p. 37), and C. hodgsoni (p. 48). LOCALITIES. The material of Cephalodiscus obtained on the “ Terra Nova ” Expedition was dredged from the following localities : — Station 90. — July 25th, 1911 ; from Summit, Great King, Three Kings Islands, New Zealand, S. 14° W., 8 miles ; 100 fms. (183 m.) ; dredge; bottom — rock. Station 295. — January 27th, 1913 ; Ross Sea, 73d 51' S., 172° 57' E. ; 190 fms. (348 m.) ; Agassi/. trawl. Station 314. — January 23rd, 1911 ; 5 miles N. of Inaccessible Island, McMurdo Sound ; 222-241 fms. (406-441 m.) ; Agassiz trawl ; bottom — mud. Station 316. — February 9th, 1911; off Glacier Tongue, about 8 miles N. of Hut Point, McMurdo Sound; 190-250 fms. (348-457 m.) ; Agassiz trawl; bottom — undecomposed animal remains and mud. Station 338. — January 23rd, 1912 ; entrance to McMurdo Sound, 77: 13' S., 164' 18' E. ; 207 fms. (379 m.) ; Agassiz trawl; bottom — mud. Station 339. — January 24th, 1912; entrance to McMurdo Sound, 77° 5' S., 164° 17' E. ; 140 fms. (256 m.) ; Agassiz trawl ; bottom — mud. Station 340. — January 25th, 1912; entrance to McMurdo Sound, 76° 56' S., 164° 12' E. ; 160 fms. (293 m.) ; Agassiz trawl; bottom — mud. Station 348. — February 13th, 1912; off Barne Glacier, McMurdo Sound; 200 fms. (366 m.) ; Agassiz trawl ; bottom — mud. Station 355. — January 20th, 1913; Ross Sea, 77° 46' S., 166° 8' E. ; 300 fms. (549 m.) ; Agassiz trawl. Station 356. — January 22nd, 1913 ; off Granite Harbour, entrance to McMurdo Sound; 50 fms. (92 m.) ; Agassiz trawl ; bottom — mud. Of these ten stations, one (Station 90) lies to the north of New Zealand, and the others are in Ross Sea. Station 295 lies E.S.E. of Coulman Island, and the remaining eight stations are in McMurdo Sound, Stations 338, 339, 340 and 356 being at the entrance to the Sound, while Stations 314, 316, 348 and 355 are farther south, to the west of the southern part of Ross Island. CEPHALODISCUS — RIDEW OOD. 13 The yield from the several stations is set out in the table below, a cross indicating that material of the species named at the top of the column was obtained from the station enumerated at the left end of the horizontal line in which the cross occurs. Station. C. evansi. C. nigrescens. C. densus. C. hodgsoni. 90 X — — — 295 — — X — 314 — X X X 316 — X X X 338 — — — X 339 — — X X 340 — — — X 348 — X — X 355 — X X X 356 — — X — HISTORICAL INTRODUCTION. The material of Cephalodiscus dodecalophus, dredged by the “ Challenger ” in 1876 from the Straits of Magellan, and described by MTntosh and Harmer (14) in 1887,* remained for many years the sole representative of the genus Cephalodiscus. In 1903 Andersson (I) announced the rediscovery of Cephalodiscus by the Swedish South-Polar Expedition, and Harmer f notified the discovery of newr species by the “ Siboga ” Expedition. In his report on the Pterobranchia of the “ Siboga ” Expedition published in 1905 Harmer (10) described three new species, C. gracilis, C. sibogae, and C. levinseni, obtained from the east and south-east of Asia. In the same year Lankester ( 1 2) gave a brief description of a species, C. nigrescens, obtained in Ross Sea by the “Discovery.” Material of this species had apparently been obtained as far back as 1841 or 1842 by the “Erebus” and “ Terror,” but it remained in the British Museum unidentified and undescribed until 1912 (19). * The full report was preceded by a preliminary notice in the “ Annals and Magazine of Natural History,” (5), X, 1882, pp. 337-348, and another in the Report of the British Association meeting of 1882, published 1883, pp. 596-597. The first published figures of Cephalodiscus, made from drawings supplied by Prof. M'Intosb, appeared in Sir E. Ray Lankester’s article, “ Polyzoa,” in the “ Encyclopaedia Britannica,” ed. 9, XIX, 1885, figs. 8-10. The “Challenger” material has been reinvestigated by numerous workers, among whom may be mentioned Harmer (Zool. Anzeiger, Bd. XX, 1897, pp. 342-346 ; Siboga Report, Monogr. XXVI, bis, 1905), Masterman (Quart. Journ. Micro. Science, new ser., Yol. XL, pt. 2, 1897, pp. 340-366 ; Zool. Anzeiger, Bd. XX, 1897, pp. 443-450; Tx-ans. Roy. Soc. Edinb., Yol. XXXIX, pt. 3, Dec. 1898, pp. 507—527 ; Quart. Journ. Micro. Science, new ser., Yol. XLVI, 1903, pp. 715-727), Cole (Journ. Linn. Soc., Zool., Vol. XXVII, no. 175, 1899, pp. 256-268), Schepotieff (Bergens Museums Aarbog, 1905, no. 8, pp. 20; Zool. Jahrb., Abth. fur Anat., Bd. XXIY, 1907, pp. 553-608 ; ibid. Bd. XXY, 1908, pp. 405-491), and Ridewood (Quart. Journ. Micro. Science, new ser., Vol. LI, 1907, pp. 221-252). f Zool. Anzeiger, Yol. XXYI, 1903, p. 593, 14 “TERRA NOVA” EXPEDITION. In 1906 Ridewood ( 1 6) published an account of C. gilchristi from South Africa, and in 1907 (17) gave detailed descriptions of Lankester’s species, C. nigrescens, and a new species, C. hoclgsoni, also of antarctic origin. In 1907 Andersson (2) produced his report on the Pterobranchia of the Swedish South-Polar Expedition, in which he described five new species ( C. aequatus, C. inaequatus, C. solidus, C. densus and C. varus) from the South-American sub- antarctic region and the Graham Land part of the Antarctic Ocean. An Indian species (C. indicus ) was described by Schepotieff (22) in 1909, and an antarctic species ( C . under ssoni) from near Graham Land was described by Gravier in 1912 and 1913 (5, 6 and 8). In 1913 Harmer and Ridewood (11) published an account of C. agglutinans obtained by the “ Scotia ” from the Falkland Islands. Thus, disregarding the possibility of some of these specific names being synonyms,* fifteen species of Cephalodiscus have up to the present been described. The material now under consideration, obtained. on the British Antarctic (“ Terra Nova”) Expedition of 1910-1913, includes large quantities of C. hodgsoni and C. nigrescens and a fair amount of C. densus , all from Ross Sea ; also two pieces of a new species, C. evansi, from New Zealand. CEPHALODISCUS AND RHABDOPLEURA. Genus Cephalodiscus , MTntosh. — Zooids small, living as a social community within a secreted coenoecium ,f from the orifices or ostia of which they can emerge at will ; coenoecium with a common branching cavity, or with a separate tubular cavity for each zooicl and its buds. Body of the zooid consisting of three parts, with separate divisions of coelom ; first part a shield, in the form of a flattened preoral disc, with a curved red line, used for crawling and for secreting the coenoecium, containing a single coelomic cavity, which opens to the exterior by a pair of pores ; second part a collar- region, with a paired coelomic cavity, which opens to the exterior on each side by a ciliated canal ; collar produced antero-dorsally into a lophophore of several pairs of arms, each arm with two rows of tentacles ; collar produced laterally and ventrally into an oral lamella ; central nervous system situated in the ectoderm at the bases of the arms ; third part, the largest, a trunk-region, with closed coelomic cavity, primarily paired, containing the alimentary canal and gonads. Alimentary canal U-shaped, mouth ventral, behind the stalk of the shield ; anus on the antero-dorsal surface of the trunk, near the bases of the arms. One pair of pharyngeal pores or gill-slits, near the collar- * For arguments showing that G. inaequatus is the same as C. hodgsoni, see ] | , pp. 559-563. A suggestion that C. aequatus also is not distinguishable from C. hodgsoni is put forward on page 59 of this report. Evidence that C. rarus is but a lax and straggling form of C. densus is adduced on pp. 39-40. The possibility of C. anderssoni also being the same as C. densus is discussed on page 40. As regards the two diminutive species of Demiotliecia, Harmer writes (10, p. 4) that “ the possibility is not excluded that C. sibogae is the male form of C. gracilis | See footnote, p. 57. CEPHALODISCUS — RIDEWOOD. 15 canals. A more or less tubular notochord projecting from the antero-dorsal wall of the pharynx ; below it, in the shield, a pericardial sac, with heart. Gonads simple, one pair, opening by short ducts near the anus. Trunk prolonged posteriorly or ventrally, according to the decree of extension of the zooid, into a stalk, with a terminal sucker, around the edge of which buds are produced. in consequence of the discovery of new species of Cephalodiscus since the time when the first description was written by MTntosh, the original generic diagnosis (13, p. 348) has from time to time been emended by various authors.'* In essential details the diagnosis given above, while adapted for the inclusion of all the species at present known, is substantially the same as those that appear in recent writings on the subject. Shield-pores, gill-slits and oral lamella are not known in the reduced males of Cephalodiscus sibogae (10, p. 6), but the neuter zooids of that species conform with the generic diagnosis in these respects. Females are not known. The only other genus of the group Pterobranchia recognised is Rhabdopleura, the existence of which was known f some ten years before Cephalodiscus was dredged by the ‘"Challenger'’ in 1876. The colony of Rhabdopleura is usually found growing upon a shell or stone. It is diminutive in size, with the zooids living in the tubes of a branching system, one zooid in each tube. The tubes are formed by lateral eruption from older tubes, and do not arise independently of the older tubes as is the case in the species of Cephalodiscus that are included in the sub-genera Idiothecia and Orthoecus. The tubes of these species of Cephalodiscus, however, resemble those of Rhabdopleura in that their length is increased by additions to the free edge. The increments are intermittent in both, so that the successive rings, and parts of rings, are distinct to the eye, although not readily separable by dissection. The mode of budding is different in the two cases, and the system of branching differs as a direct consequence of this. Whereas in the species of Cephalodiscus in question (those of the sub-genera Idiothecia and Orthoecus) the young zooids, whether developed as buds or as larvae arising from fertilised ova, settle down and secrete each a tube of its own, in Rhabdopleura there is a proliferating stolon which * As showing the extent of the alterations that have had to be made in the generic diagnosis of Cephalodiscus since 1882, a transcript of the original diagnosis may be of some historical interest. “ Coenoecium consisting of a massive, irregularly -branched, fucoid secretion resembling chitine, hispid with long spines of the same tissue, and honeycombed throughout by irregular apertures, channels, and spaces, in which the separate and independent polypides occur singly or in groups. Lophophore richly plumose, with an enormous buccal shield and large oral lamella, the mouth opening between the two. Anus on the anterior dorsal prominence, behind the plumes. Two large eyes abutting on the ovaries. The homologue of the funiculus is short and quite free, its tip serving for the development of buds.” (|3, p. 348.) f Rhabdopleura was first dredged in 1866 by G. O. Sars off the Lofoten Islands, at a depth of 120 fathoms ; it was mentioned by M. Sars in 1868 under the name Haliophus mirabilis, but was not described. New material, dredged by Canon Norman off the Shetland Islands from a depth of 90 fathoms, was described by Allman in 1869 (Quart. Journ. Micr. Sci., n.s., IX, 1869, pp. 57-63). Many papers have since appeared dealing with the structure of Rhabdopleura, and several new species have been described. 16 “TERRA NOVA” EXPEDITION. gives off buds in regular succession behind a terminal zooid. Each bud remains in that part of the axial or growing part of the colony in which it is formed, and becomes partitioned olf by the formation of a transverse septum. It breaks through the side of the portion of the tube within which it is now imprisoned, and forms a lateral tube that is increased gradually in length by the secretion of successive rings at its edge. The bud enlarges and differentiates further, and becomes ultimately an adult zooid. The adults found in the open tubes do not themselves produce buds, and thus differ from the zooids of Idiothecia and Orthoecus. As regards the general structure of the zooids, Cephalodiscus and Rhabdopleura have much in common ; the body is divisible into the same regions in both — a shield, a collar-region with oral lamella and lophophoral arms, a comparatively large trunk- region, and a stalk connected with the trunk. In Rhabdopleura the arms are two in number, each with two rowrs of tentacles. In Cephalodiscus the arms may be as many as eight or nine pairs ( C. agglutinans ) and as few as three pairs (C. indicus). The male zooids of C. sihogae are of interest in this connection, for they have only one pair of arms, without tentacles, but the neuter zooids have four pairs of arms, with tentacles (10, pp. 6, 8 and 84). Although gill-slits do not exist in Rhabdopleura , the position that these would occupy is clearly indicated by a pair of ciliated grooves. The alimentary canal of Rhabdopleura resembles that of Cephalodiscus in the relative positions of the mouth and anus, in the flexure of the gut, and in the presence of a notochord and a gastric caecum. The position of the central nerve-mass is another point in common. SUB-GENERA OF CEPHALODISCUS. Sub-genus Demiothecia, Ridewood. — Colony branching. Each ostium of the coenoecium leading into a cavity wrhich is continuous through the colony, and is occupied in common by the zooids and their buds. Transverse sections of the branches showing the central cavity surrounded by a wall of coenoecial substance, usually of irregular thickness, and sometimes with inwardly projecting bars and ridges. Sub-genus Idiothecia, Ridewood. — Colony branching. Each ostium of the coenoecium leading into a tube which is occupied by one zooid and its buds. The tubes embedded in common coenoecial substance, and disposed at a more or less constant angle to the surface ; either blind at their inner ends, or connected up in the middle of the branch. Sub-genus Orthoecus, Andersson. — Colony not branching, but in the form of a cake, or cone, or mass of irregular shape. Each ostium of the coenoecium leading into a tube which is occupied by one zooid and its buds. The tubes embedded in common CEPHALODISCUS — RIDEW OOD. 17 coenoecial substance, either for their whole length or towards their blind ends only ; either closely set and parallel, more or less vertical, or irregularly bent and straggling. To the sub-generic diagnoses Andersson (2, pp. 8, 10, 11) adds certain characters of the zooids, which I consider it desirable to omit, preferring to leave the three sub- o-enera to be determined by coenoecial characters alone. The characters mentioned bv Andersson may be dealt with as follows : — Firstly, the number of the arms borne by the zooids. In Orthoecm he gives the number as eight pairs ; but he quotes no numbers for the other two sub-genera, thus rendering the application of the character incomplete. The occurrence of eight parrs is not confined to Orthoecus, for the number in Cephalodiscus ( Idiot hecia ) nigrescent is sometimes eight pairs (more usually seven pairs, sometimes less). The number of arms, moreover, in certain species is not constant, notably in C. (I.) nigrescens (17, p. 32), C. (I.) agglutinans (II, p. 548), and C. ( D .) hodgsoni (17, p. 55 ; 18, p. 230 ; II, pp. 560-562 ; and pp. 62-64 of this report). One is rather led to suspect that past investigators have been prone to accept as the normal number of arms in a species the greatest number found in a zooid, and have assumed that in the zooids in which smaller numbers occur some of the arms have been lost. Doubtless in many of these cases accidental injury during life, or violent contraction when the zooids are plunged into the preservative fluid, or rough treat- ment of the zooids by shaking during transit, may account for the number being below the maximum ,• but in the course of the present investigation there have come to light numerous instances of a complete lophophoral system, with less than the maximum number of arms found in the majority of the zooids of the particular colony, about which there can be no doubt that all the arms that the zooid possessed at the time of death are still present. The arms in such instances form a continuous series side by side, with no gaps except the usual interval between the arms of the median pair. Sometimes there is to be seen between the outermost arm and the edge of the oral lamella a very small outgrowth, with or without lateral tentacles, representing an incompletely grown arm. The arms of Cephalodiscus develop successively in pairs (18), and not simultaneously, and it may be that the zooid in question is a young one in which all the arms are not yet developed. Some of the zooids, however, in which a rudimentary arm is found are of full size and with mature gonads, and it is more likely that the arm is permanently arrested in its development (text-fig. 12, p. 63). In other cases the vestigial arm is not the outermost or that nearest the edge of the oral lamella ; it may be No. 3 or No. 4, for instance. Such may possibly be accounted for as the unsuccessful result of an attempt at the regeneration of an arm lost by injury during life. It is only in some instances that rudimentary arms such as these are found ; in others the series is complete, without any gaps to suggest the loss of arms 18 “TERRA NOVA” EXPEDITION. or a deficiency in their number, and yet the number may be short of the maximum found in the species. Secondly, the presence or absence of globular end-swellings, with refractive beads, at the ends of the arm-axes. These swellings, Andersson points out, are commonly present in Demiothecia (wanting in C. sibogae and in the hinder pairs of arms of C. gracilis ), while absent in Idiothecia and Orthoecus. (See p. 21 of this report.) Thirdly, the presence or absence of a short stalk to the membrane that encloses the free ovum. The stalk, he states, is present in Demiothecia (except in C. gracilis, and possibly in C. sibogae, of which females are not known), and absent in Idiothecia and Orthoecus. But in C. (I.) gilchristi the stalk is known to be present in some cases (16, p. 188, and 3, p. 240). Finally, the position of the ova within the ovary, whether they lie externally to the cavity of the ovary, as in Demiothecia (e.g., 2, plate 7, fig. 63, C. inaequatus), or bulge into the cavity, as in Orthoecus (e.g., 2, plate 7., fig. 64, C. solidus). This is a character the discovery of which we owe to Andersson himself (2, pp. 7 and 81-82) ; the application of it, however, he leaves incomplete so far as concerns Idiothecia, no species of which were obtained on the Swedish South-Polar Expedition. A re-examination of the serial sections of the type-material of C. ( Idiothecia ) nigrescens obtained by the “Discovery” shows an agreement between that species (text-fig. 3, p. 37) and Andersson’s C. ( Orthoecus ) solidus (2, fig. 64). Whether the ova lie externally to the ovarian cavity or bulge into it is a difference apparently due to the fact that in Orthoecus and in C. ( Idiothecia ) nigrescens the proliferation of the ovicells occurs unilaterally, on the side of the oviduct nearest the shield (text-fig. 3, p. 37; also 17, plate 5, fig. 40), whereas in C. ( Demiothecia ) hodgsoni the small ovicells are more uniformly distributed around the lumen of the oviducal end of the ovary. The difference can only be made out in sections taken through the thin end of the ovary, for in the broadest part not more than one or two, rarely three, ova can be seen, and any space between them is just as likely to be an artefact, caused by a shrinkage of the ova, as it is to be the natural lumen of the ovary. It is clear from the small size of the ova in Andersson’s figures 63 and 64 that the sections there represented were taken near the oviduct, and he definitely states in two places (2, pp. 81 and 82) that it is difficult to trace the ovarian cavity backwards into the ovary. ' Limiting the sub-generic diagnoses to characters of the coenoecium, it will be noticed that the diagnosis for Idiothecia has had to be altered since it was founded (17, p. 10), so that it may include C. agglutinans, the inner ends of the coenoecial tubes of which are not blind. The affinities of C. agglutinans are so obviously with C. levinseni, C. gilchristi, and C. nigrescens, rather than with species of the sub-genus Demiothecia , that if the sub-genera are to represent natural divisions of the genus, CEPHALODISCUS — RIDEWOOD. 19 and not merely to form an artificial key for the identification of species, suck amend- ment of the diagnosis was necessary (11, pp. 556-559). The diagnosis of Demiothecia , also, has been recast so as to exclude C. agglutinans, the original diagnosis (17, p. 8) simply stating that there is a continuous cavity in the coenoeeium, with all the ostia leading into it. While there is a substantial difference between Demiothecia and Idiotliecia as redefined above, the independence of Orthoecus from Idiotliecia is not so well marked. The forms that are referred to the sub-genus Idiotliecia can, on the whole, be distin- guished by having a branched coenoeeium, with the youngest tubes at the apices of the branches, whereas those of Orthoecus have a more cake-like coenoeeium, with the youngest tubes around the edge or at some part of it, and the oldest and longest tubes in the middle of the mass. As showing how ill-defined is the limit between Orthoecus and Idiotliecia , it may be pointed out that Schepotieff (22, p. 435) places his species C. indicus in the sub- genus Idiotliecia, whereas the coenoeeium does not branch, but is cake-like, with all the ostia of the tubes set upon the upper surface. Possibly he was deterred from placing the species in the sub-genus Orthoecus by the fact that the tubes are not closely set, but have a fair amount of coenoecial substance between them ; also possibly by the fact that Orthoecus is stated in the original diagnosis to have eight pairs of arms, wdiereas C. indicus has not, and so by default the newr species had to be put in the sub-genus Idiotliecia, in the diagnosis of which the number of arms is not stated. Regarding the occupation of each of the tubes in a colony of Idiotliecia $by one zooid and its buds, it is to be noted that Gilchrist (3, p. 237) was fortunate enough to be able to study the zooids of C. (I.) gilchristi in a living condition, and observes that when they are drawn out of their tubes by means of a pair of forceps there appear to be more zooids than one in each tube. “ In several cases,” he remarks, “ three or four zooids were found projecting from the aperture of one tube, part of their visceral region and all the buds being contained within the tube. These were sexually mature, with well-developed gonads, and, as far as could be made out, were quite independent of each other, though where they were embedded in the tube the buds were more or less interwoven, forming a solid mass, which was difficult to disentangle without making sure that no connection was broken.” This is an observation which conflicts so strongly with preconceived notions, based upon the examination of preserved material, that one is disposed to read the passage in close conjunction with Gilchrist’s earlier statements in the same paper, that, in the living colonies which he was able to study, “ the zooids and buds were on the surface of the coenoeeium,” the zooids were found either “ at the entrance of the tube, on the spine, or even at some distance from the tube,” “ the general appearance suggested that the buds were acting as anchoring individuals for the fully-developed zooid,” “ on one or two occasions the whole cluster of buds and zooid was observed suspended VOL, IV. E 20 TERRA NOVA” EXPEDITION. in mid-water by the much drawn-out and attenuated stalk of a bud, which was securely attached to the coenoecium by its sucker -like proboscis,” “ some of the large buds were seen to break at this point (proximal end of the stalk), and ultimately many isolated buds were found scattered over the general surface.’ This freedom to wander over the surface, which had previously been noted by Andersson (2, p. 15) in living zooids of C. inaequatus, and had been suspected by the present writer from a study of preserved material of C. hodgsoni (17, p. 51), suggests at first that in those species of Cephalodiscus in which there are tubular cavities in the coenoecium (i.e., the species of the sub-genera Idiotliecia and Orthoecus) there is no rigid allotment of tubes to particular zooids, and that a zooid mag retreat into a tube other than that from which it emerged. If this point be ceded, the main question to settle is — what is the normal condition ? It may probably happen that, when alarmed, two or more zooids and their buds may withdraw for safety into the same tube, but it does not follow that they would remain thus associated for any length of time. From a study of the large amount of material that Dr. Gilchrist was good enough to send to the British Museum in 1905, in which material there certainly is a limitation of one zooid and its buds to each tube, I am disposed to believe that in natural conditions such a relation is maintained, but that in what one may term “laboratory conditions” the disturbed zooids take shelter in the tubes that happen to be nearest and most convenient. Another answer to the question is furnished by Harmer in his appendix to Gilchrist’s paper (3, p. 244), that the occurrence of several zooids in the tubes “ indicates, in all probability, that some of the buds have assumed an adult character.” Even in C. (I.) agglutinans, in which the tubes of the coenoecium communicate at their inner ends with a labyrinthic tubular system in the middle of the branch (11, p. 538), the probability is, so far as can be judged from a study of preserved material, that each tube that opens externally is the dwelling of but one zooid and its buds (11, pp. 540 and 559). ZOOIDS OF CEPHALODISCUS. Notwithstanding considerable differences in the form of the coenoecium in the various species of Cephalodiscus recognised, the anatomical structure of the zooids is remarkably similar throughout the genus. Only in the reduced male zooids of C. sihogae is there any marked departure from the average structure ; these have but one pair of arms, which are devoid of tentacles ; there are no gill-slits and no shield- pores, and the alimentary canal is reduced. With this exception, the slight differences that are noticeable between the various species are those affecting the size of the body, the degree of pigmentation of the surface, the proportions of the shield, stalk, stomach and gastric diverticulum, the length of the testes, the number and shape CEPHALODISCUS — RIDE WOOD. 21 of the arms, the mode of termination and the characters of the epidermis of the arm-axes, and the measurements of the notochord. The extreme mobility of the shield in life leads to very considerable differences in the appearance of the shields of zooids of the same species after death, and this renders a comparison of the shields of the different species a matter of practical difficulty ; one can only utilise the organ for taxonomic purposes by taking an average of measurements of a large number of selected examples. The structure of the shield is essentially the same in all : the posterior lobe is thinner than the anterior, and the curved red line is constant. Regarding the number of the arms of the lophophore, it is to be remarked that in some species the numbers fluctuate considerably (p. 17), and it is not safe to trust to results obtained by the cutting into serial sections of a few zooids. The more reliable method is to examine a large number of individuals, each zooid being dissected in diluted glycerine * under a Zeiss binocular erecting microscope (oculars — No. 2 or No. 4 ; objectives — No. a3) by means of two mounted needles (English size No. 12 or No. 14), the points projecting not more than 5 mm. so as to avoid undue flexibility. The shield of the zooid is first removed, and the arms are detached one at a time by severing their bases. The body is then dissected open to extract the gonads, so that the same dissections may serve for determining the sex. A cover-glass is put on, and ringed with zinc-white cement. Such dissections, if the slides are placed in trays so that they may remain horizontal, will keep for a considerable time, and are thus available for subsequent study. The mode of termination of the arm-axes, whether possessing end-swellings with refractive beads or not, is a fairly good character for specific distinctions, although not for defining sub-genera (see p. 18). In the zooids of C. liodgsoni obtained by the “Terra Nova’- it occasionally happens that an arm has no end-swelling (text-fig. 12, B, p. 63). The deficiency does not necessarily affect the same arms in different zooids, and it may be unilateral, one arm alone, perhaps, out of a total of ten or twelve being thus defective ; or, on the other hand, several arms of the same zooid may be without end-swellings. The condition rather suggests that the extremity has been lost during life, and the end of the arm has healed up without the end- swelling being regenerated. Although in species of Idiothecia and Ortmecus isolated instances occur, possibly one arm in six or eight zooids, in which the extremity of the arm-axis is slightly enlarged (e.g., C. gilchristi , 16, p. 185, text-fig. 2, B ; C. agglutinans, II, p. 549, text-fig. 4, C ; C. nigrescens, 17, plate 5, fig. 23; and C. densus, text-fig. 6, F, p. 45 of this report), yet this is a form of termination clearly distinct from the bulb-like end-swellings, studded over with highly refractive beads, that are present with such * Equal parts of pure glycerine and distilled water, with a small crystal of thymol dropped into the bottle a few weeks before the fluid is used. E 2 22 ‘'TERRA NOVA” EXPEDITION. regularity on the arms of C. dodecalophus (17, p. 4, text-fig. 1) and C. liodgsoni (17, plate 5, figs. 32-35) ; and so far as present knowledge goes, these characteristic “end-swellings” are confined to species of the sub-genus Demiothecia, In C. gracilis they do not occur regularly in the adult, although they are present on the first pair, and sometimes on the second and third pairs of arms of the buds, and occasionally persist in full-grown zooids (10, pp. 20 and 94). They do not occur on the arms of neuter zooids of C. sibogae, but in the reduced males, particularly young males, the refractive beads are met with in profusion along the whole course of the two arms (10, pp. 21 and 84). In Orthoecus Andersson (2, p. 11) states that end-swellings do not occur, and one consequently suspects that there must be some inaccuracy in his figure of the zooid of C. ( 0 .) rarus (2, plate 4, fig. 22), for it shows end-swellings far more distinct than those represented in his figures of C. (D.) inaequatus (plate 4, figs. 17-21). Independent examination at the British Museum (Natural History) of zooids of C. solidus and C. densus obtained on the Swedish South-Polar Expedition confirms the statement by Andersson that end-swellings with refractive beads are not present. The peculiarities of the epidermis of the arm-axis noted by Andersson in dealing with his three species of Orthoecus (2, pp. 11-12) do not as specific characters inspire one with confidence, for so much depends upon the degree of extension of the individual arms at the time of death of the zooid. The cells in question are tall epithelial cells on the neural surface of the axis, opposite the ciliated groove, and, according to the observations of Andersson, they occur to a different extent in the three species. In C. solidus the tall cells are present on the extremity of the arm and along the distal half of the length of the neural surface of the axis ; they fail to stain with haemalum and eosin, and in serial sections remain markedly transparent. In C. densus and C. rarus the peculiar epithelium does not occur at the actual extremity of the axis ; in C. densus the cells for a short distance near the end are slightly taller than elsewhere, and stain feebly or not at all ; in C. rarus the epithelium of the distal half of the arm is composed of tall cells, those nearer the extremity remaining unstained or but feebly stained, while the others, more irregular in their disposition, stain blue. Examination of zooids of the specimens of these three species that wrere received by the British Museum (Natural History) from the Stockholm Museum shows that the differences recorded are not constant. The actual tip of the axis of the arm of C. solidus really resembles that of C. densus and C. rarus in not having the tall epithelium conspicuously developed; in Andersson’s fig. 10 the end of the axis is so strongly incurved that the actual extremity is not clearly seen. In one zooid of C. solidus, dissected and mounted in glycerine in such a manner that all the sixteen arms can be examined separately, there is considerable variety among the arms so far as the disposition of the tall epithelium is concerned, and the differences are to some extent associated with differences in the degree of contraction of the arms. Two CEPHALO DISCUS — E IDE WOOD, 23 of the arms have much the same character as that shown in Anderssou’s figure ; the end of each is incurved, the tall epithelium occurs on the distal half of the axis, and not on the proximal half. Two arms show the tall epithelium disposed along the axis almost to the base, and two others are in fairly full extension, longer than the average, with the end of the axis tapering to a blunt point. The general condition of these last arms recalls that shown in Andersson’s figure of C. densus (2. fig. 8) rather than that shown in the figure of C. solidus (fig. 10). There is no doubt that the epithelium of the neural surface of the axis in the three species of Orthoecus under consideration is different in character from the general epithelium of the aponeural face and the tentacles, but how far the enlargement of the cells is unnatural and due to the particular fixing reagents employed can only be determined by a comparison of zooids from portions of the same colony fixed by different methods (cf. C. (I.) nigrescens, p. 36). The degree of pigmentation of the body is a character that must be utilised with caution, for Gilchrist has recently shown (3, p. 239) that The living zooids of C. gilchristi are blackish, and that the colour dissolves out into the preservative fluid ; if the fluid is changed frequently, the colour of the zooid becomes pale, with just a dark anterior edge to the shield. Andersson has, moreover, observed that the zooids of C. inaequatus are red when alive, the trunk being darker than the other parts (2, plate 1) ; but in the preserved state they are reddish brown, pale brown or cream-coloured. Partial solution of pigment by the action of preservative fluids had been previously suspected in zooids not known in the living state ( C. nigrescens, I 7, p. 25, zooids fixed in picric acid solution), and it is surmised in the case of C. cigglutinans (II, p. 544). Of further interest in this connection is the fact that in preserved material of C. solidus the zooids in the deeper parts of the colony are blackish, but those in the more superficial tubes, and zooids that have been removed from their tubes, fade — presumably by the action of light — to a pale brown, or yellow (ochreous), or cream-colour. Oddly enough, the red line in the shield and the red mass at the oviducal end of the ovary seem to retain their colour in all the usual preservative fluids, and to resist the fading effect of light. The size of the body of the zooid is a character to which little value is attached by Andersson, who has had the good fortune to study the zooids alive. The body of Cephalodiscus is so mobile and protean that, whereas in preserved material the stalk is not infrequently short, thick, and wrinkled, and connected with the ventral surface of the body at about two-thirds or three-quarters of the length from the mouth, in a well-extended living zooid the body tapers off posteriorly into the stalk, which is of considerable length and slenderness, and the body has no caecal end such as is present in preserved specimens. It may here be pointed out, however, that most of the material that investigators have to deal with is material that is sent to museums and laboratories in a preserved state, and it certainly is to a large extent possible to compare the sizes of dead zooids with a fair degree of conviction that the measurements 24 “TERRA NOVA” EXPEDITION. are of value. By selecting zooids that are moderately extended, not too attenuated, and not too distorted by excessive contraction, it is possible to take three useful measurements — -from the caecal end of the body to the front of the body (bases of the arms, or the anus), from the caecal end of the body to the free extremities of the arms, and, thirdly, the diameter of the body. In the Report of the National Antarctic (“Discovery”) Expedition (17, pp. 9-11) the measurements of the zooids are given as from the front of the shield to the extremity of the body ; but owing to the fact that the shield may be displaced, distorted or injured, it is now found that for practical purposes the measurement from the anal end of the body, at the bases of the arms, to the caecal end of the trunk is more reliable for purposes of comparison of the zooids of different species. Gravier has pointed out (8, p. 73) that Cephalodiscus affords a good instance of gigantism in polar regions, both as regards the size of the zooids and the bulk of the edifices that they build. He notes the large size of C. solidus , C. anderssoni, and C. nigrescens, which are antarctic forms, in contrast with the diminutive character of the tropical forms C. indicus (6°-ll° N. ),* C. gracilis (1° N.), and C, sibogae (4° S.). While the thesis is correct in the main, it must be admitted that it does not apply with precision throughout the genus, for C. hodgsoni is an antarctic form, and is not much larger than the sub-antarctic C. dodecalophus. Further, the zooids of the new species ( C. evansi) obtained on the British Antarctic (“Terra Nova”) Expedition are about twice as long as those of C. gilchristi, the material of which was dredged from the same latitude ( C. evansi, 34° S. ; C. gilchristi, 33°-35° S.). And C. levinseni, which was obtained at the same distance north of the Equator (32° N.) as C. evansi and C. gilchristi wrere obtained south of it, is intermediate, in the length of its zooids. between these two species, but the zooids are thinner than those of both species. It is doubtful whether the relative length of the stalk as found in preserved zooids has much value, for in life the stalk is now known to vary enormously ; but the length of the stalk in preserved zooids is in a sense a measure of its muscularity, and the general appearance of zooids killed rapidly in the fixing fluid is largely determined by the relative contraction of the stalk. It is quite conceivable, however, that if a living colony were divided into two, and one part plunged immediately into a fixing fluid, while the other part was fixed later, when the zooids were in a moribund con- dition, the zooids of the former would exhibit a shorter and more contracted stalk than those of the latter part. A character to which Andersson directs attention (2, pp. 7-8, 32) is the occurrence of three longitudinal thickenings in the nerve-tract on the ventral side of the stalk in the species of Demiothecia examined by him, and the presence of a single thickening in the species of Orthoecus. While this may possibly prove to be a useful character for the identification of a species, it does not follow a natural grouping of the species * Tbe latitudes are here given to the nearest degree; for more precise data, see Synopsis, pp. 66-77. CEPHALODISCUS— RIDEWOOD. 25 of Cephalodiscus known, for it has been shown (II, pp. 553-554) that the triple thickening occurs in C. ( D .) dodecalophus, C. (D.) hodgsoni, C. (D.) aeqmtus, and C. (I.) nigrescens, and the single thickening in C. (D.) gracilis , C. ( D. ) sibogae, C. (1.) levinseni, and C. (I.) agglutinans. In C. (/.) gilchristi the thickening is triple, and in C. (/.) evansi it is single. The character is difficult of application, owing to the fact that the lateral thickenings in the species with a triple nerve-tract are never very pronounced ; they are not separate nerve- tracts, but merely slight thickeninos of a continuous sheet of nervous tissue that is but feeblv differentiated O in any case, and is still sub-epidermal in situation. The length and diameter of the notochord are characters which Andersson (2, pp. 58-62) utilises as a means of discriminating between the various species of Cephalodiscus obtained on the Swedish Expedition. He points out, moreover, that the anterior or distal end is swollen and bent upwards in the species of Demiothecia, but is not swollen nor sharply bent in the species of Ortho ecus. In C. ( Idiothecia ) nigrescens the extremity is neither swollen nor bent (17, p. 34, text-fig. 10), but in C. (I.) gilchristi the end is bent upwards, though not swollen (16, p. 186, text-fig. 3). A section taken through the middle of the length of the notochord of Cephalodiscus is usually oval, the vertical diameter being less than the horizontal diameter ; but observations made on several series of sections of zooids of C. densus obtained by the “Terra Nova” go to show that in different zooids of the same species a transverse section through the notochord may vary in shape from a circle to a very elongated ellipse, and the only measurement that can in any way be regarded as constant is the mean of the vertical and horizontal diameters (see p. 48). The length of the testis is a character which has been rather neglected for purposes of classification, yet it seems to have some value. In some species ( e.g ., C. hodgsoni, 1 7, plate 5, fig. 41) the two testes are globular, ovoidal or pyriform, whereas in other species the testes are several times as long as they are wide (e.g., C. nigrescens, 17, plate 5, fig. 37, and Orthoecus, 2, p. 84). Although the character may prove useful in the determination of a species, it does not, one must admit, follow what may be regarded as the natural affinity of the different species ; for the testes of C. (I.) gilchristi (16, p. 189, text-fig. 5) are short, like those of C. ( D.) hodgsoni, and not long, like those of C. ( 1 .) nigrescens ; and the testes of C. ( D .) sibogae (10, plate 7, fig. 76) are long, whereas in other species of Demiothecia in which the testes are known they are short. When the testes are of the long variety, there may be in a zooid either two moderately long testes (length five to eight times the width) or one very long testis and a shorter one ; the longer testis frequently extends to the caecal extremity of the body, and may project into the stalk. Immature zooids must be avoided in determining the relative length of the testes, for young testes are comparatively short in the zooids of species characterised by the possession of long testes in the adult state. * 26 “TERRA NOVA” EXPEDITION. There is not a marked difference between the ovaries in the different species ; the' ovaries are usually globular or pyriform, the length being rarely more than three or four times the width, although in young ovaries this proportion is sometimes exceeded. The longest ovaries occur in the species of Orthoecus. The early or late stage at which the arms make their appearance in the buds is a character which may prove of some assistance in discriminating between the various species of Cephalodiscus. Buds with three or four pairs of developing arms have a relatively small shield in some species, and a very large shield in others. In C. dodecaloplius and C. hodgsoni the first pair of arms reach the front edge of the shield at a stage when three pairs of arms are recognisable (18, p. 231, text-fig. 4, G and H ; and p. 225, text-fig. 2, G), and the buds of C. aequatus and C. inaequatus in material collected on the Swedish South-Polar Expedition have been found to agree very closely with them (11, p. 556). The same relation evidently obtains in C. gracilis (10, plate 1, figs. 4 and 9, and plate 3, figs. 30-32), although the particular stage with three pairs of arms developing is not figured. On the other hand, in C. agglutinans (11, p. 555, text-fig. 5, D), C. nigrescens and C. gilchristi (18, p. 236, text-fig. 6, D, and p. 243, text-fig. 9, G), and in C. solidus (21, text-fig. 3, D) and C. densus (text-fig. 8, F, p. 47 of the present report) the arms at the stage when three pairs are developing are but insignificant bead-like outgrowths from the collar- region, and are very remote from the edge of the shield. CEPHALODISCUS EVANSI, n. sp. Cephalodiscus ( Idiothecia ) evansi, n. sp. — Colony massive, branching ; largest piece known, a single branch, 47 mm. high, 19 mm. wide, not including peristomes. Coenoecium friable, cream-coloured, speckled, containing large quantities of shells of Foraminifera and fragments of shells of Molluscs and Ecliinoids. Ostia at the ends of tubular peristomes that project 2 • 5 to 4 • 5 mm. beyond the surface of the branch ; ostium set obliquely, sometimes squarely, at the end of the peristome, without definite lip or spine. Each ostium leading into a tube that ends blindly in the middle part of the branch. Width of cavity of the tubes, 0 • 6 to 0*8 mm. ; length of tube, including peristomial part, 10 to 25 mm., but shorter than 10 mm. at the apex of the branch. Zooids — length from free ends of the arms to end of body of a fairly extended zooid, 3 * 5 mm. ; length from bases of arms to end of body, 2 * 3 mm. ; width of body, 0 ' 7 mm. Colour (of zooids preserved in formalin solution) pale green or white, with the red line of the shield very conspicuous. Arms usually eight pairs, sometimes nine or seven pairs ; no end-swellings with refractive beads. Male and female zooids occurring in the same colony. Hermaphrodite zooids (with one ovary and one testis) not known. Testes long. Free ova not known. Buds several to each zooid, up to nine. x CEPHALODISCUS — RIDEWOOD. 27 Material of Cephalodiscus evansi was obtained from Station 90, off New Zealand ; from summit of Great King, Three Kings Islands, S. 14° W., 8 miles; Lat. 34° 15' S. ; Long. 172° 4' E. : 100 fms. (183 m.) ; July 25th, 1911 ; bottom, rock. The species is named C. evansi after Commander E. R. G. R. Evans, C.B., D.S.O., who did valuable work as second officer of the shore-party of the British Antarctic (“ Terra Nova ”) Expedition, and has more recently given further evidence of his sterling character, resource and enterprise while in charge of H.M.S. “ Broke. ' The material consists of two pieces only, the larger (plate 2, fig. 1) measuring 47 mm. high and 19 mm. wide, not including the projecting peristomial tubes. This is clearly a branch of a large, rather massive colony, and the smaller piece may possibly have been broken off from its basal end ; but the two pieces do not fit accurately together, owing to a rounding off of the surfaces, occasioned, no doubt, by the shaking and attrition within the bottle which the specimens must have experienced before they were handed over for investigation. C. evansi is an arenaceous form of Cephalodiscus, resembling C. agglutinans in general appearance, but with peristomial tubes projecting 3 or 4 mm. from the general surface of the coenoecium, whereas in C. agglutinans it is only the short spines that project. In the present species there are no spines. The ostia are occasionally set transversely at the ends of the peristomial tubes, but more usually the tubes are obliquely cut, sometimes, as in the upper part of the larger specimen (plate 2, fig. 1), so obliquely as to suggest the presence of a lip at the outer- most part of the edge. The superimposed layers of coenoecial substance constituting this projection are, however, continued all round the ostium, so that there is not the same definite growth of lip as occurs in C. nigrescens, C. solidus, and C. levinseni. While the peristomial tube is strongly ringed by lines of growth, the rings become fainter below as one follows the tube inward, and they disappear long before the blind end of the tube is reached (text-fig. 1). The peristomial tubes measure from 2 • 5 to 4 • 5 mm. in length, but at the apex of the branch they are shorter than the average ; the peristomes are here paler and more delicate, and more crowded than those on the sides. All these features indicate that the youngest tubes occur at the free end of the branch, and that the mode of growth of the colony is probably the same as has been suggested in the case of C. nigrescens (17, p. 23). The general resemblance of the branch to one of the more slender branches of C. agglutinans is but a superficial one, and is mainly due to the multitude of extraneous particles embedded in the coenoecial substance. The foreign particles are rather less abundant than in C. agglutinans ; they are mostly white in colour, and consist of shells of foraminifera and young gastropods, fragments of molluscan shells, both gastropod and lamellibranch, spines of echinoids and silicious sponge-spicules. The form of Cephalodiscus obtained by the “ Scotia ” was designated C. agglutinans because the characteristic appearance of the coenoecium was due to the agglutination, VOL. IV. F 28 “TERRA NOVA ” EXPEDITION. within the soft coenoecial substance between the inhabited tubes, of large quantities of shells and fragments of shells. It was not imagined, however, that the habit of agglutination was confined to that species (11, p. 542), for it had already been observed, though to a less marked extent, in the case of C. sibogae (10, plate 2, figs. 17, 18), C. densus (2, p. 12), C. solidus (2, p. 11), C. aequatus (2, p. 10), C. nigrescent* (19, p. 551 ; and 8, p. 75), and C. anderssoni (6, pp. 147-8 ; and 8, p. 80). The presence of large quantities of shell-fragments within the soft coenoecial substance of C. evansi gives a speckled, cream-coloured appearance to the branch, and renders it extremely friable. Each peristome is continued within the substance of the branch as a tube that is inhabited by a single zooid and its buds. The deeper, embedded parts of the tubes are composed of extremely thin films of tough coenoecial substance, transparent, and of a pale amber colour, in marked contrast with the jelly-like, soft, colourless coenoecial substance that fills in the intervals between the tubes. The tubes themselves do not include in their walls any of the shell-fragments that occur between them in the substance of the branch, but some of the shelly particles are so tightly affixed to the outer surface that in tracing out a tube great care has to be exercised in order to avoid tearing it, and this in spite of the fact that its thin wall is so tough. The extreme transparency of the tubes of the colony adds to the difficulty of tracing them out. Good results could undoubtedly be obtained by adopting the method of decalcification that proved so useful in tracing out the tubular labyrinth of C. agglutincins (11, p. 538), but the small amount of material available rendered the method undesirable, for it would have meant the destruction of all the zooids in the piece of colony thus treated. Since only two pieces of colony were obtained, the tracing of the tubes and the extraction of zooids for study were effected mainly ujion the smaller piece. The larger and more shapely piece (plate 2, fig. l) is thus left comparatively uninjured. The tubes do not anastomose as they do in C. agglutinans, but each ends blindly in the middle part of the branch, a feature in which the present species resembles C. nigrescens, C. levinseni, and C. gilchristi. The tubes are not uniformly distributed in the interior, for while the majority are separated from one another by soft coenoecial substance and shell-fragments for about twice the width of a tube, other tubes are in places actually in contact with one another (plate 5, fig. 9). The surface-film of soft coenoecial substance is definite and continuous, but in the middle part of the branch the amount of secreted material holding together the particles of shell, etc., is very small, so that the tubes seem to be traversing a mass of shell-fragments only. The thin superficial layers that cover the soft common coenoecial substance are continued outward on the sides of the peristomial tubes, as shown in text-fig. 1, a. The longest tube dissected out measures 21 mm. from the ostium to the blind end, CEPHALODISCTJS— RIDEWOOD. 29 but there is evidence to show that some tubes are longer than this, probably 25 mm. ; some tubes are as short as 10 or 12 mm. ; at the apes of the branch they are even shorter than 10 mm. The width oi the cavitv of the tubes varies from 0*7 or 0 • 8 mm. j at the peristomial end to 0 • 6 or 0 ' 7 mm. in the middle and inner parts. The blind end of a tube has a very thin wall ; it is sliohtlv dilated into an irregular bulb, as in C. nigrescens, and may have as many as five or six curved septa (text-fig. 1, A, s). This part of the tube is frequently bent at an angle to the length of the tube. The deeper parts of the tubes — i.e., those nearest the axis of the branch — are in places curiously wrinkled and distorted ; in other places irregular longitudinal ridges project inward, and yet the outer surface of the tube is smooth (text-fig. 1, r). In rare instances a zooid can be seen without cutting into the coenoecium, but in most eases the zooids have retreated to the middle of the length of their tubes. The zooids are white in colour, some of them with a greenish tinge on the trunk, but not on the arms. The curved red line of the shield shows up very strongly by contrast with the rest of the body. Owing to the whiteness of the zooids they are, in a teased up piece of branch, difficult to distinguish among the numerous shell-fragments, but the red line of the shield is so conspicuous that it indicates the presence of a zooid even to the unaided eye. Eight zooids in a fair state of extension were measured, and the averages of the figures obtained are : — length from the free ends of the arms to the caecal end of the trunk — 3 • 5 mm. ; length from the bases of the arms, or from the anus, to the end of the trunk — 2 • 3 mm. ; width of the body — 0 • 7 mm. The length of a moderately well-extended stalk is 3 or 4 mm. The arms are mostly eight pairs, sometimes nine or seven pairs ; they are colourless, and have Text-fig. 1. — Cephalodiscus evansi. A — side view of a comparatively short tube dissected out from the common coenoecial substance ; B and C — transverse sections of a tube in the region of internal ridges ; D — transverse section of a part of a tube free from such ridges. (X 12.) a, level of the general surface of the branch : the part of the tube above this point is the peristomial tube, the part below it is surrounded by common coenoecial substance ; r, internal longitudinal ridges ; s, septa in the blind end of the tube, F 2 30 “TERRA NOVA” EXPEDITION. C.2. O.L no end-swelling with refractive beads. In extended arms the axis is slender and straight ; in contracted arms the axis shows a thickened epidermis on the neural surface, extending over the distal half or two-thirds of the length, and in many cases the extremity of the axis is incurved towards the grooved, aponeural surface. The outer edge of each lobe of the oral lamella is scalloped, with six, sometimes seven, well-marked projections. Male and female zooids occur in different tubes of the same branch of the colony. Hermaphrodite zooids, having one ovary and one testis, were not found. The ovaries are much the same as occur in C. nigrescens (17 , plate 5, figs. 38, 39, 40), and have red oviducts. No free ova were found in the tubes. The testes are long, either one very long testis and a shorter one, or both of moderate length. The spermatozoa have long, tapering heads, and resemble those of C. densus (text-fig. 7, F, p. 46), but are a little smaller. The material when collected was dropped into formalin solution, and was not specially preserved with a view to microscopical investigation, but the following features can be made out from the examination of serial sections and dissected preparations. The notochord is straight, and not bent up at the tip ; its length is about 0*24 mm. The thickest part is situated at about one-sixth of the length from o.l. -- c.c.m.- c.c. 8 s- C.3. f— Text-fig. 2.- — Ceplialodiscus evansi. Composite figure of camera lucida drawings of three consecutive sections taken parallel to the face of the shield. ( X 150.) Thickness of the sections — 0 • 007 mm. c.2, collar coelom ; c.3., trunk coelom ; c.c., collar canal ; c.c.m., collar canal muscle ; c.n.m., central nerve - mass ; g., cavity of gut ; g.s., gill-slit ; /., base of the lophophore, o.l., oral lamella. the tip, and tapers off anteriorly from this region. The transverse sections of the proximal and distal parts of the notochord are mostly circular or elliptical in outline, but those taken through the middle of the length are in some zooids nearly circular, while in others they are so compressed dorsoventrally that the cavity is practically obliterated, and the outline is very broad in comparison with its height. Taking in each case the mean of the vertical and horizontal diameters of the transverse sections, the average measurements are- near the distal extremity 0- 043 mm., in the middle of the length O' 027 mm., near the base 0 ' 021 mm. The collar-canals (text-fig. 2, c.c.) are large and curved, with both external and CEPHALODISCUS — RIDEWOOP. 31 internal openings directed forwards ; the two orifices are approximately of the same size. The nerve-tract in the stalk is single. Several buds occur at the end of the stalk of the zooids, the greatest number found being nine. The stalks are long, and there is reason to suspect that the zooids and buds were dead and in a relaxed state at the time when the pieces of colony wrere placed in the preservative solution. The buds are white, with the red line very prominent, except in the youngest buds. In the middle of the shield is a circular grey area. In the late development of the arms in the buds, or, in other words, the large size of the shield at the time when the arms begin to appear, C. evansi resembles agglutinans (1 I, text-fig. 5, p. 555), C. gilchristi ( 1 8, text-fig. 9, p. 243), C. nigrescens (18, text-fig. 6, p. 236), and C. solidus (21, text-fig. 3). CEPHALODISCUS NIGRESCENS , Lankester. Cephalodiscus ( Idiothecia ) nigrescens , Lankester. — For diagnosis, references and list of recorded specimens, see page 73. MATERIAL COLLECTED. Material of C. nigrescens wras obtained from Stations 314, 316, 348 and 355, all of which are in McMurdo Sound, Ross Sea. For details as to depths and dates, and the precise position of each of these Stations, see page 12. The material from the first three stations is considerable in amount : — from Station 314 eighteen bottles, from Station 316 three bottles, from Station 348 eight bottles. The collection from Station 316 was preserved in formalin solution; of the material collected from Station 314 and Station 348 some was preserved in formalin and some in 40 over-proof alcohol. The material from Station 355 consists of a single small incipient colony, not more than 30 mm. high, with the tubes short and crowded, and with the coenoecial substance unusually transparent. The collection contains a number of fine pieces of colony, and a photograph of the largest of these, obtained from Station 316, is reproduced in plate 1, fig. 2. The height of this specimen is 150 mm., and the width 130 mm. It is rather more straggling, and less closely branched than the large piece obtained on the National Antarctic (“Discovery”) Expedition, and figured in the reports of that expedition (17, plate 1). Another piece, from Station 348, is interesting as showing lateral branches all in the same stage of growth, set upon a thick stem which seems to be considerably older (plate 1, fig. ] ). The general appearance of the whole mass rather suggests that there has been a cessation and resumption of growth in the colony. The tubes of the lower part of the main stem are occluded and covered over with general coenoecial substance (cf. 17, p. 22), and all the tubes that are in a state of occupation by zooids seem to be of a newer growth. Whether the zooids are survivors of the original stock, or immi- grants that have settled upon a “ dead ” coenoecium, it is difficult- to say. 32 “TERRA NOVAi” EXPEDITION. The third piece shown on plate 1 (fig. 3) is a small portion of a colony with rather long, thin, tongue-shaped lips to the peristomial tubes. The apex of this branch was evidently in a state of rapid growth when the specimen was dredged, as may be judged by the crowding and the delicacy of the tubes at the extremity. A colony of C. nigrescens from Station 316 is represented in figs. 4 and 3 of plate 2, photographed against a light and against a dark background. In the former the black zooids lodged in their tubes show up very distinctly, and the pale grey, opalescent coenoecial substance tends to disappear, whereas in the latter photograph the coenoecial substance is very conspicuous and the zooids are visible in a few places only. These two photographs show how difficult it must be to compare adequately illustra- tions by different authors of specimens of Cephalodiscus illuminated in different ways and reproduced by different methods. The middle of the uppermost three branches of fig. 4 shows the strange twisting of the tubes about one another which is not infrequently seen in this species. The figure of the longitudinal section of a branch of a colony of C. nigrescens given in the “ Discovery ” Expedition report (17, plate 4, fig. 10) is but a diagram, and represents all the tubes as lying strictly in a radial plane ; as a matter of fact the inner ends of the tubes almost always twist about one another somewhat in the axial part of the branch, and it is only their distal halves that lie in the radial plane. The twisting of the inner portions of the tubes is shown clearly in 1 7, plate 3, fig. 4. In much of the material obtained by the “Terra Nova” the twisting is very pronounced, particularly in what I take to be rapidly grown branches, branches with rather short tubes, separated by very soft and perfectly clear and hyaline common coenoecial substance. Such branches are extremely fragile, and require to be handled with the greatest care. From Station 314 were obtained some very slender branches (plate 2, fig. 5) measur- ing only 5 to 9 mm. across, not including the peristomial tubes. These are isolated branches, giving no indication as to the shape of the whole colony of which they formed part ; a colony composed of many slender branches such as these could scarcely be brought up in the trawl in a perfect condition. It is interesting to contrast these attenuated branches with the more massive trunks that were obtained from the same Station, some of them as bulky as the piece figured in plate 2 of the “ Discovery ” Expedition report (17, plate 2, fig. 2). The single small piece of C. nigrescens from Station 355 is of considerable interest as being what may be regarded as a recently established or young colony. It measures 30 mm. in height and 24 mm. in maximum width ; the coenoecial substance is paler and more transparent and gelatinous in appearance than usual, and has not the custom- ary grey look of the larger pieces of colony. The tubes are rather short and crowded. Only about a dozen zooids are present, although the tubes are more numerous : the probability is that some of the zooids escaped at the time of capture, or during the process of sorting and preserving the material. The zooids of this colony are all of approximately the same age, and almost of full CEPHALO DISCUS— RIDE' WOOD. size ; and one may surmise that they had grown up from a group of larvae, rather than from buds. Each has several voung buds of its own. The zooids have no gonads large enough to be recognised in dissected preparations ; three were cut into serial sections, and one proved to possess a pair of extremely minute ovaries, while the other two had very minute gonads that were probably ovaries, but were too young to determine with certainty. Yet one of these latter zooids had a large bud possessing a pair of gonads distinctly recognisable as ovaries. Three pieces of colony of C. nigrescent were dredged with C. hodgsoni growing upon them. They were obtained from Stations 314, 316 and 348 respectively. In the last case the C. nigrescens is the more bulky part of the mass, and gives support to the coenoecium of the other species ; in the other two cases the branches of C. nigrescent are delicate. The coenoec-ia of both species contain zooids that were evidently alive at the time of dredging, and the specimens are interesting as showing that two different species of Cephcdodiscus can live harmoniously together. ZOOIDS. As regards the bulk of the material of C. nigrescens collected, dredged from Stations 314, 316 and 348, it is to be noted that zooids removed from different tubes of the same colony, and all fairly well extended and apparently full-grown, vary con- siderably in their measurements. The measurements of six zooids taken at random from a piece from Station 348 are as follows, the first being the length in millimetres from the ends of the arms to the caecal end of the body, the second the distance from the bases of the arms to the caecal end of the body, and the third the width of the body. The constancy in the last measurement is probably due to the uniformity in the width of the coenoecial tubes within which the zooids died. 6 • 0—4 • 2—0 • 9 4 • 5—2 1 7—0 • 9 5 • 5 — 3 • 7 — 0 • 8 4 • 0—2 • 6—0 • 9 5-0— 3-3— 0-9 3 '2—1 -7—0-9 Serial sections were cut of zooids selected from representative pieces of colony from these stations, for the purpose of verifying the identification of the species, but no new observations have resulted that are worth recording. As in the case of the zooids of C. nigrescens obtained by the “ Discovery ” (17, p. 32) the number of the arms is not absolutely constant, but seven pairs is the number found in the majority of cases. The arms of C. nigrescens present considerable differences in appearance according to the mode of fixation and preservation employed. In well-preserved zooids that have been fixed either in formalin solution (5 per cent.) or in alcohol (70-75 per cent.) the black pigment is confined to certain epidermal cells, notably the tall epithelium that occurs on the neural surface of the arm — two broad black bands, with a narrow white band between them, except at the tip of the axis ( e.g . 17. plate 5, figs. 23-24) — and in 34 “TERRA NOVA” EXPEDITION. scattered small cells alono; the distal halves of the tentacles of the arm. The tentacles, nevertheless, are on the whole markedly pale in comparison with the axis. In badly preserved alcohol-fixed material there is a uniform black stain that pervades the whole of the tissues, and even spreads to a certain extent to the coenoecial substance. The arms of such material, when teased apart and examined in dilute glycerine with intense illumination, appear of a greenish brown colour in the less black parts. The colour can be reduced partially or completely by the action of diluted Eau de Javelle (1 in 30) for about two hours, but the tint of the arm after such treatment is uniform. If the action be stopped before the colour has disappeared entirely, there is no indication of the paired black band along the axis. Why some of the alcohol-preserved material in the present collection is good and some is bad is not apparent. It may possibly be that in some cases the material had been allowed to remain so long in the trawl or in the bucket before being placed in alcohol that the zooids had died, and had begun to undergo disintegration, whereas the material that proves to be in good state of preservation was fixed with more expedition. Another possible explanation is that the strength of the alcohol used for preserving the material was not in every case maintained, that the pieces of colony, carrying naturally a large amount of sea-water, were dropped into bottles of alcohol, and closed down and brought home without changing the spirit subsequently. That this may have happened is probable from the fact that one bottle of badly fixed material, which according to the label should contain 70 per cent, alcohol ( i.e ., about 35 over-proof), had fluid which was salt to the taste, and gave an alcohometer-reading of 32 under- proof. The fluid in another bottle of alcohol-material yielded a reading of 45 under- proof. Both of these bottles contain material the zooids of which consist of little more than a black powdery mass. The formalin -fixed material from the same station (Station 314) is in excellent preservation. Yet another possibility is that some, or all, of the material that proves to be in bad condition may have been frozen solid before it could be adequately attended to with a view to preservation. Alcohol as a preservative for Cephalocliscus is not by any means to be condemned, for very good results can be obtained by its use. The method most likely to yield good fixation and preservation is to drop freshly-trawled pieces of colony into a bottle half-full of 40 over-proof alcohol (73 per cent, by weight), to pour away the fluid the next day, and fill up with 40 over-proof alcohol, and to repeat the process a fortnight later. The method, however, may be regarded as extravagant, for the fluid poured out on the second day would contain so low a percentage of alcohol as to be scarcely worth redistilling. Alcohol, even if not stronger than 40 over-proof, causes a little shrinkage of the coenoecial substance, whereas formalin solution seems to produce no alteration in the shape of the coenoecium. One can easily test this by placing a piece of formalin- preserved colony of C. hodgsoni into 40 over-proof alcohol ; in about ten minutes the CEPHALODISCUS — RIDE WOOD. spines will exhibit a rather shrivelled appearance. It passes oft' later to a slight extent, but not entirely. »/ A word may here be said in favour of formalin as a preservative for Cephalodiscus. The fluid can be made up readily with sea-water to the required strength — 5 per cent, yields good results, but 8 or 10 per cent, would probably be better if it is not proposed to open the bottle again until the end of the expedition — it penetrates well, and causes little distortion of the zooids. Although the finest histological details cannot be worked out in such material, the serial sections cut from it enable one to work out all the ordinary anatomical features, and no difficulty is experienced in staining the sections adequately if haematoxyhu be used in conjunction with Orange G. A further advantage that formalin has over alcohol as a preservative is that, wrhere it is desirable to transmit material to investigators residing in other countries than that from which the expedition started, the question of the payment of customs dues does not arise. An effort should, however, in each case be made to fix a few living zooids in appropriate hardening fluids, such as Perenyi’s fluid, and corrosive-sublimate-acetic-acid-alcohol, and transfer them after a few hours to 40 over-proof alcohol. The uniform staining of the arms in badly preserved alcohol-material is important in its bearing upon the specimens of C. nigrescens supposed to have been dredged on the “Erebus” and “Terror” Antarctic Expedition (19), which was most probably fixed in alcohol, the only common preservative of the time. The zooids do not show the paired black band on the arms, and there is some staining of the coenoecium. The uniform blackness of the arms puzzled me at the time when I was investigating the specimens, the coenoecial characters of which were obviously those of C. nigrescens ; and considerable light is thrown upon the problem by the discovery, among the “ Terra Nova ” material, of different pieces of colony from the same station, some preserved in formalin and showing pale tentacles and a paired black band on the axis of the arms, and other pieces poorly fixed in alcohol and having the arms uniformly black throughout, and the coenoecial substance also tinted. In formalin-fixed material of C. nigrescens there seems never to be a staining of the coenoecial substance ; thin branches are nearly hyaline and transparent, while thick pieces have at most an opalescent appearance. A spreading of the black colour may be seen in zooids fixed in corrosive sublimate solution. While most of the material of C. nigrescens collected by the “ Terra Nova ’ was fixed and preserved either in formalin solution or in alcohol, some zooids were extracted from freshly dredged colonies, and fixed in corrosive sublimate solution, and later transferred to 70 per cent, alcohol. In the arms of this latter material there is a fairly uniform dark greenish brown coloration, and the black bands of the axis are in many cases not discernible. It is as though the solid pigment-granules had been dissolved, and the resulting black fluid had imparted a uniform stain to the axis and tentacles alike. The remarks in the preceding paragraphs refer to the appearance of the arms G VOL. IV. 36 “TERRA NOVA'1 EXPEDITION. examined whole under a low magnifieation. In serial sections certain differences due to methods of fixation are to be noticed that are not obvious in whole arms. In zooids of C. nigrescens fixed in Perenyi’s fluid (material collected by the “Discovery ’) the pigmented cells of the epidermis of the neural surface of the arm are large, but not appreciably swollen ; they are of a deep brown colour, nearly black, or of a rich warm brown, with scanty black specks. In material fixed in corrosive sublimate solution (material collected by the “ Terra Nova ”) the pigmented epidermal cells appear greatly swollen and somewhat burst, and the pigment is in the form of a few scattered black spots. In material fixed in picric acid solution (material collected by the “ Discovery ”) the cells are swollen and have a vacuolated appearance ; they are clear, and show scarcely any pigment. In formalin-fixed material the cells of the neural surface ol the axis, although large, are not appreciably swollen, and the pigment is abundant and jet- black, or very dark brown. In the report of the National Antarctic (“ Discovery ”) Expedition it is stated (17, p. 25) that in serial sections of material fixed in picric acid solution the pigment cells of the arm-axis are considerably swollen and not brown or black in colour, but each cell shows its one or two small black spheres as in material fixed by other methods. No mention is made, however, of the appearance of the arms examined whole. A recent examination of material of the “ Discovery " Expedition that was fixed in picric acid solution and followed by alcohol, shows that the arms, axis and tentacles alike, are stained a uniform very dark brown, and the paired longitudinal black band is not visible. The epithelium of the neural surface of the axis of the arm does not appear unduly swollen, as it does in the serial sections. The acid seems to have the effect of dissolving and spreading the pigment in fresh material only, for the experiment was tried of soaking for 50 hours in a saturated solution of picric acid in 60 per cent, alcohol, with 2 per cent, of glacial acetic acid added, the dissociated arms of some zooids well preserved in alcohol. No visible change resulted ; the two black bands remained exactly as before. Similar well-preserved arms were soaked for the same length of time in a 10 per cent, solution of glacial acetic acid, and others in an aqueous solution of corrosive sublimate, with a little acetic acid, with similar negative results. Free ova in the material of C. nigrescens collected by the “ Terra Nova ” are similar to those previously recorded (17, p. 46), but are a little larger. They are mostly found in the blind end of an inhabited tube, and are usually single ; rarely there are two in the same tube. Six free ova were measured, and yielded the following results : — three were 0 * 8 x 0 * 7 mm. , two were l’O x 0 * 6 mm. , and the last 0 • 9 x 0 • 6 mm. The mode of extrusion of the ova from the ovary in Cephalodiscus is still an unsolved problem ; the oviducal aperture is extremely small in comparison with the great size of the ovum, and gives no indication of capacity for great dilatation. Masterman (15, p. 512) has suggested that the oviducts merely serve to admit the CEPHALODISCUS — RIDE WOOD. 37 spermatozoa into the ovary, and that the ova are liberated by the death of the mother and the subsequent disintegration of the tissues ; in other words, the ova are not laid, but survive the parental body. This view finds some confirmation from the fact that Andersson has observed the presence of spermatozoa in the ovary of C. clensus (2. p. 85), and Gilchrist records that in C. gilchristi the ovaries are sometimes discharged by the rupture of the body-wall, without the death of the mother necessarily resulting ; but he guards himself by a statement that the condition he is describing may have been brought about by pressure in the trawl-net (4, p. 192). In the case of C. hodgsoni it is by no means unusual to find zooicls with the dorsal body-wall ruptured and the intestine rent and partially evaginated, but this is a condition which one rather attributes to the violent contraction of the body on coming into contact with the fixing fluid. That in the case of C. nigrescent the eggs are definitely laid is probable from the frequency with which, in material obtained on the British Antarctic Expedition, a free ovum is found among the buds at the blind end of a tube inhabited by a female that still possesses two perfect ovaries. There can scarcely be any doubt in such cases that the free ovum found was produced by the zooid occupying the tube. In C. nigrescent the ova bulge into the cavity of the ovary (text-fig. 3), as in C. solidus (2, plate 7, fig. 64), and do not lie externally to the cavity, as they do in C. hodgsoni ( = inaequatus ; 2, fig. 63). The character as a means of dis- criminating species is not very satisfactory, for in the broadest part of the ovary the ova are so closely compacted that either no cavity is seen, or one cannot satisfactorily decide whether the cavity is the true lumen of the ovary or an artefact. It is only near the external opening of the ovary, where the ova are comparatively small, that the point can be settled. In the section represented in text-fig. 3 the ova are seen to be proliferated from one side of the oviduct (the side nearest the shield), and they bulge into the cavity of the ovary, much in the same way as is figured for C. solidus by Andersson (2. plate 7, fig. 64). CEPHALODISCUS DENSUS , Andersson. Cephalodiscus ( Orthoecus ) densus, Andersson. — For diagnosis, references and list of recorded specimens, see page 75. MATERIAL COLLECTED. Material of C. densus was obtained from McMurdo Sound, in Ross Sea, at Stations 314, 316, 339, 355 and 356, and from off Coulman Island at Station 295. For details as to depths and dates, and the precise positions of these Stations, see page 12. Text-fig. 3. — Cephalo- discus nigr escens, transverse section through the oviducal end of the ovary, showing unilateral proliferation of the ovicells. ( X 230.) o 8 “ TERRA NOVA” EXPEDITION. From each of the Stations 314, 316, 355 and 356 one bottleful of C. dens as was dredged ; the collection from Station 339 is contained in two bottles, and that from Station 295 in three bottles. The material from Station 314 is preserved in alcohol, as also is a part of the collection from Station 295. From the material obtained at Station 339 two living zookls were extracted by Mr. D. Gr. Lillie, and fixed in corrosive sublimate solution, and subsequently transferred to alcohol (plate 5, figs. 7 and 8). Tin- rest of the collection was fixed and preserved in formalin solution. The finest specimen collected is a large, complete, cake-like, nearly spherical colony, measuring 100 mm. in height, and 130 mm. across (plate 3, fig. 6). This was obtained from Station 356. The specimen is extremely difficult to handle, for the coenoecial substance is very soft and spongy in consistency, and when the mass is lifted out of the formalin solution, even with the mouths of the tubes uppermost, the fluid drains out. and the colony shrinks to about a third of its former bulk. On putting the specimen back again into the fluid it rapidly regains its original size. The base is comparatively smooth, and was probably resting upon a level muddy bottom, from which it was easily lifted by the trawl. The substance of the base is uniform, and shows no tubes, but the probability is that the tubes come to within a few millimetres of the actual under-surface. The colony has not been bisected, and so the course of the tubes within it can only be conjectured, but from an examination of other specimens in the collection it is probable that the tubes that reach the summit extend down vertically, and end blindly near the base, and are thus about 90 mm. in length. The other tubes are directed more or less obliquely, and those in the lowest parts, above the edge of the base, are nearly horizontally disposed. These tubes, judging from other specimens, would not be more than 30 or 40 mm. in length. The longest tubes actually traced out, in specimens other than that mentioned in the preceding paragraph, are 70 mm. in length. The shortest tubes found are 15 mm. long, but there are in all probability tubes shorter than this ; indeed, there is no reason why tubes should not occur as short as 5 mm., for a newly established zooid would not be likely in the first instance to secrete a tube much longer than its own body. The width of the cavity of the tubes is mostly 1 • 0 to 1 • 2 mm. ; near the ostium the width is usually a little greater than in the rest of the tube. The width of the cavity agrees with that of authentic material of C. densus examined by Andersson and now at the British Museum (Natural History), and it is evident, from the fact that Andersson states the “ Durchmesser ” of the tubes of C. densus to be l-5 mm., that of C. rarus 1'5 mm., occasionally up to 2’ 0 mm., and that of C. solidus slightly over 2’0 mm. (2, pp. 11-13), that he is quoting the external diameter of the free portions of the tubes. The width of the cavity I find to be less liable to fluctuation than the external diameter, and consequently better suited for systematic purposes. There are in the collection many pieces such as might result from the tearing up of a complete colony like that shown in plate 3, fig. 6, pieces about 70 or 80 mm. in CEPHALODISCUS — RIDE WOOD. 39 height and 50 to 80 mm. in width, with long tubes disposed in more or less parallel series, and loosely bound together by a very soft, spongy, common coenoecial substance that continues for 5, 10 or 15 mm. beyond the blind ends of the tubes, and extends to within 10 or 20 mm. of the ostia, thus leaving freely projecting lengths of tube which in the case of species of ldiothecia would be termed peris tomial tubes. From the characters both of the coenoecium and of the zooids there is no question that these pieces, like the complete colony, belong to C. ( O .) densus Andersson (2, p. 12). In the same bottles with such pieces, and in several instances actually continuous with them, are pieces such as those shown in plate 3, figs. 2, 4 and 5, in which the common coenoecial substance is less abundant, and although it binds together the basal ends of the tubes, leaves a much greater extent of the ostial ends of the tubes free. The tubes are less obviously in parallel arrangement, and the ostial ends are extremely delicate, and give one the impression of being but recently formed. One is thus disposed to regard such pieces as constituting the rapidly growing edges of large cake-like colonies, or outlying offsets from them, and to assume that if they had been allowed to remain undisturbed, the intervals between the tubes would have been gradually closed up by common coenoecial substance, and that the margins of the ostia would have become thicker and darker, more like the middle parts of the tubes, and that later, when growth became less active, the colony would resemble that shown in fig. 6. The piece shown in fig. 1 of plate 3 has even less union between the tubes than those reproduced in figs. 2, 4 and 5, and the tubes are still more irregular ; and in fig. 3 is shown a piece of colony in which the tubes are extremely irregular in their course, rambling and straggling, and only united at their basal ends. This last piece, and other similar pieces, of which there are several, I am unable to distinguish from C. rarus Andersson (2, p. 12). From a study of Andersson’s published descriptions, and a critical examination of authentic material of C. densus and C. rarus obtained on the Swedish South-Polar Expedition, and received by the British Museum (Natural History) from the Stockholm Museum in exchange for other specimens, I am convinced that Andersson’s two species stand in exactly the same relation to one another as the pieces of “Terra Nova” material represented in figs. 6 and 3 of plate 3. Andersson bases his distinction between the two species almost entirely upon the characters of the coenoecium, and such differences as he finds between the zooids of the two species (e.c/., the thickened epidermis of the axis of the arms, discussed on p. 22 of the present report) are such as might well occur within the limits of the same species. Assuming, as one is justified in doing, that a young, newly established colony is composed of short tubes, scarcely longer than the zooids themselves, the probabilities are that the tubes, as they are increased in length, may adopt a rambling course, being unimpeded by anything in the immediate vicinity, and the result will be a straggling 40 “TERRA NOVA” EXPEDITION. set of tubes such as those in fig. 3 of plate 3. But as the buds of the pioneer zooids separate off and construct tubes of their own, they incommode the earlier zooids, and also one another; and since the number of tubes is still increasing, the oldest tubes, in the middle of the mass, will be made longer and straighter and more closelv set. and intermediate, soft, coenoecial substance will be secreted to fill up the intervals between the tubes. Thus the colony ultimately may come to have the form of a hemisphere, or a cake with the edges thinner than the middle part. The middle would consist of long, straight, almost parallel tubes (except in their lowest, first-formed parts), and the edges of the cake would be composed of shorter, less straight, and more radiating tubes, less closely cemented together. Arguing along these lines, one may draw a distinction between the eoenoecia of Orthoecus and Idiothecia as regards their modes of growth : for in Idiot hecia the youngest zooids and the shortest tubes are to be found at the free ends of the branches, whereas in Orthoecus the youngest zooids and the shortest tubes are found around the margin of the mass, and the colony in consequence does not adopt a branching form. One may describe the mode of growth of the colony of Idiothecia as apical, and that of Orthoecus as peripheral. With such a complete gradation of specimens between the rarus form of Orthoecus (fig. 3) and the densus form (fig. 6) as is now available for study in the collection made by the “ Terra Nova,” it becomes clear that one of the two specific names should disappear. Since on page 12 of Andersson’s monograph (2) the diagnosis of C. densus precedes that of C. rants , I adopt the former name to include the latter ; so that C. rarus is in the present report regarded as a synonym of ( '. densus. The whole of the material of Orthoecus collected bv the “ Terra Nova ” I consider to be C. densus. Incidentally it may be mentioned that if C. rarus becomes a synonym of C. densus , it is doubtful whether C. anderssoni Gravier (6 and 8) can stand as a distinct species. The characters of the coenoecium are not very different from those of the specimens of Orthoecus obtained by the “ Terra Nova,” the prevailing feature of Gravier’s species being apparently a grouping of the tubes in clumps of four or more, which stand out more or less distinctly from the other clumps (infra, p. 76). The zooids are stated to be in a poor state of preservation, and to have a certain pigmentation of the stalk and of the shield. The number of arms in C. anderssoni is not mentioned, but buds are said to have arms up to six pairs. Although there is so close a resemblance between the eoenoecia of the forms that have been described by Andersson as C. densus and C. rarus, and by Gravier as C. anderssoni, the species termed C. solidus by Andersson (2, p. 11) is markedly distinct from these. The common coenoecial substance of C. densus is extremely soft and spongy, and liable to disintegration, whereas that of C. solidus is firmer and more gelatinous, rather resembling that of C. nigrescens in its consistency. A piece of colony of C. densus shaken up violently in a bottle half full of water would separate into its CEPHALODISC U S— RIDE WOOD. 41 individual tubes, but a piece of C. solidus similarly treated would break across the tubes almost as readily as along them. The thick peristomial lip of the tube of C. solidus finds no equivalent in C. densus, and the blackish colour of the zooids marks them oft' from those of the latter species. The entanglement and inclusion of diatoms and sand-grains in the softer material of the coenoecium, to which Andersson alludes (2), is a character which one comes to regard with suspicion. A study of specimens of C. nigrescens leads to the conclusion that if foreign particles are abundant they will in all probability be included in the secreted coenoec-ial substance : and although the specific name agglutinans was applied to the form of Cephalodiscus dredged by the Scotia " on the Scottish National Antarctic Expedition (11) because the material collected includes so much in the way of shell-fragments, yet this inclusion may be purely local, and it is quite possible that in other, more rocky, parts of the ocean-floor the zooids of the species might build up a perfectly clear and transparent coenoecium. This question is dealt with in more detail in the section on C. evansi (p. 27). The part of the material from Station 295 that was fixed in alcohol — the other two bottles contain formalin — consists of half a dozen fragments measuring 45 to 60 mm. in height and 25 to 50 mm. in width, evidently pieces broken from a large mass. The tubes are fairlv straight, and are bound loosely together by soft coenoecial substance except for their uppermost 8 or 10 mm. The freely projecting parts are of a darker tint than the embedded parts of the tubes, the reverse of what is found in some other portions of material of this species ( e.g ., plate 3, figs. 1-5). One peculiarity of this material is the frequent obliquity of the ostium and the sudden widening of the tube at its free end (text-fig. 4, A). A similar obliquity of the free end of the tube has been noted by Gravier in the material that he describes as C. anderssoni (6); and it is also seen in some of the tubes of Andersson ’s fig. 6 of C. rarus (2), although not in fig. 5. In Gravier’s material it is the laterally placed tubes that end in this manner. It is probable, therefore, that the pieces of C. densus from Station 295 are lateral fragments of a very large cake-like colony. The longest tubes dissected out measure about 55 or 60 mm., and have an average internal diameter of 1*1 mm. Curved septa occur in the basal parts of the tubes (text-fig. 4, D). In certain pieces of colony there is a tendency for the tubes to exhibit externally projecting flanges around the ostium, and also lower down the tube. A few of this kind are shown in fig. 6 of plate 5. They indicate what may be interpreted as a slow rate of growth of the tube, whereas the smooth, colourless terminal portions of tubes, seen in figs, i-5 of plate 3, represent a rapid lengthening of the tube. A few tubes with flanged mouths are to be seen in the left-hand part of fig. 6 of plate 3. In the pieces of colony in which the upper parts of the tubes are delicate and colourless (figs. 1-5, plate 3), the basal parts are yellow or brown in colour, tough, and with a distinct sheen, or sometimes an iridescent glint ; the same holds good in the 42 TERRA NOVA” EXPEDITION. case of long parallel tubes such as belong, one assumes, to the middle of a large cake- like colony. One tube dissected out is of interest in that it shows, applied to its side, a much shorter tube adhering by its basal, blind end (text-fig. 4, C). The interpretation of this association, I take it, is that the part of the main tube where the short tube is Text-pig. 4. — C. densus, tubes dissected out from the colony. ( X 3J-.) The tubes are rendered dia- grammatically, and are drawn shorter than they should be in proportion to their width. A — a tube with a bulbous blind end, and an oblique ostium ; Station 295. B — a tube with a number of septa, the highest being more than halfway up ; the ostium is square to the axis of the tube ; Station 316. C — a young tube applied to the side of the upper portion of an older tube ; Station 316. D — a tube with a zooid that has sealed itself in by means of a flat transverse septum ; numerous curved septa occur in the lower part of the tube ; the ostium is moderately oblique ; Station 295. tubes of the piece of material in which 1 occluding septa. attached was at the time not enclosed in common coenoecial substance as it is now, and a young zooid settled upon the side and began constructing its own tube. As the colony continued to enlarge, both of the tubes were lengthened at their ostial ends at about the same rate, and common coenoecial substance was de- posited higher and higher up between and around the tubes, until the present level was reached. Another tube encountered is of in- terest in that it contains what appears to be a “ dormant” zooid (text-fig. 4, D). The zooid occupies a position about half- way up the tube, and rests upon the uppermost of a series of curved septa. Immediately above the arms of the zooid is a thin, flat septum, and upon this rest a number of free sand-grains. In the case of partially sealed-up tubes of colonies of C. solidus it has been sug- gested (21) that the material when brought up in the trawl was left for some time before being dropped into the pre- servative solution, and the zooids, for their own protection, rapidly secreted a loose operculum of coenoecial substance at the mouth of the tube. In the present instance the suo-o'estion does not seem to uo apply, for the closing partition is thin, like the curved septa in the basal end of the tube, and the neighbouring is one was found do not exhibit similar CEPHALODISCUS — RIDEWOOD. 43 ZOOIDS. The zooids vary a good deal in size, and the wide range may be judged from the following measurements. The first is the length in millimetres from the free ends of O O the arms to the caecal end of the body, the second the length from the bases of the arms to the end of the body, and the third is the width of the body. Station 295 6‘0 — 3'2 — 0‘ 8 4-5 — 2*4 — 0-7 3-7 — 2-0 — 0-8 3- 5 — 1-9 — 0-9 Station 316 7 ' 4 — 4 • 3 — 0 • 9 7-0 — 5-0 — 0-9 6*2 — 3-9 — 0-8 4- 8 — 3-0 — 0-9 Station 339 5*5 — 3- 5 — 0’9 5- 0 — 3-4 — 1*0 Text-fig. 5. — Ceplialodiscus densus ; Station On an average based upon observations on Swedish Expedition specimens, Australasian Expedition specimens (21), and “ Terra Nova” specimens, one may quote the measurements for zooids of C. densus as : — 4 to 7— 2 to 4—0 '8 to 1*0 mm. The general aspect of the zooids may be seen from figs. 7 and 8 of plate 5, which are reproductions of photographs of zooids removed from their tubes alive, and specially killed in corrosive sublimate solution. The colour of the zooids is in some cases distinctly orange or brown, but in most it is ochreous, or pale brown, or greyish white. The shield of C. densus differs in no important respects from those of other species of Cephalodiscus. When fairly well expanded it presents a roughly circular outline, and the mean of the antero-posterior and the lateral diameters is about l'l mm. When dissected off and examined by transmitted light it shows a dark area in the VOL. IV. H 44 “TERRA NOVA” EXPEDITION. middle of the main or anterior lobe, and a rather dark band in the posterior lobe, touching the red line. The marginal part of the posterior lobe is thinner and more transparent than the marginal part of the main lobe (text-fig. 5, B). The lateral indentations are less pronounced than those exhibited by C. nigrescens, C. hodgsoni and C. dodecalophus (17, text-fig. 9, C, p. 27, and text-fig. 17, E and F, p. 54); but this is largely a question of the degree of contraction in which the zooid died ; the smoothness of outline shown in text-fig. 5 is only to be seen in zooids that have died in a flaccid condition. The lateral notches are frequently better marked in buds than in adults (see text-fig. 8, F, p. 47). The oral lamella varies considerably in its outline in different zooids. In some, presumably those that have died in a relaxed condition, the margin is even, with a very slight wavy curve ; in others (text-fig. 5, A) the seven, rarely eight, sinuosities or scallopings of each lateral lobe are very pronounced. The figured specimen was dissected from the zooid shown in plate 5, fig. 7. The arms, when in a state of full extension (text-fig. 6, B and C), measure about 3 mm. If dissected apart, they tend to lie upon their sides on the slide, but occasionally one finds an arm with the twro rows of tentacles lying in the same plane (Gl). The axis of the arm does not terminate in an end-swelling with refractive beads, so characteristic a feature in C. dodecalophus and C. hodgsoni, and it is necessary to avoid drawing a false conclusion from moderately contracted arms, such as F, which have a slightly enlarged, incurved tip to the axis. Such extremities have no highly refracting beads, and they occur singly, not more than one, for instance, in the set of sixteen arms of a zooid ; most zooids have none at all. Similar occasional enlarged terminations are met with in C. gilchristi (16, p. 184, and text-fig. 2, B), and something approaching them is sometimes found in C. nigrescens (17, plate 5, fig. 26) and C. agglutinans (11, p. 549, text-fig. 4, C). The arms are usually eight pairs in number, but two zooids were found with seventeen arms, and several with fewer than sixteen. Interest attaches to some of these latter cases, for they occasionally show diminutive, or arrested arms, not more advanced in their development than those that occur in late buds, such as the developing arms shown in text-fig. 8, J, K, L, p. 47. Two of these arrested arms are represented in text-fig. 6, H and J, and another in the lower part of figure B, showing its size in relation to the full-grown arms. The alimentary canal does not differ materially from that of other long-bodied species, such as C. nigrescens. The stomach extends nearly to the end of the trunk ; the gastric caecum projects so far forwards between the pharynx and intestine that its extremity is visible in transverse sections that include the gill-slits. Male and female zooids and hermaphrodites (with one ovary and one testis) occur in different tubes of the same colony. The testes, although of the long, as opposed to the globular variety, are not so long as those of C. nigrescens and C. solidus. Perhaps “ elongate pyriform ” would describe them better. The total length is about CEPHALODISCUS— RIDEWOOD. Text-fig. 6. — Arms of Ccphalodiscas densus ; x 35. A, a contracted arm, with short tentacles near the extremity. B, three arms in a state of full extension, and another very diminutive arm at their base. C and D, typical examples of extended and moderately extended arms. E, a contracted arm, differing from A in having longer tentacles near the extremity. F, an arm with a slightly enlarged and incurved tip simulating an “end-swelling.” G, an arm seen in flat view. H and J, arrested arms, very small arms occurring sporadically among arms of full size ; another is shown in the lower part of B. 46 “TERRA NOVA” EXPEDITION. four times the width (text-fig. 7, A), sometimes five or six times, but immature testes are relatively short (D). The spermatozoa (text-fig. 7, F) are very similar to those of C. gilchristi (16, text-fig. 5, J, p. 189) ; the tail is two and a half times as long as the head. The head is rather long (0*005 mm.), and tapers to a point. The ovaries are large in mature zooids (text-fig. 7, B), and show one, some- times two, large, heavily yolked ova in the posterior end. The oviducal end has the Text-fig. 7. — Gonads of Ceplialodiscus densus. A and B, testis and ovary of the same zooid. C, a young ovary. D, ovary and testis of a full-sized but immature zooid. E, the two ovaries of a full-sized but immature zooid. F, ripe spermatozoa. Fig. F is X 1320, the other figures are X 66. usual irregular patches of red pigment. Young ovaries appear as shown in figures C and D, very young ovaries as in E. When the gonads are smaller than this — figures A-E are drawn to the same scale — it is not possible to decide in dissected preparations whether they are ovaries or testes, and recourse must be had to the study of serial sections. CEPHAL0DISCT7S— RIDEWOOD. 47 The liberated ova found in the tubes of the colony are free, not attached by a stalk ; they measure about 0 • 8 by 0 • 7 mm. As a rule not more than two ova occur in one tube, but in one instance, from Station 316, three ova were found in a tube that was inhabited by a hermaphrodite individual, the ovary and testis of which were both large, and apparently equally mature. Text-fig. 8. Cepliulodiscus densus ; buds; x 55. A, a very young bud, side view. B, a slightly older bud, ventral view. C, a bud just before the appearance of the red line of the shield, dorsal view. D, an older bud, in side view, with the posterior lobe of the shield turned forward. E, a bud of about the same age, dorsal view, showing the first two pairs of arms. F, an older bud, with three pairs of arms developing. G , a bud with the fourth pair of arms appearing ; the posterior lobe of the shield is turned forward, as in D. H, the oral lamella and the five pairs of arms of an older bud. J, a similar preparation from a bud with six pairs of arms — one arm on the left side of the figure is concealed from view. K, the largest arm of an older bud possessing thirteen arms. L, another arm of the same bud. The part of the material from Station 295 that was fixed in alcohol contains zooids in an excellent state of preservation, and these were studied in detail by means 48 “TERRA NOVA” EXPEDITION. of serial sections. Sections cut parallel to the sagittal plane show that the notochord has a slight but fairly constant sigmoid flexure, the heart being attached at one of the concavities, at about one-fourth of the length from the anterior end. The pericardium extends some little distance beyond the extremity. The longest of six notochords measured is O' 33 mm., and the shortest O' 23 mm. in length; these figures agree tolerably well with those of Andersson (2, p. 59), wdio gives the length in C. rarus and C. clensus respectively as O’ 28 and O’ 24 mm. The extremity of the notochord is not appreciably swollen (cf. Demiothecia), nor is it sharply bent (cf. C. gilcliristi, 16, p. 186, text-fig. 3). The anterior two-thirds of the notochord is in longitudinal sections well defined and easy to follow, but the hindermost part is difficult to trace with precision. The longest notochords are also the thickest. There is a considerable range of variation in the thickness of the same notochord at different points, and a transverse section through the middle of the length of a notochord varies in shape in different zooids from a circle to a dorsoventrally compressed ellipse. The mean of the horizontal and vertical diameters of such a section stands, however, fairly constant at somewhere between O’ 040 and O’ 045 mm. (cf. 2, p. 62). BUDS. The buds on each zooid are numerous, commonly eight, rarely as few as five, occasionally as many as thirteen or fourteen. In the late development of the arms of the buds the species resembles C. nigrescens (18, text-fig. 6, p. 236), C. agglutinans (11, text-fig. 5, p. 555), C. gilcliristi (18, text-fig. 9, p. 243) and C. solidus (21, text- fig. 3), and differs from the various species of Demiothecia — e.g. C. dodecalophus (18, text-fig. 4, p. 231) and C. hodgsoni (18, text-fig. 2, p. 225). In text-fig. 8 are given illustrations of buds of C. densus in a graded series, from those without any traces of arms (A — C) to one showing the sixth pair in process of development (J). The tentacles of the arms make their first appearance when the bud has five pairs of arms (H.) The oldest bud found had a series of thirteen arms, the largest of which is shown in figure K. CEP HA L O DISCUS HODGSONI , Ridewood. Cephalodiscus ( Demiothecia ) hodgsoni, Ridewood. — For diagnosis, references and list of recorded specimens, see page 67. MATERIAL COLLECTED. Material of C. hodgsoni was obtained from Stations 314, 316, 338, 339, 340, 348 and 355, all of which are in McMurdo Sound, in Ross Sea. For details as to depths and dates, and the precise positions of these Stations, see page 12. From each of the CEPHALODISCUS— RIDEWOOD. 49 Stations 316, 338 and 340 one bottleful of material was dredged ; the collection from Station 314 occupies seven bottles, that from Station 348 seven bottles, that from Station 339 live bottles, and that from Station 355 live bottles. Six of the twenty- seven bottles contain alcohol, the rest formalin solution. Some loose zooids from colonies collected at Station 314 were fixed in corrosive sublimate solution, and subsequently transferred to alcohol. The finest colony of the whole collection is one which was obtained from Station 355. As now mounted on a sheet of Mass in a flat-fronted bottle for exhibition in the O public galleries of the British Museum (Natural History), it measures 250 mm. in height, and 150 mm. in greatest width. It is mostly of a rich amber colour, but the extremities of the branches are pale and nearly transparent. The coenoecium contains very few zooids. Illustrations of selected pieces of colony of C. hodgsoni are given in plate 4, figs. 1-3, and plate 5, figs. 1-5. Comparing fig. 1 with figs. 2 and 3 of plate 4, there is to be seen a marked difference in general “habit” or “ facies” which might easily influence one into regarding the specimens as belonging to different species, and it is only after prolonged and careful study of the whole of the material of Demiothecia collected by the “Terra Nova” that I have come to the conclusion that there is no specific difference between them, and this in spite of the fact that one of the forms has, in a majority of the zooids, twelve arms, whereas in zooids of the other form the number is more commonly ten. The specimens reproduced in plate 4 are specially selected because they show the differences between the two forms in a most marked degree — they are extreme cases ; but while in the majority of instances there is not much difficulty in deciding to which form any particular specimen belongs, there are a few cases which, in the features of the coenoecium, are strictly intermediate in character, and there are other pieces of colony which exhibit the distinguishing features of the one form in certain parts and those of the other form in other parts. The pieces of intermediate character and those of composite character are scarce, but they are sufficient to prevent any hasty division of the material into separate species. The pieces in question have no zooids, and so appeal cannot be made to the number of the arms. Since cases are known of two distinct species of Cephalodiscus growing in continuity ( C . hodgsoni and C. nigrescens, page 33), the possibility is not excluded that the composite pieces referred to above were being built up simultaneously by the zooids of two species of Demiothecia living in harmony ; but I do not believe this is a correct explanation. In any case, it would not account for the origin of pieces of colony of intermediate character. For purposes of reference I propose to allude to the form exemplified in fig. 1 of plate 4 as Form A, and that illustrated by figs. 2 and 3 as Form B. Form A is more common than Form B ; it was obtained from the seven Stations 314, 316, 338, 339, 340, 348 and 355, and Form B from Stations 314, 339 50 “TERRA NOVA” EXPEDITION. and 348. It will be noticed that Form B was not obtained from any station that did not yield Form A. Even within the limits of Form A the diversity of coenoecial characters is very considerable. In figs. 3, 4 and 5 of plate 5 are shown three rather extreme instances. The piece shown in fig. 5 has spines that are much more thickly set and crowded than those of the large piece shown in fig. 1 of plate 4 ; and, judging from the delicacy of the spines at the upper extremity, active growth was proceeding there when the specimen was dredged. The lower part of the specimen is of a deep amber colour, the upper part is pale and transparent. Fig. 3 of plate 5 illustrates a piece of colony that is much branched and straggling, without a definite main axis, and with long, stiff spines ; the whole is of a deep amber colour, and there is no evidence that growth of the branches was taking place at the time of capture. The piece shown in fig. 4, however, wdiile exhibiting the same general type of branching as that shown in fig. 3, is smaller, extremely pale, and transparent, and is what I consider to be a rapidly proliferating part of a colony. These three pieces have been selected for figuring because they exhibit the extreme features mentioned above. Whatever doubts I may have had regarding the propriety of including Forms A and B within the same species, I have no hesitation in concluding that these three pieces are specifically identical, and that they belong to Form A of the species. The amount of material collected and now available for study is large, and the occurrence within it of every intermediate gradation of coenoecial facies between the extremes here illustrated warrants the grouping together of the specimens within the same species. Nor are any differences that could be regarded as justifying a splitting of the species to be observed in the zooids of the specimens. AVhile I have not seen in one and the same colony the features represented severally by figs. 3, 4 and 5, it is not difficult from the abundant material available to pick out colonies which in different parts are like figs. 3 and 4, or like figs. 4 and 5, or like figs. 3 and 5. The specimens shown in figs. 3 and 4 were obtained from Station 348 ; that shown in fig. 5 is from Station 355. In fig. 2 of plate 5 is shown a piece of Form A from Station 338 differing from that represented in fig. 5 in being rather more stoutly built, of a still darker amber colour, and without the indications of growth at the apex. In general aspect it is strikingly different from the more straggling piece shown in fig. 1 of plate 4. In marked contrast with it is another (fig. 1, plate 5), from Station 355, exhibiting a very definite main axis, with less closely set spines, and with two rapidly growing regions, one at the apex and one about halfway down. This piece is clearly a piece of Form A, but on comparison with fig. 3 of plate 4 it will be noticed how the uppermost part of it tends to approach Form B. Of the three pieces of colony of C. liodgsoni referred to on page 33 as growing in continuity with the coenoecium of C. nigrescens, that from Station 348 and that from Station 314 are of Form B, whereas that from Station 316 is of Form A. CEPHALODISCUS— EIDEWOOD. 51 FORMS A AND B OF C. HODGSON I. An attempt to describe differences so intangible, subtle and elusive as those subsisting between the Forms A and B of C. hodgsoni will necessarily appear unsatisfactory and unconvincing : the statements must take the form of broad generalisations, followed immediately by qualifications that rob them of a large proportion of their significance. The generalisations are essential if any distinc- tion at all is to be drawn between the two Forms, and the qualifications are necessary to prevent the several distinctions from being regarded as arbitrary and absolute. Speaking in very general terms, one may say that the coenoecium of Form B is constructed on a larger scale than that of Form A; the inhabited cavities are larger, and the spines, particularly the terminal spines, are longer and thicker, and exhibit less forking ; the terminal ostia are larger and more funnel-shaped, and lateral ostia are less frequent. In Form B there is rather less difference in colour between the apical and basal parts of the colony than in Form A. The rich amber tint of the older parts of the coenoecium of Form A does not appear in any of the specimens of Form B in the collection ; the older parts of the colony of the latter are dull brown, while the most recently secreted parts are almost colourless. The inhabited cavities in the coenoecium are on the whole rather wider in Form B than in Form A. Measurements of two typical pieces taken from the same bottle of material dredged at Station 348 show that in Form B the cavity has an average diameter of 4’ 5 mm. ; in some places, however, the width is as large as 6 mm., and in others as small as 2 • 5 mm. In Form A the average diameter of the cavity of the coenoecium is 3 mm., but in some places the width is as great as 5 mm., and in others as small as 1 * 5 mm. In Form B the ostia are mostly terminal, funnel-shaped and large, measuring 2*5 to 3 ' 5 mm. by about 1 • 5 to 2 mm. ; there are other ostia at the ends of short side branches, and these are terminal in a sense ; the few really lateral ostia that occur, situated on the side of the axial part of the colony, are smaller, sometimes as small as 1 • 6 by 1 • 4 mm. With the terminal ostia are associated two or three, sometimes four spines. In Form A the ostia are more uniform iu size, and measure about 1*5 to 2 mm. by 1*0 to 1*5 mm. ; they are much more commonly lateral than terminal. Newly formed ostia measure about 3 by 2 mm., but they are subject to reduction in size later. In both Forms the terminal ostia are difficult to measure by reason of the sloping spines that support the sides. In the “ Discovery ” Expedition report is given a diagrammatic representation of what a cast of the interior of the coenoecium of C. hodgsoni would look like ( I 7, plate 4, fig. 22) ; this sketch, as may be judged from the preponderance of terminal ostia, clearly applies to Form B, VOL. IV. I 52 “TERRA NOVA” EXPEDITION. It may seem inconsistent to state, as is stated above, that the average diameter of the inhabited cavity of the coenoecium in Form B is 4-5 mm., and that the ostia are funnel-shaped and measure only 3' 5 by 2 mm., but in the first place it is the general disposition of the spines around the ostium that is largely responsible for creating the impression of a funnel, and it is only where the oval ring is actually complete, at the bases of these spines, that the ostium can be said to occur ; it is only here that any measurement can be taken. In the second place, it will be readily understood how, by the swelling of the cavity just below a branching of the main axis or a lateral axis of the colony (17, plate 4, fig. 22), the cross-section of the cavity is, on the whole, larger than the size of a single ostium. The common mode of multiplication of the ostia of Form B is by the bifurcation of the ostium at the end of the tube (text-fig. 11, p. 56, and explanation, pp. 56-57). There are also terminal ostia in Form A, but as the axial tube grows longer, the tendency is for a cross-bar to be secreted between two spines on one side of the ostium, and the new orifice, after reduction to about 3 by 2 mm., may remain as a lateral, sessile ostium when the tubular axis of the branch grows longer ; or, by the addition of material to its edges, the oval ring may grow into a short lateral branch with a terminal ostium ; or it may close up altogether (p. 57). There are occasionally found, running across the internal cavity of the coenoecium of Form A, solid bars of about the diameter of spines, as though the growing end of a dwelling-tube had encountered an oblique spine or a connecting bar passing right across its terminal ostium, and instead of the obstruction having the effect of dividing the ostium into two ostia, the tube had continued to grow beyond it, leaving it as a trabecula or prop passing across the dwelling-cavity. Bars traversing the coenoecial cavity in this manner are not the exact equivalents of the irregularities found in C. clodecalophus , for in that species the smooth continuity of the inner surface of the wall is broken by irregular partitions, pockets and excavations (17, text- fig. 4, p. 8). The spines of Form B, particularly the terminal spines, are longer and thicker than those of Form A, the average diameter being 1*0 mm. as against an average diameter of 0 • 7 5 mm. in the latter material. But one can without difficulty pick out slender spines of Form B that are thinner than specially selected thick spines of Form A. In basal parts of the colonies the differences in the spines are less marked than in the distal parts. In most cases the spines of Form B are more pale and transparent than those of Form A, but in rapidly growing apices of branches of Form A the coenoecial substance of both dwelling-tube and spines is exceptionally delicate, pale and transparent (e.g. in fig. 4 of plate 5, and in the uppermost part of fig. 5) ; in such parts the spines are very thin, sometimes not more than 0 ' 4 mm. across. The lines seen in the spines, marking the planes between the successive increments, are in Form B to a large extent longitudinally disposed, whereas in Form A they more readily suggest the successive application of short caps of coenoecial substance ; CEPHALODISCUS— RIDEWOOD. 53 in other words, the cap-like increments, which really occur in both Forms, extend farther down the spine in Form B than in Form A. One of the distinguishing features of Form A is the readiness with which the spines fork ; in Form B they fork also, but less commonly than do those of Form A. The mode of forking of a spine in Cephalodiscus is not dichotomous, but by the application of a new spine to the side of an older one (1 7, plate 4, fig. 21), and the new spine may be longer than the part of the older spine that lies distally to the forking. The greatest amount of branching that I have noticed is a case in which one spine-base carried a total of seven spine-tips. The general impression, gained by a casual glance, that there are more spines related to each ostium in Form A than in Form B, is due to the fact that the spines are more forked in the former. The basal parts of the spines are of about the same number in both — namely, if the ostia occur in a group of two, three, or four, usually two to each (text-fig. 10,/) ; if they occur singly, about three spine-bases to each. Counting the spine-tips in Form A, there would be about four or five to each solitary ostium. In Form A there is a tendency for cross-bars to connect up the spines, sometimes resulting in a complicated meshwork of bars and spines. In Form B the occurrence of such cross-connections between the spines is so uncommon that one has to search diligently to find them. Many of the spines that are seen projecting freely from the surface of the coenoecium are embedded for a considerable proportion of their total length in the coenoecial wall*' ; this is particularly the case in Form A, but the remark applies also in a lesser degree to Form B. The embedded parts, one may assume, were at one time free spines, but they have since been covered in by the zooids, who have utilised the support afforded by them as a skeletal structure against which to attach the dwelling- tube at that time in process of elongation. Reference to text-fig. 9, B, will show that the innermost layer of the dwelling-tube is complete, and applied to the side of the spine, but the more external layers envelop the spine as well as the inner layers of the tube. One might say, indeed, that the whole coenoecium of a colony consists of a skeleton in the form of a continuously forking spine, with an irregular dwelling-tube running along it, sometimes on the near side, sometimes on the far side, while, coming off at various points from the main spine, and also independently of it, are other spines, not necessarily thinner, which stand out as the superficial spines, and fork and have cross- bars, but are not, or are not yet, essential components of the wall of the dwelling-tubes. Perhaps the best method of explaining the differences that can be detected between Forms A and B of C. hodgsoni will be to describe in detail selected typical examples, such, for instance, as those represented in text-figs. 9 and 10. * If a piece of colony without zooids be transferred through the usual grades of alcohol to benzene, and paraffin, and then taken back through benzene and alcohol rather rapidly, so as not to dissolve away all the paraffin, the main or axial supporting spines show up very distinctly. I 2 54 “TERRA NOVA” EXPEDITION. In text-fig. 9 is shown a typical end-portion of a colony of Form A. The piece includes a total of seven ostia and fifteen spine-tips. The characteristic forking of the spines is seen, also a cross-bar to the left of spine c. At the base of the spine marked b is a large terminal ostium opening upward, and to the left of this are two lateral ostia, one on the near side and one on the far side of the base of spine c. These ostia, one may conclude from a study of the successive layers of growth of coenoecial substance, were preceded by a large terminal ostium, but the development of the cross-bar (a) had the effect of dividing it into two ; the right-hand parts of the openings became closed up, and there now remain two lateral ostia. In the lower part of the figure is a lateral ostium on the farther side. A typical lateral ostium occurs on the near side on the Text-pig. 9. — A — a typical end-portion of a colony of C. hodgsoni, Form A ; Station 314. ( x 3J.) For explanation of the lettering, see text, p. 54. B — transverse section of an internode of the axis of a colony of Form A, through a part where the cavity is at a minimum and no branches occur. Sometimes two spines are embedded in the wall, on opposite sides of the dwelling-tube. right, and a smaller one, in the distance, high up on the right. To the left of this last ostium there is, on the far side, a terminal ostium not shown in the figure. In text-fig. 10 is shown a typical end-portion of a colony of Form B. The piece includes a total of nine ostia and twenty spine-tips, but not all of these are visible from the side drawn. The spines are seen to be a little longer and thicker than those shown in text-fig. 9, and such forking as occurs is, in the particular piece selected for illustration, on the embedded parts of the spines, and not on the free portions of the spines as in text-fig. 9. In the left lower corner is an ostium (a), surrounded by three CEPHALODISCUS— RIDEWOOD. 55 spines. This is a terminal ostium, set on the end of a short lateral branch ; there is a similar ostium on the far side, not visible in the view taken. Near the middle of the lower part of the figure are two spines standing out obliquely towards the observer ; these are not related to any ostium, though they probably were when first secreted. Text-fig. 10. — A typical end-portion of a colony of C. hodgsoni, Form B; Station 348. (x 3^.) For explanation of the lettering, see text, p. 54. Constituting the whole right-hand portion of the lower half of the figure is a broad, short side branch with twro terminal ostia ( b and c), one of which (c) shows the beginning of a division into two, a small spine marking the position of such division. A similar 56 “TERRA NOVA” EXPEDITION. ostium in a state of division occurs between the spines d and e ; one small spine at the division is seen, and another slopes away from the observer in the distance. There are on the far side of the upper part of the specimen — not visible from the front — two ostia and two spines ; there is also a spine in the lower part of the specimen not seen in the view taken. So that, counting each dividing ostium as two ostia, and counting all the spines, small and large alike, there are in the figured specimen a total of nine ostia and twenty spines. Text-fig. 11. — Ends of branches of C. hodgsoni, Form B, partially diagrammatic, showing the mode of origin of new terminal ostia by the bifurcation of pre-existing ostia ; Station 339. ( X 3J.) For explanation of the lettering, see text, p. 56. Figure / of text-fig. 10 is a diagram of an end-view of a group of four ostia of Form B, with the associated seven spines cut short. There are, that is to say, roughly twice as many spine-bases as there are ostia, a proportion which conforms with that found in the piece of colony described in the preceding paragraph. A common mode of origin of new terminal ostia in Form B, by the bifurcation of pre-existing ostia, is explained in text-fig. 11. A simple ostium such as is seen at a grows wider as the dwelling-tube and the spines lengthen, and acquires the form of an oval CEPHALODISCUS— RIDEWOOD. 57 funnel, the spines being situated usually, but not invariably, at the ends of the longer diameter. A pair of new spines begin to develop on the edges of the oval, as at b, and after the tube and the four spines have lengthened, as at c d , a connection is established between the two new spines, so as to divide the oval orifice into two nearly circular ostia. As the two tubes, each now with two spines, become longer, and the ostia grow more funnel-shaped, a time comes when the two parts are short separate branches such as are seen at e and / in the lower part of the figure. A common mode of origin of new ostia in Form A, on the other hand, is by two diverging spines becoming connected, fairly low down, by a cross-bar such as is seen to the left of spine c in text-fig. 9. Applications of coenoecial substance reduce this triangular aperture to an oval one, and the dwelling-tube grows up along the side of the spines as far as the cross-bar, and ultimately beyond it, leaving the orifice as a lateral ostium. Commonly the triangular space is closed up completely, and the sheet of substance becomes part of the wall of the dwelling-tube. SPECIES OF THE SUB-GENUS DEMI 0 THE Cl A. The sub-genus Demiothecia, Ridewood (17, p. 8) was founded to include those species of Cephalodiscus in which the cavity of the coenoecium* is continuous through- out the colony, and to exclude those species in which each perfect zooid dwells in a tubular cavity of its own. Until the appearance of Andersson’s report (2) the sub- genus included the species C. dodecalophus, MTntosh, C. gracilis , Harmer, C. sibogae, Harmer, and C. hodgsoni, Ridewood ; to these Andersson added two species, C. aequatus (2, p. 9) and C. inaequatus (2, p. 10), which are less clearly differentiated than the four preceding species. The species of Demiothecia are extremely difficult to distinguish, either by the characters of the coenoecium or by those of the zooids. C. gracilis and C. sibogae may be separated off from the others by reason of the diminutive size of the coenoecium, and also of the zooids, although the latter are not readily distinguishable the one from the other, because of the fact that male zooids of C. gracilis and female zooids of C. sibogae are not known. Indeed, Harmer (10, p. 4) states that the possibility is not excluded that C. sibogae is the male form of C. gracilis. As regards the remaining species the difficulty of discriminating between them is very considerable. The material dredged on the Swedish South-Polar Expedition from Stations 59 * Coenoecium. Despite the arguments brought forward (17, p. 20, footnote) against the use of tho term coenoecium to designate the secreted dwelling of the Pterobranchia, the alternative term suggested — tubarium — has not met with general acceptance. While the word tubarium is appropriate in the case of Rhahdojdeura, Ortlioecus and Idiothecia, its suitability when dealing with species of Demiothecia is less obvious, and this may account in some measure for the term not having come into general uso. While still maintaining the inappropriateness of applying to a free secretion, which when once solidified has no longer any organic connection with the body, a term employed in the first instance to designate the locally thickened cuticle of the hinder part of the body in Polyzoa, I feel that no useful purpose is likely to be served by continuing the use of a term that fails to meet with general approval. 58 “TERRA NOVA” EXPEDITION. and 73, and referred by Andersson to tbe species* C. dodecalophus, is stated by that author to be more massive and provided with longer spines than the type-material of C. dodecalophus figured by MTntosh (14, plate 1, and plate 7, fig. 1), and in these respects his material approaches C. inaequatus (2, p. 9). The single piece of C. dode- calophus obtained from Station 58 is more slender and with thinner spines, but it is very different in appearance from the typical material obtained by the “ Challenger." Andersson himself was apparently somewhat doubtful about the identification (2, p. 9). As regards C. inaequatus, arguments have been adduced (11, pp. 559-563) to show that this is not distinguishable from C. hodgsoni, and should be regarded as a synonym of it. Andersson, on the other hand, while of opinion that C. aequatus approaches C. hodgsoni (2, p. 8), says nothing of the close relationship that exists between C. inaequatus and C. hodgsoni. The difficulty of instituting any comparison between the coenoecia of the species under consideration from an examination of the published figures is increased by the fact that the illustrations look so unlike the real objects. The figures of C. hodgsoni (17, plate 2, fig. 1, A and B) are reproduced from photographs, but they give a very imperfect suggestion of the appearance of the coenoecium, the effect of “ depth ” and “ distance ” in the photograph having been lost in reproduction by the “ half-tone ” process. The figure of C. inaequatus given by Andersson (2, plate 2, fig. l) is reproduced by a different process, apparently a lithographic process, and the background is white instead of black. This figure and those of the “Discovery” material of C. hodgsoni represent the coenoecium as of natural size, but owing to the different methods of reproduction a fair comparison cannot be made. Fortunately some small pieces of the “Antarctic” material of C. inaequatus from Stations 5 and 94 of the Swedish Expedition were sent to the British Museum from Stockholm, and it was thus possible to effect a direct comparison with the “ Discovery ” material of C. hodgsoni. C. aequatus Andersson admits to be very near C. hodgsoni (2, p. 8), but he distinguishes it largely by the fact that C. hodgsoni has hermaphrodite individuals as well as male and female, a not very substantial difference, considering that the whole of the material of C. aequatus was obtained on the same date and from the same locality. Further, it may be pointed out, hermaphrodite individuals in C. hodgsoni are not very common ; it may very possibly be that, in the material of C. aequatus collected, hermaphrodites were present, but escaped detection. Again, C. aequatus is stated by Andersson (2, p. 9) to approach so nearly to C. dodecalophus that he long hesitated before deciding that it should constitute a distinct species, and the decision turned mainly on the fact that up to the present no male zooids of C. dodecalophus have been found, whereas in the material of C. aequatus male and female zooids occur in approximately equal numbers in the same coenoecium (2, p. 9). As an argument logically applied, this would mean that anyone so fortunate as to dredge a colony of C. dodecalophus containing male zooids wmuld have to admit that, since it contained male zooids, it could not be C. dodecalophus. CEPHALODISCUS — RIDEWOOD. 59 The possibility of C. aequatus being the same as 0. dodecalophus Andersson considers to be reduced by the fact that the station from which the former was obtained lies in the antarctic region, whereas all the known material of C. dodecalophus has been dredged from sub-antarctic localities, with distinctly higher bottom-temperature than that of Station 94 of the Swedish South-Polar Expedition. The coenoecium of C. aequatus is stated by Andersson (2, p. 10) to agree in the main with that of C. inaequatus, but is less strongly developed — the coenoecium of C. inaequatus is “ kraftiger entwickelt.' C. aequatus is not figured by Andersson, and a piece of a colony from Station 94 of the Swedish Expedition, received by the British Museum from Stockholm, is taken from a part too near the base of the colony to give one a clear idea as to what the branching of the colony is like. So far as one can judge from this piece, however, C. aequatus approaches Form B of C. hodgsoni rather than Form A. It has thicker and straighter spines, with less netting by cross-bars than is common in Form A ; but while, on the whole, Form B is of larger build than Form A, C. aequatus is smaller, with the inhabited cavities of the coenoecium narrower. The zooids of C. aequatus measure about 2 • 0 mm. from the free ends of the arms to the end of the body, which is roughly the size of the zooids of Form B, whereas those of Form A are, on an average of a large number of cases, slightly larger. If it should subsequently transpire that there is no very strong reason for separating C. aequatus from C. inaequatus , and consequently from C. hodgsoni, it might be regarded as a diminutive example of Form B. The conclusion would be supported by the general purplish red colour of the zooids in both, and by the number of arms being in a majority of the zooids twelve, in both sexes, in C. aequatus and Form B of C. hodgsoni. Re-examination of the material of C. hodgsoni obtained on the “ Discovery ” Expedition goes to show that most of it, including the type-specimen (Specimen A ; 17, plate 2, fig. 1, A) is of Form B ; the inhabited cavity of the coenoecium has a tendency to dilate into funnels and gaping terminal ostia, and the spines are fairly coarse, with little looping up by cross-bars to form meshes and lateral ostia. But the larger piece figured (Specimen B; 17, plate 2, fig, 1, B) approaches Form A*, as also do Specimens C and F ; they show a certain amount of forking of the spines and netting by cross- bars, a profusion of small lateral ostia, and a smaller coenoecial cavity than is typical of Form B ; but none of the three is typically Form A, their terminal spines are rather long, and more resemble those of Form B in thickness and straightness. At the British Museum (Natural History) there are two pieces of C. inaequatus obtained on the Swedish South-Polar Expedition, one from Station 94 and the other from Station 5. That from Station 94 is clearly Form A, the other is doubtfully Form B. It is strange that in the latter material males are rare, whereas in that from Station 94 they are common (equal in number to the female zooids — 2, p. 85) ; this is VOL, IV, * But one cannot judge from the published figure. E GO “TERRA NOVA” EXPEDITION. just the reverse of what is found to occur in the “ Terra Nova : material, for in Form A almost all the zooids are female, whereas pieces of colony of Form B with male individuals are as common as pieces with female zooids. It may be that at a certain season of the year a colony changes sex, by the newly produced buds being of a different sex from their parent, although up till that time the buds had been of the same sex. A male zooid with a female bud is recorded in the “ Discovery ” Expedition report (17, p. 58). EMPTY COENOECIA OF CEPHALODISCUS. A large proportion of the material of C. liodgsoni collected is “dead,” consisting of coenoecium only, with no zooids. The “dead” pieces of colony are somewhat more brittle than those in which zooids occur, and the coenoecial substance is not bright and clear, but usually rather turbid and feebly opalescent. The coenoecial substance of Cephalodiscus, although shown in Andersson’s report (2, p. 20) to be albuminoid in character, seems to be remarkably resistant, and probably remains more or less unaltered at the bottom of the sea for years after it has ceased to be inhabited by zooids. In order to test the imperishable character of the coenoecial substance, a piece of colony of C. liodgsoni , Form A, free from zooids, was soaked in distilled water for two years exposed to the air, but lightly covered to exclude the dust. At the end of this period no difference was to be seen between the piece thus treated and a control portion of the same colony kept in alcohol. For the purpose of this experiment alcohol-fixed material was chosen in preference to that preserved in formalin solution, for it is well known that formalin has upon certain organic substances, such as gelatine, the effect of converting them into other substances that are resistant to the organisms of putrefaction.* It is unfortunate that in Andersson’s report (2, pp. 19-20) it is not stated what was the method of preservation of the material upon which Prof. Morner conducted his investigations. In the case of future collecting expeditions it might be well for the biologist in charge to prepare some coenoecial substance with a special view' to chemical investigation. For this purpose the various species of Demiothecia are better adapted * To verify this statement, a sheet of gelatine was soaked in water for twenty minutes to soften it, and then divided into two parts. One part was left in water. On the third day putrefaction was well advanced : the gelatine had swollen considerably more than at first, and the water emitted an extremely foul odour. On the ninth day there was no longer any offensive smell ; the gelatine had practically disappeared, and all that remained in the water was a light flocculent sediment ; at the end of three weeks the sediment had gone, and all that could be seen was a colourless slime, probably a growth of Leptothrix or Cladotlirix. The other part of the sheet of softened gelatine was transferred to a 4 per cent, solution of formalin for three hours, and then given several changes of water for two days to remove the last trace of free formalin. At the end of three weeks no change had occurred ; the appearance of the gelatine was the same as before, and there was no foul smell or other evidence of putrefaction. At the end of the fourth week, however, there were signs of disintegration, and at the end of the fifth week the gelatine had disappeared. The water was a little turbid, and the smell, though slightly unpleasant, was not offensive. CEPH ALO D ISCTTS— RID E W 00 D . 61 than those of Idiothecia and Orthoecus, because of the ease with which the coenoecial substance can be freed from the remains of the zooids. The preparation of the material might be conducted as follows : — the branches, after soaking in several changes of distilled water to remove the salt and the sand-grains, are cut up into small pieces to ensure that all portions of zooids are removed ; after drying quickly at a temperature of about 50° C., the pieces are put into a dry bottle or tin, and stored until, on the return of the expedition, the substance can be handed over to a chemist for analysis. It does not necessarily follow that because no zooids are found in it a coenoecium was unoccupied or “ dead ” at the time of dredging. In the largest colony that was obtained — that mentioned first in the list of material on page 49 — there are very few zooids ; but even if there were none at all, one would not gain the impression that the colony was “ dead," for the extremities of the branches are very pale and delicate, and have the appearance of being but recently secreted. The scarcity of zooids in this specimen, and the absence of zooids in some similar but smaller pieces of colony in the collection, are rather to be accounted for by the ease with which the zooids of the various species of Demiothecici may become dislodged if the specimen remains for any length of time in the trawl. ZOOIDS OF C. II 0D GS ONI. A fairly exhaustive examination of zooids from pieces of colony selected from the various bottles in which the collection arrived fails to justify any very definite generalisation regarding the relations subsisting between (a) the colour of the zooids, (/>) the number of arms of the zooids of Forms A and B, and (c) the two sexes of these Forms. In his diagnosis of C. inaequatus Andersson (2, p. 10) states that female individuals have five pairs of arms and males have six pairs. In the original account of C. hodgsoni (I 7, p. 55) it is observed that “ the normal number of plumes is twelve, but ... a full- sizecl polypide, . . . with well-developed ovaries, may have only ten fully-grown plumes.” The suggestion that there might? be some correlation between the number of arms and the sex of the zooid had not occurred to the author, whose paper appeared before that of Andersson. In the report on the Pterobranchia of the Scottish Antarctic Expedition (I 1, pp. 560-562) the question is discussed in connection with the synonymy of the two species just mentioned. In order to ascertain if in the material collected by the “ Terra Nova ” any relation could be found to exist between the number of the arms of a zooid and its sex in Forms A and B respectively, not less than 150 zooids were dissected, in the manner described on page 21. Of these zooids, 96 were of Form A, and 54 of Form B. One result of this examination is the discovery that very rarely are male and female zooids found in the same colony ; the only instance of associated sexes is in a piece of material of Form A from Station 339, with zooids of both sexes in mature K 2 62 “TERRA NOVA” EXPEDITION. condition. The only hermaphrodite zooids that were found, and the only male zooids of Form A, were from this colony. Another observation is that all the female colonies have zooids the majority of which are in an extremely immature condition, the gonads in some cases being so minute that the only means of deciding the sex was to stain them and examine them under the high power of the microscope. Of the pieces of Form B examined, one had zooids with minute ovaries, and arms varying from eleven to twelve ; another had zooids with testes, and arms from eleven to twelve ; another had zooids with minute ovaries, and twelve arms ; another had zooids with small ovaries, and arms from nine to eleveu ; another had zooids with ripe testes, and arms from ten to twelve ; another had zooids with extremely minute gonads, recognisable in most but not all cases as young ovaries, and arms from ten to twelve. In the piece of colony last mentioned (from Station 339) minute or arrested arms are unusually common ; they are reckoned in the counting as if they were full arms. Of the pieces of colony of Form A examined, that referred to in the fourth paragraph of this chapter had zooids with ripe ovaries and testes, and ten arms : another piece had zooids with gonads too minute to identify with certainty, but the indications were in favour of their being young ovaries, and the arms, though definitely ten in twenty-three zooids (in two of them there were eight large and two minute arms), were doubtfully nine in one case and certainly eleven in another. Another piece of colony was almost identical with the last ; fifteen zooids were examined, and in all of them the gonads were extremely minute, in two cases definitely recognisable as ovaries ; the arms were in all cases ten, counting dwarf arms as fully developed arms. Another piece was similar to the last in the zooids having ten arms each, but of the nine zooids examined two had fully developed ovaries, one had minute ovaries, and the other six had gonads too small to enable one to decide the sex. Another piece had zooids possessing ten arms each, and of the eight zooids examined one had large ovaries, five had small ovaries, and two had minute gonads of doubtful sex. Another piece had zooids with ten arms and extremely minute, doubtful gonads. Another piece had zooids with ten arms in thirteen cases, and doubtfully nine in two others, and gonads that were minute ovaries in all the zooids except four, in which they were too small to decide the sex. Another piece had zooids with ten arms in thirteen cases and doubtfully ten in two others, and minute gonads, recognisable as young ovaries in two cases only. Reviewing the results set out above, it is seen how difficult it is to establish any certain correlation between the three variables — (1) number of arms, (2) sex, (3) Forms A and B. Form A seems to be characterised by the possession of ten arms, even in the few males that were encountered ; but one zooid of the total of ninety-six investigated had undoubtedly eleven arms. In the case of Form B, pieces of colony with male zooids are as abundant as those with female zooids, and although in a CEPHAL0D1SCUS — RIDE WOOD, 63 proportion of cases the arms are eleven, ten, or even nine in number, yet there are many instances in which both male and female zooids have twelve arms each. These conclusions, however, are rather discounted by the fact that in reckoning the arms, arrested arms were searched for and were counted as complete arms, for arrested arms are more common in Form B than in Form A. If arrested arms were Text-fig. 12. — Cephalodiscus liodrjsoni, arms of adult zooids. A, a moderately contracted arm, in flat view. B, an arm without an end-swelling. C, a moderately extended arm, in side view. D, an extended arm, with small end-swelling. E, a contracted arm, with large end-swelling. F, a small but complete arm found among others of full size. G, a smaller arm, less complete. H, three arrested arms from one zooid. J, two arm-bases, presumably the remains of injured arms. K, a similar arm-base, with a few tentacles of full length. X 48. All are from Station 339, except D and E, which are from Station 348. Figures C-H are of Form B, figures A, B, J and K are of Form A. 64 TERRA NOVA” EXPEDITION. disregarded, and only functional arms counted, the numerical disproportion would he less striking. A selection of arrested arms and damaged but healed arms is shown in text-tig. 12, F-K, drawn to the same scale of magnification as a number of normal arms (A, C, D), and two other arms, one (E) with an exceptionally large end-swelling, and the other (B) without an end-swelling. On the whole, considering the difficulty experienced in drawing any decisive distinction between Forms A and B based upon their coenoecial characters (pp. 51-57), the generalisation that zooids of Form A have mostly ten arms in both sexes, and those of Form B mostly twelve arms in both sexes, does not seem to warrant the recognition of Forms A and B as distinct species of the sub-genus Demiothecia. Further, the investigation of the present material does not lend support to the thesis made by Andersson (2, p. 10) in respect of C. inaequatus that there is any regularity in the possession of ten arms by female and twelve arms by male zooids. It may be noted here that in the related species, C. aequatus, Andersson (2, p. 9) reports the number of arms as twelve in both sexes ; but it has already been explained (p. 59) that Form B cannot be regarded as the same as C. aequatus, since although the zooids agree fairly well in size and colour, the coenoecium of Form B is much larger than that of C. aequatus , the coenoecium of Form A being intermediate between the two. Even if Form B were considered as a giant form of C. aequatus, and pieces of colony of composite character (Form A and Form B combined) were regarded as the product of two species of zooids living harmoniously together, but not interbreeding, yet the pieces of colony of intermediate character, pieces that cannot be definitely relegated to the one Form or the other, still offer an insuperable difficulty. The colour of the zooids does not afford any real assistance in the matter, for while one can say in a general way that, in both material preserved in alcohol and that preserved in formalin, the zooids of Form B are of a purplish brown or a dull crimson, while those of Form A are usually of a dull earthy brown (raw umber), fading in time to greyish brown, oclireous or cream-colour, yet there are significant exceptions in some pieces of Form A with distinctly red zooids — e.y. that shown in plate 5, fig. 3, and the piece of colony from Station 339 with both male and female zooids ; in this last the red colour is not confined to the zooids of one sex. That differences in colour are not due to the particular fluid employed for preserva- tion is clear from the fact that in a bottle of material in which both Form A and Form B occur the zooids of the latter are distinctly more crimson than those of the former. Serial sections were cut of selected zooids of Forms A and B, but the study of them does not elicit any constant differences between the two Forms, nor any new features of general morphological interest. No differences are to be distinguished between the buds of Forms A and B. The buds of C. liodgsoni have previously been dealt with in a fair amount of detail CEPHALOmSCUS— RIDEWOOD. 6 5 (17, pp. 60-62, and 18, pp. 224-230), and there are no new observations to record from the present material. In material of Form A from Station 355 several zooids are found with a rounded or bluntly pointed anterior end destitute of shield, arms, oral lamella and mouth ; the stalk, on the other hand, is complete, with apparently healthy buds. Similar defective zooids have been previously noted in two species of Cephalodiscus besides C. hodgsoni (C. gracilis, 10, p. 93; C. dodecalophus, 14. plate 3, fig. 4; C. hodgsoni, 17, p. 53 and plate 6, fig. 52). Instances of the double budding of the same stalk, similar to those observed in material of C. nigrescens of the “Discovery” Expedition (17, plate 7, figs. 69 and 73), were found in material of C. hodgsoni from Station 348. Free ova occur in numerous instances, and are attached to the inner surface of the wall of the colony by a short stalk of coenoecial substance, which also spreads over the egg-shell. They measure usually about 0 • 7 by 0 * 5 mm. , but some were found in which the length is as great as 1 • 0 mm. There seems to be no difference in size between the free ova of Forms A and B. In the original description of C. hodgsoni the size is stated to be O' 47 by 0'42 mm. (17, p. 58), but free ova were found in one colony only (Specimen D) of the “ Discovery ” material, and the size is probably understated ; the ova may have been exceptionally shrunken. FOREIGN OVA FOUND ON THE COENOECIUM. In material of C. hodgsoni from Stations 314, 316, 338 and 355 are to be found, among the spines on the exterior of the coenoecium, solitary, oval, transparent eggs, each with a small red body within, and a quantity of thick fluid, that is clear and glairy in formalin, but coagulated in alcoholic material. The eggs measure 4 * 3 by 2‘8 mm., and thus cannot be mistaken for eggs of C. hodgsoni. The shell is smooth, tough, flexible and transparent, and each egg is smeared over with layers of coenoecial substance. These are clearly foreign eggs that have been laid by the parent upon the branches of the Cephalodiscus, and have subsequently been covered over with coe- noecial substance by the zooids, and they are only worthy of mention here from the fact that they are conspicuous objects, and cannot fail to attract the attention of anyone who at some future time may have occasion to examine this particular material. Although the eggs are fairly numerous, only two were found sufficiently advanced in their development to enable one to say definitely that they are eggs of a gastropod mollusc with a spiral shell. Mr. G. C. Robson, who has been good enough to examine the shells, identifies the mollusc as a kind of Valuta. 66 TERRA NOVA” EXPEDITION. SYNOPSIS OF THE SPECIES OF CEPHALODISCUS AT PRESENT KNOWN, AND LIST OF ALL RECORDED SPECIMENS. Material of the following species has not been examined by the present writer : — C. gracilis, C. sibogae, C. levinseni, C. anderssoni, C. indicus ; the particulars given are drawn from the works cited under the respective headings. Statements given here that do not occur in the diagnoses of the original authors have been introduced either as the result of a study of their published figures and text, or from an independent examination of material by the present writer. C. indicus was placed by its author (Schepotieff, 22, p. 435) in the sub-genus Idiothecia, but with sub-generic diagnoses as here adopted the species falls within the sub-genus Orthoecus — see p. 19. The geographical distribution of the various species of Cephalodiscus is illustrated by the two maps in plate 6. SUB-GENUS DEMI 0 THE CIA, Ridewood. Sub-genus Demiothecia, Ridewood. — Colony branching. Each ostium of the coenoecium leading into a cavity which is continuous through the colony, and is occupied in common by the zooicls and their buds. Transverse sections of the branches showing the central cavity surrounded by a wall of coenoecial substance, usually of irregular thickness, and sometimes with inwardly projecting bars and ridges. C. clodecalophus, MTntosh. — Colony irregularly branched and straggling ; some of the branches fusing up to form a network, the cross-bars being usually solid. Maximum size of known specimens — 215 mm. high, 195 mm. wide. Coenoecium pale orange-brown or colourless. Distance from one branch or cross-connection to the next about 22 mm. ; width of a branch, not including spines, 3 • 5 to 6 mm. Cavity of coenoecium continuous, but partially divided up by bars, ridges and incomplete parti- tions ; occupied in common by the zooids and their buds. Ostia numerous, irregularly placed, oval, 1’5 by O’ 7 mm., sessile — i.e. without peristomial tubes — sometimes even sunk below the general surface. Spines not obviously related to the ostia, simple, forked, or branched, length — 3 to 8 mm., rarely as much as 15 mm., width about 0 ’ 5 mm. Zooids— length from free ends of the arms to end of body of a fairly extended zooid — 2 ’ 0 mm. ; length from bases of arms to end of body— 1 * 3 mm. ; width of body — 0 ’7 to 0 * 8 mm. Colour (of preserved zooids) whitish or cream- coloured, but in 1882 the arms of the zooids of the “Challenger” material were pink or violet. Arms usually six pairs, sometimes five pairs ; axis of each arm with an end-swelling, with refractive beads. Testes — (males not known). Free ova O’ 33 mm. across, attached by a stalk. Buds one or two, rarely three, to each zooid. CEPHALODISCUS— RIDEWOOD. 67 Locality — Straits of Magellan ; Lat. 52' 45' 30" S. ; Long. 73° 46' 0" W. ; dredged on the “Challenger" Expedition. 1873-1876. Station 311 ; 245 fms. (448 m.) ; January 11th. 1876 ; several pieces. Described by AY. C. M‘ In tosh and S. F. Harmer, 14. For preliminary notices see footnote on page 13. Material referred to C. dodecalophus has since been obtained from three localities, on the Swedish South- Polar Expedition (“ Antarctic ”) : — Station 58. — Lat. 52' 29' S. : Long. 60J 36' W. ; between Falkland Islands and Burdwood Bank ; 108 fms. (197 m.) ; Sept. 11th, 1902; one piece was dredged, and a small portion of it was saved from the “ Antarctic ” before it sank. Station 59. — Lat. 53° 45' S. ; Long. 61° 10' W. ; between Falkland Islands and Burdwood Bank ; 75-82 fms. (137-150 m.) ; Sept. 12th, 1902 ; two pieces dredged, but only a small portion saved. Station 73. — Lat. 543 55' S. ; Long. 67° 41' W. ; east of southern end of Tierra del Fuego, Beagle Channel ; 44-128 fms. (80-235 m.) ; Oct. 30th, 1902 ; some pieces dredged, but none saved ; some embryos and larvae were saved. This material is described by K. A. Andersson, 2. The piece of material figured (2, plate 2, fig. 3) does not exhibit the characteristic features of the cocnoecium mentioned in the diagnosis given above ; if correctly referred to C. dodecalophus, it may be a basal portion of a young colony, or a particularly bushy piece from the end of a branch. The specimens mentioned above are referred to by Andersson in a preliminary note published in the Zool. Anzeiger of 1903 ( j ). He gives to the note the title “ Eine Wiederentdeckung von Cephalodiscus,” and makes no suggestion as to the possibility of the new material being of a different species from that which up till then was the only species known. Of the four localities mentioned two agree with Station 58 and Station 73 of the full report of 1907 (2) ; the other two are : — Off Cape Seymour ; Lat. 64° 28' S. ; Long. 56° 43' W. ; 82 fins. (150 m.); January 16th, 1902; and on the Burdwood Bank ; Lat. 53° 53' S. ; Long. 58° 47' W. ; 82 fms. (150 m.). These localities do not agree precisely with any of those mentioned on the first page of his full report, but they are presumably the same as Station 5 and Station 59, in which case the material obtained from the former is that to which he gave the name C. inaequatus, and the material from the latter, as also the material from Station 58 and Station 73, is that which he identified as C. dodecalophus. C. hodgsoni, Ridewood. — Colony irregularly branched, without cross-bars connecting up the branches. Maximum size of known specimens — 250 mm. high, 150 mm. wide (“ Terra Nova ” specimen). Width of a branch, not including spines, mostly 4 to 6 mm. , but some terminal branches as narrow as 3 mm. , and some basal stems as thick as 9 mm. Cavity of coenoecium irregular, continuous, and occupied in common by the zooids and their buds, inner surface smooth, cavity sometimes traversed by solid bars. J, VOL, JV. 68 “ TERRA NOVA” EXPEDITION. Older parts of the coenoecium of a deep amber colour, younger parts colourless. Ostia numerous, irregularly placed, oval, lateral or terminal ; terminal ostia rather funnel- shaped, about 3 by 2 mm., lateral Ostia smaller, mostly sessile ; each ostium with from two to five long, more or less radiating spines, simple, forked or branched ; length of spines up to 20 or 25 mm., but the spines of rapidly growing terminal branches commonly shorter (8 to 12 mm.) ; spines frequently connected by cross-bars, forming an imperfect mesh- work ; width of spines — 0 * G to 1 ’ 0 mm. , in terminal branches sometimes not more than 0 ’ 4 mm. Zooids — length from free ends of the arms to end of body of a fairly extended zooid — 2 ’0 to 3 ’ 2 mm. ; length from bases of arms to end of body — 1 • 0 to 1 • 8 mm. ; width of body — 0 • 6 to 1 * 0 mm. Colour (of preserved zooids) crimson-brown or pale brown, fading to ochreous and cream-colour. On the assumption that C. inaequatus is synonymous with C. hodgsoni, colour of living zooids red or crimson, the body darker and more brown than the arms. Arms five or six pairs ; axis of each arm with an end-swelling, with refractive beads. Male and female zooids and hermaphrodites (with one ovary and one testis) occurring in the same colony, but males or females usually preponderating. Testes short. Free ova measuring O' 5 by 0‘4 mm., up to O’ 8 by O’G mm., attached by a stalk. Buds two, sometimes three or four, to each zooid. Localities — Ross Sea, east end of Barrier ; Lat. 78° 16' 14" S. ; Long. 1 97 41' 47" E. (= 162° 18' 13" W.) ; 100 fms. (183 m.) ; Jan. 29th, 1902; five pieces. Oft' the Barrier; Lat. 78° 20' 30" S. ; Long. 174° 13' 12" W. ; 300 fms. (549 m.) ; Jan. 27th, 1902 ; one piece. In McMurdo Sound, South of Ross Island ; Lat. 77° 50' 30" S. ; Long. 166° 44' 45" E. ; 130 fms. (238 m.); one piece May 18th, 1903, and one piece June 3rd, 1903 ; dredged on the National Antarctic Expedition (“ Discovery”). Described by W. Gf. Ridewood, 1 7. Material of C. hodgsoni was obtained in the Graham Region on the Swedish South- Polar Expedition of 1901-1903; it is described under the name C. inaequatus, but evidence has been brought forward (11, pp. 559-563) to show that it is not distinguishable from C. hodgsoni. For the localities of this material see C. inaequatus [infra). Material of C. hodgsoni was obtained on the Australasian Antarctic Expedition of 1911-1914. From Station 1, in Commonwealth Bay, Adelie Land ; Lat. 66° 50' S. ; Long. 142° 6' E. ; 354 fms. (647 m.) ; Dec. 22nd, 1913 ; several pieces. From Station 8, off Queen Mary Land; Lat. 66° 8' S. ; Long. 94° 17' E. ; 120 fms. (219 m.) ; Jan. 27th, 1914; several pieces. From Station 12, off Shackleton Glacier, Queen Mary Land ; Lat. 64° 32' S. ; Long. 97° 20' E. ; 110 fms. (201 m.) ; Jan 31st, 1914; several pieces. Described by W. G. Ridewood. 21. CEPHALODISCUS — RIDE WOOD. 69 Material of C. hodgsoni was obtained on the British Antarctic (“Terra Nova”) Expedition of 1910-1913, from McMurdo Sound, Boss Sea, Stations 314, 316, 338, 339, 340, 348 and 355, at depths between 140 fms. (256 m.) and 300 fins. (549 m.), on dates between Jan. 23rd, 1911, and Jan. 20th, 1913 ; many pieces. Described in the present report, pp. 48-65. C. inaequatus, Andersson. — The species is here regarded as a synonym of C. hodgsoni (q.v.) The evidence for this view is given in I 1 , pp. 559-563. Localities — Graham Begion, S.E. of Seymour Island ; Lat. 64° 20' S. ; Long. 56° 38' W. ; 82 fms. (150 m.) ; Jan. 16th, 1902 ; and North of Joinville Island; Lat. 62° 55' S. ; Long. 55° 57' W. ; 57 fms. (104 m.) ; Dec. 21st, 1902 ; dredged on the Swedish South-Polar Expedition (“Antarctic”), Station 5 and Station 94; several pieces. Described by Iv. A. Andersson, 2. C. aequatus, Andersson. — Nearly related to C. hodgsoni and to C. dodecaloplius. Coenoecium closely resembling that of the latter species, less robust than that of C. inaequatus (= C. hodgsoni, vide supra). Largest piece— 120 mm. high; coenoecium containing diatoms in great numbers. Zooids— length from free ends of the arms to end of body of a fairly extended zooid — 2 • 0 mm. ; length from bases of arms to end of body — 1 • 1 mm. ; width of body — 0 • 7 mm. Colour of zooids preserved in alcohol crimson-brown, fading to pale brown. Arms six pairs in both sexes ; axis of each arm with an end-swelling, with refractive beads. Male and female zooids occurring in the same colony, in approximately equal numbers ; hermaphrodites (with one ovary and one testis) not known. Testes short. Free ova measuring 0*4 by 0‘35 mm., attached by a stalk. Buds two to four on each zooid. Locality — Graham Begion, North of Joinville Island ; Lat. 62° 55' S. ; Long. 55° 57' W. ; dredged on the Swedish South-Polar Expedition (“ Antarctic ”), Station 94 ; 57 fms. (104 m.) ; Dec. 21st, 1902 ; a few pieces. Described by K. A. Andersson, 2. C. gracilis, Manner. — Colony very small and delicate, irregularly branched; the single specimen known of prostrate habit, supported by calcareous foreign objects (branches of the Polyzoan Tnhucellaria). Width of branches — 1'0 to 1 • 5 mm. Coenoecium almost colourless, the orange-coloured zooids within imparting a yellow tint to the colony. Cavity of coenoecium continuous, occupied in common by the zooids and their buds. Ostia at the ends of the branches, funnel-shaped, 0 • 8 mm. across ; margin produced into three, four or five spines ; length of spines on the average — 2 mm., rarely as much as 4 mm. ; width of spines — O' 16 mm. Zooids — length from free ends of the arms to end of body of a fairly extended zooid — 1 • 3 mm. ; length from bases of arms to end of body — 0 1 8 mm. ; width of body — 0 • 4 mm. Colour (of preserved L 2 70 TERRA NOVA” EXPEDITION. zooids) orange, stalk with black pigment in the median line of the anterior surface. Arms five pairs ; small end-swellings, with refractive beads, on the first pair of arms of the bud, and occasionally on second and third pairs also, sometimes persisting in the adult. Testes — (male zooids not known). Free ova 0 • 3 to O’ 4 mm. in length, not stalked. Buds in great profusion on the zooids. Locality— East coast of Borneo ; Lat. 0° 34 *6' N. ; Long. 119° 8 b' E. ; obtained on the “ Siboga ” Expedition, Station 89 ; on coral-reef between tide-marks : June 21st, 1899 ; one piece. Described by S. F. Harmer, 10. C. sibogae, Harmer. — Colony very small and delicate, orange-coloured (in alcohol), irregularly branched ; size of the single specimen known — 21 mm. high, 34 mm. wide, with a dense basal encrusting part of the coenoecium attached to a stone ; coenoecial substance with numerous foreign inclusions. Width of a branch — 1‘0 to 1 • 5 mm. Cavity of coenoecium continuous, occupied in common by the zooids and their buds. Ostia at the ends of very short side-branches, more or less alternate on opposite sides of the branch, funnel-shaped, 0 - 5 mm. across; margin produced into two or three spines ; length of spines — O’ 8 to 1 - 5 mm.; width — 0‘ 12 mm. Zooids — females not known. Free ova not knowrn. Neuter zooids — length from free ends of the arms to end of body — 1 • 3 mm. ; length from bases of arms to end of body — 0 ' 8 mm. ; width of body— -0 • 3 mm. Colour blackish. Arms four pairs, no end-swellings with refractive beads. Buds in great profusion. Male zooids with conical body, passing continuously into the stalk ; alimentary canal vestigial ; testes long, occupying most of the trunk ; arms one pair, without tentacles, surface with crowded refractive beads, at least in young males ; colour blackish. Zooids having the same general structure as the neuters sometimes found with functional testes. Locality — Northern entrance of Buton Strait, oft’ S.E. of Celebes ; Lat. 4° 20' S. ; Long. 122° 58' E. ; dredged on the “Siboga” Expedition, Station 204; 41-51 fms. (75-94 m.) ; Sept. 20th, 1899 ; one piece. Described by S. F. Harmer, 10. SUB-GENUS IDI 0 THE CIA , Ridewood. Sub-genus Idiothecia, Ridewood. — Colony branching. Each ostium of the coenoecium leading into a tube which is occupied by one zooid and its buds. The tubes embedded in common coenoecial substance, and disposed at a more or less constant angle to the surface ; either blind at their inner ends, or connected up in the middle of the branch. C. agglutinans, Harmer and Ridewood. — Colony massive, branching ; largest pieces known measuring 100-115 by 45-55 mm. ; width of a main stem — 20 to 45 mm. ; width of a single branch — 10 to 25 mm. Coenoecium friable, cream-coloured, CEPHALODISCUS — RIDE WOOD. 71 speckled, opaque, containing large quantities of shells of Foraminifera, fragments of shells of Mollusca and spines ofEchinoids, and rounded particles of slate. No peristomial tubes ; each ostium with a single thick lip or spine, that projects about 3 mm. beyond the general surface. Average distance from one ostium to that nearest it — 3 or 4 mm., but less at the free extremity of the branch. Each ostium leading into a tube that communicates at its inner end with a tubular labyrinth in the middle part of the branch. Width of cavity of the tubes — l’O mm. Zooids — length from free ends of the arms to end of boclv of a fairlv extended zooid — 4 • 5 mm. : length from bases of arms to end of body — 3 * 2 mm. ; width of body — 0 ■ 8 mm. Colour (of preserved zooids) blackish. Arms eight or nine pairs, no end-swellings with refractive beads. Testes long. Females and hermaphrodites not known. Free ova not known. Buds on each zooid many, up to seven pairs ; late buds, with more than five pairs of arms, not known. Locality — Burdwood Bank. South of Falkland Islands ; Lat. 54° 25' S. ; Long. 57c 32' W. ; dredged by the “ Scotia " on the Scottish National Antarctic Expedition of 1902-1904, Station 346 ; 56 fms. (102 m.); Dec. 1st, 1903 ; several pieces. Described by S. F. Harmer and W. G. Ridewood, 1 1. C. gilchrifiti, Ridewood.— Colony of fairly long branches, some of the branches connected across by solid bars ; maximum size of known specimens — 190 mm. by 110 mm. ; distance from one branch to the next — 30 mm. on an average, maximum 55 mm., minimum 17 mm. ; width of terminal branches, not including spines, about 5 to 8 mm. ; width of basal stems about 10 mm. Colour brown or pale brown. Ostia numerous, except on the main stems ; mostly sessile, but those on terminal branches with short, rarely long, peristomial tubes. Each ostium leading into a tube that ends blindly in the middle part of the branch. Width of cavity of the tubes — l’O to 1 ’ 2 mm. ; length — 3 to 5 mm. Spine-like processes of the coenoecium numerous, slender, simple, occasionally forked ; on an average one spine in relation with each tube, alongside which it can be traced for some distance into the interior of the branch ; length of the freely projecting portion of each spine about 10 mm.; width — 0*3 to 0’4 mm. Zooids — length from free ends of the arms to end of body of a fairly extended zooid — 1 ’6 to 1 ’8 mm. ; length from bases of arms to end of body — 1 ’0 to 1 ’ 2 mm. ; width of body — 0 ’7 to 0 ’ 8 mm. Colour when alive blackish, with colourless arms ; colour of preserved zooids brown, with a broad blackish margin to the anterior part of the shield. Arms usually six pairs ; no end-swellings with refractive beads. Male and female zooids occurring in the same colony ; hermaphrodites not known. Testes globular or pear-shaped. Free ova about 0’4 mm. across, sometimes attached by a short stalk. Buds two to five on each zooid. Localities — Off Cape St. Blaize, Knysna Heads, Cape St. Francis, East London, and other localities on the South coast of Cape Colony, South Africa, within a range of Lat. 33°-35° S.; Long. 22°-28° E. ; dredged by a Cape Government trawler under 72 “TERRA NOVA” EXPEDITION. Dr. J. D. F. Gilchrist, from depths varying from 29 to 130 fms. (53 to 238 m.) ; during the years 1900-1904 ; many pieces. Described by W. G. Ridewood, 1 6. Material of C. gilchristi has since been obtained, mainly from Mossel Bay ; see J. Stuart Thomson, “Nature,” Nov. 5th, 1914, p. 273 ; and J. D. F. Gilchrist, 3 and 4. C. levinseni, Harmer. — Colony reddish orange, slightly branched, size of the single specimen known — 132 mm. high; width of a branch, not including peristomial tubes — 3 to 5 mm. ; width including peristomes about 12 mm. Ostia numerous, each leading into a tube that ends blindly in the middle part of the branch. Peristomial tubes projecting beyond the general surface about 4'0 mm.; abaxial edge of the peristome ending in a blunt lip which is continued as a ridge or “ midrib ” down the outer (abaxial) surface of the peristome. Length of the tubes, including peristomial part, about 6*0 mm. ; width of cavity — 0*Gto O’ 7 mm. Zooids — length from free ends of the arms to end of body of a fairly extended zooid — 2 • 5 mm. ; length from bases of arms to end of body — 1 '5 mm. ; width of body — 0*4 to 0*5 mm. Colour* (of preserved zooids) whitish. Arms six pairs ; no end-swellings with refractive beads. Testes — (male zooids not known). Free ova measuring 0*3 to 0 * 4 mm. in length, not stalked. Buds few on each zooid. Locality — Off the West coast of Japan, at the South end of the Corea Strait ; Lat. 32° 10' N. ; Long. 128° 20' E. ; 100 fms. (183 m.) ; sent to the Copenhagen Museum ; one piece. Described by S. F. Harmer, 1 0. C. evansi, n. sp. — Colony massive in build, branching ; largest piece known a single branch, 47 mm. high, 19 mm. wide, not including peristomes. Coenoecium friable, cream-coloured, speckled, containing large quantities of shells of Foraminifera and fragments of shells of Molluscs and Echinoids. Ostia at the ends of tubular peristomes that project 2*5 to 4*5 mm. beyond the surface of the branch; ostium set obliquely, sometimes squarely, at the end of the peristome, without definite lip or spine. Each ostium leading into a tube that ends blindly in the middle part of the branch. Width of cavity of the tubes— 0*6 to 0*8 mm.; length of tube, including the peristomial part, 10 to 25 mm., but shorter than 10 mm. at the apex of the branch. Zooids — length from free ends of the arms to end of body of a fairly extended zooid — 3 * 5 mm. ; length from bases of arms to end of body — 2 * 3 mm. ; width of body— 0 * 7 mm. Colour (of preserved zooids) pale green or white, with the red line of the shield very conspicuous. Arms usually eight pairs, sometimes nine or seven pairs ; no end-swellings with refractive beads. Male and female zooids occurring in the same colony ; hermaphrodite zooids (with one ovary and one testis) not known. Testes long. Free ova not known. Buds several to each zooid, up to nine. CEPHALODISCUS — RIDE WOOD. 73 Locality — New Zealand, from summit of Great King, Three Kings Islands, S. 14° W., 8 miles; Lat. 34° 15' S. ; Long. 172° 4' E. ; dredged on the British Antarctic (“Terra Nova”) Expedition, Station 90; 100 fms. (183 m.); July 25th, 1911 ; two pieces. Described in the present report, pp. 26-31. C. nigrescens, Lankester. — Colony massive, branching ; maximum size of known specimens — 190 by 115 mm., with twelve branches ; maximum width of a single branch — 32 mm., minimum width of a branch— 5 to 9 mm. ; branches roughly cylindrical, stout branches with rounded or tapering apices ; colour greyish, translucent. Ostia at the ends of short, tubular peristomes, but in main stems of the colony frequently sessile ; abaxial edge of the peristome produced into a blunt lip, but with no prominent ridge continued downward from the lip. No spines. Each ostium leading into a tube that ends blindly in the middle part of the branch. Width of cavity of the tubes — 1 • 2 to 1 ‘ 3 mm. ; length — 12 to 17 mm. on an average, but longer (20 to 26 mm.) in thick stems, and very short (4 to 5 mm.) at the tips of the branches ; the deeper parts of the longest tubes shut off by curved septa so that the habitable part of the tube is reduced (commonly 8 to 14 mm. in length). Zooids — length from free ends of the arms to end of body of a fairly extended zooid — 4 to 6 mm. ; length from bases of arms to end of body — 2 ’5 to 3 • 5 mm. ; width of body — about 0 * 9 mm. Colour (of preserved zooids) blackish. Arms usually seven pairs, sometimes six or eight pairs ; in well-preserved material fixed in formalin solution or in alcohol the tentacles pale and the axis rather broad, with two longitudinal black bands, but in material fixed in picric acid solution, and in badly fixed alcoholic material, the tentacles and axis of a uniform dark brown or black colour ; no end-swellings with refractive beads. Male and female zooids and hermaphrodites (with one ovary and one testis) occurring in the same colony. Testes long. Free ova— O’ 6 by O’ 5 mm. up to O’ 9 by 0*7 mm., not stalked. Buds from two to nine on each zooid. Locality — N.E. coast of Coulman Island, Victoria Land; Lat. 73° 18' S. ; Long. 170° 0' E. ; dredged by the “ Discovery ” on the National Antarctic Expedition ; 100 fms. (183 m.) ; Jan. 13th, 1902 ; several pieces. Described by E. Bay Lankester, 12 ; and by W. G. Ridewood, 17. Material of C. nigrescens had previously been obtained, but it remained in the British Museum unidentified until it was described by Ridewood (19) in 1912. This material consists of three pieces, which there is reason to believe were dredged on the “Erebus” and “Terror” Antarctic Expedition, in Ross Sea, within 3° longitude of Coulman Island, from a depth of about 300 fathoms (549 m.), either in January 1841 or February 1842. Material of C. nigrescens was obtained on the Second French Antarctic (“ Pour- quoi Pas?”) Expedition, from South of Jenny Island, near Adelaide Island, Graham 74 “TERRA NOVA” EXPEDITION. Region ; Lat. 68° S. ; Long. 70° 20' W. Paris ((58° W. Greenwich) ; 137 fms. (250 m.) ; Jan. 21st, 1909 ; many pieces, the largest measuring 120 to 150 mm. in height. Described by C. Gravier, 8. Material of C. nigrescent was obtained on the Australasian Antarctic Expedition of 1911-1914, from Station 8, off Queen Mary Land ; Lat, 66° 8' S. ; Long. 94° 17' E. ; 120 fms. (219 m.) ; Jan. 27th, 1914 ; one piece. Described by W. G. Ride wood, 21. Material of C. nigrescent was obtained on the British Antarctic (“ Terra Nova ”) Expedition of 1910-1913, from McMurdo Sound, Ross Sea. Stations 314, 316, 348 and 355, at depths between 190 fms. (348 m.) and 300 fms. (549 in.), on dates between Jan. 23rd, 1911, and Jan. 20th, 1913 ; many pieces. Described in the present report, pp. 31-37. SUB-GENUS OR THOE C US, Andersson. Sub-genus Orthoecus, Andersson. — Colony not branching, but in the form of a cake, or cone, or mass of irregular shape. Each ostium of the coenoecium leading into a tube which is occupied by one zooid and its buds. The tubes embedded in common coenoecial substance, either for their whole length or towards their blind ends only ; either closely set and parallel, more or less vertical, or irregularly bent and straggling. C. solidus , Andersson. — Colony not branching, in the form of a bulky mass or cake, measuring up to 250 or 300 mm. across, and 100 mm. high ; the mass con- sisting of closely set, vertically disposed tubes of uniform diameter, each with an ostium at the upper end, and terminating blindly below ; the tubes pale brown, united, except for a short distance at their upper ends, by common coenoecial substance which may contain sand-grains, fine sponge-spicules and diatoms. The common coenoecial substance moderately firm, like that of C. nigrescent, not soft and spongy like that of C. densus. Upper end of each tube produced on one side into a short, thick, upright lip. Tubes in the middle of the colony measuring up to 100 mm., those at the sides shorter; width of cavity of tube — 1*2 mm.; near the ostium — 1*4 mm. ; in narrowest parts — l'O mm. Zooids— length from free ends of the arms to end of body of a fairly extended zooid — 4 to 5 mm. ; length from bases of arms to end of body — 3*0 to 3*5 mm.; width of body— -1*0 mm. Colour (of preserved zooids) blackish ; but colour of zooids removed from their tubes and preserved in 70 per cent, alcohol, not kept in the dark, fading to raw umber or yellow ochre. Arms usually eight pairs, axes not known to possess the two black bands seen in arms of C. nigrescent ; no end-swellings with refractive beads. Male and female zooids and hermaphrodites (with one ovary and one testis) occurring in the same colony. CEPHALODISCUS— RIDEWOOD. 75 Testes long. Free ova measuring O’ 9 by 0'7 mm., not stalked. Buds many to each zooid, lip to fourteen. Locality — Graham Region, North of Joinville Island ; Lat. 62° 55' S. ; Long. 55° 57' AY. ; dredged on the Swedish South-Polar Expedition (“Antarctic”), Station 94 ; 57 fms. (104 m.); Dec. 21st, 1902; several large pieces were dredged, but only two fragments were saved from the “ Antarctic ” before it sank. Described by K. A. Andersson, 2. Material of C. solidus was obtained on the Australasian Antarctic Expedition of 1911-1914, from Station 2; off Mertz Glacier, Adelie Land; Lat. 66° 55' S. ; Long. 145° 21' E. ; 318 fms. (582 m.) ; Dec. 28th, 1913 ; five pieces. Described by W. G. Ridewood, 21. C. densns, Andersson. — Colony not branching, in the form of a bulky mass or cake, measuring up to 70 mm. high and 95 mm. across (Swedish Expedition material), and 100 mm. high and 130 mm. across (“Terra Nova” Expedition material); the mass consisting of closely set, vertically disposed tubes of uniform diameter, each with an ostium at the upper end, and terminating blindly below ; the tubes sometimes straggling, not closely set, nor straight. The tubes united by common coenoecial substance for a part of their length, particularly near their bases. The common coenoecial substance very soft and spongy, and liable to disintegration, unlike the more gelatinous substance of C. solidus and C. nigrescens. Ostium set transversely, rarely obliquely, to the end of the tube, the margin without lip, uniform, thin, sometimes with irregular external flanges. Tubes colourless, or pale brown ; sometimes encrusted with sand-grains or diatoms ; measuring up to 60 or 70 mm., perhaps more, in the middle of the colony, towards the edges gradually shorter; width of cavity of tube — DO to 1’2 mm., slightly greater near the ostium ; external diameter of free part of a tube about 1 • 6 mm. Zooids — length from free ends of the arms to end of body of a fairly extended zooid — 4 to 7 mm. ; length from bases of arms to end of body — 2 to 4 mm. ; width of body — 0 • 8 to 1 • 0 mm. Colour (of preserved zooids) greyish white, ochreous, orange or brown. Arms usually eight pairs ; no end-swellings with refractive beads. Male and female zooids and hermaphrodites (with one ovary and one testis) occurring in the same colony. Testes long, elongate pyriform. Free ova measuring 0‘ 8 by 0-7 mm., not stalked. Buds many to each zooid, from five to fourteen. Locality — Graham Region, S.E. of Seymour Island ; Lat. 64° 20' S. ; Long. 56° 38' W. ; dredged on the Swedish South-Polar Expedition (“Antarctic”), Station 5; 82 fms. (150 m.) ; Jan. 16th, 1902 ; a few pieces. Described by K. A. Andersson, 2. The material described by Andersson (2) as C. rarus is by the present writer regarded as belonging to the species C. densus. For localities see C. rarus (infra). M VOL. IV. 76 TERRA NOVA” EXPEDITION. Material of C. densus was obtained on the Australasian Antarctic Expedition of 1911-1914, from Station 8 ; off Queen Mary Land ; Lat. 6G° 8' S. ; Long. 94c 17' E. ; 120 fms. (219 m.) ; Jan. 27th, 1914 ; one piece. Described by W. G. Ridewood, 2 1 . Material of C. densus was obtained on the British Antarctic (“Terra Nova”) Expedition of 1910-1913, from McMurdo Sound, Ross Sea, Stations 314, 316, 339, 355 and 356, at depths between 50 fms. (92 m.) and 300 fms. (549 m.) ; on dates between Jan. 23rd, 1911, and Jan. 22nd, 1913; several pieces; also from Ross Sea, E.S.E. of Coulman Island, Station 295 ; Lat. 73° 51' S. ; Long. 172° 57' E. ; 190 fms. (348 m.) ; Jan. 27th, 1913 ; several pieces. Described in the present report, pp. 37-48. C. rarus, Andersson. — Cephalodiscus rarus is regarded by the present writer as not specifically distinct from C. densus. See pp. 39-40 of this report. Localities — Graham Region, S.W. of Snow Hill Island ; Lat. 64° 36' S. ; Long. 57° 42' W. ; 68 fms. (125 m.) ; Jan. 20th, 1902 ; and North of Joinville Island ; Lat. 62° 55' S. ; Long. 55° 57' W. ; 57 fms. (104 m.) ; Dec. 21st, 1902 ; dredged on the Swedish South-Polar Expedition (“ Antarctic ”), Station 6 and Station 94 ; two pieces from the latter station, and a fragment from the former. Described by K. A. Andersson, 2. C. anderssoni, Gravier. — Cephalodiscus anderssoni is stated by Gravier (8, p. 85) to be closely allied to C. rarus, differing from it in the tidies being less independent and isolated. The present writer doubts whether C. anderssoni is specifically distinct from C. densus. Colony in the form of a ragged, bulky mass, measuring up to 105 mm. across and 125 mm. high, the mass consisting of not very closely set, radially disposed tubes of uniform diameter, each with an ostium at the upper or outer end, and terminating * blindly below or towards the middle of the colony ; the tubes comparatively free in their outer parts, in their basal parts cemented by soft common coenoecial substance in groups of four or five, sometimes more ; the bases of the groups all cemented into a central basal mass of irregular form. Ostium set transversely, sometimes obliquely, to the end of the tube ; the margin without definite lip, uniform, thin. The tubes partly encrusted with sand-grains ; measuring up to 70 mm. in the middle of the colony, those at the sides shorter ; external diameter of the free parts of the tubes — 1 '5 to 1 ’ 7 mm. ; width of cavity of tube about 1 * 1 mm. Zooids — length from free ends of the arms to end of body — 4 to 4 • 5 mm. Colour of zooids not stated ; stalk slightly pigmented, shield pigmented, except the posterior lobe. Arms in buds up to six pairs ; number in adults doubtful owing to poor state of preservation of the material ; no end-swellings. Female zooids and hermaphrodites (with one ovary and one testis) occurring in the same CEPHALODISCUS — RIDE WOOD. 77 colony ; males not found. Testes — length not stated. Free ova measuring 1 ■ 0 to 1 ‘ 1 mm. by 0‘55 to O' 6 mm., not attached by a stalk. Buds three or four, or as many as eight, to each zooid. Locality — South of Jenny Island, near Adelaide Island, Graham Region ; Lat. 68° S. ; Long. 70° 20' W. Paris (68° W. Greenwich); dredged on the Second French Antarctic Expedition (“ Pourquoi Pas ? ”) ; 137 fms. (250 m.) ; Jan. 21st, 1909 ; many pieces. Described by C. Gravier, 6 and 8. C. inclicus, Schepotieff. — Colony diminutive, in the form of a small round plate ; diameter — 7 to 10 mm., height — 3 to 4 mm. ; bright orange when fresh, but pale when preserved in alcohol. Ostia with slight marginal projections and folds, but without definite peristome or lip ; ostia eight to fifteen in number, all on the upper surface, each leading into a tube that is blind at its lower end. Width of cavity of tube — 0 ■ 7 mm. Zooids — length from free ends of the arms to end of body of a fairly extended zooid — 2 ‘ 2 mm. ; length from bases of arms to end of body — -1 • 1 mm. ; width of body — 0 ‘ 7 mm. Colour (of preserved zooids) pale, shield slightly pigmented. Arms three pairs, no end-swellings with refractive beads. Male and female zooids occurring in the same colony, hermaphrodites not known. Testes long. Free ova not stalked, size not stated. Buds several to each zooid. Localities — Bay of Belligemma (Weligama), on South coast of Ceylon ; Lat. 5° 35' N.; Long 80c 16' E. ; collected from coral-reefs in the coast-zone; 5g to 8 fms. (10 to 15m.); in the spring of 1908 ; four pieces. Also from Mahe, Malabar coast of India ; Lat. 11° 25' N. ; Long. 75° 21' E. ; collected on the great barrier-reef; 8 to 11 fms. (15 to 20 m.) ; in the spring of 1908 ; one piece. Described by A. Schepotieff, 22. SUMMARY. Of the four species of Ceplialodiscus collected on the expedition one is new to science, and to it is given the name C. evansi. See pp. 26-31. The material of Orthoecus collected goes to show that what has been described as C. ( 0 .) varus (Andersson, 2) is but a lax and straggling form of C. ( 0 .) densus. See pp. 39-40. In the abundant material of C. hodgsoni collected two forms may be distinguished, designated respectively Form A and Form B ; they merge insensibly the one into the other, and cannot be regarded as distinct species or even varieties. See pp. 51-57. A list is given of all recorded specimens of Ceplialodiscus. See pp. 66-77. m 2 78 “TERRA NOVA EXPEDITION. REFERENCES. 1. Andersson, K. A. — “ Eine Wiederentdeckung von Cejjhalod iscus.” Zool. Anzeiger. Vol. XXVI. no. 697, 1903, pp. 368-369. 2. Andersson, K. A. — “ Die Pterobranchier der Schwedischen Siidpolar -Expedition, 1901-1903/' Wiss. Ergebn. Schwedischen Siidpolar-Expedition, Vol. V, Stockholm, 1907, pp. 1-122, eight plates. 3. Gilchrist, J. D. F. — “ Observations on the Cape Cephalodiscus (C. gilchristi, Ridewood), and some of its Early Stages,” with an Appendix by S. F. Harmer. Ann. Mag. Nat. Hist., ser. 8, Vol. XVI, 1915, pp. 233-246, one plate. 4. Gilchrist, J. D. F. — “On the Development of the Cape Cephalodiscus ( C . gilchristi, Ridewood).” Quart. Journ. Micr. Sci., n.s., Vol. LXII, pt. 2, 1917, pp. 189-211, two plates. 5. Gravier, C. — “ Sur les Pterobranches rapportes par la seconde Expedition antarctique fram;aise.” Comptes Rendus Acad. Sci., Vol. CLIV, Seance du 28 Mai, 1912, pp. 1438-1440. 6. Gravier, C. — “ Sur une espece nouvelle de Cephalodiscus ( C . anderssoni, nov. sp.) provenant de la seconde Expedition antarctique fran5aise.” Bull. Mus. Hist, nat., Paris, 1912, Vol. XVIII, no. 3, pp. 146-150, two text-figures. 7. Gravier, C. — “ Sur la repartition geographique des especes actuellement connues du genre Cephalodiscus, MTntosh.” Bull. Mus. Hist, nat., Paris, 1912, Vol. XVIII, no. 3, pp. 151-153. 8. Gravier, C.— Deuxieme Expedition antarctique franchise, 1908-1910 ; Sci. nat., Documents scientifiques ; Pterobranches; Paris, 1913, pp. 71-86, five text-figures. 9. Harmer, S. F. — Appendix to Report on Cephalodiscus. “Challenger” Reports, Zool., Vol. XX, pt. LXII, 1887, pp. 39-47, four text-figures. 10. Harmer, S. F. — -“ The Pterobranchia of the Siboga Expedition, with an Account of other Species.” Resultats des Explorations entreprises aux Indes Neerlandaises Orientales en 1899-1900 a bord du “Siboga”; livr. XXII, monogr. XXVI bis ; Leiden, 1905, pp. 132, fourteen plates and two text-figures. 11. Harmer, S. F., and Ridewood, W. G. — “The Pterobranchia of the Scottish National Antarctic Expedition (1902-1904).” Trans. Roy. Soc. Edinb., Vol. XLIX, pt. 3, no. 7, 1913, pp. 531— 565, two plates and five text-figures. 12. Lankester, E. R. — “On a New Species of Cephalodiscus (C. nigrescens ) from the Antarctic Ocean.” Proc. Roy. Soc. London, 1905, Vol. LXXVIb, pp. 400-402, one plate. 13. MTntosh, W. C. — “ Preliminary Notice of Cephalodiscus, a new Type allied to Prof. Allman’s Bhabdopleura, dredged in H.M.S. ‘Challenger.’” Ann. Mag. Nat. Hist., ser. 5, Vol. X, 1882, pp. 337-348. 14. MTntosh, W. C. — “ Report on Cephalodiscus dodecalophus.” “ Challenger ” Reports, Zool., Vol. XX, pt. LXII, 1887, with Appendix by S. F. Harmer, pp. 48, seven plates and six text- figures. 15. Masterman, A. T.— “ On the further Anatomy and the Budding Processes of Cephalodiscus dodecalophus.” Trans. Roy. Soc. Edinb., Vol. XXXIX, pt. 3, Dec. 1898, pp. 507-527, five plates. 16. Ridewood, W. G. — “A new Species of Cephalodiscus ( C . gilchristi) from the Cape Seas.’ Marine Investigations, South Africa, Vol. IV, Cape Town, 1906, pp. 173-192, three plates and five text-figures. CEPHALODISCTJS — RIDEWOOD. 79 17. Ridewood, W. G. — ■“ Pterobranchia ; Cejohalodiscus.” National Antarctic Expedition [“Discovery”], Natural History, Vol. II, London (Brit. Mus.), 1907, pp. 1-67, seven plates and seventeen text-figures. 18. Ridewood, W. G. — “ On the Development of the Plumes in Buds of Cephalodiscus.” Quart. Journ. Micr. Sci., n.s., Vol. LI, 1907, pp. 221-252, eleven text-figures. 19. Ridewood, W. G. — “On Specimens of CepTialodiscus nigrescens supposed to have been dredged in 1841 or 1842.” Ann. Mag. Nat. Hist., ser. 8, Vol. X, 1912, pp. 550-555, one text-figure. 20. Ridewood, W. G. — See Harmer, S. F., 11. 21. Ridewood, W. G. — The Pterobranchia of the Australasian Antarctic Expedition of 1911-1914. (? Date. The manuscript was posted to Sydney on Sept. 7th, 1917. — W. G. R.) 22. Schepotieff, A. — •“ Die Pterobranchier des Indischen Ozeans.” Zool. Jahrb., Abth. f. Syst., Vo’. XXVIII, pt. 4, 1909, pp. 429-448, two plates. CEPHALODISCUS — RIDE WOOD. 81 INDEX. Arms, epidermis of, 22-23, 30, 33, 35-36. ,, in buds, 26, 31, 48. ,, number of, 17, 21, 29, 33, 44, 49, 62. „ reduced, or defective, 17, 44, 62-64. Budding, mode of, 15, 65. Buds, arms of, 26, 31, 48. „ C. densus, 48. „ C. evcinsi, 31. „ C. hodgsoni, 64-65. Dcmiothecia, 16, 19, 57, 66. Empty coenoecia, 60. End-swellings of arms, 18, 21-22, 44, 64. Foreign ova, 65. Forms A and B of C. hodgsoni, 49-57, 64. Gelatine affected by formalin, 60, footnote. Geographical distribution, 24, 66. Gigantism, 24. Historical introduction, 13. Idiothecia , 16, 19, 40, 70. Localities, 12. Material collected, 12. „ C. densus, 37. ,, C. evansi, 27. ,, C. hodgsoni, 48-49. ,, C. nigrescens, 31. Nerve-tract of stalk, 24—25. Notochord, 25, 30, 48. Oral lamella, 30, 44. Orthoecus, 16, 19, 40, 74. Ova, foreign, 65. „ free, 18, 36, 47, 65. Ovaries, 18, 26, 36-37, 46. Pigmentation of arms, 33-36. „ „ body, 23. Preservation of material, 12, 34-36, 61. References, 78. Rhabdojjleura, 14—16. Shield, 21, 43. Size of zooids, 23-24. Spermatozoa, 30, 46. Stalk, length of, 24. ,, nerve-tract of, 24. Sub-genera of Cephalodiscus, 16. Summary of I’esults, 77. Synonyms, 14, footnote, 40. Synopsis of known species, 66-77. Testes, length of, 25, 44. Zooids, defective, 65. ,, dormant, 42. ,, general review, 16, 20-26. ,, habits, 19-20. „ of C. densus, 43. ,, ,, C. evansi, 29. ,, , , C. hodgsoni, 61. ,, ,, C. nigrescens, 32, 33. ,, size of, 23-24. Cephalodiscus, generic diagnosis, 14. C. aequatus , 58-59, 64, 69. ,, agglutinans, 20, 70. „ anderssoni, 40, 76. ,, densus, 37, 75. „ dodecalophus, 58, 66. „ evansi, 26, 72. „ gilchristi, 19, 71. „ gracilis, 57, 69. „ hodgsoni, 48, 58, 59, 67. „ inaeqicalus, 58, 59, 61, 64, 69. „ indicus, 66, 77. ,, levinseni, 72. „ nigrescens, 31, 73. „ rams, 39, 40, 76. ,, sibogae, 57, 70. ,, solidus, 40, 74 Coenoecium, characters of, 18, 27, 32, 39-40, 50, 51-57, 60. „ empty or “ dead,” 60. ,, inclusions in, 27-28, 41. ,, use of word, 57, footnote. LONDON : PRINTED BY WILLIAM CLOWES AND SONS, LIMITED, DUKE STREET, STAMFORD STREET, S.K. 1, AND GREAT WINDMILL STREET, W. 1. PLATE I. Ceph alod incus n i (jrescens . Fig. 1. — Ceplialodiscus nigrescens. Station 348. Nat. size. Fig. 2. — G. nigrescens. Station 316. Nat. size. Fig. 3. — C. nigrescens. Station 348. Nat. size. Photographs by Mr. C. Butterworth (Figs. 1 and 3) and Mr. H. G. Herring (Fig. 2). Brit. Antarctic (Terra Nova) Exped. 1910. Brit. Mus. (Nat. Hist.) Zoology, Vol IV. Cephalodiscu3,Pl I G .MV Woodward del . et lith.. " Hiithimp PLATE II. Cephalodiscus evansi .... Figs. 1, 2. „ nigrescens ...... 3-5. Figs. 1-2. — Cepltalodiscus evansl. Station 90. Nat. size. Fig. 3. — C. nigrescens , a piece of colony viewed against a black background. Station 316. Nat. size. Fig. 4. — The same piece viewed against a light background. Fig. 5 .— C. nigrescens, three slender branches. Station 314. Nat. size. Photographs by Mr. C. Butterworth (Figs. 1 , 2, 4 and 5) and Mr. H. G. Herring (Fig. 3). Brit . Mus . (Nat . Hist .) Brit . Ant ar otic ( ! err a N ova) ExpecL.1910. Zoology, VoL. IV. C ephalo dis cus , PI . II . G.M Woodward del et.lith. ffutli.Trap. PLATE III. Ce}tlialodiscus dcnsus. Figs. 1-5. — Ccjahalodiscus dcnsus, rather lax fragments of colony. Station 316. Fig. 6. — C. dcnsus, a complete, dense, cake-like colony. Station 356. Nat. size Photographs by Mr. C. Butterworth. Nat. size. Brit. Antarctic (Terra Nova) Exped. 1910. Brit. Mus. (Nat. Hist.) Zoology, Yol. IV. Cephalodiscus, PI. EE. G.M Woodward del. London Stereoscopic Co. imp Cephalodiscus, Plate IV. PLATE IY. Cephalodiscus hodgsoni, Forms A and B. Fig. 1. — Cephalodiscus hodgsoni, Form A. Station 348. Nat. size. Fig. 2 .—C. hodgsoni, Form B. Station 348. Nat. size. Fig. 3.—C. hodgsoni, Form B. Station 339. Nat. size. Photographs by Mr. C. Butterworth (Fig. 1) and Mr. J. H. Leonard (Figs. 2-3). Brit. Antarctic (Terra Nova) Exped. 1910 Brit. Mns. (Nat. Hist.) Zoology, Vol.IV Cephalodiscus, PI. IV. n < \/r Cephalodiscus hodgsoni, Form A . . . . Figs. 1-5. ,, densus 6—8. ,, evansi ...... Fig. 9. Fig. 1. — Cephalodiscus hodgsoni, Form A, a long, slender piece of colony. Station 355. Nat. size. Fig. 2. — C. hodgsoni , Form A, a dense, brambly piece of colony. Station 338. Nat. size. Fig. 3. — C. hodgsoni, Form A, a piece of colony without definite main axis. Station 348. Nat. size. Fig. 4. — C. hodgsoni, Form A, a similar, but apparently younger piece of colony. Station 348. Nat. size. Fig. 5. — C. hodgsoni, Form A, a dense piece of colony, differing from that shown in Fig. 2 in having an apparently newly-grown upper portion. Station 355. Nat. size. Fig. 6. — C. densus, a small portion of colony, showing external flanges to the tubes. Station 314. Nat. size. Figs. 7-8. — C. densus, two zooids extracted alive by Mr. D. G. Lillie, and specially killed in corrosive sublimate solution. Station 339. X 6. Fig. 9. — C. evansi, four tubes partially dissected out. Station 90. Approximately X 8. Photographs by Mr. C. Butterworth (Figs. 1-2), Mr. J. H. Leonard (Figs. 3-5), and Dr. W. G. Ridewood (Figs. 69) . -Brit. Mus. (Nat Hist.) Brit. Antarctic (Terra Nova) Exped. 1910. Zoology.Vol.IV. Cephalodiscus, PI V G.M. Woodward del PLATE VI. Maps showing the geographical distribution of Cephalodixcus ; for study in connection with the synopsis of the species at present known, pp. 66-77. Abbreviations. Austr. — Australasian Antarctic Expedition, 1911-1914 ; see Ridewood, 21. Chall.— “ Challenger ” Expedition, 1873-1876; see MTutosh, 14. Discov. — National Antarctic (“Discovery”) Expedition. 1901-1904 ; see Ridewood, 17. Er. & Terror. — “Erebus” and “Terror” Antarctic Expedition, 1839-1843; see Ridewood, 19. The exact locality from which Ceplialodiscus was obtained is doubtful ; in the map two alternatives are given. French.— Second French Antarctic Expedition, 1908-1910; see Gravier, 8. Scotia. — Scottish National Antarctic (“ Scotia”) Expedition, 1902-1904 ; see Harmer and Ridewood, 1 | Siboga. — “Siboga” Expedition in the Dutch East Indies, 1899-1900; see Harmer, 10. Swed. — Swedish South Polar Expedition, 1901-1903; see Andersson, 2. Terra N.— British Antarctic (“Terra Nova”) Expedition, 1910-1913 ; see Ridewood, the present report. Cephalodiscus.PI. VI Brit. Mus. (Nat. Hist.) Brit. Antarctic (Terra Nova) Exped. 1910 Zoology, Vol. IV. hodgsoni, Ausir.Stat. 12. hodgsoni, Austr Stat. 1. solidus, ■ Auotr. Stat. 2. Austr. j hodgsoni Stat 8 ) nigrescens \ densus gilchrlstl, see Rldewood, 18 (hodgsoni, Discou. & Terra N. ! T”'° " evansl, Terra N. Stat. 90. nigrescens, uiscov. densus, Terra N. Stat. 295. rarus (= densus) _ Stat. 6. inaeq. ( = hodg.) j aequatus < solidus ( 04 rarus (= densus) ' nigrescens, Er <4 Terror hodgsoni Discou. anderssoni, nigrescens, French. dodecalophus, Slued.' Stat. 73. George EbilipA SonLtd GFOGRAPHIOAI agglutlnans, Scotia, Stat. 346. dodecaloohus, Swed. Stat. 58, 59. ' niPTRI PI I T I n M nc nrnn * i nnirni m Ihe London Gecgraj.ihi.caJ Institute dodecalophus. Chall Stat. 311 iSTORY REPORT. Pp. 83-94. ERMA AND ENTEROPNEUSTA. ajjU Tec JKCClf, I/kil., li. 'Scaftflu Stir it-, Civ.”. pv1 S^nitfOy'W, 1; g5g Q0rtft60 Qttueeum (Qtafuraf J^teforg) “Terra Nova” Report. This is No. *3 of 2 5 copies of Zoology , Vol. IV, No. 3. — Echinodcrma {Part II), printed on special paper. S3 V ECHINODERMA (part ii.) AND ENTEROPNEUSTA. LARVAE OF ECHINODERMA AND ENTEROPNEUSTA.* BY E. W. MacBRIDE, D.Sc., LL.D., F.R.S., Professor of Zoology in the Imperial College of Science and Technology, London, S.W. WITH TWO PLATES. I. — Introduction ..... II. — Distribution ..... III. — Preservation and Methods of Preparation IV. — Description of the Species — ECHINODERMA : A. Holothurioidea B. Asteroidea .... ENTEROPNEUSTA TAGE 83 84 84 85 90 91 V .—Literature 94 I -INTRODUCTION. Through the kindness of my friend Dr. S. F. Harmer, F.R.S., Director of the Natural History Departments of the British Museum, I obtained the privilege of examining and reporting on the Echinoderm and Enteropneust larvae brought back by the “ Terra Nova” Expedition. A considerable number of specimens were collected ; but they all belong to four species, of which only one, a Bipinnaria, is new. Two species of Auricular ia are * Manuscript received July 24, 1918 [S. F. FI.]. VOL. IV. N 84 “TERRA NOVA” EXPEDITION. represented in the collection, and one of Tornaria. Of one of the species of Aurictda/ia, specimens in different stages of growth were obtained, and I was thereby enabled to settle some disputed points in the embryology of Holothurioidea which are of general interest. As Mr. Simpson and myself (8) had formerly described Echinoplutei and Ophioplutei from the same part of the world, all four types of Echinoderm larvae have now been recorded by me from this region of the Antarctic. II. — DISTRIBUTION. The specimens were captured at two groups of stations, the members of each group being close to one another, so that all the larvae are derived from two localities, of which only one was Antarctic. The other locality is in the vicinity of Three Kings Islands, a group of islets situated a short distance north of North Cape, New Zealand. In this second locality the Tornaria and one of the species of Auricularia [A. nudibrancliiata ) were obtained. The Antarctic locality comprises a group of stations situated between 69° 50' and 72° S., and between 1G6° and 168° 50' W. The area within which these stations lie is therefore roughly a rectangle of two-and-a-lialf decrees of longitude in width, and of two-and-a-half degrees of latitude in height. o o o o III— PRESERVATION AND METHODS OF PREPARATION. In all the bottles, except one, which were handed over to me, the larvae were contained in formalin solution. In one case the specimens had been placed in picric acid, but they were in fragments when the bottle came into my hands. All the larvae, therefore, which I was able to examine, had been preserved in formalin. This circumstance had a determining influence on the methods of preparation which I adopted. It is a well-known fact that formalin gives poor results so far as histological structure is concerned, but that it renders the tissues very transparent. Hence all the specimens which I examined were stained in bulk and mounted whole. Two stains were employed, both in strong alcoholic solution, viz. : eosin and light green. The latter, which, like eosin, is an aniline dye, gave far more satisfactory results than eosin because it produced a much more intense coloration. After being stained the specimens were transferred to absolute alcohol, to which a few drops of oil of cloves were added every day for a couple of weeks. Then the mixture of oil and alcohol was placed in an open vessel and allowed to evaporate, and, as a result, in a couple of days only pure oil of cloves remained. To this an equal amount of solution of Canada balsam in Xylol was added, and in this mixture the specimens were left for a day, after which they were ready to be transferred to pure balsam on the slides. It is to be noted that formalin readily becomes partially oxidized to formic acid, and so any calcareous structures which the specimens originally contained had been dissolved before they came into my hands. ECHIXODERMA, II— MacBRIDE. 85 IV -DESCRIPTION OF THE SPECIES. ECHINODERMA. A. HOLOTHURIOIDEA. 1. Auricularia antarctica, MacBride. Eigs. 1-3, 5, 6. Station 270. 69° 51' S., 166° 17' W., surface to 600 metres. Two large specimens. „ 272. 71° 35' S., 166’ 11' W., 80 metres. Several specimens. „ 271. 71° 29' S., 166’ W., 80 metres. Many specimens. ,, 275. ,, ,, 160 metres. Many specimens. „ 281. 71° 41’ S., 166’ 47' W., 80 metres. Many specimens. „ 283. 71’ 39' S., 166’ 47' W., 80 metres. Many specimens. „ 285. 71’ 49' S., 167’ 32' W., surface to 600 metres. Two large specimens. ,, 288. 71’ 59' S., 168° 43' W., 60 metres. Many specimens. This species was originally described by me from a single specimen (9). It was subsequently found by Mortensen amongst the larvae collected by the German South Polar Expedition ( 1 4). In the collection which is the subject of the present memoir it is represented by numerous specimens of all sizes, from about ' 8 mm. to 5 • 0 mm. in length. The diagnostic characters of A. antarctica as originally determined by me were three : viz. (1) The presence of numerous calcareous “ wheels ” ; (2) the great extension of the post-oral loop of the longitudinal ciliated band, which goes so far forward that it is separated from the prae-oral loop by a narrow groove in which the mouth lies ; (3) the great width of the larva from front to back (Fig. 2), as a result of which the anterior pole, i.e., the spot where the prae-oral loop leaves the main portion of the longitudinal ciliated band, is situated far behind the mouth instead of being directly anterior to it, as is the case with most Auricularia larvae. To these characters I can now add a fourth : viz., the intestine is produced into two well-developed pouches directed ventrally, between which the ventral curvature of the stomach is contained. These pouches (int. I .) are well seen in Figs. 3, 2, and 6. The characteristic calcareous “ wheels ” had been dissolved by the acidity of the preserving fluid, but the other two characteristics are very strongly marked, and quite sufficient to determine these larvae as belonging to A. antarctica. In many of the specimens little disc-shaped groups of cells can be made out [calc., Figs. 1 and 3) which represent the organic bases of the vanished “ wheels.” The exact mode of development of the coelom in Holothurioidea has been the subject of considerable differences of opinion. The older authors described its first beginning as a sac given off from the blind apex of the archenteron, as in other groups of Echinoderms (Selenka, 17, 18). This sac speedily acquired a communication with the exterior by the development of a tubular extension, the pore-canal, which fused with the dorsal ectoderm. In some species, at any rate, the pore-canal was developed before the alimentary canal had been completed by the union of the stomodaeum with the archenteron. N 2 8G TERRA NOVA” EXPEDITION. According to Selenka(18) and Metschnikoff (10) the coelomic sac then became divided into anterior and posterior portions, of which the former gave rise to the hydrocoele or rudiment of the water-vascular system, whilst the latter became divided transversely into two parts which became applied to the right and left sides of the alimentary canal, and constituted the posterior coelomic sacs of the larva (Fig. 5, r.p. c., 1. p. c.). The critical stages in which the posterior portion of the single sac became changed into right and left posterior coelomic sacs were first figured by Metschnikoff ( 1 0). Bury, who examined stages in the development of the larva o{ Synapta digitata ( 1 ), came to a different conclusion as to the manner in which the coelomic sac developed. According to him, when it has become divided into anterior and posterior portions the anterior division does not become directly converted into the hydrocoele. The hydrocoele on the contrary grows out as a bud from its hinder aspect, which remains connected with the main sac, termed by Bury the anterior coelom. The narrow neck « >1‘ connection between hydrocoele and anterior coelom was identified by him with the primary stone-canal. Bury compared the anterior coelom of Synapta to the axial sinus of Asteroidea, and surmised that it gave rise to the “ internal madreporite of the adult Holothuroid. Bury’s view has not been confirmed by subsequent workers. Ludwig (5) examined the development of Cucumaria planci, and came to the conclusion that the anterior portion of the primitive coelomic sac was directly converted into the hydrocoele ; in later stages he found what he termed an “ ampulla,” i.e., a thin-walled expansion on the course of the pore-canal; but he maintained that this was a secondary formation. Newth (15), who worked at Cucumaria saxicula, came to much the same conclusion. Clarke (3), who studied the development of Synapta vivipara, also describes the anterior division of the coelomic sac as being directly converted into the hydrocoele. Now it is to be remembered that Synapta digitata is the only species of Holothuroid with a typical larva, the complete life-history of which has been worked out. The species of Cucumaria have yolky eggs, with shortened life-histories, and larvae which do not show the features of an Auricularia ; whilst the young of Synapta vivipara , as the name of the species implies, develop into the adult form inside the maternal body. We are therefore permitted to surmise either that the observers who worked with these rapidly developing eggs overlooked stages which Bury was able to see in the slowly developing larvae, or else that these stages were actually suppressed in the more modified development. For instance, one might imagine that what Clarke, Ludwig, and Newth regard as a c< pore- ” or “ stone-canal,’ connecting the incipient hydrocoele with the exterior, might represent the vesicle which Bury terms the anterior coelom in a collapsed form. Under these circumstances it was a pleasure to me to discover that the specimens of Auricularia antarctica in the collection constituted a series from which it was possible to reconstruct the entire history of the coelom. In the youngest specimen the coelom is a simple rounded sac opening to the exterior by a pore-canal. In a specimen ECHINODERMA, II— MacBRIDE. S7 slightly older, such as is shown in Fig. 3, the first rudiment of the hydrocoele (hy.) can be made out as a slight thickening on the posterior border of the coelomic sac. In a slightly older larva (Fig. 1) the hydrocoele has not advanced much in development, but from its hinder aspect a small stalked cellular bud can be seen growing out (Fig. 1, post. c.). This is the rudiment of the posterior coelom. Fig. 2 represents a more advanced larva viewed from the side. In it the hydrocoele has become as large as the anterior coelom (as the primitive coelomic sac may now be termed), and is delimited from it by a constriction (Fig. 2, hy.). The posterior coelom has now developed into a long band lying at the side of the stomach. This stage corresponds with that represented in text-fig. 1 in Mortensen’s latest paper (14), but Mortensen has quite misunderstood it. He overlooked the rudiment of the posterior coelom altogether, and has figured a sac entirely detached from the anterior coelom as the posterior enterocoele. This sac, which lies above the junction of the stomach and intestine, has nothing to do with the coelom, but is one of the intestinal pouches characteristic of Auricularia antarctica. In Fig. 6, a still older larva is represented, seen from the dorso-lateral aspect. The hydrocoele has now begun to curve ventrally round the oesophagus. The rudiment of the posterior coelom shows a constriction tending to divide it transversely into front and hind portions. Of these the first is the rudiment of the left posterior coelom, and the hinder one is the rudiment of the right posterior coelom. This stage, as mentioned above, has been figured by Metschnikoff ( 1 0) in his description of the Auricularia of Synapta digitata, but has not been described in connection with any other Auricularia, although Clarke has described it in the embryo of Synapta vivipara. In Fig. 5 is represented a dorsal view of one of the oldest larvae in the collection. The hydrocoele (hy.) has now begun to develop thickened lobes, which are the rudiments of the primary tentacles ; it is sharply marked off from the anterior coelom (ant. c.), with which it is connected by a strongly constricted neck of communication. The two posterior coelomic sacs (r. p. c., 1. p. c.) have now taken up their definitive positions to the right and left of the stomach. A peculiar larval nervous system was described by Semon ( 1 9) in the larva of Synapta digitata. This system consisted of two bands of fibrils running from the corners of the mouth down the grooves which separate the post-oral loop from the main longitudinal portions of the ciliated band. This nervous system is present in Auricularia antarctica (Figs. 1 and 2, l. new.), but it seems to be shorter than in the Auricularia of Synapta digitata. It will be seen that my survey of the specimens of Auricularia antarctica entirely confirms Bury’s account of the development of the coelom in Synapta digitata ; and this account may therefore, I think, be taken as correctly describing the normal sequence of events in Holothurioidea. We have evidence in other groups of Echinoderms that the normal processes may become profoundly modified as a result of the hurrying on of development, and in this way we may probably account for Clarke’s, Ludwig’s, and Newth’s results. The question as to which species, or even family, of Holothurioidea Auricularia 88 “TERRA NOVA” EXPEDITION. antarctica belongs is one which is at present insoluble. In my original description (9) I suggested that Auricularia antarctica might be the larva of some member of the sub-order Synaptida, since calcareous wheels are characteristic of many genera belonging to that division, though not of Synapta itself. Mortensen dissents from this view on the following grounds, viz. : (l) Very few Synaptida are recorded from the Antarctic region ; (2) Auricularia antarctica shows many points of resemblance to Auricularia nudibranchiata ( v . infra), and the latter species is regarded by Mortensen as the larva of one of the Elasipoda (Elpidiida) ; (3) the Elasipoda sometimes possess “ wheels ” amongst their calcifications. When dealing with A. nudibranchiata I shall give grounds for regarding its supposed Elasipodan affinities to be entirely unfounded, though I agree with Mortensen in recognising some striking resemblances in structure between these two species of Auricularia. I adhere, however, to my original suggestion for this reason. According to Ludwig (6) the wheels of Elasipoda differ from those of Synaptida in having the hub perforated; and on this account he regards them as more primitive structures. Now although I described the wheels of Auricularia antarctica as having perforated hubs, yet a renewed examination of the type specimen in the Natural History Museum has convinced me that I was mistaken. What I took for holes appear to be glittering calcareous asperities, and hence Mortensen was right in casting doubt on my original description in regard to this point. But I utterly fail to understand why he states that perforated hubs are unknown, and why he omits to observe that wheels with solid hubs are found in Synaptida only. Had my original statement been confirmed there would have been some ground for a suggestion of Elasipodan affinity. On the ground that a young Holothuroid, which he diagnoses as a Synallactid, was obtained by the German Expedition, Mortensen finally inclines to the belief that A. antarctica may be the larva of this form. As no Synallactid is known to have wheels, this suggestion may be regarded as very improbable. Synaptida are burrowing Holotliuroids, and must, therefore, in most cases evade the dredge ; and the fragments of the two forms already recorded from the Antarctic area probably give no correct suggestion of their abundance. All the specimens of Auricularia antarctica were obtained at depths of between 80 to 160 metres below the surface; the parent must, therefore, be an inhabitant of deep water, and the suggestion that it is a mud-inhabiting Synaptic! gains in probability. 2. Auricularia nudibranchiata, Chun. Fig. 4. Station 92, from summit, Great King, S. by "VV., 24 miles (off N. end of New Zealand), surface. One specimen. In 1896 Chun (2) described a new form of Auricularia, of which he captured many specimens in the vicinity of the Canary Islands. This larva was characterised by the extraordinary complication of the ciliated band, the primary processes of which were bent into numerous secondary processes, which gave to the creature the appearance of a Nudibranch mollusc. For this reason Chun bestowed on it the name Auricularia nudibranchiata. The larva possessed three other characteristics : (1) It had embedded ECHINODERMA, II— MacBRIDE. 89 in it numerous calcareous wheels ; (2) the post-oral loop of the ciliated band extended so far forward that it was separated from the prae-oral loop only by a narrow groove ; (3) the intestine was produced into a median ventral pouch extending forward beneath the stomach. Chun described the anterior division of the larval coelom as a large sac with finger- like outgrowths, which he interpreted as the rudiments of the water- vascular canals of the adult Holothuroid. In 1911 Ohshima (16) described specimens of what he termed “ a large Auricularia allied to A. nudibranchiata” which he captured off the south coast of Japan. The larvae appear to me to be identical with Chun’s species, but to represent a more advanced stage of its development than Chun secured. Ohshima points out that the ramifications of the anterior coelom described by Chun have nothing to do with the lobes of the hydrocoele, for in his larvae the hydrocoele was a clearly defined, thick-walled structure lying at the side of the oesophagus beneath the anterior coelom, with which it was connected by a narrow neck. In a word, the distinction between hydrocoele and anterior coelom, described by Bury in the larva of Synapta d'n/itata, is equally clear in Auricularia nudibranchiata. The specimen of this remarkable larva captured by the “ Terra Xova " Expedition is in an advanced stage of development, and measures at least a centimetre in length, being more than twice the length of Chun’s oldest stage. On examining it I was able to confirm Ohshima’s account, and to add certain details which are shown in Fig. 4. This figure represents only a portion of the monster, including the hinder part of the oesophagus, the front part of the stomach and the adjacent coelomic sacs. The anterior coelom can be seen to open to the exterior by a ciliated duct, the pore-caxal (p. c.), situated at its posterior end. This duct runs along the wall of the coelom for some distance, and gives the impression of liaviug been originally a groove in this wall which has become closed oil’. This is the way in which a considerable portion of the pore-canal is formed in the larva gjf Asterina gibbosa. In front and ventrally, the anterior^ coelom communicates with a well-marked hydrocoele, showing the thickened elevations which are the rudiments of its primary lobes. When the posterior portion of this larva is examined, it is easy to see the structure described by Chun as a median ventral pouch of the intestine. It is, relatively to the size of the larva, of much less size than in the earlier stages described by Chun. Further, in my specimen it is seen not to be a single pouch, but to consist of a pair of closely adpressed pouches which, I have no doubt, are homologous with the intestinal pouches of Auricularia antarctica. On the ground that a single intestinal diverticulum has been described in certain Elasipoda, Chun and Mortensen draw the conclusion that this larva belongs to the Elasipoda. This single diverticulum is supposed to be a rudimentary representative of the gill- (lung-) trees of other Holothurioidea ; but the discovery that the pouch is paired deprives this argument of its ground, and A. nudibranchiata might be the larva of any Holothuroid with internal gills, if these pouches are the rudiments of gills. There is, however, grave reason for doubting this hypothesis. When examining a collection of post-larval young of Holothuroids from the Antarctic, which almost certainly 90 “TERRA NOVA” EXPEDITION. belonged to the genus Cucumaria, I discovered what appeared to he the rudiment <>f the gill-trees, which in this genus are extremely well developed. This rudiment, a single pouch-like outgrowth of the intestine, only appeared in the oldest specimens— in the younger there was no trace of it (9). It is, therefore, exceedingly unlikely that a rudiment of these gills should appear in the Auricularia larva ; they probably only begin to develop when the cloacal function of sucking in and ejecting water has been established. I conclude, therefore, that the intestinal pouches of A. antar<‘ti' ,/ ;md A. nudibranchiata have nothing to do with gills; and, if this be admitted, there is no reason why A. nudibranchiata, like A. antarctica, should not be the larva of a Synaptid, to which group its wheels naturally ally it. D. ASTEROIDEA. 3. Bipinnaria antarctica , nor. Figs. 7, 8. Station 274. 71° 29' S., 166" W., 80 metres. 11 specimens. ,, 284. 71° 49' S., 167 32' W., 80 metres. 3 specimens. ,, 290. 72 S., 168" 17' W., 60 metres. Many specimens. In the same locality, and at the same depths at which Auricular ia antarctica was captured, numerous specimens were obtained of a large Bipinnaria in various stages of development, ranging from a stage in which no hvdrocoele can be detected to one in which not only the lobes of the hydrocoele, but also the aboral spines of the future starfish are well developed. The calcareous matter of these spines had, of course, been dissolved by the acidity of the preservative; but the organic base of the spines remained, and their shapes were consequently well defined in the specimens. A careful search through Mortensen’s synopses of known Echinoderm larvae (12, 13) failed to disclose any species to which this Bipinnaria could be referred, and hence I feel justified in creating for its reception a new species, with the name Bipinnaria antarctica. The new species is characterised by long arms, which do not attain, however, the proportionate length of those of the Bipinnaria of Luidia. Its most distinguishing feature is, however, the crenulation of the ciliated band, which is well shown in Figs. 7 and 8. This is a character which has so far been recorded of no species of Bipinnaria. It may be compared to the formation of secondary loops on the ciliated band of Auricularia nudibranchiata, and of secondary tentacles on the ciliated band of Tornaria grenacher i ( v . infra), but the folding is slighter than in the case of either of these two larvae. This folding or crenulation of the ciliated band is confined to those portions of its course which are situated on the arms of the larva. Each of these arms has a groove on its surface bordered by ridges which are surmounted by the ciliated band. Under- lying the groove are three bands of longitudinal muscle-fibres (Figs. 7 and 8, long, muse.), one in the centre and one on each side. There are two main types of Bipinnaria larva which may be termed the Asterias- type and the Luiclia-tjipe respectively. Larvae of the first type have no median ventral ECHIXODERMA, II, AXD EXTEROPXEUSTA — MacBRIDE. 91 arm in front of the mouth ; in their older stages they develop three peculiar club- shaped arms on the region of the prae-oral lobe, between the prae-oral ciliated band and the median dorsal arm. Such larvae are then termed Brachiolaria, and they can hold on to the substratum by these new arms. Eventually they fix themselves permanently bv means of an adhesive disc situated in the centre of the three extra arms, and the whole front region of the larva is gradually absorbed. Larvae of the second type have a median ventral arm which carries a loop of the prae-oral ciliated band. So far as is known they never develop into Brachiolaria, and although the life-history of none of them has been completely worked out, yet the observations of M. C. Delap (4) on the larva of Luidia seem to indicate that, when metamorphosis approaches, the hinder portion of the larva containing the stomach and intestine, and surrounded by the circle of arm-rudiments, becomes amputated from the front half and sinks to the bottom as the future starfish, whilst the front half remains swimming. As an inspection of Fig. 7 demonstrates, Bipinnaria antarctica belongs to the Luidia type of larvae, for the median ventral arm is well developed, and in the most advanced specimen no trace of Brachiolaria arms has appeared. It will be noticed that each of the aboral plates of the future starfish carries two large conical spines. Now amongst the starfish described by Ludwig (7), in his description of the collection brought back from Cape Horn and Antarctic waters by the “ Belgica,” there occurred a specieis of the genus Cheiraster on which Ludwig bestowed the name C. gerlachei. Some very small specimens of this species were obtained, which must have metamorphosed only a short time before. In these young starfish the terminal plates covering the tips of the arms each bear two large conical spines resembling those borne by Bipinnaria antarctica. Although Cheiraster gerlachei was obtained in longitude 82 W., it was found at about the same latitude (about 70° S.) as the larvae ; and as between the two localities there extends a comparatively short stretch of open Antarctic sea, it is quite likely that Cheiraster gerlachei inhabits the whole of this area, and therefore we may assign Bipinnaria antarctica to this species. Cheiraster belongs to the family Archastcridae, which is closely allied to the Astropectinidae to which Luidia belongs, and it is practically certain that it must have a larva of the Luidia type. ENTEROPNEUSTA. 4. Tornaria grenadier i, Spengel. Figs. 9, 10. Station 130, off Three Kings Islands, N. end of Xew Zealand, surface. Two specimens. Two specimens of a species of Tornaria were obtained by the Expedition in a haul made in the vicinity of Three Kings Islands, i.e., in the same neighbourhood as that in which Auricular m nudibranchiata was obtained. One specimen was in process of metamorphosis, but the other was in the height of larval development, although unfortunately somewhat mutilated. The metamorphosing specimen is shown in Fig. 10. VOL. IV. 0 92 “TERRA NOVA” EXPEDITION. It will be . seen that the longitudinal ciliated band has almost entirely disappeared, but that the strong posterior transverse ciliated band has persisted. From an examination of the longitudinal ciliated band in the mutilated specimen, and of the vestiges of it in the metamorphosing larva, it is clear that this band was produced into secondary tentacle-like processes. It can also be seen that the stomach (stom.), instead of being, as in other species of Tornaria, a globular sac, is an elongated tubular organ. Now in respect of both these organs this larva agrees with a species termed by Spengel (20) T. grenadier i, which was captured in mid-Pacific, between the Sandwich and Marshall Islands, and we may therefore assign our larva to this species. Spengel gives the name grenacheri to the species because he regards it as absolutely identical with a larva captured by Prof. Grenadier, near the Cape Verde Islands in the South Atlantic ; he also asserts that it is identical with a Tornaria first recorded by Weldon (21) from the Bahamas, the development of which was subsequently worked out by Morgan (II), who terms it the Nassau larva. Spengel also mentions that an extremely similar larva was captured by Driesch in the Indian Ocean. If all these larvae really belong to the same species, we should have to infer the existence of an Enteropneust worm with a distribution encircling the earth. When, however, we consider that Auricularia nudibranchiata has been captured both in the South Atlantic and the South Pacific, we see that it too must belong to some Holothuroid with an equally wide distribution. Now both Synaptida (wheel-bearing Holothurioidea) and Enteropneusta are creatures which burrow in sand and mud, and on bottoms like these at considerable depths, conditions must be extraordinarily uniform over enormous ranges, so that we should expect the species inhabiting them to be equally widely distributed. There is, therefore, nothing inherently improbable in Spengel’s- theory ; but there are one or two points in the structure of the larva which we are discussing which make me doubt whether it is identical with the Nassau larva, although that larva doubtless belongs to an allied species. When we examine Morgan’s figures we observe that the stomach in the Nassau larva, although elongated, is not so tubular as in the typical T. grenacheri Moreover, Morgan figures in the metamorphosing Nassau larva the rudiments of the first gill- pouches, but no trace of a notochord. In our specimen no trace of gill-pouches can be detected, but a well-marked diverticulum is seen to arise from the oesophagus. The tip of this pouch (rich.) curves back towards the base of the anterior coelom or proboscis-cavity (ant. c.), and I think there can be no doubt that this is the rudiment of the notochord (Fig. 10). It appears more probable therefore that T. grenacheri and the Nassau larva belong to two distinct but allied species. Fig. 9 is intended to elucidate a point in the anatomy of the Tornaria larva, of which I can find no clear account either in Spengel’s or in Morgan’s descriptions. It is well known that in the adult Enteropneust there is a structure which has been variously termed “glomerulus” or “proboscis-gland.” This is a modification of the epithelium of the posterior or basal wall of the proboscis-cavity, where this is indented by the notochord. The cells in this area become thickened and are richly supplied with blood-vessels, and it has been surmised with much probability that the EXTEROPXEUSTA— MacBRIDE. 93 whole structure is an excretory organ. What seems clearly to be the beginning of this organ is shown both in Fig. 9 and Fig. 10 ( gl . ep.). In Fig. 9 the neighbourhood of the proboscis-pore (d. p.) is shown on an enlarged scale. It will be seen that the proboscis-cavity is prolonged into two postero-dorsal tubes or “ horns ” (r. h. and I. h.), of which only the left communicates with the exterior by the proboscis-pore. Beneath the right horn lies the completely closed pericardial sac (per.), which in life is contractile. The analogy of the proboscis-pore and the adjacent pericardial sac to the maclreporic pore and the madreporic vesicle of Echinoderms (which in the Bipinnciria larva is contractile) has often been pointed out. Beneath the pericardium is a blood-space (Fig. 9, II) which becomes the heart of the adult ; the roof of this space is formed by the contractile floor of the pericardial sac. Around this blood-space there is a crescentic mass of deeply-staining and obviously glandular tissue, which must be a modification of the ventro-lateral walls of the peri- cardial sac. To this structure I propose to give the name of heart-gland (Fig. 9, //. gl.). The only reference to it in the works of Morgan and Spengel which I can find is a statement that the cells forming the lateral walls of the pericardial sac become “ pear-shaped,” and they seem to regard the structure as the first rudiment of the proboscis-gland; but the complete distinctness of the two structures is obvious from an inspection of either Fig. 9 or Fig. 10. We can only surmise that the heart-gland has some function in connection with the blood, and perhaps is an organ of internal secretion. 94 “TERRA NOVA” EXPEDITION. V— LITERATURE. (1) Bury, H. — 1895. “The Metamorphoses of Echinoderms.” Quart. Journ. Micr. ScL VoL XXXVIII, p. 45. (2) Chun, C. — 1896. “ Awricularia nudibranchiata.” “ Atlantis. Biologische Studien fiber pelagische Organismen.” Bibl. Zool. Vol. VII, Heft 19, p. 55. (3) Clarke, H. — 1898. “ Synapta vivipara,a Contribution to the Morphology of Edunoderms.” Mem. Boston Soc. Nat. Hist. Vol. V, p. 53. (4) Delap, Maud C. — 1905. “Notes on the Plankton of Valencia harbour.” Appendix No. 7 to Part II. Report on Sea and Inland Fisheries of Ireland. (5) Ludwig, H. — 1891. “Zur Entwickelungsgeschichte der Holothurien.” Sitzb. Akad. Wias. Berlin, p. 603. (6) Ludwig, H. — 1892. “Die Seewalzen.” Bronn’s Classen und Ordnungen des Thier-Reichs. Vol. II, Abth. 3. Buch 1. (7) Ludwig, H. — 1903. “ Seesterne.” Resultats du Voyage du S.Y. “ Helgica.” Antwerp. (8) MacBride, E. W., and Simpson, J. C. — 1908. “ Echinoderm larvae.” Nat. Antarctic [“Discovery”] Exp. (Nat. Hist.) Vol. IV. (9) MacBride, E. W. — 1912. “On a collection of young Holothuroids.” Ibid. Vol. VI. (10) Metschnikoff, E. — 1869. “Studien fiber die Entwicklung der Echinodermen und Nemertinen.'* Mem. Acad. St. Petcrsbourg. Series 7. Vol. XIV. No. 8. (ID Morgan, T. H. — 1894. “The development of Balanoglosstu .” Journ. Morph. Vol. IX, p. 1. (12) Mortensen, Tii. — 1898. “ Die Echinodermen-Larven der Plankton Expedition.” Ergebn. Plankt. Exped. Vol. II, J. (13) Mortensen, Tii. — 1901. “ Die Echinodermen-Larven.” Nordisches Plankton. Vol. VII. (14) Mortensen, Th. — 1914. “Die Echinodermen-Larven der Deutschen Siid-Polar Expedition.” Vol. XIV. (15) Newth, H. G. — 1916. “ The development of Cucumaria saxicoln.” Proc. Zool. Soc. (London), p. 631. (16) Ohshima, H. — 1911. “Note on a gigantic form of Awricularia allied to A. nudibranchtata .” Annot. Zool. Jap. Vol. VII, p. 347. (17) Selenka, E. — 1876. “ Zur Entwicklung der Holothurien.” Zeitschr. wiss. Zool. Vol. XXVII, p. 155. (18) Selenka, E. — 1883. “Studien iiber die Entwicklung der Thiere.” Heft II. “ Die Keimbliitter der Echinodermen.” Wiesbaden. (19) Semon, R. — 1888. “ Die Entwicklung von Synapta digitate, und die Stammesgesehichte der Echino- dermen.” Jen. Zeitschr. Vol. XXII, p. 175. (20) Spengel, J. W. — 1894. “ Die Enteropneusten des Golfes von Neapel und der Angrenzenden Meeresabschnitte.” Fauna und Flora des Golfes von Neapel. Mon. 18. (21) Weldon, W. F. R. — 1887. “Preliminary Note on a BaJanoglossus larva from the Bahamas.” Proc. Roy. Soc. (Lond.). Vol. XLII, p. 146. LONDON : PRINTED BY WILLIAM CLOWES AND SONS, LIMITED, DUKE STREET, STAMFORD STREET, S.E. 1, AND GREAT WINDMILL STREET, W. 1. Echinoderma, Part II, and Enteropneusta, Plate I. PLATE I. Auricularia antarctica ,, nudibranchiata . Figs. 1-3, 5, G. Fig. 4. List of Abbreviations employed. a. d. antero-dorsal process of ciliated band. ad. cil. adoral ciliated band. an. anus. ant. c. anterior coelom. calc, cells forming organic basis of calcification. Icy. hydrococle. int. intestine. ini. d. intermediate dorsal process of ciliated band. int. 1. intestinal lobe. 1. nerv. larval nervous system. 1. p. c. left posterior coelom. m. mouth. tn. p. madreporic pore. ocs. oesophagus. p. c. pore-canal. p. d. postero-dorsal process of ciliated band. j>. 1. postero lateral process of ciliated band. jj. o. post-oral process of ciliated band. post. c. posterior coelomic sac. pr. o. prae-oral process of ciliated band. r. p. c. right posterior coelom. stom. stomach. The figures are described in the order of the ages of the larvae represented. Fig. 3. — Ventral view of a young Anrieularia antarctica, showing the first stages in the development of the liydrocoelc (hi/.) ; calc, organic bases of calcification. Magnification 22 diameters. Fig. 1. — Oblique ventral view of a somewhat older specimen of A. antarctica ; 1. nerv. larval nervous system ; post. c. a bud of cells, growing from the posterior border of the hydrococle, which is the first rudiment of the posterior coelomic sac. Magnification 22 diameters. Fig. 2. — A still older specimen of A. antarctica viewed from the left side; int. 1. lobes of the intestine. The hydrocoele (%.) is a large sac separated from the anterior coelom (ant. c.) by a constriction. The posterior coelomic sac (post, c.) is represented by a band of cells extending backwards over the stomach. All the processes of the longitudinal ciliated band are well seen. Magnification 22 diameters. Fig. 6. — -Specimen of A. antarctica , slightly older than that shown in Fig. 2, viewed from the left dorso- lateral aspect. The posterior coelomic sac is in the act of dividing into right and left posterior coeloms (r. p. c., 1. p. c.). Magnification 22 diameters. Fig. 5. — One of the oldest specimens of A. antarctica viewed from the dorsal aspect. The hydrocoele ( hy .) shows the beginnings of the primary lobes ; it is sharply marked from the anterior coelom (a n t. c. ), with which it is connected only by a constricted canal. The posterior coelom is completely divided into right and left posterior coeloms, which are situated at the sides of the stomach. Magnification 22 diameters. Fig. 4. — A portion of the specimen of Auricularia nudibranchiata viewed from the dorsal aspect, showing oesophagus, stomach, and adjacent coelomic cavities ; p. c. the pore-canal running along the posterior aspect of the anterior coelom. Magnification 16 diameters. Brit. Mus. (Nat. Hist.) Brit. Antarctic (Terra Nova) Exped. 1910. Echinoderma and Zoology.Vol.IV. Enteropneusta, PI. 1. nerv. I.nerv.-S a.d.cil.l stom. ad.cil stom ant.c stom a i_ e f \ -A.. J . Engel Terzi del. et Hth. Bale & Banielsson.L^ V ^ •'^ww *v» B ip i nn aria antarct i ca Tornaria grenacheri PLATE II. Figs. 7, 8. Figs. 9, 10. List of Abbreviations employed. ab. p. aboral plates of starfish. a. d. antero-dorsal arm of Bipinnaria. ad. cil. adoral ciliated band. an. anus. ant. c. anterior coelom of Tornaria ; conjoined right and left anterior coeloms of Bipinnaria. ap. apical plate of Tornaria. cil. ep. ciliated epithelium. cil. r. ciliated ring of Tornaria. coll, collar-groove of Tornaria. d. p. dorsal pore of Tornaria. gl. ep. glandular epithelium of base of proboscis cavity or anterior coelom of Tornaria. H. heart of Tornaria. It. gl. heart-gland. hy. hydrocoele. hit. intestine. Z. a. c. left anterior coelom of Bipinnaria. l. h. left horn of proboscis-cavity or anterior coelom of Tornaria. long. muse, longitudinal muscle -fibrils. Z. p. c. left posterior coelom. m. d. median dorsal arm of Bipinnaria. m. v. median ventral arm of Bipinnaria. nch. rudiment of notochord of Tornaria. oc. eye-pit of Tornaria. oes. oesophagus. p. d. postero-dorsal arm of Bipinnaria. per. pericardial sac of Tornaria. p. 1. postero-lateral arm of Bipinnaria. p. o. post-oral arm of Bipinnaria. pr. o. prae-oral arm of Bipinnaria. r. a. c. right anterior coelom of Bipinnaria. r. h. right horn of anterior coelom of Tornaria. r. p. c. right posterior coelom. stom. stomach. Fig. 7. — A specimen of Bipinnaria antarctica viewed from the ventral aspect. The postero-lateral and post-oral arms of the right side (left side of the figure) are broken off, and their former positions are indicated by dotted lines ; ab. p. organic bases of the aboral (terminal) plates of the future starfish, each carrying two conical spines ; ant. c. union of right and left coelomic sacs in the prae-oral lobe ; long. muse, longitudinal muscle-fibres of the arms. Magnification 16 diameters. Fig. 8. — A portion of one of the arms of another specimen of B. antarctica viewed from the side ; cil. cp. ridge of ciliated epithelium bordering the groove. Magnification 140 diameters. Fig. 10. — A specimen of Tornaria grenacheri, Spengel, in the act of metamorphosis, viewed from the dorsal and posterior aspects ; a. p. apical plate, with oc. the eye-pits ; cil. ep. remnants of longitudinal ciliated band ; cil. r. posterior ciliated ring ; d. p. dorsal pore ; gl. ep. glandular epithelium — rudiment of proboscis-gland ; nch. rudiment of notochord ; coll, beginning of collar-groove. Magnification 22 diameters. Fig. 9. — The dorsal pore and adjacent organs of the specimen shown in Fig. 10 viewed from the ventral and anterior aspect ; If heart ; h. gl. heart-gland ; per. pericardial sac ; r.h. and Z. h. right and left horns of the anterior body-cavity. Magnification 60 diameters. Brit. Mias. (Nat. Hist.) Brit. Antarctic (Terra Nova) Exped. 1910. Z o olo gy, Vol. IV. E chino derma and Enteropneusta, PI . II. m. ab.p. long. muse Fig. 8. oes. r.a.c. coll. int. oes. nch. stom Fxg.9. cil. ep. cil. r. per r.h. J. Engel Terzi del. et Tit.E . Fig. 10. Bale & Daniels seen. (grtfief (Museum (QUfuraf JE)^orS)- “Terra Nova” Report. This is No. 3 of 25 copies of Zoology , W. /K, Afa. 4, Rhabdopleura , printed on Special paper. 95 37 A? A' I * 5° z JdpAL h\s>S rhabdopleura; By J. R. NORMAN. WITH SIX FIGURES IX THE TEXT. PAGE I. — Introduction and Historical Review ...... 95 II.— Species and Their Distribution . . . . • • • 96 III. — Material Collected by the “Terra Nova” . . . . . 98 IV. — References ........... L02 I.— INTRODUCTION AND HISTORICAL REVIEW. I have to thank Sir Sidney Harmer, K.B.E., F.R.S., for entrusting me with this investigation, and Professor E. W. MacBride, F.R.S., for permitting me to carry out the work under his supervision in the Huxley Research Laboratory of the Imperial College of Science, and for the frequent and valuable advice that he has given me during its progress. Rhabdopleura was dredged in 1868 by A. M. Norman and Jeffreys, and described by Allman ( 1 , 2) in the following year as Rhabdopleura normani, a peculiar form of Polyzoa. Rhabdopleura had been previously dredged by G. 0. Sars (19, p. 23) in 1866, mentioned by M. Sars f in 1868 as Halilophus mirabilis, and fully described by G. 0. Sars (18, 19) in 1872 and 1874 as Rhabdopleura mirabilis. In 1874 Lankester (14) expressed the opinion that Rhabdopleura was related to the Lamellibranch Molluscs. In 1882 McIntosh \ published the preliminary notice of Cephalodiscus , in which he suggested that Rhabdopleura was probably closely related to that genus ; and when Harmer (8) showed in 1887 that Cephalodiscus was related to Balanoglossus, the systematic position became more clear. The thorough investigation * Manuscript received April 28, 1921. [S.F.H.] f "Fortsatte Bemaerkninger over det dyriske Livs Udbredning i Havets Dybder,” Christiania Vid. Selsk. Forh., 1868-1869, p. 255. t Ann. Mag. Nat. Hist. (5), X., 1882, pp. 337-348. VOL. TV. P “TERRA NOVA” EXPEDITION. ■ 96 made by Lankester (15) in 1884, and the subsequent work of Fowler (5, 6, 7), with “ Challenger ” material, further strengthened the view that Rhabdopleura and Cephalodiscus were closely related. The unillustrated publications of Conte and Vaney (3, 4) in 1902 contain results which stand in direct opposition to those obtained by all other workers. The most recent work on Rhabdopleura has been carried out by Schepotieff (20-24), whose series of papers published from 1904 to 1909 contain detailed accounts of the morphology of the organism, together with important conclusions relating to the buds and method of budding, and to the affinities of the genus. These publications complete the proof of the fundamental resemblance between Rhabdopleura and Cephalodiscus. The relations between the two genera were summarised by Ridewood (17) in 1907. Other investigators such as Hincks (11), Jullien (12, 13), and Xorman (16), have published accounts of Rhabdopleura, but have only dealt with it from a systematic standpoint. II -SPECIES AND THEIR DISTRIBUTION. In Rhabdopleura, as in Cephalodiscus, the characters of the ccencecium (tubarium *) form the principal means of distinguishing the species. A number of species have been described, but it is not yet possible to say how far all of these are distinct. Schepotieff (22, p. 471) notes that the form of the coencecium tends to vary considerably even in a single colony, and suggests that probably all the specimens of Rhabdopleura known (in 1906) belong to the same species, i.e. R. normani. The occurrence of Rhabdopleura among the “ Terra Nova " dredgings is interesting in that it extends the geographical range of the genus, which appears to be almost world- wide. The extreme limits from which it has been recorded are West Greenland y and South Australia (Harmer 9, p. 23.) J The following represents a synopsis of all the species of Rhabdopleura which have been described, together with the localities and depths from which material of these species has been obtained : — 1. R. normani, Allman (2) [? = R. mirabilis, Sars]. — Characterised by the great length of the free, upright peristomes. Greenland ; Shetland Islands ; Norwegian coasts ; Tristan d’Acunha. Depth varies considerably ; 5 m. (22, p. 468) to 500 m. 2. R. compacta, Hincks (11). — Peristomes short; tubes set closely side by side, forming small, crust-like colonies. Coasts of Ireland and Brittany. 100 m. 3. R. grimaldii, Jullien (12). — Peristomes short, each continuing a creeping portion of equal length ; two delicate “ tubes ” running longitudinally along the free * Lankester (15), p. 624. f Norman, A. — “ Notes on the Natural History of East Finmark.” Ann. Mag. Nat. Hist. (7), XII, 1903, p. 101. ! This was merely a fragment, and could not be referred to any particular species. RHABDOPLEUR A -NORMAN. 97 wall of the creeping portion, converging anteriorly, and ending at the base of the peristome ; pectocaulus not extending into the recumbent portion of the zooecium ; ccenoecium brown. Azores. 318 m. 4. R. manubialis, Jnllien (13). — Peristome twice as long as the attached portion ; longitudinal “ tubes ” of R. grimaldii absent ; pectocaulus extending into the recumbent portion of the zooecium ; ccenoecium yellow. Azores. 318 m. 5. R. striata, Schepotieff (24). — Peristome composed of plates arranged in regular circles ; pectocaulus paler than in R. normani, and hollow, so that the body-cavities of separate individuals communicate with one another. Indian Ocean (Coast of Ceylon). 2 to 3 m. 6. R. anniilata, n.sp. Three Kings Islands, north of New Zealand ; islands of Wowoni and Buton (near Celebes). 183 to 549 m. On the suggestion of Sir Sidney Harmer, a re-exam- Fic, ^ B -Rhabdopleura normani. Portions of peristomes ( X 56). A, Specimen from Norway (Harmer 1 0). B, Specimen from Tristan d’Acunha. ination of the Rhabdopleura material, dredged on the “ Challenger ” Expedition, was undertaken with a view to ascertaining whether the “ Challenger ” material from Tristan d’Acunha described by Fowler (5, 6. 7) is identical with R. normani. Through the courtesy of Mr. R. Kirk- patrick, of the British Museum (Natural History), I have examined authentic material of R. normani from Norway and the Shetlands, collected by A. M. Norman. I have also had the advantage of studying preparations made by Sir Sidney Harmer, from material from Norway, and now in the Cambridge Museum. The specimens from Norway in the British Museum, and those from Tristan d’Acunha, are all attached to colonies of Lopliohelia prolifera, and Fowler (3, p. 293) remarks, concerning the material from the latter locality, that '‘it is not a little remarkable that the two species thus associated should be apparently identical with the species of these genera which are associated off Norway and the Shetlands, though separated byabout 98° of Latitude.” The peristomes of the specimens from Tristan d’Acunha (Fig. 1, B) very closely resemble those prepared by Harmer from specimens from Norway (Fig. 1, A), one of which was figured by him in the Siboga Report (10, PI. 11, Fig. 19). The remaining parts of the coencecium also appear to be similar. The peristomes, however, differ from those of the colonies collected by Norman, which appear to vary considerably, both in different colonies, and even in the same colony. This variation affects the diameter of the peristomes, the thickness p 2 D ABC Fig. 2. — Rhabdopleura normani. Optical sections of peristo- mial tube-walls ( x 185). A-C, Specimens from Nor- way (Norman Collection). D, Specimen from Tristan d’Acunha. 98 “TERRA NOVA” EXPEDITION. of the peristomial walls, and the character and extent of the lateral projections of the peristomial walls (Fig. 2). It would appear, therefore, that if the characters of the coenoecium may be taken as a reliable means of distinguishing the species, as Harmer ( 1 0, p. 127) thought probable, the colonies of Rhabdopleura from Tristan d'Acunha are referable to R. normani. The following table of measurements of the diameters of the peristomes, and thickness of the peristomial walls, is compiled from measurements made by Harmer (10, p. 127), and SchepotiefE (20, p. 6), and from measurements made by me from “Norman ” material in the British Museum (Natural History), and from “ Challenger ” material. Harmer. Internal diameter of peristomial tube .... .... 255-270/a Thickness of wall at the middle of a tube-ring 16-24/a Normau Coll. SchepotiefE. Brit. Mus. (Nat. Hist.) 180-185/a under 8/a 130-200/a 3-10/a “ Challenger ” Material. 180-260/a 2-18/a III -MATERIAL COLLECTED BY THE “TERRA NOVA.” The material of Rhabdopleura collected by the “ Terra Nova ” Expedition was obtained at two stations lying to the north of New Zealand. Station 90 — July 25th, 1911 ; from Summit, Great King, Three Kings Islands, New Zealand, S. 14° AY., 8 miles, 100 fathoms (183 m.), rock. Station 91 — July 26th, 1911 ; from Summit, Great King, Three Kings Islands, New Zealand, S. 10° AY., 25 miles, 300 fathoms (549 m.), rock. The material consists of two small stones and one Lamellibranch shell dredged from 100 fathoms, with the Rhabdopleura colonies in a broken condition ; and one larger stone dredged from 300 fathoms, with an extensive colony in better condition. All the material was preserved in 70 per cent, alcohol. The state of preservation was very poor, and the zooids were found to be much decomposed, rendering an examination of any structural details impossible. A description of the ccenoecium, therefore, was all that could be undertaken. The specimens cannot be referred to any existing species, and the name Rhabdopleura annulata is proposed, on account of the projections of the peristomial tube- rings. RHABDOPLEURA ANNULATA, n.sp. Colony irregularly branched and attached to stones or shells. Coenoecium of a pale brown colour. Diameter of the free, cylindrical portions the same as that of the adherent portions. External projections of the peristomial rings very clearly marked. The free peristomes arise directly from the creeping part of the coenoecium, which contains the pectocaulus, and are not recumbent for any part of their length. Length RHABDOPLEURA— NORMAN. 99 of the peristomes and the number of peristomial rings varying considerably. Internal diameter of peristomes, 160-200/x. Maximum external diameter (i.e. across the angular projections of the tube rings), 2 65p,. Maximum thickness of peristomial wall, 30-60yu. Minimum thickness of peristomial wall, 9-1 3/x. Average distance between tube-rings, 50 p. Pectocaulus of a deep black colour, 27-34/x in diameter. Where the ccencecium is attached to a stone, the tubes are found on almost every side of the stone, although the greater part of the colony is confined to one surface. The colony attached to the shell is largely confined to the outer surface of the valve, although small portions of the ccencecium are to be seen on the concave surface. In all the colonies, the tubes, especially the free upright peristomes, have been considerably damaged, which damage is probably due largely to the shaking in the bottles which the specimens must have undergone before being handed over for investigation. For this reason it was impossible to follow the course of the colony for any distance, or to investigate the manner of branching, and the starting point of the colony. Many Hydroids and Polyzoa are associated with the Rhabdopleura colonies, which, in their older parts, are actually encrusted with foreign growths. The ccencecium is transparent, and coloured pale brown, but in some parts it appears dark brown or even black, owing to the decomposition of the contained zooids and buds. The creeping stem (Fig. 3), attached to the stone or shell, shows the oblique sutures which have been noticed in other species, and impart to it a zig-zag appearance. It contains the characteristic pectocaulus ( p ., Figs. 3, 4) embedded in its lower wall. The tube-wall of this part of the ccencecium is marked by lateral projections, which are not quite so prominent as those of the peristomial wall. In fragments of the creeping stem it may be observed that the pectocaulus does not always lie in the middle line of the basal wall, but may be found at the extreme edge. This is probably due to its detachment during the preparation of the slide. The average diameter of the pectocaulus is about 31/x, but measure- ments taken at various parts give diameters ranging from 27 to 34/x. The inner surface of the attached portion of the coencecium, as well as of the free, upright peristomes, is perfectly smooth. The creeping stem is divided internally into a number of compartments by transverse septa ( sp ., Figs. 4, 5), and from each of these compartments the perpendicular, free branches or peristomes are given off. These branches, which contain the zooids during life, differ from the branches described in R. normani, in that they are never recumbent for any part of their length, but arise directly at right angles from the creeping stem containing the pectocaulus (see Fig. 4). A similar arrangement was noted by G. 0. Sars (19) in his description of specimens obtained from the Lofoten Fig. 3. — Rhabdopleura annulata, n.sp. A fragment of the creeping portion of the ccencecium viewed from above ( x 56) ; p., pecto- caulus. 100 “TERRA NOVA” EXPEDITION. Islands, and it was this feature which led him to separate his specimens from those described by Allman, under the name of R. mirabilis. Lankester (15, p. 626) notices that Sars obtained his specimens “ in fragments growing upon coarse sand — attached not to a continuous support, but to one piece after another of coarse angular rock particles.” He suggests that the difference between the specimens of Sars and Allman is one which is merely due to the nature of the surface on which the colony is growing ; and considers that Sars has not given sufficient reason for separating R. mirabilis from R. normani. Schepotieff (22, p. 470) is of the same opinion, and further notes that where a Rhabdopleura colony is growing on smooth surfaces such as mussel-shells or flat stones, the peristomes tend to rise directly vertically from the creeping portion containing the pectocaulus ; whereas, where the surface is broken and irregular, such as tubes of Serpula and pieces of coral, the lateral branches containing the zooids tend to be recumbent for a part of their length. Fig. 4. — Rhabdopleura annulata, n.sp. A fragment of the creeping portion of the coenoeciuin viewed from the side, with three peristomes ( x about 56). Slightly diagrammatic ; p., pectocaulus ; sp., septum. This statement does not appear to be true in all cases, for all the colonies dredged by the “ Terra Nova ” Expedition are growing on surfaces which are partly smooth and partly broken. In all cases, however, the colonies exhibit the arrangement noticed by Sars, that is, the peristomes arise directly at right angles from the creeping stem. Further, the colonies of R. normani in the British Museum, and those dredged by the “ Challenger ” Expedition, are growing either on pieces of coral, or on a perfectly smooth shell, but the lateral branches are here recumbent for a part of their length. It would seem, therefore, that in R. annulata at any rate, this is a truly specific feature, and in no way due to the nature of the surface to which the colony is attached. The distance between the points where the peristomes are given off varies considerably, two peristomes sometimes ascending side by side, and sometimes separated by a distance of 3 to 8 mm. The peristomes are, as a rule, more or less straight, but twisted specimens are occasionally found. The height of the free peristomes varies greatly, as does the number of rings on each peristome. I have examined an extensive series of complete peristomes, and find RHABDOPLEUR A— NORMAN. 101 that the number of tube-rings varies from 9 to 47. The number of rings is undoubtedly correlated with the age of the contained zooid, which adds rings at intervals to the free edge of the peristome. The average distance between the sharply projecting peristomial rings is 50/r. These rings are more prominent than in R. normani, giving the peristome a more clearly annulated appearance. As has been noticed in other species, the rings are not complete, but are interrupted by oblique sutures ( s ., Fig. 5). They are clearly described and figured by Schepotieff (20, p. 6, PI. I, Fig. 3) in R. normani. Harmer ( 1 0, p. 8) notes that “ the suture is easily seen when it is on that side of the ring which faces the observer, less easily when it is necessary to focus through the tube in order to see it, and may be difficult to distinguish when it is on either side of the tube.” From examination of a number of peristomes it would seem that the suture is an ordinary feature of the peristomial ring, and represents the junction between the part first formed by the zooid, and that last formed. The projections of the peristomial rings are very clearly marked, and form one of the main features which distinguish the species. They stand out as much as from 30-50 p from the tube-wall, as compared with lop. in R. normani (SchepotiefE, 20, p. 6). The extent of these angular projections in R. annulata is shown in Fig. 6. In describing a specimen of Rliabcloplewa obtained from near Celebes (Stn. 204) by the “ Siboga ” Expedition, Harmer (10, p. 127) noticed the angular projections of the tube-rings. He suggests that the “ Siboga ” specimen is probably a new species characterised by the small number of tube-rings, by their comparative delicacy and thinness, and by their angular projections to the exterior. The specimen was so fragmentary, however, that he did not consider himself justified in giving it a specific name, but it seems probable Fig 6- — Rhabdopleura that it belongs to the species here described as R. annulata. sec" The measurements of the two are as follows : — Fig. 5. — Rhabdopleura annii- lata, n.sp. A complete peristome ( x 56) ; o., orifice of tube ; pc., pec- tocaulus ; s., suture of tube-ring ; sp>., septum. annu- a peristomial tube-wall ( x 185). Internal diameter of peristomial tube Thickness of wall at the middle of a tube-rin<* Specimens from “ Siboga,’ Station 204. 175-21 Oja 7‘5-11‘5/x Specimens from “ Terra Nova,” Stations 90, 91 (R. annulata, n.sp.). 160-200/x 9-13 ft 102 TERRA NOVA” EXPEDITION. IV.— REFERENCES. 1. Allman, G. J. — 1869. “ Re]3ort on Shetland Dredgings.” Rep. Brit. Ass. Adv. Sc., p. 311. 2. Allman, G. J. — 1869. “ On Rhabclopleura, a new form of Polyzoa, from Deep Sea Dredging in Shetland.” Quart. Journ. Micr. Sci., IX, pp. 57-63. 3. Conte, H., et Vaney, C. — 1902. “ Contribution a Tetude anatomique de Rhabdopleura” C. r. Acad. Sci. (Paris), cxxxv, pp. 63-65. 4. Conte, H., et Vaney, C. — 1902. “ Recherches sur le bourgeonnement de Rhabdopleura.” I. cil. pp! 748-750. 5. Fowler, G. H. — 1892. “ The Morphology of Rhabdopleura Normcmi Alim." Festschr. 70 Geburtstage R. Leuckarts, pp. 293-297. 6. Fowler, G. H. — 1893. “ Note on the Structure of Rhabdopleura .” Proc. Royal Soc. London, LII, pp. 132-134. 7. Fowler, G. H. — 1904. “ Notes on Rhabdopleura Normani Allman.” Quart. Journ. Micr. Sci., XLVIII, pp. 23-31. 8. Harmer, S. F. — 1887. Appendix to Report on Cephalodiscus. ‘‘Challenger” Reports, Zook, XX, pt. LXII, pp. 39-47. 9. Harmer, S. F. — 1904. “ Hemichordata.” Camb. Nat. Hist., VII, pp. 21-32. 10. Harmer, S. F. — 1905. “ The Pterobranchia of the Siboga Expedition, with an Account of other Species.” Siboga-Expeditie, livr. 22, monogr. 26 bis, pp. 132. 11. Hincks, T. — 1880. “ A History of the British Marine Polyzoa.” London, pp. 577-582. 12. Jullien, J. — 1890. “Description d’un Bryozoaire nouveau du genre Rhabdopleura .” Bull. Soc. Zool. France, XV, pp. 180-183. 13. Jullien, J., et Calvet, L. — 1903. “ Bryozoaires provenant des Campagnes de rHirondelle.” Res. Camp. Sc. Prince de Monaco, Fasc. 23, Bryozoaires, pp. 23-25. 14. Lankester, E. R. — 1874. “ Remarks on the Affinities of Rhabdopleura .” Quart. Journ. Micr. Sci., XIV, pp. 77-81. 15. Lankester, E. R. — 1884. “ A Contribution to the Knowledge of Rhabdopleura .” Quart. Journ. Micr. Sci., XXIV, pp. 622-647. 16. Norman, A. M. — 1894. “ A month on the Trondhjem Fjord.” Ann. Mag. Nat. Hist. (6), XIII, pp. 131-132. 17. Ridewood, W. G. — 1907. “Pterobranchia; Cephalodiscus .” Nat. Ant. Exp. (“ Discovery ”), Nat. Hist., Vol. 2, London (Brit. Mus.), pp. 12-20. 18. Sars, G. 0. — 1872. “ Some Remarkable Forms of Animal Life ” I, Christiania Univ.-Program for the first half-year 1869, pp. 1-18. 19. Sars, G. O. — 1874. “ On Rhabdopleura mirabilis (M. Sars).” Quart. Journ. Micr. Sci., XIV, pp. 23-44. 20. Schepotieff, A. — 1904. “ Zur Organisation von Rhabdopleura.” Bergens Mus. Aarbog, No. 2, pp. 21. Schepotieff, A. — 1905. “ Ueber Organisation und Knospung von Rhabdopleura.” Zool. Anz., XXVIII, pp. 795-806. 22. Schepotieff, A. — 1906. “ Die Pterobranchier.” Zool. Jahrb., Abt. Anat., XXIII, pp. 463-534. 23. Schepotieff, A.— 1908. “ Die Pterobranchier.” Zool. Jahrb., Abt. Anat., XXV, pp. 405-494. 24. Schepotieff, A. — 1909. “ Die Pterobranchier des Indischen Ozeans.” Zool. Jahrb., Abt. Syst., XXVIII, pp. 429-448. 1-21. BRITISH MUSEUM (NATURAL HISTORY). BF1TISH AHTARCTIG (" TERRA HOVA ) EXPEDITION, 1910 NATURAL HISTORY REPORT. i $ ZOOLOGY. VOL. IV, No. 5. Pp. 103-193. BIRDS. P. R. LOWE, O.B.E., BA., M.B. (Cantab.), and N. B. KINNEAR (Assistant-Keepers in the Zoological Department). WITH TWrENTY-FOUR TEXT-FIGURES AND SIXTEEN PLATES. LONDON « PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM SOLD BY B. Quaritch, Ltd., 11 Grafton Street, New Bond Street, W.l ; Dulau & Co., Ltd., 32 Old Bond Street, London, W.l; Oxford University Press, Warwick Square, London, E.C.4 and AT The British Museum (Natural History), Cromwell Road, S.W.7 1930 [All rights reserved,] Price Twenty Shillings. [Issued, 26 th July, 1930) British flDuseum (IRatural Ibistorp). This is No. • of 25 copies of “ Terra Nova ” Zoology , FI0/. /K, No. Birds , printed on Special paper. CM BIRDS BY P. R. LOWE, O.B.E., B.A., M.B. (CANTAB.), and N. B. KINNEAR, Assistant-Keepers in the Zoological Department. WITH TWENTY-FOUR TEXT FIGURES AND SIXTEEN PLATES. PAGE 1. Introductory Note ........... 103 2. List of Species ........... 104 3. Index . . . . . . . . . . . . . 191 I. INTRODUCTORY NOTE. During the voyage of the “ Terra Nova ” observations on birds were made and recorded by several members of the Expedition, but chiefly by Dr. E. A. Wilson. Had he survived he would no doubt have produced a report as valuable as that which he had previously written on the birds of the first “ Discovery ” expedition. On the return of the “ Terra Nova ” Dr. Wilson’s notes, together with the specimens collected on the voyage, were placed in the hands of the late Mr. W. R. Ogilvie-Grant, and under his direction the plates now published were reproduced from sketches by Dr. Wilson. Once again, however, death intervened, and it is only now, after many years, that Mr. P. R. Lowe and Air. N. B. Kinnear have been able to bring together the material left by their pre- decessors and to add to it the results of their own studies. The delay has, however, made it possible to institute careful comparisons of the rich collections of Antarctic and Sub-antarctic Birds now in the Museum with those in other Museums, especially those of Edinburgh, Tring, and Paris. The results of these comparisons, set out at length in the tables here given, have enabled important conclusions to be reached regarding some supposed sub-species and their geographical distribution. W. T. CALM AN, Keeper of Zoology. iv. 5. 1 104 “TERRA NOVA” EXPEDITION List of Species and Sub-species obtained or recorded 1 . Aptenodytes forsteri. 2. Pygoscelis adeliae. 3. Sterna sp. 4. Gygis alba alba. 5. Catharacta skua inter cedens. 6. Catharacta skua lonnbergi. 7. Catharacta skua maccormiclci. 8. Oceanodroma castro. 9. Oceanites oceanicus. 10. Fregetta tropica melanogaster . 11. Fregetta tropica tropica. 12. Bulweria bulwerii. 13. Puffinus tenuirostris brevicaudus. 14. Calonectris kuhli edwardsii. 15. Adamastor cinerea. 16. Thalassoeca antarctica. 17. Priocella antarctica. 18. Procellaria aequinoctialis. 19. Pterodroma lessoni. 20. Pterodroma arminjoniana. 21. Pterodroma mollis mollis. 22. Pterodroma maeroptera macroptera. 23. Pagodroma nivea nivea. 24. Macronectes giganteus. 25. Daption capensis. 26. Halobaena caerulea. 27. Prion sp. 28. Pelecanoides urinatrix exsul. 29. Heteroprion desolatus alexanderi 30. Diomedea exulans. 31. Diomedea epomophor a. 32. Thalassarche melanophris. 33. Thalassarche chrysostoma. 34. Thalassarche chlororhynchos. 35. Phoebetria palpebrata. 36. Phoebetria fusca. 37. Sula sula. 38. Sula leucogaster. 39. Fregata wilsoni. 40. Fregata minor nicolli. 1. Aptenodytes forsteri (Emperor Penguin). Text-figs. 1 and 2* Aptenodytes forsteri Gray, Ann. Mag. Nat. Hist, xiii, p. 315, 1844 (77° S., 180° E.) ; Ogilvie Grant, Cat. Birds Brit. Mus. xxvi, p. 626, 1898 ; Wilson, Nat. Ant. Exped. “ Discovery,” N. H. ii, Aves, p. 1, 1907. MATERIAL OBTAINED No. 62. £2.iv.l911. Cape Evans. No. 64. ^ 31.i.l912. Found on berg off Cape Evans. Moulting. No. 66. $ ll.ii.1912. Cape Evans. No. 70. Cape Koyds. Four unlabelled specimens from “ Terra Nova ” Expedition, 1911. * It should be mentioned that many of the text-figures have been reproduced from rough sketches in Dr. Wilson’s diary. In some cases it has been difficult to determine the species actually represented. Probably there was doubt in Wilson’s own mind, as any one will understand who has attempted to identify birds at sea. BIRDS— LOWE AXD KINNEAR 105 Dr. Wilson wrote very fully on the Emperor Penguin in his report on the “ Discovery ” Expedition, and the few extracts from his Zoological diary given below are all notes he recorded during the present expedition. Bee. 9, 1910. — 65° 8' S., 177° 41' W. A solitary one in its first or second year on an ice-floe at 3 p.m. Bee. 19, 1910. — About 67° 50' S., 178° W. One young and one old bird. Bee. 29, 1910. — Two young birds seen. One had the throat blackish-grey and the neck whitish ; the other was younger, throat white and neck-patch grey. Both lopped into the water at once o£E a floe. I also saw one adult alone. Bee. 31, 1910. — 72° 54' S., 174° 55' E. Open water and hove to under the lea of the pack ice in a southerly gale. One old one alone and four young ones in the pack ice, also one other young one all fairly full-grown. Jan. 3, 1911. — Numbers increasing. Jan. 4, 1911. — Five miles E. of Cape Crozier. Came up to Ross Barrier about five miles east of Cape Crozier and then proceeded west. We did not see one Fig. 1. — Emperor Penguin (Aptenodytes forsteri). Adults, feeding and swimming in the water. alive, either adult or young, until we attempted to make a landing at this junction of rock-cliff and barrier, when we found a half-grown young one in partial down standing on a pressed-up dirty piece of sea-ice, about 6 feet above water and near it an adult Emperor asleep. [Text-fig. 2.] The chick was larger than any I have seen in the down, and had already begun to shed the down on the head, fore-breast and back(?) and the wings were quite free from down. The adult looked very sick of life. The chick walked away from the edge of the ice. We might have knocked off and so have obtained it, had there not been so much swell that any attempt to land or even approach the edge of the ice was impossible. The ice on which these two birds stood was raised well above the sea and was bent upwards so as to form a sort of cave in which the swell rose and fell. The ice was very dirty sea-ice, a mere remnant of the old winter’s bay-ice, and hanging beneath it were the remnants of several dead Emperor Penguin chicks as well as an old one. The toes of these dipped every now and again in the swell. Here and there dead Emperor chicks were floating in the open water, evidently washed out of old bay-ice. [See Text-fig. 2.] 106 “TERRA NOVA” EXPEDITION This was the only visible sign of the Emperor Penguins, but it established a new fact in the life-history and falls in with what we had believed, namely that the chick is very slow in shedding its first plumage and the egg has to be laid and hatched out during the winter months. After landing Captain Scott and his party, Commander Pennell returned to New Zealand in the “ Terra Nova,” and during this and subsequent voyages he and other members of the ship’s company made careful notes in the zoological log of all the birds seen. On January 31, 1911, in 78° 29' S., 170° 56' W., there is an entry, “ counted about twenty at the head of Discovery Bay with Adelies, which were seen in the water but not the Emperors. On shore they were intermingled. Penguin-tracks were numerous.” When near Cape Colbeck in 76° 56' S., 159° 01' W., on February 2, many were seen but no more were observed after the 26th of the month, when — “ Four were seen on the floe ” off the N. coast of Victoria land in 68° 57' S., 158° 53' E. During the second voyage south in the following year, the first Emperors were met with on January 1, 68° 44' S., 178° 55' E., three being seen “ with no yellow on the throat.” On the 6th in 74° S., 171° 18' E., another immature bird was noted and from that date onwards odd birds were recorded daily till 76° 03' S., 165° 55' E., on January 10, when an adult female was captured. On February 23, 75° 43' S., 164° 20' E., the following entry appears in the log : “ Stopped in heavy pancake ice off Drygalski Barrier Tongue. Twelve Emperor Penguins in a group standing on the ice right alongside. All had yellow necks and were very large birds. Seen in the water later BIRDS— LOWE AND KIN NEAR 107 they came right lip imder the counter, making a very distinct blow like a seal, but not quite so loud. They kept their beaks open all the time their heads were out of the water, which on this occasion was never more than five to ten seconds, although their bodies were sometimes on the surface much longer. Altogether, in the afternoon and evening, well over a hundred Emperors were seen. All close enough to see well were adult birds with yellow necks. Their plumage looked quite good.” The last examples of this penguin were noted in 75° 20' S., 166° 30' E., when two were seen on March 6. On the final voyage south no Emperors were met with till 70° 39' S., 166° 17' W., was reached on December 30, 1912, when three were seen early in the morning. Next day three adults were observed close to the ship, three more were seen on the following day, and on January 5 in 71° 48' S., 166° 48' W., an adult was noted “ yellow neck, moulting. In a very unhappy state.” Three days later (71° 41' S., 167° 04' W.) the entry runs : “ Three Emperor Penguins together on a small berg. One, the largest, in middle of moult, the other two clean looking. All with yellow on neck.” Apparently all adults. On the 13th, “ One with yellow on neck, one young without sign of yellow,” an adult and immature bird, were noted, and next day in 72° 45' S., 172° 51' W., another immature example is recorded. There are no further entries in the log after January 15, when the “ Terra Nova ” was in 73° 48' S., 177° 15' W. 2. Pygoscelis adeliae (Adelie Penguin). Plate XVI and Text-figs. 3-6. Catarrhactes adeliae Hombron and Jacquinot, Ann. Sci. Nat. (2), xvi, p. 320, 1841 (Adelie Land) ; Pygoscelis adeliae Ogilvie-Grant, Cat. Birds Brit. Mus. xxvi, p. 632, 1898; Wilson, Nat. Ant. Exped. “ Discovery,” N. H. ii, Aves, p. 36, 1907. Material obtained No. 49. 15.xii.1910. 67° 23' S., 177° 58' W. No. 50. ? 13.xii.1910. 67° 28' S., 177° 58' W. No. 51. 15.xii.1910. 67° 23' S., 177° 58' W. No. 54. 22.xii.1910. 68° 41' S., 179° 28' W. No. 55. 22.xii.1910. 68° 41' S., 179° 28' W. No. 56. 22.xii.1910. 68° 41' S., 179° 28' W. No. 57. N 6h O to to to &2 H ^ 02 ft to o ft to 2 to to to 6s to to O CO H to to a to to to ai < to to Eh « to <3 CO « to to 02 I— I to to <3 to to to < to to to i-O m 13 • ~ o piS £ rP o Ph 4^ rQ © © <+4 r*l tol ^ . c3 rj ^ I CQ r2 2 ■c ^ ^ &» ^*0 lO o ?i S S - fi :i fi ^ ^ fi ?i t*h 04 C5ONHQ0l>OHt'fliMX)O ^ o LO o ^ ^ o lO ^ ^ ^ ^ ^ O CO H< o 1.0 o o o + + + + ^OXiOOhhNOOCOOW O >0 O >0 OQOXcCXhhhhDCOX Q 00 CO 00 COCOCOCOCOCOCOCOCOCOCOCOCO CO CO CO CO CD CO 05 H o i> oo oo CO CO CO CO o o CQ 03 • P • p P .O 02 43 oo ? P p © • o =3U o B © P ? cq o ^ K ^tj H . :Z. I'g : c S 3 1^1 I9to £ «.s g s „ . > £ CQ - 2 73 o M »coMc3 8 8 a to E3 <5 o o . p . eg p ; o J? to © ! ft 1 'S +S S' CS o to rq 02 to £to~j d d © § -to ft c I c3 . © to ^0-3 w H . jq « fc) < O #p 6 GO s d to « to d to : W to 1 I "3 to - •> g HH ce to ft * to S d torss^:sir;isi J .. , "so £Z} M »S #, r- P ^ 9 Ph H O o3 P P £-P bD »o o l> 00 01 04 05 > CO H H O O 22§ .£ £ >* ©4 ©4* f— i ©4 (M OI CH-OfOOf Of^QO^oOf O Of^oOf^ Oh fQOfQ ti s rH P p »o P CO rP yi P rP fH CD P 0 43 P CO *43 iO rH © oD r C 04 © CQ * CD © CQ 00 43 O © 0) £ ^ CO CO 43 O £ ^ p 04 P "04* P P-H 0 CD O lO ksH 04 trH in IrH 04 04 GQ 04* bo i-H bO »o bJo'Zj bD 'H 1— 1 P * 05 £ 00 .3 ^ 05 •S 00 •S - 04 ^ 04 cq 04 04 ^ 04 g * H 00 hS ft ‘tH H 05 05 05 10 o tH a CO £ © P <5 Ph O w 73 o o O Pm e3 o 00 cp CD 117 British Museum specimens denoted by registered numbers. | From feathers on forehead to tip. $ Greatest depth at base. MEASUREMENTS OF CATHARACTA SKUA CLARKEI. 118 TERRA NOVA” EXPEDITION 5 s' a°S « c6 ft H t'fflNt'MiNIMTfOOaHCOOO 5 a &« 0'S J3-: cP5 <13 o 1 o H * + OHWtlffiHH®10000«fl rH rH rH rH rH 0> r-H <0 rH rH <0 <0 (N & a o3 © PQ o £ CO ^ g2§2§gg§§§§8§ 3 -rf 2 :a- *■ lO 'B 00 ’ ffl.jffiSSSiafflQffl® B X X :a X X X X ooooooooooooo I ft) o o 02 "a >> o P5 C5 Ph o H C5 w H p O 02 §°2 Hi 2 Is O'o PS ^•s 3 A 42 OT T? ■*FP a 2 £ o * Pi P 42 & 'U A4 -p> P CD c3 © T3 “ GO P c« pQ 5 P O U1 o rO ce W ce CQ H co IXNCOIOIOO^h CO0 + + ‘-*3 B • • . . ?g © • • . . a> ft ft X W ■si i 1 * § M © . OPh w d h aj n CD ri 43 111 .!* 6 - 5 . cs § M M © . . OOPh © > o O ^'S o ce a xi fe ^ e_. H o«H bo „r5 d 3 ^ 49 .9 02 H A rH • r- < •a 'B 'B lO ©4 pH io 16 tH t> t> I> CD i> O CD CD o re a do i© W « O O Tt< 1© TtH 1© *© rt< 05 CD O CD l> ^ ^ lO ^ LO o O CO 52 CM l> pH i© ^-h — < -h O 05 ^ 05 00 00 ^- ^ TtH rt^ CO ^ CO CO CO CO jH q 2 3 ■g o © S © q -q S q © c3 ^ ► ^ ^ © c3 MS - * - "SB d w d pq ■nj t-j < q c3 Ph CO ai n tj g a A c3 co ■3 M © a >» o -G> © p O § £ U1 H T5 a © -a S I § p-< a} o TJ C cS a. - - Odd d d d 05 05 05 ©i H H H cod :d :d :d g 2 MHh .q cd ©i ^ .q ,rH. rH I-H CM X CO CO CO d d d ©i 05 05 CO 05 1 o oo oo O O 05 O Q 05 C5 05 © a fee fcj o fiq o eg fe5 Eh fee ^-H p LB CO S Eh o s ce s ^3 Eh S O P o XJ1 H X; p 3 m Oh P m < P m H S CS3 O S3 o Q fee o lO 00 h tH A W5 .g 8 Ph P a q q CJ o § GG O rO P c$ w a o CO ID CO S §) p w q O 05 o o o co tJH 00 oo Tt< 00 00 05 00 -t- GO rH !> GO •p o 05 r> GO > t> H CO rH r—l o ,rt d CD .q ■P CM >© — 1 H CH-O-f-O O O O O o 2 g 5 CD co ^ 2 00 ^ ^ ZD 2 &()l^ C5 r- 1 o -q ^ 05 00 h d h Ol 00 t> ^ O GO QJ 00 ( q cd I g' CD © O h O 00 rH GO C5 00 00 *> i C© i — i ^ l> CD i bO A J 120 ‘‘TERRA NOVA” EXPEDITION dg3 COM ^Qt^l>P*GOCO ^GO^COOCJOQOOCOCOt^fMN OOl>GOOOOOI>l>I>l>t>I>I>GOOO O Tjl M 00 GO l> 00 4-3 — p-pq fi-s CD > 4-3 P3 LO CMHIOO50OH LO lQLOiOlQlO»OlOlO (N 00 CO »C LO lOTfWHOO^CO CO 00 S H a l> l> l> CO CO rH P I-H nH rH rH !— I P „ rH LO lO P «,►«,*»,#> tH GO o o o d o CP CO - LO LO d r~l CO S ^ rH CQ '- CO ^H CO CO CO CO CO co p CO rH rH o D CD P 00* LO r-H c4 oi oi c Tjj l> lO LO io LO LO rH r-H rH rH CO CO pH oi CO o o :0 o o 50 CP CP CP o •g a> CP - LO CP hH CP CP CP CP CP S " CP 00 oo 00 CP s 00 00 GO 00 GO pH rH s-H rH rH rH rH rH rH rH EH i-h iH H rH rH rH 0 a bo S3 BIRDS— LOWE AND KINNEAR 121 SUMMARY length of Wing. length of Bill. Depth of Bill. length of Tarsus. Falklands Islands . . 394-365 52—17 21-19 69-66-5 Tristan da Cunha and Gough Island 396-380 55-50-5 24-20-5 75-5-70 South Orkneys 422-401 56^:6 — 79-71 South Georgia 422-395 59-49 24-22 77-71 South Shetlands 415-375 52-46 23-18-5 75-63 Kerguelen . . 422-395 57-5-49-5 24-20 78-71-5 New Zealand area . . 447-395+ 59-48 26-5-21 84-74 6. Catharacta skua lonnbergi (New Zealand Skua). Catharacta antarctica lonnbergi Mathews, Nov. Zool. xviii, p. 212, 1912 (New Zealand Seas) ; Megalestris antarctica (partim) Saunders, Cat. Birds Brit. Mus. xxv, p. 319, 1896. The extracts which are given below from Dr. Wilson’s and Commander Pennell’s diaries we consider to refer to the New Zealand race of the Great Skua, which, as far as we know, has not been observed so far south before. Sept. 25, 1910. — 46° 30' S., 116° 07' E., South of King George Sound, S.W. Australia. One flying round the ship early this morning for an hour or two. Nov. 29, 1910. — Off Port Chalmers. One seen. Feb. 20, 1911. — N. of Cape Adare. Heavier birds than McCormick’s Skua and darker altogether, but owing to the great variation in the latter cannot be logged with certainty. Feb. 21, 1911. — 68° 41' S., 168° 29' E. Seen. Easily recognised by their heavier build. Feb. 22, 1911.-69° 10' S., 164° 30' E. Seen. Feb. 23, 1911. — 69° 29' S., 162° 48' E. Four seen together on and above one floe. March 18, 1912. — 64° 03' S., 160° 12' E. Larger and darker than McCormick’s. Bee. 15, 1912. — Off New Zealand coast. Seen. Bee. 17, 1912.-49° 12' S., 178° 14' E. Three seen. Feb. 6, 1913.-54° 22' S., 164° 49' E. Seen. Feb. 8, 1913. — Off East coast New Zealand. Seen. 7. Catharacta skua maccormicki (McCormick’s Skua). Stercorarius maccormicki Saunders, Bull. Brit. Orn. Club, iii, p. 12, 1893 (Victoria Land) ; Megalestris maccormicki Saunders, Cat. Birds Brit. Mus. xxv, p. 321, 1896 ; Wilson, Nat. Ant. Exped. “ Discovery,” N. H. ii, Aves, p. 64, 1907 ; Catharacta maccormicki wilsoni, Mathews, “ Birds of Australia,” ii, p. 495, 1913 (Coat’s Land). MATERIAL OBTAINED No. 53. December 22, 1910. Pack-ice, 68° 41' S., 179° 28' W. A. Cherry- Garrard. Weight 3 lbs. — . G Cape Adare. 20. xi. 1911. G. Murray Levick. 122 “TERRA NOVA” EXPEDITION No. 136. £ Cape Evans. 31.xii.1912. A. Cherry-Garrard. Weight 2 lbs. 14f oz. No. 137. — „ 31.xii.1912. „ „ „ 2 lbs. 12 oz. No. 138. „ 31.xii.1912. „ „ „ 3 lbs. \ oz. On December 7, 1910, 61° 22' S., 179° 56' E., Wilson writes in his diary : — “ One with a distinctly paler body than in C. antarctica and rather smaller in size. I feel pretty certain as to the identification.” Five days later in 67° 28' S. another was seen of which there was no doubt about the identification, and there are notes of single birds up to January 4, 1911, in 75° 3' S., 173° 41' E. When five miles East of Cape Crozier on January 4, Wilson notes McCormick’s Skua was very abundant as was the case also in McMurdo Sound, and between January 6 and 11 he made the following interesting observation at Cape Evans : — “ Very abundant, some still hatching eggs, and I found two in one nest. Some had young chicks just hatched or just hatching. One nest had one young chick hatched out and one just breaking out of the egg the day we landed. This was at Cape Evans, which we used to call the Skuary. The birds were as usual bathing in the freshwater lochs there.” Commander Pennell has a number of records of this Skua in the zoological log of the “ Terra Nova.” After leaving Cape Evans it was seen occasionally up till March 4 in 67° 11' S., 160° 47' E., S.W. of Balleny Islands. On the voyage south again in December, 1911, there is an entry in the log for the 27th, 64° 56' S., 175° 30' W. : “ Skua Gulls — Appeared to be large and dark, probably Antarctic Skua Gull,” which doubtless refers to the previous species. From January 3, 1912, when off Cape Adare, till January 18, in 77° 21' S., 164° 12' E., McCormick’s Skua was seen almost daily, and on January 20 the entry in the log is as follows : — “ The noise made by the Skuas as they quarrel over food is almost identical with that of a flock of ducks in a farmyard. McCormick’s Skua and Adelie Penguins are always with us, and therefore not logged.” The last bird seen on this voyage was on March 16 in 66° 44' S., 164° 48' E. On the final journey south there are only four entries in the log, and all in latitudes in which this Skua was seen on the previous voyages. Systematic Notes. — Two examples of this Skua were obtained by the “ Scotia ” Expedition in the Weddell Sea and these Mathews has described as a separate race — C. s. wilsoni — on account of their smaller size. We have examined these two birds and M. Berlioz of the Paris Museum has sent us the measurements of seven obtained by Dr. Charcot’s Expedition in the South Shetlands. As will be seen from the accom- panying table they are, on the average, slightly larger, not smaller, but the difference does not seem to us sufficient to warrant separation as a separate race. MEASUREMENTS OF CATHARACTA SKUA MACCORM ICKI. BIRDS— LOWE AND KINNEAR 123 — Cl O C\J Ol C<4 r-i ^ OQ --XXXDOO^OXX L^O'^T‘T^TtlTtllOLCTt,lO'^'^ - tt Mr + -f C D X H r I O CO O O — C — <00C1C1©O^hOC1^-< do do dd£d < w c w m m rJi O Ph m * § o - G -22 fc-r ■ — . .S 5 " > -22 © --< C "3 o - « c3 ''PH o g y. _ P G CC 6-t fe c3 > o d o ^ c3 « 0 <3 d ◄ 53 d ^ d ft d d "S ^ ■s< ■d a x §< ra ■ © Th Ch © : ^ P x"t5 H §- r - X x1 'd 63 2 .-h a, a x ft - H GO W GC - DO H GO „ j §» £ g K ^ O N Sh O OT © • — E-i Q C5 ■ > O £ ' Cv o " y ~ 6 £ coso £53 £53 - = 6 3 £ O © O _© cc H gc H CO o p ^ OJ s 2“ _ > r3 0^5 > -3 > -h ._• ft h r— i :a ,_, :a 2 r x. ft ft X ft ft V, .A x ft :a -I iji ia « :a -I x ^ :a xmhhim x oo :o co — in ’"o’io^^o^o^OfafOfCH-afa)- I> l> CO Cl f-H l> r- 00 04 I> 3 o rt WOMhOhO «(Nrt«NNINN fflwooioio®® © GO CO OOODhhhh ^h©00 '3 .® a £ J . • ® si ft ^ 72 S X .2 8 :.h a >i ftc ® s ® » „ q o ft 43 £ % jti JD © -T-l -ft • w ofibPOf^ o ci o 2 C3 $ POP §§!!§-§§ •->■•—< . a i Hr £• O 3 o 'I •". 00 00 00 l> l> CD X lO X IO CD H CO CO CO CO CO CO CO co CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO •g g O O 00 05 0 00 00 CM 0 0 05 0 00 00 0 0 0 05 0 05 l> C5 05 r>* OS CO CO CM CM CO CM CM CO CO CO CM CO CM CM CO CO CO CM CO CM CM CM CM CM oo cm WWW o 00 CM •— X 00 00 CM CM S I I s IN MS t'OOCi co co co co 43 .2 * H3 d eg PQ 44 o eg 2 "O d eg ^3 £ .3 .X £ £ <3 H T3 eg ■73 ;a H 43 «*2 d o 02 o o o o CO CO CO CO * "3 d eg 44 MS Pn J- 3 03 d eg 44 d s oo o ^ 9) o ^|t3 ^ § § *•— « *2 W d S © M J5 O eg« -4> PQ T3 44 •d o oo eg t-< 2 G ■d © 2 ° O 4-» o d £ 4» 43 bp co o O -£ - eg JQ G p" eg 'O PQ * ° £ •g S . 1 is g£ w © • TJ 44 ■gJ G 40 4-> a © ° -2 i " of ^ -O O © 4-< ft O O J h J *11 SIS « a 3 i & 03 MH O * O C . • 2 -o i: ■“ „ 3 o S3 3.2 * £3 0 0-3 S *> O P _ g J3 £3 ■g SJ § cd -X TJ -j> « A ~ j* jh *T3 G G d eg eg *SS 44 d s 44 44 d H eg Oh O * © bO TJ g ^ ^ I S c; o 3 a bga< H d'« 6D Ja a”s | § & S“ « -x-g g'l .si £ l© pd 5S rd bD O bfi © 3 g* e3 c3 >> O OH o G :g 4 a . ® • * o SPh" h3 £3 TO H „.X •> X I 1 »• •'73 •> *- § s U1 C o 05 :£ o 05 CD fc*-J -.0 * 1C o 05 M f-H CO O <^)0*0*00+0*0*0^0,000fc040*0 HNMH HHHW O rH IO H H H Cl A> -X a jo P4 (H - c3 Q p * o u « jd -X jq ■“ S:«b £S^° ~ 3 0 jh 2 o ft. £3 SQ <3 « 3 J4 O O •« « X 0 3& * ^ o -b-2i 5? : ^ jd jb j< ^ 5 O G G 4f -2 55 S5S^ ^ £ 9 H o ' :1 a 0 PQ w 1 « !« H ca O H X H *4 x H « JO &0 3 X O a o -a not aft. Also on the following day, in 40° 58' S., 34° 41' E. : “ One or two. One was astern to-day with a black Macronectes, and a Majaqueus. The three sizes were very distinct.” On September 6, 1910, Commander Pennell mentions : “ Two or three seen, have only seen single birds before, since leaving the Cape.” And while doing survey work on the “ Terra Nova,” in the winter of 1911 (July-September), off Three Kings Island, New Zealand, there is frequent mention of this petrel in the zoological log. There are also several other entries seen at different times in 1912 as far south as 59° 28' S., 169° 33' W., on December 24, and on January 29, 1913, in 69° 56' S., 170° 52' E. 23. Pagodroma nivea (Snowy Petrel). Text-figs. 18 and 19. Procellaria nivea Forster, “ Voy. Round World,” i, pp. 96-98, 1777 (51° 50' S., 21° 3' E.) ; Pro- cellaria Candida Peale, U.S. Explor. Exped. viii, p. 295, 1848 (64° S., 104° W.) ; Pagodroma nivea (Gm.) Salvin, Cat. Birds Brit. Mus. xxv, p. 419, 1896 ; Wilson, Aves, Nat. Ant. Exped. “ Discovery,” N. H. ii, Aves, p. 88, 1907 ; Pagodroma confusa Mathews, “ B. of Aust.” ii, p. 177, 1912 (Cape Adare). MATERIAL OBTAINED No. 16. $ 14.xii.1910. 67° 28' S., 177° 58' W. E. A. Wilson. “ Shot in pack.” No. 48. 15.xii.1910. 67° 28' S., 177° 58' W. E. A. Wilson. Weight 9 ozs. No. 79. 10.xii.1910. 66° 38' S., 179° 04' W. E. A. Wilson. Shot in pack. Bill black, iris dark brown, legs and webs blackish-grey. The first Snowy Petrel met with on the voyage south was on December 8, 1910, in 63° 20' S., 177° 22' W. ; the following day a considerable number were seen on the 10th in 66° 38' S., 178° 47' W. “ Five or six were round the ship all day. Those we shot exhibited the differences in size, weight and bill very well indeed ; some had the TTt 1 7 _ D-rntr fri aaJ BIRDS— LOWE AXD KINNEAR 143 bill large and thick, in others it was small and finely made.” This petrel continued to be seen daily, and up till January 4, 1911, five miles East of Cape Crozier, we find the entries, such as “ a few,” “ several all day,” a small flock,” in Dr. Wilson’s notebook. In the “ Terra Nova’s ” zoological log book there are similar entries to the Fig. 18. — Snowy Petrel ( Pagodroma nivea). Feeding among the “ brash ” ice. above till 64° 23" S., 161° 39' E., was reached on March 8, after which none was recorded. On this voyage several interesting observations were made which are worth quoting : — Feb. 2, 1911. — Near Cape Colbeck. A Snowy Petrel settled on the water, then dipped completely under for a quarter of a minute with almost closed wings. Feb. 13, 1911. — 50 miles N.W. of Cape Adare. One settled on the water to pick up something thrown from the ship. It settled for about two minutes, but kept its wings extended above its back the whole time until it rose again. Feb. 28, 1911.-68° 14' S., 160° 38' E. Snowy Petrels settling on the ice in small flocks, walking in a very ungainly waddling fashion, and shuffling about a lot before they nestled down and were still. March 4, 1911.-67° 11' S., 160° 47' E. In the afternoon two Snowy Petrels came on board, one being so gorged that it could hardly fly. On being picked up it vomited Euphausia and four large eggs. Some flying about 200 feet above the ship were twittering almost exactly like Linnets, etc. They twitter when they fly in flocks fairly high up, in winter : this is the first time I have heard this note. Fig. 19.— Snowy Petrel ( Pagodroma nivea). Studies in flight. When the “ Terra Nova ” went south in December, 1911, Snowy Petrels were IV. 5. 6 MEASUREMENTS OF PAGODROMA NIVEA. 144 “TERRA NOVA” EXPEDITION p o a S P ®-u> t»£S O P kS 2« o ll o O O 00 00 GO ^ Tf< CO CO CO l> o CO ^ IO *? /yv CO O ^ CO T* Cl Cl r* ^ Tj< ^ ^ CO ^ CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO £5 CO CO IO y* ~ Cl CO ,*01 CO /»•«> Ml AT M AM CO t> o o o C3QQOO (M d Cl yf CO CO wo wo CO *.0 I -C io IO Cl Cl Cl Cl Cl Cl Cl Cl Cl S p'se io cb io IO CO Cl Cl - io Cl © Cl rH rH rH rH »o coo 3oS =3 o o w £ p M w o (3 , 2 .. • IJJSB g^3 a ^ o o o « 0Q 00 8 p „ X * w 5? d d d d d T3 45 CO d d •*P5 d .ti) 3 - O w c o : ti) P M w >• w *T3 T3 : <3 S co a tn £ w 45 o P P i T3 i H3 X x w w H3 05 p *4 *T3 45 Si •< 45 P o H "O 45 P M W w P M W w w K W to 45 H & H M W. o o , b K w o * eo TJ C C d ► X W W T3 a g o P< X w o d ® d £ 30 M W Hi . . ^ 45 rd 3 O wSfl •2 p o , X Sd« djfb , 03 ' io.g 3 C o o 02 a> e £ "O > E > o o cs 2 i. a : ' «2 . js £ - rt ~ 5 =s 2~ go - v v HH( i V-V& §§u = = a 5 ” i. „ 3 3 o I -a J= O 4> (C o t- ■- rt HSS Kr: - H M 2 .s-a £ > o 3 sa w T3 • o 'a p g x o< d ~ a b & £ ej O H O ll PQ - ^ W. * o W fc T3 d d 0 3 rQ • O w p ►W cd F o S? * d W H ’ • c t: c . o °C . 0.3= X > W1^ .S < ■ H £ W t • O 33 .H g. ■ = M a s : “ b § > 5 g . s.a • M P ■ m . o T3 d d w * ^ o o d 1 o *« Maj OD v goo «r wT t. d T3 o P d O IO ^ ^ g COM^ « 00 ^ / IO Cl © xn Zd o o ^ q c Cl !> eg ^ co CO CO O H p . £ 2 .b o< IO o o . HH p l> c6 o _ • 2iow GO ■ T3 • C s r-pM : 03 lo o "* I> ° r-I l> 05 P i — ico v r v°'n § o CO ® >o i> W • v . — >CO o o Ph d T3 <1 05 P d O CO v I'* io 1 ° d S S v A_d ►05 ‘3 O dH 5 Sh § ■*“ S 0> ■§2.1 • 2 r ji W goc § W P . o GO “ CO • ® c ■j: ^ d v 2 “ oo r<3oj Cl CO GJ (75 ° ° o i> °o <* R co co O K ■ W T3 ► ^ § ‘ 2 w . § g §3 P 05 © ^ ® P .2 O a >PiO C5 05 00 S rl GJ 05 00 05 00 o i-H 05 Cl o o 00 Cl o Cl o Cl s 05 05 T*H o Cl o 05 Cl o 05 Cl o Cl o 00 05 X ci n 05 00 * ‘x :d 1 X 1 05 X X 05 Cl GO 00 r-H •r« 05 rH 05 00 rH 'x 05 2 X rH •rH i-H Tj3 00 i-H i“H Cl rH Cl CD rH 00 r— < CO X f— 1 CO rH Cl CO f— 1 *-* Cl Cl rH rH Cl t*3 CO rH X 05 *o *o *o | I o 1 o+ o o 1 o CM- *o O o *o 1 1 zn S n ^ s ^ ^ 2 00 GO 00 O C5 ^ C5 2< io cb oo oo o H IO H rt H H C5 ^ 5 § :d oo f-i !hXh t>rHH *0 0*0 0*0 a : d o m P a -2 ® .a P CO o O o eo «33 O o «s bJD d | I 00 S g O g CO CO C5 ir> C5 • • OT co CO O 3 © 3 i-h eo a g a g g 3^ co ^ci io • p H .3 rH O h2h22 Cj io* cj 05 I> CO IO 00 CO Cl IO IO o o 05 05 ci d w § oo d r"1 bO 5 1 00 ' hH CO CO CO io o 05 oo a IO H ” g O CQ d S.S 2h 60 . W « ” 06 w i— 1 l> |> r-H f— I CO i-H MHCl Cl I— 3 M Tji rH ^ O O ^ 05 00 05 05 00 00 io co co Cl i b o 05 ^ a ®3t>n cs ej j os os -h ^ O K CO CO ^ OM SrtHO N. ei S sq oj N. *2 ^ bo^ ^ «J CO CO cNNCD O -< -g Tj( rt oo^aooa a . d I 05 CQ Cl d io io CO CO 05 Cl IO > ^ o •d oo 05 hH pH Measured from feathers on forehead to tip. t Measured at greatest depth behind the uarial tubes. MEASUREMENTS OF PAGO DRUM A NIVEA. BIRDS— LOWE AND KINNEAR 145 S Sp 3 2 1 *© w S O c' § ^143*43 8 •— z -P ^ N 2| o GO feC d d pO . 0 >, — 4 c4 c4 : £ § t- 1 d 73 " >» pQ . ^++ o .7 "3 Oh o ^ d c3 •a ^ if • It ^ 2 73 ' ® 3 £ 1 0 c3 £h ^ c3 O 3 * 3 gs a w s ^ § t- 04 04 : id co 04 id O © O 04 i> r- 04 04 + OOOCIC5 _ dX © © - rt* fl n W Cl ^ 04 I CO I W 04 LO © CO © 04 CO I O ^ Tt< t' X O | I» 04 04 04 04 04 04 OJ 03 I 2 04 -s —s ©. © Cl 04 04 73 .2 2 0 rQ pd -4-> Ph 0 73 , J3 X c3 OQ 75 O ‘2 & x w ~ a p- X W •o d . O) • ^ '•X a c3 -a O - . c*-. 73 0 Ph • *73 *73 0 «D ^ X K d o w . W'jtW co ~ d doc4 CS • *s a = ■p 0 ;-g^ |^r o : : co 7 7 SjSo : : : > .3 •o 44 , c6 . d . d CO :a « co r> ce 73 0 CD 43 co 73 d d O isH pd c=H 0 P.O S 43 la £ g c< '4H O CO 53 04 73 3 o -4> d i-r* £ d 5 P-i O ^ ° & M-H pH 73 73 CD 0 a a CO CO d c3 CD 0 S S * 43 . d . 0 gg c3 hP r d _, ! 4 § 0 a 1 p<£ ' .2> ui . >b co LO o rf CO -S 04 tQ 4 rd . M * co £ co PP CO ©'CO t-H s*Jh*3 r-o“«g X - CO o 0 CO ^ CO ^ 73 d 73 - d 3 * o c o £ £ d CO CO O J2; P4 o Cl 3 p ( Cl o O o c HQClCl! :d > > > 8 Cl Cl Cl Cl Cl . O O Cl O Cl o 2 Cl Cl O Cl O Cl TfH s8 lo o l> 7 rf T* > R > > d > :7 7 *o CO *7 LO) 7 T* QC :3 > ^ Cl !g2§§o jm m .4 > . ‘7 tJH *7 id • C. M (01 M ' *o *o *o *o *o *o *0*0*0*0*0*0 0 *000*0*00 d £ *> 2 K 3 -e s 0 3 & ; : ; :::::: : : :::.*: g d 0 co d i ! :::::: : : : : : : : S - pd CO a S£ CO o ^ „ „ „ „ _ 2 ^ % O f— 1 IQ p Q -q ^ CQ I 04 ^ r— o 4 .S g co oo os 3 -< '“t PH 2 < : : : 2 d 0 CO d ; .* S ► ra X *43 C'* X H-* CO . o 04 x — l !§j 3 | 22« - a 146 “TERRA NOVA” EXPEDITION again first met with in about the same latitude as the previous year, and continued to he seen almost daily throughout the voyage to Cape Evans and back, the last observed being in 65° 33' S., 161° 37' E., on March 17. Under the date of January 2, 1912, 70° 02' S., 175° 31' E., the following entry appears in the zoological log : — “ Saw several Snowy Petrels, settled on the water with their wings stretched out flat on the surface, splashing and fluttering like sparrows in the dust." Again on February 16 in McMurdo Sound between Butter Point and Cape Evans : — “ Large flocks of a dozen or more birds together. These birds kept higher than usual, flying about the same height as our mast-heads. Odd petrels have been seen from time to time during the last week.” During the third voyage south this beautiful petrel was not met with till 69° 28' S., 166° 17' W., on December 29, as the ship entered loose pack-ice, and the whole time the “ Terra Nova ” was south of that latitude this bird was in evidence, but after 69° 56' S., 174° 52' E., was reached on January 29 it was not seen again. Systematic Notes. — Both Dr. Sharpe and Dr. E. A. Wilson have drawn special attention to the great variations in size of this petrel, but from the material at their disposal they were not able to prove that more than one species or sub-species existed. The birds collected during the voyages of the “ Terra Nova ” do not help in the solution of the problem, and till specimens are collected on the different breeding grounds, it is unlikely that any satisfactory division of this petrel can be made. We know that the Snowy Petrel breeds at Cape Adare, where Borchgrevink found it nesting in November, 1899, and it is also known to breed in the South Shetlands, South Georgia, and South Orkneys, from November to the end of December or early in January. Dr. Wilson’s statement that McCormick of the “ Erebus ” and Webster of H.M.S. “ Chanticleer ” have described the nesting of this petrel is an error. McCormick, it is true, saw numbers of birds on Louis Philippe Land on January 2, 1843, but he was unable to land. Sir James Ross, however, obtained some eggs on Cockburn Island on January 6, 1843. Webster, the surgeon of H.M.S. “ Chanticleer,” mentions that a number of petrels, etc., were breeding on Deception Islands in January, 1892, but gives no description of the nesting habits nor does he actually mention whether eggs were collected. Mathews has divided the Snowy Petrel into two species and one sub-species, as follows : — Pagodroma nivea nivea Forst. Type locality 52° S., 20° E. Range south of South America and South Seas. Pagodroma nivea Candida Peale. Type locality 64° S., 104° W. Range South Pole, Cape Adare to Australia. Pagodroma confusa Math. Type locality Cape Adare, Victoria Land. In the “ Birds of Australia ” Mathews separated the large-billed birds on account of their size. “ Upon sorting them,” he writes, “ into groups according to localities, it BIRDS— LOWE AND KINNEAR 147 was apparent that some other conclusions must be arrived at, as large and small birds were present from the same place, but it seemed that the large ones were constantly larger, while the small were as regularly smaller, and that no intermediates were extant.” “ Careful measurements confirm this; nine birds giving : Bill 20-22 by 9-5-10*5 ; wing 251-263 ; tarsus 30-32 ; middle toe 33-35 ; middle claw 10-11 mm. ; while four others gave : Bill 24-26 by 12-14 ; wing 297-304 ; tarsus 38-40 ; middle toe 39-42 ; middle claw 13 mm. The whole being from Cape Adare, Victoria Land, and there- abouts.” Presumably Mathews based his conclusions on a selection only of the birds collected by the " Southern Cross,” “ Discovery,” and Morning ” Expeditions, and had he been able to include the additional specimens brought back by the “ Terra Nova,” as well as those collected by the naturalists on the “ Erebus ” and “ Terror,” which in many cases, till this investigation was begun, had very incomplete data, he would have been unable to separate the two species so easily. With the exception of a single male from Petermann Island, Graham Land, in the Paris Museum, which has a bill of 24 mm., all the examples from the Weddell Sea, Graham Land, South Georgia, South Shetlands, and South Orkneys have bills of under 23 mm. in length, and it is these birds we presume that Mathews considers to be the typical Pagodroma nivea nivea of Forster, but how it is to be distinguished from small Boss Sea birds — his P. n. Candida Peale — we are at a loss to understand. As we have already said, till further specimens are taken on their breeding grounds, we do not see how any useful attempt can be made to divide this petrel into species and races. That there is more than one species, a large and a small, is highly prob- able, but at present we can but give detailed measurements of the specimens in the British, Tring, Royal Scottish, Paris and New York Museums (see Table on pp. 144-5, in which British Museum specimens are indicated by registered number). 24. Macronectes giganteus (The Giant Petrel, or Nelly). Text-fig. 20. Procellaria gigantea Gmelin, Syst. Nat. i, pt. ii, p. 563, 1789 (Staten Island) ; Ossifraga alba Potts, Trans. New Zealand Inst, vi, p. 152, 1874 (Foveaux Straits, New Zealand) ; Ossi- fraga gigantea Salvin, Cat. Birds, xxv, p. 422, 1896 ; Wilson, Nat. Ant. Exped. “ Discovery,” N. H. ii, Aves, p. 93, 1907 ; Macronectes giganteus solanderi Mathews, “ Birds of Australia,” ii, p. 187, 1912 (Falkland Islands) ; Macronectes giganteus halli Mathews, loc. cit. p. 187, 1912 (Kerguelen Island) ; Macronectes giganteus wilsoni Mathews, loc. cit. p. 189, 1912 (Ross Sea) ; Macronectes giganteus forsteri Mathews, loc. cit. p. 189, 1912 (Valparaiso Bay) ; Macronectes giganteus dovei Mathews, Austrl. Av. Rec. iii, p. 54, 1916 (Sydney). MATERIAL OBTAINED No. 4. $ 18.U912. 77° 22' S., 165° 22' E., McMurdo Sound. J. PI. Mather. Bill whitish horn-colour ; legs and feet dusky becoming almost black at edges of webs ; toes dark horn shading into black. 148 “TERRA NOVA” EXPEDITION No. 5. $ 18.i.l912. 77° 22' S., 165° 22' E., McMurdo Sound. J. H. Mather. Bill whitish-horn colour ; legs and feet dusky becoming almost black at edges of webs ; toes dark horn shading into black. “ Stomach absolutely full of rope yarns.” No. 6. 18.i. 1912. 77° 22' S., 165° 22' E., McMurdo Sound. J. H. Mather. Bill whitish-horn-colour ; legs and feet dusky, becoming almost black at edges of webs ; toes dark horn shading into black. Stomach “ full of garbage, chiefly rope yarns, and one large smooth pebble about the size of a haricot-bean.” No. 7. $19.1.1912. 77° 15' S., 165° E. No. 8. $ 24.i.l912. 77° 05' S., 164° 117' E. “ Stomach quite empty.” No. 9. $ 27.i.l912. 77° 05' S., 164° 30' E. J. H. Mather. Iris dark reddish-brown ; bill pale yellowish- grey ; legs and toes dark grey ; webs whitish-grey ; claws white, tinged with bluish-green at base ; inside of mouth and tongue white. No. 10. $ 31.i.l912. 77° 32' S., 165° 38' E. J. H. Mather. Bill whitish horn-colour ; legs and feet becoming almost black at edges of webs, toes dark horn, shading into black. No. 63. 29. i. 1912. Cape Evans. H. Ponting. Weight 8 lbs. 6| ozs. No. 65. 9.ii. 1912. Cape Evans. H. Ponting. Iris brown ; beak pink, shading into bluish-brown ; legs brown ; webs bluish-grey ; nails black and white horn. Systematic Notes. — The typical bird, Macronectes giganteus giganteus, comes from Staaten Island, and G. M. Mathews regards examples from South Orkney as being identical. We have no specimens from either of these localities, but we have examined a series of six skins from the South Shetlands which may be considered typical. According to Mathews, also : “ At the Falkland Islands there breeds a uniformly coloured dark bird, almost black, as Wilson puts it — which is smaller than the southern bird. Its bill is very pale and clear and — according to Wilson — lemon-yellow,” and this race he has named M. g. solanderi. As far as we know, Wilson never landed at the Falklands nor did he bring back any specimens from the vicinity of these islands, and, as we point out below, birds of all shades are found in these islands. There are no specimens from the Falkland Islands in the Tring Museum and only one in the British Museum — an unsexed greyish-brown bird with deformed bill, collected by Leconte in 1867-68. Macronectes g. lialli is the name given by Mathews to the sub-species found on the Kerguelen Islands. Of this bird he says in his original description : — “ On Kerguelen Island breeds another uniform phase, which is easily separable by its longer and more massive bill, and while shorter in the wing than the Antarctic forms, has the tarsus and toes fully as long. Its general colouration is brown, while all specimens I have examined have more or less whitish faces.” Apparently the specimens which Mathews examined when making the above BIRDS— LOWE AND KINNEAR 149 description were the series of five birds, two from Kerguelen and three from Crozets — in the British Museum, since he had no examples in his own collection, nor are there anv in Lord Rothschild's. All the birds in this series are entirely dark brown, except for the throat and forehead, which are whitish, and with the exception of one female skin from Possession Island the bills and tarsi are above the average in length. This race may perhaps be worthy of recognition. Ossifraga alba, the name given by Potts to a white example of the Great Petrel from Foveaux Strait, is used by Mathews as Macronectes g. albus for birds found in the New Zealand Seas and breeding on the Chatham, and perhaps Antipodes and Campbell Islands. This race, Mathews says, is uniformly dark-coloured, darker than the Kerguelen birds, with a much smaller bill and shorter tarsus. White examples of this race Mathews considers to be albinos, and if that is the case then albinos are very prevalent, since in the twenty-five examples of this bird we have examined in the British and Tring Museums, ten are white with odd black feathers and all have horny- coloured bills and grey legs and feet, with the exception of one bird in the British Museum from Snares Island, in which they are yellow. In Lord Rothschild’s collec- tion there is a pure white bird from Snares Island in which both the bill, feet and legs are yellow, and this bird may perhaps be an albino, but certainly the other birds cannot be classed as such any more than can the white birds found anywhere else. As will be seen from the table of measurements, New Zealand birds are certainly smaller on the average than those from Kerguelen, but neither in colour nor size do they differ in any appreciable extent from those inhabiting the Ross Sea. Birds from Sydney have been named by Mathews as Macronectes g. dovei, “ differing from M. g. giganteus in smaller size,” but no details are given how this sub-species can be distinguished from the New Zealand one. We have measured a series of sixteen birds from Broken Bay, New South Wales, formerly in Mathews’s collection and now at Tring, and though the largest birds are certainly larger than the largest from South Shetlands, the average measurement is Fig. 20. — Giant Petrel, or Nelly ( Macronectes giganteus). Alighting on the ice. 150 “TERRA NOVA” EXPEDITION practically the same. This series was obtained in March, August, and September, and all the birds with one exception are labelled as females, and also all of them are dark except one female, which is white with odd black feathers. M . g. wilsoni Mathews, from the Ross Sea is described as : “of very large size and notably paler colouration, a majority being almost white and to which has been attached the Macquarie Island birds.” The Ross Sea birds, as will be seen from the table of measurements, cannot be called very large ; the largest bill is 101 mm. in length and the longest tarsus 100 mm. as against bills of 103 and 106 and tarsi of 98 and 100 in birds from the Sub-Antarctic Islands and New Zealand respectively. We think, however, that it is a gross exaggera- tion to say that “ a majority being almost white,” a statement certainly never made by Wilson, who in his report says, that roughly he noted between 66° 7' S. and 78° S., Dark birds, about 60 ; Intermediate, 14 ; White, 18. We have had at our disposal for examination, either in the British Museum or in Lord Rothschild’s collection at Tring, a very fine series of this Giant Petrel (for lists, see below), about which various opinions have been held in the past in regard to the significance of the several colour-phases which it exhibits. This Giant Petrel, “ Bone- breaker,” “ Nelly,” “ Stink-pot,” or “ Quebranta-huessos ” of the early Spanish voyagers, is nearly as large as a Wandering Albatross. It combines in its large and rather clumsy person the disagreeable characters of a scavenger, a vulture, and a rapacious hunter of smaller marine birds. It is circumpolar in its area of distribution, which comprises a belt extending roughly as far north as 35° and as far south as the northerly fringes of Antarctica. Within this circumpolar belt the Giant Petrel, as far as is known, does not breed either in Antarctica or on any other continental land, its nesting-places being confined to Southern Atlantic or sub-Antarctic islands, such as Tristan da Cunha, Gough Island, and the Falkland Islands ; South Georgia, the South Orkneys and South Shetlands ; Kerguelen, Prince Edward Island, Marion Island, the Crozets and Heard Island ; and the sub-Antarctic islands of the New Zealand quadrant, such as the Macquaries, Aucklands, and Campbell Island. Two subsidiary zones or belts have been distinguished by some authors in this extreme southern area of distribution — a sub- Antarctic and an Antarctic, both being circumpolar ; while correlated with each zone it has been maintained that there is a more or less definite colour phase. It has, for example, been sought to show that in the northerly belt (South Atlantic) a more or less uniformly brown or dark bird exists, while in the more southerly, or antarctic zone, a pure white form has been evolved, or is in process of evolving. Ogilvie Grant (in MSS.) even went so far in some preliminary work on this report as to recognise two definite species corresponding to the two zones. Thus he writes : — “ I am able to distinguish two species of Giant Petrel, (1) a more southern form, with dimorphic plumage — a greyish-brown phase and a white phase — met with from about 45° south latitude southwards to about 75° ; BIRDS— LOWE AND KINNEAR 151 (2) a more northern dark sooty-brown form with a lemon-yellow bill, met with from about 35° to about 55° south latitude.” While there may be some truth in the assertion that the percentage of dark birds preponderates in the northern zone, while that of the white phase rises higher in the south, we feel sure that the solution of the problem this Petrel presents is comprised in the life-history of one species only and the relative proportions of some four or more colour-phases which characterise it. In the oceanic islands enumerated above, huge communities of breeding birds are annually concentrated, and while we believe that these colour-phases may vary in proportion geographically, we also believe that each phase is represented in every breeding area of any extent. As we shall point out further on, we also believe that at least four of these colour-phases are constant and may be described as “ all whites,” “ all browns or darks,” “ white heads,” and “ grey necks.” When Wilson in the “ Discovery ” was passing down south past the Macquarie Islands and the Aucklands, he noted Macronectes in large numbers, “ almost all of which seemed to be somewhat small and grey, instead of brownish-black, as though they were perhaps the hen birds or the young of a nesting colony.” But, as far as we are aware, there is no difference, certainly none which could be noted at sea, between the juvenile and fully adult plumage of Macronectes, and the only conclusion we can draw is that the proportion of “ grey-necks,” relative to other phases, about the islands south of New Zealand may be prominent, but this conclusion, of course, is drawn in the most tentative and suggestive way, for it is obvious that the whole subject requires a great deal more investigation. The little that has yet been done seems to confirm the opinion we have expressed as to the presence throughout the whole area of dis- tribution of this species of at least four colour-phases. For instance, Mr. G. E. Ains- worth, who was left with a party on Macquarie Island during Sir Douglas Mawson’s Antarctic Expedition, and who studied the bird-life of that island, says : — “ The colour ranges through various shades from almost pure white to a dark greyish-brown, some even appearing almost black! ” (italics ours). So that it is clear we get the full range of colour, or perhaps better, the two extremes, dark and light, even in this southern zone of the bird’s range. Passing to the Falklands, that is to say, to the penultimate northern distribution of the species, we find Mr. Rollo H. Beck, in the American Museum Journal (1917, xviii, No. 456), publishing a very fine photograph of a sitting colony of these birds, in which at least a hundred examples are shown. In this photograph there is nothing to belie the statement that “ grey- necks ” and “ white-necks ” are the commonest phases present and commoner than “ all blacks.” Wilson’s observations made on the “ Discovery ” in regard to the Giant Petrel cannot unfortunately be quoted in full here, but as to the relative distribution of the various phases of this bird he has some interesting notes made during the course of the voyage and while in the ice. The conclusion he came to was that “ the white form, although seen from time to time in the more temperate region of the Southern Ocean, iv. 5. 7 152 “TERRA NOVA” EXPEDITION is really very much more abundant, both absolutely and relatively, in the ice. And not only this, but that the abundance of the intermediate forms has also some relation to locality and climatic differences.” He goes on to say that “ in a voyage of 140 days, covering many thousands of miles of the sub-Antarctic ocean, only one White Giant Petrel was seen among several hundred of the uniformly darker ones, giving a very small percentage, and also that the percentage of intermediate forms is almost as small, amounting to three or four in all, or less than a half per cent. Whereas if we compare this with the proportion of White to Dark and Intermediate birds in latitudes where ice conditions are persistent, we see that in a total of about a hundred birds observed during half as many days, in a voyage covering only about 4,000 miles, the percentage of intermediate birds rises to 23^ per cent., and of white to as much as 30 per cent. Thus : — Between 33° S. and 66° 7' S., we observed : — Dark birds. Intermediate. White. At least 500 4 1 Whereas, between 66° 7' S. and 78° S. we observed : — Dark birds. Intermediate. White. About 60 14 18 The fact that the relative abundance of the white phase to others increases as we proceed south, is partly confirmed by the observations of Mr. Bennett (in litt.). For example, he states that in the Falklands (New Island) the white phase is represented by a population of 2 per cent, only, whereas in the South Shetlands (Deception Island) “ I have many times counted up to twelve hundred, and the proportion of whites is, I am now satisfied, 12| per cent.” Again, writing of New Island (Falklands), he says “ the white ones are scarce, grey-necks and white-necks abundant, and sooty-black are equally numerous.” At the South Shetlands grey-necks amounted to 40 per cent, and “ all brown ” to 50 per cent. We shall refer again to the “ all-browns,” which we suspect are faded “ sooty-blacks.” Other observers have quoted figures which seem to fall into line with Wilson’s observations. For example, Eagle Clarke (Scottish Antarctic Expedition, Ibis, 1906, p. 172) estimates the all-white form on the South Orkneys at 2 per cent. ; while the same phase around Graham’s Land, according to Mr. Burn Murdoch, was estimated at 5 per cent. Moreover, all observers seem to be agreed that the more northerly bird is a dark bird and this seems especially to be applicable to the Kerguelen Island area. It must be remembered, however, that none of the old observers paid any par- ticular attention to problems of this kind, so that the relative proportions of colour- phases present in any island community were never noted. Thus, although Eaton and Kidder were on Kerguelen in connection with the “ Transit of Venus ” operations, we learn little or nothing from their observations. Indeed, if we may judge from John Nunn’s “ Narrative of the Wreck of the 4 Favorite,’ ” written in 1850, the white phase BIRDS— LOWE AND KINNEAR 153 of Macronectes was actually present on the island in his time. Mr. Robert Hall, how- ever (Ibis, 1900, p. 27), records finding “ some twenty-one fledglings, as large as full-grown geese, in a rookery on Kerguelen, all in grey down (? bluish-grey).” This, it may be noted, is in contrast to the pure white down of the “ white phase.” He says, too, “ I found several young birds which had just lost their grey down, and had assumed a shining black plumage ” (italics ours), but as we shall again note similar chicks in blue-grey down occur in the South Shetlands alongside chicks in pure white down and both respectively and immediately assume a pure black and a pure white plumage, which does not appear to differ in any respect from a fully adult plumage. We cannot, therefore, feel sure that the white phase does not, or at any rate did not, occur on Kerguelen although we have no specimens in the British Museum. We do learn, how- ever, from Layard (“ Birds of South Africa ”) that “ a white variety is common up the west coast (of South Africa) towards Walfish Bay.” It may be stated here that former writers, Butler being included in the number, originally regarded the white phase of Macronectes as an example of albinism, while even Mathews in his “ Birds of Australia ” had not entirely relinquished the idea, as may be noted below. That the “ all-white ” phase without any trace of black feathers scattered about the body as in the more normal “ white phase,” is not due to albinism is, I think, conclusively proved by Menegaux (Exped. Antarctique Franyaise, 1903-5), who records an example of this kind taken in the South Shet- lands, as having an iris “ brun-marron.” It was taken with two others of the normal white phase. When G. M. Mathews was engaged on the work just mentioned, it may also be noted that the opinion still prevailed that “ the oldest birds are pure white, and the youngest are the darkest coloured ; but the latter are much lighter than the lightest coloured northern birds, where also the youngest are darkest ” (cf. vol. ii, p. 1888). This we feel convinced is not correct. As regards our own conclusions, it seems first of all quite clear, judging from examples of chicks in down taken on the same day from the same colony and the same locality on the South Shetlands, that the chick of the Giant Petrel moults directly from the down into a phase of plumage which is identical with that of the adult. There are, for example, in the British Museum two big chicks in down from these islands ; in one the down is pure white, while at the base of the down feathers pure white feathers, in all respects similar to adult feathers, are coming through carrying the down feathers on their tips. Every here and there pure black feathers may be seen, a,s in adults of the “ white ” phase, scattered indiscriminately among the white plumage. In the other chick, the down is a blue-grey or greyish-slate colour, and similarly the feathers coming through and carrying the down on their tips are deep black and cannot be distinguished from adult feathers. They correspond to the “ black ” or “ all brown ” phase. It is clear, therefore, that the “ white ” and “ black ” colour -phases have no relation to the age of the bird. They breed true de natu and so continue fixed 154 “ TERRA NOVA” EXPEDITION throughout the life of the bird. That the “ all brown ” phase, alluded to by Mr. Bennett in his description (in litt.) of South Shetland rookeries, represents nothing more than faded “ sooty-blacks ” seems most probable, although only actual study on the spot will settle the point. There is good reason to believe from the series of skins we have studied and from the notes of observers in the field, especially those of Mr. Bennett of the Falkland Islands, that in addition to “ white ” and “ black ” phases there are at least two other colour-phases, which, in the words of Mr. Bennett, may be described as “ white-necks,” “ grey-necks,” and “ all-browns.” As we have just said, we should be prepared to find that the “ all-browns ” were merely faded “ all-blacks.” Unfortunately, we have no chicks in down with which to prove that these last colour-phases originate de natu , as in the “ white ” and “ black ” phases ; and this because collections of birds in the field were often made in the past in such a haphazard fashion, while the importance of collecting chicks and other plumage phases was over- looked. Nevertheless, we believe that it will eventually be demonstrated that the “ white-necks,” “ grey-necks,” and “ all-browns ” do not represent phases of plumage correlated with age, but correspond to distinct colour categories. In support of this we may say that we have examined every bird of the large series measured for moult, and though there was ample evidence of moulting we did not see any juvenile down. In every case where a bird was moulting, whether blackish- brown, grey, white, etc., the new feathers coming in were similar though naturally of a richer or darker colour as fresh feathers are. In any breeding community of Giant Petrels all the evidence seems to point to the fact that all these four or five colour-phases occur together, except possibly in Kerguelen, where it appears from the accounts of voyagers and naturalists that the “ black ” or “all brown ” phase has become entirely predominant. It seems clear, in fact, that in the various breeding or other communities at present studied, the relative proportions of the colour-phases represented may vary either geographically, seasonally, or according to mendelian ratios. Since writing the above we have read Mr. L. Harrison Matthews’ remarks on Macronectes in his account of the birds of South Georgia included in the “ Discovery ” Reports. Through the courtesy of Dr. Stanley Kemp we are enabled to reproduce them here : — “ There are three main phases of plumage in the Giant Petrel — dark, light, and intermediate. The intermediate one is the most common in South Georgia, the next in numbers being the dark one, while the light one is represented by a very small percentage. The colour phases mingle indiscriminately in nesting. The few white examples found nesting in South Georgia were all paired with dark or intermediate ones. In the South Orkney Islands, where the white phase is much more numerous, some white ones were seen paired together, but this was evidently haphazard and only due to the higher proportion of them, as others were seen paired to intermediate birds.” BIRDS— LOWE AND KINNEAR 155 m § a zn O 55 Q X <1 zn ►—I H 3 ^ 3 3 o 55 ~ i— t h a <5 5Q P P fe; =Q fel P O P 55 o »5 o a o GO a a a 3 a a P 02 <5 a §•"3 = « ■= >> © (h 60 A 3 >» a S- ' “3 x ■ § 2 J. ^ d ft 3 a> -— “ > ; o « | •*' =s 2 60 ft r? >>-2 >> £ 3 ' O i-Q 1 “ ce •« :a O ja -+^ J. ■ „© o « § 3 £ £ &HS p ^ p - © ^ -O g-O m l to _h GO I « . SP © v5 2 "© g| bi > . Ha '-* • ^ © ^ >>£P © G 05 CD 60 g ft c -5 3 ° ES — ^ O ft <- .0033 PHO« 2 ■£ ft n-g §• .^o "dWQ O © ^5 c3 — Pi eg o © m ^ © O r© • p" ^ 7© 32^' v P 2 ce £ "h -H O .2 13 02 )0 O .2 •H c ^ M « 2 S pq r r £ o x’ * CO * * • rH O 00 o T* «-H CO 'h 00 rH Tt< : j io im im I t- irj O Cl CJ © Cl CJ go go ©S co cd »o oi o CIOOt-COOO^OQO ©5 GO CO GO GO GO GO 00 O* X W o >> ce pq o a ce pq •- © tJD-2 c o •© ° © © o © s ■g o 2 C © ce a © >> . P3 *-. • © o £?• a ®W 2 S © © «2 ^ ^ ce .-p ce <£ pq ~ > £ *rT> © ^ -p> © if-s s ® pq 2 & .S ° 60^ ©3 C ce 'Zt ce © > rt, £ o T 2 £ >> 2 -p w -p ce g -P P CQ O c .'5" •* g 6o-a p ^ C fi b#n ce o -i3 © - ^ ce - ce ce ce h o p p r© P3 _ O © © kg © o V, ^ ^ Pi C fcfi © c ce T3 bD-0 s ^ rr o O t ^ a 3 6 '“’ Q , P o OT — ce © 1-0 s _ ^ — I O QJ o 3 o 02 60 fW C © 2 ^ 3 * ? . & c g S ft 2 >> ^ ft -ft -» C ft 48 'S • p © ce © ce P3 , I— K'V • ft ® &:. © *: pq pq o P3 -P o 2 2 2 ce cc • co x; I co _u ©i o ©5 :© -H rH X . rH CO 00 GO* co co ©5 ©5 o+- o o o 0hn CO GO 00 CO CO CO TjH TfH ©5 ©5 ©5 :© :© :© M X X rH* r O d CO co rH ©e co cd cd ^ cq oi rx>. P rH CO o i o i Tii 35 0) PI -P 111 c >>2 d CO . a P 4) ,pH P o PH X W 2 d a s o 2 2 © w - o a d kH w d d § 23 OQ 02 « Ph - • )>> a g £5 +3 -p H-l cS >© c 22 g P< © ft © 22 g P, © 22 S - ^ .a 3 -2> d © -p i p d bijd 2 c- .O p | s -*a «ra s IS p. £2 © 5 °2 P4 Pi c3 T3 T3 'd o (3 > o -*-> "5 o -p> .Pi Pi ^ 2= P=, P -P> ■«H3 g iv2 B ?»S 2 S' * a t£ P. h 4) . ^^3 2 ® .2 ^ £J=J g ^ s ^ « n$ >> ^ -ggj cig £> -C «2 £ ’a "* 2 ° £ C 2 .'" -Hg > . q 2^0 j.Eo M 2 2 O Js! P^P^PJSP ^ rt ^ ci ^ d -2 >> O P C3 r-H ^ '“■3 S I g 1 1 .* I ^ | % e'Sgl © .5 c .a 3 "B “* bcjS rt ^ © -P> « ^ '" © -P ^ 'P 2 © £ © > ^ .-« o °1) . X ^.<6 _ _ >,J2 o>2 a fcs2f£ bCi*d bJD > cS Q -sl'l •s gB 8 M aw O P5 P <8 -a « p *s ■I 5 IS CO «a © 'S a 03 s © g . w >>T? ■£ &§l s.3^ 2 £ cc TJ O . sis til .j: x. tf ^ © O >• £ . 2 0 0© £ 22 — 2 « n >*•> cp 3 -P> +3 — ■8 (. <1.5 P3 W P3 „ ® g fl M O g “2 S "3 °sl P5 . fl o O ‘S ^ P p o • o cS H3 ^ hS Q s s •O'c > C_i HI o v d c3 fcS 2 • 2 p P 2 p . © 2 © O © co t> g 2 x 2 C o 2 2 S 2 > g g Ph d O 2 w o H o d 'd <1 © 24 d O o d co 2 o 2 ^ o © rd ft 00 t» g o co o m ■ O Ph d d o CH cS O w |zj o © GO P l> • d 1^.3 m 6 © O O g o O d 2 2 O m % a H o d >> o Ph © Ph d O d d < © Ph d O CO r-H ^ CM CM O o 05 05 CM* a CO 00 00 00 00 1 6 00* rH »— • r-H CO d rH rH Th Tin TjH Th >o ic d CO co rH d 2 l-H r-H r-H CM o o o rH O o O 05 05 05 05 05 05 05 05 05 05 05 05 H rH rH rH rH rH CM rH rH rH rH rH H H o 05 t- 05 05 o 05 05 00 05 05 05 05 00 05 00 C5 »o GO o 05 LO CM d 00 VO CO 05 h?H CO CO Tt< CO CO CO CO CO CO CO CO CO CO CO CO o o o o o uo »o o o »o o 00 CO o o o 05 05 00 05 05 05 o 00 o 00 rH < 2 M 2 w 'A w <5 < co • - co J 5 w d d oT d d oT oT w s > d P ** . d bp k o p O > d P bb d > p o p O o _p O >> o i 8 2. o FP BIRDS— LOWE AND KINNEAR 157 > x r3 5 > 3 d St-O >. © .2? •- -2 sz 3d I 'g'g-S 3 Q *3 - 33 -2 2 - &c-f> | - S « 1 r 3 2 £ j§ ** -3 — ^ >> - ~ ©> — § > g. ^ o ^ ©-d | X J S ^ SC ^ sc s ft. 3 © fl © © *3 ©ft v* .-^ c ^ >> © 5c 5 © ^ > .3 d ■*» tbS ® c 6 o 2 c VS _© -** _© •** 9 — • * s ^ ^ « s H s J3 ^ -s «s ©>>Q’©p p°ip'*^ -.2 •- -3 ~ -y* ^ ft© ^ "l r^l -2 ^i"T" tc S£-^ — S © 4* © -ft> © -^ © ^ 5 5 c 525 5 1 T^ ® - -i _= -£> „• ||| 1 . J >> c 2ft d .g > | »’? “a. O m fi „ -9 2 -t> _ > .3 .3 O O " > ft "3 CJjS - "T — -3 Aoi'S ° *■3 °"3 >>0 • s ^ >, o 2 - ° £ 13 g A 5bJ5 ft £6 £q # o5 £ 2 E ^ S S o J |J4 c3 * : S ho ^ “ ■ a £-3-^ -w d „ “'32° i = oA-2 © ^ ft © ft ^ © ft 3 21 2 . ^ to-; — 2 J g* 3 — Sb g cn •>, , > £ g 2 S .3 2A -a 2 ' _© O -3 ! — ; J3 <2 i <1(5 3 — -3 a o ■3 "3 o ® 5 2 g ^ c „ . - >> g g - o . S g 2 so «s 2 6C 2 J3 C ^ 2*^*2 c J a >, 3 c % P o v 2 3 M ft 03 0 O . r3 S ►, 2-3 § r— ■ ftj © p. p b-SSo i.s.gg 033 t. o a; £ S O ^ <3 S ft 4°** 2 ci ©* ” 2.2 02 .sl f § .ft! ftH "§ o • 2 S3 -8 5 .1 © c 7S ^9 g ^ 5P ■S.S © ^ •n 3 ^ O O jS ^ 3 x o © © t- t- .Q > > O O ft pq r ^ © 5= -g >>.-2 o © jz r ■ ** ft - ^ • ■“■ -a .-e * o < O C3 •a St3(g o o . o bO50 <£ — , cs . " t. -+-1 jft • Cj OflWW >% Ph eg o a c8 > w © ©H c3 o ft. « b2 o S CQ © o S ^ W o >o m . r v T5 02 § §b 0^1 o CO o l— r- I> L'' m co o o ft. ft. s © § S H ft .W (~ t- .=3 t~ C' ft ^ a> 1-1 a ■Sft s c3 o .a © O P " <3 2^ 5 o C2 i© ©©© odd LO lO lO rj 1—i o o o o .K o GO © © © © r-l '"H i-H rH H * Measured from feathers on forehead to tip. f Measured at greatest depth behind the narial tubes. 158 “TERRA NOVA” EXPEDITION -a fl a £ • § O m a p p m P P i— i P H a * o f-4 pi o c3 s ® Jfl *r« CO 0) **H pd p ■g 5 S o o p* z« •°3 g tf S si c o -a £ so „ d ° <= ~s * *2 -* : : 5 : ; : i : : : : : :|i§ 2 cs S o w J= is ill to 00 73 'S so to so i.s.s .5 ^;£d:£ is 3 3 9 O O O O .SSSS I I f 5 i s o 5 O.US.-2 -a •* a £ s' 5“ S § 3 .13 .-s .■« QQQQQfl M fl *M 1 rfl fl n £i w oq Ph PQ =0 to &h fe; to **s to SQ 6h to fc} to g P o m H £i p a p P P 02 <1 P e4'« = O 05 o ,q M o o Ph 60 9 « F m 9 t> ed T3 fl c3 P i cd pq M Tf TfH Tj< H< JO H< H< OJrHOlOiOtrHOJClHClOJOlOlOJrHO d :g d d d :1 d d :g d d d d d d d -rp^-p-ps > -p.b -p WHOJWIOH^- , ^ lO r-1 pH 05 l> Tj*V OOOOtOOOOOOOOCM'OO'O ; ;t3 • • • fl) • Pi . X tj h ■3 g* .1 |w gw o -2 9 d ^ osH O (_i '"IS . . cO J (5» „ . , 170 50 124 1913.12.26.5 .. 6 ? 1910 “ On the way to Australia ” R. McConnell 165 49 120 1880.12.10.1 .. — — “ On the wav to New Zealand ” Sturgeon 168 54 120 1880.11.18.683 6 ll.xi.75 37° 23' S., 83° 1' W “ Challenger ” . . 156 47 — 1880.11.18.682 6 1.74 Kerguelen (type of chionoptera) 166 53 126 1841.748 6 — 99 • • • • • • Possession Is., Crozets . . Antarctic Exped. 170 54 126 1909.11.16.18 .. 6 13.xii.07 Dr. Collett 169 53 122 1909,11.16.17 .. — 13.xii.07 163 51 120 1880.11.18.684 • xii.73 Marion Is. 174 55 118 1901.1.7.60 imm. 9 29.x. 98 44° 26' S., 37° E., near Prince Edward Is. Hanson 160 49 116 1901.2.5.4 6 24.x. 98 42° 23' S., 20° 32' E., S. of Cape » 176 52 124 31. Diomedea epomophora (Royal Albatross). Plate VIII, figs. 10 and 11. Diomedea epomophora Lesson, Annales Sci. Nat. Paris, vi, p. 95, 1825 (Campbell Is.) ; Diomedea regia Buller, Trans. New Zealand Inst, xxiii, p. 234, 1891 (Campbell Is.) ; Salvin, Cat. Birds Brit. Mus. xxv., p. 443, 1896; Wilson, Nat. Ant. Exped. “Discovery,” N. H. ii, Aves, p. 110, 1907; Diomedea epomophora maccormicki Mathews, “ Birds of Australia,” ii, p. 261, 1912 (Auckland Is.). The following notes in Commander Pennell’s diary undoubtedly refer to the Royal Albatross : — July 11, 1911. — 42° 06' S., 175° 13' E. Absolutely snow-white above and below except the tips of the long primaries on the upperside, which are black. The purest white Albatross I have seen. Bill flesh-colour. Aug. 6, 1911. — 34° 08' S., 171° 53' E. (West of West I., Three Ivings.) Two settled on the water close to the ship, feeding, and remained a long time. The Mollymauks were afraid of them and never attempted to dispute the possession of food with them as they did continuously amongst themselves. Sept. 6, 1911. — 34° 25' S., 172° 10' E. (Off Three Kings Is.) One fine specimen. We have already expressed our opinion when discussing D. exulans that this last albatross is genetically distinct from the Royal Albatross, D. epomophora. The following description of plumage is based on material in the British Museum : — (a) Nestling. — Covered with pure white down. (h) Juvenile. — (With down still attached to feathers “ coming through ”) : general coloration white — dorsum white with no dark vermiculations, but scapular and inter- BIRDS— LOWE AND KIXXEAR 1(57 scapular feathers with dark terminal or subterminal bands or drops ” ; entire head and neck pure white ; tail nearly pure white but with a certain amount of vermiculation ; first primary with distal half dark. (c) Adult. — Similar to juvenile but with terminal and subterminal bands of scapulars and interscapulars less heavy ; first primary with distal third dark. MEASUREMENTS OF DIO MEDEA EPOMORPEORA. Reg. No. Sex. Date. Locality. How acquired. Length of bill. Greatest depth of oilmen. Tarsus. Notes. 1900.2.22.3 . . — 20.x. 1899 Campbell Is. Sitting on nest. 183 46 132 1895.3.1.1. .. — 24.iv.1894 New Zealand Sutcliff . . 185 53 131 1891.6.16.5 .. — 20.xi.1840 Endeby Is. Auck- land Is. McCormick 168 46 123 1891.6.16.4 .. — 30.xi.1840 ,, ,, ft 179 54 130 1901.1.7.62 .. Juv. 4.x. 1899 Campbell Is. “ Southern Cross ” 180 51 130 1901.1.7.63 .. Juv. — ft ft 170 46 122 1901.1.7.61 Juv. 7.x. 1899 f9 ft ft 172 45 120 Tring . . 9 15.vii.1827 36° S., 52° 30' W. . . Buenos Museum 172 50 123 6 — New Zealand Buffer Coll. 184 55 130 S — >> . . ,, 174 50 117 ” • • — — South Pacific, “ Off Australia.” 172 54 124 ” . . 9 — New Zealand (type of I). e. rothschildi) * 172 51 115 t, • • • • i — No locality Buffer Coll. No. 8 176 51 119 9 — „ No. 9 174 51 119 ,, 9 — New Zealand „ D 169 48 114 6 — ,, „ c .. 173 53 121 i 12.vii.1827 35° 41' S., 53° W. ft 178 53 121 Mottled tail. ft • • • • S 13.iii. 1875 New Zealand — Mel- bourne. Von Hugel 162 50 111 Upper parts brown ; throat white ; breast brown ; belly paler ; tail and wings brown. ft • • • • 6 13.iii.1875 ft ft " ” " 172 52 139 Coloration as above. * We think this is not an exulans. 32. Thalassarche melanophris (Black-browed Mollymauk). Plates IX ; XIV, fig. 6, and XV, figs. 4 and 5. Diomedea melanophris Temminck et Laugier, Planch. Color. d’Ois, v, Xo. 456, pi. 115, 1828 (Cape of Good Hope) ; Diomedea melanophrys Salvin, Cat. Birds Brit. Mus. xxv, p. 447, 1896 ; Wilson, Xat. Ant. Exped. “ Discovery,” N. H. ii, Aves, p. Ill, 1907 ; Thalassarche melanophris impavida Mathews, “ Birds of Australia,” ii, p. 267, 1912 (Tasmania). MATERIAL OBTAINED No. 43. $ 22.x. 1910. 44° 25' S., 160° E. Bill yellow with orange tip ; legs, toes, and webs pale greyish flesh-colour. No. 45. <$ ditto. ditto. ditto. No. 46. ditto. ditto. ditto. Systematic Notes. — The three specimens brought back by the Expedition should be, according to Mathews, T. m. impavida , inhabiting the New Zealand and Australian Seas. This race is said by the describer to be distinguished from the birds breeding on iv. 5. 9 168 “TERRA NOVA” EXPEDITION Kerguelen by “ its shorter bill, the more pronounced black in front, above and behind the eye, and the greyish wash over the lores.” In regard to size the birds from the Australian — New Zealand area, as will be seen from the comparative table of measurements, average smaller, and if we take the range of the length of the bill of the birds from the two areas we find those from Kerguelen are 114 mm. to 120 mm., while the others run smaller, 106 to 119 mm. None of the Kerguelen specimens has the greyish wash on the lores which is found in a certain number of the Australian — New Zealand examples, while the black in front, over, and behind the eye is on the whole more distinct in the latter than in the former. The colour of the bill does not seem to differ, but the feet of Kerguelen birds are given as fleshy-red with webs of the same colour, while those from Australia and New Zealand are recorded as flesh-grey, a light grey or greyish-flesh. We are not, however, prepared to place much reliance on these colour differences, since they may be con- nected with age and sex, and until further specimens of different ages are available it does not appear desirable to keep these forms separate. The original specimen to which the name melanophris was given by Temminck came from the Cape of Good Hope, and Mathews considers that the typical form breeds on Gough Island, Tristan da Cunha, and the Falklands, but on what grounds we do not know. There is a single skin in the British Museum and two at Tring from the Cape of Good Hope, while in the last-mentioned museum there is one specimen from the Falkland Islands, but in neither are there any birds from Gough or Tristan da Cunha. It seems to us that the Cape specimens agree equally with the Kerguelen examples as with the single bird from the Falklands, and till a series has been compared from the last-named locality it will not be possible to decide the question. Three skins from South Georgia in the British Museum have the longest bills of any we have measured, but the birds from Chili and Peru hardly seem to be distinguish- able from the Kerguelen specimens, and the differences in the bills claimed by Mathews are not borne out by our measurements. In regard to the colour differences there seems to be a considerable amount of individual variation. Wilson’s first entry in regard to this species is as follows : — Aug. 8, 1910. — 33° 26' S., 5° 34' W., West of Cape Town. Diotnedea melanophrys. Two which I saw were billing one another exactly as though they were adult and young. The adult bird has the typical yellow bill with orange tip : the other a yellow bill with a black tip. On August 10, in 35° 27' S., 0° 39' E., this Albatross was again noted and it con- tinued to be seen till August 14 when off Cape Town, under which date Wilson notes : “ The Albatross settled before the heavier squalls and did not appear on the wing at those times. We came into Simon’s Bay in a very squally fresh gale.” BIRDS— LOWE AND KINNEAR 169 On September 11 Wilson left Cape Town in the “ Corinthic ” for Australia and immediately fell in with this species again, noting : “ Two or three adults with yellow bill and orange tip.” On the following day, in 36° 26' S., 22° 47 ' E., he remarks : “ Ten or more all day. A marked difference in size of what appeared to be this bird in the distance but probably was not. One bird — a young of D. melanophrys ? — had a pure white head and a pale yellow bill with a black tip.” This Albatross was seen every day till September 29, in 43° 48' S., 146° 01' E., and all appear to have been adults, since Wilson in nearly every case notes : “ yellow bill and orange tip.” Pennell, in the " Terra Nova ” during the voyage from the Cape to Melbourne, noted five of these birds on September 21, S.E. Cape of Good Hope, and again on September 27, when he remarks : "Many ; they settle in the water when things are thrown overboard and then, from that position, dive completely under after them if necessary." After the Terra Nova ” left New Zealand on her southward voyage this species was again met with, first on November 30 off South Island when several adults were recorded, up to December 7, in 61° 22' S., 179° 56' E. On all occasions except two adults were noted, but on two occasions young birds with dark bills were also seen, viz : — Dec. 3, 1910. — “ Again very busy with the ship in a heavy gale, but saw adults all with orange-yellow bill. Very abundant and very bold in coming for food under the counter.” December 4, 1910. — “ Several adults with orange bills and one younger one with dusky yellow bill and blackish tip.” On February 13, 1911, Pennell in 71° 13' S., 171° 15' E., some fifty miles N.W. of Cape Adare, notes in the zoological log of the “ Terra Nova ” “ a few seen.” Then throughout the return voyage to New Zealand he records this bird daily in varying numbers. Under the date of March 15 (58° 25' S., 161° 22' E.) he records : “ Nine at one time, one having a curious dark bill and almost black unguis. It dives into and under water with three-quarters expanded wings, coming up with the wings still half-open.” On March 18, in 56° 28' S., 162° 58' E., the first specimen in the list was captured and full details were taken of all the soft parts : “ One caught (No. A), legs and webs lilac, tinged with red ; toenails white, iris light golden. Culmen yellowish-horn at the base gradually becoming pinker towards the unguis and end of the mandible, which are pink. The thin line of skin where the bill joins the feathers of the head is black. On the side of the lower mandible there is a small strip of pink skin : inside of the mouth and skin of the pouch below the mandible is dull lilac-colour. The gullet con- tained crustaceans of the Order Euphausiacea and a lot of garbage thrown over the ship, including a mass of paper. Many of these birds are rather quarrelsome (or perhaps peevish would be a better expression). When food is thrown over they settle on the water, keeping their wings half-expanded for a little till comfortably settled down, and peck at any others that come near, though I have never seen two really fight.” MEASUREMENTS OF THALASSARCHE MELANOPHRIS 170 TERRA NOVA” EXPEDITION a 42 O .a co -h O m *1 a m c3 02 •cog3 •- O 73 to a a « 73 © © a © d 2jj sSo tS ft a 03 . r—i CO §,«.-§ s £ § -a -a p ft bo fe ? h t d ® -a p s ®.Sfto © a Pa oq to a •a 'S .5 £ ■g* •P ® lO a g o 2o® s-» ^ 1 ■ S 4) -o5 “ © a . a 3 i a 3 o 0 3 -9 J -P> H "3 a o ? ? to c3 cS © H H a, ft a a a 02 02 73 I © rO c3 GO £ 1 S & P> bD O .a ” ■A >j.a jq ® ® +3 .&A .a ° g -£ ft r. rM *H O O m _ ^ 2H "© -p © C co ■5 t ■n a 1 ^ o3 -P) atH "2 02 ft 73 2 £ _ ... bo III *3 ® 2 . S ‘ O CL 0, d Jh CD CO .2 ^ '■d O *> . O Ph s® H >>J2 ^ £ n Hu '-ilff§ GO ^ CO 2 a § t£ £.-§ -2 * ft © rH a ,5a a A -p a . „ o fti a ^ 02 a g o s ft ft .3 a 73 § § 73 bo ftl 2 a IS 1: GO O ft c r-( -P t3 3 ft rQ 43 „ a ® ? >> C3 ® ti< ft^ a 02 o -** 73 a . a to p a ® a ” O ft ft o.S 53 ® 43 ©-aS aj®a £ © . §0 >> £ _d © o *-h ?h ar ft ?h ^ fee © > a ■* . o .ft a bo*3 3 a ® ft 3 g3© ce o tj ^ a O c3 CJ © b © o © © ^ ^ ±i ^ N ■2 © GO <— £ ^ CL ^ m § ft 02 > © | & 0.9=3 ® tao'43 >> cS .2 ^ ft S* M O 73 S3 • M si - © "£ S "o ft ? S « o p "7 S> ft ft°ft M C3 0 io 00 00 m o O r- 1 00 co 10 >o CD CD C5 O -<4( 10 1 a o O —1 10 uo •<# 00 -<3< O m CO CO 00 00 <1 PC ft <1 Q £ «5 ft CO CO CO l> IO CO CO CO ^ I I> CO I CO O T* CO CO ^ CO Tt< CO CO CO CO es o e ■S 1 00 10 10 »-H 1— H pH O o a> i-H O d •2 ! •§ p^> ft £ o O ft ® u bD S B =3 © d .2 ’43 © © 0 GO 1 43 0 1 1 ft ft 8 a .0 43 © © sd o O d w 73 1 — 1 © a& rf1 p p i C5 © © 33 a a § -a . 0 ft 3 • ft a r •2 _ 73 -pa© ©Oft ft 'ft ^ r-H o §Ho ft ft ’ ^ |ft B* © H ^ 43 © a n hi H S“ © © - o- a C5 o a G5 d a > o ft | 73 a p h 2 a o ft <3 S ft ft 5 - a o X a a © ft eh^!«5 a I a t» 'a 3 © <3 N43 * -g ft 5 ft 02 02 a I . -p • rH • a m ft 02 a ft 2 a d rd £> ft p-^ ft 2 o m 73 d N ft • *8 pa pq ^ r: o O t- ^ l> © pH •p a !Z2 a -p ' a £2 a ft 02 • • ft „• C73 ^ a ce cd « 73^ a © £ N r ft !> 02 Iftg r co 02 tK ft 02 m CO ^ CO TJH 1— 1 pH O O O 0 • Date. O GO ^H I> GO 00 00 I> H H GO • pH *rH '> '> :d 05 *> 05 nH 05 05 pH pH ‘E "S ^H 05 pH 222 rH 05 f-H X CO 05 GO K rH p-H l> H d d 05 05 I I 05 O CO co 00 05 CO 0 O l> i> co CO xjj T* CO 00 »o 00 06 O d d 05 X) g bD d IO IO H rH pH CO bD d £ CO CO pH tiD • .3^1 1— 1 05 pH O 05 00 bD bD^ .a .as IO IO O O co pH -CO 05 CO pH id 0 62 s S GO r-H 00 GO rH pH 662 05 05 rH rH 05 rH 6 00 rH 05 rH 05 rH BIRDS— LOWE AND KINNEAR 171 3 d 1 3 Cw 0 Ph © 6 > GO 62 Ph ^ fe o o & © «H-H -? » © o 02 d >> •-H ^ no O -4-3 £ GO © Ph Ph 0 -3 > © ^ ? d iorn. and Sligl but O Ph «h-h 3 d 3 £ 3 ' ^ 3 ^ ° 3 to +3 .5 .2 d d a r-i © s ~s si* « d 3 h ce Ph.o 3 02 ® © 0 S & ja-S £ O *•* ! . Sh fj* h i~ 51 * 8 3 3 -3 a co *43 .3 CO — - © .2® g S ft ° £ =3 ^d © _ =3 H >>■■£ S g ■£ 'S - ao o ft so cS a a ^3 >i 1C o ° •> >, ® d ft © 8ft * > _ © 3 S « ■“ oa 5ft iJ .3 A >> *3 .2 ft S ^ © a g« 8 & c.s © £ © a X H O -+3 o © GO c3 P-i X rT WO - d g 3 ° § g d y H ft a & § a ft m >>£ 1e a 0 ’43 d © 3 ce ce Ph Ph1^ ^ .3 45 d d 0 5 ft ? © d .2 -d © — ft SC ft! c a •h ce d 32 s ® -h> 33 X ^ © O g j>> ? +=> 3 3 Ph-03 02 02 02 3 02 •* © d a £ § M CO Ph 0 '-S 5 ©T ■hh I ° ft co © X! ^ © ft ^ ft ft 3 -* o O ft ^ 5 © ft ft ft ft © . f® S Jt>ft © P © ft SCft © •2S © © 72 32 § ft a§ft 3 ce O 3 -3 .a* ‘ 3.S o ft o O m Si 0 ft H a p o o o fy O W P d d 3 © GO 3 O H S ce W 3 o m & d © O H • »-q W P o ^ a w = W ce H « © Ph x W © to 3 eS ft o ft M w £ O O c— I 6 § C co e d © cS 5K 3 . o ^ OQ ^ to 3 ■g^a d 3 .3 © ® .S « > -3 © Oft ft 5- o =e d g - 3 ft © 1? 2 ?ft t: e r ce ^ d Ph ^ s ^ 3 .3 02 d O U1 c3 o d ■£ ph 3 d.S 3 02 * * * * * * * * CO O d o CO o o ip CO co CO pq 05 CM CM O O X T* PI 1 CM CM CM CM CM pq i— h O CO d ip Ip »p ip ip IP IP Ip IP ip ip IP ip co o I> CO o CO CM CO CO co CO rtn pq ^ X O l> I> X X o X X X X X X X X X Ip »P !> F-H CM d IP CO CO CO X X TiH d IP X CO CO CO CO CO CO CO CO co CO CO CO CO CO CM co C5 d CO o CM ip CO CM o 05 05 C5 X co IP tH o o o o H CM «— H i-H PI i— ( l-H l-H l-H l-H l-H ft F"H — —1 <1 l™1 PH ^H i-H o hr! w d s d 3 ^ c8 § d 2 © 3^ Ph £ K :wh h' . w H w H . W ■ w O} .2 • • 0 Ph O pP * d X CM IP IP o O CO o . O © 3 PhO ce M ip o X o PI O CM , GO rft. «H 3 05 CO I> I- * H 3 • ^ GO ft * d 3 r 3 m ip CM o T* S«5 s ° o § PQ 3 © 34 O H PQ 3 1 ft c8 o Cape Dyer Is., Robbin I Greenlan guelen. Kerguele Kerguele 02 CO o co co 02 o CM o Ttf 02 co CM r d 44° 23' S Kerguele pH . o 05 pH o 05 o 05 X § 'rtf 05 f>. X 05 X X 05 X X 05 X » ft M > d x X X . • 3 •3 :a pq 05 o • 3 ^ CO M X X t> CM CM PI CO .3 • 3 CO* 1^ o’ I p rt5 OfOf Of fQ Of *o Of Of Of CD CO 05 ■ o .3g £2 to 3 d I CD (M CO Oi oq c gq tj . .® -e Ph ^ aj O HD . g © h A* ° ft g 3 M® .3 | 3-3 § 1 .3 1 S ft is a a c8 m o o O d o GO Of 5 >>2 o-d ■« ® js o a 1 ». a a ^ W M * * * * * Pn 4^ 00 l i=J 3 o £ M 0 1 3 c8 o "3 H cS Ph *3 > o O co pH cq 05 pH 05 05 pH .d M CO tj5 f-H pH oq CO EQ 3 r*. M r -> fl O « * 3 ^ O ^ 3 Oh O O _ ■+3 £0 .*3 C fl5 c * £ P4 fl e 05 .5p o ■P w,o a o ■•g’S Se -3 mm73 „ -* a § ■cS-O ® 05 -fl o ® >> 73 ® C ... 03 c •s s bO 03 C 3 © .fl ® ■*'2 j *£ OJ Oh Cm >> g fe3 ** » £ T3 • p C Em C3 w_ O H> m cS ^ ® o O Bib SJ >»J= Jri§ • *3 ® * c 5 Jh ^0*^0 >>-Z ® o a? ! ~ a; -fl t- d -fl .2 ° a k ft t> -a c .2 *43 £*■> •2 2 'S bC -EC fl* O o o. 05 03 0 O .« >> 0 05 ' fl — .£5 .,2 "5 _ < o o- g C U H o^.S fla3 g •a p o .tj CO O 111 ■s.*-3 ft o © c | 13 o fl . O pfl ■3! G-ftO — -4-S 3 J fcfl* «■ — co d fl* T3 8 « ■o >> T3 " C •- <* e "i o g § Wing. 522+ 490+ 490 c ft PS << o CO »o 500 512+ GO 00 513 + 516 494 495+ 516+ 450 | GO 05 CO CO 00 CO 00 lO CO l> 00 !> 00 Q CO GO 00 CC 00 GO I> X 00 X Em CO (M X X . fl co C co !> • 5- c mj a a c 2 cs 2 s a a ■|ft5ft w a ° iz o bC PS g s c ft d ^d ft PS CM X ft co T ft. ^ u § I | Sg> 0) rg C fiaH 5ftl o^d Ph T3 Oh X W CO I 1 co • 0 O c g 1 § ■Sx fl . o ^ 05 ^ o -*C >>pp « g p° a* ®c S3 » fl rK ® | od ? g S tE H w w 05 05 >5 c ^o o O 05 X 05 05 05 05 05 X i-H X 05 X X X X X r— H • fl r^ rH i-H r-H r-H .fl X .fl ’P X (N X X X X X X t4 c 6 X CO 05 l> 05 ic (N *-o 'fl d ♦o *X) • 'fl •S»S a' c^» 1 a ■fl d *+5 •° > CM- • • • • • * Tt< • bO bC X CO I> r-H bO bO co i> bJD t> no co *> OQ rH lO bp .S •S O .9 fl •R H o .9 f-H r-H O 05 ■£ H S H 05 f-H "C Eh 2h 05 X AUSTRALIA— NEW ZEALAND AREA BIRDS— LOWE AND KINNEAR 175 .J, © © g 3 “O' r-f 6-1 ^3 3 © © * 33 fc- ~ © • — to tfc© -2 +3 CD Oh .. o ”3 ? >i ©'2« — cd © © C ^ 3-c § «e • -3 _ J2 . © -tT'S ’§ ’g : .3 o © J3 2 . a ^ -5 - 5 a a) “ a rB „ "B ' £ g CL.+ § =3 ft H ft ft ■+3 “ C S © * " g »• ^ .2 ^ 8 ^§-0 CD © T3 © Of s3 >> >s _§ ® H o ^ w w 3 .S .*-+3 "3 .2 c3 ~ ; © -3 -+3 O “s & ° >» 3 .■£ + — ~ A 60 I* ^ “It cd 3 C tn ■+3 © . d| >> >> £ H to rS ^ © ^ © £ -3 *73 ° 3 1 s §.s 73 .S ce p o K 500 O O id © o ID 495+ + !>• 04 © Cl CO D O O ^ (N^Cl^ X X © — T^I Tf1 T+ O 527 -b 4~ © I> © lO X © © T+ l© X© o £*■ CD 04 X D D O CO CO ID X t- X © 04 00 X X X XXX X X X X X X X X X CO CO ID CO l© COTPC3 13 T# M C3 CO CO cd CO ^ M CO CO CO CO X CO CO ^ CO GO X o X CO <“H >-H -H O O © © © © T±< CO © © © © © . . 6- • '3 - ° p< . £Li © *© T) K CL, w . rrt © 3 © rvT ® PhW • fl #© ^© x W W; © £ > O 73 w! 3 - © 7C d P3 . 2 c3 bfi C «« -6_ a o -+-> bD a b| s Jlo sa -*3 Z *3 O g m S ? C "3 5 O w |o ft H U o O cZ ■£ s 0 *“0 S 3 w : : b » T3 a cd N £ . — 'wdt w r C LO LO a W oq (M cd »D o w o C5 o W rK o C5 w H P O M Ph,_ , X r5 w a ~ CO vs CD o*E 5 '5b 6h O © O CO © CO © CO ^ 2o S (M :p > :B ^ :a :s :fl :s :p M M M M ^ ■i M M M © ID* © id* © o I> id i> © ©©©©©©© CO Ol CO iM (N (M (M 04 04 04 i> oo CO CO CO CO X CO 2 I o« s« a o| •"ii-l •o *0 *0 «'0 <> O w o 6 ci -i 04 04 04 id © © © O -H O Q Q Q O) w> .9 Hrt H • 60 ^ -r £6 50 50 ci £ — , © -r» c3 © G « ^ S - © gs 2 & 3 g).> §P .> 3 j= o L 'S ^ ’o 5 ^3 ^ S 2 O G TJ ■ c'.2 5 T -r E £ d 03 g G2 .a fcjQ -p ® .S •g 3 p< g .2 © O £V ro ^ P3 /. J c S P s g © G . &0-p © c$ . 3 EH © •- o g ^ "3 ^ P-1 +3 •g © © •~ © ^ >-.-d ... © £ w tc 5 © .23 ^ bC J3 3 ^sS bJO > ^ 'G g - g d ng & &J's © \g ce C/2 ■+} G .2 G* o P^G © *& © G? © ^ rG 3 "© -S G >* •- ce © . .G S bJD^3 ^ G5 © © £ G - O pP .2 P< Ph 2 03 »o 03 x a x r- CO CO CO CO ++ -'OGfM^OOCU^OCCOO O x i> w r^ w r- x co T^COCOCOCOCOCOCOCOCOCO X; 2 t> h ffi O (M CO O O lO XOXXI>©2XIO CO CO CO CO CO CO XX OhCiO^©CO0T^^*t xi>r-oooooiOOC XXCOCOT^COT^COT^Tfrt’ r^cowo-HCjoo XXXXXXX>X Oh X w >> w G3 3 7 r— H © x O co © O 5 £'-3 g 1L8 S > O- O £: •£ X ° c 3 P1 a • 3 u . 3 HOP ®oj & ce £< © — J § 1 “is * 2 „ « *8 ~> 0D GG o a G c6 2 ► o * O G ’> ft .‘ft 2 . © H TJ G <1 G . O > co >>■• >-> . flo ^ «)" c3 "i So a £ So Sg :g a »> 43 H G M . Pi -• OJ * J* d J? aM . O ft %j K CO o o 2 -Go 2 G G o3 ° uJ OW Ph ■ - M uG W W : - u . © :«?-s S3 * G ^ I s| -+-> d ^ hP . G O w -G *2 o3 ^ w c3 G ^ ’© g O 72 72 ^ /"?«?] -H r\ c* d „ „ C3 >-h © h-J s SS| f-< rrt O l T h n s fi.S 3 ^-OoS .sioS W fto fiStf -3 . ^ . (h • o ■ o 2 CO © © G G .g 02 © 3 ts F G cj WCO-fH o Hn o c .* rH CO • — > Tfl '•# ^ 0 0 0 0^.0 Q Q 02 O 3 5 o 333il3^i| CO CO CO CO -r: *G CO rt< -H o o o o l-H -H 0>0^ lO Ol I— I 03 03 H CO h 03 Cvl ^wo©i>i>i>Vo6hi> rH Ol 03 i-H i-H r-H 03 ^ i-H ojio^joxxx’xuo^x CO 03 TfH T^’ rji Tii ©2©2©iTHOOOOO— I o XXXX©i0©50©i©:0 tc X 03 l> CO 1 *p — : rti ^ 03 O O o> ©5 X X r-H 03 03 X f 30 oi 03 rH Tjj I io id uo jo r-4 co co » id ©: ©i ©i r-H cd cp . X -rt 03 "1 03* X X L . I> o _ f— i rH l> r-H X O • • • • • CO T+HJOU'XCOTt' JO JO JO JO JO CO CO CO CO CO o o - X X t X X ( i h oo q q ® q ©2 ■ U © O ©2 C2 ©2 C. ) CS JO ©5 ©J ©•' ©i ©2 ; X X X X X X X 186 “TERRA NOVA” EXPEDITION The colouring of the soft parts of the head and bill are not in life like the coloured drawing I got of them before, which was done some 12 hours after death, and was much too gaudy. In life the skin of the head and the bill is uniform tinted-body-colour, often almost white. Unfortunately the notebook in which the colour was noted got lost in the coming off. Lieut.-Commander Pennell, R.N., adds the following observation : — “ At 3.45 p.m. half a dozen were roosting on a kind of laurel bush at 1750 ft. on the S.E. side of the main hill. The sun had by this time set on that side of the hill. At 5 p.m. they were roosting on dead tree-trunks, etc., from (say) 800 ft. down.” 38. Sula leucogaster (Brown Gannet). Pelecanus leucogaster Bodd. Tabl. Planch. Enlum. p. 57, 1783 (Cayenne) ; Sula sula Ogilvie- Grant, Cat. Birds Brit. Mus. xxvi, p. 436, 1898. No examples of the Brown Gannet were obtained, but the following notes in Dr. Wilson’s diary refer to this species : — July 8, 1910. — 13° 56' N., 25° 08' W., S.E. of St. Paul’s Bocks. Four or five Brown Gannets came flying round the ship several times during the day, but none settled or came within shot. They were a uniform dull brown all over, ’ except for a variable amount of white on the breast and belly and under the wings. The bill was very pale, whitish-blue at the base, and yellovTish-wliite for the front half. The size of the bird was rather less than that of the common white Gannet of British coasts. Some looked like young adults, others, one at least, looked like an old bird in moult. We were only from 20 to 50 miles from the South-Western Islets of Cape Verde Group, between the hours of 7 a.m. and noon, so these birds almost certainly came from there. July 18, 1910.— North-West of Ascension. A Gannet came in from the open ocean in the West at sundown, sailing and swooping down from a great height in long sweeps with closed wings. It remained round the ship all night and tried to settle on the yards as we could see by the bright moonlight, but we could not see its colour, and it was gone by daylight. It gave a loud squawk rather like a young rook. This might also have been S. dactylatra Less. (= S. cyanops Sundev.) described from Ascension Id. 39. Fregata wilsoni (Wilson’s Frigate-bird). Plate I, fig. 2. Fregata ariel Sharpe, “ Birds of South Trinidad,” Ibis, p. 214, 1904 [part] ; Fregata arid wilsoni Lowe, Nov. Zool. xxxi, p. 311, 1924 (South Trinidad). MATERIAL OBTAINED No. 21. 3 S. Trinidad. 28.vii.1910. E. A. W. Iris brown; bill bluish- white ; skin between mandibles bluish ; legs and feet very pale flesh-colour. Systematic Notes. — Wilson took the above bird when the “ Terra Nova ” touched at the island on her way south. He had collected another skin on the “ Discovery ” when off the island in September, 1901, and this skin is also preserved in the British Museum. BIRDS— LOWE AND KINNEAR 187 One of us (Xovitates Zool. 1924, vol. xxxi, p. 310) has already called attention to an interesting condition of plumage presented by these two specimens, viz. a brown- headed condition. This apparently represents one of two things, viz. either a young phase of plumage which, by a curious chance, has not previously been met with by collectors throughout the entire range of the arid group, or a phase which is anomalous. We have never seen anything similar either in the British Museum or at Tring, and one of us ( loc . ait. p. 34), as indicated above, thought fit to emphasise the point by distinguishing the South Trinidad form with a new name. As will be observed he made this island form a sub-species of Fregata anel. He now considers it wiser, until further material arrives to clear up the point, to regard the South Trinidad form provisionally as a distinct species, Fregata wilsoni. The following notes from Hr. Wilson’s diary refer to two species of Frigate- birds : — July 26, 1910. — South Trinidad Island. Anchored close off the shore of South Trinidad. There was an enormous number of birds flying about and around the island to the very summits, and before landing we shot two Frigate- birds — one the greater Frigate-bird with a bright red pouch, the other the lesser Frigate-bird with a greyish-blue pouch. We also shot one or two Petrels, 0. trini- tatis, and a Gygis crawfordi or two. We then landed and went to the top of the island, where there is a pretty thick bush of gnarled trees and an abundance of land-crabs. We shot Petrels, Gannets, and Gygis crawfordi. We saw absolutely no sign of any land-bird, nor was there any gull or skua or other bird about except the two Frigate birds, but they were never within gun shot on the shore, though they were occasionally to be shot from the ship. These Frigate-birds interest themselves in the ship’s wind-vane and mast- head. They are powerful flyers and they sail very beautifully over the mast-head, picking at the small piece of flannel or canvas attached to the vane. We saw nothing of their habitat on the island ; it may perhaps be at the other- end of the cliffs. They hunted the other birds as we occasionally saw from the ship. July 28, 1910. — 20° 57' S., 29° 07' W. Three Frigate-birds followed us out for about ten miles. The greater one shows his pouch when flying. 40. Fregata minor nicolli (Nicoll’s Frigate-bird). Plate I, fig. 1. Fregata minor nicolli Mathews, Austral. Av. Rec. ii, p. 118, 1914 (South Trinidad) ; Lowe, Nov. Zool. xsxi, p. 309, 1924 MATERIAL OBTAINED d South Trinidad Island. 28.vii.1910. E. A. W. Bill horn-vellow and grey ; skin under chin and throat scarlet ; iris brown ; feet and legs bright red. It says a good deal for the close and discriminating work of Mathews and Roth- schild when we note that at the time when Wilson took this bird in South Trinidad 11* 188 “TERRA NOVA” EXPEDITION only two species of Frigate-bird were recognised in the Catalogue of Birds in the British Museum (vol. xxvi), and that this very specimen has been described (in MSS.) as an example of Fregata aquila. At the present time five well-differentiated species — F. aquila ; F. minor ; F. magnificens ; F. andrewsi ; and F. arid — can be easily recognised, to say nothing of sub-species. There seems to be no doubt that the above-noted specimen taken by Wilson is a young male of the Fregata minor group. Its head has a very peculiar streaky or mottled appearance, and this at first sight might easily be ascribed to the dark greenish feathers of maturity “ coming in ” among the white head feathers of immaturity. A close inspection, however, reveals the fact that these dark feathers are irregularly fringed with white, a condition, or phase, of plumage which we have only once noted before, viz. in a specimen collected by Nicoll from this self-same island. The white of the foreneck has the “ iron rust-staining ” which is such a peculiar feature of the plumage of certain Frigate-birds. In the case of Fregata minor one of us has already called attention ( loc . cit , p. 300) to the fact that nestling and immature examples from the Indian and Pacific Oceans invariably exhibit a tinge of rusty, in greater or less pro- portion, on some part of the head, neck, or breast. This statement was true as far as it went, but it did not go far enough ; for in reality the rusty coloration is not only a geographical peculiarity but also a specific hall-mark confined to F. minor , F. ariel, and F. andrewsi. It is even absent from F. magnificens magnificens of the Galapagos, although those islands are in the Pacific. But F. m. magnificens may be presumed to be merely an outlying colony of F. m. rothschildi, which latter race occupies the whole basin of the Caribbean Sea and Mexican Gulf, and from which the rusty coloration is also absent. The fact that F. magnificens is found in the Pacific at all may be an “ accident ” dependent on the narrowness of the Isthmus of Panama or its actual absence in the late Tertiary ; so that genetically we may feel fairly certain that the centre of dispersal of F. magnificens was the Caribbean basin or American Mediter- ranean. From the Caribbean basin F. magnificens also spreads down into the Atlantic along the eastern shores of South America and across to the Cape Verde Islands. The immature of the species typical of the genus Fregata — Fregata aquila, confined to Ascension Island— are likewise devoid of any sign of rusty coloration. The island of South Trinidad in the South Atlantic is, as far as we are aware, the sole exception to the rule that the “ rusty tinge ” on immature examples of Fregata is confined to the Pacific and Indian Oceans. On this island two species of Fregata are found. In one, F. minor, the rusty tinge is conspicuously present ; in the other, F. wilsoni, the examples we have do not show it. This absence of “ rusty ” in the West Indian area (Caribbean and Gulf of Mexico) and Atlantic, and its presence in the Indian Ocean and the Pacific seems to us to be very remarkable in an “ ocean-going ” bird like Fregata where physical or physiological barriers seem so difficult to define ; but still more remarkable seems to be the fact that where an apparent climatic barrier in the shape of the cold “ roaring forties ” has BIRDS — LOWE AND KINNEAR 189 intervened between the Indian Ocean population of F. minor and the Atlantic-South Trinidad population it has not been effective, for the distribution of the “ rusty ” factor is now, at any rate, discontinuous although it may not necessarily have been so in the past, for when the climate in the south was warmer intervening islands may possibly have formed a connecting chain. INDEX. [Synonyms are indicated by italics .] Adamastor, 104, 131. adeliae, Caturrhactes, 107. adeliae, Pygoscelis, 104, 107, 109, 110, 111. aequinoctialis, Majaqueus, 131, 135, 142. aequinoctialis, Procellaria, 104, 135. aequinoctialis, Procellaria aequinoctialis, 135, 136. aequinoctialis steadi, Procellaria, 135. alba alba, Gygis, 104, 112, PI. 6. alba, Gygis alba, 104, 112, PL 6. alba Ossifraga, 147, 149. alba, Sterna, 112. alb us, Macronectes giganteus, 149. alexanderi, Heteroprion desolatus, 104, 162. alexanderi, Thalassogeron chrysostoma, 173, 176. andrewsi, Fregata, 188. antarctica, Catharacta, 115, 122. antarctica, Catharacta skua, 116, 117. antarctica lonnbergi, Catharacta, 121. antarctica, Megalestris, 113, 121. antarctica, Phoebetria palpebrata, 180. antarctica, Priocella, 104, 134. antarctica, Procellaria, 132. antarctica, Tbalassoeca, 104, 132, 133, PI. 7. antarcticns, Fulmarus, 134. Aptenodytes, 104, 105, 106. aquila, Fregata, 188. ariel, Fregata, 187, 188. arid wilsoni, Fregata, 186. arminjoniana, Oestrelata, 138, 139. arminjoniana, Oestrelata, 137. arminjoniana, Pterodroma, 104, 137, 138, 139, 140, 141, Pis. 3, 4 and 6. auduboni, Phoebetria palpebrata, 182. australis, Daption capensis., 159. bassi, Liomedea, 177. (bassi !) Thalassarche chlororhynchos, 178. brevicaudus , Puffinus, 130. brevicaudus, Puffinus tenuirostris, 104, 130. bulweri, Bulweria, 129. Bulweria, 104, 124, 125, 129. bulwerii, Bulweria, 104, 124, 129. bulwerii, Procellaria, 129. caerulea, Halobaena, 104, PI. 3. caerulea, Halobaena, 160. caerulea, Procellaria, 160. Calonectris, 104, 130. Candida, Gygis, 112. Candida, Pagodroma nivea, 146, 147. Candida, Procellaria, 142. capensis australis, Daption, 159. capensis, Daption, 104, 159. capensis, Procellaria, 159. carteri, Thalassogeron, 179. carteri, Thalassogeron, 177. castro, Oceanodroma, 104, 124. castro, Thalassidroma, 124. Catarrhactes, 107. Catharacta, 104, 113, 114, 115, 116, 121. chilensis, Catharacta, 115. chilensis, Catharacta skua, 116. chionophara, Oestrelata, 138. chionoptera, Diomedea, 164, 165.1 chionoptera, Diomedea, 162. chlororhynchos (bassi !), Thalassarche, 178. chlororhynchos, Diomedea, 177. chlororhynchos, Thalassarche, 104, 177. 178, 179, 180, Pis. 14 and 15. chlororhynchus , Thalassogeron, 177. chrysostoma alexanderi, Thalassogeron, 173, 176. chrysostoma, Diomedea, 173. chrysostoma harterti, Thalassogeron, 173. chrysostoma, Thalassarche, 104, 173, 174, 175, 176, 180, Pis. 9, 14 and 15. cinerea, Adamastor, 104, 131. cinerea, Procellaria, 131. cinereus, Priofinus, 131. clarkei, Catharacta skua, 116, 118, 119. confusa, Pagodroma, 146. confusa, Pagodroma, 142. cornicoides (palpebrata), Phoebetria, 183. 191 192 INDEX cornicoides, Phoebetria, 180, 182. coronoides, Phoebetria, 184. crawfordi, Gygis, 187. crawfordi, Gygis, 112. cryptoleucura, Oceanodroma, 124. culminata, Diomedea, 173. culminata mathewsi, Diomedea, 173, 176. culminatus, Thalassogeron, 173. cyanops, Sula, 186. Cymochorea, 124, 126. Cymodroma, 126, 127, 128. dactylatra, Sula, 186. Daption, 104, 159. desolatus alexanderi, Heteroprion, 104, 162. Diomedea, 104, 162, 163, 165, 166. dovei, Macronectes giganteus, 147, 149. edwardsii, Calonectris kuhli, 104, 130. edwardsii, Puffinus, 130. epomophora, Diomedea, 104, 165, 166, 167, 180. epomophora maccormicki, Diomedea, 166. Euphausia, 143. Euphausiacea, 169. eximius, Thalassarche, 180. eximius, Thalassogeron, 177, 178. exsul, Pelecanoides , 161. exsul, Pelecanoides urinatrix, 104, 161, 162. exulans, Diomedea, 104, 162, 163, 164, 165, 166, PI. 8. forsteri, Aptenodytes, 104, 105, 106. forsteri, Macronectes giganteus, 147. Fregata, 104, 186, Fregetta, 104, 126, 127, 128. Fregettornis, 127. fuliginosa, Phoebetria, 180, 182, 183. fusca, Diomedea, 182. fusca, Phoebetria, 104, 182, PI. 13. Garrodia, 126. gigantea, Macronectes, 183. gigantea, Ossifraga, 147. gigantea, Procellaria, 147. giganteus albus, Macronectes, 149. giganteus dovei, Macronectes, 147, 149. giganteus forsteri, Macronectes, 147. giganteus giganteus, Macronectes, 148, 149. giganteus halli, Macronectes , 147, 148. giganteus, Macronectes, 104, 147, 149, 155, 156, 157, 158, 159, 163. giganteus, Macronectes giganteus, 148, 149. giganteus solanderi, Macronectes, 147, 148. giganteus wilsoni, Macronectes, 147, 150. glacialoides, Priocclla, 134. grallaria, Procellaria, 127. gravis, Puffinus, 130. Gygis, 104, 112, 187. Halobaena, 104, 160, PI. 3. Halocyptena, 126. halli, Macronectes giganteus, 147, 148. harterti, Thalassogeron chrysostoma , 173. Heteroprion, 104, 162. hirundinacea, Sterna, 111. howensis, Cymodroma, 127. huttoni, Phoebetria palpebrata, 180. impavida, Thalassarche melanophris, 167. intercedens, Catharacta lonnbergi, 115 intercedens, Catharacta skua, 104, 113, 116, 119, kuhli edwardsii, Calonectris, 104, 130. lessoni, Oestrelata, 136. lessoni, Pterodroma, 104, PI. 5. lessonii, Procellaria, 136. lessonii, Pterodroma, 136. leucogaster, Pelecanus, 186. leucogaster, Sula, 104, 186. leucogaster, Thalassidroma, 127. leucorhoa, Cymochorea, 124. lonnbergi, Catharacta, 115. lonnbergi, Catharacta antarctica, 121. lonnbergi, Catharacta skua, 104, 116, 120, 121. lonnbergi intercedens, Catharacta, 113, 115. maccormicki, Catharacta skua, 104, 116, 121, 123. maccormicki, Megalestris, 121. maccormicki, Stercorarius , 121. macronectes, 104, 142, 147, 148, 149, 151, 153, 154, 155, 156, 157, 158, 159, 163, 165, 183. Macroptera macroptera, Pterodroma, 104, 141. macroptera, Oestrelata, 141. macroptera, Procellaria, 141. macroptera, Pterodroma macroptera, 104, 141. magnificens, Fregata, 188. magnificens rothschildi, Fregata, 188. majaques, 131, 142. mathewsi, Diomedea culminatea, 173, 176, 177. Megalestris, 113, 114, 121. melanogaster, Cymodroma, 126, 128. melanogaster, Fregatta, 126, 128. melanogaster, Fregatta tropica, 104, 126. melanogaster, Thalassidroma, 126. melanophris, Diomedea, 167. INDEX 193 melanophris impavida, Thalassarche, 167. melanophris, Thalassarche, 104, 167, 168, 170, 171, 172, 179, Pis. 9, 14 and 15. melanophrys , Diomedea, 167, 168, 169, 177. minor, Fregata, 188, 189. minor nicolli, Fregata, 104. moestissima, Fregetta, 127. mollis mollis, Pterodroma, 104, 141, PI. 6. mollis, Oestrelata, 141, 142. mollis, Pterodroma mollis, 104, PI. 6. neglecta, Oestrelata, 139. Nesofregetta, 127. nicolli, Fregata minor, 104, 187, PI. 1. nivea Candida, Pagodroma, 146, 147. nivea nivea, Pagodroma, 104, 146, 147. nivea, Pagodroma, 142, 143, 144, 145. nivea, Pagodroma nivea, 104, 146, 147. nivea, Procellaria, 142. oceanica, Procellaria, 124. oceanicus, Oceanites, 104, 124, 125. Oceanites, 104, 124, 125, 126. Oceanitidae, 127. Oceanodroma, 104, 124, 126. Oestrelata, 136, 137, 138, 139, 142, 187. Ossifraga, 147, 149. Pagodroma, 104, 142-145. palpebrala antarctica, Phoebetria, 180. palpebrata auduboni, Phoebetria, 182. palpebrala, Diomedea, 180. palpebrala huttoni, Phoebetria, 180. palpebrata, Phoebetria, 104, 180, 182, 184, Pis. 11, 12 and 13. (palpebrata), Phoebetria cornicoides, 183 parkinsoni, Majaqueus, 135. pelagica, Procellaria, 125. Pelagodroma, 126. Pelecanoides, 104, 161, 162. Pelecanus, 184. Phoebetria, 104, 180, 182. piscatrix, Sula, 184. Procellaria, 104, 125, 127, 135. Procellaria, 124, 129, 131, 132, 141, 147, 159. Priocella, 104, 134. Prion, 104, 161. Prion sp., PL 2 Pterodroma, 104, 136, 137, 141. Pufhnus, 104, 130. Pygoscelis, 104, 107, 109, 110, 111. regia, Diomedea, 163, 164, 180. rothschildi, Fregata magnificens, 188. skua antarctica, Catharacta, 116, 117. skua, Catharacta, 115, 116. skua, Catharacta skua, 116. skua chilensis, Catharacta, 116. skua clarkei, Catharacta, 116, 118, 119. skua intercedens, Catharacta, 104, 113, 116, 119. skua lonnbergi, Catharacta, 104, 116, 120, 121. skua maccorruicki, Catharacta, 104, 116, 121, 123. skua skua, Catharacta, 116. skua wilsoni, Catharacta, 122. solanderi, Macronectes giganteus, 148. solanderi, Macronectes giganteus, 147. steadi, Procellaria aequinoctialis, 135. Stercorarius, 121. Sterna, 104, 111, 112, Sula, 104, 184, 186. sula, Pelecanus, 184. sula, Sula, 104, 184, 186. tenuirostris brevicaudus, Puffinus, 104, 130. lenuirostris , Puffinus, 130. Thalassarche, 104, 167. Thalassidroma, 126, 127. Thalassidroma, 124, 126. Thalassoeca, 104, 132. Thalassogeron, 173. trinitatis, Oestrelata, 137, 138, 139, 187. tropica, Fregetta, 127. tropica, Fregetta tropica, 104, 128. tropica melanogaster, Fregetta, 104, 126. tropica, Thalassidroma, 128. tropica tropica, Fregetta, 104, 126, 128. urinatrix exsul, Pelecanoides, 104, 161, 162. vittata, Sterna, 111. wilsoni, Catharacta skua, 122. wilsoni, Catharacta maccormicki, 121. wilsoni, Fregata, 104, 186, 187, PI. 1. wilsoni, Fregata arid, 186. wilsoni, Macronectes giganteus, 150. ivilsoni, Macronectes giganteus, 147. wilsoni, Oestrelata, 138. ivilsoni, Oestrelata, 137. PRINTED IN GREAT BRITAIN BY WILLIAM CLOWES AND SONS, LIMITED, LONDON AND BECCLES. Brit. Mus. (Nat. Hist.) Brit. Antarctic (Terra Nova) Exped.. 1910. Zoology, Vol. IV. No. 5. Birds, PI. I. E A Wilson, del B ale &. D aniels s on , Birds, PI. II. Brit. Antarctic (Terra Nova) Exped. 1910. Brit. Mus. (Nat. Hist.) Z oology, Vol. IV. No. 5 £ A. Wilson, del Bale &. Danielson . L.x<* Brit. Mus. (Nat. Hist.) Birds, PI. m. Brit. Antarctic (Terra Nova) Exped. 1910. Zoology, Vol. IV. No. 5. E- A. Wilson, del . IB ale &Danielsson,LtA Brit. Antarctic (Terra Nova) Exped. 1910. Brit. Mus. (Nat. Hist.) Zoology, Vol. IV No. 5. Birds, PI. IV. Birds, PI. IV. E . A. Wilson., del. Bale & Danielsscra , Brit. Mns. (Nat. Hist.) Brit. Antarctic (Terra Nova) Exped. 1910. Z oology, 'Vol. IV. No. 5. Birds, PI. V. E. A. Wilson, del. B ale &.D am els son , F/r y, Brit.Mns. (Nat. Hist.) Brit. Antarctic (Terra Nova) Exped. 1910. Zoology, Vol. IV. No. 5. Birds, PI. VI. E A. Wilson, del. B ale & I) aniels s on. , V Brit. Antarctic (Terra Nova) Expect 19JQ. hbis (Nat. Hist.) Zoology;' Vol. IV. No. 5^irds“Plate pL yjjj t y 9. £0 X\ .nr/ aTAJ? .sgamulcj lo 89&srfq Inai^Sib gaiwodS .sesaoitedlA lo B9ibi/t8 m\vs«to'sG.) eaoxtsdlA gahebrisAf -.8-1 .(moAqcvmoqn oAmWsQ.) aeoi«tBcfIA [bt(o3 — .01-6 .(m»W®9 osbamoKI) aaoilfidfA gnmbiiaV/ — .SI— II io. . 3$ • 2. Y 11. yvitm 5. NJ .del. 12. Buie ]De^ifQ&!»atL, Xr^- .IIIV 3j£iq— eblia PLATE VIII. Studies of Albatrosses. Showing different phases of plumage. I- 8.- — Wandering Albatross (Diomedea exulans). 9-10. — Royal Albatross (Diomedea epomophora). II— 12. — Wandering Albatross ( Diomedea exulans). Brit. Mus. (Nat. Hist.) Birds, PI. VIII Brit. Antarctic (Terra Nova) Exped. 1910. Zoology. Vol. IV. No. 5. /V > ■ j 4*. 12. ^ A. Wilson, del. Bale & Daniels son., Ln^- % Birds, PI. IX. A. Wilson., del. Bale &. Daniels son. Brit. Mns. (Nat. Hist.) Birds, PI. X. Brit. Antarctic (Terra Nova) Exped. 1910. Zoology. Vol. IV. No. 5. E A . Wilson . del 6. Bale & Danielsson . Lt Brit. Antarctic (Terra. Nova) Exped. 1910. Brit. Mus. (Nat. Hist.) Zoology, Vol. IV. No. 5. Birds, PI XI Birds, PI. XI A. Wilson, del. Bale &. Daniels: Bnt.Mus. (Nat. Hist.) Birds, PL XII Brit. Antarctic (Terra Nova) Exped. 1910. Zoology. Vol. IV. No. 5. E A. Wilson, del Bale ^Daniels son^L1.^ Brit. Mus. (Nat. Hist.) Brit. Antarctic (Terra Nova) Exped. 1910. Zoology.Vol.IV, No. 5. Birds, PI. XIII. Bale ^Daxiielssonlli1- £. A. Wilson, del. * _ Brit. Mois. (Nat. Hist.) Birds, PI. XIV. Brit. Antarctic (Terra Nova) Exped. 1910. Zoology, Vol. IV. No. 5. E .-A_Wils on , del . Bale &. Daniels son, 1&- Brit. Mns. (Nat. Hist.) Birds, PI. X V. Brit. Antarctic (Terra Nova) Exped. 1910. Zoology, Vol. IV. No. 5. E . A. "Wilson, del . Bale &: D aniels s on , Irf3- Brit. Mus. (Nat. Hist.) Birds, PI. XVI. Brit. Antarctic (Terra Nova) Exped. 1910. Zoology, Vol. IV. No. 5. **■ A.. Wilson, del. Bale &. BamelssorL, 11*^- s Jfc'.jt , BRITISH MUSEUM (NA" HISTORY). tRITISH 4NTARCTIG (‘ ’) EXPEDITION, 1910 IATURAL HISTORY REPORT. ■ Ts . - * ZOOLOGY. VOL. IV, No. 6. Pp. 195-251 REPORT ON THE TUNICATA. PART I.-DOLIOLIDA. BY WALTER GARSTANC, M.A., D.Sc. rao Irv >•. £ raorMop or : octroy in m uruvntfvnr Of usds. WITH EIGHT TEXT-FIGURES. B. M. it /'■ ...I * ■ P« . v.o PRINTED* BY ORDER OF THE TRUSTEES OF THE BR1TISD MUSEUM A'A LONDON! I msm&fk A, SOLD BY Mjtrrcjj, Ltd., II Grafton Street, New Bore Street, W.l; Dtruc * Co,, Ltd., XI Old Bond Street, London, W.l ; Oxedkd Uwiversjty The British Mdsedm (Natural u\f , ;; T i 2Th May, Press, Wab'wick Square, London, E.C.4 LND AT . ural History), Cromwell Road. S.W.7 '± .-"Y-f : -1 ■ : . ytl [AH rights rear.rv ed.\ _ .• ' i . •V * Poice Five Shillings. British HDuseum (IRatural Ibistorp). This is No. £ of 2 5 copies of “ Terra Nova ” Zoology , Vol. IV AV 6, Report on the Tunicata , Part /., Doliolida , printed on Special paper. 195 DOLIOLIDA. BY PROF. W. GARSTANG. WITH 8 TEXT-FIGURES. PAGE 1. Introduction and Summary ........ 196 2. Evolution and Classification of Doliolids . . . . . . 198 3. Myomeristic Growth-limits ........ 202 (a) The Branchial Septum ........ 202 (b) The Testis .......... 204 4. Identification and Nomenclature ....... 210 (a) Critical Survey ......... 210 {b) The Endorsement of Specific Names ..... 226 5. Old Nurses ........... 229 (a) Classification by Muscles ....... 229 (b) Identification of D. gegenbauri and D. miilleri . . . . 231 (c) Nurses without Otoliths ....... 237 (d) Early Stages of D. denticulatum ...... 239 (e) Relation of Oozooidal Types to the System .... 239 6. Systematic Summary and Records of Distribution ..... 241 (a) Systematic Summary . . . • • . .241 {b) Records of Distribution . . • . • .241 (i) Atlantic ........ 242 (ii) S. Pacific ........ 243 (iii) Southern Ocean ....... 249 7. Abbreviations .......... 250 8. Literature ........... 251 rv. 6. 1 196 “TERRA NOVA” EXPEDITION. I. INTRODUCTION AND SUMMARY. My first words upon submitting this instalment of my report on the “ Terra Nova ” Tunicata must be an acknowledgment of the great forbearance shown to me by the Trustees and Officers of the British Museum in consequence of my long delay in producing one. When I undertook the task fifteen years ago, the prospect of examining and describing a large and varied collection of Tunicata made a strong appeal to my interests. I had then, I must admit, no idea that the time estimated for this part of the work, even under the distracting conditions of the Great War and its aftermath, would prove but a fraction of that required to master all sorts of nomenclatorial difficulties that lay ahead. I have learned that the only possible way is to proceed by instalments, and to complete one section at a time. In preparing this report on the small but fascinating group of Doliolids, my work has been much lightened by the precision and care of my immediate predecessors, especially Borgert and Neumann, whose reports on the collections of the “ National,” “ Valdivia,” and “ Gauss ” will long remain classic. No really new species are here added, though I have raised a variety described by Neumann to that rank ( sigmoides , n. sp.), reduced the status of several others, viz. resistibile, Neum., krohni, Borgert ( non Herdman), and tritonis, Herd, as varieties of intermedium, mulleri and gegenbauri, and completely revised the classification. Any novelty my report contains lies rather in the completeness with which I have attempted to record the material, and in the more extensive use of measurement. Every specimen, common or scarce, large or small, has been examined, measured and recorded, and the details summarised for every station separately. I hope, therefore, that the report will serve as a store of data for subsequent use in various directions. Many, when studying some special problem, must have shared my experience of wanting to know something about the sizes, numbers, and special character of the local forms of some species or other, only to find that on these points the reports of the great expeditions were usually silent. Yet it is only from them, generally speaking, that one can expect light to be thrown on the oecology and bionomics of truly oceanic forms. My own use of the collections, apart from the mere determination of species, has been directed mainly towards a study of the variation of certain characters in the commoner species, and to an attempt to discriminate between different kinds of “ old nurse,” which hitherto have proved very refractory. In both respects distinct progress, I think, has been made, and the results justify the belief that, if biometric methods were more extensively applied to other collections, the results would amply repay the time and labour expended. The New Zealand waters investigated by the “ Terra Nova ” have a special interest for the student of the Thaliacea, since it was here, eighty years ago, that Huxley rediscovered the Doliolum denticulatum of Quoy and Gaimard, and recognised in it a new type of pelagic Ascidian. DOLIOLIDA— GARSTANG. 197 Tiie collections of the “ Terra Nova ” naturally include few Doliolids from the Antarctic region proper, but the single species represented, the D. resistibile of Neumann, was actually taken within the Antarctic Circle, further south than any Doliolid yet recorded. SUMMARY OF RESULTS. 1. Borgert’s sub-genera Doliolina (type mulleri) and Dolioletta (type gegenbauri) are raised to generic rank with certain restrictions ; a new genus, Dolioloides, is pro- posed for D. rarum ; and the original genus, Doliolum, of Quoy and Gaimard (type denticulatum), is restricted in scope. The form and relations of the alimentary canal in gonozooid and phorozooid are taken as diagnostic. 2. The muscular rings, through their connective tissue adhesions, are shown to act as obstacles to the forward extension of the branchial septum and testis during the later stages of growth, thus setting up myomeristic growth-limits, which are prominent features in the diagnosis of most species. 3. These myomeristic limits depend on the period of mesoblastic adhesions between inner and outer membranes, which may be accelerated or retarded by environmental conditions, such as temperature and food-supply. The differential characters of numerous pairs of so-called species are probably therefore no more than environmental modifications, e.g. tritonis and gegenbauri , intermedium and resistibile, mulleri and Borgert’s (but not Herdman’s) krohni, denticulatum and nationalis, valdivice and mirabilis ( =chuni ), Herdman’s krohni and Neumann’s indicum. When the “ species ” in these pairs are inseparable by constant characters, they have here been amalgamated under the prior name. This applies already to the first three pairs, and will probably apply ultimately to the others. The unbridged gap between denticulatum and nationalis is very narrow. 4. Doliolina krohni (Herdman) was originally described as possessing an oblique gill-septum (cf. D. indicum, Neumann). None of the forms of krohni described by Traustedt, Borgert, or Neumann display this character. They are here treated as (a) a polystigmatic variety of D. mulleri with ventral testis (all three authors), (b) a variety of D. intermedium with forwardly directed testis (Borgert, Neumann), and (c) a new type altogether, with forwardly directed testis, but with ectodermal filaments and a sigmoidal gill-septum (Neumann). For the last a new species, D. sigmmdes, n. sp., is here proposed. 5. In Dolioletta mirabilis (Korotneff), which now includes D. chuni Neumann, the 6th muscle-band is shown to be incomplete ventrally both in gonozooid and phorozooid. Fedele’s distinction between these two types of zooid as regards the limits of the gill- septum is invalid. 6. Contributions have been made as follows towards the identification of various types of oozooid (larval and old nurses) : (a) Dolioloides rarum. Gegenbaur’s figs. 12, 13 (1856, Taf. NVI) appear to represent 198 “TERRA NOVA” EXPEDITION. the oldest known oozooid of this species, with muscle-bands twice as wide as the inter- spaces at 5-5 mm., and with viscera still intact, thus easily admitting of discrimination from others. (b) Doliolina intermedium, var. resistibile (Neumann) possesses very slender muscle- bands throughout life in the old nurse condition (<15 mm.). Its brain lies close before M5. These points identify it with the first of Gegenbaur's unnamed oozooids (l.c. fig. 14). From his account (p. 300) the larval gut may be inferred to be S-shaped, as in Borgert’s gonozooid (1894, Taf. VI, fig. 16), not U-shaped as in D. mulleri. (c) Young stages assigned to Doliolum denticulatum possess extremely short endostyles (M2-M4), and closely resemble two figures of Keferstein and Ehlers (1861, Taf. X, figs. 3, 4). Grobben’s metamorphosing specimens attributed to this species (1882, figs. 4, 5) should probably be referred to D. gegenbauri. (d) Two types of broad-banded old nurse have been definitely distinguished by measurement of the muscle-bands. In one, attributed to Doliolina mulleri, M4 and M5 are approximately equal and widest. In another, attributed to Dolioletta gegenbauri, M3 and M4 are the widest muscles. They remain approximately equal up to a size of 6 mm., above which M3 predominates. A further distinction is to be found in the otolith, which usually drops away after death in Doliolina (both mulleri and intermedium), but rarely in D. gegenbauri. (e) The old nurses of Doliolum denticulatum are completely covered by a muscular cuirass (representing M2-M8), as figured by Grobben. 7. All specimens of all stages taken have been measured, localised, and reported upon under their separate station-numbers. II. THE EVOLUTION AND CLASSIFICATION OF DOLIOLIDS. (a) EVOLUTION. That the ancestors of the Thaliacea were sessile primitive Ascidians, with a tailed larva and a metamorphic life-history, and that, of existing Thaliacea, the Doliolids con- stitute the group most highly adapted to an active pelagic existence is here taken for granted. The evidence in support of these views has already been set out in a series of earlier papers (Garstang, 1928, 1929), so that only a few leading points, which bear on the classification of the group, need here be emphasised. Unlike the Pyro- somata and Salps, the Doliolids have lost the sinistral twist of the intestinal loop, characteristic of their Ascidian ancestors, both in larva, oozooid, and bud, and the musculature retains no sign of its originally plexiform character except in its develop- ment, which at an early stage recalls the biradiate plexus, concentrated dorsally, of the Pyrosomata fixata (cf. Neumann, Metcalf), which is also recognisable in the “ desmomyarian ” condition of the Salps (Garstang, 1929). As in Pyrosoma the cloaca DOLIOLIDA— GARSTANG. 199 has rotated completely to a position directly behind the pharynx, but, unlike Pyrosoma, both the oozooid and its buds have become adapted to an actively motile and independent life, by metamerisation of the musculature and restriction of the gill-slits to a posterior position in the pharyngo-atrial septum. All these points are functionally related to one another, and contribute to the efficiency of Doliolids as active planktonic organisms. The Salps also have succeeded in resolving the ancestral floating colonies into chains of semi-independent sexual individuals, but in them the free life of the buds is usually brief, the colonial stage is still dominant, and the task of feeding the young buds falls entirely on the sexless oozooid. Hence the metamerisa- tion of the muscles is less complete in Salps than in Doliolids, and more complete in their oozooids than in their buds. The Doliolids complete the emancipation of their sexual buds (gonozooids) by the highly specialised means of a polymorphic bud- generation, in which the two elder of the three types (trophozooid and phorozooid) remain asexual and successively take over the food-providing function from the parent oozooid, which then specialises entirely as a locomotive vehicle for its progeny (" old nurse ” stage). In harmony with its restricted functions, the oozooid shows few variations of structure, except as regards its equipment for locomotion. The sexual buds, on the other hand, have to swim, feed and reproduce sexually. They show accordingly a much greater diversity. The chief aim of the Doliolid taxonomer at present is to reconstruct out of frag- mentary material the successive phases of an unknown number of complex life-histories, with an approximate knowledge of one, and the debatable sequences of two other life histories as a guide. Direct links between the sexual and asexual generations are rarely to be found, and the assumption of some deep-seated resemblance between the structure of oozooid and blastozooid on which the life-history of Doliolina millleri was traced out by Grobben, is already known to have a very limited application in other cases. The most hopeful line of procedure is to build up an increasingly accurate classification of the bud-generation (especially gonozooids), towards the foundations of which we owe much to Borgert and Neumann, and another of the oozooids, in the expectation of being able to dovetail one into the other as the lines of cross-relationship become apparent. We can, indeed, already see that in the course of evolution, as might be expected, there have been independent adaptations of oozooid and blastozooid, involving an increasing divergence between these generations as we pass from lower to higher types. Unfortunately the reporters on previous large collections, unable to identify the oozooids before them, have made little attempt to discriminate between their various forms, still less to describe and figure them, except occasionally in the early stages, so that the great collections of the “ Challenger,” “ National,” “ Siboga ” and “ Gauss ” have been in this particular respect of little use to later workers. Although the collections of the “ Terra Nova ” are relatively small and local, I have attempted, by 200 “TERRA NOVA” EXPEDITION. means of detailed measurements, to discriminate between different types of these neglected “ old nurses,” and to leave a record which may assist further attempts in that direction. (6) CLASSIFICATION. In his revision of the results of the German Plankton Expedition, Borgert (1894) arranged the ten species which he then recognised (but in three cases with serious doubts as to their validity) under two sub-genera, according to the vertical or strongly arched character of the branchial septum of the gonozooids. This division, though marking a useful advance in method, was never really valid, for, as shown below (p. 214), Borgert misunderstood Herdman’s account of the structure of D. krolmi, in which the septum is neither vertical nor “ vorgewolbt.” Borgert’s arrangement was adopted, however, by Neumann (1906, 1913), and several additional species were incorporated, though the structure both of indicum and resistibile was at variance with the definition of their sub-genus. Moreover, the names applied by Borgert to the two divisions of the genus, if generally used, would have led to the gradual elimination of Doliolum itself, though by every rule of nomenclature this name is inseparable from its type denticulatum. Borgert’s division in reality marks a horizontal cleavage of Doliolids into low and advanced types, but a surer basis for such a division, as well as for a phyletic classifica- tion, is to be found in the arrangement of the alimentary canal. This, in the species of the Doliolina section, lies in the median sagittal plane, as also in Doliopsis (—Anchinia), while in Dolioletta, coupled with a lowering of the oesophageal aperture, the stomach is bent over so as to lie flat upon the ventral body- wall, on its originally right side, and the intestine itself is displaced to the right (cf. figs. 1, 4, 7). This torsion, as I have elsewhere pointed out (1929 b) is the exact opposite of that which prevails in the Ascidians and Pyrosoma. It follows upon the backward rotation of the cloacal cavity, and preserves the balance of the body for an active career after liberation of the asymmetrical gonads from the intestinal loop. But at each of the two levels of this morphological change there is a further distinction to be drawn. In the lower division the gut, though without its sinistral twist, still retains the loop-like form of Ascidians, with a descending limb (the oesophagus and stomach), and an ascending limb (the rectum), in every species except D. rarum of Grobben, in which the intestine is straightened out axially along the floor of the cloaca, a condition which, with certain differences of detail, is again met with in the oozooids of the higher forms. In the gonozooids and phorozooids of these forms, however (i.e. Dolioletta of Borgert), the twist of the intestine to the right takes place in two different ways, viz. as a close spire in the gegenbauri-tritonis group, and as a "wide horizontal arch in the denticidatum group, the anus remaining in the plane of the 6th muscle-band, but median in the 1st group, and definitely parietal in the 2nd. There are thus four sharply distinguishable arrangements of the gut in the DOLIOLIDA— GARSTANG. 201 gonozooids of the various species (text-fig. 1), and these I submit should form the basis of a quadripartite subdivision of the genus, as follows 1. Doliolina (s. str.). — Alimentary canal forming an upright U- or S-shaped loop in the sagittal plane, e.g. mulleri, intermedium. 2. Dolioloides (gen. nov.). — Alimentary canal extended horizontally in the sagittal plane, with sub-terminal anus, e.g. rarutn. 3. Dolioletta (s. str.). — Alimentary canal forming a close dextral coil in the middle of the cloacal floor with median anus, e.g. gegenbauri. 4. Doliolum (s. str.). — Alimentary canal forming a wide dextral arch round the cloacal floor, with anus parietal, on the right side, e.g. denticulatum. These distinctions will be seen in the sequel to be associated with definite com- binations of other and more variable characters, the more archaic forms of any character being found in the species of the first two genera, the more advanced forms in the last two, Doliolum itself usually furnish- ing the climax. Thus, to return to the gill-slits, the more primi- tive forms of branchial septum, which resemble, or deviate least from, its arrangement in the oozooid, are found only in Doliolina and Dolioloides ; the septum is extended further forwards, both dorsally and ventrally, in Dolioletta ; and acquires its maximum extension in Doliolum. Similarly in the oozooid, so far as our present knowledge goes, the muscular hoops are narrowest, or most widely separated, in Doliolina and Dolioloides, broad and closely set in Dolioletta, finally fused into a continuous sheet in Doliolum. The separation of Dolioloides from Doliolina is justified not only by the peculiarities of its gut (which are exhibited in both generations), but by the fact that in this type alone the oozooid is known to attain a considerable size (5-5 mm.) without undergoing the atrophy of the alimentary canal which seems to prevail in all the others, while, correlatively, there is still no evidence that it gives rise to asexual locomotive buds, or phorozooids. Moreover, apart from its greater validity, a primary division of the genus in terms of the alimentary canal has a great practical advantage over schemes based on variations of the gill-septum, since the distinctive characters are readily recognised 4 5 6 7 8 4 5 6 7 8 45678 45678 Fig. 1 . — Showing the arrangement of the alimentary canal in the proposed genera of Doliolidae. 202 “TERRA NOVA” EXPEDITION. even in badly damaged specimens, while the distribution of the gill-slits is difficult to determine when the specimens are contracted or injured by intrusive Copepods. Some examples of the serious confusion which has been wrought by definitions based on the gill-septum will be found below (pp. 214, 225). Before proceeding to some complicated problems of identification and nomenclature in particular species, it will be convenient to examine certain peculiarities of Doliolid growth which materially affect the question of intra-specific variation. III. MYOMERISTIC GROWTH-LIMITS. It is impossible to study either the variations of growth within a Doliolid species or the characters distinctive of related species without realising that the ring-like muscu- lature of the body exercises a kind of metameric control over the general organisation. In old nurses (oozooids), when the alimentary canal has disappeared, the few organs left are seen to be distributed among the various intermuscular spaces — the otocyst in the 3rd, the brain in the 4th, the heart and stolon in the 5th. Closely related species or varieties of sexual zooids may differ in little or nothing more than whether the series of gill-slits begins dorsally, or ends ventrally, at a particular muscle-band or its neigh- bour, or whether the testis extends forwards to the 4th, 3rd or 2nd. The ovary lies in the 5th or the 6th intermuscular space ; the anus in the 5th, 6th or 7th, etc. In the case of the axial organs of the alimentary canal this relation to the muscle- segments is doubtless largely accidental, and variations are so described merely because the muscle-rings furnish a convenient standard of reference. Thus the endostyle may be described as beginning half-way between two muscle-bands in one species, and exactly at a muscle-ring in another. Nevertheless, over and above this arbitrary use of the muscles as a standard of reference, there is a certain pseudo-metamerism of variation which is no delusion, and is particularly clear in the case of the extent of the gill-slits and of the testis. As the discrimination of species depends largely on these two points, the question demands examination, for the obliquity of the major part of the septum seems to rule out any special utility in the relation, which is merely one of limits, and either the formation of new species would appear to be a matter of mutational jumps from one muscle-ring to the next, or the apparent jumps are conditioned by certain checks to extension imposed by the presence of the muscular rings. The latter seems to be the real explanation. (a) The Branchial Septum. — What is commonly described as the branchial septum is the two-layered membrane, perforated by the gill-slits, which is slung across the cavity of the body between the pharynx in front and the cloaca behind. It is spoken of as being “ attached ” to the body-wall along a curving line which terminates dorsally and ventrally at certain points, usually close to one or other of the muscular rings. The word is unfortunate, for in point of fact the “ line of attachment ” simply marks the parietal boundary between pharyngeal and cloacal spaces, along which their respective epithelia break away from the body wall to become apposed to one another. DOLIOLIDA— GARSTANG. 203 In front of the line the pharyngeal epithelium, behind it the cloacal epithelium, are each closely united with the body- wall by fine connective-tissue strands, between which the blood circulates. In this parietal blood-sinus lie the various muscular rings, the contraction of which drives water through the body (cf. p. 240 *). The arrangement might appear at first to be mechanically imperfect, and it certainly differs from that in Salps, in which the muscle-bands are directly fused with the internal membrane, as Keferstein and Ehlers remarked long ago (1861, p. 56). Its effectiveness, however, is attested by the marvellous agility of the living animals, and clearly depends on the firmness of the connective-tissue connexions, especially along the lines of the muscular rings. Sections show that the enclosing membranes of the body-wall adhere to the muscle-bands much more closely and uniformly than they do to one another in the inter-muscular spaces, where the enclosed blood-sinus is capable of considerable changes of volume, pointing to a relative sparsity, greater length, and less rigidity in the fibres which traverse it. Now the first stages of the branchial septum are accomplished in development by the cloacal invagination sending certain diverticula forwards, dorsally, ventrally, or in both regions, on either side of the median plane (cf. Neumann, 1906, Taf. X, figs. 5, 6). These may be compared to hollow wedges driven in between ectoderm and endoderm at a stage when the muscles are present as rings of undifferentiated mesoderm attached to the ectodermal wall, but still quite independent of the pharyngeal and cloacal epithelia (Neumann, l.c. Taf. X, fig. 10 ; XI, figs. 16, 20). At this stage the hypobranchial diverticula reach the 4th muscle-band in denticulatum (l.c. X, 6), and there is no obstacle as yet to their further extension. But in gegenbauri they have only reached the 5th band (fig. 7), and this stage is followed by one of general growth and relatively greater pharyngeal expansion (fig. 8). From the magnification indicated the length of the bud (Mx to M8), and the mid-pharyngeal width (along M4) in the two cases are 0-41 X 0-30 and 0-71x0-50 mm., yielding percentage ratios of pharyngeal width to body-length of 66 per cent, and 70 per cent, respectively. The expansion of the pharynx inevitably brings its walls into contact with the muscle-rings, and so initiates the intimate binding of one to the other, on which their functioning depends. The connexion so begun blocks the path for further extensions of the diverticula, and the gill-septum accordingly persists in the position typical of gegenbauri ; but it is obvious that a slight delay in the process of expansion and adhesion would enable the ventral diverticula to cross M5 and realise the condition typical of tritonis (cf. Borgert’s corresponding figure of a gonozooid bud, 1894, Taf. VI, fig. 18).* Similarly, if the process of expansion took place in denticulatum shortly after the * Remarkably enough a figure of Uljanin’s (1884, Taf. XII, fig. 7), representing a median bud, shows this to have actually happened in the offspring of one of the great Mediterranean nurses which Neumann has claimed to be gegenbauri — a convincing proof of the specific identity of gegenbauri and tritonis. Owing to Uljanin naming it ehrenberqii, I did not detect its significance until after this chapter had been written (cf. pp. 216, 224). IV. 6. 2 204 “TERRA NOVA” EXPEDITION. stage of Neumann’s fig. 6, the result would be a specimen of nationalis, with a branchial septum “attached” ventrally behind the 4th ring (cf. Borgert’s gonozooid bud, Taf. V, fig. 8). Actually, to make a denticulatum, the process is retarded, and the gill-septum moves forward towards the next (3rd) ring before the expansion and adhesion take place. In this way we see that the differences between the members of these two pairs of so-called “ species ” are merely matters of a little “ more or less ” in the growth-rate of the diverticula, and in the rate of pharyngeal differentiation. This difference is associated with one of size, the buds of natiomlis being smaller than those of denticulatum. Neumann’s fig. 6 represents a stage actually earlier than Borgert’s fig. 8, as shown by the less differentiated condition of the brain, endostyle, peripharyngeal bands and other organs. Yet from the scale of magnification given one can determine that the oro-atrial length even of Neumann’s rudimentary bud (0-28 mm.) is as great or greater than that of the highly differentiated natiomlis bud (0‘27 mm.). The fact that the nationalis bud is a gonozooid and the other a phorozooid makes no difference to the argument, since the completed structure of these types of zooid is the same, apart from the presence or absence of gonads. There is thus a possibility — I put it no higher for the moment — that the “ specific ” differences in the limiting attachments of the gill-septum in such pairs of species as nationalis and denticulatum, gegenbauri and tritonis, intermedium and resistibile, val- divice and mirabilis ( =chuni ), are ultimately matters of differential growth-rates, and doubtless modifiable by nutrition, temperature, and similar environmental factors, as in other cases. Morphologically in the adult there is a segmental discontinuity in most of these limits, but dynamically in the bud the growth-process which underlies them — the penetrative advance of the cloacal diverticula — is continuous until it is checked by mesodermal differentiation. This operates segmentally, beneath the muscle-rings, and thus produces a discontinuity in the end-results, if there are any differences in either of the growth-rates involved. I venture to think that if the differences in question were not rendered conspicuous by the numerical jump from one muscle-ring to the next, they would be regarded as coming within the range of natural intra-specific variation, dependent, as in other cases, on environmental as well as constitutional influences. It is doubtful if the test of experiment can ever be applied to these delicate oceanic organisms. In the mean- time judgment has to be exercised on the question whether they are varietal or specific, and I propose to deal with each case on its merits (see pp. 210-226 below). (b) The Testis. — In several species of Doliolina the testis preserves very primitive relations reminiscent of conditions in the Enterogonous Ascidians, in which it is lodged in the intestinal loop. Thus in D. mulleri it lies alongside the ascending limb of the intestinal loop, but, owing to the evolutional rotation backwards of the rectum, with the cloaca, on its left, instead of its right side. Its growth here is limited to the vertical plane, in which, when it is fully developed, it forms a ventral hernia (Grobben, 1882, DOLIOLIDA— GARSTANG. 205 Taf. Ill, fig. 14). This condition is also exhibited in the D. krohni of Borgert (l.c. Taf. VI, fig. 11, 12), which is probably a variety of D. mulleri, and certainly not the original krohni of Herdman (see below, p. 214). In the D. krohni of Neumann ( =sigmoides , n. sp., also not Herdman’s krohni, see p. 214), it still lies alongside the gut, and retains the simple pear-shape, but grows forwards instead of downwards, until it meets the posterior wall of the pharynx, which is bulged forwards as if for its reception (Neumann, 1906, Taf. XIV, figs. 6, 7). In all other Doliolids (with the partial exception of D. valdivice and mirabilis- chuni) it grows outwards as well as forwards on the left side from its median rudiment, and, on meeting the body- wall, extends as a parietal organ, either straight and hori- zontally, as in D. rarum, indicum, intermedium, resistibile, nationalis and denticulatum, or obliquely upwards, as in D. gegenbauri and tritonis. In D. mirabilis ( =chuni ) it begins to grow as a parietal organ, but turns backwards on meeting the branchial septum behind M4, and in the “Terra Nova” specimen, finishes its growth alongside the stomach once more (cf. Neumann, l.c. Taf. XIII, fig. 10 ; Taf. XIV, fig. 3). In D. valdivice it never becomes parietal, but, growing directly forwards on the left side of the stomach, as in Neumann's krohni ( =sigmoides , n. sp.), it turns back on meeting the pharynx behind M5, and makes a half-circuit round the intestinal loop beneath the cloacal floor (Neumann, l.c. XIII, fig. 8 ; XIV, fig. 2). The extension of the testis as a parietal organ involves its penetration between the pharyngeal wall and successive muscle-bands, and there is clear evidence that the muscle-bands, or rather their connexions with the pharyngeal epithelium, con- stitute definite obstacles to its advance, at least in the later stages of development. The contrivance by which these obstacles have been overcome in the case of the higher species (cf. Dolioletta and Doliolum ) is remarkable. In D. denticulatum the testis, instead of enlarging as a whole, sends out a fine thread-like process in advance, which may be called the stolo perforans, or apical thread. This penetrates the spaces between a muscle-ring and the pharyngeal epithelium and may extend to a varying distance in the next intermuscular space, until it reaches another muscle-ring, where it is temporarily held up. In the meantime the posterior section of the thread enlarges to the dimensions of the hinder part of the testis, and the apical thread ultimately succeeds in crossing the next muscle-ring. So the process seems to continue, step by step, from one muscle- ring to the next, until the 2nd muscle-ring is reached or passed, while some succeed in passing even the 1st. There is, doubtless, considerable variation in the details of the process (cf. examples, st. 107, 109, p. 246). The “ Terra Nova ” material included no specimens of phorozooids carrying gonozooid buds sufficiently far advanced to enable me to study the earliest processes in situ, but a fair number of small free gonozooids still possessing peduncles was to be found, and many of these had probably been detached prematurely from their phoro- zooids. In the smallest specimens (1 mm.) the testis consisted simply of a rounded rudiment in front of the ovary in the 6th intermuscular space, or with an anterior 206 “TERRA NOVA” EXPEDITION. filiform extension up to M6. In slightly larger specimens, about 2 mm., the testis extended to M5 or even M4 ; but from 3 mm. upwards almost every stage in extent of the testis from behind M4 up to M4 was exhibited irrespective of size. At a little below this size (say 2-5 mm.) most of the gonozooids of denticulatum seem to become detached from the phorozooids to lead an independent life, but I am unable to give a figure based upon observation of specimens actually borne by their phorozooids. Grobben gives 2-5 mm. as the minimum size of Mediterranean gonozooids, and Neumann has figured a phorozooid carrying a pair of gonozooids nearly ripe for de- tachment (l.c. Taf. VII, fig. 3), from which their size may be estimated as about 2-1 mm. They show a testis with an apical thread which has crossed M6 and nearly reached M5. From the figures given below as to the extent of the testis in gonozooids of different size, it would seem to be possible for some individuals to complete the whole length of the testis before detachment, but such a proceeding must occur rarely (cf. st. 109, p. 52). The testis first completes its growth in the great majority of cases at a size of 3-4 mm., when the individuals are undoubtedly living an independent life. Length of Zooids (mm.) 1 2 3 4 5 6 Total Testis reaching or passing M2 57 322 222 11 1 613 99 99 >> m3 — 17 95 80 6 — 198 9 9 9 9 „ m4 — 22 18 3 — — 43 9 9 9 9 m5 1 16 2 3 — — | 22 99 99 „ m6 1 — — — — — 1 Totals . 2 112 437 308 17 1 877 (Average size : 3-22 mm.). Table I, showing correlation of testis length and size of body in Pacific denticulatum. These figures record for the first time the amount of variation in D. denticulatum in respect of this character, and are of considerable interest. The maximum size of denticulatum gonozooids has been given by Neumann (1913, b), after working out the “ Valdivia,” and “ Gauss ” collections, as 9 mm. Borgert, after the German Plankton Expedition, placed it at 6 mm. for the Atlantic, and earlier naturalists have assigned a maximum of 4-4 mm. (Krohn, as 2'") or 5 mm. (Uljanin) for Mediterranean specimens. The dominant sizes are clearly, therefore, much the same in the S. Pacific as in the Mediterranean, though the average is a little higher, but the extent of local variations cannot be accurately envisaged without proper biometric data. Neumann remarks on swarms of large specimens (8-9 mm.) encountered by the “ Valdivia ” in the Ben- guela stream, a notable fact in view of the relatively low temperature of W. African waters (l.c. 1906, pp. 231-235). The Atlantic specimens collected by the “ Terra Nova ” are insufficient to be DOLIOLIDA— GARSTANG. 207 representative of this immense area, but the data they yield are summarized in the following table : — Length of Zooids (mm.) 2 3 4 Total Testis reaching M2 5 3 8 ,, „ M3 — 24 4 28 ,, ,, M4 6 15 3 24 ,, ,, M5 1 3 — 4 Totals (Average size : 3-05 mm.). 7 47 10 64 Table II, showing correlation of testis length and body-size in Atlantic denticulatum. They are derived from two equatorial Stations 61 and 63 (midway between Fer- nando Noronha and Sierra Leone) and from Station 70 (Azores). The latter station yielded all the 4 mm. specimens, and seven out of the eight specimens in which the testis reached M2. If we may generalise from these various data, it would seem that the species attains its largest size and greatest sexual development in the more temperate oceanic waters, while in equatorial waters (and in the Mediterranean) there is apparently less bodily growth and the testis rarely reaches M2. How far the latter phenomenon is due to earlier maturity it is difficult to say, but a comparison of the two tables throws a little light on the problem. In the S. Pacific, half the 2 mm. individuals, and three-quarters of the 3 mm. individuals showed a testis which reached or passed M2 ; but in the Atlantic none of the 2 mm. and only one-ninth of the 3 mm. individuals attained the same condition. These differences were associated with a slight difference in the average size of the two populations. The facts mentioned show that, apart from actual rate of growth, extension of the testis lags behind that of the body in the Atlantic specimens as compared with the Pacific — which points to the existence of some special retarding influence other than general nutrition. It may be sought in two directions. Although the influence of actual maturation of the testis at these sizes was presumably excluded, the ovary was noticeably active and lobulate, so that precocious maturation of the ovary, under the influence of higher temperature, may possibly have retarded the extension of the testis. In view of the small size of the ovary, this is hardly likely. More probably a rapid rate of differentiation of the pharynx and connective tissues under the same conditions may have imposed additional obstacles to the passage of the testicular stolon. Earlier maturation (as distinct from mere extension) of the testis itself, added to these factors, would then account for eventual arrest of body-growth at a stage when the testis was still short of its potential limits. A conspicuous example of a particular testis checked in this way was furnished 208 “TERRA NOVA” EXPEDITION. by a 4 mm. specimen from Station 70 (Azores). The testis had grown straight forward as far as M4, which it must have found difficult to cross, for it there bent back on itself, and turned abruptly upwards, its apex ultimately pointing towards M4 again, but at a higher level up the left side. Although far from settling this complicated question, the considerations advanced tend to show that the form-differences involved fall well within the range of possible environmental variation, and reveal no constitutional peculiarities between the denticulatum of different regions. Neumann, indeed (1913, a., p. 223), has emphasised the occurrence of two distinct forms of denticulatum, in the smaller of which the testis is club-shaped and extended only to the neighbourhood of the 4th muscle-ring, while in the larger form the testis is cylindrical and extended to M2 or beyond it. Obviously the “ Terra Nova ” material lends no support to this idea of a bi-modal type of varia- tion centred round M4 and M2. It is true that Mediterranean naturalists have familiarised us with a small type of denticulatum (var. ehrenbergii) , having a testis ending near M4 (e.g. Krohn, Grobben, Uljanin), but Keferstein and Ehlers stated explicitly that the testis “ ragt oft bis in den dritten, ja bis in den ersten Zwischen-muskelraum hinein ” (1861, p. 63), and illustrated the latter condition in one of their drawings of a 4 mm. specimen (Taf. XI, fig. 2). The variation is accordingly continuous, and the only distinction of the Mediterranean “ race ” is that it usually fails to achieve the full extension of the testis in its smaller individuals— a failure which is also apparently true of the species in the equatorial Atlantic, though possibly to a different extent. The apparent difference in this respect between D. intermedium of the Atlantic and Mediterranean and D. resistibile of the Antarctic is probably strictly comparable. A similar variability in the length of the testis was recorded by Borgert for his new species nationalis, but without any specification of limits. In Dolioletta gegenbauri variations in the length of the testis appear to be less frequent than in denticidatum. In the few specimens of the “ Terra Nova ” collection the testis usually mounted obliquely up the left side until it reached M2, in two or three specimens only reaching M3, and in one attaining Mx. In the only descriptive account of Mediterranean gonozooids. Uljanin includes as part of their definition “ Hoden . . . bis an den zweiten Muskelreifen reichend ” (1884, p. 134, Taf. 7, fig. 5). In the closely related D. tritonis, Bitter described the position of the anterior end of the testis in Pacific specimens as “ variable,” but “ usually in fully developed state slightly behind the 2nd muscle-band ” (1905, p. 86, fig. 24), and added “ anterior end frequently, though not always, with pronounced hook.” This hooked extremity, just behind M2, is represented in Uljanin’s figure of gegenbauri. In his original description of tritonis (which he at first regarded as a variety of denticulatum), Herdman also emphasised the usual termination of the testis “ close to the 2nd muscle-band,” but added : “No previous investigators, so far as I am aware, either describe or figure the extraordinary variability in form of this anterior end of the testis ” (1883, p. 110, PI. XX, figs. 6-11). In view of my account of the mode of DOLIOLIDA— GARSTANG. 209 extension of the testis in denticulatum, I wish to draw particular attention to these remarks and figures of Herdman, as they furnish in another species some striking examples of the same type of development, and of the obstacles furnished by the anterior muscle-bands to the extension of the testis in the body- wall (text-fig. 2). His description and figures speak for themselves, and I need only add a few comments. In figs. 6 and 8 we see that the testis must have been tem- porarily arrested at M3, but ultimately suc- ceeded in penetrating under it by means of its stolonial thread. In fig. 6 it was finally arrested at M2, but in fig. 8 succeeded in traversing even this, though not without difficulty, as shown by its zigzag path. In fig. 7 the testis reached M2, but its “ stolo perforans ” turned completely round, and after re-traversing M3, was finally held up by M4, where it began to expand. In fig. 10 it traversed M2, was held up by M1; turned back, so as to form the terminal hook referred to by Ritter, while its “ stolo perforans ” continued to burrow back- wards through M2. In fig. 11 the apical thread effected an entrance below M2, but succeeded in emerging only after the most devious wanderings. The whole of these remarkable cases seem to be inexplicable except on the assumption that the connective tissue adhesions between muscle-bands and pharynx were exceedingly close and tenacious by the time the testis had reached the 3rd or 2nd muscle-ring. They at any rate support the view here taken that the high frequency of examples in which the testis terminates at a particular muscle-band, as compared with the number of intermuscular terminations, is due not to any particular virtue in this relation, but to the fact that the muscular connexions impose a greater obstacle to extension than does the looser tissue of an intermuscular space. At first sight the peculiar coilings of the testis in chuni and valdivice look like reactions to similar obstacles at a still earlier period of development, but it is the branchial septum, not a muscle-band, against which the testis is turned back in each case. These species accordingly seem to represent earlier attempts than those of gegenbauri and denticulatum to accommodate a lengthening testis, valdivice being presumably more primitive than chuni in retaining its testicular coils entirely within the intestinal area. This view accords with the more primitive position of the branchial septum in valdivice than in chuni , but another interpretation is possible (p. 219). Fig. 2. — Showing, in Doliolum tritonis, various irregularities in the anterior termination of the testis (modified after Herdman, 1883). the first 4 muscle-bands. H6-H11, Herdman’s figs. 6-11. 210 “TERRA NOVA” EXPEDITION. IV. IDENTIFICATION AND NOMENCLATURE. (a) CRITICAL SURVEY. In general I find myself in close agreement with. Neumann (Das Tierreich, 1913), especi- ally in his identification of Gegenbaur’s “ great nurse ” with Uljanin’s gegenbauri, in opposition to Grobben, Uljanin, and Borgert, who had so strangely associated it with denticulatum ( =ehrenbergii ). In some cases, however, Neumann appears to have attached too great a constancy to certain minor differences, and in one case to have made a serious error of identification, so that I propose to take his list of species in the order in which he presents them, and to make my comments under those heads. For convenience of cross-reference I divide it into four sections, corresponding to the four genera which I propose to use, in lieu of the two sub-genera of Borgert, which are added in brackets as used by Neumann. DOLIOLOIDES, gen. nov. 1. D. [ Doliolina ] rarum, Grob. (1882). (Text-fig. 3.) Neumann omits from his synonyms the third of Gegenbaur's unnamed oozooids (1856, Taf. XVI, figs. 12, 13), which was referred to this species by Uljanin (p. 131) with Neumann’s own approval elsewhere (1906, p. 209). There can be little doubt as to the accuracy of this identification, but its ac- ceptance demands some Br., Brain ; M1-Mg, muscle-bands ; B.s., Branchial septum ; Ov., Ovary ; modification of Neumann’s End., Endostyle ; T, Testis. definition of the “ Amme ” owing to the greater breadth of the muscle-bands in Gegenbaur’s larger specimen (2-2-"' =5*45 mm.) than in Neumann’s (4-3 mm.). In place of “ Muskelreifen schmal” read “ Muscle-bands narrow in young specimens (<4*5 mm.), broadening later, viz. twice as wide as the interspaces at 5*5 mm. Viscera (always ?) intact.” Fig. 3. — Dolioloides, gen. nov., showing the systematic characters of Dolioloides rarum, Grob. (gonozooid). DOLIOLIDA— GARSTANG. 211 DOLIOLINA (s. str.). 2. D. [ Doliolina ] ind/icum. Neum. (1906). (Text fig. 4.) The unique resemblance of the gill-septum of this species to that of Herdman’s D. krohni was overlooked by Neumann, who, with Borgert, misunderstood Herdman’s description (see krohni, below). It differs from that of Herdman's krohni merely in the small number of its gill-slits (5 pairs, as against 25 pairs in krohni), but the fact that well-developed phorozooids also possessed the same limited number indicates, as Neumann urged, that this is in all probability a specific character, as in Dolioloides rarum. The obliquity of the gill-septum is a primitive feature shared with Doliolid oozooids generally. 3. D. [ Doliolina ] intermedium, Neum. (1906). (Text-fig. 4.) To this species may probably be assigned the “ einzelne Individuen von Dol. krohni ” with anteriorly directed testis (text-fig. 4), to which Borgert referred in his Fig. 4. — Doliolina, s. str. Diagram showing the systematic characters of the species intermedium, Neum. (and var. resistibile, Neum.), indicum, krohni, Herd, (non Traustedt, Borgert, Neumann), mulleri (including var. krohni, Borgert), sigmoides, n. sp. (=krohni, Neumann, p.p.). original account of “ Doliolum sp.” (1894, p. 18, footnote 2), though it is not clear on what grounds he distinguished originally between them. The position of the ovary in these variants of “ krohni ” is not stated (see krohni, below). Fowler’s specimens of intermedium (1905, p. 90, as “ Dol. sp. Borgert ”) from the Bay of Biscay are of interest as connecting this species with Neumann’s resistibile iv. 6. 3 212 “TERRA NOVA” EXPEDITION. which follows (q.v.). They combined the long endostyle of intermedium with the gill relations and slender muscles of resistibile. The oozooid (“ Amme ”) of intermedium is said to be still “ unbekannt,” but I am strongly inclined to claim for it the first of Gegenbaur’s unnamed oozooids (1856, p. 300, Taf. XVI, fig. 14), which is characterised by its exceptionally elongated barrel-shape, its slender muscle-bands and wide interspaces, the muscles being in general less than half the width of the interspaces. The brain is large, situated immediately in front of M5, and separated by twice its own length from M4. The dimensions of the body given by Gegenbaur are up to 3"' in length by \"‘ in width, i.e. 6-5 by 2-2 mm. The otocyst measured O04'" (=0-09 mm.). Four of Gegenbaur’s specimens, stated to be the majority, lacked gill and ali- mentary canal, but two retained the ventral part of the gill-septum as well as the whole of the alimentary canal. This is stated to have had the same relations as in troschelii (i.e., in this case, mulleri ) with the significant qualification, “ Nur . . . dass der (Esophagus mit dem Magen hier ein starkeres Knie bildet, als es dort beschrieben ward ” (p. 301). Now one of the differences between the gonozooids of intermedium on the one hand, and mulleri or Borgert’s “ krohni ” on the other (as shown in Borgert's Taf. VI, figs. 11 and 16, for “ krohni ” and intermedium, and Grobben’s Taf. Ill for typical mulleri), is that in the former the stomach and intestinal loop are sharply bent forwards ventrally beneath the oesophagus, so that the whole alimentary canal assumes (from the right side) an S- instead of a U-shape, with a consequential sharpening of the oesophageal angle. Since Doliolina is still at an evolutional level in which the alimentary canal of oozooid and gonozooid are alike (cf. D. rarum and mulleri), the sharp bend of the oesophageal “ knee ” in Gegenbaur’s oozooid may be assumed provisionally to be an intermedium characteristic. From lack of data it is impossible at present to be certain whether this type of curvature is really distinctive of intermedium or a mere coincidence in a range of varia- tion. It is approached very nearly in the small oozooid (0-9 mm.) which Borgert attributed, though with considerable reserve, to his “ krohni ” (l.c. p. 25, Taf. VI, fig. 13). This, however, may equally well be another specimen of intermedium, provided the length of its endostyle (M. 2|-M. 4f) is not prohibitive (see under resistibile below). Apart, however, from the similarity of the gut in the two cases, there is strong reason for expecting the oozooid of intermedium to possess slender muscles, even after the atrophy of the viscera, since this condition has been shown by Neumann to persist throughout life in the closely related resistibile of the Antarctic (1913, p. 18, Taf. I, fig. 5), two old nurses of which were also taken by the “ Terra Nova ” within the Antarctic Circle. In any case it is impossible to accept Grobben’s identification of Gegenbaur’s oozooid as a “ development stage ” of D. mulleri (1882, p. 63) on account of the large size it attains without any marked broadening of the muscles. Even less possible, for the same reason, as well as from the form of the gut, is Uljanin’s identifica- DOLIOLIDA— GARSTANG. 213 tion of it (1884, p. 129), with his ehrenbergii ( =denticulatum ). Gegenbauri and tritonis are similarly excluded. Indeed, the form of the gut alone probably excludes all species of Dolioloides, Dolioletta and Doliolum; so that, among N. Atlantic and Mediterranean forms, there remain only Borgert's “ kroJini ” and intermedium as possible claimants. The close relationship of “krohni” to mulleri renders it practically certain that “ Jcrohni ” (pace Borgert) shares the life-history of the latter. Thus argument by exclusion leads to the same inference as does the positive similarity of the gut, and the analogy of resistibile. Instead, therefore, of “ Amine unbekannt ” read “ Oozooid permanently (?) of elongated barrel-shape, with slender muscle-rings ( M4. I. d. Equatorial Belt. 6° N.-6° S., 20° W.-250 W. May 12-19, 1913. St. 51-59, Fernando Noronha to Equator, and 60 (2° N.).- — None. St. 61 (2° N.). Doliolum denticulatum. 35 GZ (2-3 mm.), 2 PZ (3-4 mm.), 1 PZ (?) (2 mm.). St. 62 (4° 50' N.).— None. St. 63 (6° 10' N.). Dol. denticulatum. 8 GZ (3-4 mm.), 1 PZ (4 mm.). I. e. North Sub-Troipical. May 26-June 3, 1913. St. 64-69 (23° N.-30° N., 33° W.-35° W.).— None. St. 70 (Azores). Dol. gegenbauri. 17 GZ, 2-4 mm. ; 3 PZ, 2-4 mm. Many of the GZ had remnants of a peduncle, even at 4 mm. End. 2|-4? ; G. 3.51.5. Dol. denticulatum. 21 GZ, 3-4 mm. Do. (?) 1 Ooz., type A1; 2-5 mm. (cf. st. 83). Intermediate between Keferstein and Ehlers’ figs. 2 and 4 (l.c. PI. X.). Endostyle short, between M3 and M5 ; Intestine straight, anus under M8 ; Brain behind middle of interspace ; Otocyst just in front of M4, otolith 0-045 mm. ; Muscles exceeding half the interspaces ; M3 to M5 with their interspaces=0T2-0-20-0-14 — 0-20-0-14 mm., i.e. eurymyonic and mesoclinal. DOLIOLIDA— GARSTANG. 243 AREA II.— SOUTH PACIFIC. The Plankton stations of the “ Terra Nova’s ” Winter Cruise in 1911 were based on a centre at the Three Kings Islands, off the northernmost point of New Zealand. The area between these islands and the mainland was sampled more or less continuously during July, August and September, and the investigations were extended once in each month along three lines, one to the N.W. along the line of the submarine ridge which connects the Auckland Peninsula and the Three Kings with Norfolk Island, and two at right angles to the above, i.e. to the S.W. Although these three lines of outlying stations ran only a small fraction of the distance towards the Australian coast 1,200 miles away, from an oceanographic standpoint they were strategic lines, and I have therefore separated them from the general stock of stations concentrated in the passage, 35-40 miles wide ,[ between the Three Kings Islands and Cape Maria van Diemen. The area of this central group (b) may for brevity be termed “ Kings Passage,” and the three special lines defined as follows : — (a) The July Cross-line — a series of 9 stations (st. 71-79) which represents a general sampling of the region across the base of the “ ridge.” It began in the “ Passage,” about 10 miles E.S.E. from Great King (st. 71), rounded the Three Kings to the N.W. (st. 72-76), and from st. 77, about 8 miles W., ran away to the S.W. about 60 miles from Great King (st. 78), this furthest station being close up to the 171st Eastern Meridian (34° 39' S., 171° 6' E.). Station 79 on the return, was 10 miles less distant. (c) “ Along the Ridge,” an August series of 19 stations extended along the line of the submarine ridge from Dominions Bay on the N.E. side of the mainland (st. 94), round North Cape (st. 97), about the Islands (st. 98-111), and then away to the N.W., to stations respectively 45 miles (st. 112), 80 miles (st. 113), and 110 miles (st. 114) from Great King. This, the furthest station north, rather more than halfway between Cape Maria and Norfolk Island, lay at 32° 55' S., 170° 38' E. (d) The September Cross-line was essentially a repetition of the July line, but extended a little further at each end, 11 stations altogether. It began (st. 132-6) in Spirits Bay between North Cape and Cape Maria, ran due north to the 34th Parallel (st. 137), then past the Three Kings (st. 138), and about 20 miles to the S.S.W. (st. 139, 140), then away westwards to stations approximately 50 miles (st. 141), 80 miles (st. 142), and 110 miles W.S.W. from Great King, the last station (143) being at 34° 58' S., 170° 12' E. Two Plankton stations followed series ( e ) at North Cape (st. 145, 146), and may be regarded as a repetition in September of stations 94 and 97 in August. A year later (August-September, 1912), according to Harmer and Lillie's “ List of Stations,” Plankton was collected in the Bay of Islands on the N.E. coast (st. 148), but it apparently contained no Doliolids, unless a tube simply labelled “ Winter Cruise ? — Surface ” represents it. It contained a few Doliolum denticulatum, and some young oozooids of D. mutter i, which, apart from this, were only taken at stations 111 and 113. iv. 6. 7 244 “TERRA NOVA ” EXPEDITION. Station 148 is said in the “ List ” mentioned to have had no “ provisional station- number.” It would be interesting if a connection between the two could be established, as the Bay of Islands, it may be remembered, was intimately associated with Huxley's classical investigations on Doliolum denticulatum, which he found at the entrance of the Bay “ in considerable numbers."' Unless the tube mentioned belongs to st. 148, no Doliolids of any kind were taken by the “ Terra Nova ” on the N.E. coast ; but it should be borne in mind that all the " Terra Nova ” collections in Area II were taken during the southern winter. How far that may have affected the situation, I am unable to say. The stations follow in numerical order, which in this area is also the chronological order. (a) The July Cross-line (st. 71-79). No samples received from any of the 9 stations. One of these (st. 74) was a vertical haul (0-120 metres) with the Apstein net. I cannot, of course, be certain that no Doliolids were taken. ( b ) Kings Passage, (i), July (st. 80-93). St. 80-82. Three vertical hauls (0-100 metres). No results received. St. 83. 8 miles W.N.W. of Cape Maria. July 23, 1911. Dol. gegenbauri. 5 GZ (3-4 mm.), 1 PZ (4 mm.). End. M2 } -M4| ; G. 3.6.5 ; T reached M3 in 1, M2 in 3, and passed M4 in 1. Dol. denticulatum. 30 GZ (2-4 mm.), 5 PZ (2-3 mm.), 2 (?) mutilated. One of the 4 smallest GZ (2 mm. T M3) had a long peduncle, 2 others a basal stump. Do. (?). 1 Ooz., type A2, 1-9 mm. (cf. st. 70). As Als but stenomyonic. Otolith 0-05x0-04 mm. St. 84. 15 miles S.W. by W. from Cape Maria. July 24, 1911. Dol. gegenbauri. 1 GZ (crumpled). Dol. denticulatum. 45 GZ (2-4 mm.), 18 PZ (2-5-9 mm.). St. 85. 24 miles W.N.W. from Cape Maria. July 24, 1911. Dol. denticulatum. 49 GZ (2-4 mm.), 19 PZ (2-4 mm.). The peduncles of the PZ were small, and few carried any buds. St. 86. Close to the Islands. July 25, 1911. Dol. mitlleri (?), 1 small OZ (Bb), 3-4 mm. Crumpled and decayed ; otolith loose and brain absent. M4 and M5 widest and equal. Dol. denticulatum. 20 GZ (2-4 mm.), 6 PZ (3-5 mm.). Peduncles of PZ mostly minute, with few traces of small buds. St. 87-89. No results received. St. 92. 24 miles S. by W. from Great King. July 27, 1911. Doliolina mitlleri, 1 small OZ (Bb), 2-0 mm. Enclosed in a thick viscid test ; brain normal, but otolith absent ; blastophore stumpy, bearing 5-6 small buds. Dolioletta gegenbauri. 1 GZ (<3 mm.). Doliolum denticulatum. 31 GZ (2-4-5 mm.) ; 8 PZ (2-3 mm.). In all the GZ T reached M2 and reached or even surpassed Mx in 21. The 12 oral valves completely closed the mouth. The atrial aperture was also closed to a seeker, but the valves were invisible. The PZ had vestigial peduncles and apparently no buds. DOLIOLIDA — GARSTANG. 245 St. 93. 13 miles S.E. by S. from Great King. July 28, 1911. Dol. gegenbauri. 1 GZ, about 4-5 mm., but crumpled ; 3 PZ. All the PZ are much crumpled, but about 4-5, 5-5 and 7-5 mm. long when extended. The first has a peduncle bearing small buds, the 2nd a minute peduncle without buds, the large one no peduncle at all, but a A -shaped kink in M7 ventrally, directed forwards, representing point of reunion of the previously interrupted muscle- ring (cf. st. 107, under Dol. denticulatum). This specimen (but not the others) lacks gill-slits, while retaining endostyle and gut. Dol. denticulatum, a great swarm, analysed as follows - Body -length (mm.) Total numbers 3 3 4 5 Total Not measured Grand total (1) Gonozooids T to Mo 24 216 93 3 336 33 369 m3 4 41 1 3 94 12 106 m4 2 3 1 — 6 — 6 m5 — 2 1 — 3 — 3 Ovary, no testis — — 1 — 1 — 1 30 262 142 6 440 45 485 (GZ) (2) Phorozooids 1 19 21 8 49 52 101 (PZ) The number of gill-slits was counted in a few gonozooids of different size (cf. st. 107, and Incert. Sed. (1), p. 55), as follows : Length (mm.) Gill-slits of one side Total Dorsal Ventral 2-5 25 15 40 2-6 25 17 42 3-6 25 21 46 3-8 32 18 50 4-8 and 5-2 45 32 77 In all structural relations (apart from number of gill-slits) the smaller specimens (<3 mm.) agreed exactly with the larger, the gill-slits beginning just in front of M2 or abreast of it, the curve occupying the front half of the 5th interspace, and the slits ending ventrally exactly at M3. Endostyle always exactly M2-M4, and intestinal relations equally constant. One of the gonozooids (4-3 mm.) contained an ovum in its vitelline membrane free in the cloaca. It was slightly larger than the largest of the 5 or 6 eggs in the lobulate ovary. (c) Along the Ridge, August 1911 (st. 94, 97 and 114). St. 94, 97. N.E. Coast of Mainland. July 30 and August 3, 1911. No results. St. 98, 99, 104, 105. Vertical hauls 0-80 metres. No results. 7# 246 "TERRA NOVA” EXPEDITION. St. 100-103. 5 miles S.W. from West Island. August 4, 1-6 p.m. No results. St. 106. 5 miles S.W. from West I. August 4, 7-8 p.m. Dol. denticulatum. 2 GZ, 4 mm. ; 2 PZ, 3 and 5 mm. St. 107. Same as St. 106. August 4, 8 p.m.-5.30 a.m. Dol. gegenbauri. 1 GZ, crumpled, probably 5 mm. Do. OZ (Bb), one 7 mm. M3 and M4 are the widest muscles, M3>M4. No trace of otolith. Dol. denticulatum. 40 GZ (2-5 mm.), 13 PZ (3-5 mm.). In excellent condition. The testis extended to M2 in 24 (3-5 mm.), to M3 in 9 (2-4 mm.), to M4 in 3 (2 mm.), and to M5 in 4 (2 mm.) In one of the largest GZ the testis was swollen up to M4 and extended up to M2 as a moniliform thread. In another (also 5 mm.) which had presumably discharged its last egg, the ovary was a mere speck surrounded by a large empty membranous sac, apparently an ovarian blood-sinus in the cloacal wall, as it was also visible in other specimens as an equally large jacket around the normal ovary. In all the 4 mm. GZ (19) the ovary was well developed and more or less lobulate, except one in which both ovary and testis (to M3) were poorly developed. In a typical 5 mm. GZ the gill-slits were counted on one side as 40 dorsal and 31 ventral = 71 (cf. st. 93). The largest PZ (5 mm.) still possessed a slender peduncle, but the ventral limbs of the muscle-band (M7) were no longer interrupted, having reunited at its base, forming a A -shaped kink, directed forwards (cf. st. 93, under Dol. qeqenbauri). Do. OZ (C), one 5-5 mm. The muscular cuirass is incomplete over the brain, where there is a small fenestra. Otocyst 0-10 mm., enclosing OL, 0-05 X 0-04 mm. St. 108. 10 miles S.W. from Great King. August 5, noon-4 p.m. No results. St. 109. As 108, but through the night, 8 p.m.-8 a.m. Dol. denticulatum. 18 GZ (3-5) ; 9 PZ (3-7 mm.). In excellent condition. In a 3 mm. GZ the whole testis, though extended to M2, was a mere filament ; in three 4 mm. GZ the testis was moderately swollen up to M3 and connected with M2 only by a filament. St. 110. 12 miles W.N.W. from Great King. August 6, 9 p.m.-4 a.m. None. St. 111. Close to the Islands. August 7, 10 a.m.-l p.m. (1) Doliolina mulleri. One Ooz., type B, 1-4 mm. ; 18 OZ (Bb), 1 -5-7-0 mm. These old nurses are all in a decayed condition. All are without otoliths except the smallest, in which the otolith (0-04x0-03 mm.) is loose in the pharynx ! Some are without brains. One (3-7 mm.) appeared to have its brain in the 3rd instead of the 4th interspace, and was to have been treated as a meristic abnormality until it was found that the brain had simply broken adrift ! Yet the muscles are measurable, and in all the banded specimens from this station, except one recorded doubtfully under (3), M4 and M5 are the widest muscles, and approximately equal. (2) Dolioletta mirabilis. 1 GZ, 3-6 mm. (-(-peduncle, 2 mm.) ; 157 PZ, 3-6 mm. (+long peduncles). The gonozooid had obviously broken loose very recently from a phorozooid, probably before capture, or there would have been more in the sample. It shows a testis exactly as figured by Neumann (1906, as D. chuni), except that the re- current lobe extends back to the hinder end of the stomach, the left side of which it closely invests. M6 is incomplete ventrally both in GZ and PZ. The PZ are large, flabby, with long stalks supported by muscular bands, with very slender muscle-rings, and a remarkably loose, soft glairy test, which in these specimens is often completely peeled backwards off the body, in an inverted condition, and is either set entirely free, or, more usually, hangs from the peduncle of the phorozooid to which it adheres firmly. These dumbbell-shaped gelatinous objects frequently become entangled by their peduncular “ necks,” and produce the illusion of a fragment of blastophore DOLIOLIDA— GARSTANG. 247 carrying a series of fine buds, but only half of them are zooids ; tbe other half are empty skins. For supplementary remarks on structure, see p. 220. (3) Do. (?) An old nurse (Bb), 3-0 mm., possibly belongs here. M3 and M4 equal and widest ; body very broad and delicate. Otolith large, 0-065x0-05 mm. (cf. st. 126). (4) Dol. gegenbauri (?), 3 Ooz. type C, — 0-9, 1-2 and 1-8 mm. (Cx-Cy. (5) Dol. denticulatum, 49 GZ, 3-5 mm. ; 50 PZ, 1-7 mm. ; 13 (?) ; 7 OZ (C), 4-7 mm. The relative abundance of PZ, and their small size, are unusual and doubtless related to the presence of old nurses. The following is a survey of all the GZ and PZ that were measurable : — Mm. l 2 3 4 5 6 7 Total GZ _ 3 21 18 3 . 45 PZ 7 12 19 9 1 — 1 49 Below 3 mm., the endostyle is not completely in front of M4, as it is in later stages of growth. At 1 mm. the posterior caecum distinctly projects behind it ; at 2 mm. it reaches the hinder edge of M4, but does not overstep it ; at 3 mm. it only reaches the front edge of M4 (cf. remarks on D. nationalis, supra, p. 223). The branchial formula in the smallest phorozooids was the same as in larger ones, viz. 2-5}. 3. The three largest old nurses (6-7 mm.) are beautifully expanded and barrel- like. The sphincters are contracted, with the effect of truncating the extremities, literally like the ends of a barrel. Most of the specimens show small fenestrse in the muscular cuirass at one or more of the following places — behind the brain, in front of the otocyst, and behind the stolon. In one specimen (6-8 mm.) there is a clear membranous streak between M7 and M8 all round. The otolith shows no regular increase in size with growth of the body, and varies from 0-035x0-03 mm., in the smallest (4-0) to 0-04x0-03 mm. in the largest (7-5) with intermediates such as 0-05x0-04 mm. at 4-8 mm., and 0-04x0-035 mm. at 5-4 mm. St. 112. 33° 37' S., 171° 30' E., 45 miles N.W. from Great King. August 8. Noon-4 p.m. Doliolina mulleri. 4 OZ (Bb), 5-7 mm.-7-8 mm. M3, M4 and M5 are nearly equally broad in these old nurses, but M3 is never ■wider than the others, and in two cases is narrower. Averaged, M4 and M5 are the widest and M4 is slightly >M5. No otolith can be detected. Doliolum denticulatum. 3 OZ (C), 5-6-7-5 mm. (Otoliths 0-05 mm.) St. 113. 33° 12' S., 171° 05' E. 80 miles N.W. from Great King. August 9. 9 a.m.-noon. Doliolina mulleri. 1 GZ, 2 PZ, 1-6 mm., and 3 mm. ; 2 tailed larvge, with large caudal vesicles, in elongated cysts. 4 Ooz. type B, 0-8, 1-10, and 1-55 mm. ; 22 OZ (Bb) from 1-9 mm. to 5-6 mm. The Phorozooids are in very poor condition. The smaller has a thick dirty test, the larger none (cast ?). Endostyle M2 £-5 and M3-5 respectively. Gill-slits not countable, but numerous, 20-40 pairs, all behind M5 — i.e. var. krohni, Borg. The old nurses are also in bad condition. Only two possess otoliths, one of 1-9 mm. (OL, 0-04x0-03 mm.), the other of 4-0 mm. (OL, 0-06x0-05 mm.). Below 2 mm. the 4 central muscles (M3-M6) are equally wide ; above that size M4 and M5 predominate. A specimen of 5-4 mm. shows a rare abnormality ; M3 is completely divided into two unequal rings, so that there are 10 muscles altogether. Dolioletta gegenbauri (?), 2 Ooz. type C, 1-5 mm. (C4-C5). Dol. mirabilis (?). A small PZ, 1-0 mm. with slender, much convoluted muscles, a short spiral gut, and endostyle 2-5. 248 “TERRA NOVA” EXPEDITION. Doliolum denticulatum. 17 GZ, 1-4 mm. ; 6 PZ, 2-5 mm. ; 5 (?) ; 2 OZ (C), 3-3 and 5-0 mm. The otolith in the smaller oozooid measures 0-04 X 0-025 mm. In the other it is obscured by wrinkles. St. 114. 32° 33' S., 170° 38' E., 110 miles N.W. from Gt. King. August 10. 10 a.m.-4 p.m. Dol. mulleri. 10 OZ (Bb), 4-8 mm. All in poor condition and without otoliths. In the two smallest (4 mm.), M4-6 in one and M3-6 in the other are equal and widest ; above that size M4 is the widest, or shares that condition with M5. Dol. denticulatum. 118 GZ and 34 PZ, of sizes as below : — Mm. 2 3 4 5 6 Total GZ 11 45 57 4 1 118 PZ — 5 18 10 1 34 Do. 9 OZ (C), 4-5-6-5 mm. The otolith varies in size from 0-04 to 0-05 mm., and in shape from round to oval. (b) Kings Passage, (ii), August (11 stations). St. 116, 117. Vertical hauls (0-150 metres). August 16, 1911. No results. St. 118, 120, 122. Surface hauls, 50-mesh net. August 17-18. No results. St. 125. Off N.E. Coast. 18-mesh net. August 23. No results. St. 126. 7-8 miles, S.E. from Gt. King. 18-mesh net. August 24. 9 a.m.-noon. Dol. mirabilis. 1 PZ, without test. Do. (?). An old nurse (Bb), about 6-3 mm., thin- walled and very broad, but badly crumpled, with large otocyst (0-16 mm.) and otolith (0-06x0-045), possibly belongs here, but differs from that of st. Ill (3) in M4 and M5 being equal and widest, as in Doliolina mulleri. Dol. denticulatum. 13 GZ and 12 PZ, both 3-4 mm. St. 127-131. Off Three Kings Islands. August 25-27. No results. (d) September Cross-line. St. 132-136. Spirits Bay (between the two northern capes) . August 29-September 2. No results. St. 137. N. of the Capes at 34° S. September 5. No results. St. 138. Close to the Islands, September 5. No results. St. 139, 140. 40 miles W. of Cape Maria. September 6. No results. St. 141. 65 miles W. of Cape Maria. September 7. No results. St. 142. 95 miles W. of Cape Maria. September 8. 9 a.m. (7th)-9 a.m. (8th). Dol. mirabilis (?). A tangle of gelatinous debris, apparently the cast skins of this species (cf. st. 111). 33 Broad-banded old nurses, mostly mulleri (4-7 mm.), with M4 Dol. mulleri. and M5 subequal and widest, but all except one without otoliths, Dol. gegenbauri the rest gegenbauri (3-8-9-7 mm.) with M3 and M4 sub-equal and „ widest, mostly with otoliths. Dol. denticulatum. 7 OZ (C), 4-5-8-0 mm. The otolith in the smallest measures 0-03 mm. (round), in the largest 0-05x0-045 mm. DOLIOLIDA— GARSTANG. 249 St. 143. 120 miles W. of Cape Maria. September 9. 4 p.m. (8th)-9 a.m. (9th). Dol. mulleri (?). 1 OZ (Bb), 1*6 mm. Much crumpled, but M4 and M5 slightly wider than M3 and M6. Otohth 0-03 mm. in middle of interspace. Dol. gegenbauri. 1 PZ, 5 mm. Dol. denticulatum. 15 GZ, 2-4 mm. ; 27 PZ, 2-6 mm. IncertcB Sedis. (1) “ Winter Cruise ? — Surface.” (? St. 148, see pp. 243, 244, above.) Doliolina mulleri. 1 Ooz., type B., 1-5 mm. Agrees exactly with Grobben’s fig. 16 (1882) in proportions of muscle-bands and position of organs, except that the gut loop lies further back in the 5th interspace. In this it resembles Borgert's fig. 13 (1894, Taf. YI) of var. krohni. Endostyle from M2 1 almost to M5. Otocyst projecting from front edge of M4 ; otolith oval, 0-025x5-020 mm. Dol. mulleri. 2 OZ (Bb), 2-8 and 4-0 mm. M4 and M5 in the smaller, and M4 in the larger specimen, are the widest muscles. Otoliths not recognisable. Allowing for the slight differences in size, the two agree closely with Grobben’s figs. 17 and 18 (l.c. supra). Dol. denticulatum. 2 small GZ, 1-2 mm. ; 1 PZ, 3 mm. ; 2 (?), 1-2 mm. The smaller GZ (1-05 mm.) shows endostyle 2-4 ; gill-slits beginning dorsally at M2 and ending ventrally close behind M3 ; and testis with very short stolo perforans, not exceeding Mg. No sign of a peduncle. There are 18 pairs of gill-slits, 13 dorsal and 5 ventral (cf. st. 93). In the larger one the testis reaches M4|. (2) From “ Terra Nova ” Plankton samples. Picked out and mounted by Miss Gladys Webb, Zool. Dept., University College, London. Dol. mulleri. 1 OZ (Bb), 3-75 mm. Muscles slightly wider than interspaces, M4 the widest (0-40 mm.). Otolith under front edge of M4, nearly round, 0-038 mm. Brain 0-11 mm. long X 0-13 mm. wide. Agrees closely with Grobben’s fig. 18 (l.c.) which represents an old nurse of similar size, except that the muscles here are slightly narrower and the brain is bigger (Grobben’s scale of magnification, applied to his figure, yields 0-50 mm. for M4, and 0-07 mm. for the brain). AREA III.— SOUTHERN OCEAN. St. 178. 67° 23' S., 177° 59' W. Vertical haul, 0-500 metres. 24-mesh net. December 15, 1910. 9 p.m. Dol. intermedium. Var. resistibile, Neumann (1913). Two old nurses of relatively gigantic size (15 mm. long x 10 mm. across), and with very slender muscle-bands, yielded by this haul, are very clearly referable to Neumann’s resistibile, though for reasons already given (p. 213), I treat this form, not as an independent species, but as a cold-water variety of D. intermedium. All Neumann’s characters for the nurse of resistibile (l.c. p. 22 ; also Das Tierreich, p. 14) recur in these specimens, but the “Terra Nova” specimens show some additional features of importance. The brain lies immediately in front of M5, and is separated by the greater part of the wide interspace from M4. It looks minute in comparison with the huge body, but is actually 0-17 mm. in diameter, and round in shape, i.e. larger than in the largest oozooid of gegenbauri-tritonis which is available for comparison (viz. 0-15 mm. L. X0-16 mm. W. in a gegenbauri nurse of 17 mm.). But neither otocyst nor otohth is present in either of the “ Terra Nova ” specimens. 250 "TERRA NOVA” EXPEDITION. The first of these characters gives an additional link uniting resistibile with intermedium, if I am right in referring the first of Gegenbaur’s unnamed oozooids (1856, Taf. XVI, fig. 14) to this species (see p. 212 above). The second (absence of otolith) suggests that these old nurses were already dead when captured (cf. p. 237). They are of the maximum size recorded by Neumann. A third peculiarity shown by the “ Terra Nova ” specimens is the position and character of the heart and ventral stolon. The heart lies entirely within the 5th intermuscular space. Its anterior end is recognisable just behind M5, and it is prolonged backwards to occupy rather more than three-quarters of the space. The stolon is a slender, elongated cone, the basal half of which underlies the hindmost quarter of the heart, while its narrow tip extends backwards as far as M6, the front edge of which it actually overlaps. This position of the stolon in the hinder part of the 5th interspace is uncommon, and, so far as I am aware, is paralleled only by the case of the oozooid figured by Keferstein and Ehlers in their fig. 7 (l.c. PI. LX), which has never been satisfactorily identified, but which may turn out to be that of Dolioletta mirabilis. Unfortunately, Gegenbaur’s figure, which I identify with inter medium, does not represent the organs of the ventral side (cf. p. 212, above). The two specimens of resistibile differ in two respects. In one specimen, which has been compressed along its major axis like a concertina, or a Chinese lantern, the thin, glairy and tenacious test is very loose, and already detached from both apertures, but adheres firmly to the base of the blastophore, which carries median and lateral buds. In the other, which has been compressed from side to side, the test is completely absent, and only the stump of the blastophore persists without any buds. The oral aperture is surrounded apparently by only 10 small regular, triangular lips, which are very thin and delicate, and seem incapable of closing the orifice as the 12 valves do in other species, unless the slender sphincter contracts much more in life than in death. In both specimens both apertures gape widely. It is difficult to open out the axially compressed one ; but the other which, flattened and in formalin, actually measures 11 mm. in length by 9-5 mm. in width, can easily and naturally be stretched to 15 mm. L. with a width of 10 mm. in the middle, tapering to 4 mm. at each end. Could the specimen be expanded, as well as stretched, a width of 10 mm. in the middle would become a diameter of 6-4 mm. A contraction of the terminal sphincters, so as to close the apertures, would in effect convert this rather broad barrel into a balloon. In conjunction with the extreme slenderness of the muscles, these considerations seem to suggest a placidly drifting life, rather than one of active locomotion. VII. ABBREVIATIONS. GZ = Gonozooid. PZ = Phorozooid. OZ — Oozooid, fully developed (= old nurse lacking viscera). Ooz. = Oozooid after loss of larval tail, but still retaining its viscera. Bb, Nb, and Cc (or C), following OZ — Broad-banded, narrow-banded, and continuously-coated (or cuirassed). Mj-M9 = the 9 muscle-bands of OZ. in GZ and PZ.) M2J (etc.) = a point corresponding to of the interspace following M2 (etc.) E2|-5 (etc.) = Endostyle extending from half the interspace following M2 up to M6 (etc.). OC = Otocyst. OL = Otolith. DOLIOLIDA — GARSTANG. 251 VIII. LITERATURE. Borgert. 1894. Ergebn. Plankton Exped. Humboldt-Stiftung. Von V. Hensen. II. E.a.C. Thalia- cea : Vertheilung der Doliolen. Pp. 1-68 ; pis. v-vi, text-figs, 1-2, 1 chart, 1 diag. Fedele. 1923. (a.) Le Attivita dinamiche ed i rapporti nervosi nella vita dei Dolioli. Pubb. Staz. Zool. Napoli, IV, pp. 129-240 ; pis. i-iii, text-figs. 1-10. 1923. (b.) Identita fra Dolchinia mirabilis, Korotneff e Doliolum chuni, Neumann. Boll. Soc. Nat. Napoli, XXXY, pp. 152-158. Fowler. 1898. Contributions to our knowledge of the Plankton of the Faroe Channel. Proc. Zool. Soc. Bond., pp. 567-584. ( Doliolum , p. 580.) 1905. Biscayan Plankton. IV. The Thaliacea. Trans. Linn. Soc. Lond. Zool., X, pp. 89-101 ; pis. viii-ix. Garstang. 1928. The Morphology of the Tunicata, etc. Quart. J. Micr. Sci., LXXII, pp. 51-187 ; text-figs. 1-13. 1929. (a.) Article “ Tunicata,” Encyc. Britannica, XXII, pp. 549-555 ; 1 pi. ■ — — 1929. (b.) On the Dextricolic condition in Tunicates. Proc. Leeds Phil. Soc. (Sci. Sect.) I, pp. 506-515 ; 3 text-figs. Gegenbaur. 1856. Ueber den Entwicklungscyclus von Doliolum, etc. Z. wiss. Zool. VII, pp. 283-314 ; pis. xiv.-xvi. Grobben. 1882. Doliolum und sein Generationswechsel. Arb. Inst. Univ. Wien. IV, pp. 201-298 ; pis. i-v, text-figs. 1-2. Hastings. 1931. Great Barrier Reef Exped. IV, Tunicata., pp. 69-109; pis. i-iii, text-figs. 1-17. Herdman. 1883. Report on the Tunicata of H.M.S. “ Triton.” Trans. Roy. Soc. Edinb. XXXII., pp. 93-117 ; pis. xvi.-xx. 1888. “ Challenger” Reports, Zool. XXVII, Tunicata. Pp. 1-166 ; pis. i-xi, text-figs. 1-28. Huxley. 1851. Remarks upon Appendicularia and Doliolum, etc. Philos. Trans., CXLI, pp. 595-605 ; xviii-xix. Ihle. 1910. Siboga-Expeditie. Monographic, LVId. Die Thaliaceen, pp. 1-55 ; 1 pi., text-figs. 1-6. Keferstein u. Ehlers. 1861. Zoologische Beitrage, III. Doliolum, pp. 53-71 ; pis. ix-xii. Korotneff. 1891. La Dolchinia mirabilis. Mitt. Zool. Stat. Neapel., X, pp. 187-250; pis. xii-xiii. - — — 1904. Notes sur les Cyclomyaires. l.c. XVI, pp. 480-488 ; pi. xix. Krohn. 1852. Ueber die Gattung Doliolum und ihre Arten. Arch. Naturgesch. XVIII, pp. 53-65 ; pi. ii. Lahille. 1890. Recherches sur les Tuniciers des Cotes de France. (These), pp. 338 ; text illust. Neumann. 1906. Ergebn. Deutschen Tiefsee-Exped. (Valdivia), XII, Doliolum, pp. 93-243 ; pis. xi- xxiv ; 1 chart ; text-figs. 1-20. 1913. (a.) Deutsche Siidpolar Expedition (Gauss), XIV. Die Pyrosomen und Dolioliden, pp. 1-34 ; pis. i-iii, text-figs. 1-4. 1913. (b.) Das Tierreich., XL. Cyclomyaria, pp. ix, 36, 1 ; text-figs. 1-19. Ritter. 1905. The Pelagic Tunicata of the San Diego Region. Univ. Calif. Pub. Zool. II., 51-112; pis. ii-iii. Traustedt. 1893. Ergebn. Plankton Exped. Humboldt-Stiftung. Von V. Hensen, II, E.a.A. Thaliacea. — Systematische Bearbeitung, pp. 1-16 ; pi. i. Uljanin. 1884. Fauna und Flora des Golfes von Neapel. Monog. X, Doliolum., pp. viii, 140 ; pis. i-xii, text-figs. 1-11. PRINTED IN GREAT BRITAIN BY WILLIAM CLOWES AND SONS, LIMITED, LONDON AND BECCLES. “C BRITISH MOSEUiv (NATURAL HISTORY). S * BRITISH AHTARCTB (“TERRA SOYA") EXPEBIM, NATURAL HISTORY REPORT. ZOOLOGY:) VOL. IV. No. 7. Pp. 253-262. ‘ MS PENGUIN EMBRYOS. C. W. PARSONS, -M.A., F.R.S.E. {Lecturer in Zcobgy in the University of Glaigoa). rrf?," WITH TWO PLATES AND 'THREE TEXT FIGURES. B. M. (N.H.) ftK - ’ >','s £ ' LONDON :> JUViTl/Wil - PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM. SOLD BY E. Qcaritch, Lid., 11, Grafton Stbect, New Bonn Street, London, W.l ; Uvlxv Sc Go., Ltd,, 32, Old Bond Stele*, London, W.l; Oxford Umtebsitt Peess, Warwick Square, London, E.0.4; AND A* The British Museum (Natural History), Cromwell ROad, London, 3.W.7 rit Sejpt r-m her, 193 A Price Two Shillings, ■ v J . British flDuseum (IRatural Ibtstors). This is No. (J of 2 5 copies of “ Terra Nova ” Zoology , Vol. IV. , No. 7, Penguin Embryos , printed on Special paper. 253 PENGUIN EMBRYOS. BY C. W. PARSONS, M.A., F.R.S.E. [Lecturer in Zoology in the University of Glasgow.) WITH 3 TEXT-FIGURES AND 2 PLATES. PAGE Introduction ....... Embryos of Aptenodytes forsteri ... Embryos of Pygoscelis adelice .... Concluding Remarks ..... Acknowledgments ...... Bibliography ....... INTRODUCTION. The discovery of the Emperor penguin rookery at Cape Crozier, a collection of their eggs and the first description of the unique breeding habits of this remarkable bird, were amongst the most interesting of the fruits of the National Antarctic Expedition, 1901-1904. The report “ Aves ” (Wilson, 1907) gives a vivid account of the material. It will be noted that out of fifteen eggs enumerated (p. 2) only four contained embryos that were not decomposed : and of these, one embryo only was not fully incubated. It was imperative therefore to revisit the rookery early in the breeding season of the bird, if the hope of obtaining a collection of embryos to show a series of stages in its development was to be realised. Wilson fully understood the scientific importance of acquiring such a collection and also the almost insuperable difficulties in the way of its attainment. Both aspects of the problem are set out in his report (p. 31). With reference to the practical difficulties, he had himself discovered that the eggs are laid in the middle of the antarctic winter not on land, but on sea ice. The site chosen for the rookery is sudxthat it is practically inaccessible. Its approaches are guarded by ice pressure ridges presenting the utmost difficulty in negotiation even in the full light of iv. 7 1 253 254 257 260 261 262 254 “TERRA NOVA” EXPEDITION. summer, and in the winter darkness seemingly impassable. Nevertheless, when the opportunity presented itself — in the course of the “ Terra Nova " Expedition — he did not hesitate to make the attempt. With two companions, Bowers and C'herry-Garrard, he made the five weeks’ winter journey, beginning on June 27th, 1911, which has since been made famous in “ The Worst Journey in the World ” (Cherry-Garrard, 1922, vol. 1, pp. 230-299). The journey was successful to the extent of obtaining three eggs containing embryos that were partly incubated. In due course these passed into the hands of Dr. Assheton, and after his death into those of the late Professor Cossar Ewart. The embryos were designated in accordance with small differences in size, Nos. 1, 2, and 3, No. 1 being the smallest. Dr. Assheton prepared sections of the head and body of embryo No. 1, and also cut and mounted a sagittal series of embryo No. 2. The head of this embryo was embedded in paraffin wax, but it was never cut into sections. Some notes and drawings of all three specimens were also left by Dr. Assheton, and these have been available in the preparation of the present report. The report itself, as regards the first part — on the Emperor penguin embryos — is based upon the sectioned material specified, and on the head of embryo No. 2. 1 . Embryos of Aptenodytes forsteri. The most striking external difference between the three Emperor penguin embryos is remarked on by Cossar Ewart in his appendix to “ The Worst Journey in the World (Cherry-Garrard, 1922). Feather rudiments in the form of papillse are present only in the tail region of the youngest embryo, but in the two older specimens they spread over the body in countless numbers. The difference is seen in a related species, Pygoscelis papua, between embryos which compare with fowl epibryos incubated for approximately ten and thirteen days respectively (Parsons, 1932, pp. 144 and 146). It is fair to con- clude, therefore, that the stages represented do not indicate a considerable time interval between the two extremes, and that the morphological similarity between them, especially between embryos 2 and 3, is very close indeed. Unfortunately the preserva- tion of the material is unsatisfactory. As Dr. Assheton remarks in his notes on embryo No. 2, “ the epidermis has separated from the dermis in most places, preservation none too good, but very much better than that of embryo No. 1.” The serial sections of embryo No. 1 bear out this statement fully. In them the brain is almost entirely absent and the sections, which were cut sagittally, are torn and distorted along the line of the top of the head. Only the cartilage remains in position. That of the otocyst, base of skull, and lower jaw is specially conspicuous. The beak of this embryo is well developed. It is unusually sharp except in some sections where — at the tip — fragments of the protecting cushion of epidermal tissue, which is commonly present in penguin embryos at this stage, still adhere. The function of this cushion under ordinary circum- stances is interesting. It develops in correlation with the long beak, no doubt in order to minimise the risk of premature damage to the embryonic membranes. It may PENGUIN EMBRYOS— PARSONS. 255 therefore be regarded as a parallel development with that of neonychia which have the same function with regard to the developing claws (Agar. 1909). The body sections were cut transversely. Here again, many are torn and the details of the histology cannot be made out, although the gross anatomy may be followed with comparative ease. The series correspond closely with sections similarly cut of a fowl embrvo of approximately eleven days incubation and call therefore for no special comment. The section illustrated (text-fig. 1) is in all essential details like the figure given in “ The Development of the Chick (Lillie, 1908, fig. 190, p. 329), for example, and it demonstrates clearly that the youngest Emperor penguin embryo had reached a stage in development far enough advanced to possess a four-chambered heart and lungs actively budding air sacs. The limbs. — Separate preparations were made of the limbs of embryo No. 1. The right wins: and right leg were stained to show the cartilages and mounted whole, the left limbs were sectioned. The special interest underly- ing these preparations is due to an observation concerning the embryos of Emperor and Adelie penguins that were brought home by the National Antarctic Expedition, 1901-1904 (Pycraft, 1907, p. 19). In making a com- parison between the two species, Pycraft noted the principal features of the adult wing in the embryo Aptenodytes , and showed that they were not yet present in the wing of a slightly younger embryo Pygoscelis. In the present instance an examination of the figures — PI. I, figs 4 and 5 — reveals that adult features are certainly lacking in the distal cartilages of the wing of embryo No. 1. Thus, the pollex has not yet fused with the metacarpal of digit 2, and between the metacarpals on one side and the radiale and cuneiform on the other, the semilunar cartilage of the distal carpal 2 is still perfectly distinct. The shape of the humerus, however, is already defined in cartilage as that typical of the adult penguin. As pointed out by Lowe (1933, fig. 126, p. 515), its termination is triangular and in marked contrast with the rounded form of this part of the bone in the majority of birds. The radius, ulna and cuneiform are also completely spheniscine. With regard to the foot, the cylindrical metatarsals are characteristically free, a feature which is generally agreed to be representative of a primitive condition. Embryos 2 and 3. — The similarity between these two specimens has already been Fig. 1. — Aptenodytes forsteri. Embryo No. 1. Transverse section passing through the heart and lungs : ar aortic root, as air sac, au right auricle, db ductus Botalli, l liver, lu lung, n spinal cord, nt notochord, oe oesophagus, p pulmonary aorta, pc wall of the pericardiac cavity, ri rib, s systemic aorta, vl left ventricle, vr right ventricle. 256 “TERRA NOVA” EXPEDITION. stressed. They do not differ from one another probably by more than a day of incubation, and their morphological differences are superficial and trifling. In size embryo No. 3, as measured by Dr. Assheton, is only 0-5 mm. greater in length than embryo No. 2. There are also minor differences with regard to the distribution of the feather papillae. Embryo No. 2 is without papillae on the breast, abdomen and on the pre-axial surface of the wing, whereas embryo No. 3 is thoroughly covered with them. It is noteworthy that there are no papillae as yet distal to the tarsal joint in either specimen, and it is obvious that the superficial differences between them are so small that it is justifiable to treat the two embryos as one for the purposes of comment. With reference to the feather papillae, the sections — particularly those taken sagittally through the tail region of embryo No. 2 (PI. I. fig. 8) — show them in various stages of development up to the formation of feather filaments. The papillae consist of simple thickenings of the dermis with a thin epidermis over- laying them similar in type to those described in Gentoo penguin embryos (Parsons, 1932, p. 151), and there is no reason to suppose that their subsequent history is in any way different. If this is so the first filaments develop into ficial dissection of the right side : bf bursa Fabricii. cr crop, down feathers which are replaced gb gall bladder, i intestine, k opisthonephros, l liver, in the nestling by a second genera- lu lung, ot otocvst, tr trachea, i/s yolk sac. tion of similar feathers. The drawing of the whole embryo (PI. I, fig. 3) is composite. The head was taken out of the embedding wax and the body reconstructed from the series of sections of which fig. 8, PI. I, represents one approximately in the sagittal plane. Both were carefully drawn and compared with photographic prints made from some old negatives of the embryo. In comparison with other penguin embryos at the same stage of development — an Adelie embryo for example — the most striking feature of the Emperor embryo is not its size, but the high dome of the head. A dissection of the head was undertaken to obtain an accurate knowledge of the contour of the brain, therefore, although it was known that the brain could not be anything but exceptionally brittle ; and, in order to harden the brain as much as possible, the whole head was soaked in formalin lor some days before the dissection was commenced. The resulting figures (PI. I, figs. 1 and 2) show that there are no exceptional features : the relative proportions of the parts of the brain are the same as in Adelie embryos of the corresponding stage. Fig. 2. — Pygoscelis adelice. Embryo with feather papillrc, equivalent to a fowl of about thirteen days’ incubation. Super- PENGUIN EMBRYOS— PARSONS. 257 TL.e space between the two optic lobes is, however, very wide, and is filled with a mass of loose tissue surrounding the pineal body. This accumulation of tissue provides the extensive support for the dome-like surface and is largely responsible for the character- istic shape of the head. With regard to the thoracic and abdominal viscera, the sagittal section of embryo 2 (PI. I, fig. 8) marks out the relative positions of the organs, and a search through the whole series shows that in this respect also the Emperor penguin embryos are not different in any important particular from embryos of other species of penguin at the same stage. Text-fig. 2 illustrates a dissection of an Adelie embryo at about the same stage of development as the Emperor embryos Nos. 2 and 3. Besides showing the general dis- position of the organs, this figure illustrates a point which is very often overlooked. It plainly shows the relatively great length of the neck. In their customary standing position adult penguins conceal its presence, but when occupied in preening themselves they may be seen bending the head downwards towards one of the feet and extending the neck to do so. In this way they are able to perform the feat of scratching the top of the head commented upon in his observations of the habits of captive King penguins {Aptenodytes patagonica ) by C'ossar Ewart (Cossar Ewart and Mackenzie, 1917, p. 124). 2. Embryos of Pygoscelis adelice. The following observations are based upon thirteen Adelie penguin embryos collected by the expedition and representative of various stages in the development. The youngest specimens corresjmnd roughly to fowl embryos of five and a half days’ incuba- tion, and the oldest to a fowl of approximately seventeen days’ incubation. The general resemblance between the former and the fowl at the same stage is, of course, close : thus, the eyes are prominent and are pigmented, there are no eyelids, the choroid fissure is open and the predominating feature of the head is the enlarged mid-brain. They already show some features, however, that are characteristic of penguin embryos. The fore limb is directed at right angles to the hind limb, the hand is narrow and the limbs are approximately of equal length. The tail also is narrow and is well marked off from the trunk. While these specific features are important, they do not obscure the fundamental likeness between the early stages in the development of the penguin on the one hand and of the fowl on the other. For this reason in describing the penguin embryos it is best to apply as far as possible the established scheme of stages in the fowl as figured in the “ Normentafel ” (Keibel and Abraham, 1900). 1. Stage 29 N.T. (i.e., Normentafel). Number of specimens 2 : equivalent to a fowl embryo about five and a half days’ incubation. Both these embryos were cut into sections — one series sagittal and the other transverse — head and body separately. Dr. Assheton’s notes on these preparations contain the following comments : — “Head. — No exceptional features except the mandible, relative length and bold curvature of which are very characteristic of penguins. Neck. — Without visceral clefts except 1* 258 “TERRA NOVA” EXPEDITION. the hyomandibular, part of which remains as the external auditory meatus. Body. — Very bent, head and tail almost in contact. Heart.— Not yet fully formed, all the chambers distinct but the auricles still in communication with one another. Blood leaving the heart- passes into one of two completely separate channels divided off the truncus arteriosus. The larger ventral vessel passes forwards, bifurcates round the alimentary canal and reunites to form the dorsal aorta. The smaller dorsal vessel is continued forwards as the carotid root. Lungs. — Connected with bronchi, posterior air sacs developing. Liver. — Well formed. Kidney. — (1) Pronephros — remains, con- sisting of a pair of glomeruli without funnels, persist at the extreme anterior end of the coelome. (2) Mesonephros — well developed, no unusual features. (3) Metanephros— indicated by a condensation of cells along the inner border of a pair of diverticula of the mesonephric duct. These diverticula are the beginnings of the metanephric ducts. Gonads. — A pair of gonads recognisable. Miillerian ducts project far back into the body cavity and consist of plates of thickened peritoneum.” The above comments as they stand give an accurate description of the salient features of the two embryos in question. In some particulars, however, the nomenclature used differs from that generally adopted in this report. The mesonephros, for example, does not become functionally distinct from the metanephros until much later in the embryonic period, and the term opisthonephros is therefore used to describe the developing excretory organ in accordance with the definition laid down by Graham Kerr (1919, p. 221), and the mesonephric duct is equivalent to the Wolffian duct. It is of interest to record further that the gonads are in the form of genital folds, as the embryos are old enough to have passed the stage at which the genital rudiments first appear as ridges. 2. Stage 31 N.T. Number of specimens 1, equivalent to a fowl embryo of about eight days’ incubation. This embryo has the general characteristics of an eight-day fowl, but the short beak, nostrils and well developed tongue are features in which it resembles a later stage. A drawing was made of the specimen straightened out to show the dorsal surface (text-fig. 3 a), and it was then dissected and figured in side view (PI. II, fig. 1). Its anatomy is sufficiently displayed by this means and comment is unnecessary. 3. Stage 32 N.T. Number of specimens 1, possibly a day older than stage 31. The embryo was decapitated and the head cut into transverse sections. It shows growth of the cerebral hemispheres with the result that the mid-brain no longer' projects prominently above the general surface. The eyes are slightly advanced and eyelids begin to make their appearance. The beak is 3-5 mm. long, the limbs both measure 7 mm. over-all. The uropygial gland is becoming conspicuous. In all other features the anatomy resembles that of the preceding stage so closely as to be practically identical with it. 4. Stage 34 N.T. Number of specimens 2, equivalent to a fowl embryo of about ten days’ incubation. PENGUIN EMBRYOS— PARSONS. 259' The most noteworthy feature of these embryos is the appearance of the first feather rudiments as papillae. In one specimen the papillae are not very numerous and are well defined only in the tail region. In the other specimen they are much more widely distributed, being found on the surface of the head and neck and on the back and upper parts of the legs. In neither case are there any papillae on the wings. These appear at a very slightly later stage, as described by Anthony and Gain (1913) in an investigation of the pterylosis of penguin embryos. The two specimens differ very slightly in respect of the length of beak, the smaller specimen has a beak measuring 8 mm., and the larger one measuring 9 mm. in length. 5. The seven remaining embryos in the collection are too far advanced in develop- ment to compare with the Normentafel. With only one exception they are all fledged in feather filaments developed from the feather papillae, and they are embryos of con- siderable size. The exception is provided by the embryo figured in text-fig. 2, in which there is only one row of strongly developed filaments fringing the base of the tail. This specimen is comparable with a fowl embryo of approximately thirteen days' incubation and it bridges an otherwise serious gap between the stage 34 embryos and the older ones. The latter are equivalent in development to fowl embryos of between fifteen and seventeen days' incubation. The figure of the oldest (PI. II, fig. 2) illustrates the anatomy of all and, in addition, it shows two features of special interest. In the first place it is noteworthy that many of the feather filaments are so long that they curl at the tips. It seems probable that this may lead to weakening of the filament sheaths in prejDaration for their future disruption and release of the complete barbs of the down feather. Secondly, the skin of the feet is marked with papillae resembling feather papillae. It is known, however, that they are in fact the precursors of the horny “ scales " of the feet, and it has been shown elsewhere (Cossar Ewart, 1922) that such papillae never do develop into feathers. This is remarkable because the close morpho- logical relationship between feathers and scales seems to be perfectly well established on other evidence. Feather papillae and scale papillae must still be regarded, then, as being distinct from one another from the beginning, and corroboration of this opinion is provided by the sequence of the events in development. In contrast with the feather papillae, the scale papillae are late in making their first appearance. This would not be so if the feathers developed in the course of evolution from scales of the same kind as those which are conspicuous on the legs of adult penguins, and the explanation of the anomaly probably is that the horny scales protecting the feet of penguins and of other birds are epidermal structures of a secondary nature that are fundamentally different from reptilian scales which are the true ancestors of feathers. 260 “TERRA NOVA” EXPEDITION. CONCLUDING REMARKS. Taking a broad view of the facts as they have been adduced, both from a con- sideration of the unique group of three Emperor penguin embryos and of the series of Adelie embryos, it must be stated that neither has greatly added to our knowledge of penguin embryology. At the time when the collections were made the material brought back by more recent expeditions was not, of course, at hand, and the embryos on which the work of the present report is based had the advantage of priority. This advantage was lost owing to the death of two investigators into whose hands the material had been given, and in the meantime other collections were made which enabled the foundations of our knowledge of the subject to be laid. In the light of what was already known of the development of two species of Pygoscelis ( Pygoscelis papua and Pygoscelis antarctica) , it was most interesting to have the opportunity of examining embryos of the third species, Pygoscelis adelice. As might have been expected, however, there are w Fig. 3. — a, Dorsal view of an Adelie penguin embryo, Stage 31 — comparable with a fowl of approximately eight days’ incubation. The embryo drawn straightened out with a view particularly to showing the proportions of the tail ; b, Dorsal view of the tail region of an Adelie penguin embryo showing the first feather filaments, Stage 34 — comparable with a fowl of approximately ten days’ incubation ; c, Dorsal view of the tail of an Adelie penguin embryo fledged in feather filaments — comparable with a fowl of approximately seventeen days' incubation ; d, Dorsal view of the tail region of a fowl- fifteen days’ incubation. Note. — In c and d filaments, other than those fringing the tail, have been pulled out, leaving open sockets on the surface of the skin ; in b the mid-dorsal pit = uropygial gland ; in c and d mid-dorsal prominence with feather filaments = uropygial gland. no considerable differences between the embryos of the three species at corresponding stages in development. The Adelie embryos were distinctive because they have longer beaks in proportion than the other species, and another point of difference is in the earlier development of feather filaments on the head. The history of the develop- ment of Adelie embryos sheds no new light upon the problem of penguin ancestry, however, and unfortunately nothing decisive could be expected from the collection of three Emperor embryos so close to one another in development as those that Dr. Wilson PENGUIN EMBRYOS— PARSONS . 261 and his comrades obtained at such cost. A small point regarding this question of ancestry of penguins may not unsuitably be made here without raising all the arguments of previous papers on a controversial subject. The tail of the developing fowl is always a very conspicuous ob j ect in early embryos. It is conspicuous, too, in the embryo penguin, and if a comparison is made between early embryos of the two species there does not seem to be much difference between them as regards slenderness. In later embryos it is otherwise. Text-fig. 3 is designed to show that the fowl evolves its tail very differently from the penguin. In the latter it seems that the more primitive slender condition is retained in development longer than in the fowl, and this persistence of a primitive character is suggestive. ACKNOWLEDGMENTS. I wish to acknowledge my indebtedness to Professor Graham Kerr and Dr. Percy R. Lowe for assistance and advice in the preparation of this report, and to the authorities of the British Museum for the loan of the material. 1 also wish to record my deep sense of obligation to Miss C. Brown Kelly, by whom all the illustrations were prepared. 262 “TERRA NOVA” EXPEDITION. BIBLIOGRAPHY. Agar, W. E. 1909. On an embryonic appendage of the claws of the Amniota, probably of an adaptative nature. Anat. Anz., XXXV, pp. 373-380 ; text-figs. 1-7. Anthony, R., and Gain, L. 1912. Sur le developpement du squelette de l’aile chez le Pingouin. C.R. Acad. Sci. Paris, civ., pp. 1264r-1266 ; text-fig. Anthony, R., and Gain, L. 1913. Sur le developpement de la Pterylose chez les Pingouins. C.R. Acad. Sci. Paris, clvii., pp. 1018-1019. Cossar Ewart, J., and Mackenzie, Dorothy. 1917. The moulting of the King Penguin ( Aptenodytes patagonica). Trans. Roy. Soc. Edinb., LII, pp. 115-132 ; 2 pis. Cherry-Garrard, Apsley. 1922. The Worst Journey in the World. 2 vols. London, illust. (col.). Keibel, F., und Abraham, Karl. 1900. Normentafel zur Entwicklungsgeschichte des Huhnes ( Gallus domesticus). Jena. pp. 132 ; 3 pis. Kerr, J. Graham. 1919. Text-book of Embryology, II. Vertebrata. London, pp. xii, 591 ; text illust. Lillie, Frank R. 1908. The development of the chick. New York. pp. xi, 472 ; 1 pi. ; text illust. Lowe, P. R. 1933. On the Primitive Characters of the Penguins, and their Bearing on the Phylogeny of Birds. Proc. Zool. Soc. Lond., pp. 483-538, pis. 1-6 ; text-figs. 1-16. Parsons, C. W. 1932. Report on penguin embryos collected during the “ Discovery ” investigations. “Discovery” Reports, VI, Cambridge, pp. 139-164; pis. 1-6; text-figs. 1-9. Pycraft, W. P. 1907. On some points in the anatomy of the Emperor and Adelie penguins. National Antarctic Expedition, 1901-04, Zoology II, Vertebrata No. 3, pp. 36 ; 1 pi. ; text-figs. 1-8. Wilson, Edward A. 1907. Aves. National Antarctic Expedition, 1901-4, Zoology II, Vertebrata No. 2, pp. 121 ; 13 pis. ; text-figs. 1-46. PRINTED IN GREAT BRITAIN BY WILLIAM CLOWES AND SONS, LIMITED, LONDON AND BECCLES. PLATE I. Scale : All the figures = X 1 A, except fig. 8 = X 2. ■Aptenodytes forsteri. Embryo No. 2. 99 99 99 99 99 9 9 9 9 99 Embryo No. 1. 99 99 99 99 „ „ Embryo No. 3. 99 99 99 99 Embryo No. 2. Dissection of brain, dorsal view : pi pineal body. Dissection of head, side view : is interorbital septum, q quadrate. Entire embryo (equivalent to a fowl of approximately thirteen days' incubation). Right wing preparation : cu cuneiform, me III, meta- carpal III, rd radiale, I pollex. Right foot viewed from the inner side, preparation : fb fibulare, lia hallux, mt IV, metatarsal IV. The tail (drawing after Assheton). Left foot, view from the outer side (drawing after Assheton) . Sagittal section : bf bursa Fabricii, cr crop, gb gall bladder, i intestine, k opisthonephros, l liver, lu lung, p pulmonary aorta, pc wall of pericardiac cavity, s systemic aorta, se sternum, st stomach, tr trachea, ug opening of uropygial (or preen) gland, vom omphalomesenteric vein, ys yolk sac. Brit. Antarctic (Terra Nova) Exped Zoology Vol. IV. No. 7. PLATE II. Fig. 1. — Pygoscelis adelicc. Embryo N.T. 32, equivalent to a fowl of approximately eight days' incubation. Dissection of the right side, liver and right lung removed. lip hypophysis, if infundibulum, k opisthonephros, m Mullerian duct, r rectum, ra right allantoic artery, st stomach, ty thyroid, v ventricle of heart, y yolk stalk. II — X. Cranial nerves. „ 2. — Pygoscelis adelice. Embryo fledged, equivalent to a fowl of approximately seventeen days' in- cubation. Dissection of the right side, right lobe of the liver, right lung and right kidney removed. ar aortic root, aom omphalomesenteric artery, au right auricle, bf bursa Fabricii, ca carotid artery, ce cerebellum, cla coeliac artery, era crural artery, db ductus Botalli, gd genital duct, go gonad, hm hemisphere, la left allantoic artery, l liver, me medulla oblongata, o optic lobe, pa pulmonary artery, pi pineal body, ra right allantoic artery, sa subclavian artery, sp spleen, u ureter, vom omphalomesenteric vein. Brit. Antarctic (Terra Nova) Exped. 1910. Zoology Vol. IV. No. 7. BRITISH MUSEUM (NATURAL HISTORY). BRITISH ANTARCTIC (“TERRA NOVA") EXPEDITION, 1910. NATURAL HISTORY REPORT. . ■ • f I ZOOLOGY. VOL. IV, No. 8. Pp. 263-282. REPORT ON THE TUNICATA. PART II.— COPELATA. ft? k * j is| mp WALTER CARSTANG. MA„ D.Sc. omm* trorasoii in ik« wiumm or low. AND ELIZABETH GEORGESON, B.Sc.. Ph.D. WITH FIVE TEXT FIGURES. B. M. (N.H.) LONDON : PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM SOLD BY B. Quabjtcb. Ltd.. 11 Grafton Street, New Bond Street, London, W.l ; Dtxlau & Co., Ltd., 2 Stafford Street, London, W.l; Oxford University Paras, Warwick Sqvasf,, London, E.C.4; and AT The British Museum (Natural History), Cromwell Road, W: 11 ratjl 1935 1 -> *• h [All yvjkte reserved.] Price One Shilling and Sixpern [Jsrutd 25th A/ay, 1933} m m British flOuseum (IRatural Ibiston?). This is No. of 25 copies of “ Terra Nova ” Zoology , Vol. IV No. 8, Report on the Tunicata. Part II. — Cope lata, printed on Special paper. REPORT ON THE TUNICATA. PART II— COPELATA. BY WALTER GARSTANG, M.A., D.Sc. (Emeritus Professor in the University of Leeds), AND ELIZABETH GEORGESON, B.Sc., Ph.D. WITH 5 TEXT-FIGURES. Introduction and Summary Collecting Stations . Distribution Notes on the Species Literature PAGE 263 265 266 267 282 INTRODUCTION AND SUMMARY. This report on the Copelata supplements my report on the Doliolids in a very special manner, and bears more closely on the biology of Antarctic waters. While the Doliolids are essentially warm-water creatures, incapable, it would seem, of prolonged existence in the neighbourhood of ice, their relatives the Appendicularians — their direct descen- dants, as I have tried to show elsewhere — contribute an important and highly specialised element to the pelagic fauna of both poles (Lohmann, 1905). The German South Polar Expedition of 1901-3 made a thorough investigation of the Appendicularians in the Indo-African sector of the Antarctic, and the recent publication of their report (Lohmann and Biickmann, 1926) has enabled comparisons to be made with the results of the British Expedition in the New Zealand sector ten years later. These results are of considerable interest, and help materially to round off the story as well as to define further problems, iv. 8 — ii. 1 264 “TERRA NOVA” EXPEDITION The collections of Copelata by the “ Terra Nova ' ’ include some 2,000 specimens from 34 stations. The identification of these small creatures (many of them unavoidably mutilated in various ways) requires the making of innumerable dissections and prepara- tions, and the elucidation of their life-histories involves careful measurements on an extensive scale. I have to thank one of my former students, Miss Elizabeth Georgeson, Ph.I)., for nearly two years’ invaluable help in this task. She worked over the whole of the material and identified the greater number of the species present. J have since re-examined the whole material, and the present report has been prepared independently, but Dr. Georgeson’s name should be associated in the authorship. W. Garstang. SUMMARY OF RESULTS. 1. The Appendicularians collected by the “ Terra Nova ” are all Oikopleurids, and were derived from three separate groups of stations : (1) Atlantic, (2) South Pacific, north-west of New Zealand, and (3) the Antarctic (161° E.-166° W., 64°-72 S.). Though stations were also distributed between 36° and 64° in the New Zealand sector, no specimens were received from that area. 2. Of nearly 2,000 specimens from 34 stations not less than 1,900 belong to two species : Oikopleura longicauda, Vogt, which dominates the collections in the warm waters of the Atlantic and South Pacific, and was found at every station represented, and 0. valdivice, Lohmann, which similarly predominates in the Antarctic collections, and was found at every station in that area. 3. For 0. longicauda it is shown that the maximum and average size attained in the South Pacific are higher than in the tropical Atlantic, and that the difference is associated with an earlier ripening of the ovary in the latter area. 4. For 0. valdivice it is shown that a very small proportion attain maturity of the ovary before the final stage of growth (3-4 mm. body-length), while the records suggest a definite short breeding season late in the Antarctic summer. This is correlated with the evidence of seasonal variations in nmnber published by the German South Polar Expedition. The sequence of stages in development of the gonads is given, and certain peculiarities of the house-rudiments are described. 5. The other species found in the collections are Oiko'pleura cophocerca. rufescens, and dioica, and a few individuals of the less familiar genera Stegosoma, Folia, and Pelagopleura, the first of these from Atlantic and Pacific stations, the two last from the Antarctic. 6. A new species is proposed for the specimens of Stegosoma and similar specimens already described from Japan, which differ from typical representatives of S. magnum from the Atlantic and Indian Oceans (S. conogaster, n. sp.). 7. The single specimen of Folia is much larger than any previously described, and in consequence of structural peculiarities is assigned to a new species (F. gigas, n. sp.). COPELATA — G A R S T AX G AND GEORGESON. 265 8. The three specimens of Pelagopleura are referred to LohmamTs species P. magna, and yield additional information on various points of structure. The presence of food- remains in each individual corroborates LohmamTs view that the species is a normal inhabitant of Antarctic deep water, to which it appears to be confined. 9. A mutilated specimen which combines large oral glands with two continuous rows of amphichordal cells is described but cannot be referred to a known genus. 10. Attention is drawn to the failure of the “ Terra Nova ” in the New Zealand sector to obtain specimens of Oikopleura gauss ica and Pelagopleura australis, which were taken repeatedly by the German Expeditions in the Indo- African sector of the Antarctic. COLLECTING STATIONS. The stations at which Appendicularians were collected fall into three groups, represented on Plates I. II and III of Harmer and Lillie's “ List of Collecting Stations ” (Brit. Antarctic £' Terra Nova " Exped.. Zool. vol. II, No. 1, 1914), as follows : I. ATLANTIC. 1. Equatorial (Plate I). May, 1913. Stat. 52. — 5° S. 27° 15' W. Depth 2 metres. May 12, 1913. „ 57. — 4° 30' S. 27° 16' W. Surface. May 13. 2. S. Atlantic (Plate III). April 1913. Stat. 310. — 36° 57' S. 51° 21' W. Surface. April 21. II. SOUTH PACIFIC, NEW ZEALAND WATERS (Plate II). July 17 to August 24, 1911. Stat. 76. — 8 miles W. from Summit, Gt. King. Surface. July 7. ,, 85. — 24 miles W.N.W. from C. Maria van Diemen. 2 metres. July 18. ,, 88. — 10 miles S. from Summit, Gt. King. 1 metre. July 22. ,, 102. — 5 miles S.W. from West I., Three Kings. Surface. Aug. 4. ., 107. — 5 miles S.W. from West I., Three Kings. Surface. Aug. 4. ,, 111. — Off Three Kings Is. Surface. Aug. 7. ,, 114.-32° 55' S., 170° 38' E. Surface. Aug. 10. „ 115.-34° 32' S., 172° 20' E. Surface. Aug. 16. „ 126.— 34° 13' S., 172° 15' E. Surface. Aug. 24. III. SOUTHERN OCEAN, NEW ZEALAND SECTOR, S. of 36° S. (Plate III). No specimens were received from any of the open sea stations north of 64° S. 1. Dec. 1910. [Working a passage south through exceptionally thick pack-ice: “terribly heavy floes ” (Scott’s Journal).] Stat. 178.-68° 23' S., 177° 59' W. 0-500 metres. 24-mesh. Dec. 15. ,, 180. — 68° 26' S., 179° 08' W. 100 metres. 24-mesh. Dec. 22. 2. March 1911 (the only station E. of 180°). Stat. 195. — 65° 14' S., 161° 24' E. 2 metres. Mar. 6. 3. Dec. 1912 and Jan. 1913. Stat. 264.-64° 33' S., 166° 30' W. 20 metres. Dec. 26. „ 267.-66° 30' S., 166° 8' W. Surface. Dec. 27. ,, 269. — 68° 37' S., 166° 14' W. Surface. Young fish trawl. Dec. 28. la 266 “TERRA NOVA ” EXPEDITION. 3. Dec. 1912 and Jan. 1913 — continued. Stat. 270.-69° 51' S., 166° 17' W. „ 271.-71° 23' S., 166° 3' W. „ 272.-71° 35' S., 166° 01' W. „ 274.-71° 29' S., 166° 0' W. „ 275.-71° 29' S., 166° 0' W. „ 276.-71° 41' S., 166° 47' W. „ 281.-71° 41' S., 166° 47' W. „ 282.-71° 41' S., 166° 47' W. „ 283.-71° 39' S., 166° 47' W. „ 284.-71° 49' S., 167° 32' W. „ 285.-71° 49' S., 167° 32' W. „ 286.-71° 44' S., 167° 57' W. „ 287.-71° 44' S., 167° 57' W. 0-600 metres. 24-mesh. Dec. 29, Surface. Dec. 31, 1912. 80 metres. Jan. 1, 1913. 80 metres. Jan. 3. 160 metres. Jan. 3. 0-1750 metres. Jan. 5. 80 metres. Jan. 6. 0-1000 metres. Jan. 7. 80 metres. Jan. 7. 80 metres. Jan. 8. 0-600 metres. Jan. 8. 10 metres. Jan. 9. 80 metres. Jan. 9. 60 metres. Jan. 11. 24 metres. Jan. 12. 60 metres. Jan. 12. „ 288.-71° 59' S., 168° 43' W. „ 289.-72° S., 168° 17' W. „ 290.-72° S., 168° 17' W. DISTRIBUTION. OIKOPLEURID/E 1. Oikoplewa longicauda, Vogt (= O. spissa, Fol). 535 specimens (66 tailless). In every sample forwarded from the Atlantic and South Pacific areas, viz. : Atlantic : st. 310 (37° S.), 52 and 57 (5° S.), all in April and May 1913. South Pacific (32°-36° S.) : st. 76, 85, 88, 102, 107, 111 , 1 14, 1 15, 126, all in July and August 1911. 2. O. valdivicc, Lohmann. 1,172 specimens (547 tailless). In every sample forwarded from the Southern Ocean, south of 36° S. (except st. 310 above). The stations are all in the New Zealand sector between 161° E. and 166° W., and between 64° and 72° S., viz. : Dec. 1910.— st. 178 and 180. March 1911. — st. 195 (the only station E. of 180°). Dec. 26, 1912.— Jan. 12, 1913.— st. 264, 267, 269, 270, 271, 272, 274, 275, 276, 281, 282, 283, 284, 285, 286, 287, 288, 289, 290. 3. O. cophocerca, Gegenbaur (non Fol). Several specimens at st. 52 (Atlantic). 4. O. rufescens, Fol. Several at st. 52 (Atlantic), one at st. Ill (South Pacific), and several of a distinct variety at one of the Antarctic stations the original station number of which is illegible. The tube contained many O. valdivice, and was of the “ E ” series (i.e., 264-290). 5. O. dioica, Fol. Common at st. 52 (Atlantic) ; two specimens only at st. 107 (South Pacific). 6. Folia gig as, n. sp. One specimen at st. 274 at 80 metres (Antarctic). 7. Stegosoma conogaster, n. sp. Two solitary specimens at st. 310 (South Atlantic) and 76 (South Pacific), both at the surface. 8. Pelagopleura magna, Lohmann, 1926. Two at st. 275 (160 metres), one at st. 276 (0-1750 metres), both Antarctic, 71° S. 9. Gen. et sp. indet. One mutilated specimen with the preceding at st. 275 (see notes, p. 281). COPELATA— GARSTANG AND GEORGESOX. 267 NOTES ON THE SPECIES. 1. Oikopleura longicauda (Vogt, 1854). — This small species is usually at once detected in a mixed collection by its characteristic dorsal veil, which suggested Langerhans’ expressive term ' velij era (1880). Its dominance in the “Terra Nova’s” Atlantic and South Pacific samples, and its absence from the Antarctic stations, is but a further confirmation of its world-wide value as a warm-water indicator. Lohmann writes “ Die Deutsche Sud-polar Expedition fand sie in warmen Wasser des Atlantischen und Indischen Oceans uberall, sie fehlte aber vollstandig im Antarctischen Meere " (1926, p. 146). It was “ weitaus die hdufigste Art ” in the “ Siboga’s ” collections from the Indian Archipelago (Ihle. 1908, p. 112). It is “ the most common species ” off the Pacific coast of Japan, especially “ in summer when the south wind prevails and the Black Current sweeps nearer the land than in other seasons ” (Aida, 1907, p. 24). The varying state of the oikothelium ( = oikoplastic epithelium) of Appendicularians renders these little creatures difficult to measure with precision, but the following table sums up the range of size, as indicated by the body-length, in the measurable specimens taken by the “ Terra Nova ” (Miss Georgeson's measurements). Length (mm.) Atlantic Stations (April and May) South Pacific (July-August) 0-3 1 1 04 2 IS 0-5 11 35 0-6 17 33 0-7 7 46 0-8 i 19 0-9 — 11 1-0 — 9 Totals . 39 172 Average size . 0-55 mm. 0-65 mm. So far as the figures go, they indicate that the species attains a larger maximum and average size in the South Pacific than in the tropical Atlantic. Although the Atlantic numbers are very small, each of the three stations here combined yields the same modal size (0-6 mm.), so that the result is probably not without significance. It is associated with an earlier maturation of the ovary, the species, as usual in Otkopleura, being protandric. In the following table the size at which the ovary was recognisable is tabulated for the Atlantic stations and for a combination of the two most productive stations in the South Pacific, viz., stations 107 and 115. 268 ‘ TERRA NOVA ” EXPEDITION. Atlantic (April-May) South Pacific (July-Aug.) Length Number Ovary Per Number Ovary Per (mm.) measured present cent. measured present cent. 0-3-0-5 14 9 66 •46 11 24 0-6 17 10 59 25 10 40 0-7 7 6 86 39 29 74 0-8 1 1 100 12 8 67 0-9-1 -0 — — — 17 16 94 Totals 39 26 172 74 None of our specimens attained the maximum size of 1-20 mm. recorded by several other workers (Fol, Langerhans, Aida) ; but the measurements given above corroborate Lohmann’s remarks as to the variability of this species in regard to the size at sexual maturity. He mentions the occurrence of very small individuals with ripe gonads (0-3-0-45 mm.), in contrast with the usual size which he gives as from 0-55-0-8 mm. (1926, p. 146). This variation, while general, is now seen to have also a regional character, the majority of Atlantic specimens (mainly tropical) becoming fully mature at a distinctly smaller size (about 0-5 mm.) than those in New Zealand waters (0-7 mm.). The difference is doubtless a function of the external temperature, as in the corres- ponding case of Doliolids (Part 1 of this Report, 1933, pp. 206-7). As regards the growth of the tail, Aida gives 4-06 mm. as the length of the tail in his largest specimen, 1-20 mm. in body-length, a ratio of 3-38 : 1. Fol, from living material, assigns a tail-length of 4-2 to a longicauda of the same size, a ratio of 3-5 : 1. Both these ratios are small in comparison with the few measurements I have made of smaller specimens : Body-length Tail-length Ratio 0-45 2-4 1 : 5-3 0-6 3-5 1 : 5-8 0-75 4-3 1 : 5-7 It is possible that at first the tail grows more rapidly in length than in breadth, and in later stages more in breadth than in length (cf. 0. valdivice, p. 272). 2. 0. valdivice, Lohmann (1905) (Text-fig. la-k, 2a-b). — As longicauda dominates the warmer seas, so valdivice was the characteristic Appendicularian of the ice-zone traversed by the “ Terra Nova.” But a problem of some importance is raised by the fact that the German South Polar Expedition, working along the ice-edge in the adjacent quadrant, about 90° E., found not one, but two dominant species, of which 0. gaussica was even more abundant than O. valdivice. At the “ Gauss's ” winter station COPELATA— GARSTANC4 AND GEORGESON. 269 in 66° 2'S., 89° 38' E.. Appendicularians were collected through an entire year (Feb. 1902- Feb. 1903), and 83 per cent, of the Oikopleurids consisted of 0. gaussica, against only 14 per cent, of 0. valdivice. Subsequently, when the ice broke up. and the ship drifted for two months with it (Feb. 9-April 8, 1903), these proportions were reversed (38 per cent, gaussica as against 59 per cent, valdivice) ; but the fact remains that both species were present throughout the year in this quadrant, and neither Dr. Georgeson nor I have been able to find a single indubitable gaussica in the “ Terra Nova ” collections. Stragglers of other species from distant tropical or subtropical latitudes drifted into the Terra Nova’s ” nets, but gaussica, which we may assume was still present in the neighbouring quadrant, was completely absent from surface and deep-water hauls alike, both in the summer of 1910-11 and two years later in 1912-13. The problem is somewhat comparable with that of 0. vanhoffeni and labradoriensis in the North Polar Sea, the latter species having a wider temperature range, while the former is purely Arctic. In the present case, however, both species are strictly Antarctic, and the evidence of any difference in their thermal relations has hitherto been very slight. The highest temperature of the sea at the “ Gauss’s ” winter station was — 1-5°, except in January, when it rose to + 0-7°. This was one of the only three months when both species did not occur together, but it was valdivice, not gaussica, that was missing. During the two months’ drift of the ship, when valdivice exceeded gaussica in numbers, the temperature never exceeded — 1-4°. Similarly in the open sea (“ Westwindtrift ”), clear of the ice-pack, both species fell off greatly in numbers, and only stragglers were obtained north of the Antarctic Circle, valdivice, it is true, a little farther north and more frequently than gaussica — the “ Valdivia ” having in fact first taken the former species north of Kerguelen (52° 47' S.) at a surface temperature of + 2°. On the other hand the “ Gauss ” took gaussica on one occasion as far north as 58° 26' S. at a surface temperature of + 1-8°. These facts certainly indicate that, while both species are essentially ice-lovers, valdivice can endure a slight elevation of temperature more easily than gaussica. The entrance to the Ross Sea between 170° E. and 170° W., where the “ Terra Nova’s ” valdivice stations were clustered, notoriously tends to be freed of ice sooner and longer than any other part of the Antarctic circle. Indeed it was this knowledge which determined the vessel’s attempted line through the pack on her outward voyage. It indicates presumably a stronger influence from tropical inflows. However slight the difference between the two species in their relations to temperature, this feature of the Ross passage may be assumed to have favoured valdivice, as compared with gaussica, in collections concentrated in those waters ; but the complete absence of gaussica in both years appears to indicate that here at any rate there must be a continual inflow from the Pacific, intersecting the Southern Ocean with a constant barrier of relatively warm salt water that never gives gaussica a chance to establish itself. Further observa- tions on the distribution of the two species and their relations to physical factors are obviously desirable. lb -270 “ TERRA NOVA ” EXPEDITION. Lohmann gives the maximum body-length of 0. valdivice as 3-50 mm. This size was exceeded by an appreciable number of the “ Terra Nova ” individuals, the largest of which measured 4-1 mm. The following table gives an indication of size-distribution Fig. 1 .—Oikopleura valdivice, Lohm. a, dorsal, b, left, and c, right aspects of an immature individual (1-3 mm. body-length) from which the dorsal oikothelium has been removed. d-k, successive stages in the growth of the gonads, li shows the testis only ; k shows both glands in situ, the ovary still as a thin triangular disk (ventral view). in the whole collection, those collected in 1910 showing the same modal (2 mm.) and submodal (1 mm.) lengths as the larger samples of 1912. In addition the numbers are recorded of specimens with developed ovaries (Miss Georgeson’s dissections and measurements) : Body-length (mm.) Number measured Number with ripening ovaries Per cent. <1 87 — 0-0 1 263 7 2-6 2 428 22 5-1 3 231 17 7-4 4 6 6 100-0 Totals . 1015 52 These figures contrast strikingly with those for the preceding species, and suggest that in 0. valdivice the attainment of female maturity is a slow process with a marked seasonal periodicity. It is obvious that at the period when these samples were collected COPELATA— GARSTANG AND GEORGESON. 271 (i.e., the Antarctic midsummer) the great bulk of the population was still immature, while the sudden increase in ovarian maturity that accompanies the last increment of size suggests that a further period of growth would be necessary before the bulk of the population would be ready for breeding. This means that the breeding season for 0. valdivice must fall in the latter half of the Antarctic summer, and not in the spring as is usual with marine animals in temperate latitudes. These inferences accord well with the elaborate quantitative investigations of the “ Gauss '' at its winter station in 1902-3, which revealed a sudden increase and maximum density of small pelagic organisms (Tintinnids, Radiolaria, Fritillariid Appendicularians) in March, but of Oikopleurids (i.e., gaussica and valdivice) not until April, i.e., three months later than the period of the “ Terra Nova's ” collections. A slight initial increase beginning in January may well be correlated with the reproduction of the more precocious individuals of our table (cf. Lohmann u. Buckmann, l.c. pp. 216, 217, fig. 54). This late flourishing of Appendicularians in polar waters appears to be equally marked in the Arctic region, though the evidence is more fragmentary. Lohmann and Biick- mann (l.c. p. 216) cite figures for the west coast of Greenland from countings by Van- hofi’en, from which it appears that a climactic period is shown there by Fritillaria in October and by Oikopleura in November, the rise from the low winter level beginning in each case 1-2 months earlier. In the Spitzbergen region the climax is probably earlier. The “ Nautilus collected a number of Appendicularians there between Aug. 30 and Sept. 4, 1931, which came into our hands for examination. All belonged to the species 0. vanhojfeni, which is said to attain a maximum body-length of 7 mm. These specimens, which ranged from 0*5 to 2-5 mm. in length, were completely immature, neither testis nor ovary being recognisable. They would accordingly seem to be the offspring of a breeding season a month or two earlier, i.e., July or July-August, though the data are too isolated for a safe conclusion. In any case, so far as Appendicularians are con- cerned, it seems clear that the first half of the short Arctic or Antarctic summer is required for completing the growth to maturity of a scanty winter population, and that reproduction is mainly by a short breeding season that falls in one of the later summer months, typically September or October in the Arctic region, February or March in the Antarctic. Lohmann has briefly characterised the development of the gonads from a flat disk, “ with central separation of the triangular ovarian part ” (1905, p. 360). The earliest noticeable rudiment, at 1-0-1 -5 mm. body-length, has the form of a slender isosceles triangle, with truncated apex directed upwards and backwards, which is divided into three nearly equal parts, testes externally and ovary in the middle (text-fig. la-d). The lateral testes, however, are connected basally by a transverse connecting piece (“ isthmus ”), so that the testis as a whole forms a slender frame round the central triangular ovary, and is interrupted only at the truncated apex, where the tip of the ovary separates the tips of the two testicular limbs. 272 “ TERRA NOVA ” EXPEDITION. At about 2 mm. body-length the rudiment has broadened basally into an equilateral triangle, stilt showing the same general arrangement (text-fig. le). At about 2-5 (2-0-3-0) mm. the apical angle has become obtuse, and the testicular base of the triangle is produced into a pair of short wings which apply themselves to the sides of the intestinal knot (text-fig. If). The connecting isthmus tends to curve outwards below and behind the base of the ovary (text-fig. lh). The whole testicular rim thickens, at first uniformly, then with increasing effect on its postero-lateral wings. All this time the ovary remains as a thin semi-transparent triangular plate within the testicular frame. At about 3*0 mm. the lateral parts of the testis broaden and thicken underneath the ovarian triangle, while the isthmus remains slender (text-fig. lg). The ovary is still a thin triangular plate, slightly produced at its two outer corners or nearly crescentic, and covering the median gap between the lateral lobes of the testis, which it partly overlaps on either side. At 3-5 mm. the postero-lateral wings of the testis have extended forwards so as to embrace the sides of the gut-knot almost completely (text-fig. lh). The ovary begins to thicken in the middle and penetrate inwards between the two lobes of the testis, but preserves its primitive relations externally. At 4-0 mm. the ovary is itself a solid thick, more or less lobulate organ wedged in between the testicular lobes, and more or less completely enveloped by them, so that it can only be clearly seen by dissection. When massively developed (3-0-4-0 mm.) the apex of the gonadial mass protrudes dorsally to a slight extent behind the oikothelial plate. It is morphologically interesting to note that in quite small specimens (1-2 mm.), provided they are well-preserved and uncontracted, this large genital hump, with its short, blunt, apical horn, can be seen to be preformed, while the gonadial rudiment is a mere triangular shield lodged within its otherwise empty blastoccelic cavity (text-fig. lb, c). In most cases, however, the thin epithelium clothing the genital hump has collapsed inwards except where it overlies the rudimentary gonadial plate, so that the apex of this shield ventrally and that of the stiff oikothelium dorsally project behind the body like a pair of opposed valves, giving a characteristic bilabiate appearance to most of the preserved specimens. The relations of body-length to tail-length are approximately as follows, but subject •to variation : Tail-length (mm.) .5 7 9 11 14 16 17 Body-length (mm.) . 1-2 1-5 2-0 2-5 3-0 3-5 4-0 When the tail has reached a length of 14-15 mm. it appears to undergo a relatively greater increase in breadth than in length, though I have made no measurements to confirm this point. The subchordal cells are clear and distinctive in well-preserved specimens, usually 7 or 8 (6-10) in number. COPELATA— GARSIANG AND GEOBG-ESON. 273 The para-pharyngeal glands (“ oral-glands ") are very large and oblong in shape in most of our specimens, whether large or small. They are usually as long as the endostyle, but begin in front of it and terminate about the same distance from its hinder apex (text-fig. la). Detached rudiments of “ houses '* (unexpanded of course) were found in many of the samples. They are usually coated inside and out with particles of planktonic debris which obscure their structure, and two points, one on either side of the centro- dorsal area, are peculiarly adhesive, so that many of the houses exhibit a pair of irregular fibrous tufts at these points. As Lohmann has described and figured a pair of gelatinous “ beaker-shaped ” papillae near the same points in the house-rudiment of 0. albicans (1896. Taf. XII, figs. 8. 9), it was desirable to investigate these adhesive points, which seemed in many cases to be directly produced into two or three fine filaments. The houses are readily cleaned by shaking them violently in a tube of hot water, and are then very interesting objects, showing the structural elements of the food-trap and the crossed fibres of the filtering windows very beautifully. The house-rudiment (text-fig. 2a, b) as a whole appears to be made of much firmer material than that of 0. albicans. It presents a very definite shape, resembling that of a low Prussian helmet, and is stiffened regularly by special gelatinous bands. Referring for descriptive purposes to Lohmann and Buckmann's nomenclature of the oikothelium (l.c. figs. 4, 9, 13), the most conspicuous feature is the pair of elevated oval gelatinous patches in front which cover the fibres and pipes of the food-trap (c', c'"). Below these the suboral region protrudes in front of the morphological mouth ( = the ultimate exhalant aperture) like a visor in front of a mediaeval helmet, and is supported by two half-lioops of gelatinous material, one above the other, of which the upper (morphologically anterior) is much the stronger, and is yellowish in colour. These bands appear to be the product of the horizontal rows of cells forming the front part of L. and B.'s “ Kehlge- gend, ' and may be appropriately distinguished as the upper and lower suboral bands. The filtering windows, corresponding to Eisen's oikoplasts (d), lie near the lower concave edge of the middle of the helmet on each side ; and from their hinder margin there runs backward on each side a strong gelatinous bar, the pair of which meet and fuse with one another in the median or sagittal plane, together forming a single postero-lateral half -hoop between the two windows. This seems to be the product laterally of “ Ihle’s Field " (1), and dorsally of the middle row of large cells of area (i) in front of the “ posterior rosette ” (n). Behind this postero-lateral half-hoop in the median line a transversely oval gelatinous papilla is often recognisable, forming the centre of a weaker crescent of festooned elements wedged in between the stronger parallel bands. Behind this again, and forming the main part of the posterior “ neb ” of the helmet, is a bilobed terminal plate of thick jelly, with tapering lateral wings, apparently divided in the middle line by a longitudinal streak, which is really a deep-lying fibre. On either side of this streak is a curious double scroll, which forms a kind of fleur-de-lys pattern, plainly formed by the “ posterior 27-1 “ TERRA NOVA EXPEDITION. rosette ” (n). The parallel rows of elongated marginal cells (m) give rise to the thin edge of the posterior neb of the helmet. Between Fol s oval (c) and Eisen s oikoplasts (d) there runs an oblique tow (g) of cells on each side of the oikothelium. to which another pair of oblique lateral bands of jelly correspond in the house. These are united by a kind of gelatinous buckle in the mid-dorsal line, the product apparently of the front row of large cells of area (1) imme- diately behind the anterior rosette (h). This itself gives rise to a sheet of clear jelly Fig. 2. — Oilcopleura valdivice, Lohm. a, dorsal, and b, left side views of two house-rudiments, detached from the oikothelium and cleaned to show the principal gelatinous and membranous elements. ALP. AP . ASB. CP . DB . FC . F.O . FW . L.Hk M . OB . PLB PMP PO . PS . PSB . Antero-lateral Papilla. . Adhesive Patch (in a covered by a quadrangular scale, in b produced into fibrils). . Anterior Suboral Bar (yellowish). . Central Papilla, divided internally by a median thread or septum (from " Anterior Rosette"). . Dorsal Buckle (clear stiff jelly), connecting the two Oblique Bars. . Festooned Crescent. . Fobs Oval, containing the series of filter-pipes. . Filtering Windows, with crossed fibrils. . Lohmann’s “ II aiilu ngsJcorper,’ ’ the principal group of them. . Mouth (future exhalant aperture and valve). (In b its position is indicated, but it is not actually visible from the side.) . Oblique Bar. . Postero-lateral Bar (from Ihle’s Field ”). . Postero-median Papilla. . Prse-oval. . Posterior Scroll (from “ Posterior Rosette "). The median partition is a deep-lying fibre. . Posterior Suboral Bar. at the apex of the helmet, the centre of which is marked by a characteristically bisected oval area. The pair of adhesive 'patches mentioned earlier lies on either side of this centre of the anterior rosette, from which they are separated only by a narrow tract of clear jelly on either side. They seem accordingly to be the product of the large cell or cells which form the apex of “ Martini’s Field ” (f). The patches themselves are sometimes produced into 1-3 slender adhesive fibrils, sometimes mere amorphous deposits on the outer edge of a curious valve-like scale, sometimes recognisable only by their adhesiveness. If reference now be made to L. and B.s figures of the oikothelium of 0. albicans (l.c. COPELATA — GARSTANG AND GEORGESON. ‘275 figs. 12. 13) and to Lofimann's earlier figures of the house-rudiment (1896, pi. xii, figs. 5. 7, 8), it will be seen that the gelatinous “ beaker-shaped " papillae are probably, as these authors suggest (p. 82), the product of the giant-cells named “ Leuckart’s Field " (Fig. 13, o), a group which is supposed to be absent from the oikothelium of all other species. As we know nothing of the function of these curious papillae, or of the adhesive patches of raid ivies, it would be unprofitable to go minutely into the question of their possible correspondence, but it may be pointed out that the mantle " of small cells, which isolates Martini's field in gaussica and raid ivies, is absent in vanhojfeni and albicans, so that a general homology between the specialised Leuckartian cells of albicans and the apical cells of Martini s field in other species is not impossible. 0. cophocerca. which in other respects is an almost direct connecting link between albicans and cold ivies, shows an intermediate condition also in this area of the oikothelium (L. and B., l.c., fig. 14). In it there is a peculiarly enlarged cell dorsolaterally which certainly corresponds to the apical Martini-cell in gaussica and valdivice, as the incipient " mantle " above it shows : but this mantle is so rudimentary and Martini's field in general so diffuse that the large cell may equally well be claimed through vanhoffeni as a homologue of one or more of the Leuckartian cells of albicans. A comparative study of house-rudiments in connection with the pattern of the oikothelium is clearly indicated as the next desirable step in Appendicularian morphology. 3. 0. cophocerca, Gegenbaur (1855, not Fob 1872). — This is a purely warm-water form. We have occasionally suspected its occurrence in Antarctic samples, but the specimens have mostly resolved themselves into young valdivice which closely approach it in various respects. 4. 0. rufescens, Fol (1872). — This is another warm- water species and its occurrence in typical form at stations 52 and 111 calls for no comment. But the few Antarctic specimens doubtfully referred to it have peculiarities which with fuller material might justify the definition of a new species intermediate between 0. rufescens and dioica. The left stomach is ovate in outline as in rufescens, but horizontally placed and unusually shallow, so that the full extent of the rectum is visible below it. The oesophagus enters it postero-dorsally as in dioica. The right stomach is even shallower than the left, and provided with a distinct ventral csecum as in rufescens which overhangs the outer wall of the rectum. The latter runs horizontally, projects slightly in front of the stomach, and terminates by a short anal papilla on its ventral side, not, as usual, by its extremity. Ihle also refers to specimens from the Indian Archipelago with a horizontal rectum (1908, p. 14). Unfortunately the tails of our specimens are missing, so that the evidence of their subchordal cells is unobtainable. It will be seen that the reference of these specimens to rufescens is at most an approximation, and cannot be regarded as extending the range of typical members of the species. 5. 0. dioica, Fol (1872). — This commonest species of many coastal waters is capable 276 “TERRA NOVA” EXPEDITION of enduring a wide range of temperature and salinity. In the Atlantic it is recorded from the South Equatorial stream to the North Sea, but avoids the cold Labrador and Benguela currents. It seems to be relatively scarce in the Indian Ocean (Zanzibar- Kanal, Lohmann), but occurs on both sides of the Pacific. The Atlantic specimens at station 52, though small, were typical, and showed the two subchordal cells without fail. The two specimens from New Zealand waters (station 107) were unfortunately tailless. 6. Folia gigas, n. sp. (text-fig. 3a-e). — It is probably significant that our single specimen of this remarkable form was taken in a net towed (nominally, at any rate) at a depth of 80 metres. Lohmann is emphatic that Folia gracilis is essentially a deep-sea species of the tropical regions, and it would be difficult to account on any other theory for the occurrence of as many as thirteen individuals in a single haul of a closing-net worked from 390 to 190 metres in the Guinea current. The “ Gauss ” took only a single specimen in the Antarctic, and that by a vertical haul of 0-2000 metres during the ship's drift in the ice. In several respects the “ Terra Nova ” Folia from station 274 is unique. It is appre- ciably larger than any of the specimens hitherto recorded (not more than twenty alto- gether). The delicate pharynx is crushed backwards upon the rectum, and so twisted that the oral region with endostyle and oral glands protrudes sideways from this region - — practically at right angles. The maximum length of previous specimens is given by Lohmann and Biickmann as 0-484 mm. This new specimen from anus to the hinder end of the stomach alone measures 1-4 mm. If this region be assumed to bear the same proportion to the total body-length as in Lohmann’s first figure (1896, Taf. XIX, fig. 1), the body-length of the “ Terra Nova ” specimen cannot have been less than 3-25 mm. Oikothelium, body-wall, gonads and tail are completely absent from this specimen, except a small piece of ovary which encrusts the posterior apex of the stomach. The complete genital glands must have filled the blastocoelic space behind the gut as in Lohmann’s fig. 2 (l.c.), since they are known to develop from a tripartite postero- ventral discoidal rudiment, as in O. valdivice. It follows that the empty blastocoelic cavity with its parietal genital shield, often figured to represent the typical form of Folia , Stegosoma, etc., is not a special adaptation for flotation, as several times suggested, but an example of the preformation in youth of an arrangement required in maturity, exactly as in 0. valdivice and its relatives (p. 272). The general conformation of the gut closely corresponds -with Lohmann's account, but the larger size of this individual permits a more exact tracing of the morphological limits, which show that the whole gut is much more normally constituted than appears from previous figures. The stomach is not a simple pyriform sac, but an asymmetrical ovoid bag set obliquely across the plane of the oesophagus and compounded of a posterior left and an anterior right lobe imperfectly straightened out. The latter gives rise in front to a true “ post-stomach ” separated by a groove ventrally, from which the short COPELATA— GARSTANG AND GEORGESON. 277 intestine arises on the right side to open into the rectum obliquely on its right side ventrally. In fact, if we take as a starting point the gut of a fusiformis with its anteriorly produced rectum, all that is necessary to produce the gut of a Folia is, as it were, to pull the rectum still further forward, thus dragging the intestine and post-stomach to the front of the right lobe of the stomach, and straightening the relations of the right to the left lobe of the stomach at the same time. Fig. 3. — Folia gigas, n. sp. — Five successive views of the single specimen, as seen by rotating it in glycerine under a binocular dissecting microscope. a, left side, showing crumpled oesophagus, with its cardiac end imbedded in a fragment of ovary, and the pointed post-cardial gastric caecum. b, dorsal view, showing the same features, the existence of a left lobe of the stomach, and the pair of large oral glands on one of which the endostyle lies. c, dorso-dextral view, showing the pair of peripharyngeal bands entering the oesophagus separately, and the small spiracular rings. d, ventro-dextral view, to show the cardiac loop of the oesophagus and the distinct nature of the right lobe of the stomach. e, ventro-sinistral view’, emphasizing the last point and showing the differentiation of a distinct “ post-stomach ” between right stomach and intestine. Note.— After figures b and e were drawn, the ovarian tissue was partly cleared away with needles to expose the cardia. This was definitely located (a, d), but the underlying tissue of the left gastric wall and caecum was injured in the process (c). Further differences from previous descriptions are to be seen in the oral glands and peripharyngeal bands. The former, instead of being klein or verhaltnissmassig Mein, are large, oblong, and conspicuous ; although the endostyle, which seems injured, is too obscure to enable us to compare the relative size of the two organs. The peri- pharyngeal bands, which are represented as meeting beneath the brain in Lohmann’s fig. 1, or half-way towards the oesophagus in his fig. 2, must in this specimen have been 278 “ TERRA NOVA ” EXPEDITION. exceedingly oblique, since they can be traced as independent bands to the very mouth of the oesophagus over the front edge of the stomach. Some of these differences may be due to imperfect observation, others to varying rates of growth, but the constant character of the oral glands at all stages of growth in 0. valdivice suggests that there may be specific differences in this respect. If there is only one species the fact that Lohmann' s specimen with massively developed testes (fig. 2) was not more than one third the size of the present one would indicate a difference in the rate of maturation in a presumably deep-sea species hardly consistent with the We therefore create a new species for the “ Terra Nova ” individual, distinguishing Folia gigas by its large oblong oral glands, its large size (3 mm. with mature ovary), and the obliquity of its peripharyngeal bands which are independent structures from endostyle to oesophagus. As the “ Gauss's ” specimen of F. gracilis had probably drifted into the Antarctic from the Indian tropics, so the original home of the “ Terra Nova's " F. gigas was probably the tropical Pacific. 7. Stegosoma conogaster, n. sp. (text-fig. 4). (Stegosoma magnum, Aida, 1907, p. 21, not Oikopleura [ Stegosoma ] magna Langerhans (1880). — The two specimens of Stegosoma found by the “ Terra Nova ” were fished in surface nets, the first during the winter cruise of 1911 off the Three Kings Islands, north-west of New Zealand (station 76, July 7th), the second in the South Atlantic on the homeward voyage in 1913 (station 310, April 21st). These specimens retain their tails, though in poor condition, but the delicate pharyngeal region is severely crumpled in each case. In the form of the alimentary canal both specimens agree closely with Aida's account of Japanese specimens (1907), the pouch- like left lobe of the stomach having the form of a bilaterally compressed oval, broad anteriorly and tapering conically behind to a bluntly pointed posterior apex. The narrow pseudo-cardiac stalk, which connects the left with the right lobe and receives the oesophagus, opens into the pouch in the middle of its dorsal edge. This form of stomach differs appreciably from the somewhat quadrangular wallet-shape figured by Chun (1888), and Lohmann for North Atlantic and Mediterranean individuals.* The tapering posterior cone is sufficiently well-marked to deserve the title of a caecum, whereas in North Atlantic forms the gastric stalk enters the pouch behind the middle (Chun, l.c., fig. 1 ; Lohmann, 1896, fig. 2), and the posterior wall of the pouch is rounded much more obtusely than in the Japanese and our specimens. In this respect the Japanese and “ Terra Nova ” specimens occupy an intermediate position between typical magnum relative uniformity of the conditions. Fig. 4. — Stegosoma conogaster, n. sp. Left side of intestinal loop. Slightly diagrammatic above, to show the slender intergastric connection, which in both specimens is hidden behind the upper edge of the left stomach owing to con- traction. * See postscript, page 281 . COP EL AT A — G A R S TAN G AND GEORGESON. 279 and the remarkable form obtained by the “ V aldivia ” in the Indian Ocean, to which Lohmann has given separate generic rank as Cliunopleura microgaster (Lohmann, 1914). In it the gastric stalk opens into the front end of the pouch and the blunt caecum of our specimens is further produced and attenuated. On these grounds we have decided to create a new species for the Japanese type of Stegosoma and to refer our specimens to that species. The chief distinctive point is the production of the posterior wall of the gastric pouch into a bluntly conical caecum. The facts indeed suggest that S. conogaster may prove to be the typical Pacific species, of which the two Terra Nova " individuals were outliers, the first, from New Zealand waters (station 76, July 1911), needing no special comment, the second (station 310, April 1913) having accidentally rounded the Horn and been caught in the north-going Falk- land current, like the single individual of Pelagopleura australis which the second German Antarctic Expedition fished in the South Brazil stream on Aug. 23, 1911 (Lohmann u. Biickmann, l.c., p. 156). Both the ” Terra Nova ” specimens are without any trace of gonads, the larger of the two probably from loss, the smaller (which retains its epithelium) from immaturity. Only the visceral region of the body admits of measurement in either case. The length of this region in the first and larger specimen (station 76), from front of oesophagus (the rectum being broken) to hindmost point of the post-stomach, is 1*3 mm., yielding an estimated total body-length of 2-9 mm. ; tail, 6 mm. That of the smaller specimen (station 310), from anus to the same point behind, is 0-65 mm., yielding a total estimated body-length of 1-2 mm. ; tail, 3 mm. Aida’s largest specimen from Japanese waters had a tail of 13 mm., with an estimated body-length of 3-8 mm. The Stegosomas are now recognised as essentially epiplanktonic and tropical, German expeditions having taken as many as 250 S. magnum in one haul in the Sargasso Sea (•; National, " 1889), and 213 in the Indian Ocean, south of Reunion (“ Gauss,” 1903). The single specimen taken by the “ Gauss ” during its drift in the ice-pack at 64° 6' S. is regarded by Lohmann as a “ tropical guest,” carried to the Antarctic by underflow from the Indian Ocean. 8. Pelagopleura magna, Lohmann (1926) (text-fig. 5 a, b). — The three specimens we refer to this species were obtained at two adjacent stations in the Antarctic (275 and 276) within two days of one another (Jan. 3rd and Jan. 5th, 1913). They all possess the enormously elongated spiracular rings by which Lohmann and Biickmann distinguish this species (1926, p. 153). The first two at station 275 are large-bodied, but difficult to measure owing to contortion of the pharyngeal region. One (A), with a tail of 16 mm., is almost or quite 4-0 mm. long, the other (B), which is tailless, actually measures 3-8 mm. over-all, but it is more mutilated, has lost its epithelium behind, and must have been larger than A when alive. The former (A) contains streaks of degenerate tissue in its expanded posterior cavity, which suggest that it had discharged its genital products before capture, the epithelium being ruptured along the mid-ventral line. Its tail, though in poor condition, exhibits “ amphichordal ” cells as small dark granules 280 “TERRA NOVA” EXPEDITION. dispersed irregularly and thinly along each side, with a tendency to form separate groups posteriorly. B exhibits a long endostyle with many pairs of large outer glandular cells, and an internal paired series of smaller cells. It is almost as long as the rectum (actually about 0-8 mm.) and nearly as wide. The stomach and other parts of the gut in both specimens are much wrinkled. The rectum in both contains faeces. The third specimen (C), taken at station 276, is much more perfect, but smaller (body-length, 2*0 mm. ; tail, 6-0 mm.). Mouth, ciliated funnel, peripharyngeal bands, and spiracular rings are much less crumpled, and can be traced out almost completely ; the brain is obscure, but the spherical organ on the left side (text-figs. 5a and 5b) may well be the otocyst or otolith ; the whole digestive tract is clear at a glance. The stomach is fully expanded as a perfectly smooth spherical sac slightly compressed from side to side. The gut is completely empty except the rectum, which, as in A and B, is filled by a loose pellet of faeces which the gut of an ordinary Oikofleura is differentiated can be identified, as in Folia , though here the right and left lobes of the stomach are completely unified, and the “ pulling forward ” of the right limb of the intestinal knot has been carried out more completely. Gonads are present in rudi- mentary form as a flat dumb-bell-shaped disk across the postero-ventral side of the stomach, as figured by Lohmann (L. and B., l.c. Abb. 43). Unfortunately the tail of this specimen has completely lost its epithelial fins, so that the blastoccele gapes widely, ex- posing the notochord, and all amphichordal cells have been washed away. In this specimen, also in B but less clearly, the external apertures of the spiracles are recognisable as broadly oval apertures on each side of the anus and slightly in front of it, having about the same diameter as the rectum. The ciliated spiracular ring, with its much greater diameter, is seen on each side to be separated from the external aperture by a distinct interval, representing the delicate wall of the outer part of the gill-pouch. The endostyle is relatively as large as in B and again has nearly the same length and width as the rectum. The peripharyngeal bands leave the endostyle a short distance behind the front end and run obliquely backwards and upwards as independent structures until they are lost in the thickened wall of the oesophagus (cf. Folia above). The German South Polar Expedition took this species on a single occasion during the drift of the £; Gauss ” in 63° 43' S., but the subsequent Antarctic Expedition of 1911-12 almost ready for discharge. All the parts into Eig. 5. — Pelagopleura magna, Lohm. a, left-side view of the smallest specimen. The right peripharyngeal band and spiracular ring are omitted to avoid confusion. The former enters the oesophagus independently of the left band. Large flame-like cilia on the spiracular ring can be seen through the gaping external orifice of the gill-slit. b, right-side view of the ciliated funnel and apparent otocyst on a larger scale. Brain and nerve-cord are not clearly distinguishable. COPELATA — GARSTANG AND GEORGESOX. 281 took it at four different places in the Weddell Sea. The “ Terra Nova's " specimens clearly support Lohmann's view that this species, like 0. valdivice, is truly and exclusively Antarctic. It is noteworthy therefore that the “ Terra Nova " failed to secure the other Antarctic species which Biickmann has distinguished as P. australis, and which the two German expeditions obtained on several occasions in the same regions as P. magna. The problem of 0. gaussica is thus reinforced. 9. A remarkable specimen was taken at station 275 along with the two Pelagopleuras just described and a number of 0. valdivice. Having a tail of about the same size and a flabby mutilated body not readily characterised, it was at first set aside as another Pelagopleura. The tail seemed to confirm this idea, for although of a type hitherto undescribed, it possessed twTo striking rows of amphichordal cells running the wfflole length of the tail, each row consisting of a continuous series of densely crowded small dark cells, wTiich together confer upon it a bilineate appearance. These amphichordal cells are only known in twro genera, Pelagopleura and the closely allied Althojjia , but no species of either genus has been described with continuous rows of them, though, as stated above, one of the “ Terra Nova " specimens shows an approach to this condition. The body is torn to shreds and lacks any continuity between its parts, but the pharyngeal remnants include a pair of wTell-developed oblong and lobulated oral glands with a short oval endostyle (apparently broken) between them. As oral glands are absent from all these abnormal genera except Stegosoma and Folia, which possess subchordal, not amphichordal cells in their tails, we are compelled to report the specimen without any name at all. The combination of characters, few as they are, is unique and baffling. The organs present in the mutilated body, apart from those mentioned, are (1) a short arched oesophagus connected with the pharyngeal remnants at one end but torn from its connection with the stomach which is missing, except for a few shreds ; (2) fragments apparently of the right side of an intestinal knot ; and (3) an anal papilla, firmly fixed in the thin tenacious remnants of the post-pharyngeal epithelium, but torn away from the rest of the rectum which, if present, is merely part of the mutilated tissues representing the right side of the gut already mentioned. The epithelial fins of the tail have almost entirely disappeared, so that the blastocoele gapes widely. It may be important to note that the amphichordal cells are here supported by a definitely gelatinous matrix, but for which they would almost certainly have disappeared as in one of our Pelagopleuras. Postscript. During a recent visit to the Biological Station at Bermuda I have examined various specimens of Stegosoma that came up in the townettings. All conform to the type of S. magnum, and show no approach to the conical type of stomach of S. conogaster, thus confirming the distinctness of the Pacific species. An account of the Appendicularians of Bermuda, some of which are large forms of great interest, is in course of preparation. W. G. 30.iii.35. 282 An . C . CF . End G . GS . Ht . Int . L . M . MS . Oes OG . Ot . Ov . Pap . Ph . Pph . P.St. R. . SO . Sp.R. St. . T . “TERRA NOVA” EXPEDITION. KEY TO ABBREVIATIONS (except fig. 2). . Anus. . Caecum. . Ciliated Funnel. . Endostyle. . Gonads. . Gill Slit (external aperture). . Heart (?). . Intestine. . Left. . Mouth. . Median Seam of large cells. . (Esophagus (with C = cardiac end). . Oral gland. . Otolith (?). . Ovary. . Papilla. . Pharynx. . Peripharyngeal band. . Post-stomach. . Rectum (or Right). . Suboral Oikothelium. . Spiracular Ring of Cilia. . Stomach. . Testis. . Torn edge of body-wall (omitted for clearness in fig. lc). LITERATURE. Aida, T. 1907. Appendicularia of Japanese Waters. J. Coll. Sci. Tokyo, XXIII. 5. pp. 1-25 : pis. 1-4. Chun, C. 1903. Aus den Tiefen des Weltmeeres. Ed. 2. Jena. pp. 592 : illust. Fol, H. 1872. Etudes sur les Appendiculaires du detroit de Messine. Mem. Soc. Phys. Geneve, XXI (2). pp. 445-499 : 11 pis., text illust. Ihle, J. E. W. 1908. Die Appendicularien der Siboga-Expedition. Siboga Exped., LVI, c. pp. 123 : 4 pis., text illust. Lohmann, H. 1896. Die Appendicularien der Plankton-Expedition. Ergebn. der Plankton-Exped. II, E, c. Kiel u. Leipzig, pp. 148 : 24 pis. 1905. Die Appendicularien des arktischen u. antarktischen Gebiets, ihre Beziehungen zueinander u. zu den Arten des Gebeits der warmen Strome. Zool. Jahrb., Jena. Supplement, VIII. pp. 353-382 : pis. 11-12. 1914. Die Appendicularien der Valdivia-Expedition. Verh. Dtsch. Zool. Ges. XXIV. pp. 157- 192 : text illust. Lohmann, H. u. Buckmann, A. 1926. Die Appendicularien. Dtsch. Siidpolar Exped. 1901-3. XVIII. pp. 63-231 : 2 tab., text illust. PRINTED m GREAT BRITAIN BY WILLIAM CLOWES AND SONS, LIMITED, LONDON AND BECCLES. ■ % .V- V»\|o /?^3i JV- . V • ' - ■ • : ■ '■ ' v : . ,V -?• . '• + • . : 'i •*!,’ ■' ■ * . ' ' - ' • . . .• ;• ' . • . .**•*' ' • * ' ■ T J.'l. >,( ' ■ - i :• .'v '• : • . I . v ' • , ■ Vi"'. 'V' ,'A •' ' ' t ' ■' . • . • / . ... jj • . v . /• • 1 '• • • ; 1 i . • •• i* ,'.c- ; ...... vwS«8 •: X‘\wsfsB ■ . H '& ■